Does experience preclude leaks in laparoscopic gastric bypass?

Surg Endosc (2006) 20: 1687–1692 DOI: 10.1007/s00464-004-8253-y Ó Springer Science+Business Media, Inc. 2006

Does experience preclude leaks in laparoscopic gastric bypass? R. Gonzalez, K. Haines, S. F. Gallagher, M. M. Murr Interdisciplinary Obesity Treatment Group, Department of Surgery, University of South Florida College of Medicine, c/o Tampa General Hospital, P.O. Box 1289, Tampa, FL 33601, USA Received: 25 February 2005/Accepted: 28 March 2005/Online publication: 6 September 2006

Abstract Background: Improved outcomes of laparoscopic Rouxen-Y gastric bypass (LRYGB) have been demonstrated once pratice has moved beyond the learning curve. However, there is no evidence that experience has a favorable impact on the incidence of leaks. This study evaluated the incidence of staple-line leaks as experience accrued in a university-based bariatric surgery program. Methods: Prospectively collected data on our first 200 patients undergoing LRYGB since July 1998 were analyzed. Linear staplers were used to divide the stomach and to create a side-to-side jejunojejunostomy. A sideto-side cardiojejunostomy was created using a 21-mm circular stapler. Patient characteristics, operative data, and outcomes were evaluated chronologically with comparison of outcomes between quartiles. Results: Staple-line leaks developed in 9 (4.5%) of the first 200 patients undergoing LRYGB. Among the 200 patients were 190 women (95%). The median age of the patients was 48 years (ranges, 24–62 years), and their body mass index was 43 kg/m2 (ranges, 32–59 kg/m2). As surgeonsÕ experience increased over time, there was a significant increase in the weight of patients and the percentage of patients with previous abdominal operations. There also was a significant decrease in conversion rates and operative times. Leaks occurred in six patients at the cardiojejunostomy (3%), in two patients jejunojejunostomy (1%), and in one patient at the excluded stomach (0.5%). Of the 50 leaks that occurred in each quartile, there were in the 3 in the 1st quartile, 1 in the 2nd quartile, 2 in the 3rd quartile, 3 in the 4th quartile. The differences were not significant. There was no correlation between the number of LRYGBs, and the occurrence of a leak (p = 0.59 confidence interval )0.13–0.22).

Presented in part at the 13th International Congress and Endo Expo, the Society of Laparoendoscopic Surgeons Annual Meeting, New York, September 29 to October 2, 2004 Correspondence to: M. M. Murr

Conclusions: The incidence of staple-line leaks appears to be independent of the number of LRYGBs performed. These data suggest that surgeonsÕ experience may not eliminate anastomotic complications experienced by patients undergoing LRYGB. Key words: Roux-en-Y Gastric bypass — Morbid obesity — Learning curve — Laparoscopy — Leaks

The Roux-en-Y gastric bypass (RYGB) results in significant long-term weight loss and is the procedure of choice for the majority of bariatric surgeons [1]. Consequently, gastric bypass for the treatment of clinically significant obesity is rapidly becoming one of the most common operations in North America. It is estimated that in 2003 more than 100,000 gastric bypass surgeries were performed in the United States [11]. This escalating use of bariatric surgery mirrors the epidemic of obesity seen in developed countries as well as the universal failure of nonoperative treatments. Perhaps the single most important factor has been the development of a minimally invasive approach to gastric bypass. The laparoscopic Roux-en-Y gastric bypass (LRYGB) is reported to be advantageous over the open technique [12, 19]. Despite its potential benefits, the LRYGB is considered to be one of the most challenging laparoscopic procedures, and concerns have been raised that it may result in a relatively higher incidence of staple-line leaks. Earlier experience with LRYGB was based on a very select group of patients, and the procedure was demonstrated to be safe and effective in pioneering centers. However, there have been concerns about the safety of LRYGB performed in low-volume centers. The learning curve for LRYGB, one of the longest in laparoscopic surgery, is estimated to require 75 to 100 cases, [7, 14, 16]. Accumulation of operative experience in laparoscopic surgery is reportedly leads to improved outcomes such as decreased operative times and overall complication rates [7, 14, 16]. However, there are no

1688

specific data that examine the relationship between the learning curve and staple-line leaks in bariatric surgery. This study aimed to assess the incidence of staple-line leaks, considered one of the most dreaded complications in bariatric surgery, as experience accrued in a university-based bariatric surgery program.

Methods This study was approved by the Institutional Review Board of the University of South Florida College of Medicine. Prospectively collected data on patients undergoing bariatric surgery for clinically significant obesity in our bariatric surgery program were analyzed. We evaluated our first 200 patients who underwent primary LRYGB since our first laparoscopic procedure was performed in July 1998. Indications for RYGB were based on criteria established by the National Institutes of Health (NIH) Consensus Conference in 1991 [13], including a body mass index of (BMI) 40 kg/m2 or more a BMI of 35 kg/m2 or more with obesity-related comorbidities. Patients underwent a standardized preoperative physical, nutritional, and psychological evaluation, and data were collected prospectively. Early in our experience, we limited the LRYGB to women weighing 250 Ib or less without previous abdominal operations to avoid added technical difficulties derived from adhesions as well as the relative predominance of visceral fat in men. As we gained dexterity and became familiar with the laparoscopic technique, we increased our weight limit for LRYGB to more than 400 Ib, and have used this approach in men and women, even those with previous abdominal operations. None of these procedures were undertaken for patients with previous bariatric operations.

Operative technique We routinely perform a cholecystectomy for all patients undergoing gastric bypass. Our operative technique for LRYGB has been described previously [10]. Briefly, the angle of His is dissected, and a window is made between the neurovascular bundle and the lesser curvature, just inferior to the left gastric artery. The anvil of a no. 21 endoscopic circular stapler is introduced through a gastrotomy in the body of the stomach and exteriorized through the anterior wall of the cardia. The gastrotomy is closed with a 3.5-mm EndoGIA (U.S. Surgical, Norwalk, CT, USA), and the stomach is then divided using multiple applications of a 3.5-mm EndoGIA from the window in the lesser curvature to the angle of His, thereby resulting in a disconnected 15- to 30-ml pouch of cardia. After the ligament of Treitz has been verified, a point of maximal mobility in the proximal jejunum is identified and divided using a 2.5mm EndoGIA. The mesentery is divided using a 2.0-mm EndoGIA to provide additional mobilization to the Roux limb. It is our practice to use a 100-cm Roux limb for patients with a BMI up to 49 kg/m2 and a 150-cm Roux limb for patients with a BMI between 50 and 59 kg/m2. Patients with a BMI of 60 kg/m2 or more undergo a very long gastric bypass that involves anastomosing of the biliopancreatic and Roux limbs 100 cm proximal to the ileocecal valve. After the length of the Roux limb has been measured, a side-to side jejunojejunostomy is created with one application of a 3.5-mm EndoGIA. The enterotomies used to introduce the stapler are closed with another load of the 3.5mm EndoGIA. The mesenteric defect is closed using the Endostitch device (U.S. Surgical). The Roux limb then is brought to the upper abdomen in an antecolic, antegastric position. A side-to-side cardiojejunostomy then is completed with the no. 21 endoscopic circular stapler (Ethicon, Cincinnati, OH, USA). The circular anastomosis is tested by inflating air under water seal to detect any defects in the staple line. The cardiojejunostomy is submerged under saline solution, and the Roux limb is clamped distally. Air is insufflated under vision via a laparoscopic needle introduced into the Roux limb. We have found this method to be reliable, inexpensive, fast, and without complications. On occasion, we place additional sutures between the Roux limb and the gastric pouch to reinforce the anastomosis and reduce tension on the cardiojejunostomy.

Fig. 1. Total of Roux-en-Y gastric bypass (RYGB) procedures performed in our institution per year since 1998, divided according to the approach (open and laparoscopic). The numbers in parentheses represent the percentage of patients undergoing laparoscopic RYGB and show a significant increase (p < 0.001; r = 0.98; 95% CI, 22.97– 31.46). The percentage of patients approached laparoscopically also increased over time (p = 0.0002; r = 0.91; 95% CI, 8.02–14.12).

The potential sites of staple-line failures and leaks include the linear staple line in the divided gastric pouch, the stapled end of the Rouxlimb, the cardiojejunostomy, the linear staple line in the divided excluded stomach, the stapled gastrotomy, and the jejunojejunostomy.

Statistical analysis The number of procedures perforemed by one surgeon was chronologically correlated with the incidence of staple-line disruptions according to the quartile period during which the operation occurred. Age, sex, BMI, history of previous abdominal operations, and obesityrelated comorbidities also were divided and compared between quartiles. Complications were divided into early and late occurrence. They were considered early if they occurred less than 30 days after the operation, and late if they occurred 30 or more days afterward. The incidence of complications after LRYGB also/was evaluated chronologically, with an emphasis on staple-line leaks, comparing patient characteristics and outcomes between quartiles. Continuous parametric data were compared using the two-tailed StudentÕs t-test or analysis of variance (ANOVA). Continuous nonparametric data were analyzed using the two-tailed Mann–Whitney U test. Categorical data were compared using the two-tailed FisherÕs exact test or chi-square for trend analysis. Linear regression was used to correlate the number of procedures with the incidence of comorbidities and with operative outcomes. A p value less than 0.05 was considered statistically significant. Results are expressed as median and range.

Results A total of 624 patients underwent RYGB surgery from June 1998 to August 2004, and 200 (32%) of these procedures were performed laparoscopically. Figure 1 depicts the increase in the number of RYGB procedures (both open and laparoscopic) undertaken per calendar year. There was a significant increase in the number of RYGBs undertaken per year (p < 0.001; 95% confidence interval, 22.97–31.46) as well as a significant increase in the percentage of patients approached laparoscopically each year (p = 0.0002; 95%CI, 8.02–14.12). The median age of the patients was 48 years (range, 24–62 years). The age of the patients increased between the 1st and 4th quartiles, but did not reach statistical significance (Table 1). Only 10 of the patients were men

1689 Table 1. Comparison of patient characteristics and operative outcomes in quartilesa

Age (years) n (range) Gender (male/female) Weight (Ib)b n (range) BMI (kg/m2) n (range) Prior abdominal operations, patients n (%)b Incidence of comorbidities, patients n (%) Conversion rate, patients n (%)b Operative time (min)b n (range) Complication rate, patients n (%)

1st Quartile (n = 50)

2nd Quartile (n = 50)

3rd Quartile (n = 50)

4th Quartile (n = 50)

p Value

44 (26–60) 0/50 237 (213–313) 44 (38–52) 4 (8) 25 (50) 7 (14) 305 (170–852) 10 (20)

43 (25–62) 4/46 276 (185–411) 45 (32–56) 10 (20) 29 (68) 3 (6) 264 (120–325) 12 (24)

44 (24–57) 3/47 286 (233–435) 47 (39–57) 19 (38) 31 (72) 1 (2) 218 (183–362) 8 (16)

50 (30–60) 2/48 294 (224–458) 48 (40–59) 27 (54) 37 (74) 0 212 (105–285) 8 (16)

NS NS 0.006 NS