Surgery for Obesity and Related Disea

Technic xenY

Mark D h.D.c,Emma Patters H. F.A.C.S.e,Gregory Daki skey, M.D.e,

James .D.i,

Received December 13, 2013; accepted May 21, 2014

The last 2 decades have seen a dramatic increase in the*Correspondence: Dr. Mark Douglas Smith, 37 Duke Street, Invercar-

numbers of bariatric operations performed in the United Statesand worldwide. Reasons for this include the growing epidemic

http://dx.doi.org/10.1016/j.soard.2014.05.0361550-7289/r 2015 American Society for Metabolic and Bariatric Surgery. Published by Elsevier Inc. All rights reserved.

gill 9810, New Zealand.E-mail: mdsmith2@mac.com, mdsmith2@me.com (Mark D. Smith)Keywords: Roux-en-Y gastric bypass; Anastomotic leak; Technical factorsAbstract Background: Anastomotic leak is one of the most serious complications after Roux-en-Y gastric bypass(RYGB). Our objective was to examine the relationship between technical factors and incidence ofclinically relevant anastomotic leak after RYGB in longitudinal assessment of bariatric surgery (LABS).The setting of the study was 11 bariatric centers in the United States, university, and private practice.Methods: Patient characteristics, technical factors of surgery, and postoperative outcomes wereassessed by trained researchers using standardized protocols. Correlation of surgical factors ofpatients undergoing RYGB (n 4444) with the incidence of postoperative anastomotic leak wasassessed by univariate 2 analysis.Results: Forty-four participants (1.0%, 95% CI .7%1.3%) experienced a clinically relevantanastomotic leak. Of these, 39 (89%) underwent abdominal reoperation and 3 (7%) died. Technicalfactors associated with anastomotic leak were open surgery (Po .0001), revision surgery(Po .0001), and use of an abdominal drain (P .02). Provocative leak testing, method of gas-trojejunostomy, and use of brin sealant were not associated with anastomotic leak.Conclusions: Anastomotic leak after RYGB was rare (1.0%). Most cases required reintervention;however, the majority (93%) recovered from this event. Open surgery, revision surgery, and routinedrain placement were associated with increased leak rate. Some of these ndings may be due todifferences in preoperative patient risk. (Surg Obes Relat Dis 2015;11:313320.) r 2015 AmericanSociety for Metabolic and Bariatric Surgery. Published by Elsevier Inc. All rights reserved.INational Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MarylandjOregon Health and Science University, Portland, Oregonal factors associated with anastomotic leak after Rougastric bypass

. Smith, M.B.Ch.B.a,*, Abidemi Adeniji, Ph.D.b, Abdus S. Wahed, Pon, M.D.a, William Chapman, M.D.d, Anita P. Courcoulas, M.D. M.P.n, M.D. F.A.C.S.f, David Flum, M.D. M.P.H. F.A.C.S.g, Carol McCloE. Mitchell, M.D.h, Alfons Pomp, M.D. F.A.C.S.f, Myrlene Staten, M

Bruce Wolfe, M.D. FACSjaLegacy Good Samaritan Medical Center, Portland, Oregon

bBoehringer Ingelheim, Ridgeeld, ConnecticutcUniversity of Pittsburgh, Pittsburgh, Pennsylvania

dDepartment of Surgery, East Carolina University School of Medicine, Greenville, North CarolinaeUniversity of Pittsburgh Medical Center, Pittsburgh, Pennsylvania

fCornell University, New York, New YorkgDepartment of Surgery, University of Washington, Seattle, Washington

hNeuropsychiatric Research Institute, Fargo, North DakotaOriginal articleses 11 (2015) 313320

anyone who had undergone bariatric surgery before enroll-

review of case notes and discussion with the treating

y andof obesity [1,2], the reported effectiveness of bariatric surgeryin improving life expectancy and serious co-morbidities [3,4],and the excellent safety prole of modern bariatric surgery [5].Although multiple surgical options currently exist to promotedurable weight loss, Roux-en-Y gastric bypass (RYGB)remains one of the most commonly performed operations [6,7].Although RYGB is effective in promoting durable weight

loss [8], it may be complicated by a number of majorpostoperative events. Anastomotic leak after gastric bypass israre; however, its consequences may be devastating.Reported rates of anastomotic leak vary from .6% to 4.4%[9]. Surgical reexploration is usually required for anastomoticleak, and hospital stay is prolonged [10,11], resulting inincreased cost and morbidity. Anastomotic leak is also anindependent risk factor for early postoperative mortality [12].Factors associated with anastomotic leak include clinical (or

patient) factors and technical factors. Identied clinical factorsassociated with anastomotic leak include male sex, age, andpresence of sleep apnea [12]. Unfortunately, other than notoffering surgery to high-risk individuals, there is often littlethat can be done to reduce clinical risk. In contrast to clinicalrisk factors, technical risk factors are potentially modiable bythe operating surgeon to reduce risk of anastomotic leak.Examples of technical factors include method of constructingthe anastomoses, intraoperative leak testing, and routineabdominal drainage. However, the rare incidence of anasto-motic leak after RYGB makes it difcult for a single surgeonor center to accrue enough events to identify risk factors orinvestigate strategies to reduce its incidence. Consequently,many of the strategies employed by surgeons are either basedon basic surgical principles or extrapolated from other gastro-intestinal surgery [13,14]. A recent guideline published by theAmerican Society of Metabolic and Bariatric Surgery foundno high-quality evidence to support any intervention to reducethe incidence of anastomotic leaks [15].The Longitudinal Assessment of Bariatric Surgery (LABS)

is an 11-center consortium funded by the National Institute ofDiabetes, Digestive and Kidney Diseases (NIDDK) in theNational Institutes of Health (NIH) that conducts observationalcohort studies of bariatric surgical outcomes. These involvelargely prospective, standardized, and comprehensive collec-tions of clinical data. LABS-1 collected 30-day outcome datain consecutive patients, aged 18 years or older, undergoingprimary bariatric surgery. LABS-2 comprises more detailedand ongoing data collection in a selected cohort of patients,restricted to those who had not had prior bariatric surgery. Thepurpose of the present study was to describe the incidence andoutcomes of anastomotic leak after RYGB in LABS and toexamine technical factors associated with its occurrence.

Methods

Patients

M. D. Smith et al. / Surgery for Obesit314Patients were recruited by LABS into either of 2 cohorts,designated LABS-1 and LABS-2, at one of the 11surgeons. Method of gastrojejunostomy was recorded onthe initial data collection forms as a binary response towhether each of 3 methods (hand sewing, linear stapling, orcircular stapling) was used in constructing the anastomosis.This was recoded to reect the general use of these clinicalterms by surgeons. Linear stapled with or without handsewing is treated as predominantly linear stapled, circularstapled with or without hand sewing is classied aspredominantly circular stapled, and hand sewn with nostaples is classied as predominantly hand sewn.The primary outcome for this analysis is clinically

relevant leak (CRL). Clinically relevant leak is dened asing in LABS-2. LABS inclusion criteria and data collectionhave previously been described in detail [16]. Data werecollected for bariatric surgeries performed between March2005 and April 2009 and sent to the Data CoordinatingCenter (DCC) at the University of Pittsburgh, GraduateSchool of Public Health.The analysis included LABS patients (LABS-1 and

LABS-2 patients) who underwent RYGB, either open orlaparoscopic, including revision of RYGB.

Data collection

Data were collected prospectively on a large number ofclinical and technical factors [16]; however, the site of anidentied anastomotic leak was not part of the initial LABSdata collection protocol. Because site of anastomotic leakwas considered important to this analysis, this informationwas collected retrospectively by LABS site investigators byparticipating centers: University of Pittsburgh Medical Center(Pennsylvania), New YorkPresbyterian Hospital (Columbia-Presbyterian or Valley Hospitals or Weill-Cornell MedicalCollege; New York and New Jersey), East Carolina MedicalCenter (North Carolina), the MeritCare Health Systemsthrough the Neuropsychiatric Research Institute (NorthDakota), Sacramento Bariatric (California), University ofWashington Medical Center or Virginia Mason MedicalCenter (Washington), and Oregon Health and Sciences Uni-versity or Legacy Good Samaritan Medical Center (Oregon).The LABS protocols and consent forms were approved by theInstitutional Review Board at each institution.

Protocols

LABS-1, a study of 30-day outcomes, included allconsecutive patients at least 18 years of age who consentedto participate. LABS-2 involves long-term follow-up anddata collection. Accordingly, in contrast to LABS-1, non-consecutive patients who would be able to undertake therequired follow-up were selected for recruitment, excluding

Related Diseases 11 (2015) 313320patients undergoing readmission or reintervention for asuspected anastomotic leak, where the presence of a leak

was subsequently conrmed by the investigator or LABSadjudication subcommittee on review of the medical record.Anastomotic leaks diagnosed on the basis of radiology oranalysis of abdominal drain contents, which did not requirereadmission or reintervention, were not considered CRLsfor the purpose of this analysis.

Statistical analysis

Baseline demographic characteristics are presented asfrequencies and percentages for categorical variables andcompared between groups (e.g., participants with and with-out anastomosis test) using 2 test or the exact equivalent(e.g., Fishers exact test) as appropriate. Continuous charac-teristics such as the duration of surgery are summarized withmedians and interquartile ranges and are compared betweengroups using the Wilcoxon rank-sum test. Incidence of CRLin the overall sample and in the subgroups is presented asfrequencies and percentages. The association between base-line demographic characteristics, co-morbidities, intraopera-tive factors, and routine leak testing with the likelihood ofclinically relevant leak was evaluated using 2 test or its

exact version (when the expected frequencies were smallerthan 5), separately for each variable. Pair-wise comparisonsbetween various methods of anastomosis were conductedusing Holms step-down method to adjust for multiplicity.We did not consider a multivariable adjusted analysis

because the outcome CRL was rare (approximately 1%) andhence the sample size was too small to conduct a statisti-cally powered adjusted analysis. For all statistical analyses,Statistical Analysis Systems (SAS) version 9.3 (SASInstitute; Cary, NC) was used. A cut-off of o.05 was usedto determine statistical signicance.

Results

The analysis included 4444 patients who had RYGBsurgery in LABS consortium (LABS-1 and LABS-2).Patient clinical characteristics are presented in Table 1.

Clinically relevant leaks

Of the included 4444 patients, 44 suffered a CRL (1%).The most common site of leak was the gastrojejunostomy

Table 1Participant demographic characteristics, overall and by anastomosis test

Characteristic Total (N 4444) Patients without anastomosis test(n 636)

Patients with anastomosis test(n 3803)

P

Surgery performed n % n % n % o.0001

Laparoscopic gastric bypass 3841 86.4 379 59.6 3458 90.9Open gastric bypass 603 13.6 257 40.4 345 9.1

18.0

811641871525217.8

654321 50.5 2024 53.219164

493143

40594

62312

503132

Technical Factors and Anastomotic Leak after RYGB / Surgery for Obesity and Related Diseases 11 (2015) 313320 315Patient age (yr), range 18.075.0Median (Q1, Q3) 45.0 (36.0,53.0)Age (yr)o30 431 9.73039 1126 25.34049 1302 29.35059 1179 26.560 406 9.1

BMI (kg/m2), range 17.8107.4Median (Q1, Q3) 46.9 (42.6, 52.6)BMI (kg/m2)o35 66 1.535o 40 462 10.440o 50 2346 52.850o 60 1174 26.460 396 8.9

MaleNo 3574 80.4Yes 870 19.6

White/Caucasian (41 missing)No 509 11.6Yes 3894 88.4

Hispanic (2 missing)No 4186 94.2Yes 256 5.8

Current or recent smoker (13 missing)No 3766 85.0Yes 665 15.0BMI body mass index.30.0 982 25.810.1 332 8.7

.0477.5 3078 80.922.5 725 19.1

o.00016.3 468 12.493.7 3296 87.6

o.000198.1 3558 93.61.9 244 6.4

o.000179.2 3258 85.920.8 533 14.123.9 1025 27.08.2 354 9.3

79.7 19.2107.4 .00648.1 (43.1, 53.6) 46.8 (42.5, 52.4)

.0530.9 58 1.58.5 407 10.775.0 18.073.0 .05744.0 (35.5, 52.0) 45.0 (36.0, 53.0)

.0412.7 349 9.225.8 962 25.329.4 1113 29.3

(28/44, Table 2), followed by the jejunojenostomy (7/44).Clinical characteristics of the patients with and withoutCRL are presented in Table 3. Although not statisticallyadequately powered (because of low incidence rate), weexplored the bivariate association (unadjusted) betweenpatient and surgeon characteristics with incidence of CRL.In unadjusted analysis, no demographic characteristics weresignicantly associated with the incidence of CRL.Of the 44 patients with a CRL, 39 required reoperation

(17 laparoscopic, 12 open). The median time betweenRYGB and reintervention was 5.5 days (range 125 d).Median length of stay was 7.0 days in those with a CRLcompared with 2.0 days for those without. Approximately

method of gastrojejunostomy is unlikely to affect CRL at

Table 3a. Clinically relevant leak by demographic characteristics

Characteristic N CRL P

n %

Age (yr) .46

o30 431 2 .53039 1126 8 .74049 1302 15 1.25059 1179 13 1.160 406 6 1.5

BMI (kg/m2) .07o35 66 2 3.035o 40 462 7 1.540o 50 2346 16 .750o 60 1174 12 1.0

60 396 7 1.8Male .88No 3574 35 .9Yes 870 9 1.0

White/Caucasian (41 missing) .61No 509 4 .8Yes 3894 40 1.0

Hispanic (2 missing) .76No 4186 41 .9Yes 256 3 1.2

Current or recent smoker (13 missing) .13No 3766 41 1.1Yes 665 3 .5

Albumin (g/dL) (394 missing) .65Low:o3 g/dL 24 0 0Normal: 36 g/dL 4026 35 .9

CRL clinically relevant leak; BMI body mass index.Table 3b. Clinically relevant leak by health status

Characteristic N CRL P

n %

Hypertension .39No 1897 16 .8Yes 2547 28 1.1Hypertension therapy (75 missing) .77No medication 263 4 1.5Single medication 1099 11 1.0Multiple medication 1110 13 1.2

Diabetes .10No 2849 23 .8Yes 1595 21 1.3

Type of diabetes medication (4 missing) .42No diabetes medication 232 1 .4Single oral medication 476 5 1.1Multiple oral medication 369 7 1.9Insulin (with or without oral meds) 514 8 1.6

Congestive heart failure .27No 4348 42 .9Yes 96 2 2.1

Asthma .89No 3373 33 1.0Yes 1071 11 1.0

Inability to walk 200 ft (1 missing) .19

M. D. Smith et al. / Surgery for Obesity and Related Diseases 11 (2015) 313320316sites other than the gastrojejunostomy, sensitivity analysiswas conducted by restricting the analysis to leaks occurringat the gastrojejunostomy only. In this analysis the incidenceof CRL at gastrojejunostomy was .31% versus .64% versus.92% for predominantly linear-stapled versus predominantly

Table 2Site of clinically relevant leak*

Site of leak Total (N 43)Gastric pouch staple line 5 11.4Gastrojejunostomy 28 63.6Gastric remnant staple line 2 4.6Jejunojejunostomy 7 15.9Small intestine other 1 2.3Others 4 9.1

*Clinically relevant leak is a conrmed leak at gastric pouch staple lineor gastrojejunostomy.

Multiple sites may apply.1 participant missing site of leak form.List of others (n 4): colon; intraabdominal abscessexploratory98% of those with no CRL had a length of hospital stayo7days, compared with 50% with a CRL.The mortality rate was 6.8% (3/44) among those with a

CRL compared with .3% (13/4400) among those withno CRL.

Technical factors

A CRL was more common among those undergoing arevision RYGB compared with a primary surgery(Table 4), and after open compared with laparoscopicRYGB. Placement of an abdominal drain was alsoassociated with an increased incidence of CRL. The useof brin sealant was not associated with CRL in LABS.These associations persisted when analysis was restrictedto leaks occurring at the gastrojejunostomy or gastricpouch only.In the initial analysis CRL was observed less often after a

linear stapled gastrojejunostomy compared with circular-stapled or hand-sewn gastrojejunostomy. However, becausesurgery negative; gastric fundus; esophageal leak.No 4359 42 1.0Yes 84 2 2.4

Of the 3803 with anastomosis test result, 143 (3.8%) hada positive leak test. No demographic or clinical character-istics had any trend of association with the likelihood of apositive leak test (data not shown). No statistically

Table 4a. Clinically relevant leak by technical factors

Characteristic N CRL P

n %

Surgery type o.0001

Primary surgery 4286 34 .8

urgery for Obesity and Related Diseases 11 (2015) 313320 317Table 3Continued.

Table 3b. Clinically relevant leak by health status

Characteristic N CRL P

n %

History of DVT or PE .07No 4271 40 .9Yes 173 4 2.3

Sleep apnea .42No 2187 19 .9Yes 2257 25 1.1CPAP .23No 423 7 1.7

Technical Factors and Anastomotic Leak after RYGB / Scircular-stapled versus hand-sewn gastrojejunal anastomo-sis, respectively (P .12).

Provocative leak testing

Of the 33 surgeons who had done at least 10 LABSsurgeries, 29 surgeons (87.8%) routinely tested (in Z90%cases) the gastrojejunal anastomosis during surgery. Of the3803 (85.7%) patients who underwent provocative leaktesting, 143 (3.8%) had a positive test. The anastomosis testwas done by 3 methods: 548 were tested with air, of whom10 (1.8%) tested positive; 2011 were tested with endoscopy,of whom 114 (5.6%) tested positive; and 1515 were testedwith methylene blue, of whom 20 (1.3%) tested positive(Fig. 1).

Yes 1834 18 1.0Supplemental oxygen dependent (9 missing) .93No 2165 24 1.1Yes 83 1 1.2

Ischemic heart disease .63No 4213 41 1.0Yes 231 3 1.3

Pulmonary hypertension .47No 4394 43 1.0Yes 50 1 2.0

Venous edema w/ulcerations (280 missing) .0003No 3961 34 .9Yes 203 7 3.4

Beta-blocker (61 missing) .45No 3558 33 .9Yes 825 10 1.2

Statin/lipid-lowering agent .0827No 3241 27 .8Yes 1203 17 1.4

Therapeutic anticoagulation .06No 4227 39 .9Yes 217 5 2.3

Narcotic .22No 3645 33 .9Yes 799 11 1.4

Antidepressant (61 missing) .005No 2545 16 .6Yes 1838 27 1.5

CRL clinically relevant leak; DVT deep vein thrombosis; PE pulmonary embolism; CPAP continuous positive airway pressure.

Revision surgery 149 10 6.7Neither primary or revision surgery 9 0 0

Surgery performed o.0001Laparoscopic gastric bypass 3841 29 .8Open gastric bypass 603 15 2.5

Gastrojejunostomy sealed (22 missing) .36No 1854 21 1.1Yes 2568 22 .9

Drain placed at gastrojejunostomy .02No 3711 31 .8Yes 733 13 1.8

Method of gastrojejunostomy (33 missing) .02Predominantly linear stapled 1594 9 .6Predominantly EEA 1731 17 1.0Predominantly hand sewn 1086 18 1.7

Anastomosis tested (5 missing) .06No 636 2 .3Yes 3803 42 1.1

CRL clinically relevant leak; EEA end to end anastomosis (circular)stapled.

Table 4b. Clinically relevant leak at gastrojejunostomy or gastric pouch bytechnical factors

Characteristic N CRL at GJ P

n %

Surgery type o.0001Primary surgery 4286 25 .6Revision surgery 149 7 4.7Not rst time and not revision 9 0 0Surgery performed o.0001Laparoscopic gastric bypass 3841 21 .5Open gastric bypass 603 11 1.8

Gastrojejunostomy sealed (22 missing) .999No 1854 13 .7Yes 2568 18 .7Drain placed at gastrojejunostomy .02No 3711 22 .6Yes 733 10 1.4Method of gastrojejunostomy (33 missing) .02Predominantly linear stapled 1594 6 .4

Predominantly EEA 1731 12 .7Predominantly hand sewn 1086 14 1.3

Anastomosis tested (5 missing) .07No 636 1 .2Yes 3803 31 .8CRL clinically relevant leak; GJ gastrojejunostomy; EEA end toend anastomosis (circular) stapled.

signicant difference in CRL rate was found between those

GB su

stom

%) -ve

.4%) sCR

0 confi

y met

y andundergoing positive, negative, or no provocative leak test.Median operative time (skin incision to skin closed) was

similar for those surgeries that included anastomosis testcompared with those that did not (136 versus 138 minutes,P .14).4444 RY

4439 with ana

3803 tested*548 By air 2011 By endoscopy1515 By methylene blue

143 (3.8%) +ve**10 (1.8%) Air 114 (5.6%) Endoscopy20 (1.3%) Meth. blue

3660 (96.2

52 (11 (0.7%) suspected CRL***

*Some were tested by multiple methods** 1 was tested positive by both air and endoscopy

***This person was tested positive for anastomotic leak by air

1 confirmed 3

Fig. 1. Clinically relevant leak b

M. D. Smith et al. / Surgery for Obesit318Discussion

This analysis found the incidence of CRL after RYGB inLABS is rare at 1%. Clinically relevant leaks were morecommon after open and revision surgery and when anabdominal drain was placed at the time of surgery. Nodifference in CRL at the gastrojejunostomy was observedbetween the different methods of constructing this anasto-mosis. Provocative leak testing was not associated with theincidence of gastrojejunostomy CRL in LABS. Most casesof CRL required reintervention with consequent increase inhospital stay and mortality risk; however, most cases alsorecovered from this event.The ndings of the present study compare favorably with

other large series on anastomotic leak after RYGB. A single-center series examining the outcomes of 3828 patientsundergoing RYGB over a 23-year period found an anasto-motic leak rate of 3.9% [10]. Anastomotic leaks were morecommon after revisional RYGB (8.0%); however, in contrastto LABS, leaks were more common after laparoscopicRYGB (5.2%) than open RYGB (2.6%). It is important tonote that this series includes the period when laparoscopicRYGB was introduced and thato1/3 procedures werelaparoscopic (1080/3828). In this series, the overall mortalityafter anastomotic leak was 14.7%, although mortality washigher after leak at the jejunojeunostomy (40%) than thegastrojejunostomy (9%). Sixty-eight percent of patients inthis series with anastomotic leak at the gastrojejunostomywere managed nonoperatively, with none of these requiringlater surgical intervention and no mortality in this group.Another single center series examined the outcomes of 60

patients with anastomotic leak from a total of 1764 under-

rgeries

osis test data

5 did not have data on anastomosis testof which 1 had a suspected CRL

636 not tested

2 (0.31%) suspected CRL

uspected L

1 confirmedrmed

hod of anastomotic leak testing.

Related Diseases 11 (2015) 313320going RYGB (leak rate 3.4%) [17]. Again approximatelyone third (n 573) of patients had RYGB performed via alaparoscopic approach. The authors of this series dividedleaks into localized subclinical leaks (n 12) and thosewhere the patients exhibited clinical signs of sepsis(n 48). The overall mortality rate from anastomotic leakwas 10%, 1 of the 3 patients with a leak from thejejunojejunostomy died (mortality rate 33%). None ofthe patients with subclinical leaks died.A systematic review focusing specically on patients

undergoing laparoscopic RYGB identied 10 studies includ-ing 3464 patients. A total of 71 anastomotic leaks wereidentied to give an unadjusted anastomotic leak rate of 2%.Although this anastomotic leak rate was more in keepingwith that found in LABS, the included studies may haveincorporated patients receiving surgery early in the experi-ence of laparoscopic RYGB. Because of the nature of thisreview, no analysis was conducted exploring the associationsof clinical or technical factors with anastomotic leak.A MEDLINE-based literature search was unable to

identify any large studies exploring the relationshipsbetween multiple technical factors and anastomotic leakafter RYGB. Nevertheless individual technical factors havebeen investigated in numerous clinical studies. One suchfactor is whether the gastrojejunostomy is constructed using

urgera linear-stapled, circular-stapled, or hand-sewn technique.One of the interesting ndings of the present analysis wasthe decreased overall anastomotic leak rate after linearstapled anastomoses in LABS; however, this nding failedto persist when restricted to leaks at the gastrojejunostomyonly. This corresponds with the ndings of a MichiganBariatric Surgery Collaborative survey where reported leakrates were identical for these 3 anastomosis techniques [18],although circular stapled gastrojejunostomies were associ-ated with an increased risk of anastomotic hemorrhage andwound infection. A meta-analysis of studies comparinglinear-stapled with circular-stapled anastomoses also foundno difference between anastomotic leak rates but did showan increased stricture rate with circular stapled anastomoses[19]. Because plausible explanations as to how method ofgastrojejunostomy would affect leaks at other sites arelacking, this nding from LABS may have resulted fromchance alone.The use of tissue sealants was not associated with

alteration of the anastomosis leak rate in LABS. This isconsistent with the results of a randomized trial of brinsealant after RYGB where no difference in anastomotic leakrate was observed between the treatment and control groups[20], although the use of brin sealant was associated with areduced risk of anastomotic hemorrhage.Placement of an abdominal drain was associated with an

increased incidence of CRL in LABS. This study is not arandomized trial of drain usage, and data were not collectedon the reason for drain placement. It is possible that drainswere more likely to be placed when a surgeon was concernedabout an elevated risk of CRL. These results, however, donot support the assumption that abdominal drainage reducesthe need for reintervention for CRL. This nding isconsistent with previous comparative studies showing routinedrain placement to be unnecessary after RYGB [21].Randomized trials evaluating the use of routine drainageafter other major abdominal surgery have also failed to ndbenet, even suggesting an increased complication rate withdrain placement after cholecystectomy [2225].No randomized trials could be identied of intraoperative

leak testing for RYGB; however, a randomized trial ofpneumatic testing in colorectal anastomoses found adecreased CRL rate in tested patients [13]. Three series ofpatients undergoing RYGB with routine endoscopic intra-operative leak testing report CRL rates of o1% [2628].Another study compared leak testing via methylene bluedelivered by orogastric tube with endoscopic leak testing[29]. In this study endoscopic leak testing had a higher rateof positive leak tests intraoperatively but a lower rate ofpostoperative CRL. Analysis of LABS data also indicatedmore positive leak tests after endoscopic testing comparedwith other methods. However, because only a handful ofpatients (269) were tested by multiple methods, we do not

Technical Factors and Anastomotic Leak after RYGB / Shave sufcient data to compare different methods of testing.Although no difference was found in CRL betweenparticipants undergoing positive, negative or no provocativeleak test, these ndings should be interpreted with caution.The number of CRLs was small and, as with drain place-ment, no data were collected on intraoperative events thatcould have inuenced a surgeons decision to perform leaktesting. Furthermore, the presence of a positive leak testwould usually result in the operating surgeon reinforcing orrevising the anastomosis, preventing CRL. Unfortunatelydata on the management of a positive leak test were notcollected in LABS. Difference in patient risk or technicalfactors not measured or corrected for in LABS could alsopartially explain the ndings of the present analysis.LABS is a large multicenter study incorporating university

hospitals and private clinics. Despite the large numbers ofincluded patients and comprehensive data collection, thepresent study has revealed several counterintuitive ndings.One explanation for this is the design of LABS as anobservational study rather than a randomized trial. Random-ized trials are the only study design able to account fornonmeasured confounding factors; however, large random-ized trials are expensive, time consuming, and only able todraw strong conclusions regarding the randomized interven-tion. Despite this limitation, LABS data collection and riskadjustment was prespecied to include all variables deemedpotentially clinically relevant study investigators. Neverthe-less it is possible that unmeasured clinical or technical factorsmay have contributed to some of the unexpected ndings ofthis analysis. Participants may have undergone open RYGBbecause of anticipated technical difculty; and surgeons mayhave chosen to use brin sealant, place a drain, and conductprovocative leak testing when they already had concernsabout the integrity of an anastomosis.The denition of CRL used in this study is another

potential weakness. Because of the prespecied design fordenition and capture of postoperative complications, it wasnot possible to identify anastomotic leaks that did notrequire readmission to hospital or endoscopic, radiologic, orsurgical reintervention. We hypothesize that the only likelysituation where this would occur is if a drain left at theinitial RYGB was completely successful in draining theleak. However, the nding of increased rates of CRL in apatient where a drain was placed at the initial RYGB makesthis outcome unlikely.Another potential weakness of this study is the small

number of CRLs. Although a testament to the excellentsafety prole of modern RYGB surgery, this limited ourability to perform adjusted analyses of associations. Also,univariable associations presented were not adequatelypowered. Therefore, the conclusions drawn should be takenas observations from a large prospective multicenter study.The low number of CRLs in the present study also limitsthe ability to extrapolate these ndings to centers with ahigher rate of CRL.

y for Obesity and Related Diseases 11 (2015) 313320 319The results of the present study conrm the low incidenceof CRL in modern RYGB surgery. Although revision

y andsurgery continues to present an increased risk of CRL, ourndings do not support increased leak rates after laparo-scopic RYGB compared with the open approach. In fact,the reverse now appears to be true. Although anastomoticleak remains a major complication, with prompt diagnosisand appropriate management, resulting mortality is rare.Some of the unexpected or counterintuitive ndings of

the present study might be addressed by future research.The question of whether or not to test gastrojejunalanastomosis intraoperatively would be best addressed bymulticenter randomized, controlled trials. However, todetect a doubling of anastomotic leak rate from 1% to 2%would require randomizing 45000 patients. Such trials arelogistically difcult and expensive to conduct. An alter-native is systematic review and meta-analysis of smallerrandomized trials. Universally agreed endpoints and well-designed studies would facilitate later quantitative summaryof the results.

Conclusions

Anastomotic leak after RYGB in LABS was rare, limit-ing the power of even this large multicenter study toexamine associated factors. Method of gastrojejunostomy,routine provocative leak testing, and use of brin sealantwere not associated with incidence of CRL in LABS. In thisnonrandomized study, increased incidence of CRL wasobserved after revision or open RYGB compared withprimary laparoscopic surgery. No technical factors wereidentied as associated with a reduced incidence of CRL.

Disclosures

The authors have no commercial associations that mightbe a conict of interest in relation to this article.

References

[1] Flegal KM, Carroll MD, Ogden CL, Curtin LR. Prevalence and trendsin obesity among US adults, 1999-2008. JAMA 2010;303:23541.

[2] Ford ES, Mokdad AH. Epidemiology of obesity in the WesternHemisphere. J Clin Endocrinol Metab 2008;93:S18.

[3] Adams TD, Gress RE, Smith SC, et al. Long-term mortality aftergastric bypass surgery. N Engl J Med 2007;357:75361.

[4] Sjostrom L, Narbro K, Sjostrom CD, et al. Effects of bariatric surgeryon mortality in Swedish obese subjects. N Engl J Med 2007;357:74152.

[5] Flum DR, Belle SH, King WC, et al. Perioperative safety in thelongitudinal assessment of bariatric surgery. N Engl J Med 2009;361:44554.

[6] Hinojosa MW, Varela JE, Parikh D, Smith BR, Nguyen XM, NguyenNT. National trends in use and outcome of laparoscopic adjustablegastric banding. Surg Obes Relat Dis 2009;5:1505.

[7] Santry HP, Gillen DL, Lauderdale DS. Trends in bariatric surgicalprocedures. JAMA 2005;294:190917.

[8] White S, Brooks E, Jurikova L, Stubbs RS. Long-term outcomes aftergastric bypass. Obes Surg 2005;15:15563.

M. D. Smith et al. / Surgery for Obesit320[9] Nguyen NT, Wilson SE. Complications of antiobesity surgery. NatClin Pract Gastroenterol Hepatol 2007;4:13847.[10] Lee S, Carmody B, Wolfe L, et al. Effect of location and speed ofdiagnosis on anastomotic leak outcomes in 3828 gastric bypass cases.J Gastrointest Surg 2007;11:70813.

[11] Marshall JS, Srivastava A, Gupta SK, Rossi TR, DeBord JR. Roux-en-Y gastric bypass leak complications. Arch Surg 2003;138:5203. (discussion 523).

[12] Fernandez AZJ, DeMaria EJ, Tichansky DS, et al. Experience withover 3,000 open and laparoscopic bariatric procedures: multivariateanalysis of factors related to leak and resultant mortality. Surg Endosc2004;18:1937.

[13] Beard JD, Nicholson ML, Sayers RD, Lloyd D, Everson NW.Intraoperative air testing of colorectal anastomoses: a prospective,randomized trial. Br J Surg 1990;77:10957.

[14] Ricciardi R, Roberts PL, Marcello PW, Hall JF, Read TE, SchoetzDJ. Anastomotic leak testing after colorectal resection: what are thedata? Arch Surg 2009;144:40711 (discussion 411).

[15] ASMBS guideline on the prevention and detection of gastrointestinalleak after gastric bypass including the role of imaging and surgicalexploration. Surg Obes Relat Dis 2009;5:2936.

[16] Belle SH, Berk PD, Courcoulas AP, et al. Safety and efcacy ofbariatric surgery: Longitudinal Assessment of Bariatric Surgery. SurgObes Relat Dis 2007;3:11626.

[17] Csendes A, Burgos AM, Braghetto I. Classication and managementof leaks after gastric bypass for patients with morbid obesity:a prospective study of 60 patients. Obes Surg 2012;22:85562.

[18] Finks JF, Carlin A, Share D, et al. Effect of surgical techniques onclinical outcomes after laparoscopic gastric bypassresults from theMichigan Bariatric Surgery Collaborative. Surg Obes Relat Dis2011;7:2849.

[19] Giordano S, Salminen P, Biancari F, Victorzon M. Linear staplertechnique may be safer than circular in gastrojejunal anastomosis forlaparoscopic Roux-en-Y gastric bypass: a meta-analysis of compara-tive studies. Obes Surg 2011;21:195864.

[20] Silecchia G, Boru CE, Mouiel J, et al. The use of brin sealant to preventmajor complications following laparoscopic gastric bypass: results of amulticenter, randomized trial. Surg Endosc 2008;22:24927.

[21] Kavuturu S, Rogers AM, Haluck RS. Routine drain placement inRoux-en-Y gastric bypass: an expanded retrospective comparativestudy of 755 patients and review of the literature. Obes Surg 2012;22:17781.

[22] Gurusamy KS, Samraj K. Routine abdominal drainage for uncompli-cated open cholecystectomy. Cochrane Database Syst Rev 2007.(CD006003.).

[23] Gurusamy KS, Samraj K, Davidson BR. Routine abdominal drainagefor uncomplicated liver resection. Cochrane Database Syst Rev 2007.(CD006232.).

[24] Gurusamy KS, Samraj K, Mullerat P, Davidson BR. Routineabdominal drainage for uncomplicated laparoscopic cholecystectomy.Cochrane Database Syst Rev 2007. (CD006003.).

[25] Jesus EC, Karliczek A, Matos D, Castro AA, Atallah AN. Prophy-lactic anastomotic drainage for colorectal surgery. Cochrane DatabaseSyst Rev 2004. (CD002100.).

[26] Alasfar F, Chand B. Intraoperative endoscopy for laparoscopic Roux-en-Y gastric bypass: leak test and beyond. Surg Laparosc EndoscPercutan Tech 2010;20:4247.

[27] Kligman MD. Intraoperative endoscopic pneumatic testing for gastro-jejunal anastomotic integrity during laparoscopic Roux-en-Y gastricbypass. Surg Endosc 2007;21:14035.

[28] Shin RB. Intraoperative endoscopic test resulting in no postoperativeleaks from the gastric pouch and gastrojejunal anastomosis in 366laparoscopic Roux-en-Y gastric bypasses. Obes Surg 2004;14:10679.

[29] Alaedeen D, Madan AK, Ro CY, Khan KA, Martinez JM, Tichansky

Related Diseases 11 (2015) 313320DS. Intraoperative endoscopy and leaks after laparoscopic Roux-en-Ygastric bypass. Am Surg 2009;75:4858. (discussion 488).

Technical factors associated with anastomotic leak after Roux-en-Y gastric bypassMethodsPatientsProtocolsData collectionStatistical analysis

ResultsClinically relevant leaksTechnical factorsProvocative leak testing

DiscussionConclusionsDisclosuresReferences