PHYSIOLOGY RESEARCH

Comparison Between RYGB, DS, and VSG Effecton Glucose Homeostasis

Mitchell S. Roslin & Yuriy Dudiy & Joanne Weiskopf &Tanuja Damani & Paresh Shah

Published online: 9 June 2012# Springer Science+Business Media, LLC 2012

AbstractBackground Our group has reported a high incidence ofreactive hypoglycemia following Roux-en-Y gastric bypass(RYGB) with specific interest in postprandial insulin and theratio of 1- to 2-h serum glucose levels. The purpose of thisstudy is to compare the 6-month response to oral glucosechallenge in patients undergoing RYGB, duodenal switch(DS), and vertical sleeve gastrectomy (VSG).Methods Thirty-eight patients meeting the NIH criteria forbariatric surgery who have reached the 6-month postoperativemark are the basis of this report. Preoperatively and at6 months follow-up, patients underwent blood draw to deter-mine levels of fasting glucose, fasting insulin, HbA1c, Cpeptide, and 2 h oral liquid glucose challenge test (OGTT).HOMA-IR and 1 to 2 h ratios of glucose and fasting to 1 hratio of insulin were calculated.Results All patients underwent a successful laparoscopicbariatric procedure (VSG 013, DS 013, and RYGB012).All operations reduced BMI, HgbA1c, fasting glucose, andfasting insulin. HOMA IR and glucose tolerance improvedwith all procedures. In response to OGTT at 6 months, therewas a 20-fold increase in insulin at 1 h in RYGB, which wasnot seen in DS. At 6 months, 1-h insulin was markedlylower in DS (p<.05), yet HbA1C was also lower in DS

(p<.05). This resulted in 1- to 2-h glucose ratio of 1.9 forRYGB, 1.8 for VSG, and 1.3 for DS (p<.05).Conclusions All operations improve insulin sensitivity anddecrease HgbA1c. Six-month weight loss was substantial inall groups between 22–29% excess body weight. RYGBresults in marked rise in glucose following challenge withcorresponding rise in 1-h insulin. VSG has a similar responseto RYGB. In comparison, at 6 months following surgery, DScauses a much lower rise in 1-h insulin, with this differencebeing statistically significant at p<.05. As a result, DS resultsin a less abrupt reduction in blood glucose. Although 1-h insulin is lower, DS patients had the lowest HbA1C at6 months (p<.05). We believe that these findings have impor-tant implications for the choice of bariatric procedure for bothdiabetic and non-diabetic patients.

Keywords Roux-en-Y gastric bypass . Duodenal switch .

Vertical sleeve gastrectomy . Reactive hypoglycemia .

Glucose challenge . Insulin sensitivity

Introduction

The purpose of bariatric surgical procedures is to provide along-term control mechanism that enables patients to be lesshungry, to be satisfied with a smaller amount of food, and tomaintain weight loss resulting in reduced co-morbidities.Gastric procedures restrict or reduce the volume availablefor food. Intestinal operations presumably short-circuit theintestine to reduce the amount of calories absorbed by thesmall intestine. As the field has evolved, research has shownthat there are more factors than just mechanical. Cummingset al. [1] highlighted the role of ghrelin reduction, causingappetite suppression in Roux-en-Y gastric bypass (RYGB).Levels of gut peptides such as GLP 1 and PYY have beeninvestigated. Even more fundamental is the impact of theseprocedures on glucose regulation and metabolism.

M. S. Roslin :Y. Dudiy : J. Weiskopf : T. Damani : P. ShahLenox Hill Hospital,New York, NY, USA

M. S. RoslinNorthern Westchester Hospital Center,New York, NY, USA

M. S. Roslin (*)186 E 76th Street 1st floor,New York, NY 10021, USAe-mail: mitchell.roslin@mac.com

OBES SURG (2012) 22:1281–1286DOI 10.1007/s11695-012-0686-0

For the past 20 years, RYGB has been the most popularbariatric surgical stapling procedure. RYGB is based on asmall gastric pouch containing the lesser curvature of thestomach. Several principles are believed to be responsiblefor the success of gastric bypass. The small pouch providesrestriction. Since the distal stomach and pylorus arebypassed, sweet ingestion could lead to dumping, the symp-toms of which could make patients less likely to ingest simplecarbohydrates. The intestinal bypass would provide a reducedsurface area to absorb calories. Finally, by diverting pancreaticand biliary secretions, digestion, especially of fats, would beless efficient. Trials that have compared RYGB to verticalbanded gastroplasty or LAGB have consistently demonstratedbetter weight loss for gastric bypass. These results have beenattributed to RYGB ability to theoretically deter patients fromconsuming simple carbohydrates, as well as the role of mal-absorption. Recently, Kaplan et al. [2] has shown that thedifference in results may be accounted for by the fact thatthe bypass causes less reduction in basal energy expenditurethan pure restrictive operations.

Recently, our group has published data that questionwhether bypassing the pyloric valve, the anatomy createdby a gastric bypass, is ideal for a bariatric procedure [3]. Wepresented data on patients that regained weight followinggastric bypass who complained about light-headedness andhunger 1 to 2 h following eating. We performed oral glucosetolerance testing and demonstrated a rapid rise in glucose at1 h, with a spike in insulin level, with an abrupt fall to levelsconsistent with hypoglycemia. Thus, rather than deter sweetingestion, we believe that reactive hypoglycemia causesinter-meal hunger and encourages maladaptive eating.

While there has been considerable attention on the reso-lution of type II diabetes following RYGB, other aspects ofglucose regulation have been less actively discussed. It isimportant to recognize that the majority of patients under-going bariatric surgery are not diabetic. Thus, it is essentialto have a better understanding of what effect the anatomy ofthe bypass and other bariatric procedures will have onglucose homeostasis. Furthermore, an increasing numberof patients after RYGB have been diagnosed with hyper-insulinemic hypoglycemia [4–6]. Some have had neurologicsymptoms requiring reoperations that have included eventotal pancreatectomy [5, 7]. This phenomenon, coupled withrecent studies demonstrating the beneficial effects of duode-nal exclusion for non-obese diabetics [8, 9], has brought tolight the question of total glucose and insulin homeostasisafter an intestinal bypass procedure.

Although RYGB has been the most popular bariatricprocedure, there are other alternative stapling procedures.Duodenal switch (DS) involves the creation of a sleevegastrectomy with an intestinal bypass placed beneath thepyloric valve. In comparison to gastric bypass, studies haveshown greater weight loss and resolution of comorbidities

with DS, albeit with a greater degree of malabsorption-related sequelae. The ideal length of the common channelto balance malnutrition and weight loss continues to be amatter of investigation and debate.

Recently, vertical sleeve gastrectomy (VSG), whichinvolves resection of the greater curvature only, has becomean accepted primary procedure for weight loss. Studies haveshown weight loss and diabetes resolution similar to RYGB[10, 11].

As a result, we have decided to prospectively study theeffect of oral glucose challenge and solid mixed meal glucosechallenge on bariatric surgical patients undergoing laparo-scopic RYGB, VSG, and DS.

Methods

A total of 38 patients who had bariatric surgery have beenenrolled in this non-randomized prospective trial. Patientsare above 18 years of age and meet the NIH guidelines forbariatric surgery. All patients went through the standardseminar and educational program available as part of theCenter of Excellence at Lenox Hill Hospital. All patientswere allowed to select their surgical procedure and then,following this choice, enroll in the study and sign an in-formed consent. The study endpoints were discussed withthe patients only after they have selected their procedure.Patients studied were a cohort of bariatric surgical patientsand were eligible for surgery based on NIH guidelines.Patients with or without a diagnosis of diabetes were eligiblefor study.

As part of the preoperative laboratory assessment,patients enrolled in the study had a baseline blood draw thatincluded Hgb A1C, C peptide level, as well as a 2-h oralliquid glucose challenge test (OGTT) with measurement ofinsulin levels. All patients underwent an overnight fast andwere off all medications for their blood work at every studyinterval.

All surgical procedures were performed at Lenox HillHospital by three surgeons. RYGB had an isolated gastricpouch of approximately 20 cc based on the lesser curvature,ending at the angle of His, with a 150-cm roux limb and a100-cm biliopancreatic limb. VSG was performed over a 34bougie starting 3 to 4 cm from the pyloric valve, with care toleave space at the angularis incisura. DS involves a similartechnique as VSG using a 38 bougie, with a 150-cmalimentary limb and a 125- to 150-cm common channel.

At 6 months, patients had a 2-h OGTT with identicalmeasurements as a preoperative procedure.

Using determined values, homoeostatic model assessment(HOMA) IR was calculated using the standardized equation.

All procedures and laboratory measurements were per-formed at Lenox Hill Hospital using standard techniques.

1282 OBES SURG (2012) 22:1281–1286

The study design and informed consent were approved bythe North Shore Long Island Jewish / LenoxHill Hospital IRBand registered in clinical trials.gov.

Statistical analysis was performed by using a single-factoranalysis of variance (ANOVA) to look for between-groupvariations. When ANOVA showed a statistical difference,two-tailed t-tests were used to determine differences betweensubgroups.

Results

Six-month data on 38 patients are presented. Preoperatively,the average BMI was 49.1 (36.5–81.2). Of the 26 patients,12 had RYGB, 13 had VSG, and 13 had DS. The averageBMI of RYGB was 47.3±10, VSG 45.7±8, and DS was54.1±9.2. Table 1 shows the baseline and 6-month data.

At baseline, there was no statistical difference in any pa-rameter except for BMI and weight being greatest in the DSgroup. There was no difference in HOMA, C peptide, fastingglucose, fasting insulin, or any other parameter measured.

The body weight of RYGB patients fell by 22 %. Therewas an eightfold improvement in HOMA-IR, sevenfolddecrease of fasting insulin, and 3.25-fold decrease in Cpeptide level (Table 1).

Body weight was reduced by 23 % in VSG patients.There was 2.5-fold improvement in HOMA-IR, a 1.5-foldimprovement in C peptide, and a twofold improvement infasting insulin (Table 1).

There was a 29 % reduction in body weight and fivefoldimprovement in HOMA IR in DS patients. Preoperatively,they had the highest BMI but lowest fasting insulin levelsand baseline HOMA-IR (Table 1). There was a 2.3-folddecrease in C peptide and nearly a fivefold improvementin fasting insulin.

At 6 months, all operations resulted in lower fastingglucose and fasting insulin levels. With glucose challenge,1-h insulin levels rose to 76 in RYGB and 70 in VSG, butonly 36 in DS (p<.05). Both DS and RYGB resulted inlower fasting insulin levels than VSG at 6 months; however,this difference was not statistically significant (Figs. 1 and 2).

At 6 months, although there was no difference in fastingglucose in any group, however, following challenge, itresulted in 1- to 2-h glucose ratios of 1.9 for RYGB, 1.8for VSG, and 1.3 for DS (p<.05) (Table 2; Fig. 3).

Discussion

The purpose of this study is to determine how RYGB, VSG,and DS compare following glucose tolerance testing in acohort of bariatric surgical patients. Our data demonstratethat all three procedures cause a significant weight loss with T

able

1Characteristicsat

baselin

eandat

6mon

ths

Num

berof

patients

Weigh

tBMI

HOMA

C-

peptide

Glucose,

fasting

Glucose,

1h

Glucose,

2h

Insulin

,fasting

Insulin

,1h

Insulin

,2h

HbA

1C

RYGB

Preop

erative

1228

1.9±54†

47.3±10

6.2±4.8

5.2±8.1

105.5±32

182.0±82

160.2±91

25.9±23

76.7±81

56.2±82

6.8±2.0

Postoperativ

e22

3.3±45

36.8±7

0.8±0.3

1.6±0.4

86.9±14

165.3±87

†88

.9±51

3.6±1

76.25±45

†‡

15.7±23

5.9±0.8*

†

VSG

Preop

erative

1327

9.8±70‡

45.7±8‡

6.8±8

5.8±6.8

98.2±26

146.6±61

127.5±74

25.7±26

59.9±35

46.2±48

5.8±0.6

Postoperativ

e21

4.8±50

35.3±6

2.8±7

3.9±4.5

83.0±10

131.2±59

82.2±50

11.8±25

70.2±48

32.9±63

5.4±0.4*

DS

Preop

erative

1334

2.8±63†‡

54.1‡±9.2

3.4±2.3

3.5±1.4

97.2±39

151.8±74

125.8±79

13.8±9

77.2±50

50.3±31

6.1±1.1

Postoperativ

e24

5±60

38.2±7

0.7±0.5

1.5±0.6

77.9±18

.210

2.9±61

†80

.8±31

3.4±1.6

36.6±28

†‡

15±9

5.3±0.5†

*p<0.05

usingtwo-tailedt-testforRYGBversus

VSG;†p<0.05

usingtwo-tailedt-testforRYGBversus

DS;‡p<0.05

usingtwo-tailedt-testforVSG

versus

DS

OBES SURG (2012) 22:1281–1286 1283

improvement in fasting glucose, fasting insulin, C peptide,and HOMA. While the results with all three procedures areimpresssive, our study also highlights some fascinating dif-ferences. In comparison to RYGB and VSG, following DSat 6 months after glucose challenge, there is a far lower risein insulin at 1 h. This difference is statistically significant.Additionally, the ratio of 1- to 2-h glucose is lower, alsoreaching statistical significance.

Thus, the questions are why this occurs and what is theimportance and relevance of these findings? Certainly, theDS patients in our study can produce insulin followingchallenge as their preoperative levels were the same as thoseof other groups and markedly higher than following thesurgical procedure. Additionally, as shown by the reductionin fasting glucose, the normalization of the glucose toler-ance curve, and the lowest HgBA1C measured in the study,this lower level of production certainly meets the patients’needs to achieve glucose control. This indicates that DSpatients do not produce as high of value of insulin becauseit is not required to achieve euglycemia. The possibilitiesdiscussed below account for these findings.

We first need to exclude whether the groups weredifferent or the above findings are a result of the increasedweight loss caused by the DS procedure. Preoperatively,the only significant difference between groups was thatthe DS patients started with a higher weight and BMI.There was no difference in preoperative insulin or glucosevalues. At 6 months, DS resulted in a 29-% loss of totalbody weight, compared to 23 % for VSG and 22 % forRYGB. This difference was not significant. Furthermore,it is hard to believe that these findings can be accountedfor based on the slighlty greater weight loss seen in DS, at29 % weight loss compared to 23 % in VSG. Further-more, similar findings have been suggested by Johanssonwho studied the impact of meal challenge on DS patients.He also reported a reduction in peak glucose and insulinvalues [12].

What our data do suggest is that DS results in a markedimprovement in peripheral insulin resistance. While ourstudy examined a cohort of bariatric surgical patients, someof whom were diabetic, there recently have been severalpublications regarding the impact of bariatric surgcial

Fig. 1 Mean plasma glucose levels (±SD) by timepoint during a liquid glucose challenge test preoperatively (a) and at 6 months follow-up (b)

Fig. 2 Mean plasma insulin levels by timepoint during a liquid glucose challenge test preoperatively (a) and at 6 months follow-up (b)

1284 OBES SURG (2012) 22:1281–1286

procedures and diabetes. In a meta-analysis examining theeffect of bariatric surgery and diabetes, Buchwald reported a95-% resolution rate for DS, compared to 80 % for RYGB[13]. The STAMPEDE trial demonstrated diabetes remis-sion in patients following RYGB and VSG as compared tomedical therapy. In this study, RYGB was slightly moreefficacious than VSG [14]. In a trial conducted in Italy, a95-% remission rate with open biliopancreatic diversion and75-% remission rate with laparoscopic RYGB were recentlyreported [15]. Each of these studies suggests that patientswho were diabetic for a lengthy period and on insulintherapy for five or more years were least likely to achieveremission following RYGB.

In exactly this type of cohort, Frenken [16] examined DSsurgery. Using open DS in patients who were on insulintherapy for more than 5 years, they were 100 % successfulin halting insulin therapy. If we take our data, showing alower need of insulin for euglycemia, combined with thefact that even patients who possibly have lost the ability toproduce high levels of insulin (as suggested by their lengthymaintenance on insulin therapy) improved following DSsurgery, this clearly supports a marked improvement inperipheral insulin resistance. In comparison, it has beenspeculated that the predominant effect of RYGB is increasedinsulin production secondary to an increased level of incre-tins GLP and GIP. Our data suggest that VSG behaves moresimilarly to RYGB than DS.

Our initial interest in this topic was based on the findingthat many of our RYGB patients developed reactive hypo-glycemia. We hypothesized that preservation of the pyloricvalve would prevent dumping and is maybe a better phys-iologic alternative for bariatric surgical procedures. We de-cided to study two procedures that preserved the pyloricvalve, VSG and DS, and two procedures that includedintestinal bypass, RYGB and DS. An important differencebetween RYGB and DS is in how the intestinal bypass isconstructed. With RYGB, there is a fixed alimentary limband biliopancreatic limb and a variable, and a far longercommon channel. For DS, we fix the alimentary limb, andcommon channel, with the longest limb being the variablebiliopancreatic limb. Our data certainly cannot be accounted

for by preserving the pyloric valve. Additionally, the sub-stantially better resolution of type II diabetes with openScopinaro bilio pancreatic diversion (a procedure whichdoes not preserve the pyloric valve and includes a distalbypass) suggests that some factor caused by the distal intes-tinal bypass is the most likely cause for these results. Inter-estingly, laboratory studies have recently shown that the useof ursodeoxycholic acid, a blocker of bile salts, can mitigateagainst insulin resistance caused by metabolic syndrome [17].It is possible that the role of the pyloric valve is that it permitsa more distal bypass that diverts a significant proportion ofbiliary flow to be tolerated.

For patients who are not diabetic or have not been onlengthy insulin therapy our data demonstrates that followingRYGB there is a sharp rise in 1-h glucose and with higherinsulin production a sharp decline. A ratio of 1- to 2-h glucoseshows a statistically significant difference between DS andRYGB. It has been postulated that dumping or reactive hypo-glycemia caused by sweet ingestion causes postbypasspatients to avoid sweets. While the above theory is popular,it is also speculative and no study has ever shown a correlationbetween weight loss and dumping symptoms. In contrast, it is

Table 2 Ratio during oral liquid glucose challenge test

One hour to fasting toglucose ratio

One to 2 h glucose ratio One hour to fasting insulin ratio One to 2 h insulin ratio

Preoperative Postoperative Preoperative Postoperative Preoperative Postoperative Preoperative Postoperative

RYGB 1.7±0.3 1.9±0.8† 1.20±0.2 1.94±0.4° 4.7±6.0 23.7±17.7 2.6±2.3 8.5±6.1†

VSG 1.5±0.5 1.6±0.5 1.24±0.3 1.8±0.7‡ 5.5±5.5 13.6±10.3 3.2±3.6 6.6±4.9

DS 1.6±0.3 1.3±0.4† 1.28±0.3 1.26±0.4°‡ 6.1±2.5 12.4±9.0 1.9±1.2 3.6±3.7†

°p<0.01 using two-tailed t-test for RYGB versus DS, † p<0.05 using two-tailed t-test for RYGB versus DS, ‡ p<0.05 using two-tailed t-test forVSG versus DS

Fig. 3 One hour to fasting glucose level ratio during a liquidglucose challenge test. *p<0.05 using two-tailed t-test for RYGBvs DS, **p<0.05 using two-tailed t-test for VSG vs DS

OBES SURG (2012) 22:1281–1286 1285

absolutely proven that hypoglycemia is a strong stimulus forhunger and drives food consumption. As a result, we believethat a bariatric surgical procedure that reduces insulin produc-tion, while mantaining euglycemia, would be advantageousand may result in lower rates of recidivism.

Conflict of Interest Dr. Roslin is on the SAB of ValenTX, Surgi-quest, and Scientific Intake and consultant for J&J EES, Covidien.Research grant for this study is from Covidien. Patents are assigned toJ&J and CR Bard. Dr. Dudiy has no conflict of interest. JoanneWeiskopf has no conflict of interest. Dr. Tanuja Damani has no conflictof interest. Dr. Shah is on SAB of Stryker and Transenterix andconsultant for J&J EES and Olympus.

References

1. Cummings DE, Shannon MH. Roles for ghrelin in the regulationof appetite and body weight. Arch Surg. 2003;138:389–96.

2. Stylopoulos N, Hoppin AG, Kaplan LM. Roux-en-Y gastric by-pass enhances energy expenditure and extends lifespan in diet-induced obese rats. Obesity (Silver Spring). 2009;17:1839–47.

3. Roslin M, Damani T, Oren J, et al. Abnormal glucose tolerancetesting following gastric bypass demonstrates reactive hypoglyce-mia. Surg Endosc. 2010;25:1926–32.

4. Goldfine AB, Mun EC, Devine E, et al. Patients with neuroglyco-penia after gastric bypass surgery have exaggerated incretin andinsulin secretory responses to a mixed meal. J Clin EndocrinolMetab. 2007;92:4678–85.

5. Patti ME, McMahon G, Mun EC, et al. Severe hypoglycaemiapost-gastric bypass requiring partial pancreatectomy: evidence forinappropriate insulin secretion and pancreatic islet hyperplasia.Diabetologia. 2005;48:2236–40.

6. Service GJ, ThompsonGB, Service FJ, Andrews JC, Collazo-ClavellML, Lloyd RV. Hyperinsulinemic hypoglycemia with nesidioblasto-sis after gastric-bypass surgery. N Engl J Med. 2005;353:249–54.

7. Clancy TE, Moore Jr FD, Zinner MJ. Post-gastric bypass hyper-insulinism with nesidioblastosis: subtotal or total pancreatectomymay be needed to prevent recurrent hypoglycemia. J GastrointestSurg. 2006;10:1116–9.

8. Geloneze B, Geloneze SR, Fiori C, et al. Surgery for nonobesetype 2 diabetic patients: an interventional study with duodenal–jejunal exclusion. Obes Surg. 2009;19:1077–83.

9. Ramos AC, Galvao Neto MP, de Souza YM, et al. Laparoscopicduodenal–jejunal exclusion in the treatment of type 2 diabetesmellitus in patients with BMI<30 kg/m2 (LBMI). Obes Surg.2009;19:307–12.

10. Abbatini F, Rizzello M, Casella G, et al. Long-term effects oflaparoscopic sleeve gastrectomy, gastric bypass, and adjustablegastric banding on type 2 diabetes. Surg Endosc. 2010;24:1005–10.

11. Woelnerhanssen B, Peterli R, Steinert RE, et al. Effects of post-bariatric surgery weight loss on adipokines and metabolic param-eters: comparison of laparoscopic Roux-en-Y gastric bypass andlaparoscopic sleeve gastrectomy—a prospective randomized trial.Surg Obes Relat Dis. 2011;7:561–8.

12. Johansson HE, Haenni A, Karlsson FA, et al. Biliopancreaticdiversion with duodenal switch lowers both early and late phasesof glucose, insulin and proinsulin response after meal. Obes Surg.2010;20(5):549–58.

13. Buchwald H, Estok R, Fahrback K, et al. Weight and type 2diabetes after bariatric surgery: systemic review and meta-analysis. Am J Med. 2009;122(3):248–56.

14. Schauer PR, Kashyap SR, Wolski K, et al. Bariatric surgery versusintensive medical therapy in obese patients with diabetes. N Engl JMed. 2012;366(17):1567–76.

15. Mingrone G, Panunzi S, De Gaetano A, et al. Bariatric surgeryversus conventional medical therapy for type 2 diabetes. N Engl JMed. 2012;366(17):1577–85.

16. Frenken M, Cho EY, Karcz, et al. Improvement of type 2 diabetesmellitus in obese and non-obese patients after the duodenal switchoperation. J Obes. 2011;860–69.

17. Tsuchida T, Shiraishi M, Ohta T, et al. Ursodeoxycholic acidimproves insulin sensitivity and hepatic steatosis by inducing theexcretion of hepatic lipids in high-fat diet fed KK-A(y) mice.Metab. 2011.

1286 OBES SURG (2012) 22:1281–1286