Surgical Endoscopy

, Volume 25, Issue 4, pp 1287–1292

A comparative study of handsewn versus stapled gastrojejunal anastomosis in laparoscopic Roux-en-Y gastric bypass

Authors

  • Amanda J. Kravetz
    • Department of SurgeryMichigan State University
  • Subhash Reddy
    • Department of SurgeryMichigan State University
  • Ghulam Murtaza
    • Department of SurgeryMichigan State University
    • Department of SurgeryMichigan State University
Article

DOI: 10.1007/s00464-010-1362-x

Cite this article as:
Kravetz, A.J., Reddy, S., Murtaza, G. et al. Surg Endosc (2011) 25: 1287. doi:10.1007/s00464-010-1362-x

Abstract

Background

The two basic techniques available in which to perform the gastrojejunal (GJ) anastomosis during a laparoscopic Roux-en-Y gastric bypass (LRYGBP) are stapled and handsewn. Few outcomes differences have been noted between the two to recommend one as a superior approach. We present our findings in comparison of the two methods.

Methods

This is a retrospective review of all patients who underwent LRYGBP at a single institution during a 3-year period. The two different techniques that were used were a linear stapled and handsewn anastomosis with an anastomotic diameter of 18 mm. The groups were compared for postoperative complications, including stricture, anastomotic leak, and the need for early reoperation. All patients were followed up for a minimum period of 8 months.

Results

A total of 222 patients were analyzed after excluding 4 patients: 3 for revisional surgery and 1 for conversion to open. There were 99 patients in the stapled group and 123 in the handsewn group. In both groups, patients were predominantly female. The average age was 42.63 in the stapled group and 44.33 in the handsewn group (P = 0.218). Body mass index was 48.23 in the stapled group and 47.91 in the handsewn group (P = 0.733). Stricture rate in the stapled group was 10.1% (10/99) and 4.1% (5/123) in the handsewn group (P = 0.076). Four patients from the stapled group (4.08%) and six from the handsewn group (4.88%) needed early reoperation. One patient in each group had a GJ anastomotic leak (0.9%). There were no deaths.

Conclusions

The incidence of anastomotic stricture tends to be lower with a handsewn technique with lower operative time. No difference was appreciated in the anastomotic leak or reexploration rate with either technique.

Keywords

Roux-en-Y gastric bypassStrictureAnastomotic leak

Morbid obesity is in epidemic proportion in the United States, and there is no sign of slowing down [1]. Bariatric surgery has been shown to be very effective for the treatment of morbid obesity, achieving significant weight loss sustainable long-term compared with nonsurgical treatment options [2]. Since the first report of laparoscopic Roux-en-Y gastric bypass by Wittgrove and Clark in 1994 [3], this procedure has become the most commonly performed bariatric surgical procedure [4]. There are three different techniques described for the construction of the gastrojejunal anastomosis, including circular stapled, linear stapled, and completely handsewn anastomosis. The GJ anastomosis is crucial in LRYGBP because leak in the early postoperative period can lead to high morbidity and mortality [5] with later complications, including anastomotic stenosis and stomal ulceration. The reported anastomotic leak and anastomotic stricture rates range from 0–6.6 to 0–33.3% (respectively) in the literature [610]. The optimal technique for the creation of the GJ anastomosis is still debatable. Multiple studies have examined the rate of anastomotic stricture and leak using various GJ anastomotic techniques; however, a limited number of studies compare any two techniques [7, 11]. One of these looked at comparing a circular stapled technique with a linear stapled technique [7]. Another study compared the incidence of stricture using different suture material for handsewn anastomosis [11]. One more retrospective case study, which compared 108 consecutive patients who underwent LRYGBP (87 with handsewn, 13 with circular stapled, and 8 with linear stapled anastomoses), found a lower rate of stricture in the handsewn group and no anastomotic leaks [12]. Therefore, in parallel with the limited literature to date comparing linear stapled anastomosis with handsewn anastomosis, we hypothesized that a handsewn technique will have a lower anastomotic stricture and leak rate, which was the basis of our study.

Materials and methods

Study design

Institutional Review Board approval was obtained to review retrospectively charts and medical records of prospectively collected data on all patients who underwent laparoscopic Roux-en-Y gastric bypass at a single institution from July 2006 through June 2009.

Setting

All cases met National Institute of Health guidelines for treatment of morbid obesity and were done at a single tertiary care level teaching hospital. All the procedures were performed by a single fellowship-trained bariatric surgeon (PY).

Procedure

Operations were done in an antecolic, antegastric technique according to a previously described procedure [13]. In brief, the biliary-pancreatic limb was created at 50 cm and the Roux limb was measured at 75–150 cm depending on the body mass index. Jejunojejunal anastomosis was created with a 60-mm length, white (2.5 mm) load Endo GIA™ stapler (Covidien) using a double-staple technique. The mesenteric defect was closed with nonabsorbable SURGIDAC™ suture. The greater omentum was split vertically to avoid tension on the antecolic Roux limb of jejunum. The gastric pouch was created using multiple blue loads (3.5-mm staples) of the Endo GIA™ stapler. The Roux limb of the jejunum was sutured to the gastric pouch by approximation of the seromuscular layer using 2-0 SURGIDAC™. In the linear-stapled GJ, a side-to-side gastrojejunal anastomosis was created by using a 3.5-mm blue load Endo GIA™ linear stapler fired to a length of 18 mm. The small resultant defect was sutured with running 2-0 POLYSORB™ suture. In the handsewn GJ, a side-to-side anastomosis was performed by using an 18-mm gastrotomy and enterotomy, sized by completing one full-thickness bite with the Harmonic scalpel, which is 15 mm and before that L hook cautery with a length of 3 mm, then approximating them with 2-0 POLYSORB™ suture incorporating full-thickness with running suture. The length of the anastomosis was approximately 18 mm and this does not reflect the inner diameter of the anastomosis. Patency was confirmed with the endoscope. The anterior seromuscular layer was then sutured with running 2-0 SURGIDAC™. All suturing was completed with the Endo Stitch™ device.

Every anastomosis was tested intraoperatively by passing an endoscope through the anastomosis and insufflating to distend the stomach pouch and bowel after a bowel clamp was placed on the proximal part of the Roux limb. If air bubbles were noted, the area that was leaking was reinforced with 2-0 POLYSORB™ in a single or a figure-of-eight suture at the site of the leak until no further air bubbles were noted with the laparoscope. A flat #10 Jackson-Pratt (JP) drain was placed over the anastomosis at the conclusion of each case and left in place for a minimum of 2 days after the surgery. All patients underwent an upper gastrointestinal study using water-soluble contrast material to check for leak or obstruction at both the GJ and jejunojejunal anastomosis on the first postoperative day. A bariatric clear liquid diet was started if the study was negative.

Anastomotic strictures were suspected if the patients had dysphagia for liquids and solids. All of these patients underwent upper endoscopy, and if the adult (10-mm diameter) scope could not be passed across the anastomosis, then a stricture was confirmed and dilated using serial, radial-dilating balloon dilators up to a maximum of 18 mm.

Data

Charts were reviewed and analyzed for complications, the need for additional surgeries or interventions, operative time, and stricture. Demographic data also were collected.

Statistical analysis

All statistical analysis was completed using SigmaPlot 11.2 (2008 Systat Software, Inc., San Jose, CA) with P < 0.05 being chosen a priori as statistically significant.

Results

From July 2006 to June 2009, 226 laparoscopic Roux-en-Y gastric bypass procedures were performed. A total of four patients, including three with revisional surgery and one patient who had to be converted to open gastric bypass because of technical problems, were excluded from the analysis. The conversion rate to an open procedure was 0.45%. Ninety-nine patients who underwent surgery from July 2006 to June 2008 had a stapled GJ, and 123 patients from July 2008 through the end of the study period had a handsewn GJ. All patients were required to undergo smoking cessation prior to surgery. The patient demographics can be found in Table 1.
Table 1

Demographics of the stapled and handsewn groups

 

Stapled anastomosis

Handsewn anastomosis

P value

Age (yr)

42.63 (range 21–62) (9.692; 0.979)

44.33 (range 24–64) (9.912; 0.894)

0.218

Body mass index

48.23 (range 37.5–64.5) (7.044; 0.635)

47.91 (range 35.4–71.3) (6.779; 0.685)

0.733

Operative time

160.08 (range 96–300) (47.31; 4.25)

128.54 (range 54–354) (43.48; 3.9)

<0.001*

Male gender

15/99 (15.2%)

19/123 (15.4%)

0.953

Hypertension

41/99 (41.4%)

60/123 (48.8%)

0.275

Diabetes mellitus

27/99 (27.3%)

33/123 (39.8%)

0.067

Hyperlipidemia

45/99 (45.5%)

78/123 (63.4%)

0.008*

Reflux

35/99 (35.4%)

54/123 (43.9%)

0.198

Obstructive sleep apnea

52/99 (52.5%)

49/123 (39.8%)

0.06

* Indicates statistical significance

Standard deviation; standard error of the mean

A total of ten patients required early reoperation: four (4.08%) from the stapled group and six (4.88%) from the handsewn group. In this subgroup of patients, the average age in the stapled group was 46.0 and 46.4 in the handsewn group. Body mass index (BMI) did not vary significantly; the stapled group had a BMI of 50.55 and the handsewn group had a BMI of 48.74 (P = 0.762). The average length of stay was 5.11 (range 2–12) days. This was increased over the average length of stay of all patients who did not undergo early reoperation by 2.88 days. One patient from the handsewn group underwent a negative exploration (0.45% negative exploration rate) based on a false-positive upper GI study.

Two patients had gastrojejunal anastomotic leaks for a total leak rate of 0.9%. These were both diagnosed on upper GI study on postoperative day 1 and confirmed intraoperatively during reexploration by the air leak test using the upper endoscope and with the methylene blue dye test. The anastomotic leak in the stapled group was managed by revising to a handsewn anastomosis because there was a complex defect on the posterior aspect of the anastomosis that was not amenable to simple repair. The leak in the patient who had the handsewn anastomosis was primarily repaired with a single figure-of-eight suture.

Another patient from the handsewn group had a leak from an enterotomy near the staple line of the Roux limb possibly resulting from a laparoscopic grasper during the initial surgery. One patient in the stapled group and two in the handsewn group did not show any leak, but there was evidence of localized peritonitis in the left upper quadrant. One of the patients in the handsewn group presented with signs of sepsis 1 week after the surgery. These cases were managed with peritoneal lavage and placement of drains with gastrostomy tube placement in the gastric remnant. All patients with leaks and localized peritonitis underwent an upper GI study on the first postoperative day, and all studies were negative. The patients were then started on bariatric clears, and none required the gastrostomy tube for feeding purposes. All complications were managed laparoscopically. There were no mortalities in this series. The summary of findings can be found in Table 2.
Table 2

Early reoperation

Patient

No. days to reoperation

LOS

Home with drains

Anastomosis type

Complication

1

2

6

Yes

Stapled

Peritonitis; no leak identified. Abdominal drainage and G-tube placed

2

7

12

No

Stapled

Incarcerated ventral hernia; repaired with mesh

3

2

5

Yes

Stapled

Congestion of the Roux limb; mesentery released and G-tube placed

4

1

3

No

Stapled

GJ anastomotic leak; revised

5

1

3

No

Handsewn

Enterotomy of Roux limb; repaired

6

2

4

No

Handsewn

Incarcerated ventral hernia; repaired

7

2

6

No

Handsewn

Negative diagnostic laparoscopy

8

2

5

Yes

Handsewn

Peritonitis; abdominal drainage and G-tube placed

9

9

2

Yes

Handsewn

Intra-abdominal abscess; Abdominal drainage, JP and G-tube placed

10

1

10

Yes

Handsewn

GJ anastomotic leak; revised

Ten patients (10.2%) in the stapled group and five patients (4.1%) in the handsewn group developed anastomotic stricture that required esophagojejunoscopy with pneumatic balloon dilatation (P = 0.076). Two patients with stapled anastomoses required two dilatations. In the handsewn group, one patient had two dilatations and one patient had four dilatations. All other strictures required only one dilatation. The average time to the first dilatation in the stapled group was 48.5 (range 12–102) days and in the handsewn group was 114.2 (range 29–263) days with P = 0.042. We did not experience any complications of the balloon dilatation. We were unable to identify risk factors for stricture formation.

Discussion

The effectiveness of laparoscopic Roux-en-Y gastric bypass for surgical weight loss has clearly been demonstrated and the studies have shown the safety of the procedure [14]. Anastomotic leak has been associated with a higher rate of morbidity and mortality. Anastomotic stricture is one of the most common long-term complications related to this procedure [15]. Specifically, we analyzed our patient population, which was divided by a linear stapled technique and a handsewn technique for the gastrojejunal anastomosis.

In an effort to decrease the complication rate, various investigators have analyzed the different techniques for performing gastrojejunal anastomosis. One study of open Roux-en-Y gastric bypass [16] demonstrated that handsewn anastomosis was superior to stapled anastomosis in decreasing the leak rate. A large, retrospective study published in 2008 [17] quoted a leak rate of 6%, stating that their leak rate declined over the course of several years, with zero leaks in the later part of the study. A review of multiple studies quoted a range of 0–6.6% [18] where they found that the rate of gastrojejunal anastomotic leak was identical between an open GJ and a laparoscopic technique. Marema et al. [19] also found no difference in the GJ anastomotic leak rate between laparoscopic and open Roux-en-Y gastric bypasses (0.6 and 0.2% respectively). Our leak rate was similar between the stapled and handsewn groups with a total leak rate of 0.9%, which is in accordance with the published literature.

The American Society for Metabolic and Bariatric Surgery has issued guidelines on the prevention and detection of gastrointestinal leak after gastric bypass [20], and these have been substantiated by a number of retrospective reviews and small studies. Our review continues to corroborate the finding by the ASMBS that a GJ anastomotic leak in LRYGBP is not influenced by the type of anastomosis.

Management of GJ leaks has come into question as well. A large, retrospective review of 2,675 patients from 2008 [21] advocated for nonoperative treatment of GJ anastomotic leaks, indicating that the hospital stays in the operative group were longer than in the nonoperative group. We managed all of our complications operatively, and our results suggest that this is a viable option for management with potentially minimal increase in the length of stay (1 extra day in one patient and 8 extra days in the other patient).

The incidence of anastomotic stricture has been another area for concern. The rate of stricture formation quoted in the literature ranges from 0 to 33.3% [610]. Furthermore, these rates have encompassed circular stapled, linear stapled, and handsewn GJ anastomotic techniques, and all of these techniques have reported both high and low rates of stricture formation [11], which makes the literature confusing. Carrodeguas and colleagues reviewed 1,291 patients at their institution and published a stricture rate of 7.3% [15]. All of their GJ anastomoses were completed utilizing a standard, linear stapled technique. Higa et al. [22] reported a stricture rate of 4.9% with handsewn anastomosis in a series of 1,040 cases; this corresponds with our stricture rate. Adana et al. [11] reported a randomized study comparing stricture rates using monofilament and braided suture in a handsewn GJ, and the stricture rates were 0.7 and 9.7%. This stricture rate needs to be reproduced by other groups to validate this hypothesis. Another study by Bohdjalian et al. [7] compared 25-mm, circular-stapled with linear-stapled GJ anastomoses and found that the circular-stapled technique had a higher incidence of strictures. A small study comparing 87 patients with handsewn anastomoses to 13 patients with circular and 8 patients with linear stapled anastomosis demonstrated a lower rate of strictures in the handsewn group [12].

In our study, we found that a linear-stapled technique resulted in a higher rate of strictures (10.1%), which is similar to that found by Gonzalez (albeit comparing a total of only 8 linear-stapled patients), which also corresponds to that found by Carrodeguas et al. [15] but in contrast to that found by Bohdjalian et al. [7]. At this point, it is difficult to explain the difference in the findings of stricture rates, which appear to be favoring a handsewn anastomosis. In regards to nicotine use being a potential factor in stricture formation, all patients were required to undergo smoking cessation before surgery, and this was checked by doing blood carbon monoxide levels if patients gave any history of recent smoking. We did not do any blood tests to check the smoking status during the postoperative period. None of the patients who had stricture admitted to smoking after surgery at the time of stricture diagnosis. We admit that if patients were not truthful in their reporting of smoking cessation this may have confounded our results.

Time to the onset of stricture was longer (114.2 days) in the handsewn group compared with the linear stapled anastomosis (48.5 days); this is statistically significant. Even though it is an interesting finding, it could be due to multiple reasons, including tissue ischemia due to the staple line. This may cause early stricture. In the handsewn group, there is probably less tissue ischemia, and the stricture may be due to anastomotic contracture toward the end of healing process. Animal studies may be necessary to elucidate these findings, and this definitely needs to be an area of further research.

One may question whether—given a 4% stricture rate—a larger anastomosis may be beneficial to completely avoid stricture. With a larger anastomosis, there is less restriction and over time there can be dilatation of the anastomosis, which may be a factor in weight regain. There are no studies to show the ideal diameter of the anastomosis. General thinking is to create an anastomosis that causes some restriction, but at the same time does not cause stricture. If we aim to create some restriction at the anastomosis, then it may be extremely difficult to avoid strictures completely. Again, if the strictures are due to tissue ischemia, they may be independent of the diameter of the anastomosis.

There are some limitations of this study. Our operative times were noted to improve significantly with a handsewn anastomosis. However, given the retrospective nature and the time course of this study (i.e., the stapled anastomoses were all performed during the initial part of the study and the handsewn technique during the latter part), one can argue that this may be related to the learning curve and/or to the sheer volume that we completed by the end of the study. Furthermore, the retrospective nature of the study in and of itself is a limitation, and to draw conclusive results, we will need to complete a prospective, randomized, controlled trial on this subject. Next, the overall reoperation rate was approximately 4%, which may be higher than other institutions. We added in all of our reoperations, not only for GJ anastomotic leak, but also included in our analysis one negative reexploration as well as those done for early postoperative obstruction and others. The 4% rate could be due to the learning curve, as seen in the improvement in our operative times during the course of the study, but this is difficult to conclude with certainty. Finally, because this study is a single-surgeon and single-institution study, it is difficult to extrapolate to other surgeons and institutions.

In conclusion, we believe that our study indicates that the strictures and operative times were lower in the handsewn group compared with the stapled group. Anastomotic leak rate and the incidence of early reexploration do not vary by the type of GJ anastomosis performed, and thus are equivalent. Standardized techniques compared in a randomized, prospective fashion are needed to fully delineate the optimal technique in performing the GJ anastomosis.

Disclosures

Drs. Amanda Kravetz, Subhash Reddy, Ghulam Murtaza, and Panduranga Yenumula have no conflicts of interest or financial ties to disclose.

Copyright information

© Springer Science+Business Media, LLC 2010