Lower exposure and faster clearance of bevacizumab in gastric cancer and the impact of patient variables: analysis of individual data from AVAGAST phase III trial

Abstract

Altered pharmacokinetics of antibody drugs has been reported in advanced gastric cancer (AGC). We aim to evaluate bevacizumab pharmacokinetics in AGC from the Phase III trial (AVAGAST), and explore the influence of patient variables. Bevacizumab concentrations (Cp) were measured in plasma samples taken following disease progression from 162 patients (7.5 mg/kg every 3 weeks). Predicted Cp [median and 90% prediction interval] was simulated using the population pharmacokinetic model established for other cancers (PPK model) and compared to observed Cp. Bevacizumab clearance was estimated using NONMEM and compared between subgroups. Patient characteristics of AGC are similar to other cancers except for lower body weight despite higher percentage of males. Eighty-five percent of observed Cp was below the median predicted Cp and 38% below the lower boundary of the 90% prediction interval. Median bevacizumab clearance in AGC was 4.5 versus 3 mL/day/kg in other cancers. Bevacizumab clearance was significantly faster (p < 0.05) in patients without gastrectomy (n = 42) or lower albumin. Clearance appeared to be faster in patients with lower total protein, higher ECOG scores, more metastatic sites, and poorer response. No significant difference in bevacizumab concentrations and clearance was observed between Asian and Non-Asian patients. AGC patients exhibited significantly lower bevacizumab exposure due to an approximate 50% increase in clearance versus other cancers. Bevacizumab is cleared faster in patients without prior gastrectomy. No significant difference in bevacizumab pharmacokinetics was observed between Asian and Non-Asian patients. The underlying mechanism for faster bevacizumab clearance in AGC is unknown and warrants further research.

Trial registration: ClinicalTrials.gov NCT00548548.

Conflict of interest statement

All authors are employees of Genentech and hold stock in F. Hoffmann-La Roche. In addition, Jin Jin also holds stock in Eli Lilly.

Figures

Fig. 1

Visual predictive check (VPC) comparing observed and expected (simulated) bevacizumab concentrations. Only the 151 patients with a sampling time within 50 days after the last bevacizumab dose are shown (93.2% of the 162 evaluable patients)

Fig. 2

Comparison of individual bevacizumab clearance that is predicted (expected) based on the reference population pharmacokinetic model and estimated based on observed data via Empirical Bayes estimation. Solid symbols represent the mean. Hollow symbols represent individual data. Ref: reference

Fig. 3

Significant correlation of bevacizumab clearance with patient variables a baseline body weight, b gastrectomy, and c baseline serum albumin. Solid symbols represent the mean. Hollow symbols represent individual data. The clearance in other solid tumors was calculated by taking into the account the covariate values in this study

Fig. 4

Comparison of bevacizumab exposure and clearance between Asian and Non-Asian patients a bevacizumab through concentrations collected between days 19 and 23 after the last dose and b bevacizumab clearance. Solid symbols represent the mean. Hollow symbols represent individual data. The clearance in other solid tumors was calculated by taking into the account the covariate values in this study