The impact of body composition parameters on severe toxicity of nivolumab

Background: The occurrence of severe, acute limiting toxicity in patients receiving anti-programmed cell death receptor-1 monoclonal antibodies, such as nivolumab, is largely unpredictable. Sarcopenia was found to be associated with anti-cytotoxic T-lymphocyte-associated protein 4 acute toxicity. We explore the clinical and pharmacological parameters influencing nivolumab toxicity, including body composition. Methods: From June 2015 to January 2017, all consecutive patients treated with nivolumab in our institution were prospectively included. We studied the relationship between muscle mass assessed by computed tomography, nivolumab trough level (C-min) at day 14 assessed using the enzyme-linked immunosorbent assay method, and the occurrence of immune grade III or IV toxicity or any toxicity leading to treatment discontinuation (immune-related acute limiting toxicity [irALT]). Results: In our population (n = 92) with a majority of lung cancer (72%), forty-five (51.7%) patients were sarcopenic. The median plasma nivolumab C-min at day 14 was 15.4 mu g/mL (interquartile range = 11.8-21.0). In multivariate analysis, hypoalbuminaemia (<35 g/L) was independently associated with low nivolumab C-min on day 14 (odds ratio [OR] = 0.09; 95% confidence interval [CI] = 0.01-0.59, p = 0.01) and overweight/obesity with high nivolumab C-min on day 14 (OR = 5.94; 95% CI = 1.25-28.29, p = 0.03). We observed 22 irALTs in 19 patients (21%). The most frequent irALT was respiratory (6.5%) disorders and gastrointestinal (4.3%) disorders. Patients with sarcopenia were at significantly increased risk of experiencing an irALT (OR = 3.84; 95% CI = 1.02-14.46, p = 0.047). No association was found between toxicity and nivolumab plasma C-min at day 14. Conclusions: Our results highlight the importance of assessing body composition and suggest that sarcopenia could predict severe immune-related toxicity of nivolumab in real life. (C) 2019 Elsevier Ltd. All rights reserved.