A-029 Modeling the Effect of Step Changes in Glucose on %Glycated Albumin vs Time

Abstract

Abstract Background Glycated albumin (GA) is often referred to as being a reflection of glucose exposure over the past 20–30 days. We examined this premise from a theoretical standpoint using a model for formation of GA and simulation of the effects of step changes in glucose concentration ([G]). Methods A model for glycation of albumin parallels established models for formation of hemoglobin A1c, differing only in characterization of the age distribution for albumin. Probability of survival (P(s)) of albumin at a given age (t) is based on a first-order rate constant (ke) for elimination: P(s) = exp(-ke t) [Eqn.1], where ke = 4.08e-2/d (t(1/2) = 17 d). Probability of glycation (P(g)) at a given age is given by: P(g) = (1-exp(-k[G] t)) [Eqn.2]. There is a simple analytical solution for %GA for constant [G]: %GA = 100(1-w/z), where w = ke and z = (ke+k[G]). k = 9.77e-4/d/(mmol/L) reproduces reference values for %GA [PMID: 29436378], and for which %GA = 2.41(%A1c)-0.004 (r = 0.999, for A1c range = 5%–12%). For non-constant [G], %GA must be determined by simulation of changes in P(g) using Eqn.2, calculated in small time increments, dt. Using dt = 0.01 d, we calculated changes in %GA by simulation as a function of time for two scenarios involving step changes in [G] at time = 0: A. from 10 mmol/L to 15 mmol/L (%GA ultimately moves from 19.3% to 26.4%); B. from 15 mmol/L to 10 mmol/L (%GA ultimately moves from 26.4% to 19.3%). Results Figure shows the fractional transition of %GA between respective starting points and ultimate endpoints vs time for scenarios A and B. In both cases, transition at 30 days was approximately 80% of the ultimate full transition. Conclusions Physiological model-based calculations support that description of %GA as a reflection of [G] over the past 30 days is a very reasonable one under conditions in which [G] has changed.