Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced...

9
Determinants of Diabetes Remission and Glycemic Control After Bariatric Surgery DOI: 10.2337/dc15-0575 OBJECTIVE Eligibility criteria for bariatric surgery in diabetes include BMI 35 kg/m 2 and poorly controlled glycemia. However, BMI does not predict diabetes remission, and thus, predictors need to be identied. RESEARCH DESIGN AND METHODS Seven hundred twenty-seven patients were included in a database merged from the Swedish Obese Subjects (SOS) study and two randomized controlled studies, with 415 surgical and 312 medical patients in total. Bariatric operations were divided into gastric only (GO) and gastric plus diversion (GD). RESULTS Sixty-four percent of patients in the surgical arm and 15.0% in the medical arm experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission. The best predictors of diabetes remission were lower baseline glycemia and shorter diabetes duration. However, when operation type was consid- ered, GD predicted a higher likelihood of remission and greater weight loss. Patients in remission (responders) lost more weight (25% vs. 17%) and waist circumference (18% vs. 13%) and experienced better insulin sensitivity than nonresponders. CONCLUSIONS Surgery is more effective than medical treatment in achieving diabetes remission and tighter glycemic control. Shorter diabetes duration, lower fasting glycemia before surgery, and GD versus GO procedures independently predict higher rates of remission, whereas baseline HbA 1c and waist circumference predict improved glycemic control. The results show the advantage of an early operation together with better controlled glycemia on diabetes remission independently of BMI. The prospective, controlled Swedish Obese Subjects (SOS) study showed that the short- term rate of remission of diabetes after bariatric surgery is 72% compared with 21% in subjects treated with conventional weight loss methods (1). Randomized controlled studies have revealed diabetes remission rates between 37% and 42% (2) to 73% (3), 24% and 38% (4), and 75% and 95% (5), depending on the criteria used to dene diabetes remission and baseline patient characteristics, choice of surgical therapy, and duration of follow-up (13 years). Unfortunately, such differences in diabetes remission criteria obscure important questions regarding the determinants of diabetes remission, which in turn affect eligibility criteria for patient selection for bariatric surgery. Historically, only subjects with BMI $35 kg/m 2 in the presence of uncontrolled diabetes were considered as potential candidates for bariatric and metabolic 1 CNR-Institute for Systems Analysis and Com- puter Science (IASI), BioMatLab, Rome, Italy 2 Institute of Medicine, University of Gothenburg, Gothenburg, Sweden 3 Department of Internal Medicine, Catholic University, Rome, Italy 4 Department of Diabetes and Nutritional Sciences, Kings College London, London, U.K. 5 Obesity Research Unit, Department of General Practice, and Baker IDI Heart and Diabetes Institute, Monash University, Melbourne, Victoria, Australia Corresponding author: Geltrude Mingrone, [email protected]. Received 23 March 2015 and accepted 13 August 2015. This article contains Supplementary Data online at http://care.diabetesjournals.org/lookup/ suppl/doi:10.2337/dc15-0575/-/DC1. © 2015 by the American Diabetes Association. Readers may use this article as long as the work is properly cited, the use is educational and not for prot, and the work is not altered. Simona Panunzi, 1 Lena Carlsson, 2 Andrea De Gaetano, 1 Markku Peltonen, 2 Toni Rice, 3 Lars Sj¨ ostr¨ om, 2 Geltrude Mingrone, 3,4 and John B. Dixon 5 Diabetes Care 1 IDF-ADA TRANSLATIONAL SYMPOSIUM Diabetes Care Publish Ahead of Print, published online December 1, 2015

Transcript of Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced...

Page 1: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

Determinants of DiabetesRemission and Glycemic ControlAfter Bariatric SurgeryDOI: 10.2337/dc15-0575

OBJECTIVE

Eligibility criteria for bariatric surgery in diabetes include BMI ‡35 kg/m2 andpoorly controlled glycemia. However, BMI does not predict diabetes remission,and thus, predictors need to be identified.

RESEARCH DESIGN AND METHODS

Seven hundred twenty-seven patients were included in a database merged fromthe Swedish Obese Subjects (SOS) study and two randomized controlled studies,with 415 surgical and 312 medical patients in total. Bariatric operations weredivided into gastric only (GO) and gastric plus diversion (GD).

RESULTS

Sixty-four percent of patients in the surgical arm and 15.0% in the medical armexperienced diabetes remission (P < 0.001). GO yielded 60% remission, and GDyielded 76% remission. The best predictors of diabetes remission were lower baselineglycemia and shorter diabetes duration. However, when operation type was consid-ered, GD predicted a higher likelihood of remission and greater weight loss. Patientsin remission (responders) lost more weight (25% vs. 17%) and waist circumference(18% vs. 13%) and experienced better insulin sensitivity than nonresponders.

CONCLUSIONS

Surgery is more effective than medical treatment in achieving diabetes remissionand tighter glycemic control. Shorter diabetes duration, lower fasting glycemiabefore surgery, and GD versus GO procedures independently predict higher ratesof remission, whereas baseline HbA1c and waist circumference predict improvedglycemic control. The results show the advantage of an early operation togetherwith better controlled glycemia on diabetes remission independently of BMI.

The prospective, controlled Swedish Obese Subjects (SOS) study showed that the short-term rate of remission of diabetes after bariatric surgery is 72% compared with 21% insubjects treated with conventional weight loss methods (1). Randomized controlledstudies have revealed diabetes remission rates between 37% and 42% (2) to 73% (3),24%and38% (4), and75%and95% (5), depending on the criteria used todefinediabetesremission andbaseline patient characteristics, choice of surgical therapy, andduration offollow-up (1–3 years). Unfortunately, such differences in diabetes remission criteriaobscure important questions regarding the determinants of diabetes remission, whichin turn affect eligibility criteria for patient selection for bariatric surgery.Historically, only subjects with BMI $35 kg/m2 in the presence of uncontrolled

diabetes were considered as potential candidates for bariatric and metabolic

1CNR-Institute for Systems Analysis and Com-puter Science (IASI), BioMatLab, Rome, Italy2Institute of Medicine, University of Gothenburg,Gothenburg, Sweden3Department of Internal Medicine, CatholicUniversity, Rome, Italy4Department of Diabetes and NutritionalSciences, King’s College London, London, U.K.5Obesity Research Unit, Department of GeneralPractice, and Baker IDI Heart and DiabetesInstitute,MonashUniversity,Melbourne, Victoria,Australia

Corresponding author: Geltrude Mingrone,[email protected].

Received 23March 2015 and accepted 13 August2015.

This article contains Supplementary Data onlineat http://care.diabetesjournals.org/lookup/suppl/doi:10.2337/dc15-0575/-/DC1.

© 2015 by the American Diabetes Association.Readersmayuse this article as longas thework isproperly cited, the use is educational and not forprofit, and the work is not altered.

Simona Panunzi,1 Lena Carlsson,2

Andrea De Gaetano,1 Markku Peltonen,2

Toni Rice,3 Lars Sjostrom,2

Geltrude Mingrone,3,4 and John B. Dixon5

Diabetes Care 1

IDF-A

DATR

ANSLA

TIONALSYM

POSIU

M

Diabetes Care Publish Ahead of Print, published online December 1, 2015

Page 2: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

surgery (6). The International DiabetesFederation statement of 2011 (7) madeadvances on this criterion, stating thatsurgery should be prioritized for pa-tients with type 2 diabetes (T2D) andBMI .40 kg/m2 and indicated forpatientswithBMI 35–40 kg/m2 andpoorlycontrolled diabetes (HbA1c .7.5%) de-spite fully optimized conventional ther-apy. The statement also made provisionsforpatientswithBMI30–35kg/m2, statingthat these patients should be consideredfor surgery if the target HbA1c (,7.5%) isnot achieved with conventional medicaltherapy and lifestyle modification at-tempts and if other obesity-related con-cerns are present. Therefore, according tocurrent best practice, in the presence ofpoorly controlled diabetes, BMI appearsto represent theonly criterion for bariatricsurgery eligibility. However, no direct evi-dence demonstrates that BMI is a predic-tor of diabetes remission after bariatricsurgery. In fact, the use of BMI as a se-lection condition for bariatric surgeryfor T2D treatment in obese patients(BMI .34 kg/m2) has been shown notto predict outcome (5), thus challengingthe strict BMI criteria. Indeed, a meta-analysis (8) investigating the risks andbenefits associated with surgical andmedical therapies for treating diabetesor impaired glucose tolerance in pa-tients with BMI 30–35 kg/m2 showedthat surgery is more effective thanmedical therapy.The present investigation was de-

signed primarily to calculate the 2-yeardiabetes remission rate after bariatricsurgery compared with medical treat-ment and to assess the predictors of re-mission. Diabetes remission criteriawere standardized as fasting plasmaglucose ,5.6 mmol/L without pharma-cological treatment. We also examinethe proportion of surgical and medicalpatients who achieved tight glycemiccontrol at 2 years.

RESEARCH DESIGN AND METHODS

The database from the SOS study (1) wasmerged with databases of two random-ized controlled studies (3,5), yielding acombined group of patients with T2Dwho underwent conventional medicaltherapy and a combined group who un-derwent bariatric surgery. The lifestyleinterventions in the medical arm dif-fered among the three centers involvedin the study. For the SOS study, no

attempt to standardize the recommen-dations among the Swedish recruitingcenters was made. In the study by Dixonet al. (3), patients were asked to reduceglobal energy intake, particularly of fat to,30% (including saturated fats), and en-couraged to consume low–glycemic indexand high-fiber foods. Physical activity of10,000 steps/day and 200 min/week ofstructured activity was also encouraged,including moderate-intensity aerobic ac-tivity and resistance exercise. Suggestionssimilar to those adopted in the study byDixon et al. were made in the study byMingrone et al. (5).

Baseline and 2-year characteristics ofsubjects in the medical and surgicalarms are summarized in Table 1. To har-monize the results from the three stud-ies, diabetes remission was defined asnormal fasting glycemia (,5.6 mmol/L)in the absence of any antidiabetic med-ication. Hereafter, patients with diabetesremission are identified as responders.

From the available information aboutmedication usage, a categorical variablewith three levels (0 = no medication, 1 =onlyoralhypoglycemicagents, 2= insulinor insulin plus oral hypoglycemic agents)was built. The variables analyzed weresex, age, weight, BMI, waist circumfer-ence, diabetes duration, type of antidia-betic medication, fasting concentrationsof glucose and insulin, HOMA of insulinresistance (IR), total cholesterol, HDL cho-lesterol, LDL cholesterol, triglycerides,and systolic blood pressure (SBP) and di-astolicbloodpressure (DBP).HbA1cwasalsoavailable from the study by Dixon et al. (3)and the study by Mingrone et al. (5).

Overall, the three studies pooled in-cluded four different surgical techniques:vertical banded gastroplasty (VBG), gastricbanding (GB, either adjustable or non-adjustable), Roux-en-Y gastric bypass(RYGB), and biliopancreatic diversion(BPD).We compared thepercent variationfrom baseline at 2 years of weight, BMI,and waist circumference (SupplementaryTable 1). For these variables, post hoccomparisons after ANOVA with twobetween-factor components (respondersvs. nonresponders and surgical techniquewith four levels [Supplementary Table 2])showed no significant differences be-tween VBG and GB, the two procedureswith purely gastric components (gastriconly [GO]), and between RYGB and BPD,the two techniques that divert gastriccontent distally into the small intestine

(gastric with diversion [GD]). Conse-quently, the patients were divided intotwo surgical therapy groups, and analyseswere performed comparing the two dif-ferent macrotechniques GO and GD.

To predict diabetes remission, a logis-ticmodel was used. Each aforementionedvariable was initially entered in a univar-iate logistic regression to study its abilityto independently predict diabetes remis-sion. Only predictors significantly associ-ated with remission in the univariateanalysis were entered into a multivariatemodel, and their ability to independentlypredict remission was assessed by a step-wise elimination method.

The percentage of patients attainingfasting glycemia ,7 mmol/L or HbA1c,7% at 2 years was calculated as repre-senting those achieving tight glycemiccontrol. Pearson coefficients were usedto study correlations among variables.Dependence of continuous variables(changes in glycemia, HbA1c, and HOMA-IR) on predictors was studied bymultivar-iable linear models. For each variable atbaseline and for percent change at 2years, ANOVAs (with the two aforemen-tioned between-factor comparisons)were performed. Receiver operatingcharacteristic (ROC) analyses wereused to identify the most appropriatecutoffs of the variables of interest aspredictors of diabetes remission at 2years. Multivariate ROC analysis wasalso performed to study the predictivecapability, in terms of area under thecurve (AUC), of the best variable com-bination in estimating the probability ofremission, for which the best cutoffvalue is also reported (lr.h).

RESULTS

Sample DescriptionIn total, 727 patients with T2D were an-alyzed, 415 of whom were surgicallytreated. There were 227 cases of VBG,91 of GB, 77 of RYGB, and 20 of BPD. Thedropout rate at 2 years was 10.4% in thesurgical group and 19.9% in the medicalgroup. The averagediabetes durationwas0.76 1.9, 4.16 3.3, and 9.86 6.5 yearsin patients not taking medications, takingoral agents only, or taking insulin or in-sulin plus oral agents, respectively.

Comparison Between Medical andSurgical PatientsIn the medical arm, 15% of the patientsachieved diabetes remission compared

2 Bariatric Surgery Effects on Diabetes Diabetes Care

Page 3: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

with 63.7% in the surgical arm (x2 test P,0.001). Remission in the surgical groupswas 61%, 59%, and 92% in the SOS (1),the study by Dixon et al. (3), and the studybyMingrone et al. (5), respectively; the x2

test for the associationbetween study anddiabetes remission was significant (P =0.001). This association is likely due tothe association between diabetes remis-sion and surgical technique (P = 0.003). Allpatients experienced remission in the BPDgroup (19 of 19 in the study by Mingroneet al.), 69% in the GBY group (47 of 68),59% in the VBG group (116 of 198), and63% in the GB group (55 of 87). The asso-ciation between study and surgical tech-nique was significant (P, 0.001).

Table 1 reports the comparisons be-tweenmedical and surgical treatment atbaseline and 2 years after enrollment interms of both absolute values and per-cent changes. Patients in the medicalarm presented at baseline with loweraverage weight, waist circumference,BMI, SBP, and DBP than surgical patients(P, 0.001 for all comparisons). Weight,waist circumference, BMI, glycemic con-trol, lipid profile, and blood pressurewere substantially improved after sur-gery. Supplementary Table 3 reportsthe P values from t tests related to thecomparisons between medical therapyand each surgical procedure type.

Seventy-six percent of patients in theGD group and 60% in the GO group expe-rienced diabetes remission (P = 0.016).Eighty-five percent of the patients in theGD group and 78% in the GO groupachieved tight glycemic control (fastingglycemia,7 mmol/L) without pharmaco-logical therapy; these figures increased to91% and 88%, respectively, with pharma-cological treatment. The number of pa-tients with HbA1c ,7% (HbA1c dataavailable in a subset of 67 patients) with-out pharmacological therapy was similar(90% vs. 89%) in the two groups. Therewere no statistically significant differencesbetween surgical groups. In the medicalgroup, independently of pharmacologicaltreatment, the percentage of patientswith fasting glycemia ,7 mmol/L was38% and that with HbA1c ,7%, 38.6%.

ANOVA for Responders andNonrespondersTable 2 reports baseline and end-of-study values of the putative predictivecharacteristics. To test whether baselineconditions were different between

Table

1—Varia

blesatbase

lineandafte

r2years

andperce

ntch

angein

themedica

ltre

atm

entandsu

rgica

larm

sindependentofthesu

rgical

tech

niqueuse

d

Baselin

e2years

Percentchange

Med

icaltreatmen

tSurgery

Med

icaltreatmen

tSurgery

Med

icaltreatmen

tSurgery

nMean

6SD

nMean

6SD

Pvalu

en

Mean

6SD

nMean

6SD

Pvalu

en

Mean

6SD

nMean

6SD

Pvalu

e

Age

(years)312

49.76

6.9410

48.06

6.4,0.001

dd

dd

dd

dd

dd

Weigh

t(kg)

312116.6

617.8

415122.9

620.1

,0.001

251113.7

618.4

37394.3

617.6

,0.001

25122.5

66.5

373222.4

610.8

,0.001

Waist

circumferen

ce(cm

)312

122.36

11.2413

127.86

13.0,0.001

251119.8

610.9

371106.7

614.3

,0.001

25121.3

611.9

369216.0

69.1

,0.001

BMI(kg/m

2)312

40.16

4.9415

42.06

5.0,0.001

25139.0

64.9

37332.4

65.0

,0.001

25122.5

66.5

373222.4

610.8

,0.001

Glycem

ia(m

mol/L)

3128.8

62.8

4148.7

62.7

NS

2508.2

62.9

3725.2

62.0

,0.001

25021.2

633.9

371236.4

623.1

,0.001

Insulinem

ia(m

U/L)

31125.1

617.3

41027.4

618.3

NS

25021.9

614.8

37212.4

67.7

,0.001

2491.3

651.6

368247.4

631.0

,0.001

HbA1c(%

)50

8.06

1.470

8.36

1.5NS

447.2

61.0

675.8

61.0

,0.001

4426.0

614.3

67229.3

616.8

,0.001

HOMA-IR

3119.7

67.5

40910.6

67.6

NS

2508.0

65.8

3723.1

63.1

,0.001

24920.1

656.6

367264.9

629.5

,0.001

SBP(m

mHg)

282144.4

620.4

384150.6

620.0

,0.001

225142.0

617.6

341139.7

621.6

NS

22521.6

612.0

34026.8

613.7

,0.001

DBP(m

mHg)

28287.8

612.0

38492.0

611.5

,0.001

22585.3

69.6

34184.7

611.0

NS

22522.1

611.7

34027.3

611.9

,0.001

HDLcholestero

l(mmol/L)

2931.2

60.3

3901.2

60.3

NS

2391.3

60.3

3641.5

60.4

,0.001

2304.5

615.4

34425.6

681.0

,0.001

Totalch

olestero

l(mmol/L)

3125.7

61. 2

4145.8

61.2

NS

2505.4

61.2

3725.4

61.3

NS

25021.7

615.4

37125.0

620.4

0.030

Triglycerides

(mmol/L)

3122.9

62.9

4142.8

62.1

NS

2492.4

61.5

3721.7

61.0

,0.001

2490.6

644.2

371229.7

632.7

,0.001

LDLcholestero

l(mmol/L)

2933.2

61.2

3903.3

61.0

NS

2393.1

60.9

3643.2

61.1

NS

23022.3

623.7

34421.5

636.9

NS

Diab

etesduratio

n(years)

2823.7

64.6

3843.3

64.9

NS

Pvalu

esfro

mttest.

care.diabetesjournals.org Panunzi and Associates 3

Page 4: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

Table 2—Baseline and 2-year values overall and by type of surgical procedure in patients who did or did not achieve diabetesremission

Overall GO GD

Variable RESP n Mean 6 SD n Mean 6 SD n Mean 6 SD

BaselineAge (years) 0 135 48.7 6 6.1 114 49.4 6 5.8 21 45.4 6 6.8

1 234 47.7 6 6.5 171 48.3 6 6.0 63 46.1 6 7.6Total 369 48.10 6 6.4 285 48.7 6 5.9 84 45.9 6 7.3

Weight (kg) 0 135 118.77 6 15.6 114 118.1 6 15.4 21 122.2 6 16.91 237 124.15 6 21.9 171 121.8 6 20.6 66 130.2 6 23.9

Total 372 122.20 6 20.0 285 120.3 6 18.8 87 128.2 6 22.6Waist circumference (cm) 0 135 126.29 6 11.0 114 126.1 6 10.6 21 127.3 6 13.6

1 235 127.92 6 14.0 169 127.5 6 13.6 66 129.1 6 15.1Total 370 127.33 6 13.0 283 126.9 6 12.4 87 128.6 6 14.7

BMI (kg/m2) 0 135 40.99 6 4.0 114 40.7 6 39 21 42.8 6 4.31 237 42.48 6 5.3 171 41.8 6 5.0 66 44.1 6 5.7

Total 372 41.94 6 4.9 285 41.4 6 4.6 87 43.8 6 5.4Glycemia (mmol/L) 0 134 9.71 6 2.8 113 9.6 6 2.6 21 10.4 6 3.4

1 237 8.10 6 2.4 171 7.9 6 2.1 66 8.6 6 3.1Total 371 8.68 6 2.7 284 8.6 6 2.5 87 9.1 6 3.2

Insulinemia (mU/L) 0 133 26.19 6 15.0 112 26.8 6 15.9 21 22.8 6 9.21 235 26.72 6 13.1 170 28.4 6 13.8 65 22.4 6 10.0

Total 368 26.53 6 13.8 282 27.8 6 14.7 86 22.5 6 9.7HbA1c (%) 0 15 8.02 6 0.8 12 7.9 6 0.9 3 8.3 6 0.6

1 52 8.47 6 1.6 17 7.7 6 1.4 35 8.8 6 1.6Total 67 8.37 6 1.5 29 7.8 6 1.2 38 8.8 6 1.5

HOMA-IR (mmol/L ∙ mU/L / 22.5) 0 132 11.41 6 7.8 111 11.7 6 8.3 21 10.1 6 4.71 235 9.56 6 5.5 170 9.9 6 5.7 65 8.5 6 4.7

Total 367 10.22 6 6.5 281 10.6 6 6.9 86 8.9 6 4.7SBP (mmHg) 0 123 152.63 6 20.8 102 152.1 6 19.9 21 155.3 6 25.2

1 219 149.78 6 20.0 153 148.7 6 18.6 66 152.4 6 22.8Total 342 150.81 6 20.3 255 150.0 6 19.2 87 153.1 6 23.2

DBP (mmHg) 0 123 91.71 6 11.3 102 91.8 6 10.4 21 91.2 6 15.21 219 92.27 6 11.9 153 92.0 6 11.2 66 92.8 6 13.4

Total 342 92.07 6 11.6 255 91.9 6 10.9 87 92.4 6 13.7HDL cholesterol (mmol/L) 0 128 1.24 6 0.3 107 1.2 6 0.3 21 1.3 6 0.3

1 220 1.22 6 0.3 155 1.2 6 0.3 65 1.2 6 0.3Total 348 1.23 6 0.3 262 1.2 6 0.3 86 1.2 6 0.3

Total cholesterol (mmol/L) 0 135 5.72 6 1.2 114 5.8 6 1.2 21 5.4 6 1.11 236 5.84 6 1.3 170 6.0 6 1.3 66 5.4 6 1.3

Total 371 5.80 6 1.3 284 5.9 6 1.2 87 5.4 6 1.2LDL cholesterol (mmol/L) 0 128 3.2 6 1.0 107 3.2 6 1.0 21 3.1 6 1.0

1 220 3.4 6 1.0 155 3.5 6 1.0 65 3.2 6 1.2Total 348 3.3 6 1.0 262 3.4 6 1.0 86 3.2 6 1.1

Triglycerides (mmol/L) 0 135 2.73 6 1.4 114 2.8 6 1.5 21 2.2 6 0.81 236 2.82 6 2.5 170 3.0 6 2.6 66 2.3 6 1.9

Total 371 2.79 6 2.1 284 2.9 6 2.2 87 2.2 6 1.70 122 6.21 6 6.7 102 5.3 6 5.7 20 10.8 6 9.2

Diabetes duration (years) 1 220 1.78 6 2.6 154 1.0 6 1.9 66 3.6 6 2.9Total 342 3.36 6 4.9 256 2.7 6 4.4 86 5.3 6 5.9

2 years after surgeryWeight (kg) 0 135 98.5 6 17.2 114 99.8 6 17.5 21 91.5 6 13.3

1 236 91.9 6 17.5 170 93.6 6 17.4 66 87.5 6 17.1Total 371 94.3 6 17.6 284 96.0 6 17.7 87 88.5 6 16.3

Waist circumference (cm) 0 134 110.2 6 14.3 113 111.7 6 14.3 21 102.0 6 11.61 236 104.6 6 13.9 170 106.1 6 13.0 66 100.8 6 15.2

Total 370 106.6 6 14.3 283 108.3 6 13.8 87 101.1 6 14.4BMI (kg/m2) 0 135 34.0 6 5.2 114 34.3 6 5.2 21 32.2 6 4.6

1 236 31.4 6 4.6 170 32.1 6 4.5 66 29.7 6 4.3Total 371 32.4 6 5.0 284 33.0 6 4.9 87 30.3 6 4.5

Glycemia (mmol/L) 0 135 6.9 6 2.6 114 6.8 6 2.3 21 7.2 6 3.81 237 4.3 6 0.6 171 4.4 6 0.6 66 4.2 6 0.7

Total 372 5.2 6 2.0 285 5.3 6 1.9 87 4.9 6 2.3

Continued on p. 5

4 Bariatric Surgery Effects on Diabetes Diabetes Care

Page 5: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

surgical techniques and whether re-sponders and nonresponders were differ-ent at baseline, an ANOVA with twobetween-factor components (remissionand surgical procedure) was performedfor each baseline variable (Supplemen-tary Table 4). Patients allocated to thediversionary (GD) procedures were youn-ger andmore obese and presentedwith amore compromised glycemic control,including a significantly lower baselineinsulinemia; higher, though not statisti-cally significant, glycemia; higher HbA1c,and longer diabetes history. The lowerHOMA-IR score in the GD group (8.9 64.7 vs. 10.6 6 6.9) might be ascribed tolower fasting plasma insulin levels. Base-line cholesterol and triglyceride levelswere also lower in the GD group.All interaction terms except diabetes

duration (P = 0.015) were nonsignifi-cant. Nonresponders in the GD grouphad a longer history of diabetes thanthose in the GO group.The Supplementary Data report the

mean values and SDs of the variablesby surgical procedure at baseline (Sup-plementary Table 5) and 2 years aftersurgery (Supplementary Table 6) and

the percent changes (SupplementaryTable 1). Greater effects on weightloss, fat distribution, glycemia, HbA1c,and HOMA-IR were observed afterRYGB and BPD; in contrast, the type ofbariatric surgery did not affect bloodpressure outcomes. Supplementary Ta-ble 7 reports the P values related toANOVA with variables at baseline andwith between-factor remission and surgi-cal techniques at four levels. Respondershad lower glycemia, HOMA-IR, and diabe-tes duration than nonresponders.

Predictors of Diabetes RemissionResults of the logistic multivariable mod-els are reported in Table 3. Separate anal-yses were conducted overall, for themedical group, for the surgical group,and separately for the GO and GD groups.In the total population at baseline, youn-ger age, shorter diabetes duration, lowerfasting glycemia, and no antidiabetic drugspredicted a higher probability of diabetesremission at 2 years. However, age did notremain a significant predictor in the sub-group analyses. In the surgical group, BMIwas a significant predictor but lost signifi-cancewhen type of surgery was taken into

account.When GOwas considered, diabe-tes duration, fasting glycemia, and therapyfor diabeteswere inversely correlatedwithdiabetes remission; in GD, only fasting gly-cemia was a significant predictor.

To better understand the role of BMI indetermining remission, a categorical vari-able stratifying patients into three degreesof obesity was calculated. The consideredBMI classes were BMI#35, 35,BMI#40,and BMI.40 kg/m2. Odds ratios were cal-culated considering BMI #35 kg/m2 asthe reference. Estimated risk for dia-betes remission was 2.9 (95% CI 1.43–5.80) for patients with BMI.40 kg/m2

(P = 0.003). No significant differencewas founded between classes BMI#35 and 35,BMI#40 kg/m2.

Criteria for Predicting RemissionTable 3 also reports results from the ROCanalyses. An independent ROC analysiswas performed for baseline fasting gly-cemia, diabetes duration, and BMI toprovide the best cutoff values for theprediction of remission independently ofthe surgical procedure performed. Thesecutoffs were BMI $44 kg/m2, baselineglycemia #7.65 mmol/L, and diabetesduration#1.5 years.

Table 2—Continued

Overall GO GD

Variable RESP n Mean 6 SD n Mean 6 SD n Mean 6 SD

Insulinemia (mU/L) 0 135 15.2 6 8.7 114 16.2 6 8.9 21 10.0 6 5.91 237 10.8 6 6.4 171 12.0 6 6.7 66 7.8 6 4.4

Total 372 12.4 6 7.7 285 13.7 6 7.9 87 8.3 6 4.9HbA1c (%) 0 15 6.7 6 1.6 12 6.2 6 0.8 3 8.6 6 2.5

1 52 5.5 6 0.7 17 5.8 6 0.5 35 5.4 6 0.7Total 67 5.8 6 1.0 29 5.9 6 0.7 38 5.6 6 1.3

HOMA-IR (mmol/L ∙ mU/L / 22.5) 0 135 4.9 6 4.3 114 5.2 6 4.6 21 3.0 6 1.61 237 2.1 6 1.3 171 2.4 6 1.4 66 1.5 6 1.0

Total 372 3.1 6 3.1 285 3.5 6 3.4 87 1.9 6 1.3SBP (mmHg) 0 123 143.2 6 23.1 102 143.3 6 22.1 21 142.6 6 28.0

1 217 137.7 6 20.6 151 140.2 6 22.5 66 132.1 6 13.7Total 340 139.7 6 21.6 253 141.4 6 22.4 87 134.6 6 18.6

DBP (mmHg) 0 123 85.6 6 11.6 102 86.4 6 10.9 21 82.0 6 14.21 217 84.2 6 10.6 151 85.1 6 11.5 66 82.2 6 7.8

Total 340 84.7 6 11.0 253 85.6 6 11.3 87 82.1 6 9.7HDL cholesterol (mmol/L) 0 133 1.4 6 0.3 112 1.4 6 0.3 21 1.5 6 0.3

1 230 1.5 6 0.4 165 1.5 6 0.37 65 1.5 6 0.4Total 363 1.5 6 0.4 277 1.5 6 0.35 86 1.5 6 0.4

Total cholesterol (mmol/L) 0 135 5.5 6 1.2 114 5.7 6 1.16 21 4.8 6 1.01 236 5.4 6 1.4 171 5.8 6 1.12 65 4.2 6 1.4

Total 371 5.4 6 1.3 285 5.7 6 1.14 86 4.4 6 1.3Triglycerides (mmol/L) 0 135 2.0 6 1.1 114 2.1 6 1.17 21 1.4 6 0.4

1 236 1.5 6 0.8 171 1.7 6 0.88 65 1.1 6 0.5Total 371 1.7 6 1.0 285 1.8 6 1.03 86 1.2 6 0.5

LDL cholesterol (mmol/L) 0 133 3.2 6 1.0 112 3.3 6 1.0 21 2.6 6 1.01 230 3.2 6 1.1 165 3.5 6 1.0 65 2.2 6 1.1

Total 363 3.2 6 1.1 277 3.4 6 1.0 86 2.3 6 1.0

RESP, responder (0 = nonresponder; 1 = responder).

care.diabetesjournals.org Panunzi and Associates 5

Page 6: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

Additionally, two multivariate ROCanalyses were performed on the proba-bilities determined from the two logisticmodels fitted to the surgical group. Thefirst logistic model included baselineBMI, glycemia, and diabetes duration,and the second included baseline glyce-mia, diabetes duration, type of surgery,and type of therapy. Figure 1A and Bshow the ROC curves of these multivar-iate ROC analyses. Columns G and H inTable 3 report an independent ROCanalysis separately for type of surgery,showing cutoffs for those variables iden-tified as significantly associated in themultivariate logistic regressions.

Improvement of Glucose Control WithRespect to Baseline ConditionsThe only baseline factors that signifi-cantly correlated with percent changein glycemia were glycemia, HbA1c, andHOMA-IR (P , 0.001). However, thepercent change in HbA1c, availablefrom the studies by Dixon et al. (3) andMingrone et al. (5), was significantlynegatively correlated with the followingbaseline factors: glycemia (P = 0.001),HbA1c (P = 0.001), weight (P = 0.001),waist circumference (P , 0.001), BMI(P = 0.01), total cholesterol (P = 0.02),and triglycerides (P , 0.001).

In a multivariate linear model withpercent change in HbA1c as the depen-dent variable, the significant predictorsincluded in the final model, after a step-wise elimination, were baseline waistcircumference, glycemia, and triglycer-ides (regression coefficient b = 20.31[P = 0.008], 21.46 [P = 0.02], and23.79 [P = 0.02], respectively). Whenonly the GO surgical technique was con-sidered, results were quite similar. Inthe GD group, percent change in HbA1ccorrelated negatively with baselinewaist circumference (P = 0.014), totalcholesterol (P = 0.013), and triglycerides(P = 0.006), whereas final HbA1c corre-lated only with total cholesterol (P =0.044).

Diabetes Remission, SurgicalTechniques, and Glucose ControlChangesFor the change in each variable (glyce-mia, HbA1c, etc.), an ANOVA (Supple-mentary Table 4) with between-factorcomparisons of remission and type ofsurgical technique was performed totest for differences between responders

Table

3—Predictionofdiabetesremission:logisticmodelsandROC

analyse

s

Overall

Med

ical*

SurgicaloverallC(n

=41

5)GO

GD

A(n

=72

7)B(n

=31

2)WithoutT.ofS.C1

WithT.

ofS.**

C2

D(n

=31

8)

E(n

=97

)

Variable

bPvalue

bPvalue

bPvalue

bPvalue

bPvalue

bPvalue

Logisticregressionafterstep

wiseelim

ination

Intercep

t3.71

,0.001

1.895

0.103

0.089

0.942

2.83

,0.001

3.27

,0.001

2.838

,0.001

Age

20.04

20.005

Fastingglycem

ia(m

mol/L)

20.11

80.004

20.306

0.051

20.145

0.004

20.141

0.011

20.186

0.006

Diabetes

duration

20.19

2,0.001

20.910

0.013

20.210

,0.001

20.220

,0.001

20.197

0.004

20.273

,0.001

BMI

0.059

0.036

T.ofS.

2.180

,0.001

Typeoftherapy(level1)

20.81

9,0.001

22.247

0.042

21.22

,0.001

21.364

,0.001

Typeoftherapy(level2)

20.28

20.473

0.832

0.582

21.13

0.054

21.783

0.03

Univariate

ROCanalysis

FG

H

Cutoff

(Sen

s.,Spec.)

AUC(95%

CI)

Cutoff

(Sen

s.,Spec.)

AUC(95%

CI)

Cutoff

(Sen

s.,Spec.)

AUC(95%

CI)

ROCanalysis

BMI

$44

(32%

,83%

)0.58(0.52–

0.64

)Fastingglycem

ia(m

mol/L)

#7.65

(59.1%

,73.1%

)0.69(0.63–

0.75

)#7.6(62.6%

,72.6%

)0.71(0.65–

0.77)

Diabetes

duration

#1.5(65.9%

,73.8%

)0.74(0.68–

0.79

)#1.5(77.9%

,70.6%

)0.78(0.72–

0.84)

#7.5(92.4%

,55%

)0.75(0.61–

0.90

)

Sens.,sen

sitivity;Spec.,specificity;T.ofS.,typeofsurgery(GOvs.G

D).*A

further

modelincludingpercentweightchan

gewas

tested

;resultswereb=20.41

(P=0.009

)forbasalglycem

ia,b

=20.53(P

,0.001

)fordiabetes

duration,andb=20.09

(P=0.005

).**When

typeofsurgerywas

included

inthemodel,B

MIwas

nolongersign

ificant.

6 Bariatric Surgery Effects on Diabetes Diabetes Care

Page 7: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

and nonresponders and between GOand GD procedures. All interaction termsexcept for HbA1c were not significant.

Weight Loss and Waist Circumference

Percent changes in weight were217.1 6 9.1% and 225.5 6 10.4% innonresponders and responders and220.0 6 10.0% and 230.5 6 9.1% inthe GO and GD groups (P , 0.001),and percent changes in waist circumfer-ence were 212.7 6 8.5% and 218.0 68.9% in nonresponders and respondersand214.46 8.8% and221.36 8.2% inthe GO and GD groups (P , 0.001), re-spectively. Diabetes remission is ex-plained by percent weight loss aftercontrolling for baseline variables. Afterapplying a multivariable logistic regres-sion model with a stepwise eliminationcriterion, the final model included per-cent change in weight, fasting glycemia,and diabetes duration, with P , 0.001for all variables.

Glycemic Control Improvement

As shown by ANOVA, two factors (re-sponders and type of operation) signifi-cantly affected the changes in glycemia,insulinemia, HbA1c, and HOMA-IR.Change in glycemia was smaller innonresponders than in responders(225.1 6 29.4% vs. 242.9 6 15.5%)and with GO versus GD technique(234.8 6 22.6% vs. 241.9 6 24.1%).Similar results were obtained for varia-tion in HbA1c (216.5 6 20.3% vs.

232.86 13.9% in nonresponders vs. re-sponders and 222.6 6 11.9% vs.234.1 6 18.4% for GO vs. GD, respec-tively). The interaction was highly signif-icant (P , 0.001), although in the GOgroup, nonresponders and respondersexperienced similar variations (222.0469.75% vs. 223.05 6 13.54%, respec-tively), and in the GD group, nonre-sponders experienced increased andresponders experienced strongly de-creased HbA1c values (5.64 6 37.9% vs.237.506 11.54%, respectively). Conse-quently, HOMA-IR followed the sametrend (249.3 6 38.3% vs. 273.7 617.9% in nonresponders and respondersand 261.4 6 31.8% vs. 276.5 6 15.4%for GO vs. GD, respectively).

Lipid Profile Improvement

Changes in lipid profile are summarizedin Fig. 2. Total cholesterol, LDL choles-terol, and triglycerides decreased moreafter the GD than after the GOprocedure;in addition, triglycerides decreased morein responders than in nonresponders. Ta-ble 2 reports the mean values at baselineand at 2 years by remission and type ofsurgical technique.

Glycemic Control Improvements andWeight LossThe influence of weight loss, waist cir-cumference reduction, and type of sur-gery on remission was also studied bymeans of logistic regression. The onlysignificant factor associatedwith remission

was percent change in weight (P ,0.001), which explained 12% of the var-iance; the same results were obtainedwhen change in HbA1c was related withpercent change in weight, percentchange in waist circumference, andtype of surgery in a multivariate linearmodel.

From the ROC analysis, a cutoff valueof percent weight change was identified(224%) with a specificity of 80% and asensitivity of 55%. AUC was 73.1% (95%CI 67.8–78.3%). Weight change was219.99 6 10.01% in the GO group and230.526 9.08% in the GD group. Whenthe four procedures were consideredseparately, the percentage of subjectswho experienced a weight change greaterthan the calculated cutoff was 24% forGB, 33% for VBG, 75% for RYGB, and79% for BPD.

Association Between Study Centerand Diabetes RemissionA Fisher exact test for the association be-tween study center and diabetes remissionwas significant (P, 0.001) likely becauseof the association between diabetes remis-sion and the surgical technique (P = 0.007)used in the different studies.

CONCLUSIONS

This study, based on a sample of 727patients of whom 415 underwent bar-iatric surgery, is the first large analysis tofocus on the best predictors of diabetes

Figure 1—Multivariate ROC curves from logistic regression in the overall sample of surgical patients with (A) or without (B) type of surgery aspredictor. Values related to lr.h refer to the best cutoff for probability of remission. diab_treat, diabetes treatment; diadur, diabetes duration; gb,glycemia at baseline; PV2, negative predictive value; PV+, positive predictive value; Sens, sensitivity; Spec, specificity; totrem, total remission;typerec, type of surgery.

care.diabetesjournals.org Panunzi and Associates 7

Page 8: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

remission after bariatric surgery in aCaucasian population and to highlightthe differences between so-called re-sponders and nonresponders. As a result,we have reinforced that bariatric sur-gery restores euglycemia in a significantnumber of subjects with diabetes, atleast in the short term. The major find-ings of the study are as follows:

1. Diabetes remission, defined as fast-ing glycemia ,5.6 mmol/L without

pharmacological therapy, occurredin 63.7% of the patients in the surgi-cal arm and 14.4% of those in themedical arm.

2. Ninety percent of surgically treatedpatients and ,40% of medicallytreated patients achieved tight gly-cemic control at 2 years.

3. GD procedures were more effectivethan GO procedures in terms of di-abetes remission (76% vs. 60%), theformer being associated with a higher

diabetes remission rate (not correlatedwith the use of insulin before surgery),even though diabetes duration wasmuch longer (7.5 vs. 1.5 years).

4. The “risk” of remission was not statis-tically significant between subjectswithBMI#35and35,BMI#40kg/m2.

In the surgical population as a wholeand in agreement with the literature (9–12), the longer the diabetes durationbefore bariatric surgery and the lowerusage of antidiabetic drugs, particularlyinsulin, the lower the diabetes remissionrate. The metric duration of the clinicaldiagnosis of diabetes together withthe type of medication used can beconsidered a surrogate measure of thepreservation of islet secretory capacity,suggesting that the more severe the di-abetes, the lower the remission rate.However, we emphasize that althoughduration of diabetes #1.5 years, fastingglycemia #7.6 mmol/L, and absence ofinsulin therapy predicted a better out-come after purely restrictive operations,these parameters did not influence theeffect of RYGB and BPD on diabetesremission, which also occurred whendiabetes duration was much longer (upto 7.5 years) and patients were receivinginsulin treatment and hadworse glycemiccontrol.

Overall, the best predictors of im-provement in glycemic control afterbariatric surgery were smaller waist cir-cumference, better-controlled diabetes,and lower triglyceride levels at baseline.In addition, responders lost moreweight and waist circumference afterbariatric surgery, had a greater reduc-tion in plasma triglycerides, and becamemore insulin sensitive.

Abdominal obesity (measured by thewaist circumference) is a more powerfulpredictor of T2D and cardiovascular dis-ease development than obesity itself(measured by BMI) (13–16). In anotherstudy, a waist circumference$94 cm inmiddle-aged men identified subjects atrisk for T2D and cardiovascular disease,with a sensitivity of 84.4% (17). Thesedata together with the finding that theremission rate was not statisticallydifferent in subjects with BMI 30–35 ver-sus 35–40 kg/m2 suggest that abdominalobesity metrics might be a betteranthropometric measure than BMIas an indicator of eligibility for bariatricsurgery.

Figure 2—Percent changes at 2 years with respect to baseline values of total cholesterol (A), LDLcholesterol (B), and triglycerides (C) in nonresponders and responders in the GO (open bars) andGD (solid bars) groups. Significant P values of the two factors (diabetes remission: responders vs.nonresponders; surgical procedure: GO vs. GD) are shown. NR, nonresponders; R, responders.

8 Bariatric Surgery Effects on Diabetes Diabetes Care

Page 9: Determinants of Diabetes Remission and Glycemic Control ...€¦ · 11/29/2015  · experienced diabetes remission (P < 0.001). GO yielded 60% remission, and GD yielded 76% remission.

A limitation of the present study isthat in the merged database, HbA1c val-ues were available only for the studiesby Dixon et al. (3) and Mingrone et al.(5), whereas the largest number of sub-jects were acquired from the SOS data-base (1), which started when themeasurement of glycated hemoglobinwas not routinely performed to monitorglycemic control. Dixon et al. (18)more recently demonstrated that Chi-nese patients with T2D from Taiwanshow a glycemic response to RYGB re-lated to baseline BMI, duration of diabe-tes, circulating fasting C-peptide levels,and the degree of weight loss. These re-sults in this relatively small Asian sample(154 subjects) are not fully confirmed inthe present investigation in a muchlarger sample of Caucasian subjectsand of various types of bariatric surgery.These differences may depend on eth-nicity; however, differences in the crite-ria used to define remission cannot bediscounted as a reason. Of note, in theAsian sample, patients with preoperativeBMI as low as 23 kg/m2 were included,and BMI ,27 kg/m2 predicted poor gly-cemic control (18,19), possibly indicatingthat BMI is only a clinically importantpredictive factor when considering pa-tients in the nonobese range.The present study based on baseline

and 2-year postoperative data was un-able to discriminate between early andlate effects of bariatric surgery; thus, wecannot infer the different mechanismsof action of GO versus GD operations.We found that weight loss represents amajor determinant of end-study glyce-mic control. GD procedures were muchmore effective than GO procedures inachieving diabetes remission, althoughit is not clear whether this is purely be-cause of the superior weight loss ac-hieved with the GD procedures.The current results regarding longer di-

abetes duration and prediction of diabe-tes remission (which is still possible afterGD procedures for disease duration aslong as 7.5 years) do not imply that it isuseful to wait a long time before undergo-ing bariatric surgery. The SOS study, infact, recently demonstrated that theshorter the diabetes duration, the lowerthe incidenceofdiabetes complications (20).In conclusion, the predictors of diabetes

remission vary in relation to the type ofbariatric surgery, whereas better glycemiccontrol after surgery is achieved in

individuals with smaller waist circumfer-ence at baseline. Furthermore, the effectof bariatric surgery on diabetes remis-sion seems to be independent of baselineBMI, suggesting that bariatric surgerycould be a therapeutic option for patientswith T2D and abdominal obesity, evenwith BMI between 30 and 35 kg/m2.

Duality of Interest. L.C. has received paymentsfor lectures from Johnson & Johnson. G.M. is amember of the board of Fractyl Labs, Inc. andhas received research grants from AstraZeneca.J.B.D. is a member of the board of Nestle Aus-tralia, is a consultant to Apollo Endosurgery andBariatric Advantage, and has received paymentsfor lectures from iNova Pharmaceuticals andCovidien, payments for development of educa-tional presentations from iNova Pharmaceu-ticals, and grants from Allergan. No otherpotential conflicts of interest relevant to thisarticle were reported.Author Contributions. S.P. contributed to thestudy concept and design, data analysis andinterpretation, and drafting of the manuscript.L.C. contributed to the study concept and designand critical revision of the manuscript for im-portant intellectual content. A.D.G. contributedto the data analysis. M.P. contributed to thestatistical analysis and critical revision of themanuscript for important intellectual content.T.R. and L.S. contributed to the critical revisionofthe manuscript for important intellectual con-tent. G.M. contributed to the study conceptand design, data interpretation, and drafting ofthe manuscript. J.B.D. contributed to the studyconcept and design, data interpretation, andcritical revision of the manuscript for importantintellectual content. S.P. and G.M. are the guar-antors of thiswork and, as such, had full access toall thedata in thestudyandtake responsibility forthe integrity of the data and the accuracy of thedata analysis.Prior Presentation. Parts of this study werepresented at the International Diabetes Feder-ation’s 2015 World Diabetes Congress, Vancou-ver, Canada, 30 November–4 December 2015.

References1. Sjostrom L, Lindroos AK, Peltonen M, et al.;Swedish Obese Subjects Study Scientific Group.Lifestyle, diabetes, and cardiovascular risk fac-tors 10 years after bariatric surgery. N Engl JMed 2004;351:2683–26932. Schauer PR, Kashyap SR, Wolski K, et al. Bar-iatric surgery versus intensive medical therapyin obese patients with diabetes. N Engl J Med2012;366:1567–15763. Dixon JB, O’Brien PE, Playfair J, et al. Adjust-able gastric banding and conventional therapyfor type 2 diabetes: a randomized controlledtrial. JAMA 2008;299:316–3234. Schauer PR, BhattDL, Kirwan JP, et al.; STAMPEDEInvestigators. Bariatric surgery versus intensivemedical therapy for diabetes–3-year outcomes.N Engl J Med 2014;370:2002–20135. Mingrone G, Panunzi S, De Gaetano A, et al.Bariatric surgery versus conventional medical

therapy for type 2 diabetes. N Engl J Med 2012;366:1577–15856. Gastrointestinal surgery for severe obesity.Proceedings of a National Institutes of HealthConsensus Development Conference. March25-27, 1991, Bethesda, MD. Am J Clin Nutr1992;55(2 Suppl.):487S–619S7. Dixon JB, Zimmet P, Alberti KG, Rubino F;International Diabetes Federation Taskforceon Epidemiology and Prevention. Bariatric sur-gery: an IDF statement for obese type 2 diabe-tes. Diabet Med 2011;28:628–6428. Maggard-Gibbons M, Maglione M, Livhits M,et al. Bariatric surgery for weight loss and glycemiccontrol innonmorbidly obese adultswithdiabetes:a systematic review. JAMA 2013;309:2250–22619. Jimenez A, Casamitjana R, Flores L, et al.Long-term effects of sleeve gastrectomy andRoux-en-Y gastric bypass surgery on type 2 di-abetes mellitus in morbidly obese subjects. AnnSurg 2012;256:1023–102910. Blackstone R, Bunt JC, Cortes MC,Sugerman HJ. Type 2 diabetes after gastric by-pass: remission in fivemodels using HbA1c, fast-ing blood glucose, and medication status. SurgObes Relat Dis 2012;8:548–55511. Brethauer SA, Aminian A, Romero-TalamasH, et al. Can diabetes be surgically cured? Long-term metabolic effects of bariatric surgery inobese patients with type 2 diabetes mellitus.Ann Surg 2013;258:628–636; discussion 636–63712. Arterburn DE, Bogart A, Sherwood NE, et al.A multisite study of long-term remission andrelapse of type 2 diabetes mellitus followinggastric bypass. Obes Surg 2013;23:93–10213. Lean MEJ, Han TS, Morrison CE. Waist cir-cumference as ameasure for indicating need forweight management. BMJ 1995;311:158–16114. Pouliot MC, Despres JP, Lemieux S, et al.Waist circumference and abdominal sagittal di-ameter: best simple anthropometric indexes ofabdominal visceral adipose tissue accumulationand related cardiovascular risk in men andwomen. Am J Cardiol 1994;73:460–46815. McDermott MM. The international pan-demic of chronic cardiovascular disease. JAMA2007;297:1253–125516. Wei M, Gaskill SP, Haffner SM, Stern MP.Waist circumference as the best predictor ofnoninsulin dependent diabetes mellitus(NIDDM) compared to body mass index, waist/hip ratio and other anthropometric measure-ments in Mexican Americans–a 7-year prospec-tive study. Obes Res 1997;5:16–2317. Siren R, Eriksson JG, Vanhanen H. Waistcircumference a good indicator of future riskfor type 2 diabetes and cardiovascular disease.BMC Public Health 2012;12:63118. Dixon JB, Chuang LM, Chong K, et al. Pre-dicting the glycemic response to gastric bypasssurgery in patients with type 2 diabetes. Diabe-tes Care 2013;36:20–2619. Dixon JB, Hur KY, Lee WJ, et al. Gastric by-pass in Type 2 diabetes with BMI , 30: weightand weight loss have a major influence on out-comes. Diabet Med 2013;30:e127–e13420. Sjostrom L, Peltonen M, Jacobson P, et al.Association of bariatric surgery with long-termremission of type 2 diabetes and with microvas-cular and macrovascular complications. JAMA2014;311:2297–2304

care.diabetesjournals.org Panunzi and Associates 9