Meta-analysis on the relationship between everolimus exposure and safety and efficacy.

Abstract

3099 Background: Data on everolimus exposure–safety and response relationships are available from individual studies. We performed a meta-analysis to assess these relationships using everolimus predose minimum blood concentration (Cmin). Methods: Individual patient data from 5 phase 2 or 3 oncology studies in which steady-state predose pharmacokinetic samples were taken from patients administered everolimus monotherapy 10 mg/day were pooled. In a Cox regression, the times to first grade ≥3 select adverse events (AEs) associated with everolimus and first progression-free survival (PFS) event were related to log-transformed instant or time-averaged Cmin as time-varying covariates stratified by study. Probability of tumor size reduction according to RECIST (yes vs no) was assessed using a generalized linear mixed model fitted with GEE method with log-transformed instant or time-averaged Cmin as covariates and treatment arm as fixed effect. Results: 945 patients were evaluable for efficacy, 938 for safety. Patient characteristics were relatively well balanced across studies. Regardless of whether instant or time-averaged Cmin was used, a 2-fold increase in everolimus exposure increased the risk of grade ≥3 pulmonary, metabolic, and stomatitis events (Table); no increased risk of other grade ≥3 AEs was observed. A 2-fold increase in everolimus exposure increased the likelihood of tumor size reduction; there was a trend of reduced risk of a PFS event (Table). Conclusions: A 2-fold increase in everolimus exposure is associated with a higher probability of tumor size reduction and increased risk of high-grade pulmonary, metabolic, and stomatitis AEs associated with everolimus. [Table: see text]