Evaluation of pharmacokinetic model designs for subcutaneous infusion of insulin aspart

PKPD Modeling and Simulations to Support Biosimilar Development of Biphasic Insulin Aspart 30

This paper presents an analysis of data from a comparative study of biosimilarity in terms of pharmacokinetics and pharmacodynamics in healthy volunteers using a hyperinsulinemic euglycemic clamp for reference and test biphasic insulin aspart 30 (BIAsp 30). As a result of the study, one of the secondary pharmacodynamic (PD) endpoints did not satisfy the classical criterion of 80%-125% (the lower limit for PD parameter area under the glucose infusion rate-time curve [ AUC GIR 0 - t ${\rm{AUC}}_{{\rm{GIR}}_{0 - {\rm{t}}}}$ ] turned out to be 79.5%). The main hypothesis explaining this result is that the sample size is insufficient to conduct a PD test with 90% statistical power, since the sample size has been calculated based on the coefficient of variation (CV) of pharmacokinetic (PK) parameters. To test this hypothesis, population PKPD (popPKPD) modeling and subsequent simulations of the required number of PD profiles were used. Two popPKPD models were constructed (a one-compartment double simultaneous absorption model for PK and an effect compartment E_max model for PD) to describe the PKPD data of reference and test insulins. As a result, using real data along with model-based simulation data, a biosimilarity test for PD was performed, and the lower limit for AUC GIR 0 - t ${\rm{AUC}}_{{\rm{GIR}}_{0 - {\rm{t}}}}$ became 82.6%, while the CV decreased from 31.7% to 24.1%. Thus, popPKPD modeling and simulations have been shown to be effective in interpreting and supporting the results of clinical biosimilarity trials.

The Effects of Additional Local-Mixing Compartments in the DISST Model-Based Assessment of Insulin Sensitivity

BACKGROUND

The identification of insulin sensitivity in glycemic modelling can be heavily obstructed by the presence of outlying data or unmodelled effects. The effect of data indicative of local mixing is especially problematic with models assuming rapid mixing of compartments. Methods such as manual removal of data and outlier detection methods have been used to improve parameter ID in these cases, but modelling data with more compartments is another potential approach.

METHODS

This research compares a mixing model with local depot site compartments with an existing, clinically validated insulin sensitivity test model. The Levenberg-Marquardt (LM) parameter identification method was implemented alongside a modified version (aLM) capable of operator-independent omission of outlier data in accordance with the 3 standard deviation rule. Three cases were tested: LM where data points suspected to be affected by incomplete mixing at the depot site were removed, aLM, and LM with the more complex mixing model.

RESULTS

While insulin parameters identified in the mixing model differed greatly from those in the DISST model, there were strong Spearman correlations of approximately 0.93 for the insulin sensitivity values identified across all 3 methods. The 2 models also showed comparable identification stability in insulin sensitivity estimation through a Monte Carlo analysis. However, the mixing model required modifications to the identification process to improve convergence, and still failed to converge to feasible parameters on 5 of the 212 trials.

CONCLUSIONS

The mixing compartment model effectively captured the dynamics of mixing behavior, but with no significant improvement in insulin sensitivity identification.

Modeling of Pharmacokinetic Profiles of Insulin Aspart and Biphasic Insulin Aspart 30 / 70

The study includes modeling and simulation of insulin aspart pharmacokinetics (PK). The authors used PK data of biosimilar insulins - insulin aspart and biphasic insulin aspart 30/70 - to develop a predictive population PK model for the insulins. The model was built via Monolix software taking into account the weight-based dosing and the dose and body weight effects on the parameters. The model-based simulations were performed using the R package mlxR for various administered doses and various ratios of insulin aspart forms for a better understanding of the insulin behavior. The optimal model was a one-compartment model with a combination of zero- and first-order absorptions with absorption lag for the soluble form of insulin aspart and first-order absorption for the insulin aspart protamine suspension. The assumption of identical behavior of two insulins at the distribution and elimination phases was made. The developed PK model was fitted successfully to the experimental dataand all fitted parameters displayed a moderate coefficient of variation. The PK model allows us to predict PK profiles for various doses and formulations of insulin aspart and can be used to improve the accuracy, safety and ethics of novel clinical trials of insulin. This article is protected by copyright. All rights reserved.