Hung S. Ho, MD, FACS

Associate Professor, Division of Gastrointestinal and Laparoscopic Surgery, Department of Surgery
University of California, Davis, School of Medicine
Medical Director and Attending Staff, Section of Gastrointestinal and Laparoscopic Surgery, Department of Surgery
University of California, Davis, Medical Center

First depicted in anatomic drawings in 1492 by Leonardo da Vinci, the vermiform appendix was described as an anatomic structure in 1521 by Jacopo Berengari da Carpi, a professor of human anatomy at Bologna. Appendicitis became recognized as a surgical disease when the Harvard University pathologist Reginald Heber Fitz read his analysis of 257 cases of perforating inflammation of the appendix and 209 cases of typhlitis or perityphlitis at the 1886 meeting of the Association of American Physicians. In this landmark report, Fitz correctly pointed out that the frequent abscesses in the right iliac fossa were often due to perforation of the vermiform appendix, and he referred to the condition as appendicitis.1 Among his classic observations of the disease was his emphasis on the ‘vital importance of early recognition’ and its ‘eventual treatment by laparotomy.’ It was not until 1894 that Charles McBurney first described the surgical incision that bears his name and the technique of appendectomy that was to become the gold standard for appendectomy throughout the 20th century and into the 21st.2

Figure 1. Treatment options in suspected appendicitis

Although appendectomy has traditionally been done—and largely continues to be done—as an open procedure, there has been increasing interest in laparoscopic appendectomy since the beginning of the 1990s. At present, however, the only patients for whom laparoscopic appendectomy appears to offer significant advantages are women of childbearing age, obese patients, and patients with an unclear diagnosis [see Figure 1]. Accordingly, the gold standard for surgical treatment of acute appendicitis remains open appendectomy as described by McBurney. The occasional patient with chronic appendicitis should be electively treated with the laparoscopic approach.

Operative Technique

Open Appendectomy

Figure 2. Open appendectomy: McBurney’s Point

With the patient in the supine position, general anesthesia is induced and the abdomen is prepared and draped in a sterile fashion so as to expose the right lower quadrant. The skin incision is made in an oblique direction, crossing a line drawn between the anterior superior iliac spine and the umbilicus at nearly a right angle at a point about 2 to 3 cm from the iliac spine. This point, McBurney’s point, is approximately one third of the way from the iliac spine to the umbilicus [see Figure 2]. The subcutaneous fat and fascia are incised to expose the external oblique aponeurosis. A slightly shorter incision is made in this aponeurosis; first, a scalpel is used, and then, the incision is extended with scissors in the direction of the fibers of the muscle and its tendon in such a way that the fibers are separated but not cut.

Figure 3. Open appendectomy: exposure of the abdominal cavity

The fibers of the internal oblique muscle and the transversus abdominis are separated with a blunt instrument at nearly a right angle to the incision on the external oblique aponeurosis. The parietal peritoneum is lifted up, with care taken not to include the underlying viscera, and is opened in a transverse fashion with a scalpel. This incision is then enlarged transversely with scissors. When greater exposure is required, the lateral edge of the rectus sheath is incised and the rectus abdominis retracted medially without being divided [see Figure 3].

Figure 4. Open appendectomy: mobilization of the appendix

A foul smell or the presence of pus on entry into the peritoneum is an indication of advanced or perforating appendicitis. The free peritoneal fluid is collected for bacteriologic analysis. The appendix is located by following the cecal taeniae distally. The inflamed appendix usually feels firm and turgid. The appendix, together with the cecum, is delivered into the surgical incision and held with a Babcock tissue forceps. If this step proves difficult, the appendix can sometimes be swept into the field with the surgeon’s right index finger as gentle traction is maintained on the cecum with a small, moist gauze pad held in the left hand [see Figure 4]. Care should be taken at this point not to avulse the friable and possibly necrotic appendix. To deliver a retrocecal appendix, it may be necessary to mobilize the ascending colon partially by dividing the peritoneum on its lateral side, starting from the terminal ileum and proceeding toward the hepatic flexure.

Figure 5. Open appendectomy: isolation of appendicular artery

The mesoappendix, containing the appendicular artery, is divided between clamps and ligated with 3-0 absorbable sutures [see Figure 5]. The appendix is held up with a Babcock tissue forceps, and its base is crushed with a straight mosquito arterial forceps. The mosquito forceps is then opened, moved up the appendix, and closed again. The base of the appendix is doubly ligated with 2-0 absorbable sutures at the point where it was crushed, so that a cuff of about 3 mm is left between the forceps and the tie.

The appendix is divided by running a scalpel along the underside of the forceps. The mucosa of the appendiceal stump is fulgurated with the electrocautery. The stump is not routinely invaginated into the cecum. In those rare cases in which the viability of the appendiceal base is in question, a 2-0 absorbable purse-string suture is placed in the cecum, and the stump is invaginated as the suture is tied; if this is done, palpation for a patent ileocecal valve is indicated. The operative field is then checked for hemostasis. In cases of perforating appendicitis, the right paracolic gutter and pelvis are irrigated and thoroughly aspirated to ensure that any collected pus or particulate material is removed.

The peritoneum is then closed with a continuous 3-0 absorbable suture. The fibers of the transversus abdominis and the internal oblique muscle fall together readily, and their closure can be completed with two interrupted 3-0 absorbable ligatures. The external oblique aponeurosis is closed from end to end with a continuous 2-0 absorbable suture. Scarpa’s fascia is approximated with interrupted 3-0 absorbable sutures, and the skin is closed with a continuous subcuticular 4-0 absorbable suture and reinforcing tapes (Steri-Strips).

If the wound has been grossly contaminated, the fascia and muscles are closed as described, but the skin is loosely approximated with Steri-Strips, which can easily be removed after the procedure if surgical site infection or abscess develops. An alternative approach is to leave the skin and the subcutaneous tissue open but dressed with sterile nonadherent material and then to perform delayed primary closure with Steri-Strips on postoperative day 4 or 5. A meta-analysis of 27 studies involving 2,532 patients with gangrenous or perforating appendicitis concluded that the risk of surgical site infection was no higher with primary closure than with delayed primary closure.3

Laparoscopic Appendectomy

Figure 6. Laparoscopic appendectomy: port and camera placement

The patient is placed in the supine position, with both arms tucked along the sides, and general anesthesia is induced. Decompression with an orogastric tube should be routine, as should placement of a urinary Foley catheter and use of lower-extremity sequential compression devices. The surgeon should stand on the patient’s left side, with the assistant (who operates the camera) near the patient’s left shoulder [see Figure 6]. The monitors are placed on the opposite side of the operating table so that both the surgeon and the assistant can view the procedure at all times.

The abdomen is prepared and draped in a sterile fashion so as to expose the entire abdomen. A three-port approach is routinely used [see Figure 6]. All skin incisions along the midline are made vertically to allow a more cosmetically acceptable conversion to laparotomy, should this become necessary. The suprapubic port must be large enough to accommodate the laparoscopic stapler (usually 12 mm); the other two ports can be smaller (e.g., 5 or 10 mm). The ports are placed as far away from the operative field as possible to permit the application of a two-handed dissection technique. The use of a 30° angled scope facilitates operative viewing and dissection.

With the patient pharmacologically relaxed and in the Trendelenburg position, a Veress needle is inserted into the peritoneal cavity at the base of the umbilical ligament. Aspiration and the saline-drop test are performed to ensure that the tip of the needle is correctly positioned. Pneumoperitoneum is established by insufflating CO2 to an intra-abdominal pressure of 14 mm Hg. The first port is placed at the infraumbilical skin incision, the laparoscope is inserted, and a complete diagnostic laparoscopy is performed. Once the diagnosis of acute appendicitis is confirmed by inspection, the two remaining ports are placed under direct vision. In many cases, however, the diagnosis cannot be confirmed without first placing the second and third ports and exposing the appendix. If purulent fluid is encountered, it should be carefully aspirated dry without irrigation to ensure that the infected fluid is not disseminated throughout the abdominal cavity.

The appendix is exposed and traced to its base on the cecum by using an atraumatic retracting forceps. In cases of retrocecal appendix or severe appendiceal inflammation, it is best first to mobilize the cecum completely by taking the lateral reflection of the peritoneum around the terminal ileum and up the ascending colon with an ultrasonic scalpel (e.g., the Harmonic Scalpel; Ethicon Endo-Surgery, Inc., Cincinnati, Ohio). Surrounding structures, such as the iliac and gonadal vessels and the ureter, should be clearly identified to minimize the risk of injury. Dissection of the appendix can then begin.

Figure 7. Laparoscopic appendectomy: division of mesoappendix

The tip of the appendix is grasped and retracted anteriorly toward the anterior abdominal wall and slightly toward the pelvis; the mesoappendix is thus exposed in a triangular fashion. A window between the base of the appendix and the blood supply is created with a curved dissecting forceps. The mesoappendix is divided either with hemostatic clips and scissors or with a laparoscopic gastrointestinal anastomosis (GIA) stapler loaded with a vascular cartridge [see Figure 7]. If a window on the mesoappendix cannot be safely created because of intense inflammation, antegrade dissection of the blood supply is necessary. The ultrasonic scalpel is a handy (albeit expensive) instrument for this purpose. Endoscopic hemostatic clips usually suffice to control the small branches of the appendicular artery during the course of this dissection.

Figure 8. Laparoscopic appendectomy: clearing of appendix base
Figure 9. Laparoscopic appendectomy: delivery of specimen

The base of the appendix is then cleared circumferentially of any adipose or connective tissue and is divided with a laparoscopic GIA stapler loaded with an intestinal cartridge [see Figure 8]. To ensure an adequate closure away from the inflamed appendiceal wall, a small portion of the cecum may have to be included within the stapler. To ensure proper placement of the stapler and to prevent injury to the right ureter or the adjacent small bowel, the tips of the stapler must be clearly visualized before the instrument is closed. The use of an angled scope and an articulated rotating laparoscopic GIA stapler (e.g., Roticulator; AutoSuture, Norwalk, Connecticut) will facilitate this maneuver. A noninflamed or minimally inflamed appendix can be ligated with sutures, as described earlier [see Open Appendectomy, above]. The appendix is removed from the abdominal cavity, with care taken to avoid direct contact with the abdominal wall. A mildly inflamed appendix can be delivered through one of the larger ports; a severely inflamed appendix is often too big and hence should be delivered in a specimen retrieval bag [see Figure 9].

The operative field is irrigated and aspirated dry. Hemostasis is confirmed, and the cecum is inspected to ensure proper closure of the appendiceal stump. The ports are removed under direct vision, the absence of back-bleeding from the port sites is confirmed, and the abdomen is completely decompressed. All fascial defects larger than 5 mm are closed with 0 absorbable sutures. The skin incisions are reapproximated with a subcuticular 4-0 absorbable suture and reinforcing Steri-Strips.

Special Considerations

Histologically Normal Appendix

Acute appendicitis is the most common cause of an acute surgical abdomen in the United States, and it remains one of the most challenging diagnoses to make in the emergency department. Although the use of advanced diagnostic imaging modalities (e.g., ultrasonography and computed tomography) has led to more accurate diagnosis of acute appendicitis in research settings, it has not been shown to reduce the rate of misdiagnosis of acute appendicitis in the general population.4

The incidence of histologically normal appendix in patients with clinical signs and symptoms of acute appendicitis ranges from 8% to 41%.5–14 Nonetheless, appendectomy relieves symptoms in the vast majority of these patients. When extensive sectioning is done on histologically normal specimens, it often happens that a focus of inflammation is found in only a few serial sections. This condition is known as focal appendicitis—so called because the polymorphonuclear infiltration is confined to a single focus, while the remaining appendix is devoid of any polymorphonuclear cells.15 It is not clear that all cases of acute appendicitis arise from this focal inflammation; however, such inflammatory foci may be the earliest recognizable manifestations of appendicitis in some so-called negative appendectomies. Furthermore, a substantial proportion of histologically normal appendices removed from patients with clinical signs and symptoms of acute appendicitis exhibit significantly increased expression of tumor necrosis factor-a and interleukin-2 messenger RNA (a sensitive marker of inflammation in appendicitis) in germinal centers, the submucosa, and the lamina propria.16 Therefore, appendectomy is recommended in patients with clinically suspected acute appendicitis even when the appendix does not appear inflamed during exploration.17

Laparoscopic appendectomy has not been shown to reduce the incidence of negative exploration in patients with clinically suspected acute appendicitis [see Complications and Outcome Evaluation, Open versus Laparoscopic Appendectomy, below].

Appendiceal Neoplasm

Figure 10. Management of appendiceal mass

Neoplastic lesions of the appendix are found in as many as 5% of specimens obtained with routine appendectomy for acute appendicitis.18–21 Most are benign. Preoperative detection of such conditions is rare, and intraoperative diagnosis is made in fewer than 50% of cases. Appendectomy alone may be curative for appendiceal mucocele, localized pseudomyxoma peritonei, most appendiceal carcinoids, and other benign tumors. Definitive management of an appendiceal mass unexpectedly encountered during exploration for clinically suspected acute appendicitis depends on whether the tumor is carcinoid, its size and location, the presence or absence of metastatic disease, and histologic and immunohistochemical findings [see Figure 10].

Benign neoplasms of the appendix include mucosal hyperplasia or metaplasia, leiomyomas, neuromas, lipomas, angiomas, and other rare lesions. Appendiceal adenomas tend to be diffuse and to have a predominant villous character. Mucus-producing cystadenomas predispose to appendiceal mucocele, sometimes accompanied by localized pseudomyxoma peritonei. These lesions are rarely symptomatic and are often encountered incidentally during operation; however, they may also be clinically manifested as acute appendicitis, torsion, intussusception, ureteral obstruction, or another acute condition. If the base of the appendix is free of disease, appendectomy alone is sufficient treatment.

Malignant tumors of the appendix primarily consist of carcinoids and adenocarcinomas; all together, they account for 0.5% of all GI malignancies.22 The incidence of malignancy in the appendix is 1.35%.18 Metastasis to the appendix is rare. Carcinoids are substantially more common than adenocarcinomas in the appendix: as many as 80% of all appendiceal masses are carcinoid tumors. Overall, carcinoid tumors are found in 0.5% of all appendiceal specimens, and appendiceal carcinoid tumors account for 18.9% of all carcinoid lesions.23 These tumors are predominantly of neural cellular origin and have a better prognosis than all other intestinal carcinoid tumors, which typically are of mucosal cellular origin. If the tumor is less than 2 cm in diameter, is located within the body or the tip of the appendix, and has not metastasized, appendectomy is the treatment of choice. If the lesion is at the base of the appendix, is larger than 2 cm in diameter, or has metastasized, right hemicolectomy [see 5:34 Segmental Colon Resection] is indicated. In addition, secondary right hemicolectomy is indicated if the tumor is invasive, if mucin production is noted, or if the tumor is found to be of mucosal cellular origin at final pathologic examination.24,25 Patients with metastatic appendiceal carcinoid tumors appear to have a far better prognosis than those with other types of metastatic cancers.24 Therefore, hepatic debulking for symptomatic control is indicated and justified in cases of liver metastasis.

Primary adenocarcinoma of the appendix is rare, and as yet there is no firm consensus regarding prognosis, treatment of choice, and outcome.26 Currently, the recommended treatment is right hemicolectomy: a 1993 study found that this approach resulted in an overall 5-year survival rate of 68%, compared with a rate of 20% when appendectomy alone was performed.25 The prognosis is determined by the degree of tumor differentiation and by the histologic stage. As many as one third of these patients have a second primary neoplasm, which will be located within the GI tract about half the time.

Finally, nonepithelial appendiceal tumors, though extremely rare, occur as well. Such lesions include malignant and Burkitt lymphomas, smooth muscle tumors, granular cell tumors, ganglioneuromas, and Kaposi sarcoma.

Inflammatory Bowel Disease

The appendix is frequently involved in Crohn disease and ulcerative colitis (25% and 50% of cases, respectively), but isolated Crohn disease of the appendix is rare.27–30 When a histologically normal appendix is encountered in a patient with active Crohn disease, appendectomy should be performed because of the high risk of recurrent right lower quadrant pain, fever, and tenderness. Although isolated Crohn disease of the appendix may present as acute appendicitis, it is not clear that this condition will necessarily develop into a more extensive form of Crohn disease. Appendectomy is safe in such cases because fistulas almost never develop after appendectomy in patients with isolated involvement of the appendix.

Gynecologic Conditions

It is clear that the presentation of right lower quadrant pain in a woman remains a challenge to the treating physician. Frequently, the causes can be identified by means of proper blood work or ultrasonography, but often they can be revealed only through surgical exploration. In such cases, diagnostic laparoscopy provides an excellent view of the pelvic organs, and it offers the potential for easy continuation on to laparoscopic treatment. Ovarian cysts found in premenopausal women include unilocular clear fluid cysts (e.g., follicular cysts and corpus luteum cysts), dermoid cysts, and endometrial cysts. They can be removed by making an incision on the ovary and separating the cyst from the ovarian cortex. Dermoid cysts should be removed in toto to prevent chemical peritonitis. Endometrial cysts are best evaporated with the laser: complete removal is very difficult and sometimes impossible. Torsion of the fallopian tube or the ovary can be reversed by gentle detorsion of the organ with atraumatic forceps. If there is no evidence of ischemia, no further therapy is indicated. If there is gangrene with no indication of recovery, resection is indicated. If the organ shows partial recovery within 10 minutes after the pedicle is untwisted, a second-look laparoscopy is indicated in 24 hours. Pelvic inflammatory disease should be treated on an individualized basis in accordance with the degree of inflammation, the patient’s age and desire to have children, and the microbiologic findings.

Complications and Outcome Evaluation

Open Versus Laparoscopic Appendectomy

To date, 31 reports of randomized, controlled trials comparing laparoscopic appendectomy with open appendectomy have been published as full manuscripts in English [see Table 1].31–61 These reports involved a total of 4,352 patients, of whom 2,194 underwent laparoscopic appendectomy and 2,158 underwent open appendectomy. The incidence of histologically normal appendix was similar in the two groups (14.3% with laparoscopic appendectomy versus 14.8% with open appendectomy). The conversion rate from laparoscopic appendectomy to open appendectomy was 10% (range, 0% to 23%). Laparoscopic appendectomy was associated with a lower incidence of postoperative wound infection than open appendectomy was (3.5% versus 6.7%), but it was also associated with a higher incidence of postoperative intra-abdominal abscess (2.5% versus 1.1%). The length of stay was slightly shorter after laparoscopic appendectomy (1 to 4.9 days; average, 2.7 days) than after open appendectomy (1.2 to 5.3 days; average, 3.2 days). Randomized, controlled trials carried out within the past 5 years have not led to any significant changes in the statistical picture.

In men and children with suspected acute appendicitis, laparoscopic appendectomy has no major advantage over open appendectomy.37 In women of childbearing age and in equivocal cases, laparoscopy may be valuable as a diagnostic tool, but the practice of not removing a normal-looking appendix during exploration for right lower quadrant pain is controversial. Laparoscopic appendectomy appears to offer the potential benefit of less postoperative adhesion formation, but the evidence is inconclusive in the light of the short follow-up times reported in these trials, and the higher incidence of intra-abdominal abscess formation remains cause for concern. To date, unfortunately, there have been no studies designed specifically to address reduced adhesion formation as a primary end point.

Although laparoscopic appendectomy is being performed with increased frequency, it continues to be used selectively. The laparoscopic version of the procedure is at least as safe as the corresponding open procedure, but it is undeniably more time-consuming and more costly. Moreover, it remains questionable whether the benefits of laparoscopic appendectomy—reduced postoperative pain, earlier resumption of oral feeding, shortened hospital stay, quicker return to normal preoperative activities, and lower incidence of surgical site infection—outweigh the doubled incidence of postoperative intra-abdominal abscess formation. Further randomized clinical studies focusing on the efficacy of laparoscopic appendectomy as a diagnostic tool and on the incidence of postoperative intra-abdominal abscess and adhesion formation are needed, as are additional cost analyses.



J. Graham Williams, MCH, FRCS

Consultant Surgeon
Royal Wolverhampton Hospitals NHS Trust, England

Formation of an intestinal stoma is frequently a component of surgical intervention for diseases of the small bowel and the colon. The most common intestinal stomas are the ileostomies (end and loop) and the colostomies (end and loop); the less common stomas, such as cecostomy and appendicostomy, have limited applications and thus are not considered further in this chapter.

For optimal results, it is essential that stoma creation be considered an integral part of the surgical procedure, not merely an irritating and time-consuming addendum at the end of a long operation. Accordingly, the potential requirement for a stoma should be appropriately addressed in the planning of an intestinal procedure. A great effort should be made to counsel the patient before operation as to whether a stoma is likely to be needed, what stoma creation would involve, where the stoma would be situated, and whether the stoma is likely to be permanent or temporary.

Operative Planning

Preoperative Counseling

Ideally, as soon as surgical intervention that may involve a stoma is contemplated, the enterostomal nursing service should become involved—though this may not be possible in an emergency setting. Patients often have misconceptions about the effects stoma will have on their quality of life and consequently may experience considerable anxiety. Adequate preoperative counseling helps correct these misconceptions and reduce the attendant anxiety. Enough time should be set aside to allow the counselor to explore the patient’s knowledge of the disease and understanding of why a stoma may be required. This process involves reviewing the planned operation, describing what the stoma will look like, and explaining how the stoma will function. Visual aids (e.g., videos, CD-ROMs, and booklets) can be very useful in this regard and should be freely available to patients and their families. As simple a measure as showing the patient a stoma appliance and attaching it to the abdominal wall before the procedure can be helpful in preparing the patient for a stoma. Many patients facing the prospect of stoma surgery also derive great benefit from meeting patients of similar age and background who have a stoma.

Choice of Procedure

A number of common indications for stoma formation have been identified [see Table 1]. These indications are usually associated with particular types of stoma, but the association is not always a simple or automatic one. In many situations, more than one option exists, and it can be difficult to select the best available option for a particular patient.

Loop Ileostomy versus Loop Colostomy

Defunctioning of a distal anastomosis after rectal excision and anastomosis may be achieved with either a loop ileostomy or a loop transverse colostomy. A number of nonrandomized studies1–3 and randomized control trials4–7 have been performed in an effort to determine which of these two approaches is superior. Both types of stoma effectively defunction the distal bowel; however, loop ileostomy appears to be associated with a lower incidence of complications related to stoma formation and closure, though it may also carry a higher risk of postoperative intestinal obstruction.6 The two types of stoma are comparable with respect to patient quality of life, and the degree of subsequent social restriction is influenced more by the number and type of complications than by the type of stoma formed.8

Selection of Stoma Site

A poorly sited stoma will cause considerable morbidity and adversely affect quality of life. For this reason, great emphasis should be placed on selecting the best site for the stoma on the abdominal wall. In many instances [see Table 1], it may not be possible to decide beforehand whether a colostomy or an ileostomy is to be performed. An example would be the case of a patient with a tumor in the lower rectum in which the surgeon’s intention is to perform a restorative resection covered by a loop ileostomy. In such a case, the surgeon sometimes finds that restorative resection is not technically possible and elects to perform an abdominoperineal resection or a low Hartmann resection with an end colostomy instead.

A stoma should be brought out through a separate opening in the abdominal wall, not through the main incision: there is a high incidence of wound infection and incisional hernia formation if the main incision is used as a stoma site. In general, ileostomies are sited in the right iliac fossa, sigmoid colostomies (loop or end) in the left iliac fossa, and transverse loop colostomies in either the right or the left upper quadrant. These positions are preferred because they are conveniently close to the particular bowel segments to be used for creating the various stomas. At need, however—as when finding a suitable site proves difficult because of previous scars or deformity—both the ileum and the colon can be mobilized to provide sufficient length to reach most sites on the abdominal wall.

In selecting and marking a stoma site, the following key considerations should be taken into account:

  1. A flat area of skin is required for adequate adhesion of the appliance.
  2. The patient should be able to see the stoma.
  3. Skin creases, folds, previous scars, and bony prominences should be avoided.
  4. The stoma site should not be located at the beltline.
  5. The site should be identified with the patient lying, sitting, and standing.
  6. Preexisting disabilities should be taken into account.


According to received wisdom, the stoma should be brought out of the abdomen through the rectus abdominis, so that the emerging stoma will be supported and the incidence of parastomal hernia reduced. Several studies, however, have shown that this approach is not always ideal and that the optimum site for a stoma should be selected without regard to its position in relation to the rectus abdominis.9–11 Once selected, the site is marked with an indelible pen or tattooed with India ink and a fine needle.

Operative Technique

General Principles

Most abdominal stomas are formed at the end of an open operation performed to resect bowel, drain an infectious focus, or relieve obstruction. In this setting, a midline incision is generally the most appropriate choice for gaining access to the abdominal cavity because it leaves the areas to either side of the midline available for stoma placement. Other incisions may be used as well, but more careful operative planning will be required.

A defunctioning stoma can be created without opening the abdomen by making a trephine hole and using retractors and forceps to identify the relevant bowel loop from which the stoma will be formed. I generally avoid this approach, for two reasons. First, the trephine hole invariably ends up larger than is ideal, and the greater size leads to an increased risk of parastomal hernia. Second, it is often difficult to be sure that the correct bowel loop has been identified and the correct end opened as a stoma. These disadvantages can be overcome by taking a laparoscopic approach. One port is placed though the previously marked site. A tissue forceps is passed down this port and used to grasp and orient the relevant bowel segment. If necessary, the bowel can be mobilized by means of laparoscopic dissection. The colon is then divided with a linear stapler, and the proximal end is brought out through a small trephine hole made at the port site.

The fundamental concept in stoma formation is that a stoma is simply an anastomosis between a piece of bowel and the skin of the abdominal wall. For this reason, the same basic principles that apply to intestinal anastomosis also apply to stoma formation—namely, maintaining an adequate blood supply to both sides of the anastomosis, ensuring that the anastomosis is performed without tension, and avoiding any preexisting infection. In accordance with these principles, the bowel segment used should have as much of its blood supply as possible preserved during mobilization, and mobilization should be sufficient to allow the bowel to be brought through the abdominal wall without tension and without occlusion of the blood supply at the fascial level by a too-small hole in the abdominal wall. If these criteria are not met, then either the bowel should be mobilized further or a new bowel segment should be selected. It is important to make the best possible technical choices at the time of initial stoma formation. If the correct principles are not followed at the beginning of the procedure, it is generally futile to hope that the situation will improve thereafter; the usual result is a poor stoma that requires surgical revision.

Creation of Stoma Aperture

It is wise to leave formation of the hole for the stoma until the end of the procedure because unforeseen events during the operation may necessitate a change in the type or the site of the stoma. A circular incision 2.5 cm in diameter is made at the marked site, and the skin is excised. The subcutaneous fat is parted with scissors and small retractors until the fascia of the abdominal wall is reached. The fat need not be excised: it supports the emerging stoma, and its absence would leave a potential dead space. A cruciate incision is made in the rectus sheath, initially no more than 2 cm in each direction. The muscle fibers of the underlying rectus abdominis are split in the direction of their fibers with an arterial clamp or the tips of heavy scissors. The small retractors are inserted deeper to keep the muscle fibers apart, and a small cruciate incision is made in the posterior rectus sheath with an electrocautery. A swab held against the peritoneum at the stoma site will protect the intra-abdominal organs and the assistant’s fingers from being injured by the electrocautery point.

On occasion, the epigastric vessels, which lie between the rectus abdominis and the posterior sheath, are injured. Should this occur, the simplest way of dealing with the problem is to open the posterior sheath from inside the abdominal cavity and suture-ligate the bleeding point.



The typical site for an end colostomy is the left iliac fossa, and either the sigmoid or the descending colon is used for the stoma. If the rectum has been excised, the inferior mesenteric vessels will have been divided, and the blood supply to the distal colon will come from the middle colic vessels via the marginal artery. It is not usually necessary to take down the splenic flexure to mobilize the colon adequately; however, if there is any concern regarding tension on the stoma, full splenic flexure mobilization should be performed. For a simple defunctioning end colostomy, only a few small vessels in the mesentery will have to be divided.

Figure 1. End colostomy

The colon is divided at the relevant site with either crushing clamps or a linear intestinal stapler. The adequacy of the vascular supply is checked by inspection. A nontraumatic bowel clamp or a Babcock tissue forceps is passed through the hole in the abdominal wall and used to grasp the closed-off end of the colon. Care is taken when drawing the colon through the abdominal wall to keep from twisting the colon and damaging the small vessels in the supporting mesentery. The end of the colon should sit 2 cm above the skin surface. To prevent wound contamination, the colostomy is constructed only after the skin incision has been fully closed and dressed. The closed-off end of the colon is excised with a sharp knife, and the colostomy is constructed with a small spout by everting the bowel wall. The spout helps the patient position the stoma appliance but should not protrude more than 0.5 to 1 cm above the surface of the skin. The anastomosis is performed with interrupted absorbable sutures that take bites of the full thickness of the end of the colon and the subcuticular layer of the skin. Small bites are also taken of the seromuscular layer of the emerging colon at the level of the skin [see Figure 1].

This technique is sometimes modified by closing the lateral space between the abdominal wall and the colon with absorbable sutures in an effort to prevent internal herniation of the small bowel. An alternative approach is to tunnel the colostomy to the hole in the abdominal wall via an extraperitoneal route. This approach may prevent herniation and colostomy prolapse,9 but the stoma may be slow to function and difficult to mobilize if a reversal or revision operation is performed.


A loop colostomy is usually performed as a quick and temporary method of relieving acute colonic obstruction or to cover an anastomosis in the distal colon or rectum. Whenever possible, I avoid using loop colostomies, for the following reasons.

  1. Because of the need to accommodate two pieces of bowel, a loop colostomy requires a larger hole in the abdominal wall than an end colostomy does. This is a particular concern in emergency situations, where the colon may be greatly dilated.
  2. The larger hole predisposes to formation of a parastomal hernia, which can be a problem if the stoma is not reversed.
  3. Loop colostomies are more prone to prolapse than end colostomies are, possibly as a conseqence of parastomal hernia formation.
  4. The effluent from the transverse colon can be highly liquid, and the absence of a spout with loop colostomy may lead to difficulties with appliance leakage.
  5. When a loop colostomy is used to defunction a distal anastomosis, there is a theoretical risk of damage to the marginal artery, which may be the only vessel supplying the distal side of the anastomosis.


The usual site for a loop colostomy is either the right upper quadrant (using the proximal transverse colon) or the left iliac fossa (using the left colon). The colon segment that will be used to form the stoma is identified, and peritoneal attachments are divided to provide sufficient length to reach the desired site on the abdominal wall without tension. If the transverse colon is to be used, the omentum is removed. Care is taken not to damage the marginal artery, which, if occluded, may compromise vascular supply to the distal bowel.

Figure 2. Loop colostomy

A trephine hole is made at the marked site as described [see Operative Technique, General Principles, above]. The hole is usually larger than it would be in an end colostomy; the bowel loop to be brought out is often bulky, especially when the colon is obstructed. A small window is made in the mesentery immediately adjacent to the colon wall, and a Jacques catheter is passed through this aperture. The Jacques catheter is used as a handle by which the colon loop is drawn through the trephine hole in the abdominal wall, with care taken to maintain the orientation of the colon and avoid twisting [see Figure 2, part a]. The catheter is then replaced by a plastic or glass stoma rod, which supports the loop at the level of the skin.

The main incision is closed, and the stoma is matured. A transverse incision is made in the apex of the bowel loop [see Figure 2, part b], and the two edges are peeled back and sutured to the skin edge of the trephine hole to produce a double opening [see Figure 2, parts c and d]. The bridge remains in place for 5 days, by which time the stoma is usually beginning to function properly. The rod can then be removed because by this point, the stoma is fixed in place and unable to retract into the abdominal cavity.


At one time, there was a vogue for creating a double-barrel colostomy to defunction the colon. Although the height of the vogue has passed, this type of stoma still has a place in the management of colorectal trauma. After resection of a damaged segment of the colon, the proximal and distal ends of the colon are tacked together along the antimesenteric surfaces with interrupted absorbable sutures. The resulting double end is then brought out through a trephine incision at the relevant site. The double-barrel configuration makes the colostomy easier to close: closure can be performed after mobilization by resection and a sutured anastomosis or via a double-stapled technique.



End ileostomy is most frequently performed after colectomy for inflammatory bowel disease. The most distal segment of the ileum is used (i.e., that immediately proximal to the ileocecal valve), the reason being that it is important to preserve intestinal length, both for nutritional reasons and to allow for the possibility that an ileoanal pouch may have to be fashioned in the future. In certain instances, it is necessary to create an end ileostomy from a more proximal segment of the ileum.

Figure 3. End ileostomy: preparation of terminal ileum

The terminal ileum is mobilized, a large avascular window is opened between the ileocolic vessels and the ileal branches of the superior mesenteric vessels, and the ileocolic vessels are divided where they branch from the superior mesenteric vessels. The terminal ileum is usually supplied by two arcades of vessels, which join the ileocolic vessels adjacent to the cecum. These arcades must be divided as close to the ileocolic vessels as possible to preserve the blood supply to the terminal ileum [see Figure 3]. The ileocecal fold (Treves’s fold) is dissected away from the terminal ileum, which can then be divided flush with the ileocecal valve, either with a linear stapler or with a knife between bowel clamps.

The trephine incision is created at the previously marked site, and a Babcock tissue forceps is passed into the abdominal cavity and used to grasp the divided end of the ileum. The terminal ileum and the supporting mesentery are gently eased through the aperture, with the mesenteric surface oriented superiorly, until 5 cm of ileum protrudes above the abdominal skin. The cut edge of the ileal mesentery is secured to the peritoneum of the back of the anterior abdominal wall, along the line of the lateral border of the rectus abdominis, with an absorbable suture. This measure helps stabilize the stoma and is thought to prevent stoma prolapse, volvulus, and internal herniation around the stoma.

Figure 4. End ileostomy: placement of sutures

The stapled end of the ileum is excised to produce a fresh bleeding end. The emerging ileum is then everted to yield a spout about 2.5 cm long. This is accomplished by placing a suture on either side of the mesentery and a third suture on the antimesenteric side, which lies inferiorly. The superior sutures take bites of the serosa of the emerging ileum, 5 cm from the cut end of the bowel, and the inferior suture includes a serosal bite 4 cm from the cut edge [see Figure 4]. When the sutures are tied, an everted spout is created that points downward into the ileostomy appliance.12 The mucocutaneous anastomosis is then completed with a series of interrupted absorbable sutures.


Figure 5. Loop ileostomy

A loop ileostomy is employed to rest the distal bowel or to protect an anastomosis. The ileal loop used should be as distal as possible while still maintaining adequate mobility; if there is any tension, a more proximal loop may be required. The technique of loop ileostomy formation is similar to that of loop colostomy formation. A Jacques catheter is used to draw the loop through the abdominal wall trephine hole, ideally with the proximal limb in the lower position [see Figure 5, part a]. Care is taken to distinguish the proximal and distal limbs of the loop and to keep from rotating the loop during its passage through the abdominal wall. A marking suture is useful for identifying the proximal side of the loop. A supporting rod may be used, but it is not necessary, and it can hinder the fitting of the stoma appliance.

The ileostomy is created by making a circumferential incision around 80% of the distal limb at the level of the skin, with the mesenteric side preserved [see Figure 5, part b]. The cut edge of the proximal limb is then everted to create a spout for the ileostomy [see Figure 5, part c]. A Babcock tissue forceps is sometimes used to apply gentle traction to the mucosal side of the proximal limb. The cut edge of the ileum is anastomosed to the skin with a series of interrupted subcuticular absorbable sutures. The distal limb is sutured flush with the skin. On the proximal side, several sutures take bites of the serosa of the emerging ileum at skin level. The corners of the incision in the ileum are drawn around the proximal limb of the ileostomy to accentuate the spout effect and create a thin, semilunar distal limb opening [see Figure 5, part d].

An alternative approach is to create a divided loop ileostomy, which some consider superior to a conventional loop stoma.13 The construction technique for this stoma is similar to that of its conventional counterpart. The distal limb of the ileostomy is divided with a linear cutting stapler after the loop is brought through the abdominal wall. The closed distal end is tacked to the side of the emerging spout of the proximal end below skin level, and the proximal end is fashioned into an everted spout as in a conventional end ileostomy. A divided loop ileostomy is slightly more difficult and expensive to construct than a conventional loop ileostomy, but it has the advantage of achieving complete defunctioning of the distal bowel (because there is no chance that the ileostomy contents will spill over).


A loop-end ileostomy can be useful in cases where the ileum and its supporting mesentery are grossly thickened and the surgeon is encountering difficulty in preparing a sufficient length of well-vascularized ileum for a conventional end ileostomy. In a loop-end ileostomy, the ileum is prepared as in a conventional end ileostomy, but the vascular arcades are left undisturbed. A small window is made in the mesentery 5 to 10 cm proximal to the closed end of the ileum, and a nylon tape or a Jacques catheter is used to draw this distal ileal loop through the abdominal wall. The stapled closed end of the ileum lies just within the abdominal cavity. The ileostomy is then constructed in essentially the same manner as a conventional loop ileostomy.


A split ileostomy is created by bringing out the two cut bowel ends at different sites. The proximal end is usually terminal ileum, but the distal end may be either ileum or colon, depending on the indication for stoma formation. This procedure forms a mucous fistula, and only a small stoma appliance is usually required. The distal end can be either included in the closure of the abdominal wound or brought out through a separate trephine hole on the opposite side of the abdomen from the ileostomy. The advantage of a split ileostomy is that it completely defunctions the bowel without the risk of intra-abdominal leakage from a closed distal stump. The disadvantage is that it is more difficult to close: closure usually necessitates reopening of the main incision.


A continent ileostomy involves formation of a reservoir and placement of a nonreturn nipple valve, which is emptied regularly via a catheter, so that the patient need wear only a small cap appliance. The surgical technique is demanding and beyond the scope of this chapter; it is described more fully elsewhere.14

Stoma Closure

Loop Ileostomy

Closure of a loop ileostomy is usually a simple local procedure that does not require the main incision to be opened. The operation is easier to perform if a period of at least 12 weeks is allowed to elapse between formation of the stoma and closure so that there is time for edema and inflammatory adhesions to settle. Dissection is facilitated by injecting epinephrine (1:100,000 solution) into the subcutaneous plane around the stoma.

Figure 6. Stoma closure: loop ileostomy

An incision is made in the peristomal skin 2 mm from the mucocutaneous junction [see Figure 6, part a]. The incision is deepened into the subcutaneous fat until the serosa of the emerging bowel appears. Sharp dissection is continued circumferentially in this plane, dividing the fine adhesions between the bowel and its mesentery and the subcutaneous fat [see Figure 6, part b]. Blunt dissection should be avoided because it can easily lead to serosal tears. Some difficulty may be encountered at the fascial level, and care must be taken with the dissection if adhesions are particularly dense. Eventually, the peritoneal cavity is entered, and the remaining adhesions are identified with a finger and divided.

The emerging ileal loop is withdrawn from the abdominal cavity, and the mucocutaneous junction and the rim of skin are excised. The everted proximal end of the stoma is unfolded [see Figure 6, part c]; some sharp dissection is usually required to accomplish this. The freshened edges of the enterotomy are then approximated with interrupted seromuscular absorbable sutures [see Figure 6, part d]. Sometimes, a limited ileal resection is required if the stoma site is in poor condition, and a conventional end-to-end anastomosis is performed to restore intestinal continuity. It is possible to close a loop ileostomy with a double-stapled technique; however, there does not appear to be much advantage in doing so. Two randomized trials and a nonrandomized study comparing suture closure with stapled closure yielded conflicting results with respect to complication rates,15–17 but both randomized trials reported that extra costs were incurred when staples were used. Once the enterotomy is closed, the loop of ileum is returned to the abdominal cavity, and the stoma site is closed with interrupted nonabsorbable sutures.

A divided loop ileostomy is closed in the same manner as described above. Care should be taken to identify the closed distal end and to fully mobilize both limbs of the ileum from the abdomen. The closed distal end is separated from the proximal limb, and the staple line is excised to yield a fresh end. The proximal end is unfolded and a simple end-to-end anastomosis is performed with interrupted sutures. There may be a significant size discrepancy between the two limbs. Again, a double-stapled technique may be employed as an alternative closure method.

Loop Colostomy

A loop colostomy is closed in much the same manner as a loop ileostomy after the emerging colon is mobilized away from the subcutaneous fat and the abdominal wall by means of sharp dissection. Transverse closure is achieved with interrupted absorbable sutures.

Postoperative Care

A clear stoma appliance is cut to the proper size and placed on the stoma before the patient leaves the operating room. A degree of edema is to be expected in the first week. In addition, the stoma may appear somewhat dusky; this is a sign that the aperture in the abdominal wall is the correct size. It often happens that the patient becomes alarmed at the initial appearance of the stoma and requires reassurance that the stoma will look better as time passes.

When the stoma starts to function, the clear appliance is changed for the chosen appliance, and the patient is instructed in how and when to empty the pouch. When confident with this aspect of stoma management, the patient is instructed in how to cut the plate to the correct size and how to change the stoma bag or flange (if he or she is using a two-piece appliance). An ileostomy works throughout the day, often showing an increase in activity after meals. The appliance will therefore require regular emptying, a task that some patients find inconvenient. A loop transverse colostomy may be as unpredictable as an ileostomy in this regard; however, a sigmoid colostomy may have a more predictable activity, similar to the frequency of normal bowel movements. Some patients find that quality of life is improved by irrigating the colostomy with water instilled via a special appliance. This procedure induces a full colonic clearout and allows the patient to wear a less obtrusive cap appliance for 24 to 48 hours, until irrigation is repeated.

Detailed discussion of stoma care is beyond the scope of this chapter. A key role is played by the enterostomal therapist, who is an important point of contact for the patient, providing advice, instruction, and emotional support in the postoperative period. Skin complications are common, and most can be managed by the enterostomal therapist. Many such complications result from contact between the peristomal skin and digestive enzymes; common causes include poor appliance fit and stoma retraction. Skin problems can usually be resolved by means of simple measures such as switching to a different appliance, using a convex flange, applying barrier cream, or filling dips in the peristomal skin with stoma paste. Given that surgical complications such as fistula formation and parastomal hernia may present as skin problems, it is important that the surgeon and the enterostomal therapist work closely together in addressing these problems.


Wound infection after stoma closure is common. Drainage of the incision with a small corrugated drain can help reduce the incidence of such infection. Some surgeons leave the stoma site open and allow it to heal by second intention.

Incisional hernia can develop in the stoma site, and its incidence is increased by wound infection in the postoperative period. Because the defect is relatively narrow, the hernia can lead to significant symptoms. Repair is usually necessary.

Breakdown of the anastomosis lying beneath the incision will lead to a fecal fistula, with discharge from the stoma site. If the fistula is simple and there is no distal obstruction, it is likely to heal spontaneously. Expert nursing is required to manage the fistula effluent while healing occurs to prevent damage to the surrounding skin.

If there is a complex inflammatory mass at the closure site, spontaneous healing is less likely. Laparotomy may be required, with resection of the stoma site and reanastomosis or further stoma formation, depending on the patient’s condition.


Complications after stoma formation are frequent and varied [see Table 2] and can adversely affect quality of life. The complication rate has been reported to be about 25% after a colostomy formation and as high as 57% after an end ileostomy18 and 75% after a loop ileostomy.19 Cumulative complication rates at 20 years have reached 76% in patients undergoing ileostomy for ulcerative colitis and 56% in those undergoing ileostomy for Crohn disease.20 As noted [see Postoperative Care, above], many complications can be successfully managed with enterostomal care.18 This is fortunate because the results of surgical correction are often unsatisfactory, with many patients requiring further surgical revision of their stomas.21

Careful assessment is warranted when a patient presents with stomal complications. Such complications may be interrelated or may have a different cause from what initial examination suggests. For example, skin damage may be a result of a poorly fitting appliance, but the poor fit may itself be caused by a parastomal hernia or a flush ileostomy. Furthermore, stomal complications may arise from renewed activity of the underlying disease (e.g., recrudescence of Crohn disease21–23 or recurrence of cancer).


Mild ischemia of the stoma is common in the early postoperative period but usually resolves within a few days. More profound ischemia can result in necrosis of all or part of the circumference of the bowel end used to form the stoma. Satisfactory healing of the stoma depends on an adequate blood supply. Problems with the blood supply are more common with end stomas than with loop stomas; likely causes include excessive division of mesenteric blood vessels, tension on the stoma from inadequate mobilization, and a too-narrow aperture through the abdominal wall that constricts the vessels at the fascial level. It is a good idea to prepare the relevant bowel segment for use in a stoma some time before the end of the operation so that any problems with the blood supply will be evident before the stoma is fashioned. An obviously ischemic stoma should be revised at the time of operation. Such revision may include mobilization of a more proximal bowel segment.

Patchy necrosis that is confined to the mucosa can be managed expectantly and usually heals by second intention. Complete necrosis of an ileostomy is an indication for urgent revision. Necrosis of a colostomy may not necessitate revision if the segment is short. However, a fistula may form at the fascial level, or stenosis may develop as the necrotic segment heals.


Stenosis of the stoma is a consequence of postoperative ischemia. Mild stenosis can be managed with simple dilatation and may not cause many symptoms, particularly if the effluent is liquid. Substantial stenosis of a colostomy can lead to subacute obstruction that must be managed with surgical revision. Sometimes, revision can be accomplished as a local procedure. A disk of skin that includes the stenosed stoma site is excised. The distal colon is mobilized and sutured to the new skin opening. In most instances, however, it is not possible to mobilize sufficient length with this approach, and laparotomy is required for adequate mobilization.


Prolapse may occur with any type of stoma but is most common with loop colostomy. Patients with loop colostomies usually have a degree of parastomal hernia, which allows adequate space for prolapse of the emerging bowel. Appearances are often alarming, and symptoms are usually related to difficulties with fitting an appliance or to leakage. The best treatment option is to close the stoma (if appropriate). Another option is to divide the loop stoma, thus creating an end colostomy, and then to return the closed distal end to the abdomen. Amputation of the prolapsed stoma corrects the problem in the short term, but the prolapse often recurs quickly. Repairing a coexisting parastomal hernia can lower the risk of recurrence, but it involves a more extensive operation [see Parastomal Hernia, below]. Neither ensuring that the emerging stoma is brought through the rectus abdominis nor fixing the mesentery to the abdominal wall appears to prevent stoma prolapse.20


Figure 7. Stabilization of retracted ileostomy

Stoma retraction is more of a concern with an ileostomy than with a colostomy because of the possibility of leakage from the appliance. Retraction generally results from poor adhesion between the serosal surfaces of the everted stoma but may also reflect the presence of a parastomal hernia. If the retracted ileostomy is fixed in position, laparotomy will probably be required to correct the problem, though it is worthwhile to attempt local mobilization of the stoma after incising the mucocutaneous junction. If the retracted ileostomy is mobile, the problem can be corrected by inserting a series of interrupted absorbable sutures through the full thickness of the everted stoma to fix the walls together. A similar effect can be obtained by pulling the retracted stoma upward with tissue forceps, then fixing the walls together with several firings of a noncutting linear stapler inserted into the ileostomy, with care taken to avoid the mesentery [see Figure 7].24

Parastomal Hernia

Formation of an abdominal stoma necessarily involves creating a defect in the abdominal wall to accommodate the emerging bowel. Such defects may become enlarged as a result of tangential force applied to the edge of the opening, and this enlargement may lead to hernia formation. The tangential force is related to the radial force and the radius of the opening; in turn, the radial force is related to the intra-abdominal pressure and the radius of the abdominal cavity.25 Consequently, tangential forces are greater in larger openings in obese patients, who are thus at greater risk for parastomal hernia. Patients undergoing emergency procedures in which dilated bowel is used to form a stoma are also likely to be at increased risk for hernia formation. Care must be taken to make an opening that is just large enough for the emerging bowel. An incision that admits only two fingers is appropriate for most elective indications.

Several authors have addressed the problem of enlargement of the stoma opening by reinforcing the opening with a prosthetic ring or a sheet of Marlex mesh.25,26 One randomized trial compared the incidence of parastomal hernia in patients undergoing conventional end colostomy with the incidence in patients undergoing colostomy with insertion of a partially absorbable lightweight mesh between the posterior rectus sheath and the rectus abdominis.27 At 12 months, eight of the 18 patients with a conventional colostomy showed evidence of parastomal hernia formation, compared with none of the 16 with a mesh-reinforced colostomy.27

There remains some controversy over the issue of where the stoma site should be located in relation to the rectus abdominis. Some authors claim that hernia formation is less frequent when the stoma emerges through the rectus abdominis28–30; however, other authors dispute this claim,9–11 and a clinical and radiologic study of paraileostomy hernia found no differences in incidence between stomas brought out through the rectus abdominis and stomas brought out more laterally.31

The incidence of parastomal hernia formation varies widely among published studies [see Table 3]. This wide variation reflects both differences in length of follow-up and differences in the methods used to identify parastomal hernias. Given that many hernias are small and asymptomatic, the true incidence of hernia formation may well be higher than the reported figures. It is generally accepted, however, that paracolostomy hernias are more common than paraileostomy hernias. It is unclear why this is so, but the reason is likely to involve the size of the opening in the abdominal wall.

Parastomal hernias are often asymptomatic, and in obese patients, they may not be apparent on clinical examination. Patients usually present with an unsightly bulge at the stoma site, but they may also have other symptoms, such as leakage around the stoma appliance, skin problems, or difficulty in irrigating a colostomy. Rarer presenting symptoms include intestinal obstruction and strangulation of the bowel loop within the hernia. Clinical examination usually suffices for making the diagnosis, particularly when performed with the patient standing. Small hernias in obese patients can be a challenge to diagnose; in this setting, computed tomographic scanning limited to the stoma area can be helpful.31

Surgical repair of a parastomal hernia often yields disappointing results and should be considered only if the patient’s symptoms are troublesome. Many patients manage reasonably well by wearing a suitably adapted appliance and a support belt. When surgical repair is indicated, it follows one of three possible approaches:

  1. Local repair. This approach to hernia repair is the simplest of the three but also the least successful.32–34 The stoma is mobilized, and the sac is identified and removed. The defect in the fascia of the abdominal wall is narrowed around the emerging bowel with a series of interrupted nonabsorbable sutures. The repair is completed by recreating the mucocutaneous anastomosis.
  2. Repair with prosthetic mesh. Mesh repairs have become increasingly popular as different meshes have become available and as surgeons have become aware of the advantages of these materials in hernia surgery. The mesh can be inserted intra-abdominally,35,36 in the preperitoneal plane [see Figure 8],32,37 or in the subcutaneous plane.38,39 Regardless of where the mesh is inserted, the basic principle is the same—namely, to achieve and maintain a narrowing of the stoma site by surrounding the emerging bowel with a sheet of mesh in which a hole is cut to accommodate the stoma.
  3. Stoma relocation. The stoma can be moved to a fresh site on the abdominal wall without reopening the main incision. The stoma is fully mobilized, and a new hole is made in the abdominal wall. A plane is developed between the peritoneum and the abdominal contents by means of blunt finger dissection between the existing stoma site and the new one. The mobilized stoma is then passed through the new hole.40 If difficulties are encountered, a laparotomy will be required. An alternative approach is to reroute the stoma through a new fascial defect while maintaining the existing skin aperture. The original fascial defect is repaired with mesh.41


The best method of repair has not been established.42 Most published studies have included relatively few patients who were followed for a relatively short time. With longer follow-up, recurrence rates as high as 76% have been reported. Local repair is associated with the highest recurrence rate,43 and stoma relocation carries an increased morbidity (from incisional hernia at the original stoma site).44 Nor is mesh repair free of problems: intra-abdominal placement of mesh is associated with a significant risk of adhesions to the mesh and of small bowel obstruction.45 The risk of mesh infection is highest when the mesh is placed in a superficial position through a parastomal incision.

At present, the best approach is to tailor repair to the individual patient’s condition and situation. For more specific recommendations, randomized trials of the different methods of parastomal hernia repair will be required. There is a growing amount of evidence in favor of inserting prosthetic mesh at the time of stoma formation in an effort to reduce the incidence of this complication.


Conditions that may cause intestinal obstruction after stoma formation include stenosis of the stoma, parastomal hernia, postoperative adhesions, and recurrent disease (e.g., Crohn disease in the proximal ileum or recurrent cancer). Management depends on the cause of the obstruction. Retrograde contrast studies are useful for identifying the site and determining the likely cause of obstruction.


A fistula may form adjacent to a stoma as a consequence of inadvertent full-thickness placement of a suture through both walls of the stoma during formation, pressure necrosis at skin level from a tightly fitting stoma appliance, or recurrent disease, especially Crohn disease in the ileum proximal to the stoma. Surgical treatment usually involves laparotomy and reformation of the stoma at a new site.

Other Complications

Other, less common complications arising after stoma formation include bleeding, perforation, skin ulceration, and the development of cancer [see Table 4].


Neil J Mortensen, MD

Professor of Colorectal Surgery, Chairman General and Vascular Surgery, Nuffield Department of Surgery, John Radcliffe Hospital, Headington, Oxford, United Kingdom

Shazad Ashraf, MD

Bobby Moore Fellow (CRUK), University of Oxford and Cancer and Immunogenetics Lab, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Headington, Oxford, United Kingdom



The creation of a join between two bowel ends is an operative procedure that is of central importance in the practice of a general surgeon. Despite the potentially disastrous consequences that can arise from leakage of an intestinal anastomosis, joining bowel segments is one of the procedures that junior surgeons often perform in the emergency setting. With proper supervision and appreciation of fundamental concerns, however, there is little difference between the outcomes of anastomoses performed by trainees and those performed by established surgeons.1

To minimize the risk of potential complications, it is imperative to adhere to several well-established principles [see Table 1]. The main ones relate to the creation of a tension-free join with good apposition of the bowel edges in the presence of an excellent blood supply.2 The importance of surgical technique is underscored by the wide variations of anastomotic leakage rates among surgeons.3

The frequency of anastomotic leakage ranges from 1 to 24%.4–6 The rate of leakage is generally considered to be higher for elective rectal anastomoses (12 to 19%) than for colonic anastomoses (11%).7–9 The consequences of postoperative dehiscence are dire and include peritonitis, bloodstream infection, further surgery, creation of a defunctioning stoma and death. A threefold rise in mortality was seen (from 7% to 22%) in the St. Mary’s Large Bowel Cancer Project, when anastomotic leakage occurred.7 Also other problems encountered include a significant increase in hospital stay, and there is a proportion of patients who do not go on to have their stomas reversed.4

This chapter is divided into three broad sections. The first discusses factors that influence intestinal anastomotic healing. The second analyzes different technical options for creating anastomoses. The third and final section concentrates on the operative techniques that are currently used in constructing intestinal anastomoses.

Intestinal Healing

The process of intestinal anastomotic healing mimics that of wound healing elsewhere in the body in that it can be arbitrarily divided into an acute inflammatory (lag) phase, a proliferative phase, and, finally, a remodeling or maturation phase. Collagen is the single most important molecule for determining intestinal wall strength, which makes its metabolism of particular interest for understanding anastomotic healing. During the proliferative stage, fibroblasts become the predominant cell type, playing an important role in laying down collagen in the extracellular space. At the epithelial level, the crypts undergo division to cover the defect on the luminal surface of the bowel. The density of collagen synthesis is in a constant state of dynamic equilibrium, which is dependent on the balance between the rate of synthesis and that of collagenolysis. After surgery, degradation of mature collagen begins in the first 24 hours and predominates for the first 4 days. This is caused by the upregulation of matrix metalloproteinases (MMPs), which are an important class of enzymes involved in collagen metabolism.10 This family includes 20 zinc-dependent endopeptidases, among which is collagenase (MMP-1).10 In vivo use of MMP inhibitors has been found to increase the strength of intestinal anastomoses by up to 48% at postoperative day 3, which suggests that these enzymes may be important in determining the risk of leakage.11 Sepsis is thought to increase the level of transcription and activity of these enzymes, thereby potentially leading to problems in the early postoperative period. In an animal model where bacterial peritonitis was induced, increased MMP levels were seen on postoperative day 3, coinciding with a fall in the bursting pressure.10 However, no increase in anastomotic dehiscence was found in comparison with the control group. By postoperative day 7, collagen synthesis becomes the dominant force, particularly proximal to the anastomosis. After 5 to 6 weeks, there is no significant increase in the amount of collagen in a healing wound or anastomosis, though turnover and thus synthesis are extensive. The strength of the scar continues to increase for many months after injury.

Figure 1. Tissue Layers of the Jejunum

The cross-linking between collagen fibers and their orientation are the major factors that determine the tensile strength of tissues. Bursting pressure is used as a quantitative measure to grade the strength of an anastomosis in vivo. This pressure has been found to increase rapidly in the early postoperative period, reaching 60% of the strength of the surrounding bowel by 3 to 4 days and 100% by 1 week.12,13 The submucosal layer is, in fact, where the tensile strength of the bowel lies, as a consequence of its high content of collagen fibers [see Figure 1]. Therefore, in constructing a hand-sewn intestinal anastomosis, it is imperative that this layer is included when extramucosal bites are taken. Collagen synthetic capacity is relatively uniform throughout the large bowel but less so in the small intestine: synthesis is significantly higher in the proximal and distal small intestine than in the midjejunum. Overall collagen synthetic capacity is somewhat less in the small intestine. Although no significant difference has been found between the strength of ileal anastomoses and that of colonic anastomoses at 4 days, colonic collagen formation is much greater in the first 48 hours.14 It is noteworthy that the synthetic response is not restricted to the anastomotic site but appears to be generalized to a significant extent.15 The presence of the visceral peritoneum on the bowel wall also has an influence on the ease with which two bowel ends can be joined. This effect is highlighted by the increased technical difficulty of joining extraperitoneal bowel ends, for example the thoracic esophagus and the rectum.

Systemic Factors

Dehiscence has been linked adversely with increasing age.16,17 This connection may be secondary to a number of factors, including the presence of comorbid conditions, malnutrition, and vitamin deficiency. A study employing an in vivo model of severe protein malnutrition, which can occur in advanced cancer, demonstrated a reduction in tissue collagen, as well as in the bursting pressure of colonic anastomoses.18 However, the introduction of parenteral nutrition has not been shown to enhance anastomotic healing.18 Several factors that are known to inhibit collagen synthesis, such as vitamin C deficiency, zinc deficiency, jaundice, and uremia, have a detrimental effect on tissue healing.16 A critical stage in collagen formation is the hydroxylation of proline to produce hydroxyproline; this process is believed to be important for maintaining the three-dimensional triple-helix conformation of mature collagen, which gives the molecule its structural strength. The amount of collagen found in a tissue is indirectly determined by measuring the amount of hydroxyproline, though no significant statistical correlation between hydroxyproline content and objective measurements of anastomotic strength has ever been demonstrated.19 Vitamin C deficiency results in impaired hydroxylation of proline and the accumulation of proline-rich, hydroxyproline-poor molecules in intracellular vacuoles.

In high doses, corticosteroids have been associated with poor healing. One study employing therapeutic doses, however, reported no difference in leakage rates was found between control subjects and those treated with steroids.17

Local Factors

Blood flow is a critical factor in tissue healing. The increased vascularity of the bowel wall is the reason why gastric and small bowel anastomoses heal more rapidly than anastomoses involving the esophagus or the large bowel. In preparing the bowel ends for anastomosis, it is imperative that the mesentery be prepared carefully and not dissected too far from the bowel edge. Mesenteric compromise, secondary to overly enthusiastic dissection or inappropriate suturing, may result in a reduction of perianastomotic blood flow. The avoidance of tension at the anastomosis is also critical; this is achieved through the appropriate mobilization of the splenic flexure. Other factors that influence blood flow at the site of anastomoses are hypovolemia and the viscosity of the blood.20 Radiation may damage the microcirculation and thereby predispose to poor healing.17

Technical Options for Fashioning Anastomoses

Over the past 160 years, numerous different materials have been used to join one bowel end to another, including substances such as catgut and stainless steel. Newer materials include monofilaments and absorbable sutures. In the past 30 years, stapling devices have been embraced enthusiastically by the surgical community. The main attraction of these devices lies in their ability to create a robust anastomosis in a relatively short space of time. Their main drawback, as with any technologically advanced device, is their cost. More important, there continues to be some controversy regarding which of the two methods of creating an anastomosis yields better clinical outcomes.21 Accordingly, the following sections review the relative merits of hand-sewn and stapled anastomoses.

Suturing: Technical Issues

Choice of Suture Material

Apart from inert substances, most foreign materials will evoke an inflammatory reaction in the human body. Surgical sutures are no exception. Studies that have examined the relative abilities of different suture materials to elicit such a reaction have found that silk has a potent ability to cause a cellular infiltrate at the site of the anastomosis that may persist as long as 6 weeks after implantation.22 Substances such as polypropylene (Prolene), catgut, and polyglycolic acid (Dexon) evoked a milder response.22,23 There is little difference between absorbable and nonabsorbable sutures with respect to the strength of the anastomosis.

The ideal suture material is one that elicits little or no inflammation while maintaining the strength of the anastomosis during the lag phase of healing. This ideal substance has yet to be discovered, but the newer generation of sutures, which include monofilament sutures and coated braided sutures, represent a substantial advance beyond silk and other multifilament materials.

Continuous versus Interrupted Sutures

Figure 2. Stitches Commonly Used

Both continuous and interrupted sutures are commonly used in fashioning intestinal anastomoses [see Figure 2]. Retrospective reviews have not shown interrupted sutures to have any advantage over continuous sutures in a single-layer anastomosis.24–26 Oxygen tension and blood flow, as discussed previously, are critical factors in anastomotic healing. Animal studies indicated that perianastomotic tissue oxygen tension was significantly lower with continuous sutures than with interrupted sutures.27 This finding was correlated with an increased anastomotic complication rate and impaired collagen synthesis and healing with continuous sutures in a rat model.28 A prospective, randomized trial compared the continuous single-layer technique for intestinal anastomosis (in the small bowel and colon) with the two-layer interrupted technique. No significant difference was seen in the leakage rate (3.1% for the single-layer technique versus 1.5% for the two-layer). The added advantages of reduced operating times and cost were observed. In a case series review that included 3,027 patients who had single-layer anastomoses done with continuous sutures, fistula rates ranged from 0 to 6.8% (mean fistula rate, 1.7%), which demonstrated that this suture technique could be safely performed.29

Single-Layer versus Double-Layer Anastomoses

Figure 3. Double-Layer End-to-End Anastomosis

The technique of double-layer anastomosis, originating from work done by Travers and Lembert,29 has been used traditionally for more than 100 years. A double-layer anastomosis consists of an inner layer of continuous or interrupted absorbable suture and outer layer of interrupted absorbable or nonabsorbable suture [see Figure 3]. The technique of single-layer anastomosis was championed because of potential advantages such as reduced operating time and lower cost. The main issue to consider, however, is safety. A randomized trial comparing the single- and double-layer techniques of anastomosis found no evidence that there was an increased risk of leak with the single-layer approach.29 However, the study groups (65 and 67 patients) were too small for any significant difference to be detected; the authors stated that a multicenter trial comprising at least 1,500 patients would be required to establish whether any such difference existed. Furthermore, a 2006 meta-analysis that addressed this issue by analyzing six trials with data from 670 patients (299 in the single-layer group, 371 in the double-layer group) concluded that there was no evidence that two-layer anastomoses yielded a lower rate of postoperative leakage than single-layer anastomoses.30

Stapling: Technical Issues

Choice of Stapler

Surgical stapling devices were first introduced by Hültl in 1908; however, they did not gain popularity at that time and for some time afterward because the early instruments were cumbersome and unreliable. The development of reliable, disposable instruments over the past 30 years has changed surgical practice dramatically. With modern devices, technical failures are rare, the staple lines are of more consistent quality, and anastomoses in difficult locations are easier to construct.

Three different types of stapler are commonly used for fashioning intestinal anastomoses. The transverse anastomosis (TA) stapler is the simplest of these. This device places two staggered rows of B-shaped staples across the bowel but does not cut it: the bowel must then be divided in a separate step. The gastrointestinal anastomosis (GIA) stapler places two double staggered rows of staples and simultaneously cuts between the double rows. The circular, or end-to-end anastomosis (EEA), stapler places a double row of staples in a circle and then cuts out the tissue within the circle of staples with a built-in cylindrical knife. All of these staplers are available in a range of lengths or diameters. Staplers may be used to create functional or true anatomic end-to-end anastomoses as well as side-to-side anastomoses. The original staplers were all designed for use in open procedures, but there are now a number of instruments (mostly of the GIA type) available for use in laparoscopic procedures. The staples themselves are all made of titanium, which causes little tissue reaction. They are not magnetic and do not cause subsequent difficulties with magnetic resonance imaging (MRI).

In a functional end-to-end anastomosis, two cut ends of bowel (either open or stapled closed) are placed side by side with their blind ends beside each other. If the bowel ends are closed, an enterotomy must be made in each loop of bowel to allow insertion of the stapler. A cutting linear (i.e., GIA) stapler is then used to fuse the two bowel walls into a single septum with two double staggered rows of staples and to create a lumen between the two bowel segments by dividing this septum between the rows. A noncutting linear (i.e., TA) stapler is then used to close the defect at the apex of the anastomosis where the GIA stapler was inserted. An alternative, and cheaper, method of closing the defect is to use a continuous suture. The cut and stapled edges of the bowel should be inspected for adequacy of hemostasis before the apex is closed. Some authors suggest cauterizing these edges to ensure hemostasis31; however, given that electrical current may be conducted along the metallic staple line to the rest of the bowel, it is probably easier and safer simply to underpin bleeding vessels with a fine absorbable suture. It is also important to offset the two inverted staple lines before closing the apex.32

Figure 4. Anastomosis with Noncutting Linear Stapler

True anatomic end-to-end stapled anastomoses may be fashioned with a linear stapler by triangulating the two cut ends and then firing the stapler three times in intersecting vectors to achieve complete closure [see Figure 4]. The potential drawback of this approach is that the staple lines are all everted. It is often easier to join two cut ends of bowel with an EEA stapler, which creates a directly apposed, inverted, stapled end-to-end anastomosis. However, circular staplers can be more difficult to use at times because of the need to invert a complete circle of full-thickness bowel wall. In addition—at least at locations other than the anus—they typically require closure of an adjacent enterotomy.

Staple Height

TA and GIA staplers are available with a variety of inserts containing several different types of staples. These inserts vary with respect to width, the height (or depth) of the closed staple, and the distance between the staples in the rows. They are designed for use in specific tissues, and it is important to choose the correct stapler insert for a given application. In particular, inserts designed for closing blood vessels should not be used on the bowel, and vice versa. With TA and EEA staplers, it is possible to vary the depth of the closed staples by altering the distance between the staples and the anvil as the instrument is closed. The safe range of closure is usually indicated by a colored or shaded area on the shaft of the instrument. Thus, if full closure would cause excessive crushing of the intervening tissues, the stapler need not be closed to its maximum extent. A 1987 comparison of anastomotic techniques that used blood flow to the divided tissues as a measure of outcome found that the best blood flow to the healing site was provided by stapled anastomoses in which the staple height was adjusted to the thickness of the bowel wall.33 The next best blood flow was provided by double-layer stapled and sutured anastomoses, followed by double-layer sutured anastomoses and tightly stapled anastomoses, in that order.

Single-Stapled versus Double-Stapled Anastomoses

To accomplish many of these anastomoses, intersecting staple lines are created. Initially, some concern was expressed about the security of these areas and about the ability of the blade in the cutting staplers to divide a double staggered row of staples. Animal studies, however, demonstrated that even though nearly all (> 90%) of the staple lines that were subsequently transected by a second staple line contained bent or cut staples, the integrity of the anastomosis was not compromised in any way, nor was healing adversely affected.34,35

Hand-sewn Versus Stapled Anastomoses

Titanium staples are ideal for tissue apposition at anastomotic sites because they provoke only a minimal inflammatory response and provide immediate strength to the cut surfaces during the weakest phase of healing. Initially, tissue eversion at the stapled anastomosis was a major concern, given that everted handsewn anastomoses had previously been shown to be inferior to inverted ones; however, the greater support and improved blood supply to the healing tissues associated with stapling tend to counteract the negative effects of eversion. In fact, one study found that bursting strength for canine colonic end-to-end anastomoses was six times greater when the procedure was performed with an EEA stapler than when it was done with interrupted Dacron sutures.36

In 1993, a randomized multicenter trial studied 440 patients who had either hand-sewn or stapled anastomoses after ileocolic resection for cancer.37 The patients were assessed both clinically and by imaging for the presence of a leak (consisting of a contrast enema at about 10 days after the operation). The overall leakage rate in the hand-sewn group was 8.3%, which compared unfavorably with the 2.8% rate in the stapled group. A possible explanation for the higher rate in the hand-sewn group might have been surgical inexperience with the variety of suture techniques used in the study (end-to-end and end-to-side with either continuous or interrupted sutures). In a study from the West of Scotland and Highland Anastomosis Study Group that included data on 732 patients at 5 centers, the rate of radiologically proven leakage was significantly higher in the sutured group (14.4% versus 5.2%); however, no difference was seen with respect to clinical leaks, morbidity, or postoperative mortality.38 A 1998 meta-analysis comparing the hand-sewn and stapled techniques of intestinal anastomosis addressed 13 trials published from 1980 to 1995.39 For colorectal anastomoses, no significant differences were seen in mortality, total leakage rate, clinical leakage rate, radiologic leakage rate, tumor recurrence rate, or incidence of wound sepsis. Strictures and technical problems, however, were more common in the stapled group. A subsequent meta-analysis that reviewed data from 955 patients with ileocolic anastomoses reported a significant reduction in the overall leakage rate and the clinical leakage rate when stapling was employed.40

Even when the anastomosis had to heal under adverse conditions (e.g., carcinomatosis, malnutrition, previous chemotherapy or radiation therapy, bowel obstruction, anemia, or leukopenia), no significant differences have been demonstrated between stapled and hand-sewn anastomoses. Stapling has, however, been shown to shorten operating time, especially for low pelvic anastomoses. Cancer recurrence rates at the site of the anastomosis have been reported to be higher or lower depending on the technique used. Certainly, suture materials engender a more pronounced cellular proliferative response than titanium staples do, particularly with full-thickness sutures as opposed to seromuscular ones, and malignant cells have been shown to adhere to suture materials.41,42

Unusual Techniques

In 1892, Murphy introduced his button, which consisted of a two-part metal stud that was designed to hold the bowel edges in apposition without suturing until adhesion had occurred.43 Thereafter, the stud was voided via the rectum. Several modifications of this technique have been described since then, primarily focusing on the composition of the rings or stents. In particular, dissolvable polyglycolic acid systems have been developed. These so-called biofragmentable anastomotic rings leave a gap of 1.5, 2.0, or 2.5 mm between the bowel ends to prevent ischemia of the anastomotic line.

The use of adhesive agents such as methyl-2-cyanoacrylate to approximate the divided ends of intestinal segments has been studied as well.44 There was only a moderate inflammatory response at the wound, which persisted for 2 to 3 weeks. Leakage rates were high, however, and many technical problems remained (e.g., how to stabilize the bowel edges while they underwent adhesion). Fibrin glues have also been employed in this setting. Although these substances are not strong enough to hold two pieces of bowel in apposition, they have been used to coat a sutured bowel anastomosis in an effort to reduce the risk of anastomotic failure. To date, no controlled clinical trials have confirmed that this approach is worthwhile.

Factors Contributing to Failure of Anastomoses

Type and Location of Anastomosis

Since the introduction of stapling, there has been an increase in the number of extremely low anterior resections being performed routinely. The literature seems to suggest that rectal anastomoses are more prone to leakage than more proximal joins are.7–9 A retrospective review of risk factors in patients undergoing rectal resection for cancer found that low anastomoses, defined as being 5 cm or less from the anal verge, were associated with a 6.5-fold increased risk of leakage when compared to anastomoses that were more than 5 cm from the anal verge.45 Another study showed that the leakage rate was increased in patients undergoing low colorectal anterior resection in the absence of a proximal stoma (the leakage rate was 17%, which fell to 6% when a stoma was present).46 The technique of total mesorectal excision (TME), which is now standard for rectal cancer operations, has reduced the local recurrence rate to 5% at 5 years.47 However, the incidence of anastomotic leakage in patients undergoing TME for low anterior resection is higher in the absence of a defunctioning stoma (25% versus 8%).48 The Rectal Cancer Trial on Defunctioning Stoma (a randomized multicenter trial) studied the outcomes of 234 patients undergoing low anterior resection with a defunctioning stoma.4 The overall leakage rate was 19.2%; however, the rate in the group that had a stoma was only 10.3%, compared with 28.0% in the group that did not. Therefore, it is safe practice to cover a low anterior resection with a defunctioning stoma.

Patient Preparation

Patients who present in the emergency setting are usually compromised in terms of hydration status, typically as a consequence of sepsis, obstruction, or a combination of the two. Preoperative fluid optimization is always necessary and may require the aid of intensivists. Before an elective procedure, the patient is assessed with regard to systemic diseases (e.g., cardiovascular, respiratory, or diabetes), and anemia is corrected. Adequate preoperative antibiotic prophylaxis has been shown to reduce the risk of postoperative infection in all types of bowel surgery and must be given at the start of the operation [see 1:1 Prevention of Postoperative Infection]. Some patients require additional steroids perioperatively [see 8:10 Endocrine Problems].

Mechanical bowel preparation (MBP) has been thought to be an essential component of colorectal surgery for more than 100 years.49,50 Emptying the bowel before elective operations on the colon was traditional until about 5 years ago, and indeed, this practice was recommended by the Association of Surgeons of Great Britain and Ireland (ASGBI) until relatively recently.51,52 The evidence for MBP was derived from observational studies showing that mechanical clearance of feces from the bowel was associated with reduced morbidity and mortality in colonic surgery.53 Proponents of MBP listed several advantages, such as reduction in intraluminal bacterial load, prevention of potential anastomotic disruption by fecal pellets and also easier handling of bowel.54 In recent years, however, there has been a shift in practice regarding the use of MBP.

A Swiss randomized clinical trial published in 2005 studied the effect of MBP on patients undergoing left-side colorectal resection with primary anastomosis.54 The anastomotic leakage rate proved to be lower in the group that did not receive MBP than in the group that did. Furthermore, the former group spent less time in hospital and exhibited less extra-abdominal morbidity (e.g., pneumonia and cardiac-related problems). These results seemed to agree with the findings of a Cochrane review.52 Two large randomized trials published in 2007 compared the outcomes of patients who underwent MBP (with either polyethylene glycol or sodium phosphate) and patients who did not.55,56 One recruited 1,431 patients undergoing elective colorectal surgery from 13 centers.55 Leakage was defined by the onset of significant symptoms and corroborated by means of imaging. The rate of leakage was 4.8% in the MBP group, which was not significantly different from the 5.4% rate in the non-MBP group. The other trial examined 1,343 patients from 21 centers and found no significant differences in outcomes (such as cardiovascular problems, general infections and surgical site infections) between the MBP group and the non-MBP group.56 However, in a a subsequent meta-analysis57 that examined data from 10 randomized trials conducted over the past 24 years, the rates of both anastomotic leakage and wound infection were significantly higher in the MBP group than in the non-MBP group (5.1 versus 2.6% and 8.2% versus 5.5%, respectively).57 Possible explanations of these findings include immune changes in the colonic mucosa that might impede wound repair.57 In view of the currently available evidence, some surgical institutions, including our own, have chosen to adopt a policy of employing fluid restriction and enemas rather than MBP before elective colorectal surgery.58

Enemas are given to patients undergoing anterior resections to ensure that fecal matter does not impede the use of stapling devices. It is advisable for patients to stop eating solid food 24 hours before the operation. Many trials have confirmed the benefits of giving IV antibiotics over the perioperative period.59 However, there is some evidence to suggest that there is an increased risk of Clostridium difficile–associated diarrhea with the use of cephalosporin, penicillin, and clindamycin.60–62 Prophylaxis of thromboembolism [see 6:6 Venous Thromboembolism] is mandatory in all patients scheduled to undergo intestinal anastomosis. There is very little evidence in the literature to indicate that thromboembolism has any direct effect on anastomotic leakage rates. However, mesenteric venous thrombosis (MVT) accounts for one-tenth of acute mesenteric ischemic events.63 The extent of thromboses is variable, with the worst outcome being mesenteric infarction necessitating urgent repeat laparotomy. Inadequate prophylaxis may increase the risk that MVT will occur postoperatively, especially in patients with other risk factors (e.g., a hypercoagulable state, previous thrombosis, or a history of smoking).

Associated Diseases and Systemic Factors

Anemia, diabetes mellitus, previous irradiation or chemotherapy, malnutrition with hypoalbuminemia, and vitamin deficiencies are all associated with poor anastomotic healing. Some of these factors can be corrected preoperatively. Malnourished patients benefit from nutritional support delivered enterally or parenterally before and after operation [see 8:22 Nutritional Support]. Well-nourished patients appear not to derive similar benefits from such support.64

Resections for Crohn disease appear to carry a significant risk of anastomotic dehiscence (12% in one prospective study) even when macroscopically normal margins are obtained.65 With the lifetime risk of repeated resections, strictureplasty has therefore become an attractive alternative to resectional management of Crohn disease even in the presence of moderately long strictures, diseased tissue, or sites of previous anastomoses. This approach allows preservation of more of the length of the small intestine.

The glucocorticoid response to injury may attenuate physiologic responses to other mediators whose combined effects could be deleterious to the organism.66 In animal experiments, wound healing, as measured by the bursting pressure of an ileal anastomosis 1 week after operation, was optimal at a plasma corticosterone level that maintained maximal nitrogen balance and corresponded to the mean corticosterone level of normal animals.67 Both supranormal and subnormal cortisol levels resulted in significantly impaired wound healing, probably through different mechanisms. It is believed that slow protein turnover is responsible for delayed anastomotic healing in adrenalectomized animals, whereas negative nitrogen metabolic balance is responsible for increased protein breakdown and delayed healing in animals with excess glucocorticoid activity.67,68

Lifestyle factors have also been associated with an increased risk of leakage. A Danish prospective study of 333 consecutive patients undergoing colorectal resection collected lifestyle information by means of a questionnaire.69 The overall leakage rate was 15.9% (53 out of 333). Smoking was found to be associated with an increased risk of anastomotic leakage (relative risk 3.18), as was alcohol consumption exceeding 35 units a week (relative risk 7.18).

Controversial Issues in Intestinal Anastomosis

Inversion versus Eversion

The question of the importance of inversion (as described by Lembert in the early 1800s) versus eversion of the anastomotic line has long been a controversial one. It has been argued that the traditional inverting methods ignore the basic principle of accurately opposing clean-cut tissues. In the late 19th century, Halsted proposed an interrupted extramucosal technique, which has since been assessed in retrospective3 and prospective65 reviews and found to have a low leakage rate (1.3% to 6.0%) in a wide variety of circumstances. A 1969 study reported greater anastomotic strength, less luminal narrowing, and less edema and inflammation with everted small intestinal anastomoses in dogs.70 Subsequent laboratory and clinical studies have not confirmed these findings and, in fact, have often yielded quite the opposite results: lower bursting pressure, slower healing, and more severe inflammation have all been associated with an everted suture line.71–73 A 1970 trial, however, conclusively demonstrated the importance of inverting the cut edges of bowel in colorectal surgery. In this study, the rate of fecal fistula formation was far higher in the group that had everted suture anastomoses (43%) than in the group with inverted suture anastomoses (8%).74

Nasogastric Decompression

Routine nasogastric decompression in patients undergoing a procedure involving an intestinal anastomosis remains controversial. In retrospective75 and prospective,76 randomized, controlled trials, routine use of a nasogastric tube conferred no significant advantage in terms of reducing the risk of anastomotic leakage. In fact, there was a trend toward an increased incidence of respiratory tract infections after routine gastric decompression.77 Nonetheless, one study found that nearly 20% of patients required insertion of a gastric tube in the early postoperative period.76 If the choice is made not to place a nasogastric tube routinely, it is important to remain alert to the potential for gastric dilatation, which can develop suddenly and without warning.

Abdominal Drains

Fecal contamination from bowel surgery is a dreaded complication. Peritoneal drainage is of the subject of considerable controversy, with two basic schools of thought dominating.78 The first school accepts the possibility that drains may help with diagnosis by serving as an early warning system for either anastomotic leakage or bleeding and makes the point that evacuating blood and serous fluid will reduce the risk of abscesses. The other school is skeptical about the benefits, arguing that drains may irritate the peritoneum, thus increasing the production of serous fluid, and may provide a route whereby microbes can enter the peritoneal cavity. In addition, there is a potential risk that the drain may physically impede the movement of the omentum and adjacent organs and consequently may hinder the body’s innate ability to wall off any possible infection. Finally, it is thought that drains are at high risk for blockage.

Even before World War I, the old dictum ” when in doubt, drain” was called into question by Yates, who wrote that the peritoneal cavity could not be effectively drained because of adhesions and rapid sealing of the drain tract.79 Six decades later, one study showed a dramatic increase in the incidence of anastomotic dehiscence (from 15% to 55%) after the placement of perianastomotic drains in dogs.80 This increase was associated with a significant increase in mortality. A 1999 study of pelvic drainage after a rectal or anal anastomosis showed that prophylactic drainage did not improve outcome or reduce complications.81 Yet another study reported the severe inflammatory reaction caused by drains at anastomoses.82

Despite these findings, many surgeons elect to place an intra-abdominal drain in the pelvis after an anterior resection or a coloanal anastomosis because of the higher than usual risk that a fluid collection will develop. Drainage is rarely helpful, or indeed easy, after a gastric or small bowel anastomosis. Drains are indicated, however, after emergency operations for peritonitis or trauma in which it was necessary to close or anastomose damaged or inflamed bowel.

Operative Techniques for Selected Anastomoses

The following section covers the essential preliminary steps before a bowel anastomosis and then describes three generic operations involving the small and large bowel. These procedures illustrate many of the general principles previously discussed (see above).

Patient Positioning and Incision

Patients must be positioned on the operating table in a manner that is appropriate for the planned operation. Most abdominal operations are performed through a midline incision of adequate length with the patient supine. For pelvic procedures, the patient is placed in the lithotomy position to allow access to the abdomen and the anus; care must be taken to position the legs and feet in the stirrups correctly, without excessive flexion or abduction and with sufficient padding to prevent pressure ulceration, thrombosis, and neurapraxia. For esophageal procedures, the patient is positioned lying on the appropriate side, and the incision of choice is a lateral thoracotomy [see 4:7 Open Esophageal Procedures]. Occasionally, the patient must be shifted to a different position during the course of an operation. Gravity can be useful for moving structures out of the way. Accordingly, it is often helpful to alter the axis of the operating table. For example, a 30° head-down or Trendelenburg position facilitates pelvic operations.

Exposure, Mobilization, and Dissection

Access is a critical determinant of the ease with which an operative procedure can be carried out. Accordingly, the incision must be made in such a way as to allow adequate exposure of the operating field. The lateral aspects of the field can be controlled by using a suitable retractor, which can be attached to the operating table. This measure allows more efficient use of operating assistants and space. Often, the colorectal surgeon is faced with operating in confined spaces, such as the pelvis. Therefore, it is important to be able to compartmentalize the abdomen. This can be achieved in a number of ways. The small bowel can be extremely difficult to handle and therefore is commonly packed away by placing wet gauze. The next stage involves bringing the bowel to the surface. In the absence of adhesions or tethering caused by disease, the small bowel is usually sufficiently mobile to allow the relevant segment to be brought out of the abdomen. Doing so makes the operation easier and allows the remainder of the bowel to be kept warm and tension free inside the abdominal cavity. Sometimes, the transverse colon and the sigmoid colon are mobile enough to be brought to the surface. More commonly, however, as with the other sections of the large bowel, the peritoneum must be divided along the lateral border of the colon and the retroperitoneal structures reflected posteriorly. Tension is rarely a problem with small bowel anastomoses, but with colonic or esophageal anastomoses, it is absolutely vital that the two ends of bowel to be joined lie together easily. For a large bowel anastomosis, this means that the splenic flexure or the hepatic flexure—or, sometimes, both—must be adequately mobilized.

Classically, the tissues around the bowel are divided with a scissors, whereas the mesentery is divided between clamps and tied with a suitable thread. Recognized tissue planes are separated by means of blunt dissection with either the fingers or a swab. Minor bleeding points are occluded with a coagulating electrocautery, though this approach is often relatively ineffective on mesenteric or omental vessels. The disadvantages of this dissection technique are that oozing from raw surfaces can be a nuisance and that the tissues beyond a tie are often bulky and leave dead tissue within the body that may act as a focus for infection and adhesions. Newer methods of dissection that make use of the ultrasonic scalpel or the bloodless bipolar electrocautery prevent these problems by coagulating a small section of tissue between the jaws of the instrument and simultaneously occluding all blood vessels up to a certain size within the tissues. Consequently, bleeding is reduced, fewer (or no) ties are needed, and only a small quantity of dead tissue results at each point. Becoming skilled in the use of these instruments often takes a little time, but the time is well spent, in that it is now possible to perform an intestinal resection without resort to a single tie.

Bowel Resection

The precise techniques involved in resecting specific bowel segments will not be discussed in great detail here. (Colonic resection, for example, is described elsewhere [see 5:34 Segmental Colon Resection].) The following discussion outlines only the general principles.


The segment of bowel to be removed must be isolated with an adequate resection margin. To this end, all surrounding adhesions are divided. Next, the mesentery is divided. The key consideration in this step is to preserve the blood supply to the two remaining ends of bowel while still achieving adequate excision of the diseased bowel. This is more easily accomplished in the small bowel than in the large bowel, thanks to the ample blood supply of the former; even so, transillumination of the mesentery and careful division of the vascular arcade are vital. In the colon, the surrounding fat and the appendices epiploicae should be cleared from the remaining bowel ends so that subsequent suture placement is straightforward.

Care should be taken to avoid two common problems. First, ties placed close to the bowel can bunch tissues excessively and thereby cause angulation or distortion of the free edge of the intestine, which can make the anastomosis difficult and threaten the blood supply. Second, because mesenteric vessels are usually tied very close to their ends, the arteries sometimes slip back beyond the ties. Such slippage results in a hematoma within the leaves of the mesentery, which can itself threaten the viability of the bowel. Generally, the bleeding vessel can be secured with a fine stitch; sometimes, however, a limited further bowel resection is the only safe course of action. Both of these problems can be avoided by using the ultrasonic scalpel or the bipolar coagulating electrocautery.

Division of Bowel

If staplers are not available, the bowel segment to be removed is isolated between noncrushing clamps placed across the intestinal lumen some distance away from the resection margin so as to limit the amount of bowel contents that can escape into the wound. Crushing clamps are then placed on the specimen side of the diseased segment at the point of the resection, and the bowel is divided with a knife just proximal and distal to the clamps. Thus, the lumen of the diseased segment is never open within the abdominal wound. Even so, the contents of the bowel between the open ends and the non-crushing clamps can leak into the wound. To minimize this problem, it is usual to isolate the working area with abdominal packs, which are sometimes soaked in an antiseptic (e.g., povidone-iodine).

One advantage of using staplers for an anastomosis is that in most instances, division of the bowel can be accomplished without opening the lumen. A linear cutting stapler (e.g., GIA) transects the bowel and seals the two cut ends simultaneously. Unfortunately, in the pelvis, it is usually necessary to employ an angulated noncutting linear stapler (e.g., TA) so as to obtain as much length as possible distal to the lesion. The proximal rectum is then clamped with a crushing bowel clamp, and a long knife is used to transect the rectum above the staple line. Even so, there remains the potential for leakage of a small amount of fecal material, which must then be suctioned away.

Simple Bowel Closure

There are occasions where the bowel damage requires simple repair rather than a formal resection, as in the case of a perforated duodenal ulcer or an iatrogenic injury sustained during adhesiolysis. Such holes may be closed in a number of ways. Most surgeons would use a double layer of absorbable sutures. Special mention should be made of the technique of strictureplasty, which is employed for a number of benign small bowel strictures (especially those resulting from Crohn disease) as a means of avoiding small bowel resection and anastomoses. In this procedure, the bowel is opened longitudinally and closed transversely with a single layer of 2-0 polyglycolic acid sutures in a Connell stitch. Excellent functional results have been achieved with this technique despite its reputation for fistula formation, which is associated with Crohn disease.

Single-layer Sutured Extramucosal side-to-side Enteroenterostomy

Figure 5. Single-Layer Sutured Extramucosal Side-to-Side Enteroenterostomy

A side-to-side anastomosis [see Figure 5] may be performed when no resection is done, as a bypass procedure (e.g., a gastroenterostomy); after a small bowel resection; when there is a discrepancy in the diameter of the two ends to be anastomosed (e.g., an ileocolic anastomosis after a right hemicolectomy);or when the anatomy is such that the most tension-free position for the anastomosis is with the two bowel segments parallel (as in a Finney strictureplasty).

Two stay sutures of 3-0 polyglycolic acid are placed approximately 8 cm apart on the inner aspect of the antimesenteric border. A 5 cm enterotomy is made on each loop with an electrocautery or a blade on the inner aspect of the antimesenteric border. If an electrocautery is used, care must be taken not to injure the mucosa of the posterior wall during this maneuver; placement of a hemostat into the enterotomy to lift the anterior wall usually prevents this problem. Hemostasis of the cut edges is ensured, and the remaining enteric contents are gently suctioned out. A swab soaked in povidone-iodine may be used at this point to cleanse the lumen of the bowel in the perianastomotic region.

A full-length seromuscular and submucosal stitch of 4-0 polyglycolic acid is placed and tied on the inside approximately 5 to 10 mm from the far end of the enterotomies. The stitch is not passed through the mucosa: to do so would add no strength to the anastomosis and would hinder epithelialization by rendering the tissue ischemic. A hemostat is placed on the short end of the tied suture, and the assistant applies continuous gentle tension to the long end of the suture. An over-and-over stitch is started in the direction of the surgeon; small bites are taken, and proper inversion of the suture line is ensured with each pass through tissue. When the proximal ends of the enterostomies are reached, this so-called baseball stitch is continued almost completely around to the anterior wall of the anastomosis. A single Connell stitch may be used to invert this anterior layer.

Another full-length seromuscular and submucosal suture of 4-0 polyglycolic acid is then inserted and tied at the same location in the posterior wall as the first. If the two sutures are placed close enough together, the short ends need not be tied together and may simply be cut off. The remainder of the posterior wall is sewn away from the surgeon in the same manner as the portion already sewn, and the corners are approximated with the baseball stitch. The anterior wall is then completed with this second suture, either with the Connell stitch or with an over-and-over stitch with the assistant inverting the edges before applying tension to the previous stitch.

When the defect is completely closed, the two sutures are tied across the anastomotic line. The stay sutures are removed, and the anastomosis is carefully inspected. Often, there is no mesenteric defect to close in a side-to-side anastomosis, but if there is one, it should be approximated at this point with continuous or interrupted absorbable sutures, with care taken not to injure the vascular supply to the anastomosis.

Double-layer Sutured end-to-side Enterocolostomy

Figure 6. Double-Layer Sutured End-to-Side Enterocolostomy

In this procedure, the end of the ileum is joined to the side of the transverse colon [see Figure 6]. The distal colon is divided with a cutting stapler so that a blind end is left. Some surgeons underpin or bury this staple line, though this practice is probably unnecessary. The proximal cut end of the intestine is similarly closed either with staples after division with a cutting linear stapler or with a crushing bowel clamp. This proximal end is brought into apposition with the side of the distal bowel segment at a point no farther than 2.5 to 5 cm from the blind end of the distal segment; this proximity to the cut end is important for prevention of the blind loop syndrome.

Stay sutures of 3-0 polyglycolic acid are placed between the serosa of the proximal limb, about 10 to 15 mm from the clamp, and the serosa of the distal limb. Interrupted seromuscular sutures of 3-0 polyglycolic acid are then placed between these stay sutures, spaced about three to six to the centimeter. These stitches may be tied sequentially or snapped and tied once they are all in place. It is crucial not to apply excessive tension, which could cut the seromuscular layer or render it ischemic. Suction is then readied. The staple line or crushed tissue on the proximal limb is cut off with a coagulating electrocautery or a knife; this maneuver opens the lumen of the proximal limb. All residual intestinal content is gently suctioned.

An enterotomy or colotomy is created on the distal limb opposite the open lumen of the proximal bowel. A full-thickness suture of 3-0 polyglycolic acid is inserted in the posterior wall at a point close to the far end of the enterotomy and run in an over-and-over stitch back toward the surgeon. The corner is rounded with the baseball stitch, and when the anterior wall is reached, the Connell stitch is used. A second full-length 3-0 suture is started at the same point on the posterior wall as the first, and the short ends of the two sutures are tied together and cut. This second suture is then run away from the surgeon to complete the posterior wall, and the anterior wall is completed with the Connell stitch. The two sutures are then tied across the anastomotic line.

A second series of interrupted seromuscular stitches is then placed anteriorly in the same fashion as the seromuscular stitches placed in the posterior wall. It is important not to narrow either lumen excessively by imbricating too much of the bowel wall into this second layer. The lumen of the anastomosis is palpated to confirm patency, and the mesenteric defect is closed if possible with either continuous or interrupted absorbable sutures. The integrity of the join can be tested by injecting 20 mL of saline into the lumen of the bowel, which is then placed under gentle pressure by simultaneously finger-clamping the bowel a couple of centimeters distal to the anastomosis and a couple of centimeters proximally.

Double-stapled end-to-end Coloanal Anastomosis

Figure 7. Double-Stapled End-to-End Coloanal Anastomosis

Resection of the distal sigmoid colon and the rectum is a common procedure. In the past, it often resulted in a permanent colostomy because of the technical difficulties associated with a hand-sewn anastomosis deep in the pelvis. The development of circular staplers reduced the technical difficulty of the operation and made possible anastomoses as far down as the anus [see Figure 7].

Proper preparation of the patient and the bowel is essential before resection of the rectum. The patient is placed in the lithotomy position with the head tilted down, and the small bowel is packed away in the upper abdomen. This positioning gives the surgeon the best access to the pelvis. The splenic flexure and all of the distal large bowel are fully mobilized along with the rectum. The proximal resection margin is determined and cleared of serosal fat, and the bowel is divided either with a GIA stapler or between crushing bowel clamps. An angled TA stapler is fired across the distal rectal resection margin, and another bowel clamp is placed proximal to it. The rectum is divided with a long-handled knife, with care taken to avoid plunging the blade into the pelvic sidewall, which could cause significant neurovascular damage. The specimen is removed and the stapler withdrawn. Adequate pelvic hemostasis is ensured.

Once the surgeon is satisfied that the bowel is sufficiently mobilized, a noncrushing bowel clamp is placed on the colon 10 to 15 cm proximal to the margin, and the crushing clamp is removed. At this stage, it is usual to create an 8 to 10 cm colonic J pouch; this measure typically yields a substantially improved functional outcome, especially in the early postoperative period in older patients.83 A whip-stitch (or purse-string suture) of 2-0 polypropylene is placed around the colotomy, and the anvil from the appropriately sized curved EEA stapler is inserted into the open end and secured in place by tying the suture [see Figure 7].The proximal bowel clamp is removed. The assistant—who may also, if desired, gently wash out the rectal stump with a dilute povidone-iodine solution—performs a digital rectal examination.

The stapler, with its trocar attachment in place, is then inserted into the anus under the careful guidance of the surgeon. The pointed shaft is brought out through or adjacent to the linear staple line, and the sharp point is removed. The peg from the anvil in the proximal colon is snapped into the protruding shaft of the stapler, and the two edges are slowly brought together. The colonic mesentery must not be twisted, and the ends must come together without any tension whatsoever. The stapler is fired. In some types of stapling guns, a crunching sound is heard. The anvil is then loosened the appropriate amount, and the entire mechanism is withdrawn through the anus. Finally, the proximal and distal rings of tissue, which remain on the stapler, are carefully inspected to confirm circumferential closure of the staple line.

The pelvis is then filled with body-temperature saline, and a Toomey or bladder syringe is used to insufflate the neorectum with air. The surgeon watches for bubbling in the pelvis as a sign of leakage from the anastomosis. If there is a leak, additional soluble sutures must be placed to close the defect and another air test performed. A rectal tube may then be inserted by the assistant or may be placed at the end of the procedure. When the anastomosis is very low or there is some concern about healing, a drain may be placed in the pelvis behind the staple line; however, as noted [see Controversial Issues in Intestinal Anastomosis, above], this practice has not been shown to be beneficial and may in fact impair healing.

As noted (see above), a 1998 meta-analysis by Macrae demonstrated an association between stapled anastomoses and an increased incidence of colorectal anastomotic strictures, in comparison with hand-sewn anastomoses.39 The cause of this association remains uncertain. Most of these patients were asymptomatic and in those who required treatment, simple dilatation was sufficient to rectify the problem. A 2007 Cochrane review did not find the risk of anastomotic stricture to be increased in patients undergoing ileocolic resection with a linear cutting stapler.40


Any anastomosis, no matter how technically sound on creation, may fail. The limiting factor may be the tissue or the resulting inflammatory sequelae that follow closure of the abdominal wall. Therefore, as for ultralow anterior resections, it is important to recognize any potential risk factors and take the measures necessary to prevent any harm that may ensue if leakage results. This may mean giving the patient a temporary stoma. Even if fecal diversion has been carried out, it is important to keep watching closely for any signs of failure and to take prompt action if such signs appear.

Anastomotic failure rates have improved over the past two centuries, and postoperative morbidity and mortality have decreased accordingly. These beneficial developments can be attributed to a combination of factors, such as better appreciation of the principles of healing, improved anesthesia, appropriate antibiotic prophylaxis, and enhanced postoperative monitoring. Currently, with the emergence of laparoscopic colorectal surgery, it is essential that the surgeon continues to practice the same principles of creating a join—good apposition of the edges, without tension and with an optimal blood supply—just as for open colorectal surgery.

Figures 1 through 7 adapted from Tom Moore.

Portions of this chapter are based on previous iterations written for ACS Surgery by Zane Cohen, MD, Barry Sullivan, MD, and Julian Britton. The authors wish to thank Drs. Cohen, Sullivan, and Britton


Liane S. Feldman, MD, FACS

Assistant Professor, Department of Surgery
McGill University Faculty of Medicine
Staff, Departments of Videoendoscopic Surgery and Surgery
McGill University Hospital Centre

Marvin J. Wexler, MD, FACS

Professor, Departments of Surgery and Oncology
McGill University Faculty of Medicine
Senior Surgeon, Department of Surgery
Royal Victoria Hospital

Shannon A. Fraser, MD

Minimally Invasive Surgery Fellow
McGill University Faculty of Medicine

Since its first description in 1990,1 laparoscopic inguinal herniorrhaphy has shown a great deal of promise; however, concurrently with its development, open anterior herniorrhaphy has evolved into a tension-free, mesh repair that is easily performed with the patient under local anesthesia and that is also associated with rapid recovery and low recurrence rates [see 5:27 Open Hernia Repair].2,3 Thus, the key question about laparoscopic inguinal hernia repair at present is whether it provides a significant advantage over the tension-free open repair now in use.

The two most common techniques for laparoscopic inguinal hernia repair involve the insertion of mesh into the preperitoneal space; one makes use of a transabdominal preperitoneal (TAPP) approach, the other a totally extraperitoneal (TEP) approach. Both approaches would appear to offer potential advantages, such as reduced postoperative pain, shortened recovery, quicker and more accurate assessment and repair of bilateral groin hernias simultaneously, and, in the case of recurrent hernia, avoidance of previously dissected and technically difficult scarred areas. In practice, however, the advantages are not invariably realized; a laparoscopic approach is not always minimally invasive, and various disadvantages accrue from the current requirement for general anesthesia, the need to traverse the abdominal cavity in the TAPP technique, and the increase in operating room time and costs.4

Meticulous attention to surgical technique is essential. Because surgeons may be unfamiliar with inguinal anatomy as viewed from inside the abdomen and because the potential for complication necessitating laparotomy is increased with the laparoscopic approach, surgeons must be proficient in laparoscopic techniques and must have a precise knowledge of anatomic relations in the region of the groin as seen from the peritoneal surface.

Since the late 1990s, laparoscopic video techniques have also been increasingly applied to the repair of incisional hernias.5–9 Laparoscopic repair of large incisional hernias resembles open repair in that mesh is inserted to cover the defect in the abdominal wall fascia [see 5:27 Open Hernia Repair]. A laparoscopic approach is theoretically attractive because an open approach usually necessitates a large incision as well as extensive and tedious wide dissection to expose the abdominal wall defect, resulting in considerable postoperative pain and a risk of wound complications—problems that a laparoscopic approach to the defect from within might minimize.

It may be many more years before the true safety and efficacy of laparoscopic herniorrhaphy can be determined and the correct indications for its use established. In the meantime, every repair performed should be subjected to careful classification, documentation, and quality-of-life assessment. Surgeons should not perform laparoscopic herniorrhaphy simply because it is relatively new or potentially economical; they should perform it only when convinced that it is anatomically and physiologically correct and logical.

In what follows, we discuss laparoscopic repair of both inguinal and incisional hernias. In addition to describing current operative techniques, we address inguinal surgical anatomy, preoperative planning, and complications. Finally, we review selected trials measuring the results of laparoscopic repair against those of open repair and comparing the outcomes of TAPP repair with those of TEP repair.

Laparoscopic Inguinal Hernia Repair

Anatomic Considerations

To most surgeons, inguinal anatomy as viewed through the laparoscope appears unfamiliar. This is particularly true for the TEP approach, in which the preperitoneal space must be developed. The surgical perspective on the pelvic anatomy from the intraperitoneal view has been elegantly described by Skandalakis and coworkers10 and has been elegantly demonstrated in cadaver dissections by Spaw and colleagues,11 whose work forms the basis of the descriptions we present in this chapter. Excellent descriptions of the preperitoneal space by Wantz12 and Condon13 are also worthy of review.

Figure 1. Anatomy of groin region

During laparoscopic herniorrhaphy, a number of structures that are usually visible during open herniorrhaphy (e.g. the inguinal ligament, the pubic tubercle, the lacunar ligament, and the ilioinguinal and iliohypogastric nerves) are not seen initially. Conversely, a number of structures that are visible only after significant dissection in the open approach are easily viewed through the laparoscope [see Figure 1]. Identification of the iliopubic tract, Cooper’s ligament, and the transversus abdominis arch is mandatory to ensure proper coverage by the prosthetic material used in the repair.

Figure 2. Laparoscopic view of anatomy of the left groin

During a TAPP repair, four important landmarks should be seen at initial laparoscopic inspection of the inguinal region [see Figure 2]: the spermatic vessels, the obliterated umbilical artery (also referred to as the medial umbilical ligament or the bladder ligament), the inferior epigastric vessels (also referred to as the lateral umbilical ligament), and the external iliac vessels. During a TEP repair, once the preperitoneal space has been established, the first easily identified landmarks are Cooper’s ligament and the inferior epigastric vessels; identification of these structures is helpful in guiding early dissection.

Spermatic Vessels

The testicular artery and vein descend from the retroperitoneum, travel directly over and slightly lateral to the external iliac artery, and enter the internal spermatic ring posteriorly. These vessels are covered only by the peritoneum and are usually well visualized as flat structures in the abdominal cavity that assume a cordlike appearance when joined by the vas deferens immediately before entering the internal spermatic ring. If no hernia is present, a mere dimple will be seen [see Figure 2]. In the TAPP approach, an indirect hernia, if present, will be immediately apparent and will have an obvious opening. In the TEP approach, the indirect sac will be seen later, after dissection of the tissue on the anterior abdominal wall. The vas deferens is best identified where it joins the spermatic vessels. From there, the vas can be traced back medially as it courses over the pelvic brim and falls into the pelvis and behind the bladder. There is a small artery that runs with the vas deferens and is not well seen or known. It is white and cordlike in appearance and can usually be seen just beneath the peritoneum.

Obliterated Umbilical Artery

The obliterated umbilical artery is an unfamiliar but sometimes prominent structure that is seen in the TAPP approach. It courses along the anterior abdominal wall toward the umbilicus, often with an apparent mesentery. It is most prominent in the region of the medial inguinal space. This ligament is most readily identified when the umbilical laparoscope is directed toward the pelvic midline, where the ligament’s bilateral structure is best seen as it is oriented toward the umbilicus. Medial retraction of this structure is usually necessary for full exposure of the medial aspect of the inguinal canal.

Inferior Epigastric Vessels

Figure 3. TEP approach

The inferior epigastric artery and vein lie in the medial aspect of the internal inguinal ring and ascend the inferior surface of the rectus abdominis. In the TAPP approach, these vessels may be difficult to visualize, particularly in obese patients. They are best identified by locating the internal inguinal ring at the junction of the vas deferens and the testicular artery and vein. At this location, the vessels exit the medial margin of the internal ring. However, they can quickly fade from view as they travel superiorly and medially along the anterior abdominal wall. In the TEP approach, early identification of the inferior epigastric vessels helps guide lateral dissection and identification of the internal ring [see Figure 3]. However, dissection in the incorrect plane when the preperitoneal space is initially established may strip these vessels off the abdominal wall.

External Iliac Vessels

Figure 4. Courses of genitofemoral and ilioinguinal nerves

The lateral spermatic vessels and the medial vas deferens merge at the internal inguinal ring and enter the inguinal canal, where they form the apex of the so-called triangle of doom [see Figure 4, part a, insert]. Beneath this triangle lie the external iliac artery and the external iliac vein. More laterally, the femoral nerve can be found. The external iliac vessels are often difficult to visualize, though in an elderly patient, a calcified pulsating artery may be prominent. Extreme care must be taken not to dissect within the triangle of doom, because such dissection can result in serious bleeding.

Cooper’s Ligament

Cooper’s ligament is a condensation of the transversalis fascia and the periosteum of the superior pubic ramus lateral to the pubic tubercle. It can be seen only in the preperitoneal space and is the first landmark that should be identified during a TEP repair. In the initial stages of a TAPP repair, with the peritoneum intact, it is often easier to palpate the ligament than to see it, but once the ligament has been identified and cleaned, its glistening white fibers are apparent. Care must be taken during dissection to avoid the tiny branches of the obturator vein that often run along the ligament’s surface. The iliopubic tract inserts into the superior ramus of the pubis just lateral to Cooper’s ligament, blending into it.

Internal Inguinal Ring

In the TAPP approach, the internal inguinal ring is normally identified by a slight indentation of the peritoneum at the junction of the vas deferens and the spermatic vessels. When an indirect hernia is present, however, a true ring or opening is easily identified, and by rotating a 30° laparoscope, the surgeon can look directly into the hernial sac or insert the laparoscope into the sac, which often allows the external inguinal ring to be identified more medially. An indirect hernial sac lies anterior and lateral to the spermatic cord at this level, as opposed to the familiar medial cord position seen in the classic exterior groin approach to open herniorrhaphy. The medial border of the internal inguinal ring is formed by the transversalis fascia and the inferior epigastric vessels. The inferior border is formed by the iliopubic tract, a distinct structure that is the internal counterpart of the inguinal ligament. Anteriorly, the internal inguinal ring is bordered by the transversus abdominis arch, which passes laterally over the internal ring and forms a very well defined visible edge. The layers of the abdominal wall constituting the lateral border of the internal inguinal ring appear the same as when viewed from the exterior approach, and this border, like all margins of the internal inguinal ring, is visible only when an indirect hernia is present.

Iliopubic Tract

The iliopubic tract originates laterally from the anterior superior spine of the ilium and courses medially, forming the inferior margin of the internal inguinal ring and the roof of the femoral canal before inserting medially into the superior pubic ramus. This tract is formed by the condensation of the transversalis fascia with the most inferior portion of the transversus abdominis muscle and aponeurosis, and it is usually sturdy along its entire course. All inguinal hernia defects lie above the iliopubic tract, either anterior or superior to it. Conversely, femoral hernias occur below the tract, either posterior or inferior to it. Fibers of the ilio-pubic tract extend into Cooper’s ligament medially, where they become the medial margin of the femoral canal. The iliopubic tract is frequently confused with the inguinal ligament. This ligament, though nearby, is part of the superficial musculoaponeurotic layer, which is not seen laparoscopically, whereas the iliopubic tract is part of the deep layer.

Femoral Canal

The femoral canal is seen only in the presence of a femoral hernia in the most medial aspect of the femoral triangle. The anterior and medial borders are formed by the iliopubic tract, the posterior border is formed by the pectineal fascia, and the lateral border is formed by the femoral sheath and vein.

Trapezoid of Disaster

Another area worthy of careful attention is the so-called trapezoid of disaster, containing the genitofemoral, ilioinguinal, iliohypogastric, and lateral cutaneous nerves of the thigh, which innervate the spermatic cord, the testicle, the scrotum, and the upper and lateral thigh, respectively. A detailed knowledge of the anatomic courses of the nerves and careful avoidance of these structures during dissection are essential [see Figure 4].

Genitofemoral nerve The genitofemoral nerve arises from the first and second lumbar nerves; pierces the psoas muscle and fascia at its medial border opposite L3 or L4; descends under the peritoneum, on the psoas major; and divides into a medial genital and a lateral femoral branch. The femoral branch descends lateral to the external iliac artery and spermatic cord, passing posteroinferior to the iliopubic tract and into the femoral sheath to supply the skin over the femoral triangle. The genital branch crosses the lower end of the external iliac artery and enters the inguinal canal through the internal inguinal ring with the testicular vessels. This branch supplies the coverings of the spermatic cord down to the skin of the scrotum. The genitofemoral nerve is the most visible of the cutaneous nerves and is sometimes confused with the testicular vessels if the latter are not well appreciated in their more medial position.

Ilioinguinal and iliohypogastric nerves When dissected from the anterior position, the ilioinguinal and iliohypogastric nerves lie between the external oblique and the internal oblique muscles above the internal inguinal ring and descend with the spermatic cord. In the abdomen, the ilioinguinal and iliohypogastric nerves arise from the 12th thoracic and first lumbar nerve roots, are more laterally located, and run subperitoneally, emerging from the lateral psoas border to pierce the transversus abdominis near the iliac crest, then piercing and coursing between the internal oblique and the external oblique muscles close to the internal inguinal ring. Aberrant branches sometimes descend with the genital nerve. The ilioinguinal nerve supplies a small cutaneous area near the external genitals.

Lateral cutaneous nerve of thigh Supplying the front and lateral aspect of the thigh, the lateral cutaneous nerve of the thigh arises from the second and third lumbar nerves and emerges at the lateral border of the psoas. There, it descends deep to the peritoneum on the iliac muscle and only comes to lie in a superficial position 3 cm below the anterosuperior iliac spine.

Preoperative Evaluation

History and Physical Examination

Preoperative assessment is necessary to determine whether a patient is a suitable candidate for laparoscopic herniorrhaphy. A careful surgical history, including both previous hernia repairs and other procedures (particularly those involving the lower abdomen), should be elicited. A cardiovascular history should also be obtained and risk factors for general anesthesia determined.

Physical examination should confirm the presence of an inguinal hernia. If the patient reports a history of a bulge but no hernia is felt on physical examination, an occult hernia may be presumed. Ultrasonography may be helpful for distinguishing an incarcerated groin hernia from other causes of inguinal swelling (e.g., lymphadenopathy or venous varix).

Selection of Patients

Indications With the evolution of the open anterior approach to tension-free prosthetic mesh repair, determining which patients will benefit significantly from laparoscopic herniorrhaphy has become increasingly important. In 2003, the Cochrane Database of Systemic Reviews4 published an update to its original 2000 report on laparoscopic versus open techniques for hernia repair, which indicated that whereas laparoscopic repairs (TAPP and TEP) generally took longer and had a higher rate of more serious complications (bowel, bladder, and vascular injuries) than open repairs, they also were associated with shorter recovery times and a lower incidence of persistent pain and numbness.14 Reduced hernia recurrence was related to the use of mesh rather than to operative technique. The overall risk of recurrence with the laparoscopic approach may be related to the surgeon’s level of experience.15We believe that patients are best served when a surgeon has several approaches at his or her command that can be applied to and, if necessary, modified for individual circumstances.

Currently, we treat primary unilateral hernias with an open anterior mesh repair, preferably with the patient under local or regional anesthesia; possible exceptions include manual laborers and athletes who desire a rapid return to vigorous physical activity. We generally reserve laparoscopic inguinal herniorrhaphy for the following clinical situations:

  1. Recurrent hernia after previous anterior repair. In such cases, a laparoscopic approach allows the surgeon to avoid the scar tissue and distorted anatomy present in the anterior abdominal wall by performing the repair through unviolated tissue, thereby potentially reducing the risk of damage to the vas deferens or the testicular vessels. This is especially true when mesh has previously been placed anteriorly.
  2. Bilateral hernias or a unilateral hernia when the presence of a contralateral hernia is strongly suspected. In such cases, a laparoscopic approach allows the surgeon to repair the two hernias simultaneously (and perhaps more rapidly) without having to make additional incisions.
  3. Repair of an inguinal hernia concurrent with another laparoscopic procedure, provided that there is no contamination of the peritoneal cavity.


The choice of laparoscopic technique depends on the patient’s history and on the type of hernia present. In general, we favor TEP repair for most patients because it does not involve entry into the abdominal cavity. The TEP approach reduces the risk of complications affecting intra-abdominal structures, theoretically decreases the risk of adhesions, maintains an intact peritoneal layer between the mesh and the intra-abdominal contents, and allows the mesh to be placed without the use of fixation. However, it has a steeper learning curve than TAPP does, and it may not be feasible in patients who have undergone lower abdominal procedures.


ContraindicationsWe do not treat acutely incarcerated hernias laparoscopically. In patients to whom general anesthesia may pose an increased risk, we prefer open anterior repair using local or regional anesthesia. In infants and young children with indirect hernias, for whom repair of the posterior canal wall is unnecessary, we recommend high ligation of the sac via the anterior approach.

Previous lower abdominal surgery, though not an absolute contraindication, may make laparoscopic dissection difficult. In particular, with respect to TEP repair, previous lower abdominal wall incisions may make it impossible to safely separate the peritoneum from the abdominal parietes for entry into the extraperitoneal plane, and conversion to a TAPP repair or an open repair may be required. Previous surgery in the retropubic space of Retzius, as in prostatic procedures, is a relative contraindication that is associated with an increased risk of bladder injury16 and other complications.17 Similarly, previous pelvic irradiation may preclude safe dissection of the peritoneum from the abdominal wall.18


Operative Planning


General anesthesia is administered routinely. Prophylactic antibiotics are unnecessary.19 The patient is instructed to void before surgery, which renders bladder catheterization unnecessary.

Patient Positioning

The patient is placed in the supine position with both arms tucked against the sides. The anesthesia screen is placed as far toward the head of the table as possible to allow the surgeon a wide range of mobility with the laparoscope. The skin is prepared and draped so as to allow exposure of the entire lower abdomen, the genital region, and the upper thighs because manipulation of the hernial sac and the scrotum may be necessary. After the laparoscope has been introduced, the patient is placed in a deep Trendelenburg position so that the viscera will fall away from the inguinal areas. Further bowel manipulation is rarely necessary, except to reduce hernial contents. Rotation of the table to elevate the side of the hernia can provide additional exposure, if necessary. A single video monitor is placed at the foot of the bed, directly facing the patient’s head.

Figure 5. Possible OR setup

The surgeon usually begins the repair while standing on the side contralateral to the defect; the assistant surgeon stands opposite the surgeon, and the nurse stands on the ipsilateral side [see Figure 5]. A two-handed operating technique provides a distinct technical advantage.


Because inguinal hernias occur in the anterior abdominal wall, visualization through the umbilicus requires that the laparoscope be angled close to the horizontal plane. The view is paralleled anteriorly by the surface of the lower abdominal wall, which may make visualization with a 0° laparoscope difficult. In addition, an indirect hernia is a three-dimensional tubular defect that can be well visualized in its entirety only with an angled lens. For these reasons, we recommend routine use of an oblique, forward-viewing 30° or 45° laparoscope. Excellent laparoscopes are currently available in 10 mm and 5 mm sizes. For the TAPP repair, one 10/12 mm and two 5 mm trocars are used. The dissection is performed with a dissector and scissors, to which an electrocautery may be attached. For mesh fixation, we currently use a spiral tacker (e.g., TACKER; U.S. Surgical Inc., Norwalk, Connecticut).

In a TEP repair, besides the equipment needed for a TAPP repair, a balloon-tipped blunt trocar is used to gain access to the preperitoneal space. We usually develop the preperitoneal space with a balloon system [see Operative Technique, Totally Extra peritoneal Repair, below], but this can also be done with blunt dissection if the surgeon prefers. Two additional 5 mm trocars are placed. The dissection is performed with two blunt-tipped dissectors.

Operative Technique

Totally Extraperitoneal Repair

The extra-abdominal preperitoneal approach to laparoscopic hernia repair, developed by McKernan,20,21 attempts to duplicate the open preperitoneal repair described by Stoppa22–24 and Wantz.12,25 In a TEP repair, the trocars are placed preperitoneally in a space created between the fascia and the peritoneum. Ideally, the dissection remains in the extra-abdominal plane at all times, and the peritoneum is never penetrated.

Step 1: creation of preperitoneal spaceWith the patient in the Trendelenburg position, the anterior rectus fascia is opened through a 1 cm infraumbilical transverse incision placed slightly toward the side of the hernia, which helps prevent inadvertent opening of the peritoneum. An index finger is inserted on the medial aspect of the exposed rectus abdominis and slid over the posterior rectus sheath. In this plane, a preperitoneal tunnel be tween the recti abdominis and the peritoneum is created in the midline by inserting a Kelly forceps (with the tips up) and performing gentle blunt dissection to the level of the symphysis pubis. A blunt 10/12 mm trocar is then secured in the preperitoneal space with fascial stay sutures.

A 30° or 45° operating laparoscope is inserted into the trocar for visualization of the development of the correct plane while insufflation of the preperitoneal space is begun, with the recti abdominis seen anteriorly and the peritoneum posteriorly. Maximal inflation pressure is 10 to 12 mm Hg to prevent disruption of the peritoneum or development of extensive subcutaneous emphysema. Blunt gentle dissection with the laparoscope is employed to develop the space sufficiently to allow placement of additional trocars.

Figure 6. TEP approach: preperitoneal distention balloon system

An alternative approach to dissection of the preperitoneal space—one that is especially helpful early in a surgeon’s experience—is to employ a preperitoneal distention balloon system (e.g., PDB; U.S. Surgical Inc., Norwalk, Connecticut). This system consists of a trocar with an inflatable balloon at its tip, which is used to develop the preperitoneal space by atraumatically separating the peritoneum from the abdominal wall. The balloon is inserted into the preperitoneal space below the umbilicus by means of an open Hasson technique and is tunneled inferiorly toward the pubis until the bone is felt with the tip of the balloon trocar. With the laparoscope in the trocar, the preperitoneal working space is developed by gradual inflation of the balloon to a volume of 1 L; the transparency of the balloon permits constant laparoscopic visualization throughout the distention process. Once the working space is created, the PDB is removed and replaced with a blunt sealing trocar. S-retractors are used to elevate the rectus and help ensure correct positioning of the trocar above the posterior fascia. The preperitoneal space is then reinsufflated to a pressure of 10 to 12 mm Hg [see Figure 6].


Step 2: trocar placement

Figure 7. TEP approach: trocar placement

After the peritoneum is dissected away from the rectus abdominis, a midline 5 mm trocar is inserted under direct vision three fingerbreadths below the infraumbilical port. A second 5 mm trocar is then inserted another three fingerbreadths below the first 5 mm trocar. Placement of the working trocars away from the pubis facilitates mesh placement, in that the bottom port is not covered by the top of the mesh and thereby rendered nonfunctional [see Figure 7]. Care must be taken not to penetrate the peritoneum during trocar placement. If the peritoneum is penetrated, the resulting pneumoperitoneum can reduce the already limited working space. If the working space is compromised to the point where the repair cannot continue (which is not always the case), the surgeon can either try to repair the rent with a suture or place a Veress needle in the upper abdominal peritoneal cavity. If such maneuvers are unsuccessful, the loss of working space may necessitate conversion to a TAPP approach.

Step 3: dissection of hernial sac

Figure 8. TEP approach: dissection of peritoneum
Figure 9. TEP approach: gas insufflation
Figure 10. TEP approach: seperation and dissection of hernial sac

Wide dissection of the preperitoneal space is then undertaken with blunt graspers in a two-handed technique by bluntly dividing the avascular areolar tissue between the peritoneum and the abdominal wall [see Figure 8]. The pubis, Cooper’s ligament, and the inferior epigastric vessels are located first and used to orient the dissection. If a direct hernia is present medial to the inferior epigastric vessels, it will often be reduced by the balloon dissector [see Figure 9]. If not, the sac and the preperitoneal contents are carefully dissected away from the fascial defect and swept cephalad as far as possible. Gentle traction is applied to expose and dissect away the attachment of the peritoneum to the transversalis fascia [see Figure 10].

The indirect space is then exposed by sweeping off the tissue lateral to the inferior epigastric vessels until the peritoneum is found. If a lipoma of the cord is present, it will be lateral to and covering the peritoneum and should be dissected out of the internal ring in a cephalad direction to prevent it from displacing the mesh.18 If there is no indirect hernia, the peritoneum will be found cephalad to the internal ring. To ensure secure mesh placement, the peritoneum is bluntly dissected off the cord structures and placed as far cephalad as possible.

Figure 11. TEP approach: retraction of hernial sac

If an indirect hernia is present, the sac will be lateral and anterior to the cord structures. A small indirect hernial sac is bluntly dissected off the spermatic cord with a hand-over-hand technique and reduced until an area sufficient for mesh placement is created [see Figure 11]. To prevent early recurrence, all attachments of the peritoneum should be dissected cephalad to where the inferior edge of the mesh will be. If a large indirect sac is not easily reduced from the scrotum, it may be transected in its superolateral edge, dissected off the cord structures, and closed with an endoscopic ligating loop. The distal sac is then left in place and not ligated.

Unlike a TAPP repair, in which any indirect hernia present is readily apparent at first inspection, a TEP repair always requires that the space lateral to the inferior epigastric vessels be dissected to make sure that there is no indirect component. This dissection should be done even if a direct or femoral hernia is identified. The medial border of dissection is the iliac vein or its overlying fat, and the lateral border of dissection is the psoas muscle. Superiorly, dissection should reach the level of the umbilicus.


Step 4: placement of mesh

Figure 12. TEP approach: placement of mesh

As a rule, we use a large (10.8 × 16 cm) piece of polypropylene mesh shaped to the contours of the inguinal region. A number of different products can also be used for this purpose, including various forms of polypropylene and several types of polyester. A marking suture is placed at the superior edge of the mesh on the concave side, which is to be apposed to the peritoneum. The mesh is wrapped around a grasper in a tubular fashion, then inserted through the umbilical trocar into the preperitoneal space. Once in the preperitoneal space, the mesh is manipulated to cover the pubic tubercle, the internal ring, Cooper’s ligament, the femoral canal, and the rectus abdominis superiorly [see Figure 12]. Tacks or sutures are not usually needed for fixation, but if they are used, they should be placed into Cooper’s ligament and the anterior abdominal wall; to prevent nerve injury, no tacks should be placed inferior to the iliopubic tract lateral to the internal ring. Some surgeons believe that with direct hernias, there is a risk that the mesh may migrate into a large defect. Accordingly, they place several tacks into Cooper’s ligament to prevent this occurrence. For bilateral hernias, two identical repairs are done, and two mesh patches are used.

Step 5: closure The operative site is inspected for hemostasis. The trocars are removed under direct vision. The insufflated CO2 is slowly released so that the mesh may be visualized as the preperitoneal fat and contents collapse back onto the mesh. The fascia at trocar sites 10 mm or larger is closed with 2-0 polydioxanone sutures, and the skin is closed with subcuticular sutures.

Transabdominal Preperitoneal Repair

Step 1: placement of trocars

Figure 13. TAPP approach: trocar placement

Pneumoperitoneum is established through a small infraumbilical incision. We generally prefer an open technique, in which a blunt-tipped 12 mm trocar is inserted into the peritoneal cavity under direct vision. CO2 is then insufflated into the abdomen to a pressure of 12 to 15 mm Hg. The angled laparoscope is introduced, and both inguinal areas are inspected. Two 5 mm ports are placed, one at the lateral border of each rectus abdominis at the level of the umbilicus, to allow placement of the camera and the instruments [see Figure 13]. The 5 mm lateral ports may be replaced with 10 mm ports if only a 10 mm laparoscope is available.

Step 2: identification of anatomic landmarks The four key anatomic landmarks mentioned earlier [see Laparoscopic Inguinal Hernia Repair, Anatomic Considerations, above]—the spermatic vessels, the obliterated umbilical artery (medial umbilical ligament), the inferior epigastric vessels (lateral umbilical ligament), and the external iliac vessels—are identified on each side.

In the presence of an indirect hernia, the internal inguinal ring is easily identified by the presence of a discrete hole lateral to the junction of the vas deferens, the testicular vessels, and the inferior epigastric vessels. Identification of a direct hernia can be more difficult. Sometimes, a direct hernia appears as a complete circle or hole; at other times, it appears as a cleft, medial to the vas deferens-vascular junction; and at still other times, it is completely hidden by preperitoneal fat and the bladder and umbilical ligaments. Visualization can be particularly difficult in obese patients, who may have considerable lipomatous tissue between the peritoneum and the transversalis fascia, or in patients whose hernia consists of a weakness and bulging of the entire inguinal floor rather than a distinct sac. For adequate definition of this type of hernia and deeper anatomic structures, the peritoneum must be opened, a peritoneal flap developed, and the underlying fatty layer dissected.26 Direct hernial defects are often situated medial to the ipsilateral umbilical ligament, and retraction or even division of this structure is sometimes necessary. Division of this structure has no negative sequelae; however, the surgeon should be aware that the obliterated umbilical artery may still be patent and that use of the electrocautery or clips may be necessary. Traction on the ipsilateral testicle can demonstrate the vas deferens when visualization is obscured by overlying fat or pressure from the pneumoperitoneum.


Step 3: creation of peritoneal flap

Figure 14. TAPP approach: dissection of left direct inguinal hernia

The curved scissors or the hook cautery is used to create a peritoneal flap by making a transverse incision along the peritoneum, beginning 2 cm above the upper border of the internal inguinal ring and extending medially above the pubic tubercle and laterally 5 cm beyond the internal inguinal ring [see Figure 14]. Extreme care must be taken to avoid the inferior epigastric vessels. Bleeding from these vessels can usually be controlled by cauterization, but application of hemostatic clips may be necessary on occasion. Another solution is to pass percutaneously placed sutures above and below the bleeding point while applying pressure to the bleeding vessel so as not to obscure the field of vision. If the monopolar cautery is used to create the peritoneal flap, the entire uninsulated portion of the instrument must be visible at all times to ensure that inadvertent bowel injury does not occur.

Figure 15. TAPP approach: dissection of peritoneum

The incised peritoneum is grasped along with the attached preperitoneal fat and the peritoneal sac and is dissected cephalad with blunt and sharp instruments to create a lower peritoneal flap [see Figure 15]. Dissection must stay close to the abdominal wall. A significant amount of preperitoneal fat may be encountered, and this should remain with the peritoneal flap so that the abdominal wall is cleared. When the correct preperitoneal plane is entered, dissection is almost bloodless and is easily carried out.


Step 4: dissection of hernial sacThe hernial sac, if present, is removed from Hesselbach’s triangle or the spermatic cord and surrounding muscle through inward traction, countertraction, and blunt dissection with progressive inversion of the sac until the musculofascial boundary of the internal inguinal ring and the key deep anatomic structures are identified. In most cases, the hernial sac can be slowly drawn away from the transversalis fascia or the spermatic cord. The sac is grasped at its apex and pulled inward, thus being reduced by inversion. The indirect sac may be visualized more easily if it is grasped and retracted medially; this step facilitates its dissection away from the cord structures.

Spermatic cord lipomas usually lie posterolaterally and are extensions of preperitoneal fat. In the presence of an indirect defect, such lipomas should be dissected off the cord along with the peritoneal flap to lie cephalad to the internal inguinal ring and the subsequent repair so that prolapse through the ring can be prevented.

A large indirect hernial sac can be divided at the internal ring if it cannot be readily dissected away from the cord structures. This step may prevent the type of cord injury that can result from extensive dissection of a large indirect sac. Division of a large in-direct sac is best accomplished by opening the sac on the side opposite the spermatic cord, then completing the division from the inside.16


Step 5: reidentification and exposure of landmarks

Figure 16. TAPP approach: identification of anatomic landmarks

Once the peritoneal flap has been created, the key anatomic landmarks mentioned earlier [see Laparoscopic Inguinal Hernia Repair, Anatomic Considerations, above] must be reidentified and exposed so that neurovascular structures can be protected from injury and the tissues required for reliable mesh fixation can be located. The pubic tubercle is often more easily felt than seen. Cooper’s ligament is initially felt and subsequently seen along the pectineal prominence of the superior pubic ramus as dissection continues laterally and fatty tissue is swept off to expose the glistening white structure. Care must be taken to avoid the numerous small veins that often run on the surface of the ligament, as well as to avoid the occasional aberrant obturator artery. The iliopubic tract is initially identified at the inferior margin of the internal inguinal ring, with the spermatic cord above, and is then followed in both a medial and a lateral direction. Minimal dissection is carried out inferior to the iliopubic tract so as not to injure the genital femoral nerve, the femoral nerve, and the lateral cutaneous nerve of the thigh [see Figure 16].

Step 6: placement of meshA 10 × 6 cm sheet of polypropylene mesh is rolled into a tubular shape and introduced into the abdomen through the 10/12 mm umbilical trocar. Prolene is preferable to Marlex in this application because it is less dense, conforms more easily to the posterior inguinal wall, and has larger pores, which facilitate visualization and subsequent securing with staples or tacks. The inherent elasticity and resiliency of Prolene mesh allow it to unroll easily while maintaining its form. The mesh is used to cover the direct space (Hesselbach’s triangle), the indirect space, and the femoral ring areas (i.e., the entire inguinal floor). We do not make a slit in the mesh for the cord.

It is our practice with the TAPP technique to tack the mesh to prevent any migration. We use an endoscopic multifire spiral tacker to secure the mesh, beginning medially and proceeding laterally. The upper margin is first tacked to the rectus abdominis and the transversus abdominis fascia and arch, with care taken to stay 1 to 2 cm above the level of the internal inguinal ring and to avoid the inferior epigastric vessels, up to a point several centimeters lateral to the internal inguinal ring or the indirect hernial defect. Extending mesh fixation to the anterior iliac spine is neither necessary nor desirable. A two-handed technique is recommended for tack placement: one hand is on the tacker, and the other is on the abdominal wall, applying external pressure to place the wall against the tacker. The tacker itself is frequently pushed against the tissues and used as a spreader and palpator. However, it must not be forced too deeply into the abdominal wall superolateral to the spermatic cord; doing so might lead to inadvertent entrapment of the sensory nerves. The tacker can be moved from the left to the right port, depending on which position more readily allows placement of the staples perpendicular to the mesh and the abdominal wall.

Once the superior margin is fixed, fixation of the inferior margin is accomplished, beginning at the pubic tubercle and moving laterally along Cooper’s ligament. The mesh is lifted frequently to ensure adequate visualization of the spermatic cord. Care is taken to avoid the adjacent external iliac vessels, which lie inferiorly. Lateral to the cord structures, all tacks are placed superior to the iliopubic tract to prevent subsequent neuralgias involving the lateral cutaneous nerve of the thigh or the branches of the genitofemoral nerve. If the surgeon can palpate the tacker through the abdominal wall with the nondominant hand, the tacker is above the iliopubic tract. The mesh should lie flat at the end of the procedure.


Step 7: closure of peritoneum

Figure 17. TAPP approach: insertion of mesh

The peritoneal flap, including the redundant inverted hernial sac, is placed over the mesh, and the peritoneum is reapproximated with the tacker [see Figure 17]. Reduction of the intra-abdominal pressure to 8 mm Hg, coupled with external abdominal wall pressure, facilitates a tension-free reapproximation. Alternatively, the peritoneum may be sutured over the mesh, but in most surgeons’ hands, this closure takes longer.

Step 8: closure of fascia and skin The peritoneal repair is inspected to ensure that there are no major gaps that might result in exposure of the mesh and subsequent formation of adhesions. The trocars are then removed under direct vision, and the pneumoperitoneum is released. The fascia at the 10/12 mm port sites is closed with 2-0 polydioxanone sutures to prevent incisional hernias. The skin is closed with 4-0 absorbable subcuticular sutures.

Postoperative Care

Patients are observed in the recovery room until they are able to ambulate unassisted and to void; if they are unable to void at the time of discharge, in-and-out catheterization is performed. Patients are advised to resume their usual activities as they see fit; driving a car is permitted when pain is minimal. Outpatient prescriptions for acetaminophen, naproxen, and oxycodone are given, and follow-up visits in the surgical clinic are scheduled for postoperative day 7 to 14. Patients who live alone, have had intraoperative complications, have significant nausea or vomiting, or experience un explained or inordinate pain are admitted overnight.

Disadvantages and Complications


Need for general anesthesia The need for pneumoperitoneum and thus for general anesthesia in laparoscopic herniorrhaphy is sometimes considered a major disadvantage. Nausea, dizziness, and headache are more common in the recovery room after TAPP repair than after Lichtenstein repair.27 It is not necessarily true, however, that local or regional anesthesia is safer than general anesthesia.28 Anesthesiology studies critically appraising anesthetic techniques for hernia surgery have shown the choice of general anesthesia over local or regional anesthesia to be safe and, in many cases, advantageous, particularly in patients who are in poor health. Furthermore, TEP repair has been successfully done with patients under epidural and local anesthesia.14,29,30

Lower cost-effectiveness A study comparing costs at North American teaching hospitals found that TEP repair cost US$852 more than Lichtenstein repair; however, this study could not quantify the cost savings arising from faster recuperation and earlier reentry into the workforce.31 Some studies have demonstrated economic savings with the use of a laparoscopic approach, in the form of fewer days of work missed and reduced worker’s compensation costs.32,33 Operating costs can also be reduced by avoiding the use of disposable instruments.34 In addition, operating time has been shown to decrease as the surgeon’s experience with the procedure increases.15,35,36


Most randomized trials comparing laparoscopic repair with open mesh repair have found the overall complication rate to be comparable between groups.14,15 In general, however, the rate of serious perioperative complications, though still low, is increased with the laparoscopic approach.15

Complications of access to peritoneal cavity A TAPP repair exposes the patient to several potentially serious risks related to the choice of the transabdominal route. Trocar injuries to the bowel, the bladder, and the vascular structures can occur during the creation of the initial pneumoperitoneum or the subsequent insertion of the trocars.14,37 Visceral injury rates reported for the laparoscopic approach, though quite low, are still about 10 times those reported for the open approach.14 Another complication related to trocar placement is incisional hernia,14 which can lead to postoperative bowel obstruction14,15,38; however, this complication can be minimized by using 5 mm trocars and a 5 mm laparoscope instead of the larger 10/12 mm instruments.

Complications of dissection Injuries occurring during dissection are often linked to inexperience with laparoscopic inguinal anatomy. If serious enough, they can necessitate laparotomy. Fortunately, such conversion is rare (< 3%).14 The most common vascular injuries occurring during laparoscopic inguinal herniorrhaphy are those involving the inferior epigastric vessels and the spermatic vessels.14 The external iliac, circumflex iliac, profunda, and obturator vessels are also at risk. A previous lower abdominal operation is a risk factor.16 The source of any abnormal bleeding during the procedure must be quickly identified. All vessels in the groin can be ligated except the external iliac vessels, which must be repaired.16

Injuries to the urinary tract may also occur. Four bladder injuries necessitating repair were documented in a collected series of 762 laparoscopic repairs by different surgical groups.37,39,40 Bladder injuries are most likely to occur when the space of Retzius has been previously dissected (e.g., in a prostatectomy). Renal and ureteral injuries identified intraoperatively should be repaired immediately. Often, however, these injuries are not apparent until the postoperative period, when they present as lower abdominal pain, renal failure, ascites, dysuria, or hematuria—all of which should be investigated promptly. Although indwelling catheter drainage may constitute sufficient treatment of a missed retroperitoneal bladder injury, intraperitoneal injuries are best treated by direct repair via either laparoscopy or laparotomy.


Complications related to meshComplications related to the use of mesh include infection, migration, adhesion formation, and erosion into intraperitoneal organs. Such complications usually become apparent weeks to years after the initial repair, presenting as abscess, fistula, or small bowel obstruction.

Mesh infection is very rare. In the 2003 Cochrane review of antibiotic prophylaxis for nonmesh hernia repairs,19 the overall infection rate was 4.69% in the control group and 3.08% in the treatment group. Thus, to prevent one infection in 30 days, 50 patients would have to be treated, and these patients would then be at risk for antibiotic-associated complications. Laparoscopic repairs were excluded from this review; however, in a meta-analysis comparing postoperative complications after laparoscopic inguinal hernia repair with those after open repair, superficial infection was less frequent in the laparoscopic groups.14 Deep mesh infection was rare in both groups. Mesh infection usually responds to conservative treatment with antibiotics and drainage. On rare occasions, the mesh must be removed; this may be accomplished via an external approach. It is noteworthy that removal of the mesh does not always lead to recurrence of the hernia, a finding that may be attributable to the resulting fibrosis.41

Mesh migration may lead to hernia recurrence. In a TAPP repair, appropriate stapling of the mesh should reduce this possibility. In a TEP repair, stapling does not appear to be necessary to prevent migration.35,42

The risk that adhesions to the mesh will form is augmented if the mesh is left exposed to the bowel. The long-term durability and effectiveness of the sometimes flimsy peritoneal coverage employed in the TAPP approach have been questioned. Even in the TEP approach, small tears in the peritoneum may expose the bowel to the mesh.


Urinary complications Injuries to the urinary tract aside [see Complications of Dissection, above], urinary retention, urinary tract infection, and hematuria are the most common complications. Avoidance of bladder catheterization reduces the incidence of these complications, but urinary retention still occurs in 1.5% to 3% of patients.15 General anesthesia and the administration of large volumes of I.V. fluids may also predispose to retention.

Vas deferens and testicular complications Wantz43 believed that the most common cause of postoperative testicular swelling, orchitis, and ischemic atrophy is surgical trauma to the testicular veins (i.e., venous congestion and subsequent thrombosis). Because spermatic cord dissection is minimized with the laparoscopic approach, the risk of groin and testicular complications resulting from injury to cord structures and adjacent nerves may be reduced.15

Most testicular complications, such as swelling, pain and epididymitis, are self-limited. Testicular pain occurs in about 1% of patients after laparoscopic repair,33 an incidence comparable to that seen after open repair.15,44 A similar number of patients experience testicular atrophy,37 for which there is no specific treatment.

The risk of injury to the vas deferens appears to be much the same in laparoscopic repair as in open repair.14 If fertility is an issue, the cut ends should be reapproximated if the injury is recognized intraoperatively.


Postoperative groin and thigh pain Unlike patients who undergo open anterior herniorrhaphy, in whom discomfort or numbness is usually localized to the operative area, patients who undergo laparoscopic repair occasionally describe unusual but specific symptoms of deep discomfort that are usually positional and are often of a transient, shooting nature suggestive of nerve irritation. The pain is frequently incited by stooping, twisting, or movements causing extension of the hip and can be shocklike. Although these symptoms can frequently be elicited in the early postoperative period, they are usually transient. If tacks or staples were used and neuralgia is present in the recovery room, prompt reexploration is the best approach.16

Persistent pain and burning sensations in the inguinal region, the upper medial thigh, or the spermatic cord and scrotal skin region occur when the genitofemoral nerve or the ilioinguinal nerve is stimulated, entrapped, or unintentionally injured. When these symptoms persist, they may result in severe morbidity.45 A more worrisome symptom is lateral or central upper medial thigh numbness, which is reported in 1% to 2% of patients and often lasts several months or longer. Whether this numbness is related to staple entrapment, fibrous adhesions, cicatricial neuroma, or mesh irritation is unknown. Numbness and paresthesia of the lateral thigh are less frequent and are related to the involvement of fibers of the lateral cutaneous nerve. These problems can be prevented by paying careful attention to anatomic detail and technique.46 Anatomic study, based on cadaveric dissections, suggests that both the genitofemoral nerve and the lateral cutaneous nerve of the thigh will be protected in all cases if no staples are placed further than 1.5 cm lateral to the edge of the internal ring.47

A great deal of attention has rightly been focused on the risk of nerve injury with laparoscopic hernia repair, as well as on ways of preventing it. At the same time, it is important to note that pain and numbness, including thigh numbness, can also occur after open repair and may in fact be more common in that setting than was previously realized. In one study, persistent groin pain was present in 9.5% of patients after Lichtenstein repair versus 5.5% of patients after TAPP.48 In the Cochrane meta-analysis, persistent pain and numbness 1 year after surgery were found to be significantly reduced with either TEP or TAPP repair.14 This finding was confirmed by a 2004 study that reported a 9.8% incidence of neuralgia or other pain at 2 years in patients who underwent laparoscopic repair, compared with a 14.3% incidence in open repair patients.15


Miscellaneous complications Laparoscopic repair apparently reduces the incidence of hematomas while increasing that of seromas.14 Lipomas of the spermatic cord, if left unreduced in patients with indirect hernias, may produce a persistent groin mass and a cough impulse that mimic recurrence, especially to an uninitiated examiner. These lipomas are always asymptomatic.

Outcome Evaluation

Although there is a large body of literature on laparoscopic inguinal hernia repair—including a variety of randomized, controlled trials—the benefits of the laparoscopic approach have not yet been clearly defined or widely accepted. Given the low morbidity and relatively short recovery already associated with the conventional operation, demonstration of any significant differences between the open mesh and laparoscopic techniques requires large study samples. The previously cited meta-analysis done by the Cochrane collaboration addressed this question.14 Forty-one trials were included in this meta-analysis, ranging in size from 38 to 994 randomized patients. The duration of follow-up ranged from 6 weeks to 36 months. The results of the meta-analysis suggested that whereas operating times were longer and the risk of rare but serious complications higher in the laparoscopic groups, recovery was quicker and persistent pain and numbness less frequent. Recurrence rates did not differ significantly.

In addition, a large randomized, multicenter Veterans Affairs (VA) study published in 2004 compared open mesh repair with lap aroscopic mesh repair (TEP, 90%; TAPP, 10%) in 2,164 patients.15 Patients with previous mesh repairs were excluded. In this study, the laparoscopic approach was associated with a higher risk of complications and, in contrast to the findings of the meta-analysis, a higher overall recurrence rate at 2 years after operation. Pain was reduced and recovery time shortened in the laparoscopic group.

Thus, there remains a degree of controversy regarding the ideal approach to and outcome for inguinal hernia repair. Accordingly, we will briefly review the salient outcomes of a number of studies that compare laparoscopic inguinal herniorrhaphy with open mesh repair.

Laparoscopic Repair versus Open Mesh Repair

Operating time The Cochrane meta-analysis suggested that overall, the average operating time was 15 minutes longer with the laparoscopic approach; however, for bilateral hernias, laparoscopic repair required no more time than open repair.14 The surgeon’s level of experience with laparoscopic technique was not explicitly stated in all of these studies, which made it difficult to assess the impact of this variable on operating time. It has been shown that with more experience and greater specialization, the differences in operating time between laparoscopic and open repair tend to decrease and become clinically unimportant.35,49

Recovery time The most significant short-term outcome measure after hernia repair is recovery time, defined as the time required for the patient to return to normal activities. One of the most frequently cited benefits of laparoscopic herniorrhaphy is the patient’s rapid return to unrestricted activity, including work. The Cochrane meta-analysis revealed that recovery time was significantly shorter after laparoscopic repair than after open mesh repair.14 In a cost comparison between TEP repair and Lichtenstein repair, recovery time was 15 days after the former, compared with 34 days after the latter.31 In the 2004 VA study, laparoscopic repair patients returned to their normal activities 1 day earlier.15

Postoperative pain After laparoscopic repair, most patients experience minimal immediate postoperative pain and have little or no need for analgesics after postoperative day 1. Patients are able to perform some exercises better after laparoscopic repair than after Lichtenstein repair.48 That patients experience less postoperative pain after laparoscopic repair than after open mesh repair has been reported in several randomized studies.27,32,40,44,50,51 In the Cochrane meta-analysis, persistent pain and numbness 1 year after surgery was significantly less after either TEP or TAPP repair than after open repair.14 In a 2003 report describing a 5-year follow-up of 400 patients treated with either Lichtenstein open mesh repair or TAPP repair, the incidence of permanent paresthesia and groin pain was lower with the TAPP approach.52 Moreover, all of the patients with pain and paresthesia significant enough to affect their daily lives were in the open repair group. A later study that evaluated postoperative neuralgia in 400 patients who underwent either TAPP repair or Lichtenstein repair reported similar findings.48

Quality of life The studies that have assessed quality of life immediately after hernia repair have tended to favor the laparoscopic approach, albeit marginally. Using the SF-36 (a widely accepted general health-related quality-of-life questionnaire), one group found that at 1 month, greater improvements from baseline were apparent in the laparoscopic group in every dimension except general health; however, by 3 months, the differences between the two groups were no longer significant.27 Another group also found no differences in any SF-36 domains at 3 months after operation.37 Yet another study, however, using the Sickness Impact Profile, found some benefit to the laparoscopic approach.53 In contrast, no postoperative differences in SF-36 domains were found in the VA study.15

Bilateral hernias Laparoscopy allows simultaneous exploration of the abdominal cavity (TAPP) and diagnosis and treatment of bilateral groin hernias, as well as coexisting femoral hernias (which are often unrecognized preoperatively), potentially without added risk or disability. Bilateral hernias accounted for 9% of the hernias reviewed in the Cochrane database.14 Operating time was longer in the laparoscopic groups than in the open groups; however, recovery time, the incidence of persistent numbness, and the risk of wound infection were significantly reduced in the former. These results are consistent with those of a prospective, randomized, controlled trial from 2003 that compared TAPP repair with open mesh repair for bilateral and recurrent hernias.54 In this study, TAPP repair not only was less painful and led to an earlier return to work but also was associated with a shorter operating time. Further prospective, randomized trials designed to compare simultaneous bilateral open tension-free repair with bilateral TEP laparoscopic repair should be undertaken.

Recurrent hernias Approximately 10% of patients undergoing hernia repair present with recurrent inguinal hernia.54Patients with recurrent hernias may potentially derive greater benefits from a laparoscopic approach through an undisturbed plane of dissection, rather than a second groin exploration via an open technique, dissecting through scar tissue and potentially causing significant tissue trauma. This is especially true when mesh was used for the previous open repair.

Open mesh repair has been associated with long-term recurrence rates of 1% or less, even when not performed by hernia specialists.2,55 If the laparoscopic approach is to be a viable alternative to open repair, it should have comparable results. With respect to short-term results, prospective, randomized trials suggest that hernia recurrence rates are comparable in laparoscopic repair and open mesh repair groups.14 However, the Cochrane meta-analysis found that reductions in hernia recurrence were effected primarily by the use of mesh rather than by any specific placement technique.14 This finding is consistent with the Cochrane meta-analysis of open mesh inguinal hernia repair versus open nonmesh repair, which indicated that tension-free mesh repair led to a significant reduction in hernia recurrence.3 In a 5-year follow-up study from 2004, the recurrence rate after laparoscopic mesh repair still was not significantly different from that after open mesh repair.34 In the VA study, the hernia recurrence rate at 2 years was higher in the laparoscopic group (10%) than in the open mesh re pair group (4%) for primary, unilateral hernias.15 In both groups, the recurrence rates were higher than generally expected. How ever, these rates were found to be affected by the surgeon’s level of experience: those who had performed more than 250 laparoscopic repairs reported a recurrence rate of 5%.

Most reported recurrences after laparoscopic herniorrhaphy come at an early stage in the surgeon’s experience with these procedures and arise soon after operation.56 The majority can be attributed to (1) inadequate preperitoneal dissection; (2) use of an inadequately sized patch, which may migrate or fail to support the entire inguinal area, including direct, indirect, and femoral spaces; or (3) staple failure with TAPP repair.


Transabdominal Preperitoneal Repair (TAPP) versus Totally Extraperitoneal Repair (TEP)

The TAPP approach is easier to learn and perform than the TEP approach, and even experienced laparoscopic hernia surgeons report more technical difficulties with the latter.15,57 Nonetheless, there is a growing body of literature to suggest that TEP repair, by avoiding entry into the peritoneal cavity, has significant advantages over TAPP repair.4 In particular, the TEP approach should reduce the risk of trocar site hernias, small bowel injury and obstruction, and intraperitoneal adhesions to the mesh. In a study that included 426 patients, TAPP repairs were performed in 339 and TEP repairs in 87, and the patients were followed for a mean of 23 and 7 months, respectively.58 Time off work was shorter after TEP. A total of 15 major complications were noted, including one death, two bowel obstructions, one severe neuralgia, three trocar site hernias, one epigastric artery hemorrhage, and seven recurrences. With the exception of the epigastric artery hemorrhage, all of these complications occurred in the TAPP group. It is possible, however, that these results can be partly explained by the learning curve, in that the TAPP repairs were all done before the TEP repairs. That six TAPP recurrences occurred in the first 31 cases, whereas only one occurred in the subsequent 395 cases, lends support to this possibility.

In a study comparing 733 TAPP repairs with 382 TEP repairs, 11 major complications occurred in the TAPP group (two recurrences, six trocar site hernias, one small bowel obstruction, and two small bowel injuries), whereas only one recurrence and no intraperitoneal complications occurred in the TEP group.59 Seven TEP procedures were converted to TAPP procedures. Time off work was equal in the two groups but was prolonged in patients receiving compensation. As in the study cited above,58 the TAPP patients were followed longer than the TEP patients, and the TAPP cases occupied the first part of the learning curve. To avoid this type of selection bias would require a randomized study.

Not all surgeons are convinced that TEP repair is the laparoscopic procedure of choice. A 1998 study compared 108 TAPP repairs with 100 TEP repairs.57 Although the TEP repairs were done only by surgeons who were already familiar with TAPP repair, many of the surgeons still encountered technical difficulties and problems with landmark identification. Overall, complications did not occur significantly more frequently in either group, but they seemed more severe in the TAPP group: four trocar site hernias, one bladder injury, and six seromas were noted in the TAPP group, compared with one cellulitis and six seromas in the TEP group. The authors concluded that because TAPP repair is easier and does not increase complications significantly, it is an ‘adequate’ procedure. The sample size may have been too small to permit detection of small differences in complication rates.

Regardless of any individual preference for one technique or the other, laparoscopic hernia surgeons ideally should be capable of performing both TEP and TAPP well. For example, a planned TEP repair may have to be converted to a TAPP repair, or a TAPP approach may be required if the surgeon is doing another intra peritoneal diagnostic or therapeutic procedure.

Laparoscopic Incisional Hernia Repair

Incisional hernias develop in approximately 2% to 11% of patients undergoing laparotomy.60,61 It has been estimated that 90,000 ventral hernia repairs are done in the United States every year.5 When prosthetic mesh is not used, repair of large incisional hernias is associated with recurrence rates as high as 63% after 10 years, compared with 32% when mesh is used.62 In addition to these high recurrence rates, even with the use of mesh, open incisional hernia repair may be associated with significant complications and a substantial hospital stay.

Initially described in 1992,9 laparoscopic repair of incisional hernias has evolved from an investigational procedure to one that can safely and successfully be used to repair ventral hernias. Taking a laparoscopic approach allows the surgeon to minimize abdominal wall incisions, avoid extensive flap dissection and muscle mobilization, and eliminate the need for drains in proximity to the mesh, thereby potentially achieving reductions in pain, recovery time, and duration of hospitalization, as well as lower rates of surgical site infection (SSI).8,63,64 In addition, the improved visualization of the abdominal wall associated with the laparoscopic view may result in better definition of the defect, the discovery of unrecognized hernia sites, and improved adhesiolysis. Improved visualization permits more precise and accurate placement and tailoring of the mesh, as is suggested by the reduced recurrence rates (9% to 12%) reported up to 5 years after laparoscopic incisional herniorrhaphy.64–66

Preoperative Evaluation

Selection of Patients

Laparoscopic incisional hernia repair may be considered for any ventral hernia in which mesh will be used for the repair. This category includes virtually all incisional hernias, in that even small 0000(< 10 0cm2) defects are known to carry a significant risk of recurrence.62 Both upper abdominal and lower abdominal incisions are amenable to a laparoscopic approach, although hernias at the extremes of the abdominal wall—abutting the pubis, the xiphoid, or the costal margins—pose a technical challenge for effective mesh fixation. The so-called Swiss cheese hernia, which comprises multiple small defects, is particularly well suited to this approach; open repair would necessitate a large incision for access to the multiple fascial defects, and small defects might not be appreciated. Incarcerated hernias can also be approached laparoscopically; however, the suspected presence of compromised bowel is a contraindication. An abdomen that has undergone multiple operations and contains dense adhesions presents a challenge in terms of both access to the abdominal cavity and access to the hernia site. If the surgeon cannot obtain safe access to the peritoneal cavity for insufflation, a laparoscopic approach is contraindicated.

Contraindications Laparoscopic incisional herniorrhaphy is contraindicated in patients with suspected strangulated bowel or loss of domain. Hernias in which the fascial edges extend lateral to the midclavicular line may make trocar placement lateral to the defect impossible. Defects in close proximity to the bony margins of the abdomen, especially those near the xiphoid, pose significant challenges for mesh fixation, though this is also true with open incisional herniorrhaphy. Patients who have undergone multiple previous operations, with or without mesh, may have dense adhesions. Patients in whom polypropylene mesh has previously been placed in an intra-abdominal position may have dense adhesions to the underlying viscera. Whether such patients are approached laparoscopically should be determined by the surgeon’s expertise.

Operative Planning


The procedure is performed with the patient under general anesthesia. Mechanical bowel preparation is not routinely used; however, it may be considered if incarcerated colon is suspected. If the defect is in the lower abdomen, a three-way Foley catheter is placed in the bladder. Sequential compression stockings are applied. Patients are routinely given heparin, 5,000 U subcutaneously, and prophylactic antibiotics.67


The patient is placed in the supine position with both arms tucked. If the hernia is in the midline, the surgeon can stand on either side of the patient, with the monitor directly opposite. If the hernia extends significantly to one side, initial trocar placement is done on the opposite side. Initially, the assistant stands on the same side as the surgeon; however, he or she may later have to move to the opposite side to help with dissection and stapling. A second monitor on the opposite side of the table is useful. If the defect is subcostal, the surgeon may prefer to operate from between the patient’s legs, with a monitor at the head of the bed.


As the wide variety of mesh materials currently available suggests, there is no one ideal mesh. Meshes may be divided into two categories: (1) polymeric meshes and (2) meshes made of specially prepared connective tissue (animal or human) [see Table 1]. The polymeric meshes are biocompatible materials made of either polypropylene, polyester, expanded polytetrafluoroethylene (ePTFE), or laminates of these. Most ePTFE meshes are engineered so that one side is porous to encourage tissue ingrowth and the other is smooth to resist adhesion formation. They may also be coated with an adhesion-resisting absorbable material.

Because laparoscopic incisional hernia repair leaves the mesh exposed to the intraperitoneal cavity, concerns have been expressed about the risk of adhesion formation and fistulization if polypropylene mesh is used. Polytetrafluoroethylene (PTFE) mesh has been demonstrated to have a reduced propensity for adhesion formation.

Additional special equipment used for incisional hernia repair includes a suture passer, a 5 mm spiral tacker (or other tacking device), and 2-0 monofilament sutures. Several tacking devices and suture placement devices have been developed to facilitate mesh fixation [see Table 2]. All work in essentially the same manner. A Keith or similar needle may also be used. Atraumatic bowel instruments are required to manipulate the bowel if lysis of adhesions is needed.

Operative Technique

In essence, the repair consists of the intraperitoneal placement of a large piece of mesh so that it overlaps the defect in the fascia and the abdominal wall. The defect is not closed. The mesh is anchored with a minimum of four subcutaneously tied transfascial sutures placed at the four corners and is further secured between the sutures with intraperitoneally placed tacks and additional sutures as needed.

Step 1: placement of trocarsBecause of the probability of extensive intra-abdominal adhesions, we begin with open insertion of a blunt 12 mm trocar. Although open insertion necessitates an often tedious dissection through several layers of the abdominal wall, it has the advantage of allowing early diagnosis and repair of any iatrogenic bowel injury. Nevertheless, good results have also been reported with insertion of a Veress needle, usually in the left upper quadrant (where adhesions are presumed to be minimal). Ultimately, trocar position is determined by the location of the hernia. For midline hernias, we usually begin on the left side of the patient and insert the first trocar lateral to the edges of the defect, about midway between the costal margin and the iliac crest.

Figure 18. Incisional trocar site placement

CO2 is then insufflated to a pressure of 12 to 15 mm Hg, and a 5 mm 30° scope is inserted. As in laparoscopic inguinal hernia repair, an angled scope is essential because dissection and repair are done on the undersurface of the anterior abdominal wall, which cannot be adequately visualized with a 0° scope. The hernial defect is visually identified, and two additional 5 mm trocars are placed on the same side under direct vision, with their precise placement dependent on the size and contours of the defect and on the locations of any adhesions. If possible, these trocars are placed superior and inferior to the initial trocar, as far laterally as possible, with care taken to ensure that the downward movement of the instruments is not limited by the iliac crest or the thigh. Lateral placement is necessary to optimize exposure of the abdominal wall [see Figure 18].


Step 2: exposure of hernial defect

Figure 19. Small hernial defects

The edges of the hernial defect are exposed by reducing the contents of the hernia into the abdominal cavity. All adhesions from bowel or omentum to the abdominal wall in the vicinity of the defect and along the full length of the previous incision should be divided. Complete adhesiolysis of abdominal wall adhesions facilitates the identification of Swiss cheese defects [see Figure 19]. Starting in the upper abdomen may be easier, in that bowel adhesions are less likely to be encountered. External pressure may also help reduce hernial contents and facilitate identification of the hernial sac. Placing the patient head down or head up and rotating the table will also aid in exposure.

Figure 20. Dissection of incarcerated small bowl

If dense adhesions are present, it is preferable to divide the sac or the fascia, so as not to risk bowel injury [see Figure 20]. Sharp dissection with scissors is recommended to prevent thermal injury to the bowel, which may not be immediately recognized. If incarcerated bowel cannot be reduced with laparoscopic techniques, an incision is made over the area of concern, and the bowel is freed under direct vision. Once this incision is closed, laparoscopic mesh placement can proceed, and there may be no need for full conversion to an open operation. Strategies for avoidance and treatment of bowel injury are discussed in more detail elsewhere [see Complications, Bowel Injury, below].


Step 3: selection of mesh

Figure 21. Marking of contours of incisional hernial defect

The contours of the hernial defect are marked as accurately as possible on the exterior abdominal wall; the edges may be delineated with a combination of palpation and visualization. All Swiss cheese defects are marked. The defect is measured after pneumoperitoneum is released to ensure that its size is not overestimated [see Figure 21]. Ideally, the prosthesis should overlap the defect by at least 3 cm on all sides. Coverage of all of the defects with a single sheet of mesh is preferred, but more than one sheet may be needed, depending on the locations of the defects and the size of the patient. The mesh sheet is laid on the abdominal wall in a position that approximates its eventual intra-abdominal position, and its four corners and those of its representation on the abdominal wall are numbered clockwise from 1 through 4 for later orientation [see Figure 22]. A mark is made on the inner side of the mesh sheet so that the surgeon can easily determine which side is to face the peritoneum once the mesh is inserted into the peritoneal cavity. If a dual-sided mesh is being used, the smooth side must be the one facing the bowel. A 2-0 monofilament suture is tied in each corner of the mesh, and both ends of each suture are left about 15 cm long.

The mesh swatch is rolled as tightly as possible around a grasping forceps and introduced into the peritoneal cavity. Small swatches can be inserted through a 12 mm trocar; for larger pieces, we remove a large trocar and insert the mesh directly into the abdomen. The trocar is then repositioned and insufflation of CO2 recommenced.


Step 4: fixation of meshOnce the mesh swatch has been introduced into the abdomen, it is unfurled and spread out, with the previously placed corner sutures facing the fascia and oriented so that the four numbered corners are aligned with the numbers marked on the abdominal wall. Small skin incisions are then made with a No. 11 blade. Through each of these incisions, a suture passer is inserted to grasp one tail of the previously placed anchoring suture and pull it out through the abdominal wall, then reintroduced through the incision at a slightly different angle to pull out the second tail. This is done for each of the four anchoring sutures, and the mesh is unfurled under appropriate tension (with care taken to avoid excessive tension and stretching). Initially, the corner sutures are held with hemostats and not tied; some adjustment to achieve optimal positioning of the mesh is often required, especially early in the surgeon’s experience.

Figure 23. Internal view of two-layer mesh fixed to abdominal wall
Figure 24. Two-handed tacking

Once the mesh is in a satisfactory position, each suture is tied and buried in the subcutaneous tissue to anchor the mesh swatch to the fascia and maintain its proper orientation [see Figure 23]. With large defects, it is usually necessary to place one or more additional 5 mm trocars contralateral to the initial ports to aid in mesh fixation. A tacker is then employed to tack the mesh circumferentially at 1 cm intervals along its edge. A two-handed technique is used, in which the second hand applies external pressure to the abdominal wall to ensure that the tacks obtain the best possible purchase on the mesh and the abdominal wall [see Figure 24]. Care should be taken to place the tacks flush because they can cause bowel injury if left protruding.68 Additional sutures are then passed at 5 cm intervals directly through the mesh with either the suture passer and a free suture or a suture on a Keith needle. Sutures should be tied taut but not tight, so as not to cause necrosis of the intervening tissue.


Step 5: closure The pneumoperitoneum is released. The fascia at any trocar site 10 mm in diameter or larger is closed. Careful closure of the site used for open insertion of the first trocar is mandatory to prevent trocar site hernia. The skin is then closed with subcuticular sutures.

Postoperative Care

The Foley catheter is removed at the end of the procedure. Unless adhesiolysis was minimal, patients are admitted to the hospital. Oral intake is begun immediately. Patients are discharged when oral intake is tolerated and pain is controlled with oral medication. Patients are informed that fluid will accumulate at the hernia site and are asked to report any fever or redness. Finally, patients are instructed to resume all regular activities as soon as they feel capable.

Special Situations

Suprapubic Hernia

For hernial defects that extend to the pubic bone, a three-way Foley catheter is inserted. After adhesiolysis, the patient is placed in the Trendelenburg position, and the bladder is distended with methylene blue in saline. The bladder is dissected off the pubic bone until Cooper’s ligament is reached. The mesh is then placed so that it extends behind the bladder and is tacked to the pubic bone, to Cooper’s ligament, or to both.

Subxiphoid or Subcostal Hernia

A hernia in which there is no fascia between the hernia and the ribs or the xiphoid (e.g., a poststernotomy hernia) poses significant challenges for fixation. Because of the risk of intrathoracic injury, the mesh is not tacked to the diaphragm. Although some surgeons perform mesh fixation to the ribs, this measure is often associated with significant postoperative pain and morbidity. In these situations, we take down the falciform ligament and lay the mesh along the diaphragm above the liver, placing tacks and sutures up to but not above the level of the costal margin. Taking down the falciform ligament may be a helpful step for all upper abdominal wall hernia repairs. The recurrence rates for subxiphoid and subcostal hernias are higher than those for hernias at other locations.69

Parastomal Hernias

As many as 50% of stomas are complicated by parastomal hernia formation,70–72 and 10% to 15% will require operative intervention for obstruction, pain, difficulty with stoma care, or unsatisfactory cosmesis. Three methods of repair have been described: primary fascial repair, repair with mesh, and stoma relocation [see 5:27 Open Hernia Repair]. Repair via a laparoscopic approach that uses ePTFE mesh has shown promising short-term results.72,73 The technique that currently seems to be the most successful is the one described by LeBlanc and Bellanger.73 Rather than lateralizing the intestine (as in the technique described by Sugar baker74), this method centralizes the intestine in the mesh by cutting an appropriately sized hole in the middle of the mesh sheet, along with a slit to allow it to be placed around the intestine. This step is repeated on a second piece of mesh, but with the slit oriented to the opposite side. The mesh is fixed with sutures and tacks in such a way that it overlaps the defect by at least 3 cm (more commonly, 5 cm) on all sides, as in other ventral hernia repairs. This method appears to minimize the risk of mesh prolapse and bowel herniation alongside the stoma. The authors reported no recurrences within their 3- to 11-month follow-up period and no morbidity. In contrast, a subsequent study reported recurrences within 12 months in five of nine patients who underwent a variation of this repair, in which a slit in the mesh (instead of a central defect) was created and only one mesh sheet (instead of two) was used to cover the defect.71

Laparoscopic parastomal hernia repair appears to be a viable alternative to laparotomy or stoma relocation, but long-term multicenter evaluation is necessary for full assessment of this technique’s value in this setting.


Overall, fewer complications are reported after laparoscopic incisional herniorrhaphy than after open mesh repair.8,63,75–77 There are, however, several specific complications that are of particular relevance in laparoscopic procedures (see below).

Bowel Injury

A missed bowel injury is a potentially lethal complication. The overall incidence of bowel injury does not differ significantly between open repair and laparoscopic repair and is generally low with either approach (1% to 5% when serosal injuries are included). It should be noted, however, that pneumoperitoneum may hinder the recognition of bowel injury at the time of operation. There have also been several reports of late bowel perforation secondary to thermal injury with laparoscopic repair.65,76–79 One study reported two bowel injuries that were not discovered until sepsis developed; these late discoveries resulted in multiple operations, removal of the mesh, prolonged hospital stay, and, in one patient, death. The incidence of bowel injury is likely to be higher with less experienced surgeons80 and in patients who require extensive adhesiolysis. In one series describing a surgeon’s first 100 cases, four of six inadvertent enterotomies were made in the first 25 cases.81

To reduce the risk of bowel injury, we strongly discourage the use of electrocauterization and ultrasonic dissection. The visualization afforded by the pneumoperitoneum, which helps place adhesions between the abdominal wall and the bowel under tension, and the magnification afforded by the laparoscope facilitate identification of the least vascularized planes. As far as possible, we avoid grasping the bowel itself, preferring simply to push it or to grasp the adhesions themselves to provide countertraction. External pressure on the hernia may also help. Larger vessels in the omentum or adhesions are controlled with clips. Some degree of oozing from the dissected areas is tolerated; such oozing almost always settles down without specific hemostatic measures.

As noted, if dense adhesions are present, it is preferable to divide the sac or the fascia rather than risk injury to bowel. In the case of densely adherent polypropylene mesh, it may be better to excise it from the abdominal wall rather than attempt to separate it from the serosa of the bowel. If there is reason to suspect that bowel injury may have occurred, immediate and thorough inspection should be carried out; if this is not done, it may be difficult or impossible to find the exact site of injury later, once the bowel has been released after being freed of its attachments. If an injury to the bowel is recognized, the extent of the injury and the surgeon’s level of comfort with laparoscopic suture repair will determine the best approach. If there is no or minimal spillage of bowel contents, the injury may be treated with either laparoscopic repair or open repair; the latter usually can be carried out through a several centimeter counterincision over the injured area. Whether a mesh prosthesis will be placed depends on the degree of contamination. If there has been minimal or no contamination, any small open incision is closed, and laparoscopic lysis of adhesions and mesh placement can continue. If the contamination is more significant, adhesiolysis is completed, but the patient is brought back for mesh placement at another date. More significant bowel injuries may necessitate conversion to open repair.

Chronic Pain

In about 5% of patients, the transfascial fixation sutures used to secure the mesh cause pain that lasts more than 2 months. In most cases, postoperative pain decreases over time; if it does not, injection of local anesthetic into the area around the painful suture may be helpful.82


Seroma formation is one of the most commonly reported complications: it occurs immediately after operation in virtually all patients, to some extent.65,80,83 Patients sometimes mistake a tense seroma for recurring incisional hernia, but appropriate preoperative discussion should provide them with significant reassurance on this point. Seroma formation seems not to be related to particular mesh types or to the use of drains.80 Virtually all seromas resolve spontaneously over a period of weeks to months, with fewer than 5% persisting for more than 8 weeks.64 They are rarely clinically significant. Aspiration may increase the risk of mesh infection.84


Overall wound complication rates have been shown to be lower with laparoscopic incisional herniorrhaphy than with open repair.65,79,80,85 In particular, SSIs appear to be reduced with the laparoscopic approach.65,79,80,85 SSI rates for open incisional repair range from 5% to 20%,63,79,80,83 whereas those for laparoscopic repair range from 1% to 8%.5,65,72,79 Mesh infection extensive enough to necessitate mesh removal is rare (incidence ~ 1%); however, SSI after laparoscopic repair, especially when PTFE is used, more frequently results in seeding of the mesh.65,80 Because this type of mesh does not become well incorporated, antibiotic treatment alone is ineffective.80

Outcome Evaluation

Several studies have demonstrated improvements in outcome measures’such as decreased postoperative pain, shorter length of stay, earlier return to work,86 and reduced blood loss—with the laparoscopic approach to incisional hernia repair.5,8,63,64,76,84,87–89 The only randomized trial published to date compared 30 laparoscopic incisional hernia repairs with 30 open repairs.6 The mean operating time was significantly shorter in the laparoscopic group (87 minutes versus 111.5 minutes), as was the postoperative length of stay (2.2 days versus 9.1 days). Another study compared 56 prospective laparoscopic incisional hernia repairs with 49 open incisional hernia repairs assessed through retrospective chart review.8 The laparoscopic and open groups were comparable in terms of patient characteristics and hernia size. Although the mean operating time was longer in the laparoscopic group (95.4 minutes versus 78.5 minutes), the postoperative length of stay was significantly shorter after laparoscopic repair (3.4 days versus 6.5 days). In contrast, some studies have found the operating time to be 30 to 40 minutes shorter with laparoscopic repair.63,76

The recurrence rate—the primary long-term outcome measure of interest—is reported to be reduced after laparoscopic repair. In studies comparing laparoscopic and open incisional hernia repair with mesh, the recurrence rates after the laparoscopic repairs ranged from 0% to 11%, whereas those after the open repairs ranged from 5% to 35%.6,8,78,86,88 In expert hands, recurrence rates are low after laparoscopic repair. For example, in a multicenter series of 850 laparoscopic incisional hernia repairs, mostly followed prospectively, the recurrence rate after a mean follow-up period of 20 months was 4.7%.77 One group found no recurrences in the laparoscopic group and two in the open group after a minimum follow-up period of 18 months.76 Another group, however, reported a recurrence rate of 18%,90 which approached that reported after open mesh repair. Recurrence has been associated with lack of suture fixation, prostheses that overlap the defect by less than 2 to 3 cm, postoperative complications, and previous repairs.64,65,77,91

The surgeon’s level of experience plays a significant role in patient outcome, as demonstrated by a group that compared the outcomes for their first 100 laparoscopic incisional hernia repair patients with those for their second 100.84 Recurrence rates after a mean follow-up period of 36 months dropped from 9% in the first 100 patients to 4% in the second 100. In addition, the second set of patients were an average of 9 years older, had a higher percentage of recurrent hernias, and exhibited more comorbidities, yet despite these added challenges, operating time was not lengthened, length of stay was similarly short, and the complication rate was no different. Another group reported similar findings for their laparoscopic incisional hernia repair learning curve.92 Operating times and complication rates in the first 32 patients were comparable to those in the second 32; however, bowel injuries were more common in the first 32.

Although the results of large randomized trials are not available yet, the evidence to date suggests that the laparoscopic approach to the repair of large incisional hernias is highly promising. The laparoscopic approach seems to be safe and compares favorably with the open operation in terms of complication and recurrence rates.


Robert J. Fitzgibbons, Jr., MD, FACS

Harry E. Stuckenhoff Professor, Department of Surgery
Creighton University School of Medicine

Alan T. Richards, MD, FACS

Associate Professor, Staff Surgeon
Creighton University School of Medicine
Staff Surgeon
University of Nebraska Medical Center

Thomas H. Quinn, PH.D.

Professor, Department of Biomedical Sciences
Creighton University School of Medicine

Abdominal wall hernias are so common that their management constitutes the largest part of the average general surgeon’s practice. In the past, personal recollections and single-center series written by experts with a vested interest in publishing good results were the principal data sources that surgeons relied on in choosing the optimum treatment strategy for a patient. In recent years, fortunately, population-based studies have provided much better data on the true failure rates associated with the various herniorrhaphies. In addition, trials designed to examine the natural history of hernias have shed some light on nonoperative treatment options.

In this chapter, we describe many different operations for abdominal wall hernias. A well-known surgical dictum states that when numerous different operations exist to treat the same disease, the perfect procedure does not exist. This dictum does not hold true for abdominal wall herniorrhaphy, however. Because the disease is so heterogeneous, many different procedures are needed to address individual patients’ needs; thus, it can be said that multiple perfect procedures exist.


In the United States, approximately 1,000,000 abdominal wall herniorrhaphies are performed each year, of which 750,000 are for inguinal hernias, 166,000 for umbilical hernias, 97,000 for incisional hernias, 25,000 for femoral hernias, and 76,000 for miscellaneous hernias.1 About 75% of all abdominal wall hernias occur in the groin. Worldwide, some 20 million groin hernias are repaired each year.2 Inguinal hernias are more common on the right side than on the left. They occur seven times more frequently in males than in females; only 8% of groin hernia repairs are performed in women. Femoral hernias account for fewer than 10% of all groin hernias; however, 40% present as emergencies (i.e., with incarceration or strangulation), and mortality is higher for emergency repair than for elective repair. In male patients, indirect inguinal hernias are the most common type, occurring approximately twice as frequently as direct inguinal hernias; femoral hernias account for a much smaller percentage. In female patients, indirect inguinal hernias are also the most common type, but femoral hernias are seen more frequently than direct hernias, which are rare in this population. Emergency operations are more frequently required for female patients. In a study from the Swedish Hernia Registry that analyzed 90,648 inguinal hernia operations (88,753 in men, 6,895 in women) between 1992 and 2003, emergency operations were more frequently needed in women (16.9%) than in men (5.0%), leading to bowel resection in 16.6% and 5.6% of cases, respectively.3 Femoral recurrences were particularly common in women whose diagnosis at the time of the primary repair was direct or indirect hernia (41.6%, compared with 4.6% for men), a finding strongly suggesting that a hernia was missed at the original procedure. Femoral hernias are also more common in older patients and in those who have previously undergone inguinal hernia repair.

The prevalence of abdominal wall hernias is difficult to determine, as the wide range of published figures in the literature illustrates. The major reasons for this difficulty are (1) the lack of standardization in how inguinal and ventral hernias are defined, (2) the inconsistency of the data sources used (which include self-reporting by patients, audits of routine physical examinations, and insurance company databases, among others), and (3) the subjectivity of physical examination, even when performed by trained surgeons. Most authorities, however, subscribe to the two-peak theory for inguinal hernias, which states that that a new diagnosis of an inguinal hernia is most likely in patients younger than 1 year and in patients older than 55 years. Clearly, though, hernias can be diagnosed across any given age group.4 A 1996 analysis of a geographically defined population in the United Kingdom estimated that the lifetime risk of having to undergo an inguinal hernia repair was 27% for men and 3% for women.5

The incidence of the most common type of ventral hernia, incisional hernia, depends on how the condition is defined. The best definition of incisional hernia is any abdominal wall gap, with or without a bulge, that is perceptible on clinical examination or diagnostic imaging within 1 year after the index operation. A definition that requires the presence of a visible bulge will lead to underestimation of the true incidence of the condition. The reported incidence of incisional hernia after a midline laparotomy ranges from 3% and 20%, and it doubles if the index operation was associated with infection. Incisional hernias are most common after midline and transverse incisions, but they are also well documented after paramedian, subcostal, McBurney (gridiron), and Pfannenstiel incisions.6 An analysis of 11 publications dealing with ventral hernia incidence after various types of incisions concluded that the risk was 10.5% for midline incisions, 7.5% for transverse incisions, and 2.5% for paramedian incisions.7 Upper midline incisions are associated with the highest incidence of ventral hernia formation, transverse or oblique incisions with the lowest. Muscle-splitting incisions probably have a lower incidence of incisional hernias, but such incisions restrict access to the abdominal cavity. Most incisional hernias are detected within 1 year of surgery; the most common cause is believed to be separation of aponeurotic edges in the early postoperative period. The male-to-female incidence ratio is 1:1, even though early evisceration is more common in males.

At present, little information is available on the risk of major complications arising from untreated abdominal wall hernias. The main reason for this scarcity of data is that surgeons are taught, first, that all hernias, even asymptomatic ones, should be repaired at diagnosis to prevent potential strangulation or bowel obstruction, and second, that herniorrhaphy becomes more difficult the longer repair is delayed. As a result, it is difficult to find a whole population in which at least some of the members do not routinely have their hernias repaired regardless of symptoms. In these circumstances, accurate estimates of the natural history of the disease are impossible.

The natural history of an untreated, minimally symptomatic inguinal hernia was addressed in a randomized, controlled trial from 2006, in which 364 men were assigned to ‘watchful waiting’ (WW), and 356 men underwent routine operation.8 Only two patients in the WW group required emergency operations for strangulation over the follow-up period of 2 to 4.5 years. This result translated into a rate of 1.8 per 1,000 patient-years (0.18%), or about one fifth of 1% for each year that the hernia remains unrepaired. The two patients who required emergency operations recovered uneventfully. The question that remained to be answered was, which group fared better overall, the WW group or the group whose hernias were repaired immediately in accordance with conventional teaching? The answer to this question was at variance with conventional assumptions. At the conclusion of the study, functional status, as measured by quality-of-life instruments and pain scales, was identical in the two groups. About one third of the patients in the WW group crossed over to undergo operative treatment, principally because of symptom progression. However, there appeared to be no penalty for delaying surgery. Before this study, most surgeons assumed that a hernia would become harder to repair the longer it remained (because of enlargement and buildup of scar tissue) and that patients whose operations were delayed would experience more complications. The investigators found, however, that postoperative complication rates were the same in patients who underwent immediate surgery as in those who were assigned to watchful waiting but had to cross over to surgical treatment.

Classification of Inguinal and Ventral Hernias

Numerous classification schemes for groin hernias have been devised, usually bearing the name of the responsible investigator or investigators (e.g., Casten, Lichtenstein, Gilbert, Robbins and Rutkow, Bendavid, Nyhus, Schumpelick, and Zollinger). The variety of classifications in current use indicates that the perfect system has yet to be developed.9 The main problem in developing a single classification scheme suitable for wide application is that it is impossible to eliminate subjective measurements so as to ensure consistency from observer to observer. The advent of laparoscopic herniorrhaphy has further complicated the issue in that some of the measurements needed cannot be obtained via a laparoscopic approach. At present, the Nyhus system enjoys the greatest degree of acceptance [see Table 1].

A classification system for abdominal wall hernias outside the groin has been proposed by Zollinger [see Table 2].10 Ventral incisional hernias are common enough to warrant their own discrete classification system. The scheme most often used for categorizing incisional hernias [see Table 3] was the result of a 1998 consensus conference held in conjunction with the European Hernia Society’s annual congress.11 This system is important in that it affords investigators a reliable means of comparing results between one procedure and another or between one center and another.

Abdominal Wall Anatomy

The skin of the lower anterior abdominal wall is innervated by anterior and lateral cutaneous branches of the ventral rami of the seventh through 12th intercostal nerves and by the ventral rami of the first and second lumbar nerves. These nerves course between the lateral flat muscles of the abdominal wall and enter the skin through the subcutaneous tissue.

The first layers encountered beneath the skin are Camper’s and Scarpa’s fasciae in the subcutaneous tissue. The only significance of these layers is that when sufficiently developed, they can be reapproximated to provide another layer between a repaired abdominal wall and the outside. The major blood vessels of this superficial fatty layer are the superficial inferior and superior epigastric vessels, the intercostal vessels, and the superficial circumflex iliac vessels (which are branches of the femoral vessels).

Figure 1. Flat muscles of the abdominal wall

The external oblique muscle is the most superficial of the great flat muscles of the abdominal wall [see Figure 1]. This muscle arises from the posterior aspects of the lower eight ribs and interdigitates with both the serratus anterior and the latissimus dorsi at its origin. The posterior portion of the external oblique muscle is oriented vertically and inserts on the crest of the ilium. The anterior portion of the muscle courses inferiorly and obliquely toward the midline and the pubis. The muscle fibers give way to form its aponeurosis, which occurs well above the inguinal region. The obliquely arranged anterior inferior fibers of the aponeurosis of the external oblique muscle fold back upon themselves to form the inguinal ligament, which attaches laterally to the anterior superior iliac spine. In most persons, the medial insertion of the inguinal ligament is dual: one portion of the ligament inserts on the pubic tubercle and the pubic bone, whereas the other portion is fan-shaped and spans the distance between the inguinal ligament proper and the pectineal line of the pubis. This fan-shaped portion of the inguinal ligament is called the lacunar ligament. It blends laterally with Cooper’s ligament (or, to be anatomically correct, the pectineal ligament). The more medial fibers of the aponeurosis of the external oblique muscle divide into a medial crus and a lateral crus to form the external or superficial inguinal ring, through which the spermatic cord (in females, the round ligament) and branches of the ilioinguinal and genitofemoral nerves pass. The rest of the medial fibers insert into the linea alba after contributing to the anterior portion of the rectus sheath.

Beneath the external oblique muscle is the internal oblique muscle. The fibers of the internal oblique muscle fan out following the shape of the iliac crest, so that the superior fibers course obliquely upward toward the distal ends of the lower three or four ribs while the lower fibers orient themselves inferomedially toward the pubis to run parallel to the external oblique aponeurotic fibers. These fibers arch over the round ligament or the spermatic cord, forming the superficial part of the internal (deep) inguinal ring.

Beneath the internal oblique muscle is the transversus abdominis. This muscle arises from the inguinal ligament, the inner side of the iliac crest, the endoabdominal fascia, and the lower six costal cartilages and ribs, where it interdigitates with the lateral diaphragmatic fibers. The medial aponeurotic fibers of the transversus abdominis contribute to the rectus sheath and insert on the pecten ossis pubis and the crest of the pubis, forming the falx inguinalis. Infrequently, these fibers are joined by a portion of the internal oblique aponeurosis; only when this occurs is a true conjoined tendon formed.12

Aponeurotic fibers of the transversus abdominis also form the structure known as the aponeurotic arch. It is theorized that contraction of the transversus abdominis causes the arch to move downward toward the inguinal ligament, thereby constituting a form of shutter mechanism that reinforces the weakest area of the groin when intra-abdominal pressure is raised. The area beneath the arch varies. Many authorities believe that a high arch, resulting in a larger area from which the transversus abdominis is by definition absent, is a predisposing factor for a direct inguinal hernia. The transverse aponeurotic arch is also important because the term is used by many authors to describe the medial structure that is sewn to the inguinal ligament in many of the older inguinal hernia repairs.

The rectus abdominis forms the central anchoring muscle mass of the anterior abdomen. It arises from the fifth through seventh costal cartilages and inserts on the pubic symphysis and the pubic crest. It is innervated by the seventh through 12th intercostal nerves, which laterally pierce the aponeurotic sheath of the muscle. The semilunar line is the slight depression in the aponeurotic fibers coursing towards the muscle. In a minority of persons, the small pyramidalis muscle accompanies the rectus abdominis at its insertion. This muscle arises from the pubic symphysis. It lies within the rectus sheath and tapers to attach to the linea alba, which represents the conjunction of the two rectus sheaths and is the major site of insertion for three aponeuroses from all three lateral muscle layers. The line of Douglas (i.e., the arcuate line of the rectus sheath) is formed at a variable distance between the umbilicus and the inguinal space because the fasciae of the large flat muscles of the abdominal wall contribute their aponeuroses to the anterior surface of the muscle, leaving only transversalis fascia to cover the posterior surface of the rectus abdominis.

Figure 2. Nerves of the lower abdominal wall

The innervation of the anterior wall muscles is multifaceted. The seventh through 12th intercostal nerves and the first and second lumbar nerves provide most of the innervation of the lateral muscles, as well as of the rectus abdominis and the overlying skin. The nerves pass anteriorly in a plane between the internal oblique muscle and the transversus abdominis, eventually piercing the lateral aspect of the rectus sheath to innervate the muscle therein. The external oblique muscle receives branches of the intercostal nerves, which penetrate the internal oblique muscle to reach it. The anterior ends of the nerves form part of the cutaneous innervation of the abdominal wall. The first lumbar nerve divides into the ilioinguinal nerve and the iliohypogastric nerve [see Figure 2]. These important nerves lie in the space between the internal oblique muscle and the external oblique aponeurosis. They may divide within the psoas major or between the internal oblique muscle and the transversus abdominis. The ilioinguinal nerve may communicate with the iliohypogastric nerve before innervating the internal oblique muscle. The ilioinguinal nerve then passes through the external inguinal ring to run parallel to the spermatic cord, while the iliohypogastric nerve pierces the external oblique muscle to innervate the skin above the pubis. The cremaster muscle fibers, which are derived from the internal oblique muscle, are innervated by the genitofemoral nerve. There can be considerable variability and overlap.

The blood supply of the lateral muscles of the anterior wall comes primarily from the lower three or four intercostal arteries, the deep circumflex iliac artery, and the lumbar arteries. The rectus abdominis has a complicated blood supply that derives from the superior epigastric artery (a terminal branch of the internal thoracic [internal mammary] artery), the inferior epigastric artery (a branch of the external iliac artery), and the lower intercostal arteries. The lower intercostal arteries enter the sides of the muscle after traveling between the oblique muscles; the superior and the inferior epigastric arteries enter the rectus sheath and anastomose near the umbilicus.

The endoabdominal fascia is the deep fascia covering the internal surface of the transversus abdominis, the iliacus, the psoas major and minor, the obturator internus, and portions of the periosteum. It is a continuous sheet that extends throughout the extraperitoneal space and is sometimes referred to as the wallpaper of the abdominal cavity. Commonly, the endoabdominal fascia is subclassified according to the muscle being covered (e.g., iliac fascia or obturator fascia).

The transversalis fascia is particularly important for inguinal hernia repair because it forms anatomic landmarks known as analogues or derivatives. The most significant of these analogues for groin hernia surgeons are the iliopectineal arch, the iliopubic tract, the crura of the deep inguinal ring, and Cooper’s ligament (i.e., the pectineal ligament). The superior and inferior crura form a ‘monk’s hood’-shaped sling around the deep inguinal ring. This sling has functional significance, in that as the crura of the ring are pulled upward and laterally by the contraction of the transversus abdominis, a valvular action is generated that helps preclude indirect hernia formation. The iliopubic tract is the thickened band of the transversalis fascia that courses parallel to the more superficially located inguinal ligament. It is attached to the iliac crest laterally and inserts on the pubic tubercle medially. The insertion curves inferolaterally for 1 to 2 cm along the pectineal line of the pubis to blend with Cooper’s ligament, ending at about the midportion of the superior pubic ramus. Cooper’s ligament is actually a condensation of the periosteum and is not a true analogue of the transversalis fascia.

Hesselbach’s inguinal triangle is the site of direct inguinal hernias. As viewed from the anterior aspect, the inguinal ligament forms the base of the triangle, the edge of the rectus abdominis forms the medial border, and the inferior epigastric vessels form the superolateral border. (It should be noted, however, that Hesselbach actually described Cooper’s ligament as the base.)

Below the iliopubic tract are the critical anatomic elements from which a femoral hernia may develop. The iliopectineal arch separates the vascular compartment that contains the femoral vessels from the neuromuscular compartment that contains the iliopsoas muscle, the femoral nerve, and the lateral femoral cutaneous nerve. The vascular compartment is invested by the femoral sheath, which has three subcompartments: (1) the lateral, containing the femoral artery and the femoral branch of the genitofemoral nerve; (2) the middle, containing the femoral vein; and (3) the medial, which is the cone-shaped cul-de-sac known as the femoral canal. The femoral canal is normally a 1 to 2 cm blind pouch that begins at the femoral ring and extends to the level of the fossa ovalis. The femoral ring is bordered by the superior pubic ramus inferiorly, the femoral vein laterally, and the iliopubic tract (with its curved insertion onto the pubic ramus) anteriorly and medially. The femoral canal normally contains preperitoneal fat, connective tissue, and lymph nodes (including Cloquet’s node at the femoral ring), which collectively make up the femoral pad. This pad acts as a cushion for the femoral vein, allowing expansion such as might occur during a Valsalva maneuver, and serves as a plug to prevent abdominal contents from entering the thigh. A femoral hernia exists when the blind end of the femoral canal becomes an opening (the femoral orifice) through which a peritoneal sac can protrude.

Between the transversalis fascia and the peritoneum is the preperitoneal space. In the midline behind the pubis, this space is known as the space of Retzius; laterally, it is referred to as the space of Bogros. The preperitoneal space is of particular importance for surgeons because many of the inguinal hernia repairs (see below) are performed in this area. The inferior epigastric vessels, the deep inferior epigastric vein, the iliopubic vein, the rectusial vein, the retropubic vein, the communicating rectusioepigastric vein, the internal spermatic vessels, and the vas deferens are all encountered in this space.13

Choice of Prosthetic Material

For most abdominal wall hernias, the procedure of choice includes the use of a prosthesis. A detailed discussion comparing and contrasting various prosthetic materials is beyond the scope of this chapter; however, some general statements may be made. As a rule, North American surgeons tend to consider polypropylene mesh the favored prosthetic material, whereas European surgeons are more likely to employ polyester mesh. Of course, the use of mesh presupposes a situation in which the prosthesis can be isolated from contact with intra-abdominal viscera by one or more layers of human tissue (e.g., peritoneum). In situations where contact with intra-abdominal viscera cannot be avoided, a standard mesh prosthesis should not be used. Either the prosthesis should be composed of a nonmesh material, such as expanded polytetrafluoroethylene (ePTFE), or a dual-layer prosthesis should be used, with a standard plastic mesh on the side facing the abdominal wall (to encourage an intense fibroplastic response) and an adhesion barrier of some type coating the peritoneal side. Numerous dual-sided prosthetics, incorporating a variety of adhesion barriers, are now available [see Table 4]. It has consistently been shown that when these materials are used, adhesions are not only less common but also less tenacious than when mesh alone is used. Often, bowel adhesions can be literally wiped from the peritoneal surface of a dual-layer prosthesis with gentle blunt traction, in sharp contrast to the typically tedious and sometimes impossible dissection of bowel loops from a mesh prosthesis. Although all of the dual-layer prostheses currently on the market are approved for decreasing adhesions to the adhesion barrier side, no manufacturer has sought approval for complete prevention of adhesions. Consequently, the long-term effects of these less severe (but still present) adhesions are unknown; further study is required to address this issue.

A number of so-called biologic prostheses have been developed that are designed to promote vessel ingrowth and eventual remodeling of tissue to resemble the native type [see Table 5]. Although biologic prostheses are much more expensive than synthetic prostheses, they may be the better choice when the operative field is contaminated or when an abdominal wall defect is so large that the prosthesis cannot be covered by skin. Clearly, more study is required before their exact place in the armamentarium of the abdominal wall hernia surgeon can be determined.

At present, there is some controversy regarding the weight of the polypropylene mesh used in abdominal wall hernia repairs. (The controversy almost certainly applies to the other types of mesh prosthesis as well.) Data from randomized studies indicate that use of a lightweight mesh results in less long-term pain than use of a normal mesh, without having any negative effect on the recurrence rate.14,15 Lighter-weight mesh also addresses the theoretical concern about the possible carcinogenic effects of polypropylene, as has been suggested by experimental studies in rats, though it should be kept in mind that there has never been a documented case of a sarcoma developing in a human being as a result of an inguinal hernia prosthesis.16 To illustrate the difference between a lightweight mesh and a normal one, a 7.5 × 15 cm piece of polypropylene mesh (Prolene; Ethicon, Inc., Somerville, New Jersey) weighs about 80 g/m2, whereas a similarly sized piece of a polypropylene-poliglecaprone 25 (Monocryl; Ethicon, Inc., Somerville, New Jersey) lightweight mesh (UltraPro; Ethicon, Inc., Somerville, New Jersey) weighs less than 30 g/m2 after absorption of the poliglecaprone 25 component. North American surgeons have been slow to accept the use of lightweight mesh for inguinal hernia repair, fearing a higher recurrence rate (as was suggested by one of the earlier randomized trials).17 Many also have some concerns about possible bias in the data, noting that the research supporting the use of lightweight mesh has been almost exclusively funded by industry. Nevertheless, the randomized trials mentioned earlier cannot be entirely discounted.

Inguinal Hernia Repair: Choice of Procedure

Practical considerations do not allow a description of every single named inguinal hernia repair in the literature. The nonprosthetic named repairs alone number more than 70.18 For the purposes of this chapter, inguinal hernia repairs may be grouped according to (1) whether the operation makes use of the anterior space, the posterior space, or both and (2) whether a prosthesis is included or omitted [see Table 6]. In reality, most of the numerous eponyms used to name inguinal herniorrhaphies refer not to fundamentally distinct operations but, rather, to relatively minor modifications of standard hernia procedures [see Table 6]. Accordingly, rather than address every known variant, we describe only the major repairs on which these variants are based.

The most important consideration in choosing an inguinal hernia procedure is the experience of the surgeon. Knowing the ideal operation for a given clinical scenario is of no significance if the surgeon is not skilled in performing it. On the assumption that the surgeon’s expertise is equal to the task, the next consideration should be to tailor the operation to the patient’s particular hernia. For example, a simple Marcy repair would be completely adequate for a pediatric patient with a Nyhus type 1 hernia [see Table 1] but not for an elderly patient who has an indirect hernia in conjunction with extensive destruction of the inguinal floor. The conventional anterior prosthetic repairs are particularly useful in high-risk patients because they can easily be performed with local anesthesia.19 On the other hand, giant prosthetic reinforcement of the visceral sac (GPRVS), especially when bilateral, necessitates general or regional anesthesia and thus is best for patients with bilateral direct or recurrent hernias or, perhaps, for patients with connective tissue disorders that appear to be associated with their hernia. If surgery has previously been done in either the anterior or the preperitoneal space, the surgeon should choose a procedure that uses the undissected space. If local or systemic infection is present, a nonprosthetic repair is usually considered preferable, though the newer biologic prostheses now being evaluated may eventually change this view. Uncorrected coagulopathy is a contraindication to elective repair.

Inguinal Hernia Repair: Operative Technique

Anterior Herniorrhaphy

Steps Common to Prosthetic and Nonprosthetic Repairs

The various anterior herniorrhaphies have a number of initial technical steps in common; they differ primarily with respect to the specific details of floor reconstruction.

Step 1: choice of anesthetic Local anesthesia is entirely adequate, especially when combined with intravenous sedation. In specialty hernia clinics, it is the approach most commonly employed. In general practice, however, general anesthesia is the rule. This approach is reasonable in fit patients but is associated with a higher incidence of postoperative urinary retention.20 If general anesthesia is used, a local anesthetic should be given at the end of the procedure as an adjuvant to reduce immediate postoperative pain. Regional (spinal or epidural) anesthesia can also be used, but it is less popular, having the highest incidence of systemic side effects (primarily cardiovascular).19

We prefer local anesthesia combined with I.V. infusion of a rapid-acting, short-lasting, amnesic, and anxiolytic agent (e.g., propofol). This technique affords the patient all the benefits of general anesthesia in terms of comfort, without the higher incidence of urinary retention seen with regional or general endotracheal anesthesia. An added benefit is that the patient can be aroused from sedation periodically to perform Valsalva maneuvers to test the repair.

The techniques and drug dosages employed by different experts vary considerably. Compounding factors include the age of the patient and the amount of I.V. sedation used. Our preference is to use a solution containing 50 ml of 0.5% lidocaine with epinephrine and 50 ml of 0.25% bupivacaine with epinephrine. The epinephrine is optional and may be omitted if the patient has a history of coronary artery disease or if there is concern about delayed bleeding. In an adult of normal size, 70 ml of this solution is injected before preparation and draping: 10 ml is placed 1 cm medial and 1 cm inferior to the anterior superior iliac spine in an attempt to block the major nerves innervating the groin area [see Abdominal Wall Anatomy, above], and the other 60 ml is used as a field block along the orientation of the eventual incision in the subcutaneous and deeper tissues. Care is taken to ensure that some of the material is injected into the areas of the pubic tubercle and Cooper’s ligament, which are easily identified by tactile sensation (except in very obese patients). The remaining 30 ml of the solution is reserved for discretionary use during the procedure.


Step 2: initial incision Traditionally, the skin is opened by making an oblique incision between the anterior superior iliac spine and the pubic tubercle. For cosmetic reasons, however, many surgeons now prefer a more horizontal skin incision placed in the natural skin lines. In either case, the incision is deepened through Scarpa’s fasciae and the subcutaneous tissue to expose the external oblique aponeurosis. The external oblique aponeurosis is then opened through the external inguinal ring. If a prosthesis is to be used, a large space is created beneath the external oblique aponeurosis from the anterior rectus sheath medially to the anterior superior iliac spine laterally to prepare for the eventual placement.

Step 3: care of the sensory nerves The iliohypogastric nerve is identified; it can be either left in situ or freed from the surrounding tissue and isolated from the operative field by passing a hemostat under the nerve and grasping the upper flap of the external oblique aponeurosis. Routine division of the iliohypogastric nerve along with the ilioinguinal nerve is practiced by some but is not advised by most, though there does not seem to be any consistent correlation with postoperative groin pain either way. The ilioinguinal and genitofemoral nerves are usually left with the cord structures. The genitofemoral nerve cannot always be identified with certainty. It will be sacrificed in those procedures that include division of the cremaster muscle (e.g., Shouldice repair).

Step 4: mobilization of cord structures The cord structures are bluntly dissected away from the inferior flap of the external oblique aponeurosis to expose the inguinal ligament (shelving edge) and the iliopubic tract. This dissection is continued over the pubic tubercle and onto the anterior rectus sheath for at least 2 cm, defining the point where the most medial edge of a prosthesis will eventually be sutured if a Lichtenstein prosthetic repair is being performed. This measure facilitates en masse lifting of the cord structures with the fingers of one hand at the pubic tubercle so that the index finger can be passed underneath to meet the ipsilateral thumb or the fingers of the other hand. Mobilization of the cord structures is completed by means of blunt dissection, and a Penrose drain is placed around them so that they can be retracted during the procedure.

Step 5: division of cremaster muscle For decades, complete division of the cremaster muscle with concomitant sacrifice of the genitofemoral nerve was common practice, especially with indirect hernias. The purpose of this step was to facilitate identification of the sac and to lengthen the cord for better visualization of the inguinal floor. It is clear, however, that adequate exposure can almost always be obtained by opening the muscle longitudinally, which reduces the chances of damage to the cord and prevents testicular descent. Accordingly, the latter approach should be considered best practice unless circumstances argue for division of the muscle.

Step 6: management of hernial sac The term high ligation of the sac is used frequently in discussing inguinal hernia repair; its historical significance has ingrained it in the descriptions of most of the older operations. For our purposes in this chapter, high ligation of the sac should be considered equivalent to reduction of the sac into the preperitoneal space without excision. The two methods work equally well and are highly effective. Proponents of sac inversion believe that this measure results in less pain (because the richly innervated peritoneum is not incised) and may be less likely to cause adhesive complications. To date, however, no randomized trials have been done to determine whether this is so.21Sac eversion in lieu of excision does protect intra-abdominal viscera in cases of unrecognized incarcerated sac contents or sliding hernia.

Many surgeons (especially urologists) believe that complete excision of all indirect inguinal hernial sacs, even when inguinal-scrotal, is important for preventing excessive postoperative hydrocele formation. The downside of this practice is that the incidence of ischemic orchitis from excessive trauma to the cord rises substantially. The logical sequela of ischemic orchitis is testicular atrophy, though this presumed relationship has not been conclusively proved. In our view, it is better to divide an indirect inguinal hernial sac in the midportion of the inguinal canal once it is clear that the hernia is not sliding and no abdominal contents are present. The distal sac is not removed, but its anterior wall is opened as far distally as is convenient. We have not observed an increased incidence of hydroceles with this approach.

Direct hernial sacs are separated from the cord and other surrounding structures and reduced into the preperitoneal space. Dividing the superficial layers of the neck of the sac circumferentially—thereby, in effect, opening the inguinal floor—usually facilitates reduction and helps to maintain it while the prosthesis is being placed. The opening in the inguinal floor also allows the surgeon to palpate for a femoral hernia. Sutures can be used to maintain reduction of the sac, but they have no real strength in this setting; their main purpose is to allow the repair to proceed without being hindered by continual extrusion of the sac into the field, especially when the patient strains.


Step 7: repair of inguinal floor Methods of repairing the inguinal floor differ significantly among the various anterior herniorrhaphies and thus are described separately under the relevant headings [see Nonprosthetic Repairs and Prosthetic Repairs, below].

Step 8: relaxing incision A relaxing incision is employed only if a nonprosthetic repair is being performed. The incision is made through the anterior rectus sheath and down to the rectus abdominis, extending superiorly from the pubic tubercle for a variable distance, as determined by the degree of tension present. A so-called hockey-stick incision oriented laterally at the superior end is a common choice. The posterior rectus sheath is strong enough to prevent future incisional herniation. The relaxing incision works because as the anterior rectus sheath separates, the various components of the abdominal wall are displaced laterally and inferiorly.

Step 9: closure Closure of the external oblique fascia serves to reconstruct the superficial (external) ring. The external ring must be loose enough to prevent strangulation of the cord structures yet tight enough to ensure that an inexperienced examiner will not confuse a dilated ring with a recurrence. A dilated external ring is sometimes referred to as an industrial hernia, because over the years it has occasionally been a problem during preemployment physical examinations. Scarpa’s fascia and the skin are closed to complete the operation.

Nonprosthetic Repairs

Marcy repair

Figure 3. Marcy repair

The Marcy repair is the simplest nonprosthetic repair performed today. Its main indication is for treatment of Nyhus type 1 hernias (i.e., indirect inguinal hernias in which the internal ring is normal). It is appropriate for children and young adults in whom there is concern about the long-term effects of prosthetic material. The essential features of the Marcy repair are high ligation of the sac and narrowing of the internal ring. Displacing the cord structures laterally allows the placement of sutures through the muscular and fascial layers [see Figure 3].

Bassini repairEdoardo Bassini (1844–1924) is considered the father of modern inguinal hernia surgery. It was during the 19th century that many of the great anatomists—Scarpa, Cooper, Hesselbach, Bogros, Retzius, Cloquet, Gay, and others—made their discoveries. By combining high ligation of the hernial sac with reconstruction of the inguinal floor (based on the principles formulated by the 19th-century anatomists), as well as taking advantage of the developing disciplines of antisepsis and anesthesia, Bassini was able to reduce recurrence and morbidity substantially. Before Bassini’s achievements, elective herniorrhaphy was almost never recommended, because the results were so bad. Bassini’s operation, known as the radical cure, became the gold standard for inguinal hernia repair for most of the 20th century.

The initial steps in the procedure are as previously described [see Steps Common to Prosthetic and Nonprosthetic Repairs, above]. Bassini felt that the incision in the external oblique aponeurosis should be as superior as possible while still allowing the superficial external ring to be opened, so that the reapproximation suture line created later in the operation would not be directly over the suture line of the inguinal floor reconstruction.22 Whether this technical point is significant is debatable. Bassini also felt that lengthwise division of the cremaster muscle followed by resection was important for ensuring that an indirect hernial sac could not be missed and for achieving adequate exposure of the inguinal floor.

Figure 4a. Bassini repair

After performing the initial dissection and the reduction or ligation of the sac, Bassini began the reconstruction of the inguinal floor by opening the transversalis fascia from the internal inguinal ring to the pubic tubercle, thereby exposing the preperitoneal fat, which he then bluntly dissected away from the undersurface of the superior flap of the transversalis fascia [see Figure 4a]. This step allowed him to properly prepare the deepest structure in his famous ‘triple layer’ (comprising the transversalis fascia, the transversus abdominis, and the internal oblique muscle).

Figure 4b. Bassini repair

The first stitch in Bassini’s repair includes the triple layer superiorly and the periosteum of the medial side of the pubic tubercle, along with the rectus sheath. In current practice, however, most surgeons try to avoid the periosteum of the pubic tubercle so as to decrease the incidence of osteitis pubis. The repair is then continued laterally, and the triple layer is secured to the reflected inguinal ligament (Poupart’s ligament) with nonabsorbable sutures. The sutures are continued until the internal ring is closed on its medial side [see Figure 4b]. A relaxing incision was not part of Bassini’s original description but now is commonly added.

Concerns about injuries to neurovascular structures in the preperitoneal space and to the bladder led many surgeons, especially in North America, to abandon the opening of the transversalis fascia. The unfortunate consequence of this decision is that the proper development of the triple layer is severely compromised. In lieu of opening the floor, a forceps (e.g., an Allis clamp) is used to grasp tissue blindly in the hope of including the transversalis fascia and the transversus abdominis. The layer is then sutured, along with the internal oblique muscle, to the reflected inguinal ligament as in the classic Bassini repair. The structure grasped in this modified procedure is sometimes referred to as the conjoined tendon, but this term is not accurate, because of the variability in what is actually grasped in the clamp. This imprecise ‘good stuff to good stuff’ approach almost certainly accounts for the inferior results achieved with the Bassini procedure in the United States.


Maloney darn

Figure 5. Maloney darn

The Maloney darn derives its name from the way in which a long nylon suture is repeatedly passed between the tissues to create a weave that one might consider similar to a mesh. After initial preparation of the groin (see above), a continuous nylon suture is used to oppose the transversus abdominis, the rectus abdominis, the internal oblique muscle, and the transversalis fascia medially to Poupart’s ligament laterally. The suture is continued into the muscle around the cord and is woven in and out to form a reinforcement around the cord [see Figure 5]. On the lateral side of the cord, it is sutured to the inguinal ligament and tied. The darn is a second layer. The sutures are placed either parallel or in a criss-cross fashion and are plicated well into the inguinal ligament below. The darn must be carried well over the medial edge of the inguinal canal. Once the darn is complete, the external oblique fascia is closed over the cord structures. The Maloney darn can be considered a forerunner of the mesh repairs, in that the purpose of the darn is to provide a scaffold for tissue ingrowth.23

Shouldice repair Steps 1 through 6 of this repair are performed essentially as previously described [see Steps Common to Prosthetic and Nonprosthetic Repairs, above]. Particular importance is placed on freeing of the cord from its surrounding adhesions, resection of the cremaster muscle, high dissection of the hernial sac, and division of the transversalis fascia during the initial steps of the procedure.24A continuous nonabsorbable suture (typically of monofilament steel wire) is used to repair the floor. The Shouldice surgeons believe that a continuous suture distributes tension evenly and prevents potential defects between interrupted sutures that could lead to recurrence.

A second wire suture is started near the internal ring, approximating the internal oblique muscle and the transversus abdominis to a band of external oblique aponeurosis superficial and parallel to Poupart’s ligament—in effect, creating a second, artificial Poupart’s ligament. This third suture line ends at the pubic crest. The suture is then reversed, and a fourth suture line is constructed in a similar manner, superficial to the third line. At the Shouldice clinic, the cribriform fascia is always incised in the thigh, parallel to the inguinal ligament, to make the inner side of the lower flap of the external oblique aponeurosis available for these multiple layers. In general practice, however, this step is commonly omitted.

The results at the Shouldice clinic have been truly outstanding and continue to be so today. For a time, the Shouldice repair was the gold standard against which all newer procedures were compared. The major criticism of this operation is that it is difficult to teach because surgeons have problems understanding what is really being sewn to what. Unless one is specifically trained at the Shouldice clinic and has the opportunity to work with the surgeons there, one may find it hard to identify the various layers in the medial flap reliably and reproducibly—a step that is crucial for developing the multiple suture lines. To compound the difficulty, modifications developed outside the Shouldice clinic have given rise to different versions of the procedure. For example, some surgeons use three continuous layers instead of four for reconstruction of the inguinal floor.


McVay Cooper’s ligament repair This operation is similar to the Bassini repair, except that it uses Cooper’s ligament instead of the inguinal ligament for the medial portion of the repair. Interrupted sutures are placed from the pubic tubercle laterally along Cooper’s ligament, progressively narrowing the femoral ring; this constitutes the most common application of the repair—namely, treatment of a femoral hernia [see Figure 7]. The last stitch in Cooper’s ligament is known as a transition stitch and includes the inguinal ligament. This stitch has two purposes: (1) to complete the narrowing of the femoral ring by approximating the inguinal ligament to Cooper’s ligament, as well as to the medial tissue, and (2) to provide a smooth transition to the inguinal ligament over the femoral vessel so that the repair can be continued laterally (as in a Bassini repair). Given the considerable tension required to bridge such a large distance, a relaxing incision should always be used. In the view of many authorities, this tension results in more pain than is noted with other herniorrhaphies and predisposes to recurrence. For this reason, the McVay repair is rarely chosen today, the main exception being for treatment of a patient with a femoral hernia or a patient with specific contraindications to mesh repair.

Prosthetic Repairs

Lichtenstein repair This operation is now considered the gold standard for inguinal herniorrhaphy. The initial preparation of the inguinal floor does not differ substantially from that carried out in a nonprosthetic repair. The transversalis fascia is not opened—a practice that has occasionally been criticized on the grounds that it might cause an occult femoral hernia to be missed. To date, however, an excessive incidence of missed femoral hernias has not been reported in men. The situation may be different in women: evidence from the large population-based Swedish study cited earlier suggests that femoral recurrence is much more common than one might assume when the entire myopectineal orifice is not addressed (as is the case with a McVay procedure or any of the preperitoneal operations).3

Figure 8. Lichtenstein repair

The key to the operation is the placement of a large prosthesis (at least 15 × 10 cm for an adult) extending from a point 2 cm medial to the pubic tubercle (to prevent the pubic tubercle recurrences all too commonly seen with other operations) to the anterior superior iliac spine laterally. The medial end is rounded to correspond to the patient’s particular anatomy, and a continuous suture of either nonabsorbable or long-lasting absorbable material is begun between the prosthesis and the anterior rectus sheath 2 cm medial to the pubic tubercle [see Figure 8]. The suture is continued laterally in a locking fashion, securing the prosthesis to either side of the pubic tubercle (not into it) and then to the shelving edge of the inguinal ligament. The suture is tied at the internal ring.

A slit is made on the lateral side of the prosthesis to create two tails, a wider one (approximately two thirds of the total height) above and a narrower one below. The tails are positioned around the cord structures and placed beneath the external oblique aponeurosis laterally to about the anterior superior iliac spine, with the upper tail placed on top of the lower. A single interrupted suture is placed to secure the lower edge of the superior tail to the lower edge of the inferior tail and the inguinal ligament—thereby, in effect, creating a shutter valve. This step is considered crucial for preventing the indirect recurrences occasionally seen when the tails are simply reapproximated. The maneuver provides a cradling effect as well, preventing direct contact between the cut edges of the prosthesis and the cord structures, which could result in damage when linear approximation is used. The suture also incorporates the shelving edge of the inguinal ligament so as to create a domelike buckling effect over the direct space, thereby ensuring that there is no tension, especially when the patient assumes an upright position. The Lichtenstein group has developed a customized prosthesis with a built-in domelike configuration, which, in their view, makes suturing the approximated tails to the inguinal ligament unnecessary.

A few interrupted sutures are placed to attach the superior and medial aspects of the prostheses to the underlying internal oblique muscle and rectus fascia. Care is taken to tie these loosely (with an ‘air knot’) and to avoid placing them laterally so as to minimize the risk of damaging the intramuscular and therefore invisible portions of the important nerves. On occasion, the iliohypogastric nerve, which courses on top of the internal oblique muscle, penetrates the medial flap of the external oblique aponeurosis. In this situation, the prosthesis should be slit to accommodate the nerve. The prosthesis can be trimmed in situ, but enough laxity must be maintained to allow for the difference between the supine and upright positions, as well as for possible shrinkage of the mesh.

If a femoral hernia is recognized, the transversalis fascia is opened and the hernia reduced to expose Cooper’s ligament. The Lichtenstein group’s approach is still to suture the inferior edge of the prosthesis to the inguinal ligament. The femoral space is then addressed by suturing the posterior surface of the prosthesis to Cooper’s ligament, thereby covering the entire myopectineal orifice, and finally by completing the superior and lateral sutures. We prefer to forgo the approximation of the inferior edge of the prosthesis to the inguinal ligament in favor of using interrupted sutures between that edge and Cooper’s ligament, much as in a McVay repair (the ‘Fitztenstein’ technique). A transition stitch is required between the inferior edge of the prosthesis, Cooper’s ligament, and the inguinal ligament on the medial side of the femoral vein. This stitch closes the femoral canal and sets the stage for the lateral side of the prosthesis to be sutured to the inguinal ligament. The rest of the operation then proceeds in the same manner as a classic Lichtenstein repair.


Plug-and-patch repair

Figure 9. Gilbert plug-and-patch repair

The mesh plug technique was first developed by Gilbert and subsequently modified by Rutkow and Robbins, Millikan, and others [see Figure 9].25–27The groin is entered via a standard anterior approach. The hernial sac is dissected away from surrounding structures and reduced into the preperitoneal space. A flat sheet of polypropylene mesh is rolled up like a cigarette, tied, inserted in the defect, and secured with interrupted sutures to either the internal ring (for an indirect hernia) or the neck of the defect (for a direct hernia).

A prefabricated prosthesis that has the configuration of a flower is commercially available and is recommended by Rutkow and Robbins. This prosthesis is tailored to each patient’s particular anatomy by removing some of the ‘petals’ to avoid unnecessary bulk. Many surgeons consider this step important for preventing erosion into surrounding structures (e.g., the bladder); indeed, such complications have been reported, albeit rarely.

Millikan further modified the procedure by recommending that the inside petals be sewn to the ring of the defect. For an indirect hernia, the inside pedals are sewn to the internal oblique portion of the internal ring; this forces the outside of the prosthesis underneath the inner side of the defect and makes it act like a preperitoneal underlay. For direct hernias, the inside petals are sewn to Cooper’s ligament and the shelving edge of the inguinal ligament, as well as to the conjoined tendon; this, again, forces the outside of the prosthesis to act as an underlay.

The patch portion of the procedure is optional and involves placing a flat piece of polypropylene in the conventional inguinal space so that it widely overlaps the plug, much as in a Lichtenstein repair. The difference with a plug-and-patch repair is that only one or two sutures—or, sometimes, no sutures—are used to secure the flat prosthesis to the underlying inguinal floor. Some surgeons, however, place so many sutures that they have in effect performed a Lichtenstein operation on top of the plug (sometimes referred to as a ‘plugstenstein’ repair).

To the credit of its proponents, the plug-and-patch repair, in all of its varieties, has been skillfully presented and has rapidly taken a significant share of the overall inguinal hernia market. It is not only fast but also extremely easy to teach, which has made it popular in both private and academic centers. A randomized, controlled trial has shown it to be equivalent to the Lichtenstein repair in terms of recurrence and morbidity.28 However, numerous case reports in the literature have described removal of plugs for pain, migration, or erosion, and as a result, the plug-and-patch repair has been the focus of considerable medicolegal scrutiny.


Posterior (Preperitoneal) Herniorrhaphy

Nonprosthetic Repairs

A key technical issue in a preperitoneal hernia repair is how the surgeon chooses to enter the preperitoneal space. In fact, within this general class of repair, it is the method of entry into this space that constitutes the major difference between the various procedures.

Many approaches to the preperitoneal space have been described. For example, the space can be entered either anteriorly or posteriorly. If an anterior technique is to be used, the initial steps of the operation are similar to those of a conventional anterior herniorrhaphy. If a posterior technique is to be used, any of several incisions (lower midline, paramedian, or Pfannenstiel) will allow an extraperitoneal dissection. The preperitoneal space can also be entered transabdominally. This approach is useful when the patient is undergoing a laparotomy for some other condition and the hernia is to be repaired incidentally. Of course, the transabdominal preperitoneal laparoscopic repair described elsewhere [see 5:28 Laparoscopic Hernia Repair], by definition, enters the preperitoneal space from the abdomen.

Reed credits Annandale as being the first surgeon to describe the anterior method of gaining access to the preperitoneal space.29 Bassini’s operation, as classically performed, is technically an anterior preperitoneal operation, but it is never discussed in this group, because in the American variant of the procedure, the preperitoneal space is not entered. Cheatle suggested the posterior approach to the preperitoneal space for repair of an inguinal hernia but used a laparotomy to do it.30 Cheatle and Henry subsequently modified the operation so as to render it entirely extraperitoneal (the Cheatle-Henry approach), which made the procedure more acceptable to surgeons.31

The preperitoneal nonprosthetic method remained popular into the second half of the 20th century, championed by proponents such as Nyhus and Condon, who emphasized the importance of the iliopubic tract as the inferior border in primary closure of direct or indirect hernia defects.32 Today, however, these operations are of little more than historical significance, because it is now universally agreed that better results are obtained in this space when a prosthesis is used. Indeed, after 1975, Nyhus and Condon began routinely placing a 6 × 14 cm piece of polypropylene mesh to buttress the primary repair for all recurrent hernias.33 When contraindications to a prosthesis are present [see Table 7], most surgeons would opt for a conventional anterior herniorrhaphy (e.g., a Bassini or Shouldice repair) rather than a preperitoneal nonprosthetic herniorrhaphy.

Prosthetic Repairs

The most important step in any preperitoneal prosthetic repair is the placement of a large prothesis in the preperitoneal space on the abdominal side of the defect. The theoretical advantage of this measure is that whereas in a conventional repair, abdominal pressure might contribute to recurrence, in a preperitoneal repair, the abdominal pressure would actually help fix the mesh material against the abdominal wall, thereby adding strength to the repair. The hernia defect itself may or may not be closed, depending on the preference of the surgeon. The strength of the repair depends on the prosthesis rather than on closure of the defect; however, such closure may decrease the seroma formation that inevitably occurs at the site of the undisturbed residual sac. Although these seromas almost always are self-limiting and disappear with time, they can be confused with recurrences by both patients and referring physicians. Accordingly, some surgeons prefer to take every step possible to prevent them.

Read-Rives repair The posterior space is accessed directly through the groin, and thus, the initial part of a Read-Rives repair, including the opening of the inguinal floor, is much like that of a classic Bassini repair. The inferior epigastric vessels are identified and the preperitoneal space completely dissected. The spermatic cord is parietalized by separating the ductus deferens from the spermatic vessels. A 12 × 16 cm piece of mesh is positioned in the preperitoneal space deep to the inferior epigastric vessels and secured with three sutures placed in the pubic tubercle, in Cooper’s ligament, and in the psoas muscle laterally. The transversalis fascia is closed over the prosthesis and the cord structures replaced. The rest of the closure is accomplished much as in a conventional anterior prosthetic repair.

Stoppa-Rignault-Wantz repair (GPRVS)

Figure 10. Orifice of Fruchaud

GPRVS has its roots in the important contribution that Henri Fruchaud made to herniology. In describing the myopectineal orifice that bears his name [see Figure 10], Fruchaud, who was Stoppa’s mentor, popularized a different viewpoint on the etiology of inguinal hernias.34Instead of subdividing hernias into direct, indirect, and femoral and then examining their specific causes, he emphasized that the common cause of all inguinal hernias was the failure of the transversalis fascia to retain the peritoneum. This concept led Stoppa to develop GPRVS, which reestablishes the integrity of the peritoneal sac by inserting a large permanent prosthesis that entirely replaces the transversalis fascia over the myopectineal orifice of Fruchaud with wide overlapping of surrounding tissue. With GPRVS, the exact type of hernia present (direct, indirect, or femoral) is unimportant, because the abdominal wall defect is not addressed.

Step 1: skin incision. A lower midline, inguinal, or Pfannenstiel incision may be used, depending on the surgeon’s preference. The inguinal incision is placed 2 to 3 cm below the level of the anterior superior iliac spine but above the internal ring; it is begun at the midline and extended laterally for 8 to 9 cm.35

Step 2: preperitoneal dissection. The fascia overlying the space of Retzius is opened without violation of the peritoneum. A combination of blunt and sharp dissection is continued laterally posterior to the rectus abdominis and the inferior epigastric vessels. The preperitoneal space is completely dissected to a point lateral to the anterior superior iliac spine [see Figure 11]. The symphysis pubis, Cooper’s ligament, and the iliopubic tract are identified. Inferiorly, the peritoneum is generously dissected away from the vas deferens and the internal spermatic vessels to create a large pocket, which will eventually accommodate a prosthesis without the possibility of rollup. In the inguinal approach, the anterior rectus sheath and the oblique muscles are incised for the length of the skin incision. The lower flaps of these structures are retracted inferiorly toward the pubis. The transversalis fascia is incised along the lateral edge of the rectus abdominis, and the preperitoneal space is entered; dissection then proceeds as previously indicated.

Step 3: management of hernial sac. Direct hernial sacs are reduced during the course of the preperitoneal dissection. Care must be taken to stay in the plane between the peritoneum and the transversalis fascia, allowing the latter structure to retract into the hernia defect toward the skin. The transversalis fascia can be thin, and if it is inadvertently opened and incorporated with the peritoneal sac during reduction, a needless and bloody dissection of the abdominal wall is the result.

Indirect sacs are more difficult to deal with than direct sacs are, in that they often adhere to the cord structures. Trauma to the cord must be minimized to prevent damage to the vas deferens or the testicular blood supply. Small sacs should be mobilized from the cord structures and reduced back into the peritoneal cavity. Large sacs may be difficult to mobilize from the cord without undue trauma if an attempt is made to remove the sac in its entirety. Accordingly, large sacs should be divided, with the distal portion left in situ and the proximal portion dissected away from the cord structures. Division of the sac is most easily accomplished by opening the sac on the side opposite the cord structures. A finger is placed in the sac to facilitate its separation from the cord. Downward traction is then placed on the cord structures to reduce any excessive fatty tissue (so-called lipoma of the cord) back into the preperitoneal space. This step prevents the ‘pseudorecurrences’ that may occur if the abnormality palpated during the preoperative physical examination was not a hernia but a lipoma of the cord.

Figure 12. Preperitoneal repair: placement of mesh

Step 4: management of abdominal wall defect. It is this step that varies most from one author to another. In Nyhus’s approach, the defect is formally repaired, and only then is a tailored mesh prosthesis sutured to Cooper’s ligament and the transversalis fascia for reinforcement [see Figure 12]. Rignault prefers to close the defect loosely to prevent an unsightly early postoperative bulge.36 In Stoppa’s and Wantz’s technique, the defect is usually left alone, but the transversalis fascia in the defect is occasionally plicated by suturing it to Cooper’s ligament to prevent the bulge caused by a seroma in the undisturbed sac.

Figure 13. Preperitoneal repair: parietalization of spermatic cord

Step 5: parietalization of spermatic cord. The term parietalization of the spermatic cord, [see Video 1] popularized by Stoppa, refers to a thorough dissection of the cord aimed at providing sufficient length to permit lateral movement of the structure [see Figure 13]. In Stoppa’s view, this step is essential, in that it allows a prosthesis to be placed without having to be split laterally to accommodate the cord structures; the keyhole defect created when the prosthesis is split has been linked with recurrences. In Rignault’s opinion, creation of a keyhole defect in the mesh to encircle the spermatic cord is preferable, the rationale being that this gives the prosthesis enough security to allow the surgeon to dispense with fixation sutures or tacks. Minimizing fixation in this area is important because of the numerous anatomic elements in the preperitoneal space that can be inadvertently damaged during suture placement.

Figure 14a. Bilateral GPRVS
Figure 14b. Bilateral GPRVS

Step 6: placement of prosthesis. Dacron mesh, being more pliable than polypropylene, conforms well to the preperitoneal space and is therefore considered particularly suitable for GPRVS. Stoppa’s technique is most often associated with a single large prosthesis for bilateral hernias. The prosthesis is cut in the shape of a chevron [see Figure 14a], and eight clamps are positioned strategically around the prosthesis to facilitate placement into the preperitoneal space [see Figure 14b].

Figure 15. Unilateral GPRVS

Unilateral repairs require a prosthesis that is approximately 15 × 12 cm but is cut so that the bottom edge is wider than the top edge and the lateral side is longer than the medial side. In Wantz’s technique, three absorbable sutures are used to attach the superior border of the prosthesis to the anterior abdominal wall well above the defect [see Figure 15]. The sutures are placed from medial to lateral near the linea alba, the semilunar line, and the anterior superior iliac spine. A Reverdin suture needle facilitates this task. Three long clamps are then placed on each corner and the middle of the prosthesis of the inferior flap. The medial clamp is placed into the space of Retzius and held by an assistant. The middle clamp is positioned so that the mesh covers the pubic ramus, the obturator fossa, and the iliac vessels and is also held by the assistant. The lateral clamp is placed into the iliac fossa to cover the parietalized cord structures and the iliopsoas muscle. Care must be taken to prevent the prosthesis from rolling up as the clamps are removed.

Step 7: closure of wound. The surgical wound is closed in accordance with anatomic guidelines once the surgeon is assured that there has been no displacement or rollup of the prosthesis.


Kugel-Ugahary repair The Kugel and Ugahary repairs were developed to compete with laparoscopic repairs. They require only a small (2 to 3 cm) skin incision placed 2 to 3 cm above the internal ring.37,38In Kugel’s operation, the incision is oriented obliquely, with one third of the incision lateral to a point halfway between the anterior superior iliac spine and the pubic tubercle and the remaining two thirds medial to this point. The incision is deepened through the external oblique fascia, and the internal oblique muscle is bluntly spread apart. The transversalis fascia is opened vertically for a distance of about 3 cm, but the internal ring is not violated. The preperitoneal space is entered and a blunt dissection performed. The inferior epigastric vessels are identified to confirm that the dissection is being done in the correct plane. These vessels should be left adherent to the overlying transversalis fascia and retracted medially and anteriorly. The iliac vessels, Cooper’s ligament, the pubic bone, and the hernia defect are identified by palpation. Most hernial sacs are simply reduced; the exceptions are large indirect sacs, which must sometimes be divided, with the distal sac left in situ and the proximal sac closed. To prevent recurrences, the cord structures are thoroughly parietalized to allow adequate posterior dissection.

The key to Kugel’s procedure is a specially designed 8 × 12 cm prosthesis made of two pieces of polypropylene with a single extruded monofilament fiber located near its edge. The construction of the prosthesis allows it to be deformed so that it can fit through the small incision; once inserted, it springs open to regain its normal shape, providing a wide overlap of the myopectineal orifice. The prosthesis also has a slit on its anterior surface, through which the surgeon places a finger to facilitate positioning.

Ugahary’s operation is similar to Kugel’s, but it does not require a special prosthesis. In what is known as the gridiron technique, the preperitoneal space is prepared through a 3 cm incision, much as in a Kugel repair. The space is held open with a narrow Langenbeck retractor and two ribbon retractors. A 10 × 15 cm piece of polypropylene mesh is rolled onto a long forceps after the edges have been rounded and sutures placed to correspond to various anatomic landmarks. The forceps with the rolled-up mesh on it is introduced into the preperitoneal space, and the mesh is unrolled with the help of clamps and specific movements of the ribbon retractors.

Both operations have been very successful in some hands and have important proponents. However, because they are essentially blind repairs, considerable experience with them is required before the surgeon can be confident in his or her ability to place the patch properly.


Combined Anterior and Posterior (Preperitoneal) Herniorrhaphy

Prosthetic Repair

Bilayer prosthetic repairThe bilayer prosthetic repair involves the use of a dumbbell-shaped prosthesis consisting of two flat pieces of polypropylene mesh connected by a cylinder of the same material. The purpose of this design is to allow the surgeon to take advantage of the presumed benefits of both anterior and posterior approaches by placing prosthetic material in both the preperitoneal space and the extraperitoneal space.

The initial steps are identical to those of a Lichtenstein repair. Once the conventional anterior space has been prepared, the preperitoneal space is entered through the hernia defect. Indirect hernias are reduced, and a gauze sponge is used to develop the preperitoneal space through the internal ring. For direct hernias, the transversalis fascia is opened, and the space between this structure and the peritoneum is developed with a gauze sponge. The deep layer of the prosthesis is deployed in the preperitoneal space, overlapping the direct and indirect spaces and Cooper’s ligament. The superficial layer of the device occupies the conventional anterior space, much as in a Lichtenstein repair. It is slit laterally or centrally to accommodate the cord structures and then affixed to the area of the pubic tubercle, the middle of the inguinal ligament, and the internal oblique muscle with three or four interrupted sutures.


Inguinal Hernia Repair: Complications


An analysis of nearly 18,000 herniorrhaphies in Sweden determined that 15% of these operations were performed to treat recurrent hernias.39 This figure is remarkably consistent with the data from most other large series. A population-based study conducted by the Rand Corporation documented recurrence rates ranging from 10% to 30%, depending on the characteristics of the hernia (e.g., a lower rate after repair of simple small hernias and a higher rate after repair of recurrent hernias).40

Because routine use of prosthetic material in herniorrhaphy is a comparatively recent phenomenon, most of these historical data have to do with sutured repairs. Many surgeons now believe that sutured repairs inevitably result in distortion of the anatomy and in tissue approximation under tension, leading to high recurrence rates. If this belief is correct, then it may be assumed that the overall hernia recurrence rate should decrease dramatically over the next several decades as the percentage of prosthetic herniorrhaphies being performed increases. Two well-controlled, highly funded, randomized trials that examined various aspects of inguinal hernia management have now lent some support to this assumption.8,41 Both trials included the Lichtenstein tension-free repair as the control operation; the recurrence rates at 2 years were 4% and 1% for that operation.

Postherniorrhaphy Pain

It is generally recognized that inguinal herniorrhaphy results in greater morbidity than was previously appreciated. Now that modern hernioplasty techniques have reduced recurrence rates to a minimum, chronic postoperative groin pain syndromes have emerged as the major complication facing inguinal hernia surgeons. In a critical review of inguinal herniorrhaphy studies between 1987 and 2000, the incidence of some degree of long-term groin pain after surgery was as high as 53% at 1 year (range, 0% to 53%).42 This complication is more likely to be observed in younger patients and in patients who report preoperative pain attributable to their hernia. Other risk factors have also been identified [see Table 8]. Chronic postoperative groin pain occurs without regard to the type of repair performed (tissue repair versus tension free; open versus laparoscopic) and does not depend on the method by which the nerves are treated intraoperatively (division versus preservation).43

Treatment is difficult and often fails entirely. The difficulty is compounded when workman’s compensation issues cloud the picture. The first possibility that must be ruled out is a recurrent hernia. As a rule, all three types of pain (somatic, neuropathic, and visceral) are best treated initially with reassurance and conservative treatment (e.g., anti-inflammatory medications and local nerve blocks); frequently, the complaint resolves spontaneously. The only exception to this rule might be the patient who complains of severe pain immediately (i.e., in the recovery room), who might be best treated with immediate reexploration before scar tissue develops. Otherwise, reexploration is scrupulously avoided in the first year after the procedure to allow for the possibility of spontaneous resolution. When groin exploration is required, neurectomy and neuroma excision, adhesiolysis, muscle or tendon repair, and foreign body removal are all possibilities. The results are often less than satisfying.

Ischemic Orchitis and Testicular Atrophy

Orchitis or atrophy may result if the testicular blood supply is compromised during herniorrhaphy. Orchitis is defined as postoperative inflammation of the testicle occurring within the first 2 postoperative days. Patients experience painful enlargement and hardening of the testicle, usually associated with a low-grade fever; the pain is severe and may last several weeks. Ischemic orchitis is most likely attributable to thrombosis of the veins draining the testicle caused by dissection of the spermatic cord. It may progress over a period of months and eventually result in testicular atrophy. This latter development is not inevitable, however. In fact, the occurrence of testicular atrophy is quite unpredictable, in that most patients with this condition have no history of testicular problems associated with the index herniorrhaphy. The vast majority of patients who experience testicular complications go on to recover without atrophy. Bendavid, in a study of the incidence of testicular atrophy at the Shouldice Hospital, found that this complication occurred in only 19 (0.036%) of 52,583 primary inguinal hernia repairs and in only 33 (0.46%) of 7,169 recurrent inguinal hernia repairs.44


Postherniorrhaphy bleeding—usually the result of delayed bleeding from the cremasteric artery, the internal spermatic artery, or branches of the inferior epigastric vessels—can produce a wound or scrotal hematoma. Injuries to the deep circumflex artery, the corona mortis, or the external iliac vessels may result in a large retroperitoneal hematoma.

Osteitis Pubis

Osteitis pubis has diminished in frequency since surgeons began to realize the importance of not placing sutures through the periosteum. In laparoscopic repairs, staples are used to attach the mesh to Cooper’s ligament, which may cause osteitis in some cases.

Prosthesis-Related Complications

The increasingly liberal use of prosthetic material in conventional herniorrhaphy and the routine use of such material in laparoscopic herniorrhaphy make the discussion of complications related directly to foreign material a timely one. Tissue response, which is variable from person to person, can be so intense that the prosthetic material is deformed by contraction. Erosion can result in intestinal obstruction or fistulization, especially if there is physical contact between intestine and prosthesis.45,46 Erosion into the cord structures has also been reported.47

The other controversial issue is the possibility of damage to the spermatic cord caused by the normal fibroplastic response to polypropylene mesh. Such damage may lead to infertility through obstruction of the vas deferens. This was the conclusion in a 2005 paper describing 14 patients attending several specialty infertility clinics.48 Nine of the patients had undergone bilateral tension-free inguinal herniorrhaphies with mesh; the remaining five had undergone unilateral repairs but also had pathologic conditions (e.g., testicular atrophy or torsion) on the opposite side. All patients underwent surgical exploration with intraoperative vasography. The vasogram identified the site of the obstruction in the inguinal region, and the surgical exploration determined that the mesh was the cause of the obstruction. These distressing findings certainly call for continued vigilance. It has been suggested, however, that there may be another explanation for infertility after mesh herniorrhaphy: this complication might be the consequence of a more traditional injury mechanism at the time of surgery, such as ligation, division, or cauterization, followed by scarring to a moist, conveniently adjacent structure (which in this case would be the mesh).49


Prostheses used for inguinal herniorrhaphies, unlike those used for ventral herniorrhaphies, rarely become infected. The reason why the groin is apparently a protected area is unclear. If a prosthesis composed of a mesh material (e.g., polypropylene) becomes infected, it can usually be salvaged with drainage alone. This should be the initial treatment for all infected mesh prostheses, with removal being reserved for refractory cases. If, however, a prosthesis composed of a nonmesh material (e.g., ePTFE) becomes infected, it can never be sterilized and virtually always must be removed. Rejection of the prosthesis because of an allergic response is possible but extremely rare. What patients call rejection in their histories is usually the result of infection.

Incisional Hernia Repair

Incisional hernias occur as a complication of previous surgery. As noted, their incidence depends on how they are defined [see Epidemiology, above]. In the literature, the incidence of incisional hernia ranges from 3% to 12% of all laparotomy incisions,6,50 and it is twice as high if the operation was associated with infection.

The root cause of incisional hernia is undoubtedly multifactorial. In the past, incisional hernias were believed to be mostly iatrogenic, related to surgical technical factors at the index operation (e.g., slippage of knots, breakage of sutures, tearing of fascia by sutures, rough handling of tissues, closure of the abdomen under tension, and poor choice of suture material).51 Today, however, it is clear that noniatrogenic factors [see Table 9] play a much larger role than was previously recognized. Nevertheless, the importance of careful attention to technical detail in the closure of any abdominal incision should not be minimized.

Surgeons’ practices in closing laparotomies tend to be far more dependent on tradition than on high-quality level I scientific evidence.52 There are, however, some general recommendations that can be made on the basis of current data and experience. Most authorities believe that the best way of preventing incisional hernias is to close the incision with a continuous monofilament nonabsorbable suture, with the stitches placed 1 cm from the skin edge and 1 cm apart. To prevent excessive tension, the length of the suture should be four times the length of the wound.52 Monofilament sutures perform better than braided sutures because bacteria tend to form colonies among the braids of multifilament sutures.54–57 Nonabsorbable suture material has the advantage of greater longevity than absorbable suture material, but it is more likely to result in sinus formation and chronic wound pain.7,52 The incidence of wound dehiscence or wound infection is not affected by the suture material or the closure method. Studies of various suture materials have determined that incisional hernia rates are essentially the same with polydioxanone as with polypropylene but may be higher with polyglactin.52

Various patient-related risk factors for incisional hernia have been identified [see Table 10].58,59 Although some controversy remains, the current consensus is that there appears to be an association between these comorbid conditions and the incidence of incisional hernia. The type of wound incurred also plays a role. Incisional herniation is most common after midline laparotomies, especially upper midline incisions, and less common after transverse or oblique incisions.6 An analysis of 11 publications addressing ventral hernia incidence after various types of incisions found the risk to be 10.5% for midline incisions, 7.5% for transverse incisions, and 2.5% for paramedian incisions.7 Over longer periods, the incidence increases, with the majority of incisional hernias developing in the first 4 years after the operation.60 It is anticipated that as the use of minimally invasive surgical techniques increases, the incidence of incisional hernia will drop. Hernias developing within 10 mm and 12 mm port sites are well documented; hernias in 5 mm port incisions are rare. At present, long-term data on the incidence and natural history of port-site hernias are lacking.61

Genetic factors are important as well: familial predisposition to incisional hernia has long been recognized by surgeons caring for patients with this condition. An increased incidence of incisional herniation in patients with certain connective tissue diseases (e.g., osteogenesis imperfecta, Marfan syndrome, and Ehlers-Danlos syndrome) has been documented. Finally, the molecular details of incisional hernia causation are now beginning to be appreciated. Type 1-type 3 collagen imbalance, abnormal matrix metalloproteinase (MMP) expression, and growth factor relations are among the molecular-level processes that are currently under intense scrutiny by the scientific community with regard to the etiology of incisional hernia.

Not every patient who presents to a surgeon with an incisional hernia is necessarily a candidate for surgical repair. There are three indications for operation: (1) a hernia that is symptomatic, causing pain, discomfort, or changes in bowel habits; (2) a hernia resulting in an unsightly bulge that affects the patient’s quality of life; (3) a hernia that poses a significant risk of bowel obstruction (e.g., a large hernia with a narrow neck).

Operative Technique

Nonprosthetic Repairs

Primary suture repair Historically, primary suture repair was the procedure of choice for most incisional hernias; prosthetic material was reserved for particularly difficult cases. In the latter part of the 20th century, large population-based studies changed this way of thinking, revealing that primary suture repair was associated with a much higher recurrence rate then most surgeons would have assumed (25% to 55%).11 Studies comparing primary suture with prosthetic repair showed that the recurrence rate was dramatically lower with the latter.62 In a randomized, controlled study from the Netherlands, even small incisional hernias (< 10 cm2) had a recurrence rate of 67% when primary suture repair was employed.63 Nevertheless, for patients who have no significant comorbid conditions [see Table 10] and who have a solitary defect less than 3 cm in diameter, primary closure with nonabsorbable suture material may be considered. Some surgeons perform a simple edge approximation after flaps have been developed on either side of the defect. Others use a Mayo ‘vest over pants’ repair.

Component separation repair This operation was initially described by Ramirez in 1990 and has become increasingly popular since then.64 Although first envisioned for treatment of hernias no larger than 10 cm, it is now being used to repair more substantial defects. Perhaps its most common application is in contaminated wounds, where a conventional prosthetic repair would be contraindicated.65–67

Figure 16. Component separation repair

A long midline incision is made through the scar to expose the hernia. The hernial sac is dissected up to its neck, deep to the fascial edge. The skin and the subcutaneous fat are dissected away from the anterior sheath of the rectus abdominis and the aponeurosis of the external oblique muscle. The aponeurosis of the external oblique muscle is transected longitudinally just lateral to the lateral side of the rectus sheath [see Figure 16]. It is important to extend the incision onto the chest wall at least 5 to 7 cm cranial to the costal margin. The external oblique muscle is separated from the internal oblique muscle as far laterally as possible. This step is safe because the neurovascular bundle (comprising the intercostal nerves and vessels) lies deep to the internal oblique muscle. The result is that the internal oblique muscle and the rectus abdominis slide medially, so that the edges of the hernial defect can be brought together without tension and sutured primarily.68 If primary closure still is not possible without undue tension, 2 to 4 cm of additional length can be gained by separating the posterior rectus sheath from the rectus abdominis. Care must be taken not to damage the neurovascular bundle that runs between the internal oblique muscle and the transversus abdominis to enter the rectus sheath posterolaterally.


Prosthetic Repairs

The use of prosthetic material to reduce tension has unquestionably reduced the recurrence rate after incisional hernia repair, especially in single-center series.69 For example, in a Finnish study of 84 consecutive patients treated with a retromuscular polypropylene mesh repair and followed for 3 years, the recurrence rate was 5%. In a separate U.S. study, no recurrences were reported in 102 patients after 28 months of follow-up.70

Figure 17. Incisional hernia repair: three positions for prosthesis

Prosthetic material may be positioned in three different ways for an incisional herniorrhaphy—namely, as an overlay (onlay), an inlay, or an underlay (sublay) [see Figure 17]. A mesh overlay may be placed on top of any of a variety of simple repairs. Although some series have reported that this approach yields acceptable results in selected patients, most surgeons feel that it offers little advantage over the simple repair that the prosthesis overlies and that it typically is associated with a similarly disappointing recurrence rate.11

Prosthetic inlay (bridging) repair became popular in the 1990s, in keeping with the tension-free ideal for inguinal herniorrhaphy. The principle underlying this technique is that for a prosthetic repair to be truly tension free, the defect should be bridged. Although this repair is theoretically attractive, it has not been nearly as successful for incisional hernias as for inguinal hernias. The recurrence rate is especially high in obese patients. Recurrences invariably develop at the mesh-native tissue interface. In the previously cited study from the Netherlands,63 the recurrence rate even with mesh repairs (mostly onlay and inlay) was 32% for large defects and 17% for small (< 10 cm2) defects.

When a hernia defect is bridged with a mesh prosthesis, every attempt should be made to isolate the material from the intra-abdominal viscera so as to prevent erosion and subsequent fistula formation or adhesive bowel obstruction. Such isolation may be accomplished with a peritoneal flap constructed from the peritoneal sac or with omentum. When contact with intra-abdominal organs cannot be avoided, ePTFE should be strongly considered for the prosthesis. Alternatively, one of the dual-layer prostheses that have mesh on one side and some type of adhesion barrier to protect the viscera on the other may be considered [see Table 4]. As yet, however, none of the dual-layer prostheses have a long enough track record to ensure that they will be safer than polypropylene alone would be.

Figure 18. Incisional hernia repair: modified onlay technique

The issue of contact between the intra-abdominal viscera and the prosthesis has been further addressed by techniques that combine some features of component separation repair with the tension-free concept. An example is the so-called keel operation of Maingot, which was popular in the middle of the 20th century. The anterior rectus sheath is incised longitudinally, and the medial edge is allowed to rotate behind the rectus abdominis. This, in effect, lengthens the posterior rectus sheath, allowing it to be closed primarily, which isolates the intra-abdominal viscera. If possible, the lateral edges of the incised rectus sheath on each side are approximated to each other. Otherwise, an inlay prosthesis may be used in the hope that the closure of the posterior sheath will render failure less likely than it would be with a simple bridging onlay repair [see Figure 18].

Sublay prosthetic repair

Figure 19. Incisional hernia repair: retromuscular approach

Sometimes referred to as the retromuscular approach, a sublay prosthetic repair is characterized by the placement of a large prosthesis in the space between the abdominal muscles and the posterior fascia—or the transversalis fascia or the peritoneum, [see Video 2] depending on what part of the abdomen is being repaired (there is no posterior fascia inferior to the arcuate line).27 In this chapter, we illustrate the version of the operation described by Flament [see Figure 19], but very similar operations have been attributed to Velamenta, Stoppa, and Wantz.71–73The operation was originally envisioned for treatment of large and multiply recurrent hernias in cases where most of the abdominal wall had to be reconstructed. Because it has proved so successful, it is now being increasingly used to repair ever smaller defects. Sublay prosthetic repair is currently considered the most effective conventional incisional hernia repair and is therefore the one against which all other procedures must be measured.

Extensive flaps are created by dissecting the skin and subcutaneous tissue off the external fascia well lateral to the hernia defect on either side. This step often allows the musculoaponeurotic components of the abdominal wall to be advanced to the point where the posterior and anterior fascial layers can be closed primarily. Once the flaps have been created, the fascia is opened at the edges of the defect, thereby affording entry into the plane between the posterior surface of the deepest muscle and the underlying peritoneum and posterior fascia. A combination of blunt and electrocautery dissection works best for creating this large space, which will eventually accommodate a sizable prosthesis (at least 5 cm long and wide).74 The posterior rectus sheaths are approximated to each other primarily if possible. If the posterior sheath cannot be approximated because of tension, then the use of ePTFE or a dual-layer prosthesis should be considered instead of the standard mesh. The prosthesis is then placed in the space beneath the muscle and secured in this position with sutures that are placed with a suture passer through small stab incisions at the periphery of the retromuscular pocket. The sutures pull the prosthesis well lateral and firmly affix it to the abdominal wall; they are then tied in the subcutaneous tissue above the fascia. To prevent excessive skin flap dissection, it is usually best to bring the two tails through the full thickness of the abdominal wall, including the skin. The tails must exit through two separate fascia sites but through the same small skin incision, so that when the knot is tied, it resides in the subcutaneous tissue. Some surgeons prefer to avoid using full-thickness sutures because of concern over the possibility of wound pain resulting from neuromuscular entrapment. The prosthesis is therefore either sutured or stapled to the posterior fascia as far laterally as possible. An increasingly popular choice for the most complicated hernias is to incorporate component separation into the procedure.

A 2006 study from Sweden confirmed the superiority of this operation, reporting a 7.3% recurrence rate for sublay prosthetic repair, compared with 19.3% for onlay mesh repair and 29.1% for suture repair.75 Sublay prosthetic repair has been successfully employed to treat massive hernias with substantial loss of domain.76 Laparoscopic incisional hernia repair was designed with the principles of this operation in mind. Indeed, current experience with laparoscopic methods has encouraged surgeons to place prosthetic material intraperitoneally even when performing conventional open ventral herniorrhaphies, so as to minimize the need for extensive abdominal wall dissection.77,78



There is overwhelming proof that tension-free prosthetic repairs yield lower recurrence rates than direct suture repairs do. In a Medline search for complications of incisional hernia repair, recurrence rates ranged from 31% to 63% for direct suture repairs and from 0% to 32% (mostly less than 10%) for prosthetic repairs.63,79 Although the primary adverse outcome of hernia repair is recurrence, the short-term morbidity of open hernia repair must also be assessed. In one meta-analysis, the overall complication rate after open repair was 27%.80 Ileus, postoperative pain, sepsis, fistulization, and necrotizing fasciitis have all been documented. A 3.5% rate of enterocutaneous fistula formation within a 3-year follow-up period has also been reported.81 Prothesis-related infection, though rare with inguinal hernia repairs, remains a major problem with incisional hernia repairs. It occurs in as many as 25% of repairs in some series, delays healing for prolonged periods, and is one of the most important risk factors for re-recurrence. Higher rates of prosthesis infection are associated with preexisting infection, ulceration of the skin overlying the hernia, obesity, incarcerated or obstructed bowel within the hernia, and perforation of the bowel during hernia repair. Seromas are common, especially when a large prosthesis is required or there has been extensive flap dissection of the subcutaneous layer from the fascia. Untreated seromas commonly become infected secondarily. Suction drains can be useful but are likely to result in prosthesis infection if left in place too long. Strategies for preventing and managing seromas are largely based on empiricism and personal opinion; objective data are virtually nonexistent. It is not always necessary to remove the mesh if infection develops. A trial of local wound care after opening the incision and debriding the infected area is warranted. Some authorities believe that ePTFE prostheses are less prone to infection; however, once infection is established, ePTFE prostheses, unlike mesh prostheses, are almost never salvageable.

A dilemma arises when a patient has a large incisional hernia and the wound is contaminated either by skin infection or by injury to the bowel during the repair at the time of adhesiolysis. In this situation, a nonabsorbable mesh would have a significant chance of becoming infected, and an enterocutaneous fistula could complicate matters further. In the past, the use of absorbable mesh made of polyglycolic acid was recommended to prevent evisceration. Granulation tissue forms over the mesh, making skin grafting possible. The mesh itself is absorbed in about 3 weeks, leaving no permanent foreign body to serve as a persistent focus of infection. Unfortunately, however, recurrence of the incisional hernia is inevitable. Currently, many surgeons prefer to use one of the newer biologic prostheses in this setting. This has now become the best indication for these very expensive materials. Long-term data are not yet available.

Periumbilical Hernia Repair


This life-threatening condition is seen in the newborn. It may also be found in the fetus during ultrasound examination of a pregnant patient. There is an all-layer deficiency of the abdominal wall, to the right of a normal umbilicus, through which the bowels protrude. There is no hernial sac. Gastroschisis is discussed in more detail elsewhere [see 9:2 The Pediatric Surgical Patient].

Omphalocele (Exomphalos)

Like gastroschisis, this condition is seen in the fetus in utero and in the newborn. It is a hernia into the umbilical cord; the hernial contents are therefore covered by Wharton’s jelly and amnion. Omphalocele is also discussed in more detail elsewhere [see 9:2 The Pediatric Surgical Patient].

Umbilical and Paraumbilical Hernia

Unlike an omphalocele, an umbilical hernia is covered by skin. If the defect is located to one side of the umbilicus, it is called a paraumbilical hernia (this variant is more common in adults). Umbilical hernias developing during childhood are congenital, whereas those developing during adult life are acquired. Accordingly, in adult patients, it is important to look for an underlying cause of increased intra-abdominal pressure (e.g., ascites or an intra-abdominal tumor). The differential diagnosis of an umbilical hernia includes a caput medusae of varices at the umbilicus from portal hypertension, a metastatic tumor deposit (so-called Sister Mary Joseph node), a granuloma, an omphalomesenteric duct cyst, and a urachal cyst.

Management of umbilical hernia is determined by the age of the patient. The majority of hernias occurring in children younger than 2 years will heal spontaneously; therefore, watchful waiting is the rule, and only symptomatic hernias are operated on. In children older than 2 years and in adults, surgical correction is required, with the type of repair employed depending on the size of the hernia. If the defect is small (< 3 cm), a direct suture repair may be performed. Alternatively, the Mayo repair may be used. A subumbilical semilunar incision is made, the hernial sac is opened, the contents of the sac are reduced into the abdomen, and the sac is excised. An overlapping or waistcoating technique is employed, in which the upper edge of the linea alba is placed so as to overlap the lower and then is fixed in this position with nonabsorbable mattress sutures. Because the Mayo repair results in increased tension, there has been controversy about its efficacy in adults. In various series, recurrence rates ranging from 1% to 40% have been reported in adult patients.82

For larger umbilical and paraumbilical hernias, particularly those in adults, a mesh repair is preferred. The sac is dissected away from the undersurface of the rectus and the linea alba circumferentially, then reduced into the abdomen. If the peritoneum remains intact, a mesh prosthesis may be placed in a subfascial position and secured with sutures. Alternatively, a mesh plug may be inserted and secured to the edges of the defect with a series of sutures.83 If the abdomen is entered, a dual-layer prosthesis with an adhesion barrier on the visceral side is recommended. In a series of 100 adult patients with a median follow-up period of 4.5 years, the recurrence rate was 11.5% for suture repairs and 0% for mesh repairs.84

Miscellaneous Ventral Abdominal Wall Hernia Repairs

Epigastric Hernia

Epigastric hernias occur through a single defect or multiple defects in the linea alba. In most patients with these hernias, only a single decussation of the fibers of the linea alba is present, as opposed to the triple decussation seen in most persons.85 The reported incidence of epigastric hernia ranges from less than 1% to as high as 5%. They are two to three times more common in men than in women, and 20% are multiple. Most defects are smaller than 1 cm and contain only incarcerated preperitoneal fat, with no peritoneal sac. For this reason, they generally cannot be visualized laparoscopically.

The usual complaint is a painful nodule in the upper midline. As a rule, reduction of the preperitoneal fat followed by simple closure of the defect resolves the complaint. Given the relatively high recurrence rate (up to 10%), some surgeons prefer to place a postage stamp-sized piece of prosthetic material in the preperitoneal space to reinforce the repair. Others bridge the defect by suturing the prosthesis circumferentially. Some authorities recommend exposure of the entire linea alba because of the incidence of multicentricity. We believe that this practice leads to unnecessary morbidity. Instead, we make a small incision with the patient under local anesthesia and explain to him or her that additional repairs may be required later.

Left untreated, an epigastric hernia can become large enough to develop a peritoneal sac into which intra-abdominal contents can protrude. Usually, however, the sac is wide, and serious complications are infrequent.

Diastasis Recti

In diastasis recti, the two rectus abdominis muscles are separated quite widely, and the linea alba is stretched and protrudes like a fin. Although the protrusion is easily reducible and almost never produces complications, many patients find it unsightly and request treatment. The usual therapy involves removing a strip of the weakened linea alba and reapproximating it; however, this could result in tension, which in turn might lead to recurrence. The alternative would be a mesh repair.

Parastomal Hernia

Parastomal hernia is one of the most common complications of stoma formation. Its incidence is much higher than is generally appreciated. There is good evidence to suggest that more than 50% of patients will eventually be found to have a paracolostomy hernia if followed for longer than 5 years.86 The rate of herniation with small bowel stomas is also discouraging, though less so than that with colostomies. The results of parastomal hernia repair are particularly dismal, with recurrence being the rule rather than the exception.

Some parastomal hernias can be accounted for by poor site selection or technical errors (e.g., making the fascial opening too large or placing a stoma in an incision), but the overall incidence is too high to be explained by these causes alone. Placement of the stoma lateral to the rectus sheath is widely touted as a cause of parastomal hernia, but high-quality scientific evidence to support this claim is not available. Obesity, malnutrition, advanced age, collagen abnormalities, postoperative sepsis, abdominal distention, constipation, obstructive uropathy, steroid use, and chronic lung disease are also contributing factors.87,88

Newer techniques for stomal construction, such as extraperitoneal tunneling, have had little impact on the incidence of parastomal hernia. Fortunately, patients tolerate these hernias well, and life-threatening complications, such as bowel obstruction and strangulation, are rare. Most are asymptomatic. Routine repair, therefore, is not recommended; repair is appropriate only when there is an absolute or relative indication [see Table 11]. If repair is considered, patients must be informed that there is a significant chance that the hernia will recur.

Three general types of parastomal hernia repairs are currently performed: (1) fascial repair, (2) stomal relocation, and (3) prosthetic repair. Fascial repair involves local exploration around the stoma site, with primary closure of the defect. This approach should be considered of historical interest only because the results are so poor. Stomal relocation yields much better results and is considered the procedure of choice by many surgeons. This approach is especially appropriate for patients who have other stomal problems, such as skin excoriation or suboptimal stomal construction. The use of a prosthesis with a stomal relocation is not generally recommended, because of the inherent danger of contamination. In the past few years, the popularity of stomal relocation has waned because of the realization that patients who undergo this procedure are subjected to a triple threat of hernia recurrence: (1) at the old stoma site, (2) at the new stomal site, (3) in the laparotomy incision used to move the stoma.

Prosthetic repair appears to be the most promising approach, but it is necessary to accept the complications inherent in the placement of a foreign body. The stomal exit site must be isolated from the surgical field to lower the risk of prosthesis infection. The prosthesis can be placed extraperitoneally by making a hockey-stick incision around the stoma, with care taken to ensure that the incision is outside the periphery of the stomal appliance. Once the subcutaneous tissue is divided, dissection proceeds along the fascia until the sac is identified and removed. The defect is then closed with an overlying prosthesis buttress sutured in place. Alternatively, the fascial defect is bridged with the prosthesis for a tension-free repair.

The extraperitoneal approach seems logical but can be technically demanding, in that it is sometimes difficult to define the entire extent of the hernia defect. Moreover, the considerable undermining involved can lead to seroma formation and eventual infection. As an alternative, an intra-abdominal prosthetic approach has been described that is theoretically attractive because it avoids the local complications of the extraperitoneal operation and incorporates the mechanical advantage gained by placing the prosthesis on the peritoneal side of the abdominal wall.89,90 Intra-abdominal pressure then serves to fuse the prosthetic material to the abdominal wall rather than being a factor in recurrence. Either ePTFE or polypropylene mesh with an adhesion barrier can be used for the prosthesis. One technique is to slit the prosthesis and create a keyhole in its center, then suture this directly around the peritoneal side of the stoma so that it widely overlaps the hernia defect. Sugarbaker’s practice is to mobilize the bowel thoroughly, then lateralize it with the prosthesis—in effect, creating a long tunnel in addition to covering the hernia defect.90 The detractors of the intra-abdominal approach argue that the risk of complications (e.g., adhesive bowel obstruction and fistula formation) outweighs the advantages. The intra-abdominal approach is particularly well suited for adaptation to laparoscopic methods.91–93

Spigelian Hernia

Figure 20. Spigelian hernia belt

A spigelian hernia, first described 400 years ago by the Flemish anatomist Adriaan van den Spiegel,94 is a hernia through a defect in the spigelian fascia. The spigelian fascia is the area between the semilunar line and the lateral border of the rectus abdominis. The majority of spigelian hernias occur just below the arcuate line, where the posterior rectus sheath becomes deficient. This region, known as the spigelian belt, is a band between the iliac crest and a line drawn 6 cm above [see Figure 20].95 These rare hernias are being reported with increasing frequency: there are more than 100 cases in the surgical literature.

A spigelian hernia may present as a bulge lateral to the rectus. However, because many of these hernias are interparietal, they may not be clinically apparent; often, they are picked up incidentally during laparoscopy. A significant percentage of patients present with an incarcerated or even strangulated hernia. If such a hernia is interparietal, the diagnosis frequently is not made until a laparotomy is done for treatment of the acute process.

The standard treatment is operative repair.96 A transverse incision is made over the bulge. The anterior rectus sheath is incised transversely, and the sac is dissected as far as its neck and either excised or inverted. The defect is then repaired with a continuous suture of nonabsorbable material. Alternatively, a mesh plug may be placed in the defect and sutured to the edges of the defect. Laparoscopic methods are increasingly being employed to repair spigelian hernias.97

Richter’s Hernia

In a Richter’s hernia, part of the bowel wall herniates through the defect. The herniated bowel wall may become ischemic and gangrenous, but intestinal obstruction does not occur. The overlying skin may be discolored. The herniated bowel wall is exposed by opening the sac, and the neck of the sac is enlarged to allow delivery of the bowel into the wound. The gangrenous patch is excised and the bowel wall reconstituted. The hernia is then repaired.

Supravesical Hernia

Supravesical hernias develop anterior to the urinary bladder as a consequence of failure of the integrity of the transversus abdominis and the transversalis fascia, both of which insert into Cooper’s ligament. The preperitoneal space is continuous with the retropubic space of Retzius, and the hernial sac protrudes into this area. The sac is directed laterally and emerges at the lateral border of the rectus abdominis in the inguinal region, the femoral region, or the obturator region. It may therefore mimic a hernia from any of these areas, and it sometimes is associated with a hernia from one of these regions. It is important to recognize this hernia during groin exploration for a suspected groin hernia and then to repair the defect appropriately.

A variant of this hernia, known as an internal supravesical hernia, may also arise. These hernias are classified according to whether they cross in front of, extend beside, or pass behind the bladder. Bowel symptoms predominate in patients with these defects, and urinary tract symptoms may develop in as many as 30%. Treatment is surgical and is accomplished transperitoneally via a low midline incision. The sac can usually be reduced without difficulty, and the neck of the sac should be divided and closed.

Lumbar Hernia

The lumbar region is the area bounded inferiorly by the iliac crest, superiorly by the 12th rib, posteriorly by the erector spinae group of muscles, and anteriorly by the posterior border of the external oblique muscle as it extends from the 12th rib to the iliac crest. There are three varieties of lumbar hernia.

  1. The superior lumbar hernia of Grynfelt. In this variety, the defect is in a space between the latissimus dorsi, the serratus posterior inferior, and the posterior border of the internal oblique muscle.
  2. The inferior lumbar hernia of Petit. Here the defect is in the space bounded by the latissimus dorsi posteriorly, the iliac crest inferiorly, and the posterior border of the external oblique muscle anteriorly.
  3. Secondary lumbar hernia that develops as a result of trauma—mostly surgical (e.g. renal surgery)—or infection.98 In the past, it was encountered relatively frequently as a consequence of spinal tuberculosis with paraspinal abscesses, but it is less common today. Surgical repair is discouraged because the natural history is more consistent with that of diastasis recti than that of a true hernia. Denervation appears to play a significant role in the pathogenesis. In other words, this ‘hernia’ really reflects a weakness in the abdominal wall more than it does a dangerous hernia defect. Therefore, appropriate repair is commonly followed by gradual eventration, which is perceived by the patient as a recurrence.

Lumbar hernias should be repaired if they are large or symptomatic. A prosthesis or a tissue flap of some kind is usually required for a successful repair. A rotation flap of fascia lata can be used for inferior lumbar hernias. Laparoscopic repair of lumbar hernias is now being performed with increasing frequency and is proving successful.99


Abdelrahman A. Nimeri, MD

Assistant Clinical Professor, Department of Surgery
University of California, San Francisco, School of Medicine

L. Michael Brunt, MD, FACS

Associate Professor, Department of Surgery
Washington University School of Medicine

The surgical approach to the adrenals has evolved substantially over the past decade with the development and refinement of techniques for performing laparoscopic adrenalectomy. At present, the majority of adrenal tumors are removed laparoscopically because minimally invasive approaches result in reduced pain, faster recovery, and fewer complications and because the rate of adrenal malignancy is low. Nevertheless, open adrenalectomy still has a role in the management of selected patients with large or malignant tumors. With both open adrenalectomy and laparoscopic adrenalectomy, several different surgical approaches to the adrenals are possible [see Operative Planning, Choice of Procedure, below]. Regardless of the particular approach followed, the keys to successful adrenalectomy are the same: proper patient selection for operation, a solid understanding of adrenal pathophysiology, and a thorough knowledge of adrenal anatomy.

Anatomic Considerations

Figure 1. Relation of adrenal glands to adjacent structures

The adrenal glands are retroperitoneal organs that lie along the superomedial aspects of the two kidneys [see Figure 1]. Each gland comprises two discrete anatomic and functional units: the adrenal cortex, which is the site for synthesis and secretion of cortisol, aldosterone, and adrenal androgens; and the medulla, which is derived from the neural crest and is the site for synthesis of the catecholamines epinephrine and norepinephrine. A normal adrenal gland typically weighs between 4 and 6 g and measures approximately 4 to 5 cm by 2 to 3 cm by 0.5 to 1 cm. The right adrenal is relatively pyramidal in shape, whereas the left is somewhat flattened and is more closely applied to the kidney. Grossly, the adrenals may be distinguished from the surrounding retroperitoneal fat by their golden-orange color, which is a result of the high intracellular lipid content. The glands have a fibrous capsule but are relatively fragile and can be easily cracked or fragmented with surgical manipulation.

Right Adrenal

Anteriorly, the right adrenal is partially covered by the liver and the right triangular ligament. The gland abuts the inferior vena cava (IVC) medially and may, in part, lie posterior to the lateral aspect of the vena cava. Inferiorly, the adrenal sits just above the upper pole of the kidney. The diaphragm forms the posterior and lateral boundaries of the gland.

Figure 2. Adrenal blood supply

The blood supply of the right adrenal is derived from branches of the inferior phrenic artery, the right renal artery, and the aorta [see Figure 2]. Typically, multiple small branches enter the gland along its superior, medial, and inferior aspects. Arterial branches from the aorta generally course posterior to the vena cava before entering the adrenal. Each adrenal is drained by a single central vein. On the right, this vein is short (1 to 1.5 cm long), runs transversely, and joins the lateral aspect of the inferior vena cava. In some cases, a more superiorly located accessory adrenal vein may enter either the IVC or one of the hepatic veins. Control of the adrenal vein is the most critical aspect of right adrenalectomy, in that the short course of this vessel makes it susceptible to tearing or avulsion from the IVC.

Left Adrenal

The spleen and tail of the pancreas overlie the anterior and medial borders of the left adrenal. The inferolateral aspect of the gland lies over the superomedial aspect of the left kidney, to which it is more closely applied than the right adrenal is to the right kidney. The inferior aspect of the adrenal is in close proximity to the renal vessels, especially the renal vein. As on the right side, the posterior aspect of the adrenal rests on the diaphragm.

The arterial blood supply of the left adrenal is similar to that of the right adrenal [see Figure 2]. The left adrenal vein is longer than the right adrenal vein and runs somewhat obliquely from the inferomedial aspect of the gland to enter the left renal vein. The inferior phrenic vein courses in a superior-to-inferior direction just medial to the adrenal and usually joins the left adrenal vein cephalad to its junction with the renal vein.

Preoperative Evaluation

Indications for Operation

The main indications for adrenalectomy are well established [see Table 1]. Any adrenal lesion that either is hypersecretory for one of the adrenal hormones or appears to be malignant or possibly malignant should be removed. In selected cases, it may be appropriate to remove adrenal metastases if they are solitary and if there is no evidence of extra-adrenal metastatic disease. Nonfunctioning adrenal lesions that appear to be benign on the basis of their size (< 4 cm) and their appearance on computed tomography or magnetic resonance imaging need not be removed unless they enlarge during follow-up. Adrenal myelolipomas and cysts usually can be diagnosed radio graphically and should not be removed unless they cause symptoms.

Most of the conditions for which adrenalectomy is indicated are amenable to a laparoscopic approach. However, the role of laparoscopy in patients with large adrenal tumors (> 6 to 8 cm) or potentially malignant primary adrenal lesions remains controversial. In the presence of a locally invasive tumor, a laparoscopic approach is contraindicated because of the need to perform en bloc resection of the tumor and any adjacent involved structures.

Common Adrenal Tumors

A brief review of the pertinent clinical and biochemical features of the various hypersecretory adrenal tumors [see Table 2] will facilitate evaluation of adrenal lesions (including adrenal incidentalomas) and planning for adrenal surgery.


Primary hyperaldosteronism is the most common form of secondary hypertension, and aldosterone-producing adenoma is the most common hypersecretory adrenal tumor. The prevalence of this diagnosis is much higher than was previously thought,1,2 reaching levels as high as 12% of hypertensive individuals in some series.1 The classic finding in primary hyperaldosteronism is hypertension in conjunction with hypokalemia, but many patients have a normal or low-normal serum potassium level. Therefore, any patient who becomes hypertensive at an early age or who has malignant or difficult-to-control hypertension should be screened for this diagnosis. Screening consists of measuring plasma aldosterone concentration (PAC) and plasma renin activity (PRA). A PAC-to-PRA ratio higher than 20 to 30, in conjunction with a plasma aldosterone concentration higher than 15 ng/dl, is suggestive of the diagnosis and should be confirmed by measuring 24-hour urine aldosterone levels while the patient is on a high-sodium diet.2 A 24-hour urine aldosterone level higher than 12 µg/24 hr in this setting is confirmatory.

Because 25% or more of cases of primary hyperaldosteronism may be idiopathic as a result of bilateral adrenal hyperplasia and should therefore be managed medically and not surgically, the next step should be imaging with thin-section (3 mm cuts) CT or MRI. The finding of a discrete unilateral adenoma larger than 1 cm on CT in conjunction with a normal contralateral adrenal is sufficient localization to allow the surgeon to proceed with adrenalectomy. If CT shows bilateral nodules, bilateral normal adrenals, or a unilateral nodule smaller than 1 cm, then adrenal vein sampling for aldosterone and cortisol should be done to determine whether a unilateral gradient of increased aldosterone production exists.3

Cortisol-Producing Adenoma

Approximately 20% of cases of Cushing syndrome are related to increased production of cortisol by an adrenal cortical tumor. Adrenal Cushing syndrome is most commonly attributable to adenoma but may also result from adrenal cortical carcinoma or primary adrenal hyperplasia. The classic features of full-blown Cushing syndrome are usually obvious and include centripetal obesity, moon facies, hypertension, purple skin striae, proximal muscle weakness, osteopenia, and amenorrhea. Not all patients present with advanced clinical signs, however, and the high prevalence of hypertension and obesity in the general population necessitates liberal use of diagnostic testing.

Suspected Cushing syndrome should be evaluated initially by measuring 24-hour urinary free cortisol levels or by administering a single low-dose dexamethasone test. If plasma cortisol does not fall to a level below 3 to 5 µg/dl the morning after administration of 1 mg of dexamethasone at 11 P.M., Cushing syndrome is a strong possibility and further testing is required. Once the diagnosis of hypercortisolism is established, plasma levels of adrenocorticotropic hormone (ACTH) should be measured to differentiate ACTH-dependent (resulting from increased ACTH production by a pituitary tumor or an ectopic source) from ACTH-independent (primary adrenal) causative conditions. The plasma ACTH level should be in the low-normal or suppressed range in patients with primary adrenal tumors, whereas it is normal or elevated in patients with ACTH-dependent Cushing syndrome. Imaging should then be carried out with CT (or MRI) to localize the adrenal tumor.


Pheochromocytoma should be suspected in any patient who experiences either severe episodic hypertension or hypertension that is associated with spells of tachycardia, headache, anxiety, and diaphoresis. Biochemical evaluation consists of measuring urinary concentrations of catecholamines and metabolites (e.g., metanephrine and normetanephrine), plasma concentrations of fractionated metanephrines, or both. MRI is our preferred imaging modality for suspected pheochromocytomas because of the typical bright appearance these tumors exhibit on T2-weighted imaging sequences. Occasionally, radionuclide imaging with iodine-123-metaiodylbenzylguanidine (123I-MIBG) or octreotide scintigraphy is necessary to localize an extra-adrenal pheochromocytoma.

Adrenocortical Carcinoma

Adrenocortical carcinomas are rare, with an incidence of approximately one per 1.5 to 2.0 million in the general population. These tumors are typically large (> 6 to 8 cm) at diagnosis, and most patients have advanced (stage III or IV) disease at presentation. Consequently, patients often have a mass or complain of abdominal or back pain. A significant percentage of patients with adrenocortical carcinoma present with evidence of hormone over-production in the form of Cushing syndrome or virilizing features. Complete surgical resection offers the only chance for cure; thus, the role of laparoscopic adrenalectomy in the treatment of adrenocortical carcinoma remains controversial [see Troubleshooting, Large Tumors, below].

Adrenal Incidentaloma

Adrenal incidentalomas are the adrenal lesions most frequently referred to surgeons and are seen on 1% to 5% of all abdominal CT scans.4 Practical, current recommendations for evaluation and management of patients with incidentally discovered adrenal lesions are available.5 The most common adrenal incidentaloma is a nonfunctioning cortical adenoma for which adrenalectomy is not usually required. All patients with adrenal incidentalomas should be screened for hypercortisolism by administering an overnight low-dose dexamethasone test and for pheochromocytoma by measuring plasma concentrations of fractionated metanephrines or urine levels of catecholamines and metanephrines. Patients who are hypertensive or hypokalemic should also undergo testing for hyperaldosteronism with measurement of plasma aldosterone and renin levels. Some patients with adrenal incidentalomas are found to have subclinical Cushing syndrome with evidence of autonomous corticoid steroid production, as demonstrated by lack of suppressibility with a dexamethasone test and by low plasma ACTH levels.6 These patients do not exhibit the classic features of Cushing syndrome but do have a high incidence of hypertension, diabetes, and osteoporosis. Adrenalectomy is generally indicated if the operative risk is suitably low. Supplemental corticosteroids should be given before, during, and after operation because contralateral adrenal function is often suppressed and adrenal insufficiency may ensue.

The nonfunctioning adrenal lesion should also be assessed for malignant potential on the basis of its size and appearance on diagnostic imaging. Cortical adenomas typically have low attenuation values (< 10 Hounsfield units) on unenhanced CT imaging and show loss of signal intensity on MRI chemical-shift imaging sequences. Needle biopsy is not useful in differentiating benign from malignant primary adrenal lesions and is rarely indicated. Adrenal biopsy should never be done unless a pheochromocytoma has first been excluded biochemically. Most experts recommend removing any nonfunctioning adrenal lesion larger than 4 to 5 cm unless the radiographic appearance of the lesion is diagnostic of a cyst or myelolipoma. Smaller tumors should be followed with imaging at 4 and 12 months after the initial presentation.

Operative Planning

Preparation for Operation

Preoperative preparation of the patient for adrenalectomy entails control of hypertension and correction of any electrolyte imbalances. Patients with a pheochromocytoma should receive 7 to 10 days of alpha-adrenergic blockade with phenoxybenzamine to minimize any exacerbation of hypertension during the operation. The usual starting dosage is 10 mg twice daily, which is increased by 10 to 20 mg/day until the hypertension and tachycardia are controlled and the patient is mildly orthostatic. Patients with Cushing syndrome or subclinical Cushing syndrome should receive perioperative dosages of stress steroids. Mechanical bowel preparation is not routinely employed.

Choice of Procedure

The retroperitoneal location of the adrenals renders them accessible via either transabdominal or retroperitoneal approaches. The choice of surgical approach in any given patient depends on a number of factors, including the nature of the underlying adrenal pathology, the size of the tumor, the patient’s body habitus, and the experience of the operating surgeon. For the vast majority of adrenal lesions, laparoscopic adrenalectomy is preferred. Most centers favor the transabdominal lateral approach to laparoscopic adrenalectomy,7 which has the advantages of a large working space, familiar anatomic landmarks, and widespread success. Some centers, however, prefer a retroperitoneal endoscopic approach.8,9 The advantages of this technique are that the peritoneal cavity is not entered, there is no need to retract overlying organs, and the incidence of postoperative ileus may be lower. The disadvantages are that the retroperitoneal approach employs a smaller working space, is more difficult to learn with fewer anatomic landmarks for orientation, and is usually restricted to tumors smaller than 5 cm.

The only absolute contraindications to laparoscopic adrenalectomy are local tumor invasion and the presence of regional lymphadenopathy. A large tumor (> 8 to 10 cm), a suspected primary adrenal malignancy, and a history of previous nephrectomy, splenectomy, or liver resection on the side of the lesion to be removed are all indicators that a case is likely to be more difficult and should be considered relative contraindications to a laparoscopic approach in all but the most experienced hands. Portal hypertension is also a contraindication to a laparoscopic approach because of the dilated collateral vessels in the retroperitoneum.

Options for open adrenalectomy include transabdominal, flank, posterior retroperitoneal, and thoracoabdominal approaches. The lateral flank approach and the posterior retroperitoneal approach have been replaced by laparoscopic approaches and are now rarely used. A posterior retroperitoneal adrenalectomy is done through a hockey-stick incision in the back, with subperiosteal resection of the 12th rib. This approach has low morbidity and yields adequate exposure of the adrenal gland, but the visual field is often limited, and there is a high incidence of residual incisional complaints. The current consensus is that open posterior retroperitoneal adrenalectomy is indicated only in patients who require bilateral adrenalectomy but are not candidates for laparoscopic adrenalectomy. Most large or malignant adrenal tumors that necessitate an open approach can be removed via an anterior abdominal incision, usually a unilateral or bilateral subcostal incision (with subxiphoid extension if necessary); a thoracoabdominal incision is rarely needed.

Operative Technique

Laparoscopic Adrenalectomy

Transabdominal Approach

Patient positioningA gel-padded bean-bag mattress is placed on the operating table before the patient enters the room. The patient is placed in the supine position, general anesthesia is induced, and sequential compression stockings are placed. A urinary catheter is inserted for monitoring of urine output, and the stomach is decompressed with an orogastric tube. Invasive monitoring is not usually necessary unless the patient has a vasoactive pheochromocytoma, in which case an arterial line is routinely placed.

Figure 3. Laparoscopic adrenalectomy: transabdominal approach

Next, the patient is moved into a lateral decubitus position with the affected side up [see Figure 3]. A soft roll is placed underneath the chest wall to protect the axilla. The bean-bag mattress is molded around the patient and the legs are wrapped in a foam pad to minimize all pressure points. The patient is secured to the operating table with tape placed across the padded lower extremities and a safety strap across the pelvis. The operating table is then flexed at the waist. The combination of the lateral position, the flexed operating table, and the reverse Trendelenburg position facilitates placement of the laparoscopic ports and provides optimal access to the superior retroperitoneum.


Equipment Our preferred instrumentation for laparoscopic adrenalectomy is as follows [see Table 3]. An angled (30°) laparoscope, preferably 5 mm in diameter, is used to optimize viewing angles. One 10/12 mm port is employed to allow insertion of a clip applier and extraction of the specimen; the other ports can all be 5 mm if a 5 mm laparoscope is employed. The principal instruments needed for dissection and hemostasis are atraumatic graspers, an L-hook electrocautery, and a medium-large clip applier. An ultrasonic coagulator is not essential for a right adrenalectomy, but it may facilitate mobilization of the splenic ligaments and dissection of the adrenal from the retroperitoneal fat during a left adrenalectomy. An endovascular stapler should be available for a right adrenalectomy because it will occasionally be needed to divide the right adrenal vein. Other essential items are a suction-irrigation device and an impermeable specimen retrieval bag.

Initial access and placement of trocarsBecause the patient is in a lateral position, initial access to the peritoneal cavity is usually achieved in a closed fashion with a Veress needle. After insufflation to a pressure of 15 mm Hg, a 5 mm direct-view trocar is placed to afford direct visualization of the peritoneal cavity. An open insertion technique may be used instead, but this approach requires a larger incision and is hindered somewhat by the bulky overlapping muscle layers in the subcostal region. Open insertion at the umbilicus is an option in some patients.

Figure 4. Laparoscopic adrenalectomy: port placement

The initial access site is generally at or somewhat medial to the anterior axillary line about two fingerbreadths below the costal margin [see Figure 4]. Subsequent ports should be placed at least 5 cm apart to allow freedom of movement externally. The most dorsal port should be approximately at the posterior axillary line. It is helpful to outline the anterior and posterior axillary lines with a marker before the patient is prepared to ensure that the ports are positioned properly. Whereas four ports are required for a right adrenalectomy, a left adrenalectomy can be done with either three or four ports, depending on the surgeon’s preference and experience. On the left side, the splenic flexure of the colon usually must be mobilized before the fourth port (the most dorsal one) can be inserted.


Right adrenalectomy Step 1: exposure of right adrenal gland and vein.The key to exposure of the right adrenal gland is extensive division of the right triangular ligament of the liver. This maneuver should be continued until the liver can be easily elevated and retracted medially and both the right adrenal and the IVC are visible. A retractor is then inserted through the most medial port to hold the right lobe of the liver up and away from the operative site.

Figure 5. Laparoscopic right adrenalectomy: anatomic exposure for right adrenalectomy

Next, the plane between the medial border of the adrenal and the lateral aspect of the IVC is developed. An L-hook cautery is used for gentle elevation and division of the peritoneum and the small arterial branches here [see Figure 5]. The adrenal is pushed laterally with an atraumatic grasper to apply traction to the dissection site; however, the gland itself should not be grasped, because it is fragile and the capsule and adrenal parenchyma are easily fractured. At all times, it is imperative to know where the lateral border of the IVC is, both to ensure that the dissection is extra-adrenal and to avoid injuring the IVC. The right adrenal vein should come into view as the medial border is dissected.

Figure 6a. Laparoscopic right adrenalectomy: schematic representation
Figure 6b. Laparoscopic right adrenalectomy: intraoperative view

Step 2: isolation, clipping, and division of right adrenal vein. The right adrenal vein is first exposed by gentle blunt spreading, and a right-angle dissector is then used to isolate enough of the vein’s length to permit clip placement [see Figures 6a and 6b]. A medium-large clip is usually sufficient for securing the vein, though sometimes it is necessary to use larger clips or even an endovascular stapler. (We use an endovascular stapler primarily in cases in which the tumor is located in the medial area of the adrenal and the vein must be taken along with a portion of the IVC junction.) Usually, two clips are placed on the IVC side and one or two on the adrenal side, depending on the length of vein available. Meticulous hemostasis throughout the dissection is important: even minimal bleeding will stain the tissue planes and make the dissection more difficult and potentially treacherous.

Step 3: mobilization and detachment of specimen. Once the adrenal vein is divided, dissection is continued superiorly and inferiorly with the L-hook cautery. The numerous small arteries that enter the gland at its superior, medial, and inferior margins can be safely cauterized, but larger branches may have to be clipped. Superiorly, as the adrenal is mobilized, the musculature of the posterior diaphragm is exposed and serves as a marker of the proper plane for the posterior dissection. Inferiorly, the dissection should stay close to the margin of the adrenal so as not to injure branches of the renal hilar vessels. The inferior dissection then proceeds in a medial-to-lateral direction as the gland is elevated off the superior pole of the right kidney. The remaining attachments to the back muscles and the retroperitoneal fat are relatively avascular and can be divided with the electrocautery.

Once the specimen has been detached, it is placed in an impermeable bag. The retroperitoneum is then irrigated and inspected for hemostasis and for secure placement of the clips on the IVC.

Step 4: extraction of specimen. The fascial opening at the 10/12 mm port site is enlarged somewhat, and the specimen bag is removed through this site. For larger tumors, a remote extraction site (e.g., the umbilicus or the suprapubic region) may be a preferable alternative. Large pheochromcytomas may be morcellated within the entrapment bag and removed piecemeal, but ideally, cortical tumors and metastatic lesions should be extracted intact to permit full pathologic examination.


Left adrenalectomy

Figure 7a. Laparoscopic left adrenalectomy: anatomic exposure
Figure 7b. Laparoscopic left adrenalectomy: intraoperative view

Step 1: exposure of left adrenal gland and vein. The splenic flexure of the colon is mobilized. The lateral attachments of the flexure are divided to allow placement of the fourth port (if needed), and the colon is then released from the inferior pole of the spleen and away from the left kidney. Next, the splenorenal ligament is incised from the inferior pole of the spleen to the diaphragm to allow full medial rotation of the spleen and provide access to the left retroperitoneum [see Figures 7a and 7b]. It is important not to dissect lateral to the kidney; doing so will cause the kidney to tilt forward and will interfere with exposure. Once the spleen is completely mobilized, it should fall medially, with minimal or no retraction needed to keep it out of the operative field. Division of the ligaments can be accomplished more quickly and with less bleeding if an ultrasonic coagulator is used.

At this point in the dissection, the tail of the pancreas should be visible, along with the splenic artery and vein. The plane between the pancreas and the left kidney is then developed. The adrenal is located on the superomedial aspect of the kidney just cephalad to the tail of the pancreas and should be visible at this point unless there is a great deal of retroperitoneal fat (as is often the case in patients with Cushing syndrome). If the adrenal gland is not readily visible, laparoscopic ultrasonography should be employed to help locate it and to delineate the surrounding anatomy, particularly the upper kidney and the renal hilar vessels. If the dissection starts too low, the renal hilar vessels or the ureter could be injured.

Figure 8a. Laparoscopic left adrenalectomy: schematic view
Figure 8b. Laparoscopic left adrenalectomy: intraoperative view

Once the adrenal is visualized, the medial and lateral borders are usually defined by means of dissection with the hook cautery and division of areolar attachments and small vessels. The dissection is then continued inferiorly to locate the adrenal vein as it exits the inferomedial border of the gland [see Figures 8a and 8b]. The inferior border of the adrenal often sits adjacent to the left renal vein, from which it can be separated by means of gentle blunt dissection and judicious use of the electrocautery.

Step 2: isolation, clipping, and division of left adrenal vein. Once the adrenal vein has been visualized, it is isolated, doubly clipped, and divided. Because the adrenal vein is usually joined by the inferior phrenic vein cephalad to its junction with the renal vein, it is often necessary to clip the inferior phrenic vein again as the dissection proceeds more proximally.

Step 3: mobilization and detachment of specimen. Once the left adrenal vein has been securely clipped and divided, the dissection is continued cephalad along both the lateral and the medial borders of the gland. Because of the surrounding retroperitoneal fat, it is advisable to use the ultrasonic coagulator for this part of the left-side dissection. Because the left adrenal is more flattened out on the superomedial aspect of the left kidney than the right adrenal is on the right kidney, more of the kidney will be exposed during dissection in a left adrenalectomy than in a right adrenalectomy. Finally, the posterior and superior attachments to the diaphragm and the retroperitoneal fat are divided.

Step 4: extraction of specimen. Once the gland is free, the retroperitoneum is inspected and the specimen extracted as in a right adrenalectomy. If there is any possibility that the pancreatic parenchyma may have been violated, a closed suction drain is left in place.


Retroperitoneal Approach

Retroperitoneal endoscopic adrenalectomy can be carried out with the patient in either a lateral or a prone position. In general, this technique is more challenging to learn than transabdominal adrenalectomy, the working space is more cramped, and it is easier for surgeons to become disoriented unless they have experience working in the retroperitoneum. On the other hand, the retroperitoneal approach allows surgeons to avoid having to reposition patients for bilateral adrenalectomy (if the prone position is used), and it may simplify access in patients who have previously undergone extensive upper abdominal procedures.

Initial access is usually achieved through open insertion of a 12 mm port into the retroperitoneum either (1) just lateral or inferior to the tip of the 12th rib (for the prone position) or (2) in the midaxillary line about 3 cm above the iliac crest (for the lateral position). A potential advantage of the lateral approach is that it can be converted to a transperitoneal approach if difficulty is encountered.

Figure 9. Laparoscopic left adrenalectomy: retroperitoneal endoscopic approach

Once the retroperitoneum is entered, a balloon device is deployed to create an initial working space, which is further developed by means of CO2 insufflation and blunt dissection. The second and third ports are then placed [see Figure 9]. The principles of dissection are the same as in a transabdominal adrenalectomy [see Transabdominal Approach, above]. Laparoscopic ultrasonography may be useful for defining the upper portion of the kidney and the adrenal gland and tumor.

Open Adrenalectomy

Figure 10. Open adrenalectomy

Of the four approaches to open adrenalectomy [see Operative Planning, Choice of Procedure, above], the anterior transabdominal approach is the preferred method for any tumors that are too large to be removed laparoscopically and for all invasive adrenal malignancies. The incision most commonly used is an extended unilateral or bilateral subcostal incision, though a midline incision is also an option [see Figure 10]. The extended subcostal incision yields exposure of both adrenal glands, as well as the rest of the peritoneal cavity. If necessary, it may be extended superiorly in the midline to the xiphoid to provide better upper abdominal exposure for full mobilization of the liver and access to the hepatic veins and the vena cava. The exposure obtained with this incision is sufficient for all but the most extensive adrenal malignancies. If the tumor involves the vena cava, the incision may be extended into a median sternotomy to provide access to the superior vena cava and the heart. The classic thoracoabdominal incision, which extends from the abdomen up through the seventh or eighth intercostal space and through the diaphragm, provides excellent exposure but is associated with increased incision-related morbidity and is rarely used.

Much of the exposure and dissection is the same as in a laparoscopic adrenalectomy; however, because open adrenalectomy is often employed for removal of particularly large tumors, some additional maneuvers may be necessary to achieve adequate exposure and vascular control. For example, it may be helpful to elevate the flank with a roll or a bean-bag mattress and then flex the operating table to open up the space between the costal margin and the iliac crest. Once the abdomen is entered, exploration is carried out for the presence of metastatic disease.

Exposure of the adrenal on the right side is achieved by dividing the right triangular ligament of the liver, as in the laparoscopic approach. The hepatic flexure of the colon is also reflected inferiorly. With large tumors, a Kocher maneuver should be performed to afford better exposure of the vena cava and the renal vessels. The remainder of the dissection proceeds in much the same manner as in a laparoscopic right adrenalectomy. For suspected adrenal malignancies, a wide resection should be carried out, with removal of periadrenal fat and lymphatic tissue and any suspicious lymph nodes. For tumors that appear to involve the vena cava, vascular control of both the IVC proximal and distal to the tumor and the renal veins should be achieved before the lesion is removed.

Open left adrenalectomy entails mobilization of the splenic flexure of the colon and division of the splenorenal ligament. The spleen, the tail of pancreas, and the stomach are reflected medially en bloc to expose the left kidney and the left adrenal. The left adrenal vein is ligated with clips or silk ties near its junction with the renal vein. The remainder of the dissection proceeds as in a laparoscopic left adrenalectomy. For left-side primary adrenal malignancies, periaortic lymphatic vessels and lymph nodes should be removed along with the specimen. If a large left-side tumor is invading adjacent structures, removal may require en bloc resection of the spleen, the distal pancreas, and the kidney.


Inability to Locate Adrenal

Figure 11. Laparoscopic ultrasonogram of enlarged adrenal gland

The adrenal is usually not difficult to find on the right side, where it should be visible once the right hemiliver has been mobilized. Important landmarks on that side are the IVC, which is medial to the adrenal, and the kidney, which is inferior to the adrenal. Once these structures have been identified, the location of the adrenal should be apparent. In contrast, the adrenal can be difficult to find on the left side, especially if the tumor is small or the patient is obese. To locate the left adrenal, the splenorenal ligament should be fully divided, and then the plane between the kidney and the tail of the pancreas should be developed, with the tail of the pancreas rotated medially. As dissection proceeds superiorly, the adrenal can be visually distinguished from the retroperitoneal fat by its golden-orange appearance. If the adrenal is not yet visualized at this point, laparoscopic ultrasonography should be used to verify the locations of the superior pole of the left kidney and the renal vessels. Ultrasonography should also be able to image the adrenal gland and tumor within the retroperitoneal fat [see Figure 11].


The best means of managing bleeding problems during adrenalectomy is prevention. Important measures for minimizing bleeding risk include obtaining good exposure of the operative field and employing meticulous dissection and gentle handling of the adrenal and surrounding structures. When bleeding does occur, it may be from the adrenal veins, the adrenal gland itself, the IVC, the renal veins, the liver, the pancreas, the spleen, or the kidney. For bleeding during laparoscopic adrenalectomy, the first maneuver should be to tamponade the bleeding site with an atraumatic instrument. If this maneuver is successful, dissection should be directed away from the bleeding site for a while, until better exposure of the area can be obtained. Major hemorrhage from the IVC or the renal veins that is not immediately controlled should be managed by prompt conversion to open adrenalectomy (see below). Lesser bleeding may also be an indication for conversion to open adrenalectomy if it obscures the tissue planes and thereby increases the risk of inadvertent entry into the adrenal gland or tumor.

Conversion to Open Adrenalectomy

Conversion to open adrenalectomy may sometimes be necessary because of bleeding, failure to progress with the dissection, or a locally invasive tumor. If the patient is in the lateral decubitus position, conversion may be accomplished by means of a subcostal incision extended into the flank. With the patient on a bean-bag mattress, the operating table can be rotated out of the straight lateral plane so that the patient comes to occupy more of a hemilateral position. If the procedure is a bilateral adrenalectomy, then either a bilateral subcostal incision or a midline incision may be employed after the patient has first been returned to more of a supine position. For this reason, it is important to extend the initial preparation and draping past the midline of the abdomen. Alternatively, if the conversion is not being done on an urgent basis because of bleeding, the port sites may be closed, and the patient may then be moved into the supine position, reprepared, and redraped.

An option that may be considered before conversion to an open procedure is the use of a hand access port. A hand-assisted technique may be particularly useful for larger, noninvasive tumors that are harder to manipulate with laparoscopic instruments. The location of the incision for the hand port may vary according to the patient’s body habitus; generally, however, an ipsilateral subcostal location medial to the working ports allows adequate hand access while preserving visualization through the more lateral ports.

Large Tumors

Large adrenal tumors (> 6 to 8 cm) are more difficult to remove than smaller ones because they are bulkier and more vascular and because they are harder to manipulate and retract. Accordingly, the dissection should stay extra-adrenal, and care must be exercised during manipulation to avoid entering the tumor. Although surgeons have not yet had a great deal of experience with hand-assisted laparoscopic adrenalectomy, it appears that this approach may be useful for exposure and retraction of large tumors and may facilitate extraction of large specimens. Laparoscopic ultrasonography should also be employed to verify that the tumor is well circumscribed and noninvasive. Surgeons who attempt to remove large adrenal tumors laparoscopically should be highly experienced in laparoscopic adrenalectomy techniques.

For large invasive adrenal malignancies, an open approach, involving a generous bilateral subcostal incision, is indicated, and the chest should be prepared in case a thoracoabdominal or median sternotomy extension proves necessary. The important principles are to obtain wide exposure of the operative field and to control all major vessels that may be involved before removing the tumor.

Obese Patients

Obese patients present a particular challenge during adrenalectomy, for several reasons: initial access is more difficult; retraction and exposure are more challenging; and the copious amount of retroperitoneal fat makes it difficult to identify the adrenal and to clearly define the margins of the gland within the retroperitoneum. Our practice is to attempt to gain initial access to the peritoneal cavity by using a closed Veress needle technique. Because resting intra-abdominal pressure may be higher in obese patients, especially if they are in the lateral position, it may be necessary to increase the CO2 pressure to 20 mm Hg temporarily until the first trocar is inserted. If it proves difficult to establish pneumoperitoneum with the Veress needle technique, the initial trocar should be placed at the umbilicus by means of an open insertion technique. The subcostal and flank ports are then inserted in the usual locations under direct vision.

On the right side, the presence of a bulky, fatty liver should be anticipated, and the locations of the port sites should be adjusted accordingly by placing them somewhat more caudad. Ample time should be taken to mobilize the liver fully so that the adrenal and the IVC can be safely accessed. On the left, the ports may be placed in the standard locations. The splenic flexure should be fully mobilized, as should the spleen and the tail of the pancreas. Laparoscopic ultrasonography is often needed to locate the gland in the retroperitoneal fat. In addition, use of an ultrasonic coagulator may facilitate division of the retroperitoneal fat.

Postoperative Care

After a laparoscopic adrenalectomy, most patients are admitted to a regular nursing unit, though some patients with pheochromocytomas will need to undergo a short stay in the intensive care unit for invasive monitoring. Patients are started on clear liquids on postoperative day 1, and the diet is advanced as tolerated. The urinary catheter is usually removed on postoperative day 1. Intravenous analgesia is switched to oral analgesia as soon as the patient can tolerate oral feeding. A complete blood count is obtained on postoperative day 1 in all patients, and electrolyte levels are monitored in patients with aldosteronomas and hypercortisolism. Patients with Cushing syndrome should be given stress doses of steroids perioperatively and should be discharged on a maintenance prednisone dosage of 10 to 15 mg/day in divided doses. These patients should be advised that it may take 6 to 12 months or longer for the contralateral adrenal to recover to the point where prednisone can be discontinued. Patients undergoing bilateral adrenalectomy will need lifelong replacement therapy with a glucocorticoid (e.g., prednisone) and a mineralocorticoid (e.g., fludrocortisone acetate, 0.1 mg/day).

Postoperative management of hypertensive medications depends on the pathology of the underlying adrenal lesion. In patients with aldosteronomas, spironolactone is stopped immediately after adrenalectomy, and the other antihypertensive agents are usually continued while blood pressure is monitored closely on an outpatient basis; further medication reductions are made as clinically warranted. In most patients with Cushing syndrome, antihypertensive medications are continued, whereas in most patients with pheochromocytomas, they are not. In both sets of patients, however, close outpatient monitoring of blood pressure should be carried out in the early postoperative period. Adrenalectomy can have a dramatic impact on hypertensive control and can lead to hypotension if medications are not appropriately adjusted.

In most routine cases, patients can be discharged within 24 hours after a laparoscopic adrenalectomy, though some patients will have to stay longer for blood pressure monitoring, for adjustment of steroid replacement therapy, or for resumption of a regular diet. After an open adrenalectomy, resumption of an oral diet takes longer, and postoperative hospital stays of 4 to 5 days are more typical.

After discharge, patients are seen in the clinic within 2 to 3 weeks for a wound check, blood pressure evaluation, and a review of antihypertensive medications. In patients who underwent adrenalectomy for an aldosteronoma, electrolyte levels and the creatinine concentration should be checked. In patients who underwent adrenalectomy for pheochromocytoma, yearly clinical and biochemical follow-up is indicated, with measurement of either plasma levels of fractionated metanephrines or urine levels of catecholamines and metanephrines. In selected patients on steroid replacement therapy who are proving difficult to wean from prednisone, an ACTH stimulation test may be necessary to assess the responsiveness of the pituitary-adrenal axis.


It appears that laparoscopic adrenalectomy has a major advantage over open adrenalectomy in terms of the incidence of postoperative complications. In a meta-analysis of 98 adrenalectomy series reported between 1980 and 2000, the overall complication rate was 10.9% with laparoscopic procedures and 25.2% with open procedures.10 This difference between the complication rates was primarily attributable to the occurrence of fewer wound, pulmonary, and infectious complications in the laparoscopic series. The most common complication of laparoscopic adrenalectomy is bleeding, which was reported in 4.7% of patients from the series reviewed in the meta-analysis. Bleeding is also the most common reason for conversion to open adrenalectomy; however, major bleeding that leads to transfusion is relatively uncommon. The risk of bleeding can be minimized by obtaining meticulous hemostasis, taking care not to grasp the adrenal gland, and handling tissue gently. If bleeding does occur, the prudent course of action is to maintain pressure on the bleeding source while obtaining better exposure or even starting the dissection in another area, rather than to resort to indiscriminate use of clips or electrocautery. The surgeon must be prepared to convert rapidly to an open procedure should major hemorrhage occur.

Other potential complications of adrenalectomy (either laparoscopic or open) include injury to the tail of the pancreas (with resultant pancreatic leakage or pancreatitis), injury to the diaphragm, and pneumothorax. Wound infections are uncommon with laparoscopic adrenalectomy. Trocar site hernias are infrequent as well, provided that the fascia is closed at all port sites that are 10 mm or larger. Deep vein thrombosis occurs in 0.8% of cases, pulmonary embolism in 0.5%.10 Pneumatic compression stockings should be used perioperatively to minimize the risk of venous thromboembolism. Renovascular hypertension from injury to the renal artery has also been reported.11,12 The operative mortality associated with laparoscopic adrenalectomy is about 0.3%.

Several cases of local or regional tumor recurrence have been reported after laparoscopic adrenalectomy. In most of these cases, the tumors removed were either suspected or unsuspected adrenal malignancies, and the extensive nature of the recurrences was probably related to aggressive tumor biology rather than to the minimally invasive surgical technique. In some of the cases, however, the pattern of recurrence, characterized by the development of multiple intraperitoneal or port site metastases, suggested that laparoscopic dissection and pneumoperitoneum might have contributed to tumor spread.13–16 One group treated three patients for recurrent pheochromocytomatosis that developed after laparoscopic adrenalectomy.17 These patients were found to have multiple small tumor nodules in the adrenalectomy bed during open reoperation after removal of apparently benign pheochromocytomas. Fragmentation of the tumor and excessive tumor manipulation during the laparoscopic dissection were considered the probable mechanisms of tumor recurrence.

These reports highlight the need for caution in approaching large, malignant, or potentially malignant adrenal tumors. Surgeons who attempt a laparoscopic approach in this setting should be highly experienced in laparoscopic adrenalectomy techniques, and the tumor should be well circumscribed and not locally invasive. The use of a hand port may be a valuable adjunct to resection in these cases. Regardless of the specific surgical approach followed, wide excision of the lesion along with the surrounding periadrenal fat is crucial for minimizing recurrence rates in this population.

Outcome Evaluation

The safety and efficacy of laparoscopic adrenalectomy for the removal of small, benign adrenal tumors have been clearly established. Rates of conversion to open adrenalectomy in high-volume centers have ranged from 3% to 13%, and operating times have averaged 2 to 3 hours.9,11,18–23 Most patients are now discharged from the hospital within 24 to 48 hours after operation. Although no prospective, randomized trial comparing laparoscopic with open adrenalectomy has been carried out, several retrospective studies have consistently shown that the laparoscopic approach is associated with decreased pain, a shorter hospital stay, and a faster recovery.24–27 Complication rates have also been low, and overall, complications appear to be less common than with open adrenalectomy.10

The results of a laparoscopic approach in patients with large (> 6 cm) adrenal tumors or malignant primary or metastatic adrenal lesions have been reviewed28; generally, the conversion rates for large or malignant tumors have been higher than those reported in other laparoscopic adrenalectomy series. Overall, tumor recurrence rates after laparoscopic adrenalectomy have been low.29–33 In one series, however, local or regional tumor recurrence developed in three of five patients with adrenocortical carcinomas that were treated laparoscopically.34 Other groups have also published anecdotal reports of local tumor recurrences after resection of unsuspected adrenal carcinomas.13–16 Whether these recurrences were related primarily to the surgical technique employed or to the underlying tumor biology is unclear. It would appear, therefore, that in most cases, primary adrenal malignancies are best approached in an open fashion unless the tumor is small and well circumscribed and the surgeon is highly experienced.


Eric C. Poulin, MD, MSc, FACS, FRCSC

Professor and Chair, Department of Surgery
University of Ottawa Faculty of Medicine
Ottawa Wilbert J. Keon Hospital

Christopher M. Schlachta, MD, FACS

Lecturer and General Surgeon, Department of Surgery
University of Toronto Faculty of Medicine

Joseph Mamazza, MD, FRCSC

Assistant Professor, Department of Surgery
University of Toronto Faculty of Medicine
Medical Director of Minimal Access Therapeutics and Diseases of Digestive Systems and Director of Minimally Invasive Surgery
St. Michael’s Hospital

Medicine is not an exact science, and nowhere is this observation more appropriate than in the operating room when a spleen is being removed.1

The first reported splenectomy in the Western world was performed by Zacarello in 1549, though the veracity of his operative description has been questioned. Between this initial report and the 1800s, very few cases were recorded. The first reported splenectomy in North America was performed by O’Brien in 1816. The patient was in the act of committing a rape when his victim plunged a large knife into his left side. As in this case, most early splenectomies were done in patients who had undergone penetrating trauma; often, the spleen was protruding from the wound and the surgeon proceeded with en masse ligation. The first elective splenectomy was performed by Quittenbaum in 1826 for sequelae of portal hypertension, and soon afterward, Wells performed one of the first splenectomies using general anesthesia; both patients died. In 1866, Bryant was the first to attempt splenectomy in a patient with leukemia. Over the following 15 years, 14 splenectomies were attempted as therapy for leukemia; none of the patients survived. In a 1908 review of 49 similar cases, Johnston reported a mortality of 87.7%.2 These dismal results led to the abandonment of splenectomy for leukemia. In 1916, Kaznelson, of Prague, was the first to report good results from splenectomy in patients with thrombocytopenic purpura.

As the 20th century progressed, splenectomy became more common in direct proportion to the increase in the use of the automobile. The eventual recognition of the syndrome known as overwhelming postsplenectomy infection (OPSI) made splenic conservation an important consideration. Partial splenectomy had initially been described by the French surgeon Péan in the 19th century. This procedure received little further study until almost 100 years later, when the Brazilian surgeon Campos Cristo reevaluated Péan’s technique in his report of eight trauma patients treated with partial splenectomy.3 Simpson’s report on 16 children admitted for splenic trauma to the Hospital for Sick Children in Toronto between 1948 and 1955 was instrumental in establishing the validity of nonoperative treatment of splenic trauma [see 7:7 Injuries to the Liver, Biliary Tract, Spleen, and Diaphragm ].4

In late 1991 and early 1992, four groups working independently—Delaître in Paris, Carroll in Los Angeles, Cushieri in the United Kingdom, and our group in Canada—published the first reports of laparoscopic splenectomy in patients with hematologic disorders.5–7 Since then, the development of operative techniques for partial laparoscopic splenectomy has tested the limits of minimally invasive surgery and encouraged clinical research into methods of simplifying the execution of the operation.8,9 The adoption of laparoscopic splenectomy has led to a gradual decrease in the indications for open splenectomy; however, both procedures are still essential components of spleen surgery.

Anatomic Considerations

Most anatomy texts suggest that the splenic artery is constant in its course and branches; however, as the classic essay by Michels made clear, each spleen has its own peculiar pattern of terminal artery branches.10

Splenic Artery

The celiac axis is the largest but shortest branch of the abdominal aorta: it is only 15 to 20 mm long. The celiac axis arises above the body of the pancreas and, in 82% of specimens, divides into three primary branches: the left gastric artery, which is the first branch, and the hepatic and splenic arteries, which derive from a common stem. In rare instances, the splenic artery originates directly from the aorta; even less often, a second splenic artery arises from the celiac axis. There are numerous other possible variations, in which the splenic artery may originate from the aorta, the superior mesenteric artery, the middle colic artery, the left gastric artery, the left hepatic artery, or the accessory right hepatic artery. As a rule, however, the splenic artery arises from the celiac axis to the right of the midline, which means that the aorta must be crossed to reach the spleen and that selective angiography is likely to be difficult at times. The splenic artery can take a very tortuous course, particularly in patients who are elderly or who have a longer artery.

Figure 1. Splenic vascularization

In his study of 100 cadaver spleens,10 Michels divided splenic arterial geography into two types, distributed and magistral (or bundled) [see Figure 1]. In the distributed type, found in 70% of dissections, the splenic trunk is short, and six to 12 long branches enter the spleen over approximately 75% of its medial surface. The branches originate between 3 and 13 cm from the hilum [see Figure 1, part a]. In the bundled type, found in the remaining 30% of dissections, there is a long main splenic artery that divides near the hilum into three or four large, short terminal branches that enter the spleen over only 25% to 35% of its medial surface. These short splenic branches originate, on average, 3.5 cm from the spleen, and they reach the center of the organ as a compact bundle [see Figure 1, part b]. Early identification of the type of splenic blood supply present can help the surgeon estimate how difficult a particular splenectomy is likely to be. Operation on a spleen with a distributed vascular anatomy usually involves dissection of more blood vessels; however, the vessels, being spread over a wider area of the splenic hilum, are relatively easy to deal with. Operation on a spleen with a bundled-type blood supply typically involves dissection of fewer vessels; however, because the hilum is narrower and more compact, dissection and separation of the vessels are more difficult.

Branches of Splenic Artery

The splenic branches vary so markedly in length, size, and origin that no two spleens have the same anatomy. Outside the spleen, the arteries also frequently form transverse anastomoses with each other that, like most collaterals, arise at a 90° angle to the vessels involved [see Figure 1].11 As a consequence, attempts to occlude a branch of the splenic artery by means of clips or embolization, if carried out proximal to such an anastomosis, may fail to devascularize the corresponding splenic segment. Before the splenic trunk divides, it usually gives off a few slender branches to the tail of the pancreas. The most important of these is called the pancreatica magna (a vessel familiar to vascular radiologists); occlusion of this branch with embolic material has been reported to result in pancreatitis. Next, the splenic artery divides into two to six first and second terminal branches, and these branches undergo two further levels of division into two to 12 penultimate and ultimate branches. Segmental and subsegmental division can occur either outside or inside the spleen. The number of arteries entering the spleen ranges from six to 36. The size of the spleen does not determine the number of arteries entering it; however, the presence of notches and tubercles usually correlates well with a higher number of entering arteries.

Figure 2. Division of splenic artery branches

A reasonable general scheme of splenic artery branches might include as many as seven principal branches at various division levels and in various anatomic arrangements: (1) the superior terminal artery, (2) the inferior terminal artery, (3) the medial terminal artery, (4) the short gastric arteries, (5) the left gastroepiploic artery, (6) the inferior polar artery, and (7) the superior polar artery [see Figure 2]. Veins are usually located behind the corresponding arteries, except at the ultimate level of division, where they may be either anterior or posterior.

First Terminal Division Branches

A classic study from 1917 found that 72% of specimens had three terminal branches (superior polar, superior terminal, and inferior terminal) and 28% had two12; the medial terminal artery was observed in only 20% of cases. When the superior terminal artery is excessively large, the inferior terminal is rudimentary, with an added blood supply often coming from the left gastroepiploic and polar vessels.

Second Terminal Division Branches

Superior polar artery The superior polar artery is present in 65% of patients. It usually arises from the main splenic trunk (75% of cases) or the superior terminal artery (20% of cases), but on occasion, it may originate from the inferior terminal artery or separately from the celiac axis (thus providing the spleen with a double splenic artery). In most instances, the superior polar artery gives rise to one or two short gastric branches; rarely, it gives rise to the left inferior phrenic and pancreatic rami. The presence and size of this artery appear to be correlated with tubercle formation, in that it is more prominent in spleens with large tubercles. The superior polar artery is frequently very long and slender and thus easily torn during splenectomy; accordingly, it was suggested in 1928 that ligation of splenic branches be started from the inferior pole of the spleen.13

Inferior polar artery The inferior polar artery is present in 82% of cases. As many as five collateral branches may arise from the splenic trunk, the inferior terminal artery, or, as noted, the left gastroepiploic artery. Inferior polar branches may have multiple origins, and they tend to be of smaller caliber than the superior polar artery.

Left gastroepiploic artery The left gastroepiploic artery, the most varied of the splenic branches, courses along the left side of the greater curvature in the anterior layer of the greater omentum. In 72% of cases, it arises from the splenic trunk several centimeters from its primary terminal division, and in 22% of cases, it originates from the inferior terminal artery or its branches; however, it may also originate from the middle of the splenic trunk or from the superior terminal artery. Characteristically, the left gastroepiploic artery gives off inferior polar arteries, which vary in number (ranging from one to five), size, and length. Typically, these branches are addressed first during laparoscopic splenectomy. When they are small, they can usually be controlled with the electrocautery.


Short gastric arteries As many as six short gastric arteries may arise from the fundus of the stomach, but as a rule, only the one to three that open into the superior polar artery must be ligated during laparoscopic splenectomy [see Figure 1].

Suspensory Ligaments of Spleen and Tail of Pancreas

Figure 3. Suspensory ligaments of the spleen

Duplications of the peritoneum form the many suspensory ligaments of the spleen [see Figure 3]. Medially and posteriorly, the splenorenal ligament contains the tail of the pancreas and the splenic vessels. Anteriorly, the gastrosplenic ligament contains the short gastric and gastroepiploic arteries. In the lateral approach to laparoscopic splenectomy [see Operative Technique, below], the splenorenal and gastrosplenic ligaments are easily distinguished, and dissection of the anatomic structures they contain is relatively simple. In the anterior approach, these two ligaments lie on top of each other, and to separate them correctly and safely requires considerable experience with splenic anatomy.

The phrenicocolic ligament courses laterally from the diaphragm to the splenic flexure of the colon; its upper portion is called the phrenicosplenic ligament. The attachment of the lower pole on the internal side is called the splenocolic ligament. Between these two structures, a horizontal shelf of areolar tissue, known as the sustentaculum lienis, is formed on which the inferior pole of the spleen rests. The sustentaculum lienis is often molded into a sac that opens cephalad and acts as a support for the lower pole. This structure, often overlooked during open procedures, is readily visible through a laparoscope. The phrenicocolic ligament, the splenocolic ligament, and the sustentaculum lienis are usually avascular, except in patients who have portal hypertension or myeloid metaplasia.

A 1937 study found that the tail of the pancreas was in direct contact with the spleen in 30% of cadavers.14 A subsequent report confirmed this finding and added that in 73% of patients, the distance between the two structures was no more than 1 cm.15 Care must be exercised to avoid damage with the electrocautery during dissection as well as damage with the linear stapler in the course of en masse ligation of the splenic hilum (a maneuver more easily performed via the lateral approach to laparoscopic splenectomy).

Laparoscopic Splenectomy

Preoperative Evaluation

Currently, we consider all patients evaluated for elective splenectomy to be potential candidates for laparoscopic splenectomy. Contraindications to a laparoscopic approach include severe portal hypertension, uncorrectable coagulopathy, severe ascites, and most traumatic injuries to the spleen. Extreme splenomegaly remains a relative contraindication as well. Because most patients scheduled for laparoscopic splenectomy have hematologic disorders, they undergo the same hematologic preparation that patients scheduled for open surgery do—namely, steroids and g-globulins (when required). Ultrasonography is performed to determine the size of the spleen. Spleen size is expressed in terms of the maximum interpole length (i.e., the length of the line joining the two organ poles) and is generally classified into three categories: (1) normal spleen size (< 11 cm), (2) moderate splenomegaly (11 to 20 cm), and (3) severe splenomegaly (> 20 cm).16 Because extremely large spleens present special technical problems that test the current limits of laparoscopic surgery, we make use of a fourth category for spleens longer than 30 cm or heavier than 3 kg, which we call megaspleens [see Table 1]. The ultrasonographer is also asked to try to identify any accessory spleens that may be present. Computed tomography is done when there is doubt about the exactness of the ultrasonographic measurement; such measurement is sometimes inaccurate at the upper pole and with spleens longer than 16 cm.

Patients receive thorough counseling about the consequences of the asplenic state. Polyvalent pneumococcal vaccine is administered at least 2 weeks before operation in all cases; preoperative vaccination against Haemophilus influenzae and meningococci is also advisable. Heparin prophylaxis for thrombophlebitis is administered according to standard guidelines, provided that there is no hematologic contraindication [see 6:6 Venous Thromboembolism]. Non steroidal anti-inflammatory drugs (NSAIDs) are often given orally before operation to minimize postoperative pain; however, on empirical grounds, NSAIDs are not used when heparin prophylaxis is employed. Platelets are rarely, if ever, required when laparoscopic splenectomy is performed for idiopathic (immune) thrombocytopenic purpura (ITP).

Operative Planning

Laparoscopic splenectomy presents special problems, such as the necessity of dealing with a fragile and richly vascularized organ that is situated close to the stomach, the colon, and the pancreas and the difficulty of devising an extraction strategy that is compatible with proper histologic confirmation of the pathologic process while maintaining the advantages of minimal access surgery. For successful performance of laparoscopic splenectomy, a detailed knowledge of both splenic anatomy and potential complications is essential. The operative strategy is largely determined by the anatomic features, which, as noted [see Anatomic Considerations, above], may vary considerably from patient to patient.17

Operative Technique

Lateral Approach

This approach was first described in connection with laparoscopic adrenalectomy and is currently used for most laparoscopic splenectomies.18 At present, the only indication for the anterior approach to laparoscopic splenectomy is the presence of massive splenomegaly or a megaspleen. Typically, this alternative approach is taken when a spleen reaches or exceeds 23 cm in length or 3 kg in weight.

Step 1: placement of trocars

Figure 4. Laparoscopic splenectomy: lateral approach

The patient is placed in the right lateral decubitus position, much as he or she would be for a left-side posterolateral thoracotomy. The operating table is flexed and the kidney bolster raised to increase the distance between the lower rib and the iliac crest. Usually, four 12 mm trocars are used around the costal margin so that the camera, the clip applier, and the linear stapler can be interchanged with maximum flexibility [see Figure 4]. The trocars must be far enough apart to permit good working angles. Some advantage may be gained from tilting the patient slightly backward; this step gives the operating team more freedom in moving the instruments placed along the left costal margins, especially during lifting movements, when it is easy for instrument handles to touch the operating table. For the same reason, it is also advisable to place the anterior or abdominal side of the patient closer to the edge of the operating table.

A local anesthetic is infiltrated into the skin at the midpoint of the anterior costal margin, and a 12 mm incision is made. The first trocar is inserted under direct vision, and a symmetrical 15 mm Hg pneumoperitoneum is created. The locations of the remaining trocars are determined by considering the anatomic configuration in relation to the size of the spleen to be excised. In most cases, the fourth posterior trocar cannot be inserted until the splenic flexure of the colon has been mobilized. Accordingly, the procedure is usually started with three trocars in place.


TroubleshootingAfter years of using the Veress needle, we now prefer the open method of inserting the first trocar. It is true that use of the Veress needle is for the most part safe; however, the small number of catastrophic complications that occur with blind methods of first trocar insertion are more and more difficult to justify. Admittedly, these complications are infrequent, and thus, it is unlikely that even a large randomized trial would be able to show any significant differences between various methods of first trocar insertion. Nevertheless, even though complications occur with the open method of first trocar insertion as well, they are very uncommon and tend to be limited to trauma to the intestine or the omental blood vessels; they do not have the same serious consequences as the major vessel injury that may arise from blind trocar insertion.

Figure 5a. Lateral approach: alternative trocar placement
Figure 5b. Lateral approach: needlescopic trocar placement

Trocar placements differing from the ones we describe may be considered. More experienced surgeons (or those simply wishing to make the procedure easier) may choose to replace one or two 12 mm trocars with 5 mm trocars [see Figure 5a]. The procedure can also be performed with only three trocars. In leaner patients, one of the trocars can be inserted into the umbilicus to gain a cosmetic advantage. The advent of needlescopic techniques has made it possible to replace some of the 5 and 12 mm trocars with 3 mm trocars. The ultimate (i.e., least invasive) technique, usually reserved for lean patients with ITP and normal-size spleens, involves one 12 mm trocar placed in the umbilicus and two 3 mm trocars placed subcostally [see Figure 5b]. This approach requires two different camera-laparoscope setups, so that a 3 mm laparoscope can be interchanged with a 10 mm laparoscope as necessary to permit application of clips or staplers through the umbilical incision once the dissection is completed. The specimen is then retrieved through the umbilicus. Because the use of 3 mm laparoscopes is accompanied by a decrease in available intra-abdominal light and focal width, a meticulously bloodless field and sophisticated surgical judgment are critical for successful performance of needlescopic splenectomy.


Step 2: search for and retrieval of accessory spleens

Figure 6a. Lateral approach: locations of accessory spleens
Figure 6b. Lateral approach: accessory spleen

The camera is inserted, and the stomach is retracted medially to expose the spleen. Then a fairly standard sequence is followed. A thorough search is then made for accessory spleens. To maximize retrieval, all known locations of accessory spleens should be carefully explored [see Figure 6aandFigure 6b]. Any accessory spleens found should be removed immediately; they are considerably harder to locate once the spleen is removed and the field is stained with blood.

TroubleshootingIt is especially important to retrieve accessory spleens from patients with ITP, in whom the presence of overlooked accessory spleens has been associated with recurrence of the disease. Remedial operation for excision of missed accessory spleens has been reported to bring remission of recurrent disease; such operation can be performed laparoscopically. The overall retrieval rate for accessory spleens should fall between 15% and 30%.

Splenic activity has been demonstrated after open and laparoscopic splenectomy for trauma and hematologic disorders19,20; accordingly, it is advisable to wash out and recover all splenic fragments resulting from intraoperative trauma at the end of the procedure. This step is particularly important for patients with ITP, in whom intraoperative trauma to the spleen is thought to contribute to postoperative scan-detectable splenic activity. As of this writing, we have recovered accessory spleens in 33% of ITP cases treated laparoscopically.


Step 3: control of vessels at lower pole, demonstration of ‘splenic tent,’ and incision of phrenicocolic ligament

Figure 7. Lateral approach: splenic tent

The splenic flexure is partially mobilized by incising the splenocolic ligament, the lower part of the phrenicocolic ligament, and the sustentaculum lienis. The incision is carried slightly into the left side of the gastrocolic ligament. This step affords access to the gastrosplenic ligament, which can then be readily separated from the splenorenal ligament to create what looks like a tent. This maneuver cannot be accomplished in all cases, but when it can be done, it simplifies the procedure considerably. The walls of this so-called splenic tent are made of the gastrosplenic ligament on the left and the splenorenal ligament on the right, and the floor is made up of the stomach. In fact, this maneuver opens the lesser sac in its lateral portion (a point that is better demonstrated with gentle upward retraction of the splenic tip) [see Figure 7].

Figure 8. Lateral approach: trocar placement around iliac crest

The branches of the left gastroepiploic artery are controlled with the electrocautery or with clips, depending on the size of the branches. The avascular portion of the gastrosplenic ligament, situated between the gastroepiploic artery and the short gastric vessels, is then incised sufficiently to expose the hilar structures in the splenorenal ligament. To accomplish this, the lower pole is gently elevated; in this position, the spleen almost retracts itself as it naturally falls toward the left lobe of the liver. At this point, the surgeon can usually assess the geography of the hilum and determine the degree of difficulty of the operation. The fourth trocar, if needed, is then placed posteriorly under direct vision, with care taken to avoid the left kidney. Caution must also be exercised in placing the trocars situated immediately anterior and posterior to the iliac crest. The iliac crest can impede movement and hinder upward mobilization of structures if the trocars are placed over it rather than in front of or behind it [see Figure 8].

Finally, the phrenicocolic ligament is incised all the way to the left crus of the diaphragm, either with a monopolar electrocautery with an L hook or with scissors. A small portion of the ligament is left to keep the spleen suspended and facilitate subsequent bagging. The phrenicocolic ligament is avascular except in patients with portal hypertension or myeloproliferative disorders (e.g., myeloid metaplasia). Leaving 1 to 2 cm of ligament all along the spleen side facilitates retraction and handling of the spleen with instruments.


TroubleshootingRemarkably few instruments are needed for laparoscopic splenectomy: most of the operation is done with three reusable instruments. A dolphin-nose 5 mm atraumatic grasper is used to elevate and hold the spleen. It is also used to separate tissue planes and vessels with blunt dissection because its atraumatic tip is easily insinuated between tissue planes. A gently curved 5 mm fine-tip dissector (Crile or Maryland) and a 10 mm 90° right-angle dissector are the only other tools required for cost-efficient dissection.

When a powered instrument is called for, we use a monopolar electrocautery with an L hook or a gently curved scissors. Alternatively, an ultrasonic dissector or a tissue-welding device may be used, albeit at a much higher cost.


Step 4: dealing with splenic hilum and tailoring operative strategy to anatomy

Figure 9a. Lateral approach: clipping vessels in hilum (1)
Figure 9b. Lateral approach: clipping vessels in hilum (2)

It is advisable to base one’s operative strategy on the specific splenic anatomy. If a distributed anatomy is present, the splenic branches are usually dissected and clipped. This is not only the least costly approach but also the simplest, in that the vessels are spread over a wider area of the splenic hilum and are easier to dissect and separate [see Figures 9aand9b].

Figure 10a. Lateral approach: stapling (device open)
Figure 10b. Lateral approach: stapling (device closed)
Figure 10c. Lateral approach: stapling (completed)

A bundled anatomy lends itself more to a single use of the linear stapler, provided that the tail of the pancreas is identified and dissected away when required. When possible, a window is created above the hilar pedicle in the splenorenal ligament so that all structures can be included within the markings of the linear stapler under direct vision [see Figures 10a, 10b, and10c]. The angles provided by the various trocars make this maneuver much easier via the lateral approach than via the anterior approach. Dissection continues with individual dissection and clipping of the short gastric vessels; occasionally, these vessels can also be taken en masse with the linear stapler. So far, we have not used sutures in this setting, except once to control a short gastric vessel that was too short to be clipped safely. This portion of the operation is performed while the spleen is hanging from the upper portion of the phrenicocolic ligament, which has not yet been entirely cut.


TroubleshootingIt is at this point in the procedure that experience in designing the operative strategy pays off in reduced operating time. Because of the many variations in size, shape, vascular patterns, and relations to adjacent organs, spleens are almost as individual as fingerprints. Accordingly, an experienced spleen surgeon learns to keep an open mind with regard to operative strategy and must be able to call on a wide range of skills to facilitate the procedure.

The surgeon should start by looking at the internal surface of the spleen. If the splenic vessels cover more than 75% of the internal surface (as is the case in 70% of patients), a distributed anatomy is present. With a distributed vascular anatomy, the vessels tend to be easier to dissect and isolate and thus can be readily (and cost-effectively) controlled with clips. On the other hand, if the splenic vessels entering the spleen cover only 25% to 35% of the inner surface of the hilum (30% of patients), the pattern is bundled. With a bundled vascular anatomy, the vessels, being fewer and closer together, can usually be controlled with a single application of the vascular stapler across the hilum, provided that the tail of the pancreas can be protected.


Step 5: extraction of spleen

Figure 11. Lateral approach: insertion of specimen bag

A medium-size or large heavy-duty plastic freezer bag, of the sort commercially available in grocery stores, is used to bag the spleen. This bag is sterilized and folded, then introduced into the abdominal cavity through one of the 12 mm trocars [see Figure 11]. The bag is unfolded and the spleen slipped inside to prevent splenosis during the subsequent manipulations. Grasping forceps are used to hold the two rigid edges of the bag and to effect partial closure. Bagging is facilitated by preserving the upper portion of the phrenicocolic ligament. After final section of the phrenicocolic ligament and any diaphragmatic adhesions present, extraction is performed through one of the anterior port sites. Extraction through a posterior site is more difficult because of the thickness of the muscle mass; usually, the incision must be opened, and more muscle must be fulgurated than is desirable.

Figure 12. Lateral approach: specimen bag position

The subcostal or umbilical incision through which extraction is to take place is extended slightly. A grasping forceps is inserted through the extraction incision to hold the edges of the bag inside the abdomen. Gentle traction on the bag from outside brings the spleen close to the peritoneal surface of the umbilical incision and then out of the wound [see Figure 12]. Specimen retrieval bags have been developed that can accommodate a normal-size spleen and thus make bagging much easier, but they are costly.

A biopsy specimen of a size suitable for pathologic identification is obtained by incising the splenic tip. The spleen is then fragmented with finger fracture, and the resulting blood is suctioned. The remaining stromal tissue of the spleen is then extracted through the small incision, hemostasis is again verified, and all trocars are removed. No drains are used. The incisions are closed with absorbable sutures and paper strips.


Troubleshooting The freezer bags can be more easily introduced into the abdomen if they are pulled in rather than pushed in [see Figure 11]. This may be accomplished by bringing out a 5 mm toothed grasper through the introduction trocar from another properly angled trocar, grasping the specimen bag, and pulling the bag back down through the trocar. A laparoscopic hernia mesh introducer may also be used.

Slipping the spleen into a freezer bag is also an acquired skill that takes some time to master. It is an important skill that is useful in many other instances where specimen retrieval is needed (e.g., in procedures involving the gallbladder, the appendix, the adrenal glands, or the colon). In addition, it is highly cost-effective, in that these commercially available bags cost only a few cents each. Admittedly, laparoscopic retrieval bags are easier to use, but their substantially higher cost can become a factor in a busy minimally invasive surgery unit. We use a powerful suction machine (-70 mm Hg) and a custom-made sharp beveled 10 mm cannula to suction splenic tissue from the plastic retrieval bag.


Anterior Approach

The anterior approach is seldom used nowadays; however, it remains the preferable approach in some patients with massive splenomegaly (21 to 30 cm long) and all patients with mega spleens (> 30 cm or > 3 kg) with the aid of hand-assisted devices [see Special Considerations, Hand-Assisted Laparoscopic Splenectomy, below]. Very large spleens are extremely heavy and difficult to manipulate with laparoscopic instruments, and it is complicated to lift them so as to gain access to the phrenicocolic ligament posteriorly. The anterior approach can also be considered if another procedure (e.g., cholecystectomy) is being contemplated; alternatively, in this situation, the lateral approach can be used, and the patient can be repositioned for the secondary procedure.

Step 1: placement of trocars

Figure 13. Anterior approach: standard trocar placement

Under general anesthesia, the patient is placed in a modified lithotomy position to allow the surgeon to operate between the patient’s legs and to allow the assistants to stand on each side of the patient. The procedure is performed through five trocars in the upper abdomen [see Figure 13], with the patient in a steep Fowler position with left-side elevation. A 12 mm trocar is introduced through an umbilical incision, and a 10 mm laparoscope (0° or 30°) is connected to a video system. A 12 mm trocar is placed in each upper quadrant, and two 5 mm trocars are inserted close to the rib margin on the left and right sides of the abdomen. Alternatively, trocars can be deployed in a semicircle away from the left upper quadrant. Trocar sites are carefully selected to optimize working angles. The 12 mm ports are used to allow introduction of clip appliers, staplers, or the laparoscope from a variety of angles as needed.

Troubleshooting With increasing experience, we find that we prefer to do as many laparoscopic splenectomies as possible via the lateral approach because it is so much easier, even with spleens that are longer than 20 cm and are readily palpable. The decision is arbitrarily made on the basis of estimated available working space. If the spleen comes too close to the iliac crest or the midline, the anterior approach should be taken instead.

Step 2: isolation of lower pole and control of blood supplyThe left hepatic lobe is retracted, and the stomach is retracted medially to expose the spleen. Accessory spleens are searched for, and the phrenicocolic ligament, the splenocolic ligament, and the sustentaculum lienis are incised near the lower pole with an electrocautery and a hook probe or with scissors. Vascular adhesions—frequently found on the medial side of the spleen—are cauterized. The gastrocolic ligament is carefully dissected close to the spleen, and the left gastroepiploic vessels are ligated one by one with metallic clips or, if small, simply cauterized. The upper and lower poles of the spleen are gently lifted with one or both palpators (placed through the 5 mm ports) to expose the splenic hilum and the tail of the pancreas within the splenorenal ligament, thereby facilitating individual dissection and clipping of all the branches of the splenic artery and vein close to the spleen. The short gastric vessels are then identified and ligated with clips or, occasionally, with staples. No sutures are used. Alternatively, the splenic artery itself can be isolated and clipped within the lesser sac before extensive dissection of the lower pole and suspensory ligaments.

Because of the segmental and terminal distribution of splenic arteries, it is easy to determine the devascularized portions of the spleen: these segments exhibit a characteristic grayish color, whereas the vascularized segments retain a pinkish hue. When the organ is completely isolated, it is left in its natural cavity, and hemostasis is verified.


Troubleshooting If one elects first to clip the splenic artery within the lesser sac, there are a few precautions that must be taken. First, the clipping must be done distal to the pancreatica magna to prevent pancreatic injury. Second, one must make sure that the splenic artery proper is clipped, not one of its branches (e.g., the superior terminal branch). This is an easy mistake to commit with the distributed type of splenic vasculature [see Figure 1] because the splenic artery itself is short and the branches can take off very early. Third, one must always keep in mind the possibility of an anastomotic branch between the major splenic branches, as described by Testut.11 Should a major terminal branch be clipped rather than the splenic artery proper, there will be no spleen ischemia if such an anastomosis is present [see Figure 1].

Yet another challenge posed by the anterior approach is that if bleeding occurs, the blood tends to pool in the area of the hilum and obscure vision even more, whereas in the lateral approach, the blood tends to flow away from the operative field. One quickly learns that there is a steep price to pay for cutting corners during the dissection. The dissection must be meticulous, especially behind branches of the splenic vein.


Step 3: extraction of spleen Given that the anterior approach is now used only in cases of massive splenomegaly or megaspleen, bagging can be problematic. The largest commercially available freezer bag we have seen measures 27 by 28 cm, and the largest spleen we have been able to bag in one of them was 24 cm long. Furthermore, an accessory extraction incision is often required; a Pfannenstiel incision gives better cosmetic results, but a left lower quadrant incision can also be used. Hand-assisted devices are used with increasing frequency in laparoscopic removal of large spleens [see Special Considerations, Hand-Assisted Laparoscopic Splenectomy, below].

If the spleen cannot be bagged, it may be fragmented in the pelvis before extraction, provided that the abdomen is copiously washed and cleaned of any residual spleen fragments before closure to prevent splenosis. Most patients with large spleens have hematologic malignancies; thus, residual splenic activity is not as crucial an issue in these patients as it would be in others.


Laparoscopic Partial Splenectomy

Concern regarding the risk of OPSI has encouraged the practice of preserving splenic tissue and function whenever possible. For this reason, partial splenectomy has occasionally been indicated for treatment of benign tumors of the spleen and for excision of cystic lesions.21 Its use has been described in connection with the management of type I Gaucher disease, cholesteryl ester storage disease, chronic myelogenous leukemia, and thalassemia major, as well as with the staging of Hodgkin disease.22,23 Partial splenectomy has also been an option in the management of splenic trauma when the patient’s condition is stable enough to permit the meticulous dissection required for the operation.24,25

Like standard laparoscopic splenectomy, laparoscopic partial splenectomy is performed with the patient in the right lateral decubitus position. Trocar placement is similar as well. The splenocolic ligament and the lower part of the phrenicocolic ligament are incised to permit mobilization of the lower pole of the spleen. If the lower portion of the spleen is to be excised, branches of the gastroepiploic vessels supplying the lower pole are dissected and clipped close to the parenchyma. An appropriate number of penultimate branches of the inferior polar artery are then taken in such a way as to create a clear line of demarcation between normal spleen and devascularized spleen. This process is continued until the desired number of splenic segments are devascularized.

Figure 14a. Partial splenectomy: cauterization of splenic capsule (1)
Figure 14b. Partial splenectomy: cauterization of splenic capsule (2)
Figure 14c. Partial splenectomy: fracture of splenic pulp
Figure 14d. Partial splenectomy: spleen surface after transection

Next, a standard monopolar electrocautery is used to score the splenic capsule circumferentially, with care taken to ensure that a 5 mm rim of devascularized splenic tissue remains in situ; this is the most important technical point for this procedure [see Figures 14a, 14b, 14c, and 14d]. The incision is then carried into the splenic pulp. Atraumatic intestinal graspers are also used to fracture the splenic pulp in a bloodless fashion. The laparoscopic L hook and scissors provide excellent hemostatic control.

Once the spleen has been allowed to demarcate, resection is remarkably bloodless, provided that the 5 mm rim of ischemic tissue is left in place. Complete control of the splenic artery is not required before splenic separation, because division occurs in an ischemic segment of spleen.9 The feasibility of leaving portions of ischemic spleen in situ has been demonstrated in a large prospective, randomized trial involving partial splenic embolization as primary treatment of hematologic disorders.26

If the superior pole is to be removed, the phrenicocolic ligament must be incised almost entirely so that the spleen can be easily mobilized and the proper exposure achieved. The short gastric branches are taken first, along with the desired number of superior polar artery branches.

Laparoscopic partial splenectomy can be performed either with or without the aid of selective preoperative arterial embolization (see below). Radiologists are capable of cannulating the desired segmental splenic arterial branch and embolizing the segment that is to be resected. We have removed the superior pole in a patient with a class IV isolated splenic injury sustained while skiing8; laparoscopic partial splenectomy was made possible largely by the accuracy of selective arterial embolization, which permitted control of the bleeding and allowed laparoscopy to be performed in unhurried conditions.27

Preoperative Splenic Artery Embolization

Preoperative splenic artery embolization is used as an adjuvant in a few patients to make laparoscopic splenectomy possible and to reduce blood loss. Although it is now infrequently used, it remains a useful tool in the armamentarium of spleen surgeons.

Figure 15. Splenic artery embolization (before and after)

Generally speaking, the technique involves embolization of the spleen with coils placed proximally in the splenic artery and absorbable gelatin sponges and small coils placed distally in each splenic arterial branch (the double embolization technique), with care taken to spare vessels supplying the tail of the pancreas [see Figure 15].

The procedure is ended when it is estimated radiologically that 80% or more of the splenic tissue has been successfully embolized. In most cases, successful embolization is achieved with both proximal and distal emboli; in a minority of cases, it is achieved with proximal emboli alone or with distal emboli alone.28

Troubleshooting Preoperative splenic artery embolization is safe, provided that two main principles are adhered to. First, embolization must be done distal to the pancreatica magna to avoid damaging the pancreas. Second, neither microspheres nor absorbable gelatin powder should be used, because particles of this small size may migrate to unintended target organ capillaries and cause tissue necrosis; only coils and absorbable gelatin sponge fragments should be used.

Postoperative Care

Postoperative care for patients who have undergone laparoscopic splenectomy is usually simple. The nasogastric tube inserted after induction of general anesthesia is removed either in the recovery room, once stomach emptying has been verified, or the next morning, depending on the duration and the degree of technical difficulty of the procedure. The urinary catheter is usually removed before the patient leaves the recovery room. The patient is allowed to drink clear fluids on the morning after the operation; when clear fluids are well tolerated, the patient is allowed to proceed to a diet of his or her choice.

If the patient has no history of ulcer or dyspepsia, one naproxen sodium tablet (500 mg) is given with sips of water on the morning before operation. Meperidine injections (1 mg/kg) are administered during the first night, followed by oral acetaminophen (1 g every 6 hours). If pain is not well controlled, coanalgesia with an NSAID is added; this combination produces the best results. Because of its side effects (i.e., nausea, vomiting, abdominal fullness, and constipation), codeine is currently avoided if at all possible. When naproxen sodium is used, prophylactic doses of subcutaneous heparin are avoided on empirical grounds, especially if the platelet count is low or platelet function is abnormal.

Patients receiving I.V. cortisone are given oral steroids on postoperative day 1 after an overlap I.V. injection; thereafter, steroids are gradually tapered. Patients are allowed to shower 12 hours after surgery and are advised to keep the paper strips covering the trocar incisions in place for 7 to 10 days. No drains are used. No limitations are imposed on physical activity, and patients are allowed to tailor their activities to their degree of asthenia or discomfort.


Postoperative complications directly related to splenectomy include intraoperative and postoperative hemorrhage; left lower lobe atelectasis and pneumonia; left pleural effusion; subphrenic collection; iatrogenic pancreatic, gastric, and colonic injury; and venous thrombosis.24–31

Successful laparoscopic splenectomy depends to a large extent on proper preparation. Recognition of anatomic elements and their arrangement is paramount. As with other laparoscopic procedures, the keys are avoiding complications and minimizing technical misadventures. Vascular structures should be cleanly isolated and dissected from surrounding fat; they then can usually be controlled with two clips proximally and distally. Staplers should be used with care and should not be applied blindly. The stapler tip should be clearly seen to be free of tissue before it is closed; otherwise, hemorrhage from partial section of a major splenic branch might occur after the instrument is released. Blind application of the stapler may also result in damage to the tail of the pancreas, which often lies in close proximity to the inner surface of the spleen. If both clips and a linear stapler are used, it is vital to prevent interposition of clips in the staple line, which will cause the stapler to misfire and possibly to jam.

Improper use of the electrocautery during the procedure can cause iatrogenic injury to the stomach, the colon, or the pancreas. In a smoke-filled environment, where controlling vessels is difficult and time consuming, blind fulguration of fat in the hilum can lead to bleeding. Structures close to the lower pole in the gastrocolic ligament can be approached more aggressively, but not those in the hilum. To prevent arcing and spot necrosis, which may result in delayed perforation and sepsis, the instrument should be activated only in proximity to the target organ.

The assistants also play an important role in preventing complications. All instruments, including those handled by assistants, should be moved under direct vision. Especially in the anterior approach, retraction of the liver and stomach and elevation of the spleen require constant concentration if lacerations and subsequent hemorrhage or perforation are to be avoided.

Special Considerations

Extraction of Specimens

Spleens removed via the anterior approach are extracted through the umbilical trocar site after finger fragmentation in a plastic bag. It is rarely necessary to enlarge the umbilical incision to more than 2 or 3 cm. When the lateral approach is used, extraction is more easily performed through one of the ports situated anteriorly. This extraction site also requires little or no enlargement. On occasion,for a spleen longer than 20 cm, a 7.5 to 10 cm Pfannenstiel incision is made, and the operator’s forearm is introduced into the abdomen to deliver the spleen into the pelvis for extraction in large fragments under direct vision.32 The abdomen is copiously irrigated before closure.

Special mention should be made of extraction of the splenic specimen from patients with malignant disease. If lymphoma or Hodgkin disease is suspected, neither preoperative splenic artery embolization nor finger fragmentation in a plastic bag should be performed, for fear of making the histologic diagnosis difficult. Extraction of intact spleens through a small left subcostal or median incision has also been employed when preservation of tissue architecture is required. Alternatively, a port site may be slightly enlarged, and a knife or a Mayo scissors may be used to furnish the pathologist with intact specimen pieces of various sizes. The various techniques of fragmentation and extraction of splenic tissue during laparoscopic splenectomy should be discussed and agreed on with the pathologist ahead of time to ensure that proper pathologic diagnoses are not compromised by either necrotic tissue (in the case of preoperative splenic artery embolization) or altered tissue architecture (in the case of finger fragmentation), especially if malignancy is suspected but not proved. In practice, however, we have found that the diagnosis is made preoperatively in more than 90% of patients with benign and malignant hematologic disease; hence, the issue rarely arises.

Hand-Assisted Laparoscopic Splenectomy

Figure 16. Laparoscopic splenectomy: hand-assisted

The term hand-assisted laparoscopic surgery refers to laparoscopic procedures performed with the aid of a plastic device inserted in a 7.5 to 10 cm wound. This plastic hand port consists of a sealed cuff that enables a hand to be inserted into and withdrawn from the abdomen without loss of pneumoperitoneum during the operation; in this way, the surgeon regains some of the tactile feedback lost in conventional laparoscopic surgery [see Figure 16]. A number of different models have been developed, some of them quite expensive. Most use either an inflatable sleeve clipped to an O-ring, a spiral inflatable valve, or a flap valve to maintain pneumoperitoneum.

The optimal placement of the incision for a hand-assisted laparoscopic splenectomy remains a subject of debate: recommended locations have included the upper midline, the right upper quadrant, the left iliac fossa, and, for very large spleens, the Pfannenstiel position. Whether the surgeon is left-handed or right-handed plays a role; most surgeons agree that the nondominant hand should be used in the device.

There are obvious advantages and drawbacks to hand-assisted laparoscopic splenectomy. The most apparent disadvantage is the cosmetic cost of a longer abdominal incision (except when a Pfannenstiel incision is employed). More generally, the use of a longer incision would seem to be at odds with the current trend toward developing surgical techniques that reduce surgical trauma as much as possible. Nevertheless, comparative studies of splenectomy in patients with large spleens (> 700 g) seem to indicate that for the most part, the hand-assisted approach yields outcomes similar to those of conventional laparoscopic splenectomy.33

Although the precise role of hand-assisted laparoscopic splenectomy remains to be defined, it is likely that this technique will find a place in the surgical management of patients with large spleens. In addition, the hand-assisted approach may be a valuable aid for surgeons who have not yet completed the learning curve for conventional laparoscopic splenectomy. Finally, this technique may render preoperative splenic embolization unnecessary for most very large spleens.

Outcome Evaluation

No randomized, prospective trials comparing open splenectomy with laparoscopic splenectomy have yet been conducted. At present, such trials are unlikely to be held, for a variety of reasons. For one thing, randomization is difficult with procedures that are still in evolution. At one end of the spectrum, laparoscopic splenectomy is done for patients with ITP, who usually are relatively healthy and have normal-size spleens. In many of these patients, needlescopic instruments (< 3 mm) can be used in conjunction with a single 12 mm port site in the umbilicus. This approach permits hospital discharge within 24 hours of operation in a significant number of cases. At the other end of the spectrum, laparoscopic splenectomy is done for patients with myeloid metaplasia and spleens longer than 30 cm. In this setting, a laparoscopic approach poses formidable challenges, and the optimal technique and its justification remain to be determined. The window of opportunity for randomized comparative trials may have been lost.

Large case series and nonrandomized comparative trials, however, have consistently reported better outcomes from laparoscopic splenectomy than from open splenectomy.34–41 For example, in one set of 528 patients [see Table 2],36–39 the rate of postoperative pneumonia was 1.1% (6/528), and no subphrenic abscesses occurred as postoperative complications. Many surgeons who have completed the learning curve associated with the procedure feel that there is still room for improvement regarding complication rates and length of stay for patients with ITP and other relatively benign conditions necessitating laparoscopic splenectomy. The more serious conditions and the mortality seen in conjunction with the procedure tend to occur in patients with advanced hematologic malignancies or megaspleens. In such cases, most of the adverse results are related to the disease state rather than to the operation, and it remains to be seen whether laparoscopic splenectomy will have a positive effect on outcome.

One of the great attractions of minimally invasive surgery has been the prospect of significant cost reductions. At this point in the development of laparoscopic splenectomy, however, we are reluctant to place too much trust in premature cost analyses that do not take into account the ‘work in progress’ nature of minimally invasive surgery. Most surgeons can now perform most laparoscopic splenectomies with simplified trays of reusable in struments. Our basic laparoscopic tray contains a few instruments and two sizes of reusable clip appliers with inexpensive clips. As noted [see Operative Technique, above], clips are used for distributed-type spleens, and single-use linear staplers are mostly used for magistral-type spleens. To reduce costs, ultrasonic dissectors are rarely used. In addition, the use of commercially available freezer bags instead of laparoscopic retrieval bags further reduces the cost of specimen extraction. Finally, even if intraoperative costs are higher with laparoscopic splenectomy, our experience is that the increase is offset by reductions in postoperative stay.

We, like most authorities, believe that as a surgeon gains experience with laparoscopic splenectomy, operating time tends to fall until it approaches that of open splenectomy. We also concur with the numerous authors who have suggested that once laparoscopic splenectomy is mastered, use of blood products tends to decrease substantially.

Open Splenectomy

Preoperative Evaluation

With the growing acceptance of laparoscopic splenectomy, the indications for open splenectomy have essentially been reduced to (1) elective removal of megaspleens and (2) treatment of splenic trauma when conservative treatment either is not indicated or has failed. In rare cases, open splenectomy may be done for iatrogenic injuries incurred during left upper quadrant surgical procedures.

Preoperative evaluation for elective open splenectomy is similar to that for laparoscopic splenectomy [see Laparoscopic Splenectomy, Preoperative Evaluation, above]. Preoperative evaluation of trauma patients is covered in more detail elsewhere [see 7:1 Initial Management of Life-Threatening Trauma]. Essentially, a coagulogram and blood typing and crossmatching are required. A preoperative CT scan will have established the size of the spleen, the grade of the splenic injury, the presence of other injuries (if any), and, in elective cases, the location and configuration of any masses or cysts.

Operative Planning

Most surgeons would agree that the lessons learned from successful performance of minimally invasive procedures have had a positive impact on the refinement of the corresponding open procedures. The principles of careful appreciation of fine anatomic details (as described for laparoscopic splenectomy) and maximal reduction of tissue trauma from retractors or excessive tissue handling should be incorporated into the planning of open splenectomy.

Total versus Partial Splenectomy

As a consequence of the recognition that splenectomy renders patients susceptible to a lifelong risk of OPSI, it is now routine practice to attempt splenic conservation. Accordingly, saving normally functioning splenic parenchyma has become the most important goal in the management of splenic injuries. In some 50% of adults (and over 80% of children), this goal can be achieved by means of nonoperative treatment. In approximately 20% of adults, splenorrhaphy and partial splenectomy are possible; splenectomy is indicated in the remainder. Partial splenectomy is also favored on occasion when excision of splenic tissue is required for the treatment of other elective conditions.

For the sake of brevity, we describe the surgical technique for total splenectomy and partial splenectomy concurrently, noting differences only where significant.

Operative Technique

Step 1: Incision

Figure 17. Open splenectomy: incisions

The patient is supine, in a reverse Tredelenburg position with a 15° tilt to the right. For maximal exposure, a midline incision is made, starting on the left side of the xiphoid process [see Figure 17]. The incision is extended below the umbilicus for a variable distance, depending on circumstances such as the size of the patient, the surgical situation (traumatic versus nontraumatic), the possibility of associated injury, and the size of the spleen. Occasionally, a left subcostal incision may be used for nontraumatic indications in patients with normal-size spleens. This incision may be extended onto the right side to form a chevron incision if necessary; however, this may impede the search for accessory spleens. Some surgeons have performed splenectomy via a thoracoabdominal approach, but most have abandoned this approach. Appropriate retraction of the left lobe of the liver and the abdominal wall is achieved with the help of surgical assistants placed on each side of the table or the use of self-retaining retractors.

Troubleshooting In trauma cases, the anesthetist should always be informed when the peritoneum is opened; release of a tense hemoperitoneum can precipitate hypotension with the loss of tamponade.

Step 2: Evacuation of Blood and Packing of the Abdomen

In trauma cases, gross blood and clots are evacuated manually with large laparotomy sponges. All quadrants of the abdomen are then packed with laparotomy pads. Standard suction equipment is not very useful for evacuating large quantities of blood from the abdomen.

Step 3: Control of Splenic Artery

Once other major injuries are excluded, the first decision to be made is whether to control the splenic artery first or to mobilize the spleen to the midline. This decision is dictated by the urgency of the clinical situation, the spleen size, and the presence of underlying disease.

Figure 18. Open splenectomy: control of splenic artery

If the decision is made to control the splenic artery first, the main splenic trunk is identified above the pancreas via an approach that leads to the lesser sac either through the gastrocolic ligament or through the avascular plane of the greater omentum above the distal transverse colon. Once dissected, the artery is controlled with a vascular loop. The main artery can also be accessed and dissected posteriorly after the spleen is mobilized [see Figure 18].

Troubleshooting One advantage of dissecting the splenic artery in the lesser sac (as opposed to the hilum) is that the splenic vein is rarely damaged, being located under the pancreas and away from the artery. Good proximal control of the splenic blood supply facilitates the performance of the more complex variations of partial splenectomy or total splenectomy for megaspleens.

Step 4: Mobilization of Spleen

If the decision is made to mobilize the spleen first, as in most trauma cases, mobilization should be carried out in a carefully planned manner; it is all too easy to compound splenic injury with ill-advised maneuvers that obligate the surgeon to perform a total splenectomy.

Figure 19. Open splenectomy: incision of phrenicocolic ligament

Gastric decompression is ensured with a properly placed nasogastric tube. The spleen is then retracted anteromedially with the left hand, with care taken to confirm proper retraction of the left abdominal wall. The phrenicocolic ligament is thereby placed on stretch, and the ligament insertion on the lateral abdominal wall serves as countertraction. The phrenicocolic ligament is then incised from the bottom up with either long scissors or the 45°-angle tip of a monopolar cautery [see Figure 19]. Efforts should be made to leave 2 cm of ligament on the spleen side and to avoid capsular injury. If the surgeon cannot put a finger behind the ligament, an assistant should elevate the ligament between the jaws of a right-angle clamp. The incision of the phrenicocolic ligament is then extended to the left crus of the diaphragm. Except in patients with portal hypertension or myeloproliferative disorders, this ligament is avascular. The left lateral portion of the gastrocolic ligament (the greater omentum) is also dissected away from the splenic flexure of the colon to facilitate mobilization of the spleen. At this point, the splenocolic ligament and the sustentaculum lienis are left alone.

Figure 20 Open splenectomy: dissection of areolar plane

After complete division of the phrenicocolic ligament, a plane is developed between the pancreas and the retroperitoneal structures with gentle blunt finger dissection. The spleen can then be delivered to the midline, where the splenectomy can be planned in an unhurried manner [see Figure 20]. Continuing splenic bleeding during this maneuver can be controlled with manual compression of the organ. The splenic pedicle may also be gently compressed between the thumb and the index finger at this stage. Laparotomy sponges are placed in the left subphrenic space.

Troubleshooting When performing elective resections of very large spleens, experienced spleen surgeons use a few tricks to simplify the procedure. In most patients with megaspleens, the suspensory ligaments have been stretched over time, allowing the surgeon much more leeway in mobilizing or turning the spleen. This greater leeway allows the surgeon to rotate the spleen from the lower pole so as to deliver it transversely into the incision. Thus, the presence of a large spleen does not always necessitate the creation of a long xiphopubic incision, because a transversely placed spleen can be extracted into the abdominal wall through a shorter incision.

Step 5a (Total Splenectomy): Planning of Resection

To devise the appropriate operative strategy, the surgeon performing open total splenectomy must address the same anatomic issues that he or she would if performing laparoscopic total splenectomy—for example, the nature of the splenic blood supply (distributed or bundled) and the distance between the tip of the pancreas and the splenic hilum. The anatomy must be appreciated before the operative strategy can be defined.

Once the spleen has been delivered into the abdominal wall, various techniques may be employed to control the blood supply. The classic approach is to serially clamp, ligate, or suture-ligate the vessels between curved clamps, starting from the lower pole. Alternatively, the vessels may be controlled with clips.

Troubleshooting We frequently use laparoscopic clip appliers to achieve vascular control in open splenectomy. The long, slender design of these devices is particularly useful in obese patients, in whom it is often difficult to achieve complete mobilization of the spleen without causing additional splenic trauma. With laparoscopic clip appliers, vessels can be safely controlled inside the abdomen. Locking plastic clips may also be used. Moreover, provided that the same precautions are taken as in laparoscopic splenectomy, a linear stapler with a vascular cartridge may be used to control the hilum in one step once the gastroepiploic branches have been controlled.

Step 5b (Partial Splenectomy): Planning of Resection

Once the spleen is appropriately placed for full evaluation and adequate hemostasis is ensured, planning for partial splenectomy can start. In trauma cases, such planning is guided by the extent of the injury, and in elective cases, it is guided by the nature of the underlying pathologic condition [see Table 3].

In most cases, as noted, the spleen can be divided into independent lobes or segments, each with its own terminal blood supply. The superior pole is supplied by the short gastric vessels, and the lower pole is supplied by branches of the gastroepiploic artery, which are known to form anastomoses with the inferior polar artery. In addition, most patients, possible variations notwithstanding, have two or three major vessels entering the hilum. Thus, there are usually four or five discrete regions or lobes that may be removed, individually or in combination, in a partial splenectomy. It should be kept in mind that the vessels supplying the spleen lie in different supportive ligaments. The vessels supplying the superior pole (the short gastrics) and the inferior pole (the gastroepiploic branches) rest in the gastrosplenic ligament, whereas the splenic branches proper lie in the splenorenal ligament along with the tail of the pancreas.

Step 6 (Partial Splenectomy): Exposure of Entire Hilum and Ligation of Appropriate Arteries

The entire hilum of the spleen is then exposed close to the parenchyma. The gastrosplenic ligament and the splenorenal ligament must be separated, with care taken to preserve the blood supply to both poles of the spleen. There is a fairly avascular area of the gastrosplenic ligament, between the short gastric vessels supplying the superior pole and the gastroepiploic branches supplying the lower pole, that must be opened; once this is done, a complete view of the entire splenic blood supply is available. The surgeon can then determine whether partial splenectomy is feasible and how many lobes he or she can resect while still leaving enough spleen tissue behind for adequate splenic function. Any number of segmental resections are possible. If the surgeon is unsure of the extent of the necessary resection, an accurate assessment can be made by temporarily compressing the splenic arterial branches.

Selected arterial branches are then carefully dissected as close to the spleen parenchyma as possible, with the understanding that the veins are situated posteriorly in close proximity. The vessels may be doubly ligated, transfixed, or clipped. If clips are used, care is taken not to dislodge them with inappropriate manipulations. Once the arterial blood supply is controlled, the affected area of the spleen will rapidly become visibly demarcated. If the devitalized area of the spleen corresponds to the intended resection, a similar technique is applied to the venous side. Access to the venous side can also be achieved from the posterior aspect of the spleen. When this approach is followed, it is helpful to identify the tail of the pancreas if possible to avoid inadvertent damage: the tail of the pancreas touches the hilum of the spleen in 30% of cases and lies within 1 cm of the hilum in 70%.

Step 7 (Partial Splenectomy): Incision of Splenic Capsule and Partial Resection of Spleen

The capsule of the spleen is incised circumferentially with a scalpel or a monopolar cautery, and a 5 mm rim of devitalized tissue is left in situ. The splenic fragments may be transected with a scalpel, scissors, a monopolar cautery, or a combination thereof. Various techniques have been used to control residual bleeding, including use of a monopolar cautery on spray current; use of a cutaneous ultrasonic surgical aspirator; use of an argon beam coagulator; suture compression, with or without Teflon pledgets; and omental pedicle packing. One low-tech way of dealing with residual hemostatic requirements is to employ the hollow part of a Poole suction device to aspirate blood while employing a coagulating monopolar current on the suction tip. In some cases, wrapping the splenic remnant in an absorbable polyglycolic mesh is useful. Our experience suggests that when enough residual devitalized tissue (i.e., at least 5 mm) is left behind circumferentially, good hemostasis is easily achieved, typically requiring nothing more than simple measures and topical agents. No drains are used unless the tail of the pancreas has been damaged, in which case a closed-suction drain is placed.

Postoperative Care

The principles of postoperative care are essentially the same for open splenectomy as for laparoscopic splenectomy [see Laparoscopic Splenectomy, Postoperative Care, above], though most authors agree that the pace of aftercare is slower with the former. It should be kept in mind that acute postoperative gastric distention occurs more frequently in children and may necessitate more prolonged gastric decompression.


The complications seen after open splenectomy are the same as those seen after its laparoscopic counterpart [see Laparoscopic Splenectomy, Complications, above]. Hemorrhagic complications may necessitate transfusion, reoperation, or both.

Although the rate of serious postoperative infection after splenic surgery is generally considered to be 8%, it is thought to be lower in patients undergoing splenorrhaphy or partial splenectomy. The lower rate is probably attributable to the presence of less severe underlying injuries, rather than to the preservation of splenic tissue. Infectious complications usually manifest themselves between postoperative days 5 and 10 and are typically diagnosed by means of physical examination, chest x-ray, ultrasonography, and CT.