Population pharmacokinetics of bevacizumab in cancer patients with external validation

Background Bevacizumab is approved for various cancers. This analysis aimed to comprehensively evaluate bevacizumab pharmacokinetics and the influence of patient variables on bevacizumab pharmacokinetics. Methods Rich and sparse bevacizumab serum concentrations were collected from Phase I through IV studies in early and metastatic cancers. Bevacizumab was given intravenously as single agent or in combination with chemotherapy for single- and multiple-dose schedules. Results Model-building used 8943 bevacizumab concentrations from 1792 patients with colon/colorectal, non-small cell lung, kidney, pancreatic, breast, prostate and brain cancer. Bevacizumab doses ranged from 1 to 20 mg/kg given once every 1, 2 or 3 weeks. A two-compartment model best described the data. The population estimates of clearance (CL), central volume of distribution (V1) and half-life for a typical 70-kg patient were 9.01 mL/h, 2.88 L and 19.6 days. CL and V1 increased with body weight and were higher in males than females by 14 and 18 %, respectively. CL decreased with increasing albumin and decreasing alkaline phosphatase. The final model was externally validated using 1670 concentrations from 146 Japanese patients that were not used for model-building. Mean prediction errors were −2.1, 3.1 and 1.0 % for concentrations, CL and V1, respectively, confirming adequate predictive performance. Conclusions A robust bevacizumab pharmacokinetic model was developed and externally validated, which may be used to simulate bevacizumab exposure to optimize dosing strategies. Asian and non-Asian patients exhibited similar bevacizumab pharmacokinetics. Given the similarity in pharmacokinetics among monoclonal antibodies, this may inform pharmacokinetic studies in different ethnic groups for other therapeutic antibodies. Electronic supplementary material The online version of this article (doi:10.1007/s00280-016-3079-6) contains supplementary material, which is available to authorized users.