Bill Poland: No financial relationships to disclose
Objectives: ALXN2060 (acoramidis, AG10), an investigational, potent, highly selective, small molecule transthyretin (TTR) stabilizer that can achieve ≥ 90% TTR stabilization, is in development for the treatment of TTR amyloid cardiomyopathy (ATTR-CM), a progressive, fatal disease in which the deposition of amyloid derived from TTR causes progressive heart failure. The objective was to describe the population pharmacokinetics (PK) of ALXN2060 in healthy subjects and subjects with symptomatic ATTR-CM, including Japanese subjects, for a New Drug Application in Japan.
Methods: A population PK model for ALXN2060 was developed using the NONMEM software, based on data from 121 healthy subjects in 5 Phase 1 studies and 209 subjects with symptomatic ATTR-CM in 4 Phase 2-3 studies, of whom 68.4% were White and 11.5% were Japanese. Covariates of potential clinical interest were selected using stepwise covariate modeling followed by model evaluation with diagnostic plots. The analysis plan specified that race effects on apparent clearance (CL/F) and apparent central volume (Vc/F) were to be evaluated as a single, multi-valued covariate or as binary covariates, one for each race category. The best fit retained only the White/non-White binary covariate, but the analysis plan was overridden with a three-category covariate—White, Japanese, and Other—to ensure capture of any distinct Japanese race effect, which was of particular relevance in this analysis.
Results: ALXN2060 PK was well described by a 2-compartment disposition model with sequential zero-order and first-order absorption. The final model included the effects of health status (ATTR-CM patient or healthy) on CL/F and Vc/F and of baseline creatinine clearance and race (White, Japanese, and Other) on CL/F. Bioavailability was found to decrease with dose, and food with dose administration increased the initial absorption duration (with high uncertainty attributed to limited relevant data) and reduced the absorption rate without significant effect on exposure area-under-the-curve (AUC). For a typical White patient at steady state, CL/F was 2.60 L/h and Vc/F was 5.52 L. Compared to White patients, the CL/F estimate was 41.4% lower in healthy subjects, 11.0% lower in Japanese patients, and 18.4% lower in non-White, non-Japanese patients. These effects correspond to 70.6%, 12.4%, and 22.5% higher steady-state AUC, respectively, which can be compared to a 35.4% interindividual variability of apparent clearance. Vc/F was 145% (2.45-fold) higher in healthy subjects, reducing their Cmax.
Conclusions: Population PK modeling was able to predict observed ALXN2060 concentrations with reasonable accuracy. Had the planned race categorization for covariate modeling not been modified to distinguish Japanese subjects, they would have been combined with Other race and assigned a much lower typical clearance; overriding the planned analysis was therefore essential to characterize Japanese subjects accurately.
