Incidence of Hypoglycemia After Gastric Bypass vs Sleeve Gastrectomy: A Randomized Trial

Incidence of Hypoglycemia After Gastric Bypass vs Sleeve Gastrectomy: A Randomized Trial Abstract Context We compared the incidence of hypoglycemia after Roux-en-Y gastric bypass (RYGB) vs sleeve gastrectomy (SG). Design, Setting, and Main Outcome Measures Randomized, open-label trial conducted at the outpatient obesity clinic in a university hospital in Rome, Italy. The primary aim was the incidence of reactive hypoglycemia (<3.1 mmol/L after 75-g oral glucose load) at 1 year after surgery. Secondary aims were hypoglycemia under everyday life conditions, insulin sensitivity, insulin secretion, and lipid profile. Results Of 175 eligible patients, 120 were randomized 1:1 to RYGB or SG; 117 (93%) completed the 12-month follow-up. Reactive hypoglycemia was detected in 14% and 29% of SG and RYGB patients (P = 0.079), respectively, with the effect of treatment in multivariate analysis significant at P = 0.018. Daily hypoglycemic episodes during continuous glucose monitoring did not differ between groups (P = 0.75). Four of 59 RYGB subjects (6.8%) had 1 to 3 hospitalizations for symptomatic hypoglycemia vs 0 in SG. The static β-cell glucose sensitivity index increased after both treatments (P < 0.001), but the dynamic β-cell glucose sensitivity index increased significantly in SG (P = 0.008) and decreased in RYGB (P = 0.004 for time × treatment interaction). Whole-body insulin sensitivity increased about 10-fold in both groups. Conclusions We show that reactive hypoglycemia is no less common after SG and is not a safer option than RYGB, but RYGB is associated with more severe hypoglycemic episodes. This is likely due to the lack of improvement of β-cell sensitivity to changes in circulating glucose after RYGB, which determines an inappropriately high insulin secretion. During the past 40 years, the prevalence of obesity has doubled worldwide, with >1.3 billion adults, equivalent to 39% of the world’s population, being overweight, and >600 million, representing 13% of the population, being obese, in 2014 (1). Long-term weight loss is difficult to maintain, despite dietary and behavioral changes; weight loss attained at 1 year in lifestyle modification trials is only 1.6 kg on average (2). In contrast, bariatric surgery is very effective in maintaining weight loss, with an average of 32 kg or greater weight reduction at 5 years after treatment (3). Recently, sleeve gastrectomy (SG) outperformed in frequency the use of Roux-en-Y gastric bypass (RYGB) worldwide, representing 49% and 43% of global bariatric operations, respectively (4). Although bariatric/metabolic surgery has resulted in numerous health benefits, including type 2 diabetes (T2D) remission (5–7), it is also associated with serious metabolic side effects. Among these, hypoglycemia is a relevant and underestimated complication of RYGB leading to cognitive impairment when plasma glucose levels fall <3 mmol/L (54 mg/dL) (8). There is increasing evidence of severe hypoglycemia after RYGB, and some cases of postoperative nesidioblastosis have been described (9). Roslin et al. (10) found that, at 6 months after RYGB, 68% of patients had hypoglycemia at 2 hours during an oral glucose tolerance test (OGTT). Based on electronic medical record data, the incidence of clinical episodes of hypoglycemia was 13% in a retrospective study in nondiabetic RYGB patients (11). In contrast, studies based on inpatient hospitalization International Classification of Disease codes (12), or on self-reported data (13), show figures <1%. Yet, hospitalization is usually necessary only when severe hypoglycemia occurs. Repeated hypoglycemic episodes have been shown to impair the normal counterregulatory stress responses to subsequent hypoglycemia in diabetic individuals (14). Similar changes have been shown after RYGB, with clinical symptoms and counterregulatory hormonal and sympathetic nerve responses to hypoglycemia significantly reduced (15). Compared with RYGB, SG seems to have a much lower occurrence of reactive hypoglycemia, reported as ∼20% (16). Conversion from RYGB to SG is currently used as an alternative to pancreatectomy to treat severe hypoglycemia after RYGB (17). Although hypoglycemia following bariatric/metabolic surgery represents an important health issue, no randomized trials have been undertaken until now to compare the effect of SG with that of RYGB in relation to the incidence and severity of hypoglycemic episodes. The primary aim of our study was to conduct a 1-year randomized trial to compare the incidence of hypoglycemia after RYGB and SG. Secondary objectives were to measure the effects of these surgical procedures on metabolic processes associated with hypoglycemic risk, namely insulin sensitivity and insulin secretion, and on cardiovascular risk factors, body weight, and fat distribution. Methods Participants From December 2012 to December 2014, we screened 175 patients eligible for bariatric surgery at the Day-Hospital of Obesity and Related Disorders of the Catholic University in Rome (Fig. 1). A total of 120 patients were enrolled and randomly assigned 1:1 to RYGB or to SG. The study was reviewed and approved by the institutional human ethics committee (A1534/CE/2012) in accordance with national guidelines and the provisions of the Helsinki Declaration, as revised in 2000. All patients provided written informed consent to participate in the study, and additional written informed consent was obtained before surgery. Figure 1. View largeDownload slide Flow diagram outlining patient enrollment and outcomes. IGT, impaired glucose tolerance. Figure 1. View largeDownload slide Flow diagram outlining patient enrollment and outcomes. IGT, impaired glucose tolerance. Inclusion criteria were: age 25 to 65 years, body mass index (BMI) >40 or 35 to 40 kg/m2 in the presence of obesity complications, or ability to understand and comply with the study process. Exclusion criteria were history of T2D; previous bariatric surgery; or history of medical problems such as mental impairment, cancer, major cardiovascular or gastrointestinal or respiratory diseases, hormonal disorders, infections, or pregnancy. Diabetic patients were excluded to avoid confounding factors such as need for hypoglycemic agents. Intervention Patients were evaluated by a multidisciplinary team (including endocrinologists, surgeons, dietitians, and nurses) at baseline and at 1,3,6,9, and 12 months after SG or RYGB. Dietary suggestions including multiple meals and reduction of simple carbohydrates were provided at the moment of the discharge from the hospital after metabolic surgery. Hypoglycemia was assessed as follows: OGTT: Plasma glucose measurements during a 3-hour standard (75 g) OGTT conducted at 1 year postsurgery. Blood samples were drawn at fasting and at 30, 60, 90, 120, 150, and 180 minutes after the OGTT. In a subset of 50 randomly selected subjects, more frequent blood sampling was performed (i.e., at fasting and at 10, 20, 30, 40, 60, 80, 95, 120, 140, 155, and 180 minutes following the OGTT). OGTTs were performed at baseline and at 1, 3, 6, 9, and 12 months after surgery. Hypoglycemia awareness questionnaire at each visit relative to episodes of hypoglycemia in the 30 days preceding the visit. Hypoglycemia awareness was assessed by the Gold score, in which participants rated their awareness of hypoglycemia from 1 (always aware) to 7 (never aware) on a linear analog scale (18) at each visit. Episodes of severe hypoglycemia during the visit or during the previous 30 days, defined by Whipple triad (18), and confirmed by a finger prick blood glucose measurement, namely the association of neuroglycopenic and/or neurogenic symptoms, signs, or both, consistent with hypoglycemia, a plasma glucose concentration <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test, and resolution of those symptoms or signs after plasma glucose increased. Hospitalization for hypoglycemia. Blinded continuous glucose monitoring using Enlite Sensor (Medtronic), which participants wore for at least 5 days for 1 year. Fifty patients were randomly selected in each group; 25 at 1 year after RYGB and 25 at 1 year after SG were involved. An episode of hypoglycemia on continuous glucose monitoring (CGM) was defined as a single glycemic level <3.1 mmol/L, which is the average value of 5 minutes’ registration. Primary outcome The primary end point was the incidence of hypoglycemia (≤55 mg/dL or 3.1 mmol/L within 3 hours after ingestion of glucose during a 75-g glucose OGTT) at 1 year after surgery. Secondary outcomes Secondary end points were changes of body weight, BMI, symptomatic hypoglycemia, lipid profile, insulin sensitivity, insulin secretion during OGTT, abdominal circumferences, and body composition 1 year after surgery. Hypoglycemic events during everyday life by CGM were added as a secondary outcome in October 2014. Safety Adverse events from bariatric surgery were noted by the study nurse at each visit and subsequently classified by the study physician for potential causality (unlikely, possibly, or likely). Randomization A total of 120 patients requiring bariatric surgery for medical reasons met the study inclusion criteria after physician screening, and were subsequently randomly assigned by a computer-generated blocked random sequence to the RYGB or SG groups in a 1:1 ratio. Sample size Previous studies reported an absolute difference in the occurrence of hypoglycemia between the two surgical procedures of ∼50%, expecting an incidence of 68% in the RYGB group (10) vs 20% in the SG group (16). With a more conservative approach, the sample size was computed to detect a 30% difference between treatments, expecting an incidence of 50% in the RYGB group vs 20% in the SG group. Under this hypothesis, with two-tailed P < 0.05 and a power of 0.90, a sample size of 50 patients per group would have been sufficient. Considering an attrition rate of 20%, the number of patients in each group was set to 60 and, thus, overall to 120. Body composition Body weight was measured to the nearest 0.1 kg with a beam scale and height to the nearest 0.5 cm using a stadiometer. Hip and waist circumferences were measured in duplicate, with patients standing in underwear, as the maximal circumference over the buttocks and between the iliac crest and the lower ribcage, respectively. Body composition was measured by dual-energy X-ray absorptiometry (Lunar-iDXA). Laboratory analyses Fasting plasma glucose was measured by the glucose-oxidase method (Analox) and plasma insulin by microparticle enzyme immunoassay (Abbott), with 1 μU/mL sensitivity and 6.6% intra-assay coefficient of variation. Architect C-peptide (Abbott) assay precision of ≤10% total coefficient of variation was used to measure C-peptide levels. Total cholesterol, high-density lipoprotein (HDL) cholesterol, and triglyceride levels were measured by standard enzymatic assays. Low-density lipoprotein (LDL) cholesterol was calculated by the Friedewald formula. Mathematical modeling Insulin sensitivity and secretion were computed in the subset of 50 patients by the oral glucose minimal model (19). Changes over time of the indices were compared by a repeated measurement analysis of variance with time as within-factor and treatment as between-factor. Statistics Hypoglycemia during OGTT at 1 year after surgery was defined as any value of plasma glucose <3.1 mmol/L (55 mg/dL). During CGM, the average number of hypoglycemic episodes per day, defined as any glucose value <3.1 mmol/L, was also computed and compared between treatments by Mann-Whitney U test. Associations between treatments and presence/absence of hypoglycemic events were analyzed by χ2 or Fisher exact tests. For continuous variables means and standard deviations were computed. Student t tests were used to assess differences between groups at baseline. Percent changes at 1 year were computed as the differences between 1 year and baseline values divided by the baseline values, multiplied by 100. Deltas in the two groups were then compared by t test. Logistic regression with stepwise elimination was applied to identify possible predictors of hypoglycemic incidence. To study the possible relationship between the dependent binomial variable, presence/absence of hypoglycemia, and its regressors (independent variables), we used a logistic regression model with a multivariate variable selection procedure and a forward selection method; the latter was used to take into account multicollinearity. In particular, we included sex, age, BMI, triglycerides and total cholesterol, insulin sensitivity and insulin secretion at baseline, as well as variations in BMI, triglycerides, total cholesterol, insulin sensitivity and insulin secretion at 1 year after surgery, including the type of treatment as a possible explanatory variable. A logistic model relating presence/absence of hypoglycemia (glycemia <3.1 mmol/L) and baseline characteristics was carried out. The variables entering the model were age, sex, BMI, insulin sensitivity, and insulin secretion at baseline. The analysis was performed considering either all patients together, independent of the type of surgery, or separately (RYGB and SG). A nonlinear mixed-effects model was applied to both glycemic “trough” and “peak” variables. The model describing the trend over time of both the variables (y) is:  y=ϕ1×exp(−λ×time)+(ϕ3−ϕ1)Parameter ϕ3 represents the values of the dependent variables at baseline, parameter λ represents the decreasing rate of the variables, and ϕ1 represents the decrement that occurred during the follow-up. P < 0.05 was considered significant. All analyses were performed in R and Matlab (R2009b). Results Overall, 120 of 175 screened patients underwent randomization (Fig. 1).The follow-up at 1 year was 92.5%. Three patients withdrew their consent before the first visit after surgery and were excluded from the analysis. Two of the 3 patients who withdrew their consent in the SG arm after the visit at 6 months and 2 patients in the RYGB arm who withdrew after the visit at 3 months did so because of job-related changes in location. One patient withdrew consent in the SG arm after the 6-month visit did because of family problems. These patients were included in the intention-to-treat analysis. Three patients who abandoned the study immediately after surgery, 1 in the RYGB and 2 in the SG arm, respectively, did not give reasons for withdrawing prematurely from the trial and did not undergo the OGTT at 1 month and were therefore not included in the study. A total of 117 patients were therefore evaluated for the primary and secondary end points: 59 who underwent RYGB, age 43.07 ± 9.17 years, and 58 who underwent SG, age 45.76 ± 9.68 years. Between-groups baseline variables (Table 1) as well as sex (33.3% men and 66.7% women) did not differ significantly between treatment groups (P = 0.133). Table 1. Variables Averages and Their SD at Baseline and 1 y After Bariatric Surgery, With Percentage Changes     SG (n = 58)  RYGB (n = 59)  Overall (n = 117)        Mean  SD  Mean  SD  Mean  SD  P Value  Weight, kg  Baseline  121.57  18.32  124.19  13.30  122.89  15.96  0.390    After 1 y  84.04  16.33  81.13  9.95  82.57  13.51  0.262    % Δ at 1 y  −30.93  7.38  −34.47  6.81  −32.72  7.28  0.010  BMI, kg/m2  Baseline  43.44  4.25  43.10  3.96  43.27  4.09  0.666    After 1 y  30.02  3.93  28.27  3.81  29.14  3.95  0.019    % Δ at 1 y  −30.74  7.20  −34.33  6.89  −32.55  7.24  0.008  HbA1c, mmol/mol  Baseline  38.17  3.13  40.00  7.40  39.12  5.80  0.095    After 1 y  34.62  2.44  34.91  4.61  34.77  3.69  0.706    % Δ at 1 y  −8.85  7.89  −10.29  11.70  −9.58  9.99  0.488  Glycemia, mmol/L  Baseline  5.15  0.50  5.09  0.56  5.12  0.53  0.590    After 1 y  4.54  0.51  4.48  0.50  4.51  0.50  0.517    % Δ at 1 y  −11.06  12.63  −11.34  12.76  −11.20  12.64  0.911  Total cholesterol, mmol/L  Baseline  5.11  0.94  4.89  0.97  5.00  0.96  0.250    After 1 y  4.25  0.74  4.00  0.77  4.12  0.77  0.128    % Δ at 1 y  −14.56  19.80  −15.51  28.62  −15.05  24.62  0.855  HDL cholesterol, mmol/L  Baseline  1.29  0.29  1.26  0.27  1.28  0.28  0.669    after 1 year  1.34  0.26  1.49  0.30  1.42  0.29  0.024    % Δ at 1 y  8.42  25.15  20.92  43.23  14.83  35.92  0.121  LDL cholesterol, mmol/L  Baseline  3.02  0.88  2.97  0.70  2.99  0.79  0.742    After 1 y  2.48  0.74  2.13  0.61  2.29  0.69  0.023    % Δ at 1 y  −14.99  28.21  −29.25  15.84  −22.31  23.68  0.008  Triglycerides, mmol/L  Baseline  1.56  0.91  1.35  0.77  1.45  0.84  0.194    After 1 y  1.13  0.44  0.97  0.44  1.05  0.45  0.083    % Δ at 1 y  −14.97  40.08  −5.55  77.56  −10.15  61.97  0.479  Systolic blood pressure, mm Hg  Baseline  127.53  10.27  129.25  15.08  128.39  12.87  0.495    After 1 y  122.36  9.68  125.55  8.77  124.05  9.30  0.078    % Δ at 1 y  −3.81  8.91  −2.22  8.20  −2.98  8.54  0.352  Diastolic blood pressure, mm Hg  Baseline  80.72  6.97  83.68  10.01  82.20  8.71  0.080    After 1 y  78.54  6.96  80.54  5.45  79.59  6.26  0.102    % Δ at 1 y  −2.38  9.28  −2.82  12.18  −2.61  10.85  0.839  Hypoglycemia assessed by OGTT, %  Baseline  0    0    0    —    At 1 y  14    29    21.4    0.047  Episodes of severe hypoglycemia assessed by questionnaire, %  Baseline  0    0    0    —    At 1 y  0    5.1    2.6    0.082  Episodes of nonsevere hypoglycemia (Whipple triad plus glycemia <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test) during the visit at 1 y  Baseline  0    0    0    —    At 1 y  0    0    0    —  Hypoglycemia assessed during CGMa  At 1 y  32    24    28    0.533      SG (n = 58)  RYGB (n = 59)  Overall (n = 117)        Mean  SD  Mean  SD  Mean  SD  P Value  Weight, kg  Baseline  121.57  18.32  124.19  13.30  122.89  15.96  0.390    After 1 y  84.04  16.33  81.13  9.95  82.57  13.51  0.262    % Δ at 1 y  −30.93  7.38  −34.47  6.81  −32.72  7.28  0.010  BMI, kg/m2  Baseline  43.44  4.25  43.10  3.96  43.27  4.09  0.666    After 1 y  30.02  3.93  28.27  3.81  29.14  3.95  0.019    % Δ at 1 y  −30.74  7.20  −34.33  6.89  −32.55  7.24  0.008  HbA1c, mmol/mol  Baseline  38.17  3.13  40.00  7.40  39.12  5.80  0.095    After 1 y  34.62  2.44  34.91  4.61  34.77  3.69  0.706    % Δ at 1 y  −8.85  7.89  −10.29  11.70  −9.58  9.99  0.488  Glycemia, mmol/L  Baseline  5.15  0.50  5.09  0.56  5.12  0.53  0.590    After 1 y  4.54  0.51  4.48  0.50  4.51  0.50  0.517    % Δ at 1 y  −11.06  12.63  −11.34  12.76  −11.20  12.64  0.911  Total cholesterol, mmol/L  Baseline  5.11  0.94  4.89  0.97  5.00  0.96  0.250    After 1 y  4.25  0.74  4.00  0.77  4.12  0.77  0.128    % Δ at 1 y  −14.56  19.80  −15.51  28.62  −15.05  24.62  0.855  HDL cholesterol, mmol/L  Baseline  1.29  0.29  1.26  0.27  1.28  0.28  0.669    after 1 year  1.34  0.26  1.49  0.30  1.42  0.29  0.024    % Δ at 1 y  8.42  25.15  20.92  43.23  14.83  35.92  0.121  LDL cholesterol, mmol/L  Baseline  3.02  0.88  2.97  0.70  2.99  0.79  0.742    After 1 y  2.48  0.74  2.13  0.61  2.29  0.69  0.023    % Δ at 1 y  −14.99  28.21  −29.25  15.84  −22.31  23.68  0.008  Triglycerides, mmol/L  Baseline  1.56  0.91  1.35  0.77  1.45  0.84  0.194    After 1 y  1.13  0.44  0.97  0.44  1.05  0.45  0.083    % Δ at 1 y  −14.97  40.08  −5.55  77.56  −10.15  61.97  0.479  Systolic blood pressure, mm Hg  Baseline  127.53  10.27  129.25  15.08  128.39  12.87  0.495    After 1 y  122.36  9.68  125.55  8.77  124.05  9.30  0.078    % Δ at 1 y  −3.81  8.91  −2.22  8.20  −2.98  8.54  0.352  Diastolic blood pressure, mm Hg  Baseline  80.72  6.97  83.68  10.01  82.20  8.71  0.080    After 1 y  78.54  6.96  80.54  5.45  79.59  6.26  0.102    % Δ at 1 y  −2.38  9.28  −2.82  12.18  −2.61  10.85  0.839  Hypoglycemia assessed by OGTT, %  Baseline  0    0    0    —    At 1 y  14    29    21.4    0.047  Episodes of severe hypoglycemia assessed by questionnaire, %  Baseline  0    0    0    —    At 1 y  0    5.1    2.6    0.082  Episodes of nonsevere hypoglycemia (Whipple triad plus glycemia <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test) during the visit at 1 y  Baseline  0    0    0    —    At 1 y  0    0    0    —  Hypoglycemia assessed during CGMa  At 1 y  32    24    28    0.533  Abbreviation: HbA1c, glycosylated hemoglobin. a CGM was performed only on a subsample of 50 subjects, 25 per group, and only after surgery. View Large Table 1. Variables Averages and Their SD at Baseline and 1 y After Bariatric Surgery, With Percentage Changes     SG (n = 58)  RYGB (n = 59)  Overall (n = 117)        Mean  SD  Mean  SD  Mean  SD  P Value  Weight, kg  Baseline  121.57  18.32  124.19  13.30  122.89  15.96  0.390    After 1 y  84.04  16.33  81.13  9.95  82.57  13.51  0.262    % Δ at 1 y  −30.93  7.38  −34.47  6.81  −32.72  7.28  0.010  BMI, kg/m2  Baseline  43.44  4.25  43.10  3.96  43.27  4.09  0.666    After 1 y  30.02  3.93  28.27  3.81  29.14  3.95  0.019    % Δ at 1 y  −30.74  7.20  −34.33  6.89  −32.55  7.24  0.008  HbA1c, mmol/mol  Baseline  38.17  3.13  40.00  7.40  39.12  5.80  0.095    After 1 y  34.62  2.44  34.91  4.61  34.77  3.69  0.706    % Δ at 1 y  −8.85  7.89  −10.29  11.70  −9.58  9.99  0.488  Glycemia, mmol/L  Baseline  5.15  0.50  5.09  0.56  5.12  0.53  0.590    After 1 y  4.54  0.51  4.48  0.50  4.51  0.50  0.517    % Δ at 1 y  −11.06  12.63  −11.34  12.76  −11.20  12.64  0.911  Total cholesterol, mmol/L  Baseline  5.11  0.94  4.89  0.97  5.00  0.96  0.250    After 1 y  4.25  0.74  4.00  0.77  4.12  0.77  0.128    % Δ at 1 y  −14.56  19.80  −15.51  28.62  −15.05  24.62  0.855  HDL cholesterol, mmol/L  Baseline  1.29  0.29  1.26  0.27  1.28  0.28  0.669    after 1 year  1.34  0.26  1.49  0.30  1.42  0.29  0.024    % Δ at 1 y  8.42  25.15  20.92  43.23  14.83  35.92  0.121  LDL cholesterol, mmol/L  Baseline  3.02  0.88  2.97  0.70  2.99  0.79  0.742    After 1 y  2.48  0.74  2.13  0.61  2.29  0.69  0.023    % Δ at 1 y  −14.99  28.21  −29.25  15.84  −22.31  23.68  0.008  Triglycerides, mmol/L  Baseline  1.56  0.91  1.35  0.77  1.45  0.84  0.194    After 1 y  1.13  0.44  0.97  0.44  1.05  0.45  0.083    % Δ at 1 y  −14.97  40.08  −5.55  77.56  −10.15  61.97  0.479  Systolic blood pressure, mm Hg  Baseline  127.53  10.27  129.25  15.08  128.39  12.87  0.495    After 1 y  122.36  9.68  125.55  8.77  124.05  9.30  0.078    % Δ at 1 y  −3.81  8.91  −2.22  8.20  −2.98  8.54  0.352  Diastolic blood pressure, mm Hg  Baseline  80.72  6.97  83.68  10.01  82.20  8.71  0.080    After 1 y  78.54  6.96  80.54  5.45  79.59  6.26  0.102    % Δ at 1 y  −2.38  9.28  −2.82  12.18  −2.61  10.85  0.839  Hypoglycemia assessed by OGTT, %  Baseline  0    0    0    —    At 1 y  14    29    21.4    0.047  Episodes of severe hypoglycemia assessed by questionnaire, %  Baseline  0    0    0    —    At 1 y  0    5.1    2.6    0.082  Episodes of nonsevere hypoglycemia (Whipple triad plus glycemia <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test) during the visit at 1 y  Baseline  0    0    0    —    At 1 y  0    0    0    —  Hypoglycemia assessed during CGMa  At 1 y  32    24    28    0.533      SG (n = 58)  RYGB (n = 59)  Overall (n = 117)        Mean  SD  Mean  SD  Mean  SD  P Value  Weight, kg  Baseline  121.57  18.32  124.19  13.30  122.89  15.96  0.390    After 1 y  84.04  16.33  81.13  9.95  82.57  13.51  0.262    % Δ at 1 y  −30.93  7.38  −34.47  6.81  −32.72  7.28  0.010  BMI, kg/m2  Baseline  43.44  4.25  43.10  3.96  43.27  4.09  0.666    After 1 y  30.02  3.93  28.27  3.81  29.14  3.95  0.019    % Δ at 1 y  −30.74  7.20  −34.33  6.89  −32.55  7.24  0.008  HbA1c, mmol/mol  Baseline  38.17  3.13  40.00  7.40  39.12  5.80  0.095    After 1 y  34.62  2.44  34.91  4.61  34.77  3.69  0.706    % Δ at 1 y  −8.85  7.89  −10.29  11.70  −9.58  9.99  0.488  Glycemia, mmol/L  Baseline  5.15  0.50  5.09  0.56  5.12  0.53  0.590    After 1 y  4.54  0.51  4.48  0.50  4.51  0.50  0.517    % Δ at 1 y  −11.06  12.63  −11.34  12.76  −11.20  12.64  0.911  Total cholesterol, mmol/L  Baseline  5.11  0.94  4.89  0.97  5.00  0.96  0.250    After 1 y  4.25  0.74  4.00  0.77  4.12  0.77  0.128    % Δ at 1 y  −14.56  19.80  −15.51  28.62  −15.05  24.62  0.855  HDL cholesterol, mmol/L  Baseline  1.29  0.29  1.26  0.27  1.28  0.28  0.669    after 1 year  1.34  0.26  1.49  0.30  1.42  0.29  0.024    % Δ at 1 y  8.42  25.15  20.92  43.23  14.83  35.92  0.121  LDL cholesterol, mmol/L  Baseline  3.02  0.88  2.97  0.70  2.99  0.79  0.742    After 1 y  2.48  0.74  2.13  0.61  2.29  0.69  0.023    % Δ at 1 y  −14.99  28.21  −29.25  15.84  −22.31  23.68  0.008  Triglycerides, mmol/L  Baseline  1.56  0.91  1.35  0.77  1.45  0.84  0.194    After 1 y  1.13  0.44  0.97  0.44  1.05  0.45  0.083    % Δ at 1 y  −14.97  40.08  −5.55  77.56  −10.15  61.97  0.479  Systolic blood pressure, mm Hg  Baseline  127.53  10.27  129.25  15.08  128.39  12.87  0.495    After 1 y  122.36  9.68  125.55  8.77  124.05  9.30  0.078    % Δ at 1 y  −3.81  8.91  −2.22  8.20  −2.98  8.54  0.352  Diastolic blood pressure, mm Hg  Baseline  80.72  6.97  83.68  10.01  82.20  8.71  0.080    After 1 y  78.54  6.96  80.54  5.45  79.59  6.26  0.102    % Δ at 1 y  −2.38  9.28  −2.82  12.18  −2.61  10.85  0.839  Hypoglycemia assessed by OGTT, %  Baseline  0    0    0    —    At 1 y  14    29    21.4    0.047  Episodes of severe hypoglycemia assessed by questionnaire, %  Baseline  0    0    0    —    At 1 y  0    5.1    2.6    0.082  Episodes of nonsevere hypoglycemia (Whipple triad plus glycemia <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test) during the visit at 1 y  Baseline  0    0    0    —    At 1 y  0    0    0    —  Hypoglycemia assessed during CGMa  At 1 y  32    24    28    0.533  Abbreviation: HbA1c, glycosylated hemoglobin. a CGM was performed only on a subsample of 50 subjects, 25 per group, and only after surgery. View Large Primary outcome Percentages of patients experiencing hypoglycemia (<55 mg/dL or 3.1 mmol/L) after OGTT at 1 year were 14% and 29% in the SG and RYGB groups, respectively (P = 0.079). The percentages of patients experiencing hypoglycemia after OGTT over time was higher in RYGB than in the SG group, with a maximum at 6 and 3 months, respectively (Fig. 2). Figure 2. View largeDownload slide Average number of hypoglycemic episodes (glycemia ≤3.1 mmol/L) following the OGTT at 1 year after sleeve gastrectomy (blue solid line) and after gastric bypass (red dashed line). Figure 2. View largeDownload slide Average number of hypoglycemic episodes (glycemia ≤3.1 mmol/L) following the OGTT at 1 year after sleeve gastrectomy (blue solid line) and after gastric bypass (red dashed line). About 5% of the RYGB and none of the SG patients referred to symptoms of hypoglycemia in their questionnaire (Table 1). Absence of Whipple triad associated with glycemia <3.9, < 3.4, or 3.1 mmol/L, as assessed by finger-prick test, excluded symptomatic hypoglycemia during the periodic visits and OGTT in all patients. In a mixed-effects model, with the occurrence of hypoglycemia as the dependent variable, both time and treatment were significant contributory factors (P < 0.001 and P = 0.01, respectively). When postsurgery weight was introduced as a covariate, time was no longer significant, with estimates of the regression coefficients of 1.50 (P = 0.018) for RYGB vs SG, 0.04 (P = 0.5) for time, and −0.04 (P = 0.020) for weight. The odds ratio for hypoglycemia was 4.5 times higher for RYGB than for SG (P = 0.018). An OGTT plasma glucose value ≤3.9 mmol/L was recorded in 67.2% of SG and in 91.5% of RYGB patients (P = 0.002). Glucose levels ≤3.3 mmol/L were 36.2% and 61% in the SG and RYGB group, respectively (P = 0.012). Glycemia ≤2.8 mmol/L was recorded in 5.2% of SG and in 18.6% of RYGB patients (P = 0.050). Glycemia ≤2.2 mmol/L occurred only in 2 RYGB patients (P = 0.50). Four of 59 subjects who completed the 1-year follow-up in the RYGB group had 1 to 3 hospitalizations for symptomatic, neuroglycopenic hypoglycemia between month 9 and 12 after surgery. Incidence of hyperglycemia during OGTT Hyperglycemia (≥10 mmol/L) during OGTT occurred in 13.8% of SG and 22% of RYGB patients (P = 0.36). The association at 1 year between hypoglycemic and hyperglycemic episodes, independent of treatment, was not significant (P = 0.56). Glycemic trough and peak after OGTT The values of both plasma glucose trough and peak during the OGTT were evaluated for each patient. Trough values decreased at 1 year and Δ was significantly different between groups (−30.9 ± 24.7 in RYGB vs −21.7 ± 21.7 in SG, P = 0.040). Also, the mixed-effects model showed a significantly greater decrement of the trough values in RYGB patients: ϕ1 = 21.0 ± 2.22 [standard error (SE)] for SG vs ϕ1 + 6.0 ± 1.88 for RYGB, P = 0.0006. Peak glycemic values decreased less in RYGB than in SG with Δ of −29.9 ± 44.06 and of −47.7 ± 30.55 (P = 0.015), respectively. RYGB patients had a numerically lower (P = 0.094) glycemic peak value at baseline, which did not reach significance, and experienced a significantly smaller decrement in the attained maximum values of glycemia: ϕ1 = 46.60 ± 4.5 (SE) for SG vs ϕ1 − 16.67 ± 6.3 for RYGB, P = 0.0087. Supplemental Figs. 1 and 2 report the average (dashed lines) and predicted values (continuous lines) of the trough and peak variables for SG and RYGB, respectively. Glycated hemoglobin Glycated hemoglobin decreased significantly at 1 year: changes were −10.29 ± 11.70% (P < 0.001) in RYGB and −8.85 ± 7.89% (P < 0.001) in the SG groups, with no significant difference between groups (P = 0.50). Also, fasting glycemia decreased significantly in both groups without significant differences between groups: % Δ −11.34 ± 12.76 in RYGB (P < 0.001) and −11.06 ± 12.63, P < 0.001 in SG; absolute values: 4.48 ± 0.50 vs 4.54 ± 0.51 mmol/L in RYGB and SG, respectively (P = 0.52). Incidence of hypoglycemia during CGM The average daily glycemic values ≤3.1 mmol/L did not differ between groups: 5.47 ± 11.35 in the SG vs 2.76 ± 6.64 in RYGB (P = 0.75) groups. Eight patients in the SG group and 6 in the RYGB group experienced 1 to 4 hypoglycemic episodes (continuous periods of hypoglycemia during CGM); overall, 14 hypoglycemic episodes were recorded in SG and 9 in RYGB patients, with an average duration of 48.1 ± 37.2 and 45.6 ± 32.8 minutes, respectively (P = 1 by Mann-Whitney U test). Fifty-seven percent of the hypoglycemic episodes in the SG group were postprandial, whereas 43% were nocturnal. In the RYGB group, however, most occurred during the night (55%), and two were postprandial and two occurred in the afternoon at some distance from meal consumption. The daily average number of hypoglycemic episodes was 1.26 ± 0.81 and 1.77 ± 2.14 (P = 0.95 by Mann-Whitney U test) in the SG and RYGB groups, respectively. The mean glycemic values during hypoglycemic episodes in the 8 SG and 6 RYGB patients were 2.73 ± 0.27 and 2.98 ± 0.06 mmol/L, respectively (P = 0.10 by Mann-Whitney U test). We report in Supplemental Table 1 the study variables for patients with and without hypoglycemia. The only significantly different variables between the two groups were weight percent changes as well as 1-year total cholesterol and triglyceride levels that were lower in the group with hypoglycemia. Insulin sensitivity and insulin secretion during OGTT Results from the minimal model analysis are reported in Table 2. The static β-cell glucose sensitivity index ϕs increased after both treatments (50.99 ± 33.68 at baseline vs 120.43 ± 98.90 ×109/min–1 at 1 year, P < 0.001), whereas there were no significant difference between RYGB and SG. A similar trend was recorded for the global β-cell glucose sensitivity. In contrast, dynamic β-cell glucose sensitivity had a different outcome; it increased significantly (P = 0.008) in the SG group (from 324 ± 481 to 933 ± 1064 ×109/min), but decreased in the RYGB group (from 646 ± 733 to 415 ± 470 ×109/min) (P = 0.004 for time × treatment interaction). Whole-body insulin sensitivity increased from 3.76 ± 3.09 to 36.26 ± 57.09 ×105 dL/kg/min−1 (pmol/L)−1, whereas insulin secretion rate (ISR) area under the curve (AUC) decreased from 140.6 ± 54.8 to 64.6 ± 24.2 nmol. Table 2. OGTT Minimal Model β-cell Glucose Sensitivity, Whole Body Insulin Sensitivity Indices, and Body Composition Variables    SG (n = 25)  RYGB (n = 25)  Overall (n = 50)      Mean  SD  Mean  SD  Mean  SD  P Value  Static β-cell glucose sensitivity, ×109/min−1  Baseline  53.67  35.55  48.30  32.22  50.99  33.68  0.58    After 1 y  114.45  100.61  126.41  98.86  120.43  98.90  0.67  Dynamic β-cell glucose sensitivity, 109  Baseline  323.93  480.87  645.69  733.21  484.81  634.81  0.07    After 1 y  933.32  1063.86  414.90  469.87  674.11  855.01  0.03  Global β-cell glucose sensitivity, ×109/min−1  Baseline  57.32  35.51  54.71  34.44  56.01  34.65  0.79    After 1 y  130.37  104.36  135.74  99.97  133.06  101.17  0.85  Whole body insulin sensitivity, 105 dL/kg per min per pmol/L  Baseline  3.28  2.80  4.24  3.34  3.76  3.09  0.28    After 1 y  36.87  59.75  35.66  55.54  36.26  57.10  0.94  Insulin secretion rate AUC, nmol  Baseline  136.11  52.26  145.20  58.02  140.65  54.84  0.56    After 1 y  66.5  22.21  62.51  26.35  64.63  24.21  0.54  Disposition index, 1014 dL/kg/min−2 per pmol/L  Baseline  202.00  259.66  217.25  200.46  209.62  229.71  0.82    After 1 y  4532.68  6329.19  3728.83  5042.44  4130.75  5677.94  0.62  C-peptide AUC over glucose AUC above basal levels, nmol/pmol  Baseline  135.39  36.87  119.66  41.00  127.52  39.39  0.16    After 1 y  179.70  118.98  228.63  175.45  204.17  150.40  0.25  C-peptide AUC over glucose AUC, nmol/pmol  Baseline  118.79  33.29  112.84  36.42  115.81  34.66  0.55    After 1 y  100.32  28.17  114.40  49.09  107.36  40.25  0.22  Waist circumference, cm  Baseline  125.76  8.50  127.64  11.53  126.70  10.07  0.515    After 1 y  101.52  9.97  97.84  10.75  99.68  10.43  0.215    % Δ at 1 y  −19.29  5.40  −23.32  5.29  −21.31  5.67  0.010  Hip circumference, cm  Baseline  131.68  9.76  133.52  10.37  132.60  10.01  0.521    After 1 y  106.12  9.80  102.72  12.23  104.42  11.10  0.283    % Δ at 1 y  −19.40  4.57  −23.16  6.08  −21.28  5.65  0.017  Lean mass, kg  % Δ at 1 y  −17.03  9.95  −15.87  10.35  −16.45  10.06  0.688  Fat mass, kg  % Δ at 1 y  −36.11  14.27  −47.81  14.77  −41.96  15.54  0.006  Variables    SG (n = 25)  RYGB (n = 25)  Overall (n = 50)      Mean  SD  Mean  SD  Mean  SD  P Value  Static β-cell glucose sensitivity, ×109/min−1  Baseline  53.67  35.55  48.30  32.22  50.99  33.68  0.58    After 1 y  114.45  100.61  126.41  98.86  120.43  98.90  0.67  Dynamic β-cell glucose sensitivity, 109  Baseline  323.93  480.87  645.69  733.21  484.81  634.81  0.07    After 1 y  933.32  1063.86  414.90  469.87  674.11  855.01  0.03  Global β-cell glucose sensitivity, ×109/min−1  Baseline  57.32  35.51  54.71  34.44  56.01  34.65  0.79    After 1 y  130.37  104.36  135.74  99.97  133.06  101.17  0.85  Whole body insulin sensitivity, 105 dL/kg per min per pmol/L  Baseline  3.28  2.80  4.24  3.34  3.76  3.09  0.28    After 1 y  36.87  59.75  35.66  55.54  36.26  57.10  0.94  Insulin secretion rate AUC, nmol  Baseline  136.11  52.26  145.20  58.02  140.65  54.84  0.56    After 1 y  66.5  22.21  62.51  26.35  64.63  24.21  0.54  Disposition index, 1014 dL/kg/min−2 per pmol/L  Baseline  202.00  259.66  217.25  200.46  209.62  229.71  0.82    After 1 y  4532.68  6329.19  3728.83  5042.44  4130.75  5677.94  0.62  C-peptide AUC over glucose AUC above basal levels, nmol/pmol  Baseline  135.39  36.87  119.66  41.00  127.52  39.39  0.16    After 1 y  179.70  118.98  228.63  175.45  204.17  150.40  0.25  C-peptide AUC over glucose AUC, nmol/pmol  Baseline  118.79  33.29  112.84  36.42  115.81  34.66  0.55    After 1 y  100.32  28.17  114.40  49.09  107.36  40.25  0.22  Waist circumference, cm  Baseline  125.76  8.50  127.64  11.53  126.70  10.07  0.515    After 1 y  101.52  9.97  97.84  10.75  99.68  10.43  0.215    % Δ at 1 y  −19.29  5.40  −23.32  5.29  −21.31  5.67  0.010  Hip circumference, cm  Baseline  131.68  9.76  133.52  10.37  132.60  10.01  0.521    After 1 y  106.12  9.80  102.72  12.23  104.42  11.10  0.283    % Δ at 1 y  −19.40  4.57  −23.16  6.08  −21.28  5.65  0.017  Lean mass, kg  % Δ at 1 y  −17.03  9.95  −15.87  10.35  −16.45  10.06  0.688  Fat mass, kg  % Δ at 1 y  −36.11  14.27  −47.81  14.77  −41.96  15.54  0.006  View Large Table 2. OGTT Minimal Model β-cell Glucose Sensitivity, Whole Body Insulin Sensitivity Indices, and Body Composition Variables    SG (n = 25)  RYGB (n = 25)  Overall (n = 50)      Mean  SD  Mean  SD  Mean  SD  P Value  Static β-cell glucose sensitivity, ×109/min−1  Baseline  53.67  35.55  48.30  32.22  50.99  33.68  0.58    After 1 y  114.45  100.61  126.41  98.86  120.43  98.90  0.67  Dynamic β-cell glucose sensitivity, 109  Baseline  323.93  480.87  645.69  733.21  484.81  634.81  0.07    After 1 y  933.32  1063.86  414.90  469.87  674.11  855.01  0.03  Global β-cell glucose sensitivity, ×109/min−1  Baseline  57.32  35.51  54.71  34.44  56.01  34.65  0.79    After 1 y  130.37  104.36  135.74  99.97  133.06  101.17  0.85  Whole body insulin sensitivity, 105 dL/kg per min per pmol/L  Baseline  3.28  2.80  4.24  3.34  3.76  3.09  0.28    After 1 y  36.87  59.75  35.66  55.54  36.26  57.10  0.94  Insulin secretion rate AUC, nmol  Baseline  136.11  52.26  145.20  58.02  140.65  54.84  0.56    After 1 y  66.5  22.21  62.51  26.35  64.63  24.21  0.54  Disposition index, 1014 dL/kg/min−2 per pmol/L  Baseline  202.00  259.66  217.25  200.46  209.62  229.71  0.82    After 1 y  4532.68  6329.19  3728.83  5042.44  4130.75  5677.94  0.62  C-peptide AUC over glucose AUC above basal levels, nmol/pmol  Baseline  135.39  36.87  119.66  41.00  127.52  39.39  0.16    After 1 y  179.70  118.98  228.63  175.45  204.17  150.40  0.25  C-peptide AUC over glucose AUC, nmol/pmol  Baseline  118.79  33.29  112.84  36.42  115.81  34.66  0.55    After 1 y  100.32  28.17  114.40  49.09  107.36  40.25  0.22  Waist circumference, cm  Baseline  125.76  8.50  127.64  11.53  126.70  10.07  0.515    After 1 y  101.52  9.97  97.84  10.75  99.68  10.43  0.215    % Δ at 1 y  −19.29  5.40  −23.32  5.29  −21.31  5.67  0.010  Hip circumference, cm  Baseline  131.68  9.76  133.52  10.37  132.60  10.01  0.521    After 1 y  106.12  9.80  102.72  12.23  104.42  11.10  0.283    % Δ at 1 y  −19.40  4.57  −23.16  6.08  −21.28  5.65  0.017  Lean mass, kg  % Δ at 1 y  −17.03  9.95  −15.87  10.35  −16.45  10.06  0.688  Fat mass, kg  % Δ at 1 y  −36.11  14.27  −47.81  14.77  −41.96  15.54  0.006  Variables    SG (n = 25)  RYGB (n = 25)  Overall (n = 50)      Mean  SD  Mean  SD  Mean  SD  P Value  Static β-cell glucose sensitivity, ×109/min−1  Baseline  53.67  35.55  48.30  32.22  50.99  33.68  0.58    After 1 y  114.45  100.61  126.41  98.86  120.43  98.90  0.67  Dynamic β-cell glucose sensitivity, 109  Baseline  323.93  480.87  645.69  733.21  484.81  634.81  0.07    After 1 y  933.32  1063.86  414.90  469.87  674.11  855.01  0.03  Global β-cell glucose sensitivity, ×109/min−1  Baseline  57.32  35.51  54.71  34.44  56.01  34.65  0.79    After 1 y  130.37  104.36  135.74  99.97  133.06  101.17  0.85  Whole body insulin sensitivity, 105 dL/kg per min per pmol/L  Baseline  3.28  2.80  4.24  3.34  3.76  3.09  0.28    After 1 y  36.87  59.75  35.66  55.54  36.26  57.10  0.94  Insulin secretion rate AUC, nmol  Baseline  136.11  52.26  145.20  58.02  140.65  54.84  0.56    After 1 y  66.5  22.21  62.51  26.35  64.63  24.21  0.54  Disposition index, 1014 dL/kg/min−2 per pmol/L  Baseline  202.00  259.66  217.25  200.46  209.62  229.71  0.82    After 1 y  4532.68  6329.19  3728.83  5042.44  4130.75  5677.94  0.62  C-peptide AUC over glucose AUC above basal levels, nmol/pmol  Baseline  135.39  36.87  119.66  41.00  127.52  39.39  0.16    After 1 y  179.70  118.98  228.63  175.45  204.17  150.40  0.25  C-peptide AUC over glucose AUC, nmol/pmol  Baseline  118.79  33.29  112.84  36.42  115.81  34.66  0.55    After 1 y  100.32  28.17  114.40  49.09  107.36  40.25  0.22  Waist circumference, cm  Baseline  125.76  8.50  127.64  11.53  126.70  10.07  0.515    After 1 y  101.52  9.97  97.84  10.75  99.68  10.43  0.215    % Δ at 1 y  −19.29  5.40  −23.32  5.29  −21.31  5.67  0.010  Hip circumference, cm  Baseline  131.68  9.76  133.52  10.37  132.60  10.01  0.521    After 1 y  106.12  9.80  102.72  12.23  104.42  11.10  0.283    % Δ at 1 y  −19.40  4.57  −23.16  6.08  −21.28  5.65  0.017  Lean mass, kg  % Δ at 1 y  −17.03  9.95  −15.87  10.35  −16.45  10.06  0.688  Fat mass, kg  % Δ at 1 y  −36.11  14.27  −47.81  14.77  −41.96  15.54  0.006  View Large The ratio between the incremental AUCs of OGTT C-peptide and glycemia increased with time (from 127.52 ± 39.39 to 204.17 ± 150.40 nmol/mmol, P < 0.001). Figure 3 reports the mean concentrations of plasma glucose, C-peptide, and insulin after OGTT in the SG and RYGB groups at baseline and 1 year. The disposition index increased significantly with both treatments (from 209.62 ± 229.71 to 4130.75 ± 5677.94 ×1014 dL/kg/min−1 per pmol/L−1, P < 0.001). Figure 3. View largeDownload slide Plasma glucose, C-peptide, and insulin concentrations before (solid line) and after (dashed line) sleeve gastrectomy (solid circle) or gastric bypass (solid square). Disposition index (i.e., the product between β-cell glucose insulin sensitivity and whole body insulin sensitivity) and ΦD and ΦS (i.e., the dynamic and static components of the insulin secretion) before and after SG (open bar) or gastric bypass (filled bar). Figure 3. View largeDownload slide Plasma glucose, C-peptide, and insulin concentrations before (solid line) and after (dashed line) sleeve gastrectomy (solid circle) or gastric bypass (solid square). Disposition index (i.e., the product between β-cell glucose insulin sensitivity and whole body insulin sensitivity) and ΦD and ΦS (i.e., the dynamic and static components of the insulin secretion) before and after SG (open bar) or gastric bypass (filled bar). Weight loss and BMI The highest weight loss change was observed at 1 year after RYGB: −34.47 ± 6.81% after RYGB and −30.93 ± 7.38% after SG (P = 0.010). Changes in weight reflected changes in BMI (Table 1) with significant between-groups difference (P = 0.008). Lipid profile Total and LDL cholesterol decreased significantly in both groups (P = 0.001 and P < 0.001 in RYGB patients; P < 0.001 and P = 0.003 in SG patients, respectively). Although total cholesterol Δ did not reach statistical significance between groups, changes in LDL cholesterol differed between the two surgical procedures at 1 year: −29.25 ± 15.84 in RYGB and −14.99 ± 28.21 in SG (P = 0.008). HDL cholesterol increased in both groups; although changes in RYGB were highly significant (20.92 ± 43.23, P = 0.004), they were borderline in the SG group (8.42 ± 25.15, P = 0.046). Δ differences between groups were not significant. Triglycerides decreased significantly only in the SG group (−14.97 ± 40.08, P = 0.019). Arterial blood pressure Systolic and diastolic blood pressure slightly decreased in both groups, with no substantial difference between groups. Only the change in Δ systolic blood pressure was significantly different from zero for SG patients (−3.81 ± 8.91, P = 0.005). At baseline, 9 RYGB and 12 SG patients had hypertension; 44% RYGB and 50% SG hypertensive patients normalized blood pressure at 1 year without changes in type and dose of medications. Stepwise regressions No variable significantly predicted the binomial outcome presence/absence of hypoglycemia of any severity, whereas Δ BMI was borderline (β = −1.70e-01, SE = 9.25e-02, P = 0.07). Instead, when the primary end point (i.e. glycemia <3.1 mmol/L) was considered, both treatments (β = 2.33, SE = 1.07, P = 0.03) and Δ-BMI (β = −0.21, SE = 0.10, P = 0.03; with the probability of hypoglycemic events increasing in the RYGB group) were significant predictors of hypoglycemia. In other words, the larger the BMI reduction, the higher the number of hypoglycemic events in response to the OGTT. For each minimal model index, ϕs and dynamic control of insulin secretion in the fed state (ϕd), a regression model was built with a stepwise procedure to take into account multicollinearity to test if these parameters could be related to some baseline variables or to variable values recorded at 1 year. Each model parameter was related to age and sex and to their baseline values, BMI, glycemia, insulinemia, and to BMI, glycemia, insulinemia, and insulin sensitivity, at 1 year. In the model for the variable ϕs, only Δ BMI was significant (β = 4.33, SE = 1.79, P = 0.020). The ϕd was related significantly to the type of surgery (β = −601.35, SE = 226.27, P = 0.011), decreasing if the patients underwent RYGB, to baseline fasting insulinemia (β = −55.98, SE = 17.42, P = 0.0025) and to the BMI at 1 year after surgery (β = 56.72, SE = 23.94, P = 0.022). Total insulin secretion was predicted directly by baseline BMI (β = 1.87e+03, SE = 8.431e+02, P = 0.032) and inversely by insulin sensitivity at 1 year (β = −1.78e+02, SE=5.35e+01, P = 0.002). Insulin sensitivity at one year was predicted by sex (β = 48.94, SE = 15.84, P = 0035), being higher in women, baseline BMI (β = 6.78, SE = 2.58, P = 0.012) and BMI at 1 year after surgery (β = −4.42, SE = 2.21, P = 0.052). Finally, we checked for baseline characteristics that could predict hypoglycemia after surgery. When patients were considered altogether, the only significant predictor was age (β = −0.061, SE = 0.025 and P = 0.015). Age was also a significant predictor in the RYGB group (β = −0.084, SE = 0.036, P = 0.021), whereas in the SG group the baseline BMI was the only significant predictor (β = 0.19, SE = 0.097, P = 0.044). Adverse events There were no deaths or excessive weight loss in any group. Reintervention was not needed. Intravenous treatment of dehydration was necessary in 4 patients after RYGB and in 3 patients after SG. Cholelithiasis occurred in 2 RYGB patients. Nutritional deficiencies (i.e., anemia, low serum ferritin, hypoalbuminemia, hypocalcaemia, raised parathyroid hormone, and low vitamin D) were noted in 18 patients of the RYGB group and in 11 of the SG group. Discussion Our study demonstrates that the 1-year incidence of hypoglycemia (defined as <3.1 mmol/L or <55 mg/dL) after an OGTT was not significantly different between RYGB and SG (29% vs 14%, P = 0.079). Everyday hypoglycemic events, as measured with CGM, were also not significantly different. Yet, our multivariate analysis showed more hypoglycemia in the RYGB group as well as differences in hypoglycemia after OGTT using higher cutoffs (≤3.9 or ≤3.3 mmol/L). A more rapid recovery in the glycemic trough was also observed in SG as compared with RYGB subjects. Severe hypoglycemia that required hospitalization occurred in 6.8% of RYGB treatments, but never in SG-treated patients. We found substantial differences on the impact of both surgeries on parameters of insulin secretion that may contribute to hypoglycemia. We used <3.1 mmol/L to define hypoglycemia as our primary end point based on evidence that glucose concentrations under 3 mmol/L are associated with clinical harm. Cognitive impairment during both experimental and spontaneous hypoglycemia in people with T1D is detectable at 3 mmol/L or less (20), and exposure to 3 mmol/L can induce symptoms as well as counterregulatory hormonal responses to hypoglycemia in clamp studies (21, 22). T2D subjects who fail to recognize hypoglycemic symptoms until their glucose falls below 3 mmol/L (54 mg/dL) are nearly 8 times more likely to experience severe episodes of hypoglycemia (23) (blood glucose below 2.8 to 3 mmol/L). A recent conjunct position statement of the American Diabetes Association and of the European Association for the Study of Diabetes supports this definition, opting for the slightly stricter criterion of <3 mmol/L or 54 mg/dL (24). We also looked at <3.9 mmol/L, or 70 mg/dL, as this value is associated with the start of counter-regulatory response in health and has been defined as hypoglycemia by many authorities in the past. In the literature, RYGB reversal was associated with resolution of hypoglycemia in 13/17 (76%) (25) and 3/4 (75%) (26) of the patients with severe hypoglycemia. Conversion from RYGB to SG is currently used as an alternative treatment strategy to pancreatectomy for severe hypoglycemia following RYGB (27, 28). However, this conversion not only often fails to resolve the hypoglycemic symptoms but is also associated with high rates of major complications and hospital readmissions, and with needs for supplemental nutrition (29). Our study shows that SG is not a safer option than RYGB when assessed in terms of incidence rate of hypoglycemia after OGTT 1 year after surgery. In a recent study Salehi et al. (30) administered a liquid mixed meal during a hyperinsulinemic-hypoglycemic clamp with a target of 3 to 3.5 mmol/L in patients who underwent RYGB, in obese patients and in healthy controls. In spite of similar glycemic levels, insulin secretion was suppressed in controls but unsuppressed in RYGB subjects (30). The minimal model estimates two indexes, the static and the dynamic sensitivity to glucose. The static sensitivity ϕs measures the effect of glucose on β-cell secretion and is the ratio between insulin secretion rate and glucose concentrations, above a threshold level, at steady state. The ϕd is a measure of the stimulatory effect on insulin secretion of the rate at which glucose increases or decreases. We found that ϕd was reduced after RYGB, whereas the static control, ϕs, was improved. Therefore, after RYGB, the patients had an increased insulin secretion in response to increased plasma glucose concentrations, but were unable to reduce the secretion of insulin when glucose concentrations were falling. The hyperglycemic peak that follows an OGTT or a meal after RYGB or SG has been attributed to the accelerated gastric emptying associated with both types of operations (31). About one-third of the variability in the glycemic response to an OGTT, either in subjects with normal glucose tolerance or T2D, is due to the variability of the gastric emptying rate (32). Gastric emptying has been described to be equally accelerated after both RYGB and SG. In fact, in rats in which gastric emptying was evaluated by 99mTc-scintigraphy, the liquid meal mixture was fully (100%) emptied in only 5 minutes after both RYGB and SG, whereas only about 6% was emptied in sham-operated animals in the same period (33). In humans (34), the half-emptying time evaluated by scintigraphy was significantly reduced after SG compared with baseline, down to <30 minutes. After RYGB, the scintigraphic gastric emptying time for both liquids and solids was also found to be <30 minutes (35). However, our data suggest that the postprandial hyperinsulinemia observed after RYGB is not the consequence of the rapid glycemic rise but rather the result of an inappropriately high insulin secretion. In fact, the hypoglycemic events were uncorrelated with the hyperglycemic peaks, which were observed after both RYGB and SG. The dysregulation in insulin secretion in conjunction with an impressive improvement of insulin sensitivity can help to explain why RYGB patients had an overall larger number of reactive hypoglycemic episodes than SG. Recently, Abrahamsson et al. (15) found significantly reduced symptoms and counterregulatory hormonal response to hypoglycemia 23 weeks after RYGB. The repeated episodes of hypoglycemia in our series can help explain why these patients are generally unaware of low glycemic levels. In fact, repeated hypoglycemia or a state of chronic hypoglycemia, such that occurring during a continuous hyperinsulinemic-hypoglycemic clamp for 4 consecutive days, are associated with a glycemic threshold for cognitive impairment that is lowered from 3.0 to 2.5 mmol/L (36). This suggests a progressive adaptation to hypoglycemia. Also, in patients with T2D, recurrent hypoglycemia reduces the glucose concentration needed for the counterregulatory response (37,). Although hypoglycemia is regarded as a life-threatening complication in diabetic individuals, very few data for nondiabetic subjects are reported in the literature. The increasing use of bariatric surgery, leading to frequent hypoglycemic episodes, makes it necessary to rethink the need for lifelong nutritional support for these patients, possibly together with a modification of the surgical technique such as, for example, a longer alimentary limb in RYGB. In fact, biliopancreatic diversion is rarely associated with hyper- and hypoglycemia (38). Looking for a predictor of hypoglycemia after gastrointestinal surgery, we found that the higher the reduction of BMI at 1 year, the higher the number of hypoglycemic events triggered by the oral glucose load. The minimal model parameter ϕd was more reduced in those RYGB patients who had the largest reduction in BMI at 1 year after surgery, suggesting that insulin secretion dysregulation is more likely in those subjects who experienced the largest weight loss. Women had a better improvement of insulin sensitivity than men; this might be a reason they were subjected to more hypoglycemic episodes. However, new, larger trials are necessary to identify those factors that can predict reactive hypoglycemia before surgery. Although the power of our study is high, an important limitation consists of the relatively short duration of follow-up (12 months), with evidence that a difference in the rate of OGTT hypoglycemia starts to appear at 6 months. In addition, we note the single-center, open-label nature of our study. However, we are continuing patient follow-up and we plan to publish the results at 3 and 5 years. Finally, lack of glucagon-like peptide 1 levels data during the OGTT is a limitation of our study because it is a major glucoregulatory player that improves β-cell glucose sensitivity. In conclusion, RYGB induced a rate of reactive hypoglycemia (<3.1 mmol/L) similar to that of SG, but more hypoglycemic events using glycemic higher cutoffs and 6.8% severe hypoglycemia with hospitalization. Hypoglycemic episodes increased in frequency over time after RYGB and appeared to be associated with an imbalance between impaired β-cell dynamic glucose sensitivity and dramatically enhanced total body insulin sensitivity. Abbreviations: Abbreviations: AUC area under the curve BMI body mass index CGM continuous glucose monitoring HDL high-density lipoprotein LDL low-density protein OGTT oral glucose tolerance test RYGB Roux-en-Y gastric bypass SE standard error SG sleeve gastrectomy T2D type 2 diabetes ϕ1 decrement occurring during follow-up ϕd dynamic control of insulin secretion in the fed state ϕs glucose sensitivity index Acknowledgments Financial Support: The study was conducted with internal funds of the Catholic University. Clinical Trial Information: ClinicalTrials.gov no. NCT01581801 (registered 20 April 2012). Author Contributions: E.C., G.M., M.R., S.P., and A.D.G. designed the study. E.C., V.S., and P.G. acquired the data. The data analysis was done by S.P. and A.D.G. Data interpretation was done by E.C., G.M., M.R., S.P., R.B., and A.D.G. E.C., G.M., S.P., M.R., and A.D.G. contributed to drafting the article. S.R.B., S.A., and A.L.B. made critical revisions. All authors approved the final version of the manuscript. E.C., G.M., V.S., and S.P. ensure that questions related to the accuracy or integrity of any part of the article were appropriately investigated and resolved. E.C., S.P., and G.M. are guarantors of the study Disclosure Summary: The authors have nothing to declare. References 1. Nguyen DM, El-Serag HB. The epidemiology of obesity. Gastroenterol Clin North Am . 2010; 39( 1): 1– 7. Google Scholar CrossRef Search ADS PubMed  2. Dietz WH Jr, Gortmaker SL. Do we fatten our children at the television set? Obesity and television viewing in children and adolescents. Pediatrics . 1985; 75( 5): 807– 812. Google Scholar PubMed  3. Puzziferri N, Roshek TB III, Mayo HG, Gallagher R, Belle SH, Livingston EH. Long-term follow-up after bariatric surgery: a systematic review. JAMA . 2014; 312( 9): 934– 942. Google Scholar CrossRef Search ADS PubMed  4. Khorgami Z, Andalib A, Corcelles R, Aminian A, Brethauer S, Schauer P. Recent national trends in the surgical treatment of obesity: sleeve gastrectomy dominates. Surg Obes Relat Dis . 2015; 11( 6): S6– S8. Google Scholar CrossRef Search ADS   5. Mingrone G, Panunzi S, De Gaetano A, Guidone C, Iaconelli A, Leccesi L, Nanni G, Pomp A, Castagneto M, Ghirlanda G, Rubino F. Bariatric surgery versus conventional medical therapy for type 2 diabetes. N Engl J Med . 2012; 366( 17): 1577– 1585. Google Scholar CrossRef Search ADS PubMed  6. Mingrone G, Panunzi S, De Gaetano A, Guidone C, Iaconelli A, Nanni G, Castagneto M, Bornstein S, Rubino F. Bariatric-metabolic surgery versus conventional medical treatment in obese patients with type 2 diabetes: 5 year follow-up of an open-label, single-centre, randomised controlled trial. Lancet . 2015; 386( 9997): 964– 973. Google Scholar CrossRef Search ADS PubMed  7. Schauer PR, Bhatt DL, Kirwan JP, Wolski K, Brethauer SA, Navaneethan SD, Aminian A, Pothier CE, Kim ES, Nissen SE, Kashyap SR; STAMPEDE Investigators. Bariatric surgery versus intensive medical therapy for diabetes--3-year outcomes. N Engl J Med . 2014; 370( 21): 2002– 2013. Google Scholar CrossRef Search ADS PubMed  8. Seaquist ER, Anderson J, Childs B, Cryer P, Dagogo-Jack S, Fish L, Heller SR, Rodriguez H, Rosenzweig J, Vigersky R. Hypoglycemia and diabetes: a report of a workgroup of the American Diabetes Association and the Endocrine Society. Diabetes Care . 2013; 36( 5): 1384– 1395. Google Scholar CrossRef Search ADS PubMed  9. Service GJ, Thompson GB, Service FJ, Andrews JC, Collazo-Clavell ML, Lloyd RV. Hyperinsulinemic hypoglycemia with nesidioblastosis after gastric-bypass surgery. N Engl J Med . 2005; 353( 3): 249– 254. Google Scholar CrossRef Search ADS PubMed  10. Roslin MS, Oren JH, Polan BN, Damani T, Brauner R, Shah PC. Abnormal glucose tolerance testing after gastric bypass. Surg Obes Relat Dis . 2013; 9( 1): 26– 31. Google Scholar CrossRef Search ADS PubMed  11. Lee CJ, Wood GC, Lazo M, Brown TT, Clark JM, Still C, Benotti P. Risk of post-gastric bypass surgery hypoglycemia in nondiabetic individuals: A single center experience. Obesity (Silver Spring) . 2016; 24( 6): 1342– 1348. Google Scholar CrossRef Search ADS PubMed  12. Marsk R, Jonas E, Rasmussen F, Näslund E. Nationwide cohort study of post-gastric bypass hypoglycaemia including 5,040 patients undergoing surgery for obesity in 1986-2006 in Sweden. Diabetologia . 2010; 53( 11): 2307– 2311. Google Scholar CrossRef Search ADS PubMed  13. Sarwar H, Chapman WH III, Pender JR, Ivanescu A, Drake AJ III, Pories WJ, Dar MS. Hypoglycemia after Roux-en-Y gastric bypass: the BOLD experience. Obes Surg . 2014; 24( 7): 1120– 1124. Google Scholar CrossRef Search ADS PubMed  14. Amiel SA, Tamborlane WV, Simonson DC, Sherwin RS. Defective glucose counterregulation after strict glycemic control of insulin-dependent diabetes mellitus. N Engl J Med . 1987; 316( 22): 1376– 1383. Google Scholar CrossRef Search ADS PubMed  15. Abrahamsson N, Börjesson JL, Sundbom M, Wiklund U, Karlsson FA, Eriksson JW. Gastric bypass reduces symptoms and hormonal responses in hypoglycaemia. Diabetes . 2016; 65( 9): 2667– 2675. Google Scholar CrossRef Search ADS PubMed  16. Tzovaras G, Papamargaritis D, Sioka E, Zachari E, Baloyiannis I, Zacharoulis D, Koukoulis G. Symptoms suggestive of dumping syndrome after provocation in patients after laparoscopic sleeve gastrectomy. Obes Surg . 2012; 22( 1): 23– 28. Google Scholar CrossRef Search ADS PubMed  17. Chen CY, Lee WJ, Lee HM, Chen JC, Ser KH, Lee YC, Chen SC. Laparoscopic conversion of gastric bypass complication to sleeve gastrectomy: technique and early results. Obes Surg . 2016; 26( 9): 2014– 2021. Google Scholar CrossRef Search ADS PubMed  18. Cryer PE, Axelrod L, Grossman AB, Heller SR, Montori VM, Seaquist ER, Service FJ; Endocrine Society. Evaluation and management of adult hypoglycemic disorders: an Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab . 2009; 94( 3): 709– 728. Google Scholar CrossRef Search ADS PubMed  19. Breda E, Cavaghan MK, Toffolo G, Polonsky KS, Cobelli C. Oral glucose tolerance test minimal model indexes of beta-cell function and insulin sensitivity. Diabetes . 2001; 50( 1): 150– 158. Google Scholar CrossRef Search ADS PubMed  20. Gonder-Frederick LA, Zrebiec JF, Bauchowitz AU, Ritterband LM, Magee JC, Cox DJ, Clarke WL. Cognitive function is disrupted by both hypo- and hyperglycemia in school-aged children with type 1 diabetes: a field study. Diabetes Care . 2009; 32( 6): 1001– 1006. Google Scholar CrossRef Search ADS PubMed  21. Heller SR, Cryer PE. Reduced neuroendocrine and symptomatic responses to subsequent hypoglycemia after 1 episode of hypoglycemia in nondiabetic humans. Diabetes . 1991; 40( 2): 223– 226. Google Scholar CrossRef Search ADS PubMed  22. Davis SN, Shavers C, Mosqueda-Garcia R, Costa F. Effects of differing antecedent hypoglycemia on subsequent counterregulation in normal humans. Diabetes . 1997; 46( 8): 1328– 1335. Google Scholar CrossRef Search ADS PubMed  23. Pistrosch F, Ganz X, Bornstein SR, Birkenfeld AL, Henkel E, Hanefeld M. Risk of and risk factors for hypoglycemia and associated arrhythmias in patients with type 2 diabetes and cardiovascular disease: a cohort study under real-world conditions. Acta Diabetol . 2015; 52( 5): 889– 895. Google Scholar CrossRef Search ADS PubMed  24. International Hypoglycaemia Study Group. Glucose concentrations of less than 3.0 mmol/L (54 mg/dL) should be reported in clinical trials: a joint position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care.  2017; 40: 155– 157. CrossRef Search ADS PubMed  25. Mala T. Postprandial hyperinsulinemic hypoglycemia after gastric bypass surgical treatment. Surg Obes Relat Dis . 2014; 10( 6): 1220– 1225. Google Scholar CrossRef Search ADS PubMed  26. Pernar LI, Kim JJ, Shikora SA. Gastric bypass reversal: a 7-year experience. Surg Obes Relat Dis . 2016; 12( 8): 1492– 1498. Google Scholar CrossRef Search ADS PubMed  27. Lakdawala M, Limas P, Dhar S, Remedios C, Dhulla N, Sood A, Bhasker AG. Laparoscopic revision of Roux-en-Y gastric bypass to sleeve gastrectomy: a ray of hope for failed Roux-en-Y gastric bypass. Asian J Endosc Surg . 2016; 9( 2): 122– 127. Google Scholar CrossRef Search ADS PubMed  28. Carter CO, Fernandez AZ, McNatt SS, Powell MS. Conversion from gastric bypass to sleeve gastrectomy for complications of gastric bypass. Surg Obes Relat Dis . 2016; 12( 3): 572– 576. Google Scholar CrossRef Search ADS PubMed  29. van Beek AP, Emous M, Laville M, Tack J. Dumping syndrome after esophageal, gastric or bariatric surgery: pathophysiology, diagnosis, and management. Obes Rev . 2017; 18( 1): 68– 85. Google Scholar CrossRef Search ADS PubMed  30. Salehi M, Woods SC, D’Alessio DA. Gastric bypass alters both glucose-dependent and glucose-independent regulation of islet hormone secretion. Obesity (Silver Spring) . 2015; 23( 10): 2046– 2052. Google Scholar CrossRef Search ADS PubMed  31. Holst JJ, Madsbad S. Mechanisms of surgical control of type 2 diabetes: GLP-1 is key factor. Surg Obes Relat Dis . 2016; 12( 6): 1236– 1242. Google Scholar CrossRef Search ADS PubMed  32. Horowitz M, Edelbroek MA, Wishart JM, Straathof JW. Relationship between oral glucose tolerance and gastric emptying in normal healthy subjects. Diabetologia . 1993; 36( 9): 857– 862. Google Scholar CrossRef Search ADS PubMed  33. Chambers AP, Smith EP, Begg DP, Grayson BE, Sisley S, Greer T, Sorrell J, Lemmen L, LaSance K, Woods SC, Seeley RJ, D’Alessio DA, Sandoval DA. Regulation of gastric emptying rate and its role in nutrient-induced GLP-1 secretion in rats after vertical sleeve gastrectomy. Am J Physiol Endocrinol Metab . 2014; 306( 4): E424– E432. Google Scholar CrossRef Search ADS PubMed  34. Kandeel AA, Sarhan MD, Hegazy T, Mahmoud MM, Ali MH. Comparative assessment of gastric emptying in obese patients before and after laparoscopic sleeve gastrectomy using radionuclide scintigraphy. Nucl Med Commun . 2015; 36( 8): 854– 862. Google Scholar CrossRef Search ADS PubMed  35. Hinder RA, Esser J, DeMeester TR. Management of gastric emptying disorders following the Roux-en-Y procedure. Surgery . 1988; 104( 4): 765– 772. Google Scholar PubMed  36. Boyle PJ, Nagy RJ, O’Connor AM, Kempers SF, Yeo RA, Qualls C. Adaptation in brain glucose uptake following recurrent hypoglycemia. Proc Natl Acad Sci USA . 1994; 91( 20): 9352– 9356. Google Scholar CrossRef Search ADS PubMed  37. Davis SN, Mann S, Briscoe VJ, Ertl AC, Tate DB. Effects of intensive therapy and antecedent hypoglycemia on counterregulatory responses to hypoglycemia in type 2 diabetes. Diabetes . 2009; 58( 3): 701– 709. Google Scholar CrossRef Search ADS PubMed  38. Mingrone G, Castagneto M. Bariatric surgery: unstressing or boosting the beta-cell? Diabetes Obes Metab . 2009; 11( Suppl 4): 130– 142. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Clinical Endocrinology and Metabolism Oxford University Press

Loading next page...
 
/lp/ou_press/incidence-of-hypoglycemia-after-gastric-bypass-vs-sleeve-gastrectomy-a-6RzvSllVYB
Publisher
Endocrine Society
Copyright
Copyright © 2018 Endocrine Society
ISSN
0021-972X
eISSN
1945-7197
D.O.I.
10.1210/jc.2017-01695
Publisher site
See Article on Publisher Site

Abstract

Abstract Context We compared the incidence of hypoglycemia after Roux-en-Y gastric bypass (RYGB) vs sleeve gastrectomy (SG). Design, Setting, and Main Outcome Measures Randomized, open-label trial conducted at the outpatient obesity clinic in a university hospital in Rome, Italy. The primary aim was the incidence of reactive hypoglycemia (<3.1 mmol/L after 75-g oral glucose load) at 1 year after surgery. Secondary aims were hypoglycemia under everyday life conditions, insulin sensitivity, insulin secretion, and lipid profile. Results Of 175 eligible patients, 120 were randomized 1:1 to RYGB or SG; 117 (93%) completed the 12-month follow-up. Reactive hypoglycemia was detected in 14% and 29% of SG and RYGB patients (P = 0.079), respectively, with the effect of treatment in multivariate analysis significant at P = 0.018. Daily hypoglycemic episodes during continuous glucose monitoring did not differ between groups (P = 0.75). Four of 59 RYGB subjects (6.8%) had 1 to 3 hospitalizations for symptomatic hypoglycemia vs 0 in SG. The static β-cell glucose sensitivity index increased after both treatments (P < 0.001), but the dynamic β-cell glucose sensitivity index increased significantly in SG (P = 0.008) and decreased in RYGB (P = 0.004 for time × treatment interaction). Whole-body insulin sensitivity increased about 10-fold in both groups. Conclusions We show that reactive hypoglycemia is no less common after SG and is not a safer option than RYGB, but RYGB is associated with more severe hypoglycemic episodes. This is likely due to the lack of improvement of β-cell sensitivity to changes in circulating glucose after RYGB, which determines an inappropriately high insulin secretion. During the past 40 years, the prevalence of obesity has doubled worldwide, with >1.3 billion adults, equivalent to 39% of the world’s population, being overweight, and >600 million, representing 13% of the population, being obese, in 2014 (1). Long-term weight loss is difficult to maintain, despite dietary and behavioral changes; weight loss attained at 1 year in lifestyle modification trials is only 1.6 kg on average (2). In contrast, bariatric surgery is very effective in maintaining weight loss, with an average of 32 kg or greater weight reduction at 5 years after treatment (3). Recently, sleeve gastrectomy (SG) outperformed in frequency the use of Roux-en-Y gastric bypass (RYGB) worldwide, representing 49% and 43% of global bariatric operations, respectively (4). Although bariatric/metabolic surgery has resulted in numerous health benefits, including type 2 diabetes (T2D) remission (5–7), it is also associated with serious metabolic side effects. Among these, hypoglycemia is a relevant and underestimated complication of RYGB leading to cognitive impairment when plasma glucose levels fall <3 mmol/L (54 mg/dL) (8). There is increasing evidence of severe hypoglycemia after RYGB, and some cases of postoperative nesidioblastosis have been described (9). Roslin et al. (10) found that, at 6 months after RYGB, 68% of patients had hypoglycemia at 2 hours during an oral glucose tolerance test (OGTT). Based on electronic medical record data, the incidence of clinical episodes of hypoglycemia was 13% in a retrospective study in nondiabetic RYGB patients (11). In contrast, studies based on inpatient hospitalization International Classification of Disease codes (12), or on self-reported data (13), show figures <1%. Yet, hospitalization is usually necessary only when severe hypoglycemia occurs. Repeated hypoglycemic episodes have been shown to impair the normal counterregulatory stress responses to subsequent hypoglycemia in diabetic individuals (14). Similar changes have been shown after RYGB, with clinical symptoms and counterregulatory hormonal and sympathetic nerve responses to hypoglycemia significantly reduced (15). Compared with RYGB, SG seems to have a much lower occurrence of reactive hypoglycemia, reported as ∼20% (16). Conversion from RYGB to SG is currently used as an alternative to pancreatectomy to treat severe hypoglycemia after RYGB (17). Although hypoglycemia following bariatric/metabolic surgery represents an important health issue, no randomized trials have been undertaken until now to compare the effect of SG with that of RYGB in relation to the incidence and severity of hypoglycemic episodes. The primary aim of our study was to conduct a 1-year randomized trial to compare the incidence of hypoglycemia after RYGB and SG. Secondary objectives were to measure the effects of these surgical procedures on metabolic processes associated with hypoglycemic risk, namely insulin sensitivity and insulin secretion, and on cardiovascular risk factors, body weight, and fat distribution. Methods Participants From December 2012 to December 2014, we screened 175 patients eligible for bariatric surgery at the Day-Hospital of Obesity and Related Disorders of the Catholic University in Rome (Fig. 1). A total of 120 patients were enrolled and randomly assigned 1:1 to RYGB or to SG. The study was reviewed and approved by the institutional human ethics committee (A1534/CE/2012) in accordance with national guidelines and the provisions of the Helsinki Declaration, as revised in 2000. All patients provided written informed consent to participate in the study, and additional written informed consent was obtained before surgery. Figure 1. View largeDownload slide Flow diagram outlining patient enrollment and outcomes. IGT, impaired glucose tolerance. Figure 1. View largeDownload slide Flow diagram outlining patient enrollment and outcomes. IGT, impaired glucose tolerance. Inclusion criteria were: age 25 to 65 years, body mass index (BMI) >40 or 35 to 40 kg/m2 in the presence of obesity complications, or ability to understand and comply with the study process. Exclusion criteria were history of T2D; previous bariatric surgery; or history of medical problems such as mental impairment, cancer, major cardiovascular or gastrointestinal or respiratory diseases, hormonal disorders, infections, or pregnancy. Diabetic patients were excluded to avoid confounding factors such as need for hypoglycemic agents. Intervention Patients were evaluated by a multidisciplinary team (including endocrinologists, surgeons, dietitians, and nurses) at baseline and at 1,3,6,9, and 12 months after SG or RYGB. Dietary suggestions including multiple meals and reduction of simple carbohydrates were provided at the moment of the discharge from the hospital after metabolic surgery. Hypoglycemia was assessed as follows: OGTT: Plasma glucose measurements during a 3-hour standard (75 g) OGTT conducted at 1 year postsurgery. Blood samples were drawn at fasting and at 30, 60, 90, 120, 150, and 180 minutes after the OGTT. In a subset of 50 randomly selected subjects, more frequent blood sampling was performed (i.e., at fasting and at 10, 20, 30, 40, 60, 80, 95, 120, 140, 155, and 180 minutes following the OGTT). OGTTs were performed at baseline and at 1, 3, 6, 9, and 12 months after surgery. Hypoglycemia awareness questionnaire at each visit relative to episodes of hypoglycemia in the 30 days preceding the visit. Hypoglycemia awareness was assessed by the Gold score, in which participants rated their awareness of hypoglycemia from 1 (always aware) to 7 (never aware) on a linear analog scale (18) at each visit. Episodes of severe hypoglycemia during the visit or during the previous 30 days, defined by Whipple triad (18), and confirmed by a finger prick blood glucose measurement, namely the association of neuroglycopenic and/or neurogenic symptoms, signs, or both, consistent with hypoglycemia, a plasma glucose concentration <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test, and resolution of those symptoms or signs after plasma glucose increased. Hospitalization for hypoglycemia. Blinded continuous glucose monitoring using Enlite Sensor (Medtronic), which participants wore for at least 5 days for 1 year. Fifty patients were randomly selected in each group; 25 at 1 year after RYGB and 25 at 1 year after SG were involved. An episode of hypoglycemia on continuous glucose monitoring (CGM) was defined as a single glycemic level <3.1 mmol/L, which is the average value of 5 minutes’ registration. Primary outcome The primary end point was the incidence of hypoglycemia (≤55 mg/dL or 3.1 mmol/L within 3 hours after ingestion of glucose during a 75-g glucose OGTT) at 1 year after surgery. Secondary outcomes Secondary end points were changes of body weight, BMI, symptomatic hypoglycemia, lipid profile, insulin sensitivity, insulin secretion during OGTT, abdominal circumferences, and body composition 1 year after surgery. Hypoglycemic events during everyday life by CGM were added as a secondary outcome in October 2014. Safety Adverse events from bariatric surgery were noted by the study nurse at each visit and subsequently classified by the study physician for potential causality (unlikely, possibly, or likely). Randomization A total of 120 patients requiring bariatric surgery for medical reasons met the study inclusion criteria after physician screening, and were subsequently randomly assigned by a computer-generated blocked random sequence to the RYGB or SG groups in a 1:1 ratio. Sample size Previous studies reported an absolute difference in the occurrence of hypoglycemia between the two surgical procedures of ∼50%, expecting an incidence of 68% in the RYGB group (10) vs 20% in the SG group (16). With a more conservative approach, the sample size was computed to detect a 30% difference between treatments, expecting an incidence of 50% in the RYGB group vs 20% in the SG group. Under this hypothesis, with two-tailed P < 0.05 and a power of 0.90, a sample size of 50 patients per group would have been sufficient. Considering an attrition rate of 20%, the number of patients in each group was set to 60 and, thus, overall to 120. Body composition Body weight was measured to the nearest 0.1 kg with a beam scale and height to the nearest 0.5 cm using a stadiometer. Hip and waist circumferences were measured in duplicate, with patients standing in underwear, as the maximal circumference over the buttocks and between the iliac crest and the lower ribcage, respectively. Body composition was measured by dual-energy X-ray absorptiometry (Lunar-iDXA). Laboratory analyses Fasting plasma glucose was measured by the glucose-oxidase method (Analox) and plasma insulin by microparticle enzyme immunoassay (Abbott), with 1 μU/mL sensitivity and 6.6% intra-assay coefficient of variation. Architect C-peptide (Abbott) assay precision of ≤10% total coefficient of variation was used to measure C-peptide levels. Total cholesterol, high-density lipoprotein (HDL) cholesterol, and triglyceride levels were measured by standard enzymatic assays. Low-density lipoprotein (LDL) cholesterol was calculated by the Friedewald formula. Mathematical modeling Insulin sensitivity and secretion were computed in the subset of 50 patients by the oral glucose minimal model (19). Changes over time of the indices were compared by a repeated measurement analysis of variance with time as within-factor and treatment as between-factor. Statistics Hypoglycemia during OGTT at 1 year after surgery was defined as any value of plasma glucose <3.1 mmol/L (55 mg/dL). During CGM, the average number of hypoglycemic episodes per day, defined as any glucose value <3.1 mmol/L, was also computed and compared between treatments by Mann-Whitney U test. Associations between treatments and presence/absence of hypoglycemic events were analyzed by χ2 or Fisher exact tests. For continuous variables means and standard deviations were computed. Student t tests were used to assess differences between groups at baseline. Percent changes at 1 year were computed as the differences between 1 year and baseline values divided by the baseline values, multiplied by 100. Deltas in the two groups were then compared by t test. Logistic regression with stepwise elimination was applied to identify possible predictors of hypoglycemic incidence. To study the possible relationship between the dependent binomial variable, presence/absence of hypoglycemia, and its regressors (independent variables), we used a logistic regression model with a multivariate variable selection procedure and a forward selection method; the latter was used to take into account multicollinearity. In particular, we included sex, age, BMI, triglycerides and total cholesterol, insulin sensitivity and insulin secretion at baseline, as well as variations in BMI, triglycerides, total cholesterol, insulin sensitivity and insulin secretion at 1 year after surgery, including the type of treatment as a possible explanatory variable. A logistic model relating presence/absence of hypoglycemia (glycemia <3.1 mmol/L) and baseline characteristics was carried out. The variables entering the model were age, sex, BMI, insulin sensitivity, and insulin secretion at baseline. The analysis was performed considering either all patients together, independent of the type of surgery, or separately (RYGB and SG). A nonlinear mixed-effects model was applied to both glycemic “trough” and “peak” variables. The model describing the trend over time of both the variables (y) is:  y=ϕ1×exp(−λ×time)+(ϕ3−ϕ1)Parameter ϕ3 represents the values of the dependent variables at baseline, parameter λ represents the decreasing rate of the variables, and ϕ1 represents the decrement that occurred during the follow-up. P < 0.05 was considered significant. All analyses were performed in R and Matlab (R2009b). Results Overall, 120 of 175 screened patients underwent randomization (Fig. 1).The follow-up at 1 year was 92.5%. Three patients withdrew their consent before the first visit after surgery and were excluded from the analysis. Two of the 3 patients who withdrew their consent in the SG arm after the visit at 6 months and 2 patients in the RYGB arm who withdrew after the visit at 3 months did so because of job-related changes in location. One patient withdrew consent in the SG arm after the 6-month visit did because of family problems. These patients were included in the intention-to-treat analysis. Three patients who abandoned the study immediately after surgery, 1 in the RYGB and 2 in the SG arm, respectively, did not give reasons for withdrawing prematurely from the trial and did not undergo the OGTT at 1 month and were therefore not included in the study. A total of 117 patients were therefore evaluated for the primary and secondary end points: 59 who underwent RYGB, age 43.07 ± 9.17 years, and 58 who underwent SG, age 45.76 ± 9.68 years. Between-groups baseline variables (Table 1) as well as sex (33.3% men and 66.7% women) did not differ significantly between treatment groups (P = 0.133). Table 1. Variables Averages and Their SD at Baseline and 1 y After Bariatric Surgery, With Percentage Changes     SG (n = 58)  RYGB (n = 59)  Overall (n = 117)        Mean  SD  Mean  SD  Mean  SD  P Value  Weight, kg  Baseline  121.57  18.32  124.19  13.30  122.89  15.96  0.390    After 1 y  84.04  16.33  81.13  9.95  82.57  13.51  0.262    % Δ at 1 y  −30.93  7.38  −34.47  6.81  −32.72  7.28  0.010  BMI, kg/m2  Baseline  43.44  4.25  43.10  3.96  43.27  4.09  0.666    After 1 y  30.02  3.93  28.27  3.81  29.14  3.95  0.019    % Δ at 1 y  −30.74  7.20  −34.33  6.89  −32.55  7.24  0.008  HbA1c, mmol/mol  Baseline  38.17  3.13  40.00  7.40  39.12  5.80  0.095    After 1 y  34.62  2.44  34.91  4.61  34.77  3.69  0.706    % Δ at 1 y  −8.85  7.89  −10.29  11.70  −9.58  9.99  0.488  Glycemia, mmol/L  Baseline  5.15  0.50  5.09  0.56  5.12  0.53  0.590    After 1 y  4.54  0.51  4.48  0.50  4.51  0.50  0.517    % Δ at 1 y  −11.06  12.63  −11.34  12.76  −11.20  12.64  0.911  Total cholesterol, mmol/L  Baseline  5.11  0.94  4.89  0.97  5.00  0.96  0.250    After 1 y  4.25  0.74  4.00  0.77  4.12  0.77  0.128    % Δ at 1 y  −14.56  19.80  −15.51  28.62  −15.05  24.62  0.855  HDL cholesterol, mmol/L  Baseline  1.29  0.29  1.26  0.27  1.28  0.28  0.669    after 1 year  1.34  0.26  1.49  0.30  1.42  0.29  0.024    % Δ at 1 y  8.42  25.15  20.92  43.23  14.83  35.92  0.121  LDL cholesterol, mmol/L  Baseline  3.02  0.88  2.97  0.70  2.99  0.79  0.742    After 1 y  2.48  0.74  2.13  0.61  2.29  0.69  0.023    % Δ at 1 y  −14.99  28.21  −29.25  15.84  −22.31  23.68  0.008  Triglycerides, mmol/L  Baseline  1.56  0.91  1.35  0.77  1.45  0.84  0.194    After 1 y  1.13  0.44  0.97  0.44  1.05  0.45  0.083    % Δ at 1 y  −14.97  40.08  −5.55  77.56  −10.15  61.97  0.479  Systolic blood pressure, mm Hg  Baseline  127.53  10.27  129.25  15.08  128.39  12.87  0.495    After 1 y  122.36  9.68  125.55  8.77  124.05  9.30  0.078    % Δ at 1 y  −3.81  8.91  −2.22  8.20  −2.98  8.54  0.352  Diastolic blood pressure, mm Hg  Baseline  80.72  6.97  83.68  10.01  82.20  8.71  0.080    After 1 y  78.54  6.96  80.54  5.45  79.59  6.26  0.102    % Δ at 1 y  −2.38  9.28  −2.82  12.18  −2.61  10.85  0.839  Hypoglycemia assessed by OGTT, %  Baseline  0    0    0    —    At 1 y  14    29    21.4    0.047  Episodes of severe hypoglycemia assessed by questionnaire, %  Baseline  0    0    0    —    At 1 y  0    5.1    2.6    0.082  Episodes of nonsevere hypoglycemia (Whipple triad plus glycemia <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test) during the visit at 1 y  Baseline  0    0    0    —    At 1 y  0    0    0    —  Hypoglycemia assessed during CGMa  At 1 y  32    24    28    0.533      SG (n = 58)  RYGB (n = 59)  Overall (n = 117)        Mean  SD  Mean  SD  Mean  SD  P Value  Weight, kg  Baseline  121.57  18.32  124.19  13.30  122.89  15.96  0.390    After 1 y  84.04  16.33  81.13  9.95  82.57  13.51  0.262    % Δ at 1 y  −30.93  7.38  −34.47  6.81  −32.72  7.28  0.010  BMI, kg/m2  Baseline  43.44  4.25  43.10  3.96  43.27  4.09  0.666    After 1 y  30.02  3.93  28.27  3.81  29.14  3.95  0.019    % Δ at 1 y  −30.74  7.20  −34.33  6.89  −32.55  7.24  0.008  HbA1c, mmol/mol  Baseline  38.17  3.13  40.00  7.40  39.12  5.80  0.095    After 1 y  34.62  2.44  34.91  4.61  34.77  3.69  0.706    % Δ at 1 y  −8.85  7.89  −10.29  11.70  −9.58  9.99  0.488  Glycemia, mmol/L  Baseline  5.15  0.50  5.09  0.56  5.12  0.53  0.590    After 1 y  4.54  0.51  4.48  0.50  4.51  0.50  0.517    % Δ at 1 y  −11.06  12.63  −11.34  12.76  −11.20  12.64  0.911  Total cholesterol, mmol/L  Baseline  5.11  0.94  4.89  0.97  5.00  0.96  0.250    After 1 y  4.25  0.74  4.00  0.77  4.12  0.77  0.128    % Δ at 1 y  −14.56  19.80  −15.51  28.62  −15.05  24.62  0.855  HDL cholesterol, mmol/L  Baseline  1.29  0.29  1.26  0.27  1.28  0.28  0.669    after 1 year  1.34  0.26  1.49  0.30  1.42  0.29  0.024    % Δ at 1 y  8.42  25.15  20.92  43.23  14.83  35.92  0.121  LDL cholesterol, mmol/L  Baseline  3.02  0.88  2.97  0.70  2.99  0.79  0.742    After 1 y  2.48  0.74  2.13  0.61  2.29  0.69  0.023    % Δ at 1 y  −14.99  28.21  −29.25  15.84  −22.31  23.68  0.008  Triglycerides, mmol/L  Baseline  1.56  0.91  1.35  0.77  1.45  0.84  0.194    After 1 y  1.13  0.44  0.97  0.44  1.05  0.45  0.083    % Δ at 1 y  −14.97  40.08  −5.55  77.56  −10.15  61.97  0.479  Systolic blood pressure, mm Hg  Baseline  127.53  10.27  129.25  15.08  128.39  12.87  0.495    After 1 y  122.36  9.68  125.55  8.77  124.05  9.30  0.078    % Δ at 1 y  −3.81  8.91  −2.22  8.20  −2.98  8.54  0.352  Diastolic blood pressure, mm Hg  Baseline  80.72  6.97  83.68  10.01  82.20  8.71  0.080    After 1 y  78.54  6.96  80.54  5.45  79.59  6.26  0.102    % Δ at 1 y  −2.38  9.28  −2.82  12.18  −2.61  10.85  0.839  Hypoglycemia assessed by OGTT, %  Baseline  0    0    0    —    At 1 y  14    29    21.4    0.047  Episodes of severe hypoglycemia assessed by questionnaire, %  Baseline  0    0    0    —    At 1 y  0    5.1    2.6    0.082  Episodes of nonsevere hypoglycemia (Whipple triad plus glycemia <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test) during the visit at 1 y  Baseline  0    0    0    —    At 1 y  0    0    0    —  Hypoglycemia assessed during CGMa  At 1 y  32    24    28    0.533  Abbreviation: HbA1c, glycosylated hemoglobin. a CGM was performed only on a subsample of 50 subjects, 25 per group, and only after surgery. View Large Table 1. Variables Averages and Their SD at Baseline and 1 y After Bariatric Surgery, With Percentage Changes     SG (n = 58)  RYGB (n = 59)  Overall (n = 117)        Mean  SD  Mean  SD  Mean  SD  P Value  Weight, kg  Baseline  121.57  18.32  124.19  13.30  122.89  15.96  0.390    After 1 y  84.04  16.33  81.13  9.95  82.57  13.51  0.262    % Δ at 1 y  −30.93  7.38  −34.47  6.81  −32.72  7.28  0.010  BMI, kg/m2  Baseline  43.44  4.25  43.10  3.96  43.27  4.09  0.666    After 1 y  30.02  3.93  28.27  3.81  29.14  3.95  0.019    % Δ at 1 y  −30.74  7.20  −34.33  6.89  −32.55  7.24  0.008  HbA1c, mmol/mol  Baseline  38.17  3.13  40.00  7.40  39.12  5.80  0.095    After 1 y  34.62  2.44  34.91  4.61  34.77  3.69  0.706    % Δ at 1 y  −8.85  7.89  −10.29  11.70  −9.58  9.99  0.488  Glycemia, mmol/L  Baseline  5.15  0.50  5.09  0.56  5.12  0.53  0.590    After 1 y  4.54  0.51  4.48  0.50  4.51  0.50  0.517    % Δ at 1 y  −11.06  12.63  −11.34  12.76  −11.20  12.64  0.911  Total cholesterol, mmol/L  Baseline  5.11  0.94  4.89  0.97  5.00  0.96  0.250    After 1 y  4.25  0.74  4.00  0.77  4.12  0.77  0.128    % Δ at 1 y  −14.56  19.80  −15.51  28.62  −15.05  24.62  0.855  HDL cholesterol, mmol/L  Baseline  1.29  0.29  1.26  0.27  1.28  0.28  0.669    after 1 year  1.34  0.26  1.49  0.30  1.42  0.29  0.024    % Δ at 1 y  8.42  25.15  20.92  43.23  14.83  35.92  0.121  LDL cholesterol, mmol/L  Baseline  3.02  0.88  2.97  0.70  2.99  0.79  0.742    After 1 y  2.48  0.74  2.13  0.61  2.29  0.69  0.023    % Δ at 1 y  −14.99  28.21  −29.25  15.84  −22.31  23.68  0.008  Triglycerides, mmol/L  Baseline  1.56  0.91  1.35  0.77  1.45  0.84  0.194    After 1 y  1.13  0.44  0.97  0.44  1.05  0.45  0.083    % Δ at 1 y  −14.97  40.08  −5.55  77.56  −10.15  61.97  0.479  Systolic blood pressure, mm Hg  Baseline  127.53  10.27  129.25  15.08  128.39  12.87  0.495    After 1 y  122.36  9.68  125.55  8.77  124.05  9.30  0.078    % Δ at 1 y  −3.81  8.91  −2.22  8.20  −2.98  8.54  0.352  Diastolic blood pressure, mm Hg  Baseline  80.72  6.97  83.68  10.01  82.20  8.71  0.080    After 1 y  78.54  6.96  80.54  5.45  79.59  6.26  0.102    % Δ at 1 y  −2.38  9.28  −2.82  12.18  −2.61  10.85  0.839  Hypoglycemia assessed by OGTT, %  Baseline  0    0    0    —    At 1 y  14    29    21.4    0.047  Episodes of severe hypoglycemia assessed by questionnaire, %  Baseline  0    0    0    —    At 1 y  0    5.1    2.6    0.082  Episodes of nonsevere hypoglycemia (Whipple triad plus glycemia <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test) during the visit at 1 y  Baseline  0    0    0    —    At 1 y  0    0    0    —  Hypoglycemia assessed during CGMa  At 1 y  32    24    28    0.533      SG (n = 58)  RYGB (n = 59)  Overall (n = 117)        Mean  SD  Mean  SD  Mean  SD  P Value  Weight, kg  Baseline  121.57  18.32  124.19  13.30  122.89  15.96  0.390    After 1 y  84.04  16.33  81.13  9.95  82.57  13.51  0.262    % Δ at 1 y  −30.93  7.38  −34.47  6.81  −32.72  7.28  0.010  BMI, kg/m2  Baseline  43.44  4.25  43.10  3.96  43.27  4.09  0.666    After 1 y  30.02  3.93  28.27  3.81  29.14  3.95  0.019    % Δ at 1 y  −30.74  7.20  −34.33  6.89  −32.55  7.24  0.008  HbA1c, mmol/mol  Baseline  38.17  3.13  40.00  7.40  39.12  5.80  0.095    After 1 y  34.62  2.44  34.91  4.61  34.77  3.69  0.706    % Δ at 1 y  −8.85  7.89  −10.29  11.70  −9.58  9.99  0.488  Glycemia, mmol/L  Baseline  5.15  0.50  5.09  0.56  5.12  0.53  0.590    After 1 y  4.54  0.51  4.48  0.50  4.51  0.50  0.517    % Δ at 1 y  −11.06  12.63  −11.34  12.76  −11.20  12.64  0.911  Total cholesterol, mmol/L  Baseline  5.11  0.94  4.89  0.97  5.00  0.96  0.250    After 1 y  4.25  0.74  4.00  0.77  4.12  0.77  0.128    % Δ at 1 y  −14.56  19.80  −15.51  28.62  −15.05  24.62  0.855  HDL cholesterol, mmol/L  Baseline  1.29  0.29  1.26  0.27  1.28  0.28  0.669    after 1 year  1.34  0.26  1.49  0.30  1.42  0.29  0.024    % Δ at 1 y  8.42  25.15  20.92  43.23  14.83  35.92  0.121  LDL cholesterol, mmol/L  Baseline  3.02  0.88  2.97  0.70  2.99  0.79  0.742    After 1 y  2.48  0.74  2.13  0.61  2.29  0.69  0.023    % Δ at 1 y  −14.99  28.21  −29.25  15.84  −22.31  23.68  0.008  Triglycerides, mmol/L  Baseline  1.56  0.91  1.35  0.77  1.45  0.84  0.194    After 1 y  1.13  0.44  0.97  0.44  1.05  0.45  0.083    % Δ at 1 y  −14.97  40.08  −5.55  77.56  −10.15  61.97  0.479  Systolic blood pressure, mm Hg  Baseline  127.53  10.27  129.25  15.08  128.39  12.87  0.495    After 1 y  122.36  9.68  125.55  8.77  124.05  9.30  0.078    % Δ at 1 y  −3.81  8.91  −2.22  8.20  −2.98  8.54  0.352  Diastolic blood pressure, mm Hg  Baseline  80.72  6.97  83.68  10.01  82.20  8.71  0.080    After 1 y  78.54  6.96  80.54  5.45  79.59  6.26  0.102    % Δ at 1 y  −2.38  9.28  −2.82  12.18  −2.61  10.85  0.839  Hypoglycemia assessed by OGTT, %  Baseline  0    0    0    —    At 1 y  14    29    21.4    0.047  Episodes of severe hypoglycemia assessed by questionnaire, %  Baseline  0    0    0    —    At 1 y  0    5.1    2.6    0.082  Episodes of nonsevere hypoglycemia (Whipple triad plus glycemia <3.9, <3.4, or 3.1 mmol/L assessed by finger prick test) during the visit at 1 y  Baseline  0    0    0    —    At 1 y  0    0    0    —  Hypoglycemia assessed during CGMa  At 1 y  32    24    28    0.533  Abbreviation: HbA1c, glycosylated hemoglobin. a CGM was performed only on a subsample of 50 subjects, 25 per group, and only after surgery. View Large Primary outcome Percentages of patients experiencing hypoglycemia (<55 mg/dL or 3.1 mmol/L) after OGTT at 1 year were 14% and 29% in the SG and RYGB groups, respectively (P = 0.079). The percentages of patients experiencing hypoglycemia after OGTT over time was higher in RYGB than in the SG group, with a maximum at 6 and 3 months, respectively (Fig. 2). Figure 2. View largeDownload slide Average number of hypoglycemic episodes (glycemia ≤3.1 mmol/L) following the OGTT at 1 year after sleeve gastrectomy (blue solid line) and after gastric bypass (red dashed line). Figure 2. View largeDownload slide Average number of hypoglycemic episodes (glycemia ≤3.1 mmol/L) following the OGTT at 1 year after sleeve gastrectomy (blue solid line) and after gastric bypass (red dashed line). About 5% of the RYGB and none of the SG patients referred to symptoms of hypoglycemia in their questionnaire (Table 1). Absence of Whipple triad associated with glycemia <3.9, < 3.4, or 3.1 mmol/L, as assessed by finger-prick test, excluded symptomatic hypoglycemia during the periodic visits and OGTT in all patients. In a mixed-effects model, with the occurrence of hypoglycemia as the dependent variable, both time and treatment were significant contributory factors (P < 0.001 and P = 0.01, respectively). When postsurgery weight was introduced as a covariate, time was no longer significant, with estimates of the regression coefficients of 1.50 (P = 0.018) for RYGB vs SG, 0.04 (P = 0.5) for time, and −0.04 (P = 0.020) for weight. The odds ratio for hypoglycemia was 4.5 times higher for RYGB than for SG (P = 0.018). An OGTT plasma glucose value ≤3.9 mmol/L was recorded in 67.2% of SG and in 91.5% of RYGB patients (P = 0.002). Glucose levels ≤3.3 mmol/L were 36.2% and 61% in the SG and RYGB group, respectively (P = 0.012). Glycemia ≤2.8 mmol/L was recorded in 5.2% of SG and in 18.6% of RYGB patients (P = 0.050). Glycemia ≤2.2 mmol/L occurred only in 2 RYGB patients (P = 0.50). Four of 59 subjects who completed the 1-year follow-up in the RYGB group had 1 to 3 hospitalizations for symptomatic, neuroglycopenic hypoglycemia between month 9 and 12 after surgery. Incidence of hyperglycemia during OGTT Hyperglycemia (≥10 mmol/L) during OGTT occurred in 13.8% of SG and 22% of RYGB patients (P = 0.36). The association at 1 year between hypoglycemic and hyperglycemic episodes, independent of treatment, was not significant (P = 0.56). Glycemic trough and peak after OGTT The values of both plasma glucose trough and peak during the OGTT were evaluated for each patient. Trough values decreased at 1 year and Δ was significantly different between groups (−30.9 ± 24.7 in RYGB vs −21.7 ± 21.7 in SG, P = 0.040). Also, the mixed-effects model showed a significantly greater decrement of the trough values in RYGB patients: ϕ1 = 21.0 ± 2.22 [standard error (SE)] for SG vs ϕ1 + 6.0 ± 1.88 for RYGB, P = 0.0006. Peak glycemic values decreased less in RYGB than in SG with Δ of −29.9 ± 44.06 and of −47.7 ± 30.55 (P = 0.015), respectively. RYGB patients had a numerically lower (P = 0.094) glycemic peak value at baseline, which did not reach significance, and experienced a significantly smaller decrement in the attained maximum values of glycemia: ϕ1 = 46.60 ± 4.5 (SE) for SG vs ϕ1 − 16.67 ± 6.3 for RYGB, P = 0.0087. Supplemental Figs. 1 and 2 report the average (dashed lines) and predicted values (continuous lines) of the trough and peak variables for SG and RYGB, respectively. Glycated hemoglobin Glycated hemoglobin decreased significantly at 1 year: changes were −10.29 ± 11.70% (P < 0.001) in RYGB and −8.85 ± 7.89% (P < 0.001) in the SG groups, with no significant difference between groups (P = 0.50). Also, fasting glycemia decreased significantly in both groups without significant differences between groups: % Δ −11.34 ± 12.76 in RYGB (P < 0.001) and −11.06 ± 12.63, P < 0.001 in SG; absolute values: 4.48 ± 0.50 vs 4.54 ± 0.51 mmol/L in RYGB and SG, respectively (P = 0.52). Incidence of hypoglycemia during CGM The average daily glycemic values ≤3.1 mmol/L did not differ between groups: 5.47 ± 11.35 in the SG vs 2.76 ± 6.64 in RYGB (P = 0.75) groups. Eight patients in the SG group and 6 in the RYGB group experienced 1 to 4 hypoglycemic episodes (continuous periods of hypoglycemia during CGM); overall, 14 hypoglycemic episodes were recorded in SG and 9 in RYGB patients, with an average duration of 48.1 ± 37.2 and 45.6 ± 32.8 minutes, respectively (P = 1 by Mann-Whitney U test). Fifty-seven percent of the hypoglycemic episodes in the SG group were postprandial, whereas 43% were nocturnal. In the RYGB group, however, most occurred during the night (55%), and two were postprandial and two occurred in the afternoon at some distance from meal consumption. The daily average number of hypoglycemic episodes was 1.26 ± 0.81 and 1.77 ± 2.14 (P = 0.95 by Mann-Whitney U test) in the SG and RYGB groups, respectively. The mean glycemic values during hypoglycemic episodes in the 8 SG and 6 RYGB patients were 2.73 ± 0.27 and 2.98 ± 0.06 mmol/L, respectively (P = 0.10 by Mann-Whitney U test). We report in Supplemental Table 1 the study variables for patients with and without hypoglycemia. The only significantly different variables between the two groups were weight percent changes as well as 1-year total cholesterol and triglyceride levels that were lower in the group with hypoglycemia. Insulin sensitivity and insulin secretion during OGTT Results from the minimal model analysis are reported in Table 2. The static β-cell glucose sensitivity index ϕs increased after both treatments (50.99 ± 33.68 at baseline vs 120.43 ± 98.90 ×109/min–1 at 1 year, P < 0.001), whereas there were no significant difference between RYGB and SG. A similar trend was recorded for the global β-cell glucose sensitivity. In contrast, dynamic β-cell glucose sensitivity had a different outcome; it increased significantly (P = 0.008) in the SG group (from 324 ± 481 to 933 ± 1064 ×109/min), but decreased in the RYGB group (from 646 ± 733 to 415 ± 470 ×109/min) (P = 0.004 for time × treatment interaction). Whole-body insulin sensitivity increased from 3.76 ± 3.09 to 36.26 ± 57.09 ×105 dL/kg/min−1 (pmol/L)−1, whereas insulin secretion rate (ISR) area under the curve (AUC) decreased from 140.6 ± 54.8 to 64.6 ± 24.2 nmol. Table 2. OGTT Minimal Model β-cell Glucose Sensitivity, Whole Body Insulin Sensitivity Indices, and Body Composition Variables    SG (n = 25)  RYGB (n = 25)  Overall (n = 50)      Mean  SD  Mean  SD  Mean  SD  P Value  Static β-cell glucose sensitivity, ×109/min−1  Baseline  53.67  35.55  48.30  32.22  50.99  33.68  0.58    After 1 y  114.45  100.61  126.41  98.86  120.43  98.90  0.67  Dynamic β-cell glucose sensitivity, 109  Baseline  323.93  480.87  645.69  733.21  484.81  634.81  0.07    After 1 y  933.32  1063.86  414.90  469.87  674.11  855.01  0.03  Global β-cell glucose sensitivity, ×109/min−1  Baseline  57.32  35.51  54.71  34.44  56.01  34.65  0.79    After 1 y  130.37  104.36  135.74  99.97  133.06  101.17  0.85  Whole body insulin sensitivity, 105 dL/kg per min per pmol/L  Baseline  3.28  2.80  4.24  3.34  3.76  3.09  0.28    After 1 y  36.87  59.75  35.66  55.54  36.26  57.10  0.94  Insulin secretion rate AUC, nmol  Baseline  136.11  52.26  145.20  58.02  140.65  54.84  0.56    After 1 y  66.5  22.21  62.51  26.35  64.63  24.21  0.54  Disposition index, 1014 dL/kg/min−2 per pmol/L  Baseline  202.00  259.66  217.25  200.46  209.62  229.71  0.82    After 1 y  4532.68  6329.19  3728.83  5042.44  4130.75  5677.94  0.62  C-peptide AUC over glucose AUC above basal levels, nmol/pmol  Baseline  135.39  36.87  119.66  41.00  127.52  39.39  0.16    After 1 y  179.70  118.98  228.63  175.45  204.17  150.40  0.25  C-peptide AUC over glucose AUC, nmol/pmol  Baseline  118.79  33.29  112.84  36.42  115.81  34.66  0.55    After 1 y  100.32  28.17  114.40  49.09  107.36  40.25  0.22  Waist circumference, cm  Baseline  125.76  8.50  127.64  11.53  126.70  10.07  0.515    After 1 y  101.52  9.97  97.84  10.75  99.68  10.43  0.215    % Δ at 1 y  −19.29  5.40  −23.32  5.29  −21.31  5.67  0.010  Hip circumference, cm  Baseline  131.68  9.76  133.52  10.37  132.60  10.01  0.521    After 1 y  106.12  9.80  102.72  12.23  104.42  11.10  0.283    % Δ at 1 y  −19.40  4.57  −23.16  6.08  −21.28  5.65  0.017  Lean mass, kg  % Δ at 1 y  −17.03  9.95  −15.87  10.35  −16.45  10.06  0.688  Fat mass, kg  % Δ at 1 y  −36.11  14.27  −47.81  14.77  −41.96  15.54  0.006  Variables    SG (n = 25)  RYGB (n = 25)  Overall (n = 50)      Mean  SD  Mean  SD  Mean  SD  P Value  Static β-cell glucose sensitivity, ×109/min−1  Baseline  53.67  35.55  48.30  32.22  50.99  33.68  0.58    After 1 y  114.45  100.61  126.41  98.86  120.43  98.90  0.67  Dynamic β-cell glucose sensitivity, 109  Baseline  323.93  480.87  645.69  733.21  484.81  634.81  0.07    After 1 y  933.32  1063.86  414.90  469.87  674.11  855.01  0.03  Global β-cell glucose sensitivity, ×109/min−1  Baseline  57.32  35.51  54.71  34.44  56.01  34.65  0.79    After 1 y  130.37  104.36  135.74  99.97  133.06  101.17  0.85  Whole body insulin sensitivity, 105 dL/kg per min per pmol/L  Baseline  3.28  2.80  4.24  3.34  3.76  3.09  0.28    After 1 y  36.87  59.75  35.66  55.54  36.26  57.10  0.94  Insulin secretion rate AUC, nmol  Baseline  136.11  52.26  145.20  58.02  140.65  54.84  0.56    After 1 y  66.5  22.21  62.51  26.35  64.63  24.21  0.54  Disposition index, 1014 dL/kg/min−2 per pmol/L  Baseline  202.00  259.66  217.25  200.46  209.62  229.71  0.82    After 1 y  4532.68  6329.19  3728.83  5042.44  4130.75  5677.94  0.62  C-peptide AUC over glucose AUC above basal levels, nmol/pmol  Baseline  135.39  36.87  119.66  41.00  127.52  39.39  0.16    After 1 y  179.70  118.98  228.63  175.45  204.17  150.40  0.25  C-peptide AUC over glucose AUC, nmol/pmol  Baseline  118.79  33.29  112.84  36.42  115.81  34.66  0.55    After 1 y  100.32  28.17  114.40  49.09  107.36  40.25  0.22  Waist circumference, cm  Baseline  125.76  8.50  127.64  11.53  126.70  10.07  0.515    After 1 y  101.52  9.97  97.84  10.75  99.68  10.43  0.215    % Δ at 1 y  −19.29  5.40  −23.32  5.29  −21.31  5.67  0.010  Hip circumference, cm  Baseline  131.68  9.76  133.52  10.37  132.60  10.01  0.521    After 1 y  106.12  9.80  102.72  12.23  104.42  11.10  0.283    % Δ at 1 y  −19.40  4.57  −23.16  6.08  −21.28  5.65  0.017  Lean mass, kg  % Δ at 1 y  −17.03  9.95  −15.87  10.35  −16.45  10.06  0.688  Fat mass, kg  % Δ at 1 y  −36.11  14.27  −47.81  14.77  −41.96  15.54  0.006  View Large Table 2. OGTT Minimal Model β-cell Glucose Sensitivity, Whole Body Insulin Sensitivity Indices, and Body Composition Variables    SG (n = 25)  RYGB (n = 25)  Overall (n = 50)      Mean  SD  Mean  SD  Mean  SD  P Value  Static β-cell glucose sensitivity, ×109/min−1  Baseline  53.67  35.55  48.30  32.22  50.99  33.68  0.58    After 1 y  114.45  100.61  126.41  98.86  120.43  98.90  0.67  Dynamic β-cell glucose sensitivity, 109  Baseline  323.93  480.87  645.69  733.21  484.81  634.81  0.07    After 1 y  933.32  1063.86  414.90  469.87  674.11  855.01  0.03  Global β-cell glucose sensitivity, ×109/min−1  Baseline  57.32  35.51  54.71  34.44  56.01  34.65  0.79    After 1 y  130.37  104.36  135.74  99.97  133.06  101.17  0.85  Whole body insulin sensitivity, 105 dL/kg per min per pmol/L  Baseline  3.28  2.80  4.24  3.34  3.76  3.09  0.28    After 1 y  36.87  59.75  35.66  55.54  36.26  57.10  0.94  Insulin secretion rate AUC, nmol  Baseline  136.11  52.26  145.20  58.02  140.65  54.84  0.56    After 1 y  66.5  22.21  62.51  26.35  64.63  24.21  0.54  Disposition index, 1014 dL/kg/min−2 per pmol/L  Baseline  202.00  259.66  217.25  200.46  209.62  229.71  0.82    After 1 y  4532.68  6329.19  3728.83  5042.44  4130.75  5677.94  0.62  C-peptide AUC over glucose AUC above basal levels, nmol/pmol  Baseline  135.39  36.87  119.66  41.00  127.52  39.39  0.16    After 1 y  179.70  118.98  228.63  175.45  204.17  150.40  0.25  C-peptide AUC over glucose AUC, nmol/pmol  Baseline  118.79  33.29  112.84  36.42  115.81  34.66  0.55    After 1 y  100.32  28.17  114.40  49.09  107.36  40.25  0.22  Waist circumference, cm  Baseline  125.76  8.50  127.64  11.53  126.70  10.07  0.515    After 1 y  101.52  9.97  97.84  10.75  99.68  10.43  0.215    % Δ at 1 y  −19.29  5.40  −23.32  5.29  −21.31  5.67  0.010  Hip circumference, cm  Baseline  131.68  9.76  133.52  10.37  132.60  10.01  0.521    After 1 y  106.12  9.80  102.72  12.23  104.42  11.10  0.283    % Δ at 1 y  −19.40  4.57  −23.16  6.08  −21.28  5.65  0.017  Lean mass, kg  % Δ at 1 y  −17.03  9.95  −15.87  10.35  −16.45  10.06  0.688  Fat mass, kg  % Δ at 1 y  −36.11  14.27  −47.81  14.77  −41.96  15.54  0.006  Variables    SG (n = 25)  RYGB (n = 25)  Overall (n = 50)      Mean  SD  Mean  SD  Mean  SD  P Value  Static β-cell glucose sensitivity, ×109/min−1  Baseline  53.67  35.55  48.30  32.22  50.99  33.68  0.58    After 1 y  114.45  100.61  126.41  98.86  120.43  98.90  0.67  Dynamic β-cell glucose sensitivity, 109  Baseline  323.93  480.87  645.69  733.21  484.81  634.81  0.07    After 1 y  933.32  1063.86  414.90  469.87  674.11  855.01  0.03  Global β-cell glucose sensitivity, ×109/min−1  Baseline  57.32  35.51  54.71  34.44  56.01  34.65  0.79    After 1 y  130.37  104.36  135.74  99.97  133.06  101.17  0.85  Whole body insulin sensitivity, 105 dL/kg per min per pmol/L  Baseline  3.28  2.80  4.24  3.34  3.76  3.09  0.28    After 1 y  36.87  59.75  35.66  55.54  36.26  57.10  0.94  Insulin secretion rate AUC, nmol  Baseline  136.11  52.26  145.20  58.02  140.65  54.84  0.56    After 1 y  66.5  22.21  62.51  26.35  64.63  24.21  0.54  Disposition index, 1014 dL/kg/min−2 per pmol/L  Baseline  202.00  259.66  217.25  200.46  209.62  229.71  0.82    After 1 y  4532.68  6329.19  3728.83  5042.44  4130.75  5677.94  0.62  C-peptide AUC over glucose AUC above basal levels, nmol/pmol  Baseline  135.39  36.87  119.66  41.00  127.52  39.39  0.16    After 1 y  179.70  118.98  228.63  175.45  204.17  150.40  0.25  C-peptide AUC over glucose AUC, nmol/pmol  Baseline  118.79  33.29  112.84  36.42  115.81  34.66  0.55    After 1 y  100.32  28.17  114.40  49.09  107.36  40.25  0.22  Waist circumference, cm  Baseline  125.76  8.50  127.64  11.53  126.70  10.07  0.515    After 1 y  101.52  9.97  97.84  10.75  99.68  10.43  0.215    % Δ at 1 y  −19.29  5.40  −23.32  5.29  −21.31  5.67  0.010  Hip circumference, cm  Baseline  131.68  9.76  133.52  10.37  132.60  10.01  0.521    After 1 y  106.12  9.80  102.72  12.23  104.42  11.10  0.283    % Δ at 1 y  −19.40  4.57  −23.16  6.08  −21.28  5.65  0.017  Lean mass, kg  % Δ at 1 y  −17.03  9.95  −15.87  10.35  −16.45  10.06  0.688  Fat mass, kg  % Δ at 1 y  −36.11  14.27  −47.81  14.77  −41.96  15.54  0.006  View Large The ratio between the incremental AUCs of OGTT C-peptide and glycemia increased with time (from 127.52 ± 39.39 to 204.17 ± 150.40 nmol/mmol, P < 0.001). Figure 3 reports the mean concentrations of plasma glucose, C-peptide, and insulin after OGTT in the SG and RYGB groups at baseline and 1 year. The disposition index increased significantly with both treatments (from 209.62 ± 229.71 to 4130.75 ± 5677.94 ×1014 dL/kg/min−1 per pmol/L−1, P < 0.001). Figure 3. View largeDownload slide Plasma glucose, C-peptide, and insulin concentrations before (solid line) and after (dashed line) sleeve gastrectomy (solid circle) or gastric bypass (solid square). Disposition index (i.e., the product between β-cell glucose insulin sensitivity and whole body insulin sensitivity) and ΦD and ΦS (i.e., the dynamic and static components of the insulin secretion) before and after SG (open bar) or gastric bypass (filled bar). Figure 3. View largeDownload slide Plasma glucose, C-peptide, and insulin concentrations before (solid line) and after (dashed line) sleeve gastrectomy (solid circle) or gastric bypass (solid square). Disposition index (i.e., the product between β-cell glucose insulin sensitivity and whole body insulin sensitivity) and ΦD and ΦS (i.e., the dynamic and static components of the insulin secretion) before and after SG (open bar) or gastric bypass (filled bar). Weight loss and BMI The highest weight loss change was observed at 1 year after RYGB: −34.47 ± 6.81% after RYGB and −30.93 ± 7.38% after SG (P = 0.010). Changes in weight reflected changes in BMI (Table 1) with significant between-groups difference (P = 0.008). Lipid profile Total and LDL cholesterol decreased significantly in both groups (P = 0.001 and P < 0.001 in RYGB patients; P < 0.001 and P = 0.003 in SG patients, respectively). Although total cholesterol Δ did not reach statistical significance between groups, changes in LDL cholesterol differed between the two surgical procedures at 1 year: −29.25 ± 15.84 in RYGB and −14.99 ± 28.21 in SG (P = 0.008). HDL cholesterol increased in both groups; although changes in RYGB were highly significant (20.92 ± 43.23, P = 0.004), they were borderline in the SG group (8.42 ± 25.15, P = 0.046). Δ differences between groups were not significant. Triglycerides decreased significantly only in the SG group (−14.97 ± 40.08, P = 0.019). Arterial blood pressure Systolic and diastolic blood pressure slightly decreased in both groups, with no substantial difference between groups. Only the change in Δ systolic blood pressure was significantly different from zero for SG patients (−3.81 ± 8.91, P = 0.005). At baseline, 9 RYGB and 12 SG patients had hypertension; 44% RYGB and 50% SG hypertensive patients normalized blood pressure at 1 year without changes in type and dose of medications. Stepwise regressions No variable significantly predicted the binomial outcome presence/absence of hypoglycemia of any severity, whereas Δ BMI was borderline (β = −1.70e-01, SE = 9.25e-02, P = 0.07). Instead, when the primary end point (i.e. glycemia <3.1 mmol/L) was considered, both treatments (β = 2.33, SE = 1.07, P = 0.03) and Δ-BMI (β = −0.21, SE = 0.10, P = 0.03; with the probability of hypoglycemic events increasing in the RYGB group) were significant predictors of hypoglycemia. In other words, the larger the BMI reduction, the higher the number of hypoglycemic events in response to the OGTT. For each minimal model index, ϕs and dynamic control of insulin secretion in the fed state (ϕd), a regression model was built with a stepwise procedure to take into account multicollinearity to test if these parameters could be related to some baseline variables or to variable values recorded at 1 year. Each model parameter was related to age and sex and to their baseline values, BMI, glycemia, insulinemia, and to BMI, glycemia, insulinemia, and insulin sensitivity, at 1 year. In the model for the variable ϕs, only Δ BMI was significant (β = 4.33, SE = 1.79, P = 0.020). The ϕd was related significantly to the type of surgery (β = −601.35, SE = 226.27, P = 0.011), decreasing if the patients underwent RYGB, to baseline fasting insulinemia (β = −55.98, SE = 17.42, P = 0.0025) and to the BMI at 1 year after surgery (β = 56.72, SE = 23.94, P = 0.022). Total insulin secretion was predicted directly by baseline BMI (β = 1.87e+03, SE = 8.431e+02, P = 0.032) and inversely by insulin sensitivity at 1 year (β = −1.78e+02, SE=5.35e+01, P = 0.002). Insulin sensitivity at one year was predicted by sex (β = 48.94, SE = 15.84, P = 0035), being higher in women, baseline BMI (β = 6.78, SE = 2.58, P = 0.012) and BMI at 1 year after surgery (β = −4.42, SE = 2.21, P = 0.052). Finally, we checked for baseline characteristics that could predict hypoglycemia after surgery. When patients were considered altogether, the only significant predictor was age (β = −0.061, SE = 0.025 and P = 0.015). Age was also a significant predictor in the RYGB group (β = −0.084, SE = 0.036, P = 0.021), whereas in the SG group the baseline BMI was the only significant predictor (β = 0.19, SE = 0.097, P = 0.044). Adverse events There were no deaths or excessive weight loss in any group. Reintervention was not needed. Intravenous treatment of dehydration was necessary in 4 patients after RYGB and in 3 patients after SG. Cholelithiasis occurred in 2 RYGB patients. Nutritional deficiencies (i.e., anemia, low serum ferritin, hypoalbuminemia, hypocalcaemia, raised parathyroid hormone, and low vitamin D) were noted in 18 patients of the RYGB group and in 11 of the SG group. Discussion Our study demonstrates that the 1-year incidence of hypoglycemia (defined as <3.1 mmol/L or <55 mg/dL) after an OGTT was not significantly different between RYGB and SG (29% vs 14%, P = 0.079). Everyday hypoglycemic events, as measured with CGM, were also not significantly different. Yet, our multivariate analysis showed more hypoglycemia in the RYGB group as well as differences in hypoglycemia after OGTT using higher cutoffs (≤3.9 or ≤3.3 mmol/L). A more rapid recovery in the glycemic trough was also observed in SG as compared with RYGB subjects. Severe hypoglycemia that required hospitalization occurred in 6.8% of RYGB treatments, but never in SG-treated patients. We found substantial differences on the impact of both surgeries on parameters of insulin secretion that may contribute to hypoglycemia. We used <3.1 mmol/L to define hypoglycemia as our primary end point based on evidence that glucose concentrations under 3 mmol/L are associated with clinical harm. Cognitive impairment during both experimental and spontaneous hypoglycemia in people with T1D is detectable at 3 mmol/L or less (20), and exposure to 3 mmol/L can induce symptoms as well as counterregulatory hormonal responses to hypoglycemia in clamp studies (21, 22). T2D subjects who fail to recognize hypoglycemic symptoms until their glucose falls below 3 mmol/L (54 mg/dL) are nearly 8 times more likely to experience severe episodes of hypoglycemia (23) (blood glucose below 2.8 to 3 mmol/L). A recent conjunct position statement of the American Diabetes Association and of the European Association for the Study of Diabetes supports this definition, opting for the slightly stricter criterion of <3 mmol/L or 54 mg/dL (24). We also looked at <3.9 mmol/L, or 70 mg/dL, as this value is associated with the start of counter-regulatory response in health and has been defined as hypoglycemia by many authorities in the past. In the literature, RYGB reversal was associated with resolution of hypoglycemia in 13/17 (76%) (25) and 3/4 (75%) (26) of the patients with severe hypoglycemia. Conversion from RYGB to SG is currently used as an alternative treatment strategy to pancreatectomy for severe hypoglycemia following RYGB (27, 28). However, this conversion not only often fails to resolve the hypoglycemic symptoms but is also associated with high rates of major complications and hospital readmissions, and with needs for supplemental nutrition (29). Our study shows that SG is not a safer option than RYGB when assessed in terms of incidence rate of hypoglycemia after OGTT 1 year after surgery. In a recent study Salehi et al. (30) administered a liquid mixed meal during a hyperinsulinemic-hypoglycemic clamp with a target of 3 to 3.5 mmol/L in patients who underwent RYGB, in obese patients and in healthy controls. In spite of similar glycemic levels, insulin secretion was suppressed in controls but unsuppressed in RYGB subjects (30). The minimal model estimates two indexes, the static and the dynamic sensitivity to glucose. The static sensitivity ϕs measures the effect of glucose on β-cell secretion and is the ratio between insulin secretion rate and glucose concentrations, above a threshold level, at steady state. The ϕd is a measure of the stimulatory effect on insulin secretion of the rate at which glucose increases or decreases. We found that ϕd was reduced after RYGB, whereas the static control, ϕs, was improved. Therefore, after RYGB, the patients had an increased insulin secretion in response to increased plasma glucose concentrations, but were unable to reduce the secretion of insulin when glucose concentrations were falling. The hyperglycemic peak that follows an OGTT or a meal after RYGB or SG has been attributed to the accelerated gastric emptying associated with both types of operations (31). About one-third of the variability in the glycemic response to an OGTT, either in subjects with normal glucose tolerance or T2D, is due to the variability of the gastric emptying rate (32). Gastric emptying has been described to be equally accelerated after both RYGB and SG. In fact, in rats in which gastric emptying was evaluated by 99mTc-scintigraphy, the liquid meal mixture was fully (100%) emptied in only 5 minutes after both RYGB and SG, whereas only about 6% was emptied in sham-operated animals in the same period (33). In humans (34), the half-emptying time evaluated by scintigraphy was significantly reduced after SG compared with baseline, down to <30 minutes. After RYGB, the scintigraphic gastric emptying time for both liquids and solids was also found to be <30 minutes (35). However, our data suggest that the postprandial hyperinsulinemia observed after RYGB is not the consequence of the rapid glycemic rise but rather the result of an inappropriately high insulin secretion. In fact, the hypoglycemic events were uncorrelated with the hyperglycemic peaks, which were observed after both RYGB and SG. The dysregulation in insulin secretion in conjunction with an impressive improvement of insulin sensitivity can help to explain why RYGB patients had an overall larger number of reactive hypoglycemic episodes than SG. Recently, Abrahamsson et al. (15) found significantly reduced symptoms and counterregulatory hormonal response to hypoglycemia 23 weeks after RYGB. The repeated episodes of hypoglycemia in our series can help explain why these patients are generally unaware of low glycemic levels. In fact, repeated hypoglycemia or a state of chronic hypoglycemia, such that occurring during a continuous hyperinsulinemic-hypoglycemic clamp for 4 consecutive days, are associated with a glycemic threshold for cognitive impairment that is lowered from 3.0 to 2.5 mmol/L (36). This suggests a progressive adaptation to hypoglycemia. Also, in patients with T2D, recurrent hypoglycemia reduces the glucose concentration needed for the counterregulatory response (37,). Although hypoglycemia is regarded as a life-threatening complication in diabetic individuals, very few data for nondiabetic subjects are reported in the literature. The increasing use of bariatric surgery, leading to frequent hypoglycemic episodes, makes it necessary to rethink the need for lifelong nutritional support for these patients, possibly together with a modification of the surgical technique such as, for example, a longer alimentary limb in RYGB. In fact, biliopancreatic diversion is rarely associated with hyper- and hypoglycemia (38). Looking for a predictor of hypoglycemia after gastrointestinal surgery, we found that the higher the reduction of BMI at 1 year, the higher the number of hypoglycemic events triggered by the oral glucose load. The minimal model parameter ϕd was more reduced in those RYGB patients who had the largest reduction in BMI at 1 year after surgery, suggesting that insulin secretion dysregulation is more likely in those subjects who experienced the largest weight loss. Women had a better improvement of insulin sensitivity than men; this might be a reason they were subjected to more hypoglycemic episodes. However, new, larger trials are necessary to identify those factors that can predict reactive hypoglycemia before surgery. Although the power of our study is high, an important limitation consists of the relatively short duration of follow-up (12 months), with evidence that a difference in the rate of OGTT hypoglycemia starts to appear at 6 months. In addition, we note the single-center, open-label nature of our study. However, we are continuing patient follow-up and we plan to publish the results at 3 and 5 years. Finally, lack of glucagon-like peptide 1 levels data during the OGTT is a limitation of our study because it is a major glucoregulatory player that improves β-cell glucose sensitivity. In conclusion, RYGB induced a rate of reactive hypoglycemia (<3.1 mmol/L) similar to that of SG, but more hypoglycemic events using glycemic higher cutoffs and 6.8% severe hypoglycemia with hospitalization. Hypoglycemic episodes increased in frequency over time after RYGB and appeared to be associated with an imbalance between impaired β-cell dynamic glucose sensitivity and dramatically enhanced total body insulin sensitivity. Abbreviations: Abbreviations: AUC area under the curve BMI body mass index CGM continuous glucose monitoring HDL high-density lipoprotein LDL low-density protein OGTT oral glucose tolerance test RYGB Roux-en-Y gastric bypass SE standard error SG sleeve gastrectomy T2D type 2 diabetes ϕ1 decrement occurring during follow-up ϕd dynamic control of insulin secretion in the fed state ϕs glucose sensitivity index Acknowledgments Financial Support: The study was conducted with internal funds of the Catholic University. Clinical Trial Information: ClinicalTrials.gov no. NCT01581801 (registered 20 April 2012). Author Contributions: E.C., G.M., M.R., S.P., and A.D.G. designed the study. E.C., V.S., and P.G. acquired the data. The data analysis was done by S.P. and A.D.G. Data interpretation was done by E.C., G.M., M.R., S.P., R.B., and A.D.G. E.C., G.M., S.P., M.R., and A.D.G. contributed to drafting the article. S.R.B., S.A., and A.L.B. made critical revisions. All authors approved the final version of the manuscript. E.C., G.M., V.S., and S.P. ensure that questions related to the accuracy or integrity of any part of the article were appropriately investigated and resolved. E.C., S.P., and G.M. are guarantors of the study Disclosure Summary: The authors have nothing to declare. References 1. Nguyen DM, El-Serag HB. The epidemiology of obesity. Gastroenterol Clin North Am . 2010; 39( 1): 1– 7. Google Scholar CrossRef Search ADS PubMed  2. Dietz WH Jr, Gortmaker SL. Do we fatten our children at the television set? Obesity and television viewing in children and adolescents. Pediatrics . 1985; 75( 5): 807– 812. Google Scholar PubMed  3. Puzziferri N, Roshek TB III, Mayo HG, Gallagher R, Belle SH, Livingston EH. Long-term follow-up after bariatric surgery: a systematic review. JAMA . 2014; 312( 9): 934– 942. Google Scholar CrossRef Search ADS PubMed  4. Khorgami Z, Andalib A, Corcelles R, Aminian A, Brethauer S, Schauer P. Recent national trends in the surgical treatment of obesity: sleeve gastrectomy dominates. Surg Obes Relat Dis . 2015; 11( 6): S6– S8. Google Scholar CrossRef Search ADS   5. Mingrone G, Panunzi S, De Gaetano A, Guidone C, Iaconelli A, Leccesi L, Nanni G, Pomp A, Castagneto M, Ghirlanda G, Rubino F. Bariatric surgery versus conventional medical therapy for type 2 diabetes. N Engl J Med . 2012; 366( 17): 1577– 1585. Google Scholar CrossRef Search ADS PubMed  6. Mingrone G, Panunzi S, De Gaetano A, Guidone C, Iaconelli A, Nanni G, Castagneto M, Bornstein S, Rubino F. Bariatric-metabolic surgery versus conventional medical treatment in obese patients with type 2 diabetes: 5 year follow-up of an open-label, single-centre, randomised controlled trial. Lancet . 2015; 386( 9997): 964– 973. Google Scholar CrossRef Search ADS PubMed  7. Schauer PR, Bhatt DL, Kirwan JP, Wolski K, Brethauer SA, Navaneethan SD, Aminian A, Pothier CE, Kim ES, Nissen SE, Kashyap SR; STAMPEDE Investigators. Bariatric surgery versus intensive medical therapy for diabetes--3-year outcomes. N Engl J Med . 2014; 370( 21): 2002– 2013. Google Scholar CrossRef Search ADS PubMed  8. Seaquist ER, Anderson J, Childs B, Cryer P, Dagogo-Jack S, Fish L, Heller SR, Rodriguez H, Rosenzweig J, Vigersky R. Hypoglycemia and diabetes: a report of a workgroup of the American Diabetes Association and the Endocrine Society. Diabetes Care . 2013; 36( 5): 1384– 1395. Google Scholar CrossRef Search ADS PubMed  9. Service GJ, Thompson GB, Service FJ, Andrews JC, Collazo-Clavell ML, Lloyd RV. Hyperinsulinemic hypoglycemia with nesidioblastosis after gastric-bypass surgery. N Engl J Med . 2005; 353( 3): 249– 254. Google Scholar CrossRef Search ADS PubMed  10. Roslin MS, Oren JH, Polan BN, Damani T, Brauner R, Shah PC. Abnormal glucose tolerance testing after gastric bypass. Surg Obes Relat Dis . 2013; 9( 1): 26– 31. Google Scholar CrossRef Search ADS PubMed  11. Lee CJ, Wood GC, Lazo M, Brown TT, Clark JM, Still C, Benotti P. Risk of post-gastric bypass surgery hypoglycemia in nondiabetic individuals: A single center experience. Obesity (Silver Spring) . 2016; 24( 6): 1342– 1348. Google Scholar CrossRef Search ADS PubMed  12. Marsk R, Jonas E, Rasmussen F, Näslund E. Nationwide cohort study of post-gastric bypass hypoglycaemia including 5,040 patients undergoing surgery for obesity in 1986-2006 in Sweden. Diabetologia . 2010; 53( 11): 2307– 2311. Google Scholar CrossRef Search ADS PubMed  13. Sarwar H, Chapman WH III, Pender JR, Ivanescu A, Drake AJ III, Pories WJ, Dar MS. Hypoglycemia after Roux-en-Y gastric bypass: the BOLD experience. Obes Surg . 2014; 24( 7): 1120– 1124. Google Scholar CrossRef Search ADS PubMed  14. Amiel SA, Tamborlane WV, Simonson DC, Sherwin RS. Defective glucose counterregulation after strict glycemic control of insulin-dependent diabetes mellitus. N Engl J Med . 1987; 316( 22): 1376– 1383. Google Scholar CrossRef Search ADS PubMed  15. Abrahamsson N, Börjesson JL, Sundbom M, Wiklund U, Karlsson FA, Eriksson JW. Gastric bypass reduces symptoms and hormonal responses in hypoglycaemia. Diabetes . 2016; 65( 9): 2667– 2675. Google Scholar CrossRef Search ADS PubMed  16. Tzovaras G, Papamargaritis D, Sioka E, Zachari E, Baloyiannis I, Zacharoulis D, Koukoulis G. Symptoms suggestive of dumping syndrome after provocation in patients after laparoscopic sleeve gastrectomy. Obes Surg . 2012; 22( 1): 23– 28. Google Scholar CrossRef Search ADS PubMed  17. Chen CY, Lee WJ, Lee HM, Chen JC, Ser KH, Lee YC, Chen SC. Laparoscopic conversion of gastric bypass complication to sleeve gastrectomy: technique and early results. Obes Surg . 2016; 26( 9): 2014– 2021. Google Scholar CrossRef Search ADS PubMed  18. Cryer PE, Axelrod L, Grossman AB, Heller SR, Montori VM, Seaquist ER, Service FJ; Endocrine Society. Evaluation and management of adult hypoglycemic disorders: an Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab . 2009; 94( 3): 709– 728. Google Scholar CrossRef Search ADS PubMed  19. Breda E, Cavaghan MK, Toffolo G, Polonsky KS, Cobelli C. Oral glucose tolerance test minimal model indexes of beta-cell function and insulin sensitivity. Diabetes . 2001; 50( 1): 150– 158. Google Scholar CrossRef Search ADS PubMed  20. Gonder-Frederick LA, Zrebiec JF, Bauchowitz AU, Ritterband LM, Magee JC, Cox DJ, Clarke WL. Cognitive function is disrupted by both hypo- and hyperglycemia in school-aged children with type 1 diabetes: a field study. Diabetes Care . 2009; 32( 6): 1001– 1006. Google Scholar CrossRef Search ADS PubMed  21. Heller SR, Cryer PE. Reduced neuroendocrine and symptomatic responses to subsequent hypoglycemia after 1 episode of hypoglycemia in nondiabetic humans. Diabetes . 1991; 40( 2): 223– 226. Google Scholar CrossRef Search ADS PubMed  22. Davis SN, Shavers C, Mosqueda-Garcia R, Costa F. Effects of differing antecedent hypoglycemia on subsequent counterregulation in normal humans. Diabetes . 1997; 46( 8): 1328– 1335. Google Scholar CrossRef Search ADS PubMed  23. Pistrosch F, Ganz X, Bornstein SR, Birkenfeld AL, Henkel E, Hanefeld M. Risk of and risk factors for hypoglycemia and associated arrhythmias in patients with type 2 diabetes and cardiovascular disease: a cohort study under real-world conditions. Acta Diabetol . 2015; 52( 5): 889– 895. Google Scholar CrossRef Search ADS PubMed  24. International Hypoglycaemia Study Group. Glucose concentrations of less than 3.0 mmol/L (54 mg/dL) should be reported in clinical trials: a joint position statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care.  2017; 40: 155– 157. CrossRef Search ADS PubMed  25. Mala T. Postprandial hyperinsulinemic hypoglycemia after gastric bypass surgical treatment. Surg Obes Relat Dis . 2014; 10( 6): 1220– 1225. Google Scholar CrossRef Search ADS PubMed  26. Pernar LI, Kim JJ, Shikora SA. Gastric bypass reversal: a 7-year experience. Surg Obes Relat Dis . 2016; 12( 8): 1492– 1498. Google Scholar CrossRef Search ADS PubMed  27. Lakdawala M, Limas P, Dhar S, Remedios C, Dhulla N, Sood A, Bhasker AG. Laparoscopic revision of Roux-en-Y gastric bypass to sleeve gastrectomy: a ray of hope for failed Roux-en-Y gastric bypass. Asian J Endosc Surg . 2016; 9( 2): 122– 127. Google Scholar CrossRef Search ADS PubMed  28. Carter CO, Fernandez AZ, McNatt SS, Powell MS. Conversion from gastric bypass to sleeve gastrectomy for complications of gastric bypass. Surg Obes Relat Dis . 2016; 12( 3): 572– 576. Google Scholar CrossRef Search ADS PubMed  29. van Beek AP, Emous M, Laville M, Tack J. Dumping syndrome after esophageal, gastric or bariatric surgery: pathophysiology, diagnosis, and management. Obes Rev . 2017; 18( 1): 68– 85. Google Scholar CrossRef Search ADS PubMed  30. Salehi M, Woods SC, D’Alessio DA. Gastric bypass alters both glucose-dependent and glucose-independent regulation of islet hormone secretion. Obesity (Silver Spring) . 2015; 23( 10): 2046– 2052. Google Scholar CrossRef Search ADS PubMed  31. Holst JJ, Madsbad S. Mechanisms of surgical control of type 2 diabetes: GLP-1 is key factor. Surg Obes Relat Dis . 2016; 12( 6): 1236– 1242. Google Scholar CrossRef Search ADS PubMed  32. Horowitz M, Edelbroek MA, Wishart JM, Straathof JW. Relationship between oral glucose tolerance and gastric emptying in normal healthy subjects. Diabetologia . 1993; 36( 9): 857– 862. Google Scholar CrossRef Search ADS PubMed  33. Chambers AP, Smith EP, Begg DP, Grayson BE, Sisley S, Greer T, Sorrell J, Lemmen L, LaSance K, Woods SC, Seeley RJ, D’Alessio DA, Sandoval DA. Regulation of gastric emptying rate and its role in nutrient-induced GLP-1 secretion in rats after vertical sleeve gastrectomy. Am J Physiol Endocrinol Metab . 2014; 306( 4): E424– E432. Google Scholar CrossRef Search ADS PubMed  34. Kandeel AA, Sarhan MD, Hegazy T, Mahmoud MM, Ali MH. Comparative assessment of gastric emptying in obese patients before and after laparoscopic sleeve gastrectomy using radionuclide scintigraphy. Nucl Med Commun . 2015; 36( 8): 854– 862. Google Scholar CrossRef Search ADS PubMed  35. Hinder RA, Esser J, DeMeester TR. Management of gastric emptying disorders following the Roux-en-Y procedure. Surgery . 1988; 104( 4): 765– 772. Google Scholar PubMed  36. Boyle PJ, Nagy RJ, O’Connor AM, Kempers SF, Yeo RA, Qualls C. Adaptation in brain glucose uptake following recurrent hypoglycemia. Proc Natl Acad Sci USA . 1994; 91( 20): 9352– 9356. Google Scholar CrossRef Search ADS PubMed  37. Davis SN, Mann S, Briscoe VJ, Ertl AC, Tate DB. Effects of intensive therapy and antecedent hypoglycemia on counterregulatory responses to hypoglycemia in type 2 diabetes. Diabetes . 2009; 58( 3): 701– 709. Google Scholar CrossRef Search ADS PubMed  38. Mingrone G, Castagneto M. Bariatric surgery: unstressing or boosting the beta-cell? Diabetes Obes Metab . 2009; 11( Suppl 4): 130– 142. Google Scholar CrossRef Search ADS PubMed  Copyright © 2018 Endocrine Society

Journal

Journal of Clinical Endocrinology and MetabolismOxford University Press

Published: Mar 23, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off