Valproate protein binding is highly variable in ICU patients and not predicted by total serum concentrations: a case series and literature review.

Document Type


Publication Date



Critical Care, Neurology, Pharmacy

Journal Title


MeSH Headings

Academic Medical Centers, Adult, Aged, Anticonvulsants, Critical Illness, Female, Humans, Intensive Care Units, Male, Middle Aged, Prospective Studies, Protein Binding, Serum Albumin, Valproic Acid


STUDY OBJECTIVE: The free fraction of valproate (the pharmacologically active moiety, normally 5-10%) may vary significantly in critically ill patients, but this topic is understudied, with only four prior intensive care unit (ICU) case reports. The objective of this study was to evaluate the range of valproate plasma protein binding in ICU patients.

DESIGN: Observational study of consecutive ICU patients.

SETTING: Neurocritical and medical critical care services in a nonuniversity academic medical center.

PATIENTS: Consecutive ICU patients treated with valproate with serum albumin less than 4 g/dl.

MEASUREMENTS AND MAIN RESULTS: Simultaneous total and free trough serum valproate concentrations were measured as were serum creatinine, blood urea nitrogen, albumin, platelets, and transaminase values. The reference concentration range was 50-125 mg/L (total) and 5-17 mg/L (free). Valproate concentrations were categorized as within reference range, low, or high, and as concordant if both concentrations were in the same category. Data are reported as median (interquartile range). Fifteen patients (nine men) were evaluated. The median age was 63 (34-70) years. The valproate dose was 3 g/day (35 mg/kg/day). No patient had a valproate free fraction of 5-10%; the median was 48%, and the range was 15-89%. Total and free concentrations showed poor correlation (0.43) and were concordant in only two patients (both in the reference range). Free valproate concentration was poorly predicted by an equation correcting for albumin (r = 0.45). Suspected adverse drug events occurred in 10 patients: hyperammonemia in 7 of 12 tested (58%), elevated transaminases in 2 of 15 (13%), and thrombocytopenia in 5 of 15 (33%).

CONCLUSIONS: Protein binding of valproate was highly inconsistent in this cohort of ICU patients, and total valproate concentrations did not predict free concentrations, even when correcting for albumin. Additional research to define best practice for dosing and monitoring valproate and the relationship between free valproate concentrations and clinical or adverse effects in ICU patients is needed.



First Page


Last Page