The American Journal of Surgery 188 (2004) 796 – 800
Scientific paper
Internal hernias after laparoscopic Roux-en-Y gastric bypass Ernesto Garza, Jr., M.D., Joseph Kuhn, M.D., David Arnold, M.D., William Nicholson, M.D., Suraj Reddy, M.D., Todd McCarty, M.D.* Department of Surgery, Baylor University Medical Center, 3409 Worth Street, Suite 420, Dallas, TX 75246, USA Manuscript received July 20, 2004; revised manuscript August 7, 2004 Presented at the 56th Annual Meeting of the Southwestern Surgical Congress, Monterey, California, April 18 –21, 2004
Abstract Background: Laparoscopic gastric bypass (Lap-RYGB) is an increasingly common procedure performed for severe obesity. Internal hernias are a potential problem associated with Lap-RYGB, and little is known about the clinical presentation and the diagnostic accuracy of this potentially serious complication. Methods: A retrospective review of 1,000 retrocolic Lap-RYGB was performed to identify those who developed postoperative internal hernias. Clinical symptoms, radiologic characteristics, and operative outcomes were analyzed to determine clinical and radiologic diagnostic accuracy (including computed tomography [CT] scan and upper gastrointestinal imaging). Subsequent independent review was performed to match operative intervention with radiologic imaging and interpretation. Operative outcomes, including the hernia closure technique, hospital length of stay, and mortality were obtained. Results: Of 1,000 Lap-RYGB procedures, 45 internal hernias were identified (4.5%) in 43 patients. Hernia location included transverse colon mesentery (n ⫽ 43, 95%) or Petersen’s defect (n ⫽ 2, 5%). The most common clinical symptoms included intermittent, postprandial abdominal pain, and/or nausea vomiting (86%), although 20% had no abdominal tenderness. Initial radiologic imaging studies were diagnostic in 64%, although subsequent review of all imaging studies showed diagnostic abnormalities in 97%. CT findings suggestive of internal hernia include small bowel loops in the left upper quadrant and evidence of small bowel mesentery traversing the transverse colon mesentery. All patients with internal hernias underwent operative repair (98% performed laparoscopic). One patient had a negative laparoscopy, although the preoperative CT suggested an internal hernia was present. The mean time to intervention for an internal hernia repair was 225 days (range 2 to 490), whereas hospital length of stay was 1.2 days (range 1 to 4). No deaths were noted. Conclusions: Internal hernias after retrocolic lap-RYGB are associated with vague abdominal complaints and limited radiologic imaging results. A high index of clinical suspicion should be used in this patient population, and surgeon review of radiology imaging studies should be performed. Prompt surgical intervention is successful and can commonly be performed laparoscopically. © 2004 Excerpta Medica Inc. All rights reserved. Keywords: Internal hernia; Obesity; Laparoscopic gastric bypass; Retrocolic
Roux-en-Y gastric bypass (RYGB) is the most common surgical procedure performed in the United States for severe obesity. Laparoscopic RYGB has been increasing in popularity with favorable outcomes comparable to its open counterpart. Complications are similar between the two approaches, including small bowel obstruction, gastric outlet stenosis, anastomotic leaks, and ulcers [1]. However, the source of small bowel obstruction after lap-RYGB is more likely to be from an internal hernia rather than adhesions.
Potential internal locations include a transverse mesocolon defect; Petersen’s space, the area between the mesentery of the Roux-limb and the transverse mesocolon; and a jejunojejunostomy mesenteric defect. This article reviews the incidence of internal hernias in 1,000 lap-RYGB procedures with the presentation and diagnosis of each.
* Corresponding author. Tel.: ⫹1-214-824-7167; fax: ⫹1-214-8247167. E-mail address:
[email protected]
An institutional review board–approved retrospective analysis was performed on 1,000 patients who had a lapRYGB at a large teaching institution. Patients reviewed
Materials and Methods
0002-9610/04/$ – see front matter © 2004 Excerpta Medica Inc. All rights reserved. doi:10.1016/j.amjsurg.2004.08.049
E. Garza, Jr. et al. / The American Journal of Surgery 188 (2004) 796 – 800 Table 2 Patient symptoms
Table 1 Patient demographics Average age Average BMI Sex
797
21 (range 21–57) 49.1 (range 41–100) Male 3 (5%) female 40 (95%)
were selected based on either preoperative diagnosis of internal hernias or postoperatively. The patient’s age, sex, and body mass index were all noted (Table 1). The date of their initial surgery as well as the interval to operative repair was recorded. Initial symptoms, duration of symptoms, and physical examination findings were all noted if available. Most patients underwent extensive radiologic workup before diagnosis of their internal hernias. Upper gastrointestinal (GI) imaging, a computed tomography (CT) scan of the abdomen and pelvis, or both was performed on most patients before the diagnosis of an internal hernia. Whether either of these tests was helpful at eliciting a diagnosis was also reviewed. All operative reports were reviewed to determine the location of each internal hernia. The repair of each hernia was reviewed including type of suture and type stitch used. Lastly, postoperative care was reviewed, noting the number of postoperative days required for discharge. This was based on, patient pain, tolerance of a liquid diet, and ambulation.
Results Of 1,000 lap-RYGB procedures, 45 (4.5%) internal hernias were identified in 43 patients. Hernia location included transverse colon mesentery (n ⫽ 43, 95%) or Petersen’s defect (n ⫽ 2, 5%). No jejunojejunostomy mesenteric defects were noted in any patient with an internal hernia (Fig. 1). Many patients presented with similar complaints. The most common clinical symptoms included intermittent, postprandial abdominal pain (88%), and/or nausea vomiting (65%) (Table 2). The duration of these symptoms varied from a range of 1 to 180 days (mean 16 days). Forty-seven percent of patients presented with diffuse abdominal tenderness, whereas 20% of patients presented with a benign abdominal examination.
Fig. 1. Internal hernia sites.
Number of internal hernias Nausea Emesis Abdominal pain Nausea, emesis, abdominal pain
43 27 (61%) 27 (61%) 36 (82%) 23 (52%)
Radiologic studies included either a CT scan of the abdomen/pelvis, an upper GI, or both. CT findings suggestive of internal hernia include small bowel loops in the left upper quadrant and evidence of small bowel mesentery traversing the transverse colon mesentery. In review, 86% of patients had a CT scan done before surgery, 10% had an upper GI, 7% had both studies done before surgery, and 14% did not have either study done. Initial radiologic imaging studies were diagnostic in 60% when taking into account either study done before surgery. When CT was used alone, 64% (22/34) were positive for an internal hernia. Of the four patients with upper GIs done, three patients had prior CT scans that were read initially as negative. Subsequent review of all imaging studies showed diagnostic abnormalities in 97% of the patients. All patients with internal hernias underwent operative repair (98% performed laparoscopic). One patient had a negative laparoscopy, although the preoperative CT suggested an internal hernia was present, and one patient was initially begun as laparoscopic but then converted to an open procedure for repair. The mean time to intervention for an internal hernia repair was 225 days (range 2 to 490) (Fig. 2). Most internal hernias were repaired in similar fashion. Thirtyseven of the 43 patients (86%) had internal hernias repaired in a running fashion with 2-0 Ethibond. The remaining 6 had their hernia defects closed with interrupted suture, with 5 patients having used 2-0 Ethibond for repair and one with 2-0 silk. The length of hospital stay ranged from 1 to 4 days with a mean of 1.2 days. No recurrences or deaths were noted from any repair of these internal hernias.
Fig. 2. Days to operative intervention.
798
E. Garza, Jr. et al. / The American Journal of Surgery 188 (2004) 796 – 800
Comments The incidence of internal hernias after lap-RYGB in this series is comparable to that of Higa et al [2]. Interestingly, both studies reported an internal hernia higher than that observed with open RYGB. One possible reason for this occurrence may be because of the decreased amount of adhesions formed with laparoscopy [3,4]. This increased mobility may allow the small bowel to enter a potential hernia defect more frequently. The patients in our study all had retrocolic placement of the Roux limbs through a transmesenteric defect, and the majority had a transmesenteric defect contributing to the internal hernia. This experience prompted modification of the surgical technique to close the potential internal hernia sites more thoroughly. Unfortunately, this did not eliminate the occurrence of internal hernias and did result in the stricture of the Roux limb as it passed through the mesenteric defect. Interestingly, if we could eliminate the transmesenteric defect altogether, with antecolic placement of the Roux limb, the incidence of internal hernias after lap-RYGBs may potentially be decreased [4]. The disadvantage to this approach is a longer segment of small bowel is needed to make the Roux limb, and, thus, more tension may be placed on the Roux limb. Our current technique uses an antecolic Roux limb when possible. Symptoms of internal herniation were very vague. Nausea, emesis, and postprandial abdominal pain were among the most frequent complaints. The location of the abdominal pain correlated with the side of internal herniation (data not shown). Specifically, left upper-quadrant pain occurred when the small bowel herniated through the defect to the patients left side, whereas right upper-quadrant pain was seen when the small bowel herniated to the right side of the defect. Overall, postprandial left upper quadrant pain was the most consistent symptom. In addition to varied complaints, the time of presentation also varied among all patients. The range of days from initial operation to operative repair of their internal hernia was 2 to 490 days, with a median of 262 days. This wide range of presentations continues to allow the differential diagnosis to expand. Because of this, physicians performing lap-RYGBs must be very suspicious of internal hernias postoperatively in all those patients complaining of abdominal pain, nausea, and/or vomiting, regardless of the timeframe after surgery. When diagnostic tests are added to the workup of these patients, there still is no reliable accurate radiologic test. At our institution, CT scans of the abdomen and pelvis were taken with both oral and intravenous contrast. Sixty-four percent of our patients had CT scans that were initially read as positive for an internal hernia. In comparison to a similar study, CT imaging of the abdomen showed limited diagnostic ability. Higa et al’s study [1] of 2,000 lap-RYGBs revealed a negative CT scan in 20% of those patients with internal hernias. It is likely that this inaccuracy is in part
caused by the inexperience of radiologists diagnosing internal hernias [5,6]. The CT scan findings suggestive of an internal hernia include small bowel loops in the left or right upper quadrant, evidence of small bowel mesentery traversing the transverse colon mesentery, and/or location of the jejunojejuostomy superior to the transverse colon. In addition, crowding, stretching, and engorgement of the main mesenteric trunk to the right and signs of small bowel obstruction may be seen [5]. Our ability to recognize and diagnose internal hernias’ improved throughout the study as reflected by decreased time from initial lap-RYGB to hernia repair (Fig. 2). Internal hernias after retrocolic lap-RYGB are associated with vague abdominal complaints and limited radiologic imaging results. A high index of clinical suspicion should be used in this patient population, and surgeon review of radiology imaging studies should be performed. As more surgeons and radiologist gain experience with lap-RYGB patients and their presentation, the diagnostic accuracy of internal hernias’ should significantly increase. Whenever an internal hernia is suspected, the surgeon must act promptly and surgical intervention is important. Success can be frequently attained with minimally invasive surgical techniques.
References [1] Higa KD, Boone KB, Ho T. Complications of the laparoscopic Rouxen-Y gastric bypass: 1,040 patients—what have we learned? Obesity Surgery 2000;10:509 –13. [2] Higa KD, Ho T, Boone KB. Internal hernias after laparoscopic Rouxen-Y gastric bypass: incidence, treatment and prevention. Obesity Surgery 2003;13:350 – 4. [3] Champion JK, Williams M. Small bowel obstruction and internal hernias after laparoscopic Roux-en-Y gastric bypass. Obesity Surgery 2003;13:596 – 600. [4] Blachar A, Federle MP. Gastrointestinal complications of laparoscopic roux-en-Y gastric bypass surgery in patients who are morbidly obese: findings on radiography and CT. Am J Roentgenol 2002;179:1437– 42. [5] Blachar A, Federle MP, Brancatelli G, Peterson MS, Oliver JH 3rd, Li W. Radiologist performance in the diagnosis of internal hernia by using specific CT findings with emphasis on transmesenteric hernia. Radiology 2001;221:422– 8. [6] Blachar A, Federle MP, Pealer KM, Ikramuddin S, Schauer PR. Gastrointestinal complications of laparoscopic Roux-en-Y gastric bypass surgery: clinical and imaging findings. Radiology 2002;223: 625–32.
Discussion Sherman Smith, M.D. (Salt Lake City, UT): Although Roux-en-Y gastric bypass has been performed with increasing frequency in the United States, some new problems with this kind of surgery have been noted with the adoption of
E. Garza, Jr. et al. / The American Journal of Surgery 188 (2004) 796 – 800
the laparoscopic approaches. Although the concept, anatomy, and physiology of the open and laparoscopic procedures closely mimic each other and although patient success with dramatic weight loss and reduction of comorbid conditions occur with similar frequency, the actual shift from the open approach to the laparoscopic approach requires some innovative technical adaptations. The reported incidence of internal hernia with open approaches is extremely low, usually less than 1%. Most of these hernias have involved the Peterson defect behind the alimentary limb and the space between the mesentery leaflets at the jejunojejunostomy. Herniations through the mesocolic window has been seen in less than 0.1% of cases. Because problems with whole gut volvulus around the alimentary limb were seen with increasing frequency in antecolic positioning, Drs Miller and Goodman made a shift to the retrocolic approach in 1985, just the opposite direction from what you had for a different reason. Since then, they have seen no more cases of that type. With closure of the mesocolic window around the alimentary limb of jejunum and closure of the mesenteric leaflets space, the jejunojejunostomy has become standard practice with laudable results in patients, and the surgeons performing bariatric surgery. As the laparoscopic procedures have become more prevalent, so too have the particular problems with internal herniation. Michelle Ganier now prefers the antecolic limb but will anchor the jejunal mesentery to the transverse mesocolon and the colon. Drs Higga and Boone close their retro colic windows carefully with running nonabsorbable suture. All surgeons performing these kind of procedures have learned that fastidious attention to detail and closure of the potential spaces is mandatory and to a much more attentive degree than seem to be necessary with open gastric bypass. Dr Garza and his colleagues have aptly described in retrospective fashion what has happened to a subgroup of potentially obstructed patient after laparoscopic Roux-en-Y gastric bypass. They have stressed the need for careful review of contrast studies in the workup of these patients because the clinical picture must not be ignored in the phase of what was thought to be a normal study. Diagnostic accuracy of radiographic studies was improved from 64% to 97% when the surgeon took the time to go over the films with the radiologists. Ninety-five percent of the 45 internals seen in a review of 1,000 patients were through the transverse mesocolic window and only one patient had no findings of laparoscopy, even though the study suggested an internal hernia. I have the following questions for the authors. Number one, how often do you reexplore patients in the face of clinical suspicion of internal hernia, even when the preoperative studies are normal? Number two, what has been the incidence of internal herniation or small bowel obstructions since you have shifted to the antecolic approach? Number three, have strictures at the mesenteric window and the retrocolic approach been a problem?
799
Answer Ernesto Garza, Jr., M.D. (Dallas, TX): First, in regard to reexploration of patients with a clinical suspicion of an internal hernia, a very low threshold is used. Despite normal CT scans of patients, we will still perform a diagnostic laparoscopy if the clinical symptoms are concerning for an internal hernia. Through the same camera incision from their initial operation, the abdomen is examined a second and possible third working port is inserted to evaluate the entire small bowel and to view all the potential hernia defects. If hernias are found, the repair involves reducing the hernia and tacking the small bowel to the defect. All other potential defects are closed if they had not already been closed. On average, these reexplorations are not very time consuming (an hour). This further adds to the low threshold of reexploring patients suspected of having an internal hernia. The number of hernias since shifting to the antecolic approach has decreased remarkably. All potential defects are closed in each antecolic laparoscopic Roux-en-Y gastric bypass. Strictures at the mesenteric window have been a problem when we closed this in a running fashion. Based on clinical suspicion, nausea, vomiting, and intolerability of liquids, all these patients get seen and evaluated.
Question Phil Schauer, M.D. (Pittsburgh, PA): Because most of the surgeons in the room here are not bariatric surgeons, it is important to understand this particular problem because these folks are going to show up at your hospital in the emergency room with unusual abdominal pain and you might see one or two air fluid levels. What you should do is take that patient to surgery. You should not wait and hope it will resolve. These patients develop a closed loop obstruction and that is a devastating problem. So, you’ll be seeing a lot of this down the road as bariatric surgery continues to increase across the country. We and others have moved toward antecolic approach and have close to eliminated the problem with the transverse mesocolon herniation, which is the most common reported in your series as well. Many surgeons are reluctant to go with antecolic because of fear with excess tension on the gastric jejunojejunostomy. Now it appears that your group has started an antecolic approach. Share with us your experience. Have you had difficulty with tension with the antecolic approach and have you seen any of these internal hernias occurring after the antecolic approach?
Answer Ernesto Garza, Jr., M.D. (Dallas, TX): As far as tension, a longer segment of roux-limb is used to decrease the
800
E. Garza, Jr. et al. / The American Journal of Surgery 188 (2004) 796 – 800
amount of tension on the Roux limb. Postoperatively, we have seen some internal hernias, particularly jejunojejunostomy mesenteric defects after the studies and switching to the antecolic approach, and once again these are addressed as well. The same workup is begun, history and physical and CT scan. With any clinical suspicion, the patient goes to the operating room for evaluation.
Question Ken Murayama, M.D. (Honolulu, HI): It looked to me on your data slides that the interrupted sutures had a much lower incidence of, or accounted for less of your 42 internal
hernias, and yet your recommendations at the end were to switch to a running suture. Did I miss something?
Answer Ernesto Garza, Jr., M.D. (Dallas, TX): I think it depends on surgeon’s preference initially which type of a repair is used, but the current practice of this group of surgeons is to use a running nonabsorbable stitch to close all the defects. I think initially interrupted stitches were used, but, even still, internal hernias defects were appearing. The switch was then made to a running suture, but internal hernias still occurred. Now, the practice performs the antecolic approach when feasible to avoid internal hernias.
