birkmayer 2011 bariatric surg complications in michigan

7/29/2019 Birkmayer 2011 Bariatric Surg Complications in Michigan

1/8

ORIGINAL CONTRIBUTION

Hospital Complication RatesWith Bariatric Surgery in MichiganNancy J. O. Birkmeyer, PhD

Justin B. Dimick, MD, MPH

David Share, MD, MPH

Abdelkader Hawasli, MD

Wayne J. English, MD

Jeffrey Genaw, MD

Jonathan F. Finks, MD

Arthur M. Carlin, MD

John D. Birkmeyer, MD

for the Michigan Bariatric SurgeryCollaborative

W

ITH RATES INCREASING

over the last decade,1

bariatric surgery has be-

come the second mostcommon abdominal operation in theUnited States.Despitetrends toward de-clining mortality rates,2 payers and pa-tient advocacy groups remain con-cerned about the safety of bariatricsurgery and uneven quality across hos-pitals. In response, 2 major profes-sional organizationsthe AmericanCollege of Surgeons and the AmericanSociety for Metabolic and Bariatric Sur-geryhave implemented programs foraccrediting hospitals as centers of ex-

cellence (COE) in bariatric surgery.Standards for COE accreditation varysomewhat between the programs, butthey generally include minimum pro-cedure volume standards, availabilityof specific protocols and resources formanaging morbidly obese patients, andsubmission of outcomes data to a cen-tral registry.

Context Despite the growing popularity of bariatric surgery, there remain concernsabout perioperative safety and variation in outcomes across hospitals.

Objective To assess complication rates of different bariatric procedures and vari-ability in rates of serious complications across hospitals and according to procedure

volume and center of excellence (COE) status.Design, Setting, and Patients Involving 25 hospitals and 62 surgeons statewide,the Michigan Bariatric Surgery Collaborative (MBSC) administers an externally au-dited, prospective clinical registry. We evaluated short-term morbidity in 15 275 Michi-gan patients undergoing 1 of 3 common bariatric procedures between 2006 and 2009.We used multilevel regression models to assess variation in risk-adjusted complicationrates across hospitals and the effects of procedure volume and COE designation (bythe American College of Surgeons or American Society for Metabolic and Bariatric Sur-gery) status.

Main Outcome Measure Complications occurring within 30 days of surgery.

Results Overall, 7.3%of patients experiencedperioperativecomplications,mostofwhichwere wound problems and other minor complications. Serious complications were mostcommon aftergastricbypass (3.6%; 95%confidenceinterval[CI], 3.2%-4.0%),followedby sleeve gastrectomy (2.2%; 95% CI, 1.2%-3.2%), and laparoscopic adjustable gastric

band (0.9%; 95% CI, 0.6%-1.1%) procedures (P.001). Mortality occurred in 0.04%(95% CI, 0.001%-0.13%) of laparoscopicadjustable gastric band, 0 sleeve gastrectomy,and 0.14% (95% CI, 0.08%-0.25%) of the gastric bypass patients. After adjustment forpatient characteristics and procedure mix,ratesof serious complications variedfrom 1.6%(95%CI, 1.3-2.0)to 3.5%(95% CI,2.4-5.0)(risk difference,1.9;95% CI,0.08-3.7) acrosshospitals. Average annual procedure volume was inversely associated with rates of seri-ouscomplicationsat boththe hospitallevel(150cases,4.1%;95%CI,3.0%-5.1%;150-299cases,2.7%;95% CI,2.2-3.2;and300cases,2.3%; 95%CI, 2.0%-2.6%; P=.003)andsurgeon level (100cases,3.8%; 95% CI, 3.2%-4.5%; 100-249 cases,2.4%; 95%CI, 2.1%-2.8%;250 cases,1.9%; 95% CI, 1.4%-2.3%; P=.001). Adjusted rates of se-riouscomplicationsweresimilar inCOE andnon-COEhospitals(COE,2.7%;95% CI,2.5%-3.1%; non-COE, 2.0%; 95% CI, 1.5%-2.4%; P=.41).

Conclusions The frequency of serious complications among patients undergoing bar-iatric surgery in Michigan was relatively low. Rates of serious complications are in-versely associated with hospital and surgeon procedure volume, but unrelated to COEaccreditation by professional organizations.

JAMA. 2010;304(4):435-442 www.jama.com

Author Affiliations:Department of Surgery,and Cen-terfor HealthcareOutcomesand Policy(CHOP)(DrsN. Birkmeyer, Dimick, Finks, and J. Birkmeyer), andDepartment of Family Medicine (Dr Share), Univer-sity of Michigan, Ann Arbor; Department of Sur-gery, St John Hospital and Medical Center, Detroit,Michigan (Dr Hawasli); Department of Surgery,Marquette General Hospital, Marquette, Michigan

(DrEnglish); andDepartment of Surgery, Henry FordHospital, Detroit, Michigan (Drs Genaw and Carlin).Corresponding Author: Nancy J. O. Birkmeyer, PhD,Michigan Surgical Collaborative for Outcomes Re-search and Evaluation(M-SCORE),Departmentof Sur-gery, University of Michigan, 211 N Fourth Ave, Ste2A and 2B, Ann Arbor, MI 48104 ([email protected]).

2010 American Medical Association. All rights reserved. (Reprinted) JAMA, July 28, 2010Vol 304, No. 4 435

at CENTRE HOSPITALIER UNIVERSITAIRE NICE on July 7, 2011jama.ama-assn.orgDownloaded from
http://jama.ama-assn.org/http://jama.ama-assn.org/http://jama.ama-assn.org/http://jama.ama-assn.org/


2/8

WhetherCOEaccreditationhelps pa-tients and payers identify safer hospi-tals for bariatric surgery remains amatter of debate. Hospital procedurevolume, a core component of accredi-tation, has been linked to periopera-

tive mortality with bariatric surgery.

3-8

However, many of these studies areout-dated. As bariatric surgery has beenmore commonplace and mortality hasdeclined, previous hospital volumebenchmarks (125 per year for COEs)may be less important now than in thepast. To date, only 1 published studyhas directly compared the outcomes ofbariatric surgery at COE and non-COE hospitals, noting higher mortal-ity and equivalent morbidity at theformer.6 Because this study was basedon administrative data, however, its re-

liability in capturing hospital compli-cation rates is questionable. It also in-cluded data from only 1 year, 2005,when COE programs were just begin-ning to be implemented.

In thiscontext,we studied periopera-tive outcomes at 25 hospitals par-ticipating in the Michigan BariatricSurgery Collaborative (MBSC), a payer-funded quality improvement programthat administers a prospective, exter-nally audited clinical outcomes regis-try. In addition to comparing compli-

cation rates by procedure and amonghospitals, we examined relationshipsbetween procedure volume, COE ac-creditation, and hospital safety.

METHODS

Study Sample

This study is based on analysis of datafrom the MBSC. As described in greaterdetail elsewhere, the MBSC is a re-gional consortium of hospitals and sur-geons performing bariatric surgery inMichigan.9,10 Participation in the MBSC

is voluntary and any hospital thatperforms a minimum of 25 bariatricprocedures per year is eligible toparticipate. The MBSC now enrolls ap-proximately 6000 patientsper year from25 hospitals in its clinical registry. Par-ticipating hospitals submit data for allof their bariatric surgery patients in-cluding those undergoing gastric by-

pass, laparoscopic gastric banding, bil-iopancreatic diversion with or withoutduodenal switch, and sleeve gastrec-tomy procedures. Procedures done onan outpatient basis are included in theMBSC registry and are subject to the

same data collection requirements.In theMBSC, data for theclinical reg-istry is collected via medical record re-view for each patient at the end of the30-day perioperative period. Informa-tion collected includes demographicvariables, preoperative clinical charac-teristics and conditions, as well as peri-operative process of care and out-comes. Patient readmissions to otherhospitals are captured if it is recordedin the medical records of the hospitalperforming the bariatric surgery. Themedical record reviews are performed

by centrally trained nurse data abstrac-tors using a standardized and vali-dated instrument. Each participatinghospital is visited annually by theproject data quality coordinator to verifythe accuracy and completeness of itsMBSC clinical registry data. The col-lection of data for the purposes of par-ticipation in the MBSC has been ap-proved by the institutional reviewboards of all member sites.

For this study, we identified all pa-tients undergoing bariatric surgery be-

tween June 2006 and September 2009,which includes 15 275 patients from 25hospitals. We excluded patients under-going revisional surgery fromthis analy-sis because of the heterogeneity of thepatient population and surgical proce-dures as well as inherently higher ratesof complications for patients undergo-ing revisional surgery. We also ex-cluded patients undergoing duodenalswitch (n= 245) for confidentiality rea-sons since most of these procedureswere performed by 1 surgeon in the

state. We combined patients undergo-ing open and laparoscopic gastric by-pass procedures as there was no differ-ence in therates of major complicationswith the 2 procedures following ad-justment for patient case mix and be-cause open gastric bypass is now per-formed so rarely (5% of patientsduring the study period).

Outcomes

Data were collectedon 12 differenttypesof bariatric surgeryrelated complica-tions. Complications were groupedaccording to severity as nonlife-threatening, potentially life-threaten-

ing,or life-threatening complicationsas-sociated with residual and permanentdisability or death. Potentially life-threatening complicationsincluded ab-dominal abscess (requiring percutane-ous drainage or reoperation), bowelobstruction (requiringreoperation),leak(requiring percutaneous drainage or re-operation), bleeding (requiring trans-fusion 4 units, endoscopy, reopera-tion,or splenectomy), respiratory failure(requiring 2-7 days intubation), renalfailure (requiring dialysis while patientis hospitalized during the perioperative

period), woundinfection/dehiscence (re-quiring reoperation), andvenous throm-boembolism (deep vein thrombosis orpulmonary embolism). Complicationsresulting in permanent disabilityincluded myocardial infarction or car-diac arrest, renal failure requiring long-term dialysis,respiratory failure requir-ing more than 7 days of intubation, ortracheostomy. The MBSC end pointscommittee grades the severity of anyperioperative complications not fallingunambiguously into one of these cat-

egories. Our primary outcome mea-sure for this study was the occurrenceof a serious complication defined as po-tentially life threatening or resulting indeath or disability.

Independent Variables

Data on patient characteristics includepatient demographics, weight and medi-calhistory, andweight-relatedand othercomorbidities listed inTABLE 1. Ingen-eral, MBSC comorbidity definitions in-clude clinical documentationof the con-

dition, its treatment, or both in themedical record. Lung disease includesasthma,other obstructive/restrictive lungdisease, and home oxygen use. Cardio-vascular disease includes coronary ar-terydisease,heart rhythm disorder, con-gestive heart failure, or peripheralvascular disease. Patients withnonalco-holic fattyliver,clinicalor subclinicalcir-

HOSPITAL COMPLICATION RATES WITH BARIATRIC SURGERY

436 JAMA, July 28, 2010Vol 304, No. 4 (Reprinted) 2010 American Medical Association. All rights reserved.



3/8

rhosis, or liver transplant are consid-ered to have liver disorders.

Annual hospital and surgeon volumecategories (TABLE 2) were determinedusingacombinationofgenerallyacceptedvolume cutpoints andempiricalderiva-

tionbasedonthedistributionofpatients,hospitals,andsurgeons.Sitesweredeemedcenters of excellence if they were desig-natedassuchbytheAmericanCollegeofSurgeonsortheAmericanSocietyofMeta-bolic andBariatricSurgeons at anypointduring our study period. Two sites heldBlueCrossandBlueShieldCentersofDis-tinction status, which has similar crite-ria to the other COE accreditation pro-grams in addition to COE accreditationfrom the American College of SurgeonsorAmericanSocietyofMetabolicandBar-iatric Surgeons.

Statistical Analyses

Pearson 2 test for categorical variablesand the Kruskal-Wallis testfor continu-ous variables were used to compare pa-tient characteristics and rates of 30-daycomplicationsamongpatientsundergo-ingthe differenttypes of bariatricproce-dures.Multilevelmixed-effectslogisticre-gressionmodelswereusedtoevaluateriskfactorsfor seriouscomplications,withthelog (odds) of the outcome modeled as alinearfunctionof baselinecovariates. The

final modelsincludedallpatientrisk fac-torsthatweresignificant in multivariateanalyses(age,bodymassindex[calculatedasweightinkilogramsdividedbyheightin meters squared], male sex, mobilitylimitations,priorhistoryofvenousthrom-boembolism,andtotalnumberofcomor-bidconditions)andproceduretype(lapa-roscopic adjustable gastric band, sleevegastrectomy, or gastric bypass) as fixedeffects,andhospitalidentifierasarandomeffect to adjust for clustering of patientswithin hospitals.

Because hospital and surgeon com-plication rates can vary due to chancealone, we adjusted our estimates for re-liability. This technique adjusts hospi-tal and surgeon outcomes for randomvariation, ensuring that performance isnotoverestimatedor underestimateddueto statistical noise.11 Empirical Bayesmethods shrink the observed complica-

tion rate at each hospital or for eachsurgeon toward the overall average, de-pending on its reliability. Reliability ismeasuredon a scale of 0 (completely un-reliable) to 1 (perfectly reliable) and islargely a function of samplesize. Forthis

analysis,weusedthe randomeffectsfromthe mixed-effects models to calculaterisk- and reliability-adjusted complica-tions ratesfor each hospital. For this cal-culation, we add the overall average log(odds) of serious complications to therandom effect (since the mean is 0 bydefinition)andthentaketheinverselogitof this sum. All reported P values are2-sided, andP.05 was considered sta-tistically significant. All statistical analy-ses were performed using Stata version10.1(StataCorp, CollegeStation,Texas).

RESULTSPatient Characteristics

Thereweresignificantdifferencesacrossthe procedure types with regard to allpotential risk factors for complica-tions, including demographics, medi-

cal history, and obesity-related comor-bidity (Table 1). In general, patientsreceiving laparoscopic adjustable gas-tric bands were lower risk than pa-tients receiving gastric bypass or sleevegastrectomy. Specifically, patients re-ceiving laparoscopic adjustable gas-tric bands had significantly lower bodymass index at baseline and lower ratesof associatedcomorbid conditions. Thepredicted risk of serious complica-tions based on a logistic regressionmodel including significant multivar-

Table 1. Patient Characteristics and Predicted and Adjusted Rates of Serious ComplicationsAccording to Bariatric Procedure Type

Characteristic

Procedures, %

P

ValueaOverall

(N=15 275)

LaparoscopicAdjustable

Gastric Band(n=5380)

SleeveGastrectomy

(n=854)

GastricBypass

(n=9041)

DemographicsBMI, median (IQR)b 46 (42-52) 43 (40-49) 50 (44-56) 47 (43-56) .001

Age, median (IQR), y 46 (37-54) 47 (38-56) 47 (37-55) 45 (37-54) .001

Male sex 21 20 30 20 .001

Private insurance 72 76 80 69 .001

Medical historyMusculoskeletal disorder 76 70 68 79 .001

Cardiovascular disease 56 55 59 56 .001Hyperlipidemia 50 46 50 51 .001

Gastroesophageal reflux disease 47 46 43 48 .001

Psychological 46 44 43 48 .001

Sleep apnea 44 38 56 47 .001

Current or past smoking 39 36 41 40 .001

Diabetes 34 30 36 36 .001

Total No. comorbidities 5 29 24 31 32 .001

Cholelithiasis 27 24 29 28 .001

Lung disease 26 23 20 27 .001

Urinary incontinence 20 19 23 20 .001

Mobility problems 5 5 10 5 .001

Liver disorder 4 3 3 4 .001

Prior venous thromboembolism 4 3 5 4 .001

Peptic ulcer disease 3 2 4 3 .001History of hernia repair 3 2 5 3 .001

Predicted risk of seriouscomplicationsc

2.7 2.4 3.0 2.8 .001

Abbreviations: BMI, body mass index; IQR, interquartile range.aP values for medians calculated using a nonparametric k-sample test on the equality of medians and P values for

categorical variables calculated using 2 tests.b BMI was calculated as weight in kilograms divided by height in meters squared.c Based on a multivariate logistic regression model including all significant patient risk factors for serious complications

(age, body mass index, male sex, mobility limitations, prior history of venous thromboembolism, and total number ofcomorbid conditions).





4/8

iate predictors (age, body mass index,male sex, mobility limitations, prior his-tory of venous thromboembolism, andtotal number of comorbid conditions)was 2.4% for laparoscopic adjustablegastric band, 3.0% for sleeve gastrec-

tomy, and 2.8% for gastric bypass.Incidence of Specific Complications

Overall, 7.3%(95%confidence interval[CI],6.9%-7.7%)of patientsexperienced1 or more perioperative complications.Ratesof potentiallylife-threateningcom-plications(TABLE3)werehighestforpa-tientsundergoing gastricbypass (3.1%;95%CI, 2.8%-3.5%), followedby sleevegastrectomy(2.2%;95%CI,1.3%-3.5%),andlaparoscopicadjustablegastricband(0.78%; 95% CI, 0.56%-1.1%) proce-dures(P.001).Fatalcomplicationsoc-curred in 2 patients receiving laparo-scopic adjustable gastric band (0.04%;95%CI,0.01%-0.13%),0patientsreceiv-ing sleeve gastrectomy,and 13 patientsreceivinggastricbypass(0.14%;95%CI,0.08%-0.25%).Complicationsthat werenot fatal butthatresultedin permanentdisability occurred in 2 patients receiv-

inglaparoscopicadjustablegastricband(0.04%; 95% CI,0.01%-0.13%) and30patientsreceivinggastricbypass(0.33%;95% CI, 0.22%-0.47%).

Complicationsat thesurgical siteoc-curredin5.9%ofpatients(95%CI,5.6%-

6.3%) and were highest in patients un-dergoinggastricbypass (8.7%; 95%CI,8.1%-9.3%),followedbypatientsreceiv-ing sleeve gastrectomy (3.6%; 95% CI,2.5%-5.1%), andlaparoscopicadjustablegastric band(1.7%; 1.4%-2.1%). Infec-tionwasthemostfrequenttypeofcom-plication(3.2%;95%CI,2.9%-3.5%)andwas most common among patients un-dergoinggastricbypass (4.4%; 95%CI,4.0%-4.8%) and sleeve gastrectomy(2.5%;95% CI, 1.5%-3.7%)procedures(Table 3). The subcategory of medicalcomplications(includingvenousthrom-boembolism, cardiac,renal failure,andrespiratory)occurredin1.5% ofpatientsoverall (95% CI, 1.3%-1.7%), with theincidencevarying from0.58% (95% CI,0.39%-0.82%) in patients with laparo-scopic adjustable gastric band to 2.1%(95%CI,1.8%-2.4%)inpatientswhore-ceived gastric bypass.

Rates of reoperation ranged from0.59% (95% CI, 0.19%-1.4%) for sleevegastrectomy to 2.5% (95% CI, 2.2%-2.8%) for gastric bypass procedures(Table 3). Transfers to other medicalfacilities (0.14%; 95% CI, 0.09%-

0.22%) occurred infrequently. Hospi-tal readmission and emergency depart-ment visits occurred in 4% (95% CI,3.7%-4.3%) and 6.8% (95% CI, 6.4%-7.2%) of patients overall, respectively.Rates of both readmission and emer-gency department visits were lowest inpatients who received laparoscopic ad-justable gastric band and highest inthose receiving gastric bypass. Me-dian hospital length of stay (days) was1 (range, 0-96), 2 (range, 0-63), and 2(range, 0-148) for patients receivinglaparoscopic adjustable gastric band,sleeve gastrectomy, and gastric by-pass, respectively.

Variation in Serious

Complication Rates

Risk- and reliability-adjusted rates of se-rious complications varied from 1.6%(95% CI, 1.3%-2.0%) to 3.5% (95% CI,

Table 2. Distribution of Patients, Hospitals, and Surgeons and Predicted and Adjusted Rates of Serious Complications by Mean AnnualBariatric Procedure Volume Category

No. (%)a

Low Medium HighAnnual bariatric procedures by surgeon, mean 100 100-249 250

Patients 3664 (24) 7542 (49) 4069 (27)

Surgeons 33 (53) 22 (36) 7 (11)

Predicted riskof serious complications, % (95% CI)b 2.5 (2.4-2.6) 2.7 (2.6-2.8) 2.6 (2.5-2.7)

Adjusted serious complication rate, % (95% CI) 3.8 (3.2-4.5) 2.4 (2.1-2.8) 1.9 (1.4-2.3)

Annual bariatric procedures by hospital, mean 150 150-299 300

Patients 1346 (9) 4338 (28) 9591 (63)

Hospitals 10 (40) 9 (36) 6 (24)

COE hospitals 6 (32) 8 (42) 5 (26)

Predicted riskof serious complications, % (95% CI)b 2.7 (2.6-2.8) 2.7 (2.6-2.8) 2.6 (2.6-2.6)

Adjusted serious complication rate, % (95% CI) 4.1 (3.0-5.1) 2.7 (2.2-3.2) 2.3 (2.0-2.6)

Adjusted serious complication rates by surgeon, % (95% CI)Low-volume surgeons 4.0 (2.8-5.3) 4.4 (3.2-5.6) 3.3 (2.3-4.2)

Medium-volume surgeons 6.1 (2.2-10.0) 2.2 (1.7-2.7) 2.4 (1.9-2.9)High-volume surgeons c c 1.9 (1.4-2.3)

Adjusted serious complication rate, % (95% CI) by COE statusNon-COE 3.7 (2.1-5.2) 2.2 (0.8-3.6) 1.6 (1.2-2.1)

COE 4.4 (3.0-5.8) 2.7 (2.2-3.3) 2.6 (2.2-3.0)

Abbreviations: CI, confidence interval; COE, center of excellence.aValues are reported as No. (%) unless otherwise indicated.b Based on a multivariate logistic regression model including all significant patient risk factors for serious complications (age, body mass index, male sex, mobility limitations, prior history

of venous thromboembolism, and total number of comorbid conditions) and procedure type (laparoscopic adjustable gastric band, sleeve gastrectomy, or gastric bypass).c No observation.





5/8

2.4%-5.0%) by hospital (FIGURE). Rateswere significantly lower than the state-wide average for 3 hospitals. The ma-jority of hospitals (68%) had seriouscomplication rates between 2% and3%.

Risk of serious complications was in-versely associated with average annualbariatric procedure volume (Table 2).For surgeon volume, rates in the low-,medium-, and high-volume categorieswere 3.8% (95% CI, 3.2%-4.5%), 2.4%(95%CI, 2.1%-2.8%),and1.9%(95%CI,

1.4%-2.3%), respectively (P

fortrend=.001). For hospital volume, ad-justed ratesofseriouscomplications were4.1% (95% CI, 3.0%-5.1%), 2.7% (95%CI, 2.2%-3.2%), and 2.3% (95% CI,2.0%-2.6%) in low-, medium-, andhigh-volume hospitals, respectively (P fortrend.001). Seriouscomplicationrateswere about twice as high (4.0%; 95% CI,

Table 3. Occurrence of Specific Perioperative Complications and Adverse Events by Procedure Type

Outcome

% (95% CI)

P

ValueaOverall

(N=15 275)

LaparoscopicAdjustable Gastric

Band (n=5380)

SleeveGastrectomy

(n=854)Gastric Bypass

(n=9041)

Any complication 7.3 (6.9-7.7) 2.3 (1.9-2.7) 5.9 (1.3-7.4) 10.3 (9.7-11.0) .001

Nonlife-threatening 4.7 (4.4-5.1) 1.5 (1.2-1.8) 3.6 (2.5-5.1) 6.7 (6.2-7.3) .001Potentially life-threatening 2.3 (2.0-2.5) 0.78 (0.56-1.1) 2.2 (1.3-3.5) 3.1 (2.8-3.5) .001

Permanently disabling 0.21 (0.14-0.30) 0.04 (0.01-0.13) 0 0.33 (0.22-0.47) .001

Fatal 0.10 (0.6-0.16) 0.04 (0.01-0.13) 0 0.14 (0.08-0.25) .09

Combined serious complications b 2.6 (2.3-2.8) 0.86 (0.61-1.1) 2.2 (1.2-3.2) 3.6 (3.2-4.0) .001

Surgical site 5.9 (5.6-6.3) 1.7 (1.4-2.1) 3.6 (2.5-5.1) 8.7 (8.1-9.3) .001

Leak/perforation 0.59 (0.47-0.72) 0.07 (0.02-0.19) 0.35 (0.07-1.0) 0.92 (0.73-1.1) .001

Anastomotic leak 0.49 (0.36-0.64) 0 0 0.49 (0.36-0.64)

Perforation/other leak 0.27 (0.20-0.37) 0.07 (0.02-0.19) 0.35 (0.07-1.0) 0.39 (0.27-0.54) .002

Obstruction 1.5 (1.3-1.7) 0.26 (0.14-0.44) 0.70 (0.26-1.5) 2.4 (2.0-2.7) .001

Small bowel obstruction 0.49 (0.38-0.61) 0 0.12 (0.01-0.66) 0.81 (0.63-1.0) .001

Stricture/other obstruction 1.1 (0.93-1.3) 0.26 (0.14-4.4) 0.59 (0.19-1.4) 1.6 (1.4-1.9) .001

Infection 3.2 (2.9-3.5) 1.3 (1.0-1.6) 2.5 (1.5-3.7) 4.4 (4.0-4.8) .001

Abdominal abscess 0.45 (0.35-0.57) 0.07 (0.02-0.19) 0.47 (0.13-1.2) 0.67 (0.52-0.87) .001

Wound complication 2.7 (2.5-3.0) 0.84 (0.61-1.1) 2.2 (1.3-3.5) 3.9 (3.5-4.3) .001

Port site infection 0.30 (0.32-7.1) 0.30 (0.32-7.1) 0 0

Hemorrhage 1.5 (1.3-1.7) 0.13 (0.05-0.27) 0.59 (0.19-1.4) 2.3 (2.0-2.7) .001

Medical complication 1.5 (1.3-1.7) 0.58 (0.39-0.82) 1.4 (0.73-2.4) 2.1 (1.8-2.4) .001

Venous thromboembolism 0.39 (0.30-0.50) 0.11 (0.04-0.24) 0.94 (0.41-1.8) 0.50 (0.36-0.67) .001

Cardiac 0.10 (0.06-0.16) 0.04 (0.01-0.13) 0 0.14 (0.08-0.25) .09

Renal failure 0.31 (0.23-0.41) 0.07 (0.02-0.19) 0 0.48 (0.34-0.61) .001

Respiratory 0.99 (0.84-1.2) 0.35 (0.21-0.55) 0.47 (0.13-1.2) 1.4 (1.2-1.7) .001

UtilizationReoperation 1.7 (1.5-1.9) 0.63 (0.44-0.88) 0.59 (0.19-1.4) 2.5 (2.2-2.8) .001

Readmission 4.0 (3.7-4.3) 2.0 (1.6-2.4) 5.5 (4.1-7.3) 5.1 (4.6-5.6) .001

Transfer 0.14 (0.09-0.22) 0 0.23 (0.03-0.84) 0.22 (0.14-0.34) .002

Emergency department visit 6.8 (6.4-7.2) 3.1 (2.7-3.6) 7.5 (5.8-9.5) 8.9 (8.4-9.5) .001

Abbreviation: CI, confidence interval.aPvalues were calculated using 2 tests.b Includes potentially life-threatening, permanently disabling, and fatal complications.

Figure. Risk- and Reliability-Adjusted Serious Complication Rates by Site

Center of excellence

Noncenter of excellence

5

4

3

2

1

01 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Hospitals Ranked by Adjusted Serious Complication Rates

Ad

justedSeriousComplicationRate,

%

Based on a multivariate logistic regression model including all significant patient risk factors for serious com-plications (age, body mass index, male sex, mobility limitations, prior history of venous thromboembolism,and total number of comorbid conditions) and procedure type (laparoscopic adjustable gastric band, sleevegastrectomy, or gastric bypass). Error bars indicate 95% confidence intervals. Dotted line indicates the meanserious complication rate in the Michigan Bariatric Surgery Collaborative.





6/8

2.8%-5.3%) for low-volume surgeonsatlow-volume hospitals than for high-

volume surgeonsat high-volume hospi-tals (1.9%; 95% CI, 1.4%-2.3%). Over-all, rates of serious complications weresimilar among patients undergoing sur-gery at a COE (2.7%; 95% CI, 2.5%-3.1%) than for patients undergoing sur-gery at non-COE hospitals (2.0%; 95%CI,1.5%-2.4%). After adjustment for pa-tient case and procedure mix, there re-mained no significantdifference in ratesof serious complications at COE andnon-COE hospitals (adjusted odds ra-tio [OR], 1.27; 95% CI, 0.72-2.25;

P=.41). There also was no significantdif-ference(adjustedOR,1.34;95%CI,0.88-2.05; P=.18) in rates of serious compli-cations in the COE hospitals comparedwith the non-COE hospitals within hos-pital procedure volume categories.

Inour multivariatemodels, includingpatient risk andprocedure mix reducedvariation in serious complication ratesacrosscentersby22%(TABLE4).Individu-allyadding surgeonvolume, hospital pro-cedure volume, and COE status to thismodel reducedvariationinseriouscom-

plicationratesacrosscentersby75%,59%,and 25%, respectively.

COMMENT

In this study, we report major peri-operative adverse outcomes in a largecohort of bariatricsurgery patients.Ourresults provide information about theperioperative risks of the various types

of bariatric procedures in general prac-tice. Overall, 7% of patients experi-

enced perioperative complications. Themajority of complications were nonlife-threatening with minor woundproblems being the most frequent typeof complication. Approximately 2.5%of patients had more serious compli-cations with mortality occurring in0.12% of patients. Complication rateswere highest for patients undergoinggastric bypass, followed by sleeve gas-trectomy, and laparoscopic adjustablegastric band procedures.

Our study also suggests that the out-

comes of bariatricsurgery reported fromselect academic centers are achievablemore broadly. Rates of serious compli-cations were similar across providerswith rates between 2% and 3% for themajority of hospitalsand surgeons. Theresults of our study are similar to thoserecently reported by a select group ofhigh-volume bariatric programs par-ticipating in the National Institutes ofHealthfunded Longitudinal Assess-ment of Bariatric Surgery (LABS) Con-sortium.12 Despite similar patient popu-

lations, the overall rate of death andmajor complications are higher in LABSthan those reportedin our study. Highercomplication rates reported in LABSmay be attributable to the time peri-ods studied, which included patientsundergoing surgery between 2005 and2007 in LABS and between 2006 and2009 in Michigan.

Similar to many high-risk surgicalprocedures, procedure volume has

been shown to be an important pre-dictor of adverse outcomes in bariat-ric surgery.3-8,13,14 The results of ourstudy are similar to what others havefound regarding the magnitude of theprocedure volume effect on morbiditywith bariatric surgery. For example, astudy based on discharge claims datafrom the state of Florida (1999-2003)found approximate 2-fold differencesin adjusted rates of serious complica-tions comparing the lowest to thehighest volume strata for both hospi-

tals and surgeons.7

A limitation ofstudies based on discharge claimsdatabases is their ability to reliablycapture nonfatal complications. Mostof these prior volume outcome analy-ses in bariatric surgery are also quitedated with the most recent cohortincluding patients from 2005.6

Our results support those recentlyre-ported by Livingston6 that COE ac-creditation is not associated with lowerrates of bariatric complications. Theprior study used 2005 National Inpa-

tient Survey data to compare morbid-ity and mortality rates among 19 363bariatric surgery patientsat 24 COE and229 non-COE centers. Mortality rateswere higher at COE centers (0.17%)than non-COE centers (0.09%) andmorbidity rates were close to identical(6.3% COE vs 6.4% non-COE). ORsad-justed for procedure volume, patient

Table 4. Results of Model Fitting

Level Variable

Coefficient (P Value)

Empty Model Model 1 Model 2 Model 3 Model 4

Fixed effectsPatient level Predicted risk a 1.02 (.001) 1. 04 (.001) 1.03 (.001) 1.02 (.001)

Surgeon level Average annual surgeon volume 0.299 (.003)

Center level Average annual hospital volume 0.260 (.02)

COE status 0.166 (.50)

Model information criterionAkaike information criterionb 3616 3497 3492 3495 3499

Covariance parameterCenter Standard deviation of intercepts 0.3895 0.3438 0.1942 0.2501 0.3371

Reduction in between centervariability, %

22 75 59 25

a Based on a multivariate logistic regression model including all significant patient risk factors for serious complications (age, body mass index [calculated as weight in kilogramsdivided by height in meters squared], male sex, mobility limitations, prior history of venous thromboembolism, and total number of comorbid conditions) and procedure type(laparoscopic adjustable gastric band, sleeve gastrectomy, or gastric bypass).

bAkaike information criterion is defined as minus twice log likelihood plus 2 degrees of freedom.





7/8

risk, and teaching status were 1.76(P=.71) and 1.00(P=.97) for mortalityand morbidity, respectively. The studyby Livingston6 differedfrom ours in thatit was based on claims data that cap-tured only in-hospital complications

and also included data from only 1 year(2005) when COE programs were justbeginning to be implemented.15-17

There are a number of reasons whyCOEaccreditation by professionalorga-nizations or payers might not necessar-ily identify safer hospitalswithbariatricsurgery.First,althoughCOEapplicationsoftenaskhospitalsforratesofspecificout-comes(eg,postoperativevenousthrom-boembolism), such outcomes data aregenerallynotauditedforaccuracyorcom-pleteness and are often loosely defined.Second,aside fromminimumcaseloads,

most requirementsfor bariatricCOE ac-creditation,including the availability ofspecificprotocolsandresourcesforman-aging morbidly obese patients, are eas-ily met by most hospitals with bariatricprograms and likely have little bearingon surgical complication rates. Finally,given thehighlycompetitivemarketplacefor bariatricsurgery, COEaccreditationprogramsmaybeattractinghospitalsmo-tivatedas muchby marketingadvantageas by the desire to demonstrate and im-prove their quality.

Thisstudy hasnumerouslimitations.First, because all but 8 of the 25 hospi-tals were COE-accredited by the end ofthestudy period,we hadsuboptimal sta-tistical power fordetecting differencesinrisk between COEand non-COE hospi-tals.Basedonan levelof.05,ourstudyhad only 70% power to detect a relativeriskreductionof25%ormoreassociatedwith COE accreditation.

Second, we counted as COEs anyhospital that had received that desig-nation by the end of the study period.

In sensitivity analysis, however, we as-sessed the outcomes of patients accord-ing to whether their hospitals wereCOEs at the time of their procedures.As in our main analysis, there was nosignificant difference in rates of seri-ous complications between COE andnon-COE hospitals (adjusted OR, 1.15;95% CI, 0.92-1.43; P=.22).

When considering undergoing bar-iatric surgery,patientsshould weigh therisks and benefits of the various treat-ment options. Although we cannot yetreporton therelative effectiveness of dif-ferent bariatric procedures, prior stud-

ies suggest that weight loss is greaterwith procedures that combine both re-strictive and malabsorptive elementsthan in purely restrictive proceduressuch as the laparoscopicadjustable gas-tricband.18-24 In thefuture, ourstudy willbe able to provide information regard-ing the relative effectiveness of these dif-ferent procedures with regard to weightloss, weight-related comorbidity reso-lution, latecomplications, quality of life,patient satisfaction, and health care re-source utilization.

Our study findings may not be gen-

eralizable outside of the state of Michi-gan. These results reflect the out-comes among bariatric surgery centersthat participate in a statewide collabo-rative quality improvement initiative.The extent to which this program, stillin its first few years of existence, ex-plains the relatively low rates of seri-ous complicationsamong bariatrichos-pitals and surgeons in the state isunknown.However, these efforts go be-yond data feedback, requiring the ac-tive participation of bariatric surgeons

in quality improvement initiatives andmandatory attendance at collabora-tive meetings held 3 times each year.For this reason, we believe that the re-sults reportedin thisstudy represent theoutcomes of bariatric surgery that arepossible, but not necessarily those thatare typical in community settings.

In conclusion, the frequency of se-rious complications among patients af-ter bariatric surgery in Michigan is low.Rates of serious complications are in-versely associatedwith hospital andsur-

geon procedure volume but not COEstatus. These data may serve as usefulsafetyperformance benchmarksfor hos-pitals performing bariatric surgery. Wehope that they might also inform thedebate about the effectiveness of vari-ous approaches to ensuring high-quality care for bariatric surgery pa-tients.

Author Contributions: Dr N. Birkmeyer had full ac-cess to all of the data in the study and takes respon-sibility for the integrity of the data and the accuracyof the data analysis.Study concept and design: N. Birkmeyer, Dimick,Hawasli, English, Genaw, Finks, Carline, J. Birkmeyer.Acquisition of data: N. Birkmeyer, J. Birkmeyer.Analysis and interpretation of data: N. Birkmeyer,Dimick.

Drafting of the manuscript: N. Birkmeyer, Dimick,J. Birkmeyer.Critical revision of the manuscript for important in-tellectual content: N. Birkmeyer, Dimick, Share,Hawasli, English, Genaw, Finks, Carlin, J. Birkmeyer.Statistical analysis: N. Birkmeyer, Dimick.Obtained funding: N. Birkmeyer.Study supervision: N. Birkmeyer.Financial Disclosures: DrShare reports thathe isem-ployed part time by Blue Cross Blue Shield of Michi-ganas executive medicaldirector for Healthcare Qual-ity. DrCarlinreportsthat heis a member ofthe EthiconEndo-Surgery Provider Advisory Council. The otherau-thors report no disclosures.Funding/Support: TheMichigan Bariatric SurgeryCol-laborative is funded by an unrestricted longitudinalre-search contract from Blue Cross and Blue Shield ofMichigan and Blue Care Network. Drs N. Birkmeyerand Dimick also receive support for research relatedto the Michigan Bariatric Surgery Collaborative fromthe Agency for Healthcare Research and Quality(1R01HS018050 and 1R01HS018728) and the Na-tional Institute of Diabetes and Digestive and KidneyDiseases (1R21DK084397).Michigan Bariatric Surgery Collaborative: Individualmembers andmember institutions areavailableat http://www.med.umich.edu/mscore/mbsc.html.Role of the Sponsor: Blue Cross and Blue Shield ofMichigan andthe Agencyfor HealthcareQuality andResearch had no roles in the design and conduct ofthis study; collection, management, analysis, and in-terpretation of the data; or preparation of the manu-script.

REFERENCES

1. Santry HP, Gillen DL, Lauderdale DS. Trends in bar-iatric surgical procedures. JAMA. 2005;294(15):

1909-1917.2. Encinosa WE, Bernard DM, Du D, Steiner CA. Re-cent improvements in bariatricsurgery outcomes.MedCare. 2009;47(5):531-535.3. Courcoulas A, SchuchertM, Gatti G, Luketich J. Therelationship of surgeon and hospital volume tooutcome after gastric bypass surgery in Pennsylva-nia: a 3-year summary. Surgery. 2003;134(4):613-621.4. FlumDR, Dellinger EP.Impact of gastric bypass op-eration on survival:a population-based analysis.J AmColl Surg. 2004;199(4):543-551.5. Liu JH, Zingmond D, Etzioni DA, et al. Character-izing the performance and outcomes of obesity sur-gery in California. Am Surg. 2003;69(10):823-828.6. Livingston EH. Bariatric surgery outcomes at des-ignated centers of excellence vs nondesignatedprograms. Arch Surg. 2009;144(4):319-325.

7. Murr MM, Martin T, Haines K, et al. A state-widereview of contemporary outcomes of gastric bypassin Florida: does provider volume impact outcomes?Ann Surg. 2007;245(5):699-706.8. Weller WE, Hannan EL. Relationship between pro-vider volume andpostoperative complicationsfor bar-iatric procedures in New York State. J Am Coll Surg.2006;202(5):753-761.9. BirkmeyerNJ, BirkmeyerJD. Strategies for improv-ing surgical qualityshould payers reward excel-lence or effort? N Engl J Med. 2006;354(8):864-870.





8/8

10. Birkmeyer NJ, Share D, Campbell DA Jr, PragerRL, Moscucci M, Birkmeyer JD. Partnering with pay-ers to improve surgical quality: the Michigan plan.Surgery. 2005;138(5):815-820.11. Hofer TP,HaywardRA, Greenfield S, Wagner EH,Kaplan SH, Manning WG. The unreliability of indi-vidual physicianreport cards for assessingthe costsandquality of care of a chronic disease.JAMA. 1999;281(22):2098-2105.

12. Flum DR, Belle SH, King WC, et al; LongitudinalAssessment of Bariatric Surgery (LABS) Consortium.Perioperative safetyin the longitudinal assessment ofbariatric surgery. N Engl J Med. 2009;361(5):445-454.13. Nguyen NT, Paya M, Stevens CM, MavandadiS, Zainabadi K, Wilson SE. The relationship betweenhospital volume and outcome in bariatric surgery atacademic medical centers. Annals of Surgery. 2004;240(4):586-593.14. Zingmond DS, McGory ML, Ko CY. Hospitaliza-

tion before and after gastric bypass surgery. JAMA.2005;294(15):1918-1924.15. Latts L, Singer J. Centers of excellence. ArchSurg.2010;145(1):106-108.16. Pratt GM. Designation does matter. Arch Surg.2010;145(1):105.17. Shikora SA, Wolfe B, Schirmer B. Bariatric cen-ters of excellence programs do improve surgicaloutcomes. Arch Surg. 2010;145(1):105-106.

18. Biertho L, Steffen R, Ricklin T, et al. Laparo-scopic gastric bypass versus laparoscopicadjustable gas-tricbanding: a comparative study of 1,200 cases.J AmColl Surg. 2003;197(4):536-545.19. CottamDR, AtkinsonJ, AndersonA, GraceB, FisherB. A case-controlled matched-pair cohort studyof lapa-roscopic Roux-en-Y gastric bypass and Lap-Band pa-tientsin a singleUS centerwith three-year follow-up.Obes Surg. 2006;16(5):534-540.20. Jan JC, Hong D, Pereira N, Patterson EJ. Laparo-scopic adjustable gastric banding versus laparoscopic

gastric bypass for morbid obesity: a single-institutioncomparison study of earlyresults.J Gastrointest Surg.2005;9(1):30-41.21. Kim T, Daud A, Ude A, et al. Early US outcomesof laparoscopic gastric bypass versus laparoscopic ad-justable silicone gastric banding for morbid obesity.Surg Endosc. 2006;20(2):202-209.22. te Riele WW, VogtenJM, BoermaD, Wiezer MJ,vanRamshorst B. Comparison of weight lossand mor-

bidity after gastric bypass andgastricbanding: a singlecenter European experience. Obes Surg. 2008;18(1):11-16.23. Weber M, Mller MK, Bucher T, et al. Laparo-scopic gastric bypass is superior to laparoscopic gas-tricbandingfor treatment of morbidobesity. AnnSurg.2004;240(6):975-983.24. Tice JA,Karliner L, Walsh J, Petersen AJ,FeldmanMD. Gastric banding or bypass? a systematic reviewcomparingthe two mostpopularbariatric procedures.Am J Med. 2008;121(10):885-893.




birkmayer 2011 bariatric surg complications in michigan

Documents