Propylene glycol is a commonly used diluent in several pharmaceutical preparations, including the sedative lorazepam. Fifty critically ill patients receiving continuous-infusion lorazepam for a minimum of 36 hours were prospectively evaluated to determine the extent of propylene glycol accumulation over time, characterize propylene glycol clearance in the presence of critical illness, and develop a pharmacokinetic model that would predict clearance based on patient-specific clinical, laboratory, and demographic factors. In this cohort, the median lorazepam infusion rate was 2.1 mg/h (0.5-18). Propylene glycol concentration correlated poorly with osmolality, osmol gap, and lactate. In all, 8 patients (16%) had significant propylene glycol accumulation (>25mg/dL). When propylene glycol concentrations were >25 mg/dL, the median lorazepam infusion rate before sample collection was higher, 6.4 (1.9-11.3) versus 2.0 (0.5-7.4) mg/h (P =.0003). A linear first-order model with interoccasion variability on clearance adjusted for total body weight and Acute Physiology and Chronic Health Evaluation II score predicted propylene glycol concentration.