The impact of polypharmacy on toxicity from chemotherapy in elderly patients: Focus on cytochrome P-450 inhibition and protein binding effects

Abstract

9505 Background: Drug-drug and disease-drug interactions are significant clinical issues in the elderly. Of the elderly taking 3+ chronic medications, 33% are rehospitalized within 6 months of discharge from a hospital. Drug interactions can also be relevant to toxicity from chemotherapy, such as ketoconazole and irinotecan, via p450 inhibition, or oxaliplatin and warfarin, possibly via competition for protein binding. Methods: To assess the impact of polymedication on toxicity from chemotherapy, we analyzed a cohort of 290 patients 70+ receiving chemotherapy at Moffitt Cancer Center. In previous analyses using unweighted inhibition/induction of p450 interactions we found results approaching significance, and albumin level predicting non-hematologic toxicity. We refined our analysis by using weighted effects according to the strength of p450 inhibition/induction. We also assessed the impact of competition by highly protein-bound drugs (>50%). Stepwise logistic regression was used to assess predictors of G4 hematologic (G4H) or G3–4 non-hematologic (G3–4NH) toxicity. CART software was used to assess heterogeneity of effects. Results: In unconstrained models, higher likelihood of G4H was predicted by high competitive protein binding (p= 0.003), bone marrow invasion (p= 0.027), tumor stage (p=0.005), mean bilirubin (p= 0.025), and mean red blood cells during chemotherapy (p= 0.012). In unconstrained models, higher likelihood of G3–4H was predicted by higher BMI (p=0.048), tumor stage (p=0.029), lower mean albumin during chemotherapy (p=0.005), and higher AST before chemotherapy initiation (p=0.018). When entry order was preset with MAX2, p450 interaction, and competitive protein binding being entered first, MAX2 score (p=0.016) and competitive protein binding (p=0.009) predicted higher likelihood of G4H, but no predictor significantly influenced G3–4NH. The CART analyses indicate a dominant effect of MAX2 for G4H, and of mean albumin for G3–4NH, with no major heterogeneity. Conclusions: Our results suggest that drug distribution factors play important roles in chemotherapy toxicity in elders. The role of protein binding in the pharmacokinetics/dynamics of chemotherapy drugs should be further explored. No significant financial relationships to disclose.