Table of Contents
A prospective, open-label, randomized study was conducted at the Obesity Clinic of the University Hospitals of Leuven, Belgium. Patients were screened and considered eligible if they met all inclusion criteria (Table 1), including age between 18 and 65 years and a body mass index (BMI) of 35 kg/m or greater.2Participants had to be diagnosed with type 2 diabetes and require intensive insulin therapy with multiple daily injections, defined as requiring basal or bolus insulin injections for effective diabetes management. Further eligibility required that participants be approved for bariatric surgery by a local multidisciplinary obesity team according to National Institutes of Health guidelines.18Informed consent was obtained from all participants included in the study. The study was approved by the local ethics committee Research UZ/KU Leuven, registered at ClinicalTrials.gov (NCT01086111), and conducted in accordance with the Declaration of Helsinki.
Eligible patients were randomly assigned in a 1:1:1 ratio to a protein-sparing modified fasting (PSMF) diet alone, or RYGB or SG in combination with a PSMF diet, by a person not connected to the clinical trial. Blinding of treatment assignment was not possible due to the nature of the intervention. Once all study procedures had been performed, all patients randomly assigned to PSMF alone also underwent bariatric surgery.
The surgical interventions were performed in a single institution by the same experienced surgeon using already established techniques. Briefly, the laparoscopic RYGB surgery consisted of the creation of a small gastric pouch with a 25 mm circular stapled outlet, 120 cm of digestive tract, and 30 cm of bile duct. The laparoscopic SG surgery consisted of the resection of the fundus, gastric body, and part of the gastric antrum. This was prepared with a 32 French gastric tube, stapled from a point midway between the gastric angular notch and the pylorus to the angle of His. Regarding the dietary intervention, all participants were conditioned with a PSMF diet (Modifast® Patients were administered a PSMF diet (intensive, Nutrition & Santé, Belgium) with 800 kcal per day (very low calorie diet). All surgical patients were administered the PSMF diet from the first postoperative day until the third postoperative week, while non-surgical patients were administered the same PSMF diet for three consecutive weeks. At the time of intervention, patients were administered bolus long-acting insulin glargine (Lantas®Sanofi, France) is NPH insulin (Insulatard®Basal insulin dose was reduced by half in the surgery group, while the same dose was maintained in the PSMF group, to account for dietary adjustments immediately after surgery and the metabolic effects of surgery.
All patients underwent euglycemic (day -4) and hyperglycemic (day -1) clamps before surgery or before starting PSMF (day 0) and 3 weeks after surgery or after starting dietary therapy (days 21 and 24). After an overnight fast of at least 12 h, euglycemic and hyperglycemic clamps were performed according to a modified version of the protocol by DeFronzo et al..19,20The overall approach for each clamp is outlined in Figure 1 On the day of clamping, intake of all medications was deferred until after the procedure.
Euglycemic and hyperglycemic clamp protocols were performed before and 3 weeks after each intervention. The euglycemic clamp (top row) consisted of continuous insulin infusion (0.28 nmol/m2/min), adjusted by a variable 20% dextrose infusion to maintain blood glucose at 120 mg/dL based on bedside blood glucose monitoring at 5-min intervals. Blood samples were taken at -15, 90, 105, and 120 min to assess insulin and glucose. The hyperglycemic clamp (lower panel) involved taking blood samples at -30, -15, and 0 before starting the glucose infusion and raising blood glucose to 180 mg/dL using a stepwise glucose infusion between T0 and T170 to measure baseline glucose, insulin, and C-peptide. Blood samples were then taken at 5, 10, 120, 135, 150, 155, and 160 min to measure glucose, insulin, and C-peptide to assess beta-cell responses in phases 1, 2, and 3, respectively. Glucagon (1 mg) was injected at 150 min to assess maximal secretory capacity.
To perform the euglycemic clamp (Figure 1, top), intravenous catheters were inserted into the antecubital veins of both arms and insulin (ActRapid) was administered.® Blood samples were collected intermittently with 20% glucose solution (100 units/mL, Novo Nordisk, Denmark) before insulin injection (-15 min) for glucose and insulin measurements. Insulin was then infused at a constant rate of 0.28 nmol/m.2The infusion rate was adjusted based on bedside blood glucose measurements taken every 5 minutes (Hemocue).®Sweden). During steady state, blood samples were taken at 90, 105 and 120 min for glucose and insulin measurements. Each patient was asked to perform a hyperglycemic clamp 3 days later (Fig. 1—lower panel). Similarly, venous catheters were inserted into the cubital veins of both arms, a 20% glucose solution was infused, and blood samples were taken intermittently. To measure the baseline β-cell response (phase 0: −30 to 0 min), blood samples were taken at −30, −15 and 0 min for glucose, insulin and C-peptide measurements before glucose infusion. Plasma glucose was then measured at a concentration of 870 mg/m2/min (time: 0-5 min), 350 mg/m2/min (time: 5-10 min), 210 mg/m2/min (time: 10-14 min) and 180 mg/m2/min (after 14 min). Bedside blood glucose monitoring was performed every 5 min (Hemocue®Hyperglycemia was stabilized for the first 30 min using a clamp technique developed in the University of Cambridge, Sweden. From 15 to 150 min, a glucose target concentration of 180 mg/dL was maintained by infusing a 20% glucose solution at a variable rate based on bedside glucose values measured every 5 min. The glucose infusion rate at 150 min was then maintained until 170 min, and 1 mg of glucagon was injected intravenously. During the clamp, blood samples were taken at 5, 10, 120, 135, 150, 155, and 160 min to measure glucose, insulin, and C-peptide concentrations and to evaluate phase 1 (phase 1: 0–10 min), phase 2 (phase 2: 10–150 min), and phase 3 beta-cell responses or maximum secretory capacity (phase 3: 150–160 min).
Plasma glucose was analyzed by colorimetric hexokinase UV assay using a COBAS 8000 (Roche).®Basel, Switzerland), plasma insulin and C-peptide were measured using a COBAS 8000 analyzer (Roche®(Basel, Switzerland).
The primary outcomes of this study were changes in beta cell function and insulin sensitivity after RYGB, SG, or PSMF alone. Beta cell function was measured using the processing index (DI) and insulin sensitivity was measured using the glucose disposal rate (GDR).
From the hyperglycemic clamp, the AUC of C-peptide was calculated for each phase and expressed per minute. The insulin secretion rate (ISR) was calculated from plasma C-peptide levels using the insulin secretion (ISEC) method, with ISR phase 3 used as an index of maximal β-cell function.twenty one,twenty two,twenty threeSteady-state glucose infusion rates from the euglycemic clamp were used to assess whole-body insulin sensitivity, expressed as GDR.The DI was then calculated as a comprehensive measure of beta-cell function by multiplying the phase 3 AUC of the ISR from the hyperglycemic clamp by the GDR from the euglycemic clamp.
Statistical analyses were performed using SAS software, version 9.2 of the SAS System for Windows. Thirty patients were included in this exploratory study by convenience, with 10 patients each assigned to the three intervention groups. Data are presented as mean ± SD unless otherwise stated. Between-group comparisons of pre-intervention characteristics and post-intervention weight loss were performed using one-way ANOVA, followed by post-hoc Tukey tests for pairwise comparisons. Within-group comparisons of pre- and post-intervention characteristics were performed using paired t-tests. To compare post-intervention AUC values between groups, ANCOVA was performed with pre-intervention values as covariates, and Tukey adjustment was used for pairwise comparisons between groups. Least-squares means and corresponding 95% confidence intervals (CIs) were reported. Statistical significance was determined by p< 0.05 (two-tailed test).