OBJECTIVE: To longitudinally evaluate unbound and total serum phenytoin concentrations during intravenous phenytoin maintenance dosage and to determine the relationship among phenytoin protein binding, serum albumin, and unbound fatty acid concentrations in patients with head injuries during intensive care unit (ICU) and convalescent care. DESIGN: Serum albumin and phenytoin unbound fraction were determined twice weekly during ICU and convalescent cave in 10 patients receiving phenytoin following acute brain injury. Phenytoin protein binding was also determined in 10 healthy control subjects. MAIN OUTCOME MEASURES: Longitudinal serum phenytoin concentrations associated with dosage adjustments targeted to achieve unbound phenytoin serum concentrations between 1.0 and 2.0 mg/L were documented during ICU and convalescent care. Longitudinal phenytoin protein binding was correlated with serum albumin and unbound fatty acid concentrations in neurotrauma patients. RESULTS: ICU patients received intravenous therapy for a mean of 15.0 days. The mean ± SD initial phenytoin intravenous dosage regimen of 6.0 ± 0.7 mg/kg/d resulted in mean ± SD total and unbound phenytoin concentrations of 3.2 ± 2.3 and 0.3 ± 0.2 mg/L. Two patients had seizures associated with low phenytoin concentrations. Four patients continued to receive oral phenytoin therapy during convalescent care; phenytoin dosage requirements decreased over time in these patients. During acute and convalescent care, the phenytoin unbound fraction ranged from 6.0% to 18.3% and correlated with albumin (r2 = 0.61, p < 0.0001) but did not correlate with unbound fatty acid concentrations. The mean phenytoin unbound fraction was 10.1% and 8.9% for the ICU and convalescent patients with brain injuries, respectively, and was 7.0% for the healthy volunteers. CONCLUSIONS: Phenytoin protein binding was significantly correlated with albumin and was more variable in ICU and convalescent patients with brain injuries than in healthy volunteers. The high dosage requirements and subtherapeutic unbound phenytoin concentrations observed during acute care are best explained by increased metabolism. Phenytoin dosage requirements decreased during convalescence.