Alejandro Pérez-Pitarch(1,2), Valerie Nock(1), Andreas Huennemeyer(1), Nima Soleymanlou(3), Stefan Kaspers(4), Jan Freijer(1).
(1) Translational Medicine & Clinical Pharmacology, Boehringer Ingelheim Pharma GmbH & Co. KG, Biberach an der Riss, Germany; (2) Departamento de Farmacia y Tecnolog.a Farmac.utica, Universidad de Valencia, Valencia, Spain; (3) Boehringer Ingelheim Canada Ltd./Lt.e, Burlington, Canada; (4) Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim, Germany
Objectives: Due to its insulin-independent mechanism of action, empagliflozin is under clinical investigation as adjunct to insulin to improve glycemic control in patients with type 1 diabetes mellitus (T1DM). A previous analysis of the EASE-1 trial (4-week, randomized, blinded, placebo-controlled trial of empagliflozin as adjunct to insulin in patients with T1DM) reported that rates of symptomatic hypoglycemia with plasma glucose <3.0 mmol/l were generally lower with empagliflozin than with placebo [1]. The objective of this exploratory analysis was to develop a repeated time to categorical event (RTTCE) model in order to describe factors influencing the rate and severity of hypoglycemia.
Methods: Hypoglycemic event (HYPO) data from the EASE-1 trial were analyzed. To describe the repeated time to events (RTTE), a parametric survival model was fitted to the data using NONMEM 7.3. To simultaneously capture the severity of events, the RTTE model was combined with an ordered categorical model to form the RTTCE model [2]. The influence of insulin titration, individual predicted empagliflozin AUCs at steady state (AUCss) and other patient-related covariates were explored. The performance of the model was evaluated with Visual Predictive Checks of the Kaplan Meier curves.
Results: Reporting-rates of asymptomatic and symptomatic HYPOs proved to have a different profile and thus, asymptomatic HYPOs were excluded from the analysis. The rate of symptomatic HYPOs was best described by a Gompertz model. The shape parameter proved to be influenced by insulin titration. When insulin was not titrated, instantaneous hazard increased over time, but once insulin was titrated and patients were stabilized, instantaneous hazard decreased over time. Higher empagliflozin AUCss were related to a lower HYPO risk. The hazard of having a symptomatic HYPO increased with lower mean glucose concentrations and female sex. The simultaneous description of rate and severity of events allowed identification of an inverse correlation between them.
Conclusions: The proposed model described the symptomatic HYPOs reported in the trial and provided information regarding patient- and study-specific factors influencing the risk of having a HYPO. This work represents a first exploratory step in the direction of a model-informed exposure-risk-benefit evaluation for empagliflozin as a treatment option in patients with T1DM, and will be further refined with availability of future data.
References:
[1] Pieber TR, et al. Empagliflozin as adjunct to insulin in patients with type 1 diabetes: a 4-week, randomized, placebo-controlled trial (EASE-1). Diabetes Obes Metab. 2015; 17(10):928-35.
[2] Plan EL, Karlsson KE, Karlsson MO. Approaches to simultaneous analysis of frequency and severity of symptoms. Clin Pharmacol Ther. 2010; 88(2):255-9.
Reference: PAGE 25 (2016) Abstr 5743 [www.page-meeting.org/?abstract=5743]
Poster: Drug/Disease modeling - Safety