Model-Based Assessment of C-Peptide Secretion and Kinetics in Post Gastric Bypass Individuals Experiencing Postprandial Hyperinsulinemic Hypoglycemia

Abstract

Assessment of insulin secretion is key to diagnose postprandial hyperinsulinemic hypoglycemia (PHH), an increasingly recognized complication following bariatric surgery. To this end, the Oral C-peptide Minimal Model (OCMM) can be used. This usually requires fixing C-peptide (CP) kinetics to the ones derived from the Van Cauter population model (VCPM), which has never been validated in PHH individuals. The objective of this work was to test the validity of the OCMM coupled with the VCPM in PHH subjects and propose a method to overcome the observed limitations. Two cohorts of adults with PHH after gastric bypass (GB) underwent either a 75 g oral glucose (9F/3M; age=42±9 y; BMI=28.3±6.9 kg/m2) or a 60 g mixed-meal (7F/3M; age = 43 ± 11 y; BMI=27.5±4.2 kg/m2) tolerance test. The OCMM was identified on CP concentration data with CP kinetics fixed to VCPM (VC approach). In both groups, the VC approach underestimated CP-peak and overestimated CP-tail suggesting CP kinetics predicted by VCPM to be inaccurate in this population. Thus, the OCMM was identified using CP kinetics estimated from the data (DB approach) using a Bayesian Maximum a Posteriori estimator. CP data were well predicted in all the subjects using the DB approach, highlighting a significantly faster CP kinetics in patients with PHH compared to the one predicted by VCPM. Finally, a simulation study was used to validate the proposed approach. The present findings question the applicability of the VCPM in patients with PHH after GB and call for CP bolus experiments to develop a reliable CP kinetic model in this population.

Effects of a single 10mg dose of empagliflozin on postprandial insulin kinetics in patients with postbariatric hypoglycaemia

“…Total first-pass hepatic insulin extraction increased after empagliflozin compared to placebo (49.6±14.2 vs. 39.7±12.1 %, p=0.006), while no significant effect of empaglizflozin on basal first-pass hepatic insulin extraction was observed (79.7±7.1 vs. 81.1±6.6 %, p=0.521). Total insulin clearance, as calculated from (7), resulted to be significantly lower after empagliflozin compared to placebo (3.91±1.58 vs. 3.00±1.27 l/min, p=0.002).…”

“…C-peptide and post-hepatic insulin kinetics (required for the OCMM and IMM identification) were estimated from the data using an extended version of a recently proposed Bayesian approach (6), but these were assumed to be identical on both visits helping to avoid possible compensations against beta-cell responsivity and hepatic insulin extraction. Additionally, we also estimated total insulin clearance (composite of hepatic and posthepatic elimination) using a formula reported in Gastaldelli et al (7).…”

Altered Insulin Clearance after Gastric Bypass and Sleeve Gastrectomy in the Fasting and Prandial Conditions

“…Thus, we estimated that the exchange rate from the fast to slow compartment (k21, Figure 4a) could be up to 60% less in GB subjects, with an additional reduction of 30% in the exchange from slow to fast compartments (k12, Figure 4a) compared to the population values calculated using the formula proposed by Van Cauter [47]. The new values were derived from previous data recently published [48,49]. [30,47] and derived equation for the estimation of pre-hepatic insulin secretion rates (ISR) from C-peptide concentrations measured over time, from Cp(t), and from k values (fractional clearance rates between compartments).…”

“…Thus, we estimated that the exchange rate from the fast to slow compartment (k21, Figure 4 a) could be up to 60% less in GB subjects, with an additional reduction of 30% in the exchange from slow to fast compartments (k12, Figure 4 a) compared to the population values calculated using the formula proposed by Van Cauter [ 47 ]. The new values were derived from previous data recently published [ 48 , 49 ].…”

Inferring Insulin Secretion Rate from Sparse Patient Glucose and Insulin Measures

“…This method describes the ISR function by two rates proportional linearly with the c-peptide and glucose concentrations. Another recent estimation approach based on OGTT measurements of insulin and c-peptide has been developed to estimate the ISR time traces using two different models, for insulin and c-peptide ( Schiavon et al, 2021 ).…”