PGx7610: Genetic Evaluation of Hepatotoxicity in Pazopanib Studies

Trial description: This confirmatory study will evaluate associations between specified genetic variants and on-treatment alanine aminotransferase (ALT) elevation in pazopanib-treated subjects from 23 clinical studies. The genetic variants are the HLA B*57:01 allele, and 16 single nucleotide polymorphisms (SNPs), which were all significantly or suggestively associated with ALT elevation in pazopanib-treated subjects in a previous exploratory analysis (VEG117365, PGx6652). The primary endpoint will be maximum on-treatment ALT (in units of laboratory specific upper limit of normal [ULN]). Association will be tested using a normal linear model after appropriate transformation, adjusted for baseline ALT, clinical study and arm, ancestry principal components, sex, and age at baseline. For secondary endpoints (ever vs never on-treatment ALT>3xULN, ever vs never on-treatment ALT>5xULN, time until first on-treatment ALT>3xULN, time until first on-treatment ALT>5xULN), association will be tested using logistic regression and Cox regression, adjusted for the same covariates. The primary analysis will test association with the HLA B*57:01 allele, and secondary analyses will test association with the 16 SNPs. Co-primary analyses will be conducted in all patients treated with pazopanib, and also in the subgroup of patients on pazopanib monotherapy. The overall false positive rate for these cop-primary analyses will be controlled by an alpha spending rule allocating alpha=0.04 for the analysis of all patients and alpha=0.01 for the analysis of patients on pazopanib monotherapy.