Diabetes Remission After Gastric Bypass: To Be or Not to Be Dependent on Weight Loss

Weight loss surgery has been increasingly utilized for treatment of type 2 diabetes (T2D) in patients with obesity (1). Following Roux-en-Y gastric bypass (RYGB), one of the 2 most commonly performed bariatric surgeries, a larger proportion of patients with uncontrolled T2D achieve glycemic control compared to those who receive only medical intervention (2). While weight loss induced by calorie restriction (CR) or bariatric surgery improves fasting insulin kinetics, the glycemic effect of RYGB is more robust and occurs shortly after surgery before significant weight loss (2). Considerable research has focused on understanding mechanisms involved in weight loss independent antidiabetic effects of RYGB. Hyperinsulinemia that accompanies rapid rates of prandial glucose appearance and clearance after RYGB has been taken as evidence that surgery enhances enteroinsular axis activity improving glycemia (2). Regardless of the mechanism, short-term studies in small cohorts comparing the effect of equal weight loss on glucose homeostasis indicate that RYGB has greater antidiabetic effects than CR (3, 4).

In this issue, Purnell et al’s study (5) provides support for glycemic control, independent of weight loss, several years after RYGB. Diabetes remission over a 7-year follow-up period was compared among 827 obese subjects with T2D at the time of their enrollment in a multicenter prospective study, the Longitudinal Assessment of Bariatric Surgery, in which patients underwent either RYGB (80%) or laparoscopic adjustable gastric banding (LAGB), a purely restrictive bariatric procedure (20%). Baseline characteristics among the 2 groups were similar and so were the adherence rates to follow-up evaluation that maintained at 85% in 5 years and 50% in 7 years.

The results of the analysis by Purnell et al confirm established knowledge that complete diabetes remission, defined as glycated hemoglobin (A1C) <5.7% or fasting glucose <100 mg/dL in absence of antidiabetic medication, was greater after RYGB compared with LAGB. Complete resolution of diabetes peaked 1 to 2 years following surgery (51% after RYGB vs 23% after LAGB) when patients reached their maximum weight loss (nadir median body mass index [BMI] in RYGB vs LAGB: 30 vs 37 kg/m²), highlighting the role of weight loss in glycemic improvement. Among postoperative characteristics, weight loss was associated with diabetes remission after both procedures. However, adjusting for weight loss, RYGB remained superior to LAGB in inducing diabetes remission, indicating that factors beyond weight loss are also at play. Furthermore, diabetes relapse among the 2 surgical groups did not differ from year 2 to year 5 (~8% in RYGB vs 7% in LAGB) despite weight regain in RYGB (increased median BMI: 3 vs 0.2 kg/m² in RYGB vs LAGB).

These findings seem to be at odds with a recent study by Yoshimo et al (6) that compared T2D outcomes and glucose kinetics during fasting or fed conditions after 18% weight loss achieved by RYGB or CR and found no significant differences. Yet, these findings, in turn, are in contrast with those from previous studies where glycemic effect of RYGB were compared to CR (3, 4) or LAGB (7). Inconsistency in results among these studies may be rooted in variations in methodology or studied cohorts.

The present work by Purnell et al also corroborates previous reports that diabetes remission after bariatric surgery is greater in participants who are younger, those with shorter duration of diabetes, and individuals with lower A1C levels requiring fewer diabetic medications with no insulin requirement for adequate diabetic control before surgery. While the current guidelines (1) for treatment of diabetes in obese individuals recommend bariatric surgery, antidiabetic effects of these procedures may provide greater benefit when β-cell function is preserved in early stages of T2D.

Using adjusted risk ratio estimation including all baseline characteristics, Purnell and associates observed that greater fasting C-peptide and lower insulin levels before surgery predict diabetes remission regardless of surgery type. Using markers of glucose kinetics in a fasting state, the authors also reported that postintervention improvement in fasting C-peptide and glucose values (calculated as homeostatic model assessment for beta-cell function [HOMA]-B) was associated with better diabetes remission after RYGB whereas postsurgical improvement in fasting insulin sensitivity measured by HOMA-IR determined the likelihood of diabetes remission only after LAGB. These findings were taken by investigators as enhanced β-cell function beyond weight loss after RYGB and improved insulin kinetics secondary to weight loss after LAGB contribute to long-term glycemic control. However, this conclusion is oversimplified due to estimating β-cell function based on fasting levels of glucose, insulin, and C-peptide. Also, the method is inherently limited, since fasting markers are measured while the main effect of RYGB on β-cell secretory response is in a fed state (2-4, 7).

Laferrer et al (3), using an oral glucose tolerance test (OGTT) before and after 10% weight loss in patients with an average A1C of 6.8% treated with RYGB or CR, showed that post-OGTT glucose tolerance was improved only after RYGB, despite similar improvement in fasting glucose in both groups. Glycemic improvement after RYGB was associated with prandial hyperinsulinemia. Using similar methodology, Vetter et al (4) reported improved fasting hepatic insulin sensitivity and insulin clearance as well as better prandial β-cell secretory response adjusted for insulin sensitivity (ie, disposition index) after RYGB rather than CR (baseline A1C of 7.5%). Similarly, Dutia et al (7) demonstrated that after 10% weight loss, RYGB-treated patients had better glucose tolerance, likely due to greater post-OGTT insulin secretion, compared with LAGB-treated individuals. Fasting insulin kinetics and β-cell glucose sensitivity during intravenous glucose infusion did not differ among the 2 surgical groups. Altogether, these findings indicate that improved β-cell function in subjects with RYGB is mainly postprandial (2-4, 7). In fact, we have shown that insulin secretion in response to intravenous glucose, which has no effect on the release of gastrointestinal factors, is impaired in nondiabetic persons after RYGB compared with nonoperated individuals (8). Therefore, longitudinal studies are needed to explore the relationship between RYGB effects on islet-cell function and diabetes outcome.

Studies of long-term effects of bariatric surgery on diabetes, observational and randomized, have increased in recent years given the dramatic antidiabetic effects of procedures like RYGB. Purnell et al’s report is an example of observational studies in real-world settings showing a 4-fold greater diabetes remission in obese patients with T2D after RYGB compared with LAGB after adjusting for the amount of weight loss during 5 to 7 years.

Glossary

Abbreviations

A1C

glycated hemoglobin

CR

caloric restriction

HOMA-B

homeostatic model assessment for beta-cell function

HOMA-IR

homeostatic model assessment for insulin resistance

LAGB

laparoscopic adjustable gastric banding

OGTT

oral glucose tolerance test

RYGB

Roux-en-Y gastric bypass

T2D

type 2 diabetes

Additional Information

Disclosures: The authors have no relevant conflicts of interest.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.