This study investigated the relationship between plasma voriconazole concentrations (pVC) and risk of visual adverse events (VAEs) or liver function test (LFT) abnormalities using longitudinal logistic regression. Seven-day mean pVC were calculated from 2,925 plasma samples (1,053 patients); in each 7-day period, the presence or absence of VAEs/abnormal LFTs was analyzed as a binary outcome variable. There was a relationship between pVC and risk of VAE (P = .011) and a weaker, but statistically significant, association with risk of aspartate transaminase (AST), alkaline phosphatase (ALP), or bilirubin but not alanine transaminase (ALT) abnormalities. The odds ratios of LFT abnormalities per 1 mug/mL pVC increase ranged from 1.07 to 1.17. Maximum weekly occurrences were 10%, 8%, 5%, and 14% for AST, ALT, ALP, and bilirubin abnormalities, respectively. Receiver-operating characteristic curve analysis indicates that individual pVC cannot be used to predict subsequent LFT abnormalities.