Clinical Consequences of Occult Free Valproate Toxicity in Critically Ill Adult Patients: A Multicenter Retrospective Cohort Study
Andrew J Webb,David J Gagnon,Caitlin S Brown,Richard R Riker,Natasha D Lopez,Melanie Z Goodberlet,Michael J Schontz,Kaylee K Marino,Sacha N Uljon,Sahar F Zafar,Eric S Rosenthal
DOI: https://doi.org/10.1101/2024.05.16.24307506
2024-05-19
Abstract:Valproate has wide pharmacokinetic variability and a narrow therapeutic index. Its protein binding is unpredictable, particularly among critically ill patients who may experience unexpectedly elevated free concentrations. We sought to identify the clinical consequences and determinants of occult free valproate toxicity in critically ill adults.
We conducted a multicenter retrospective cohort study of adult patients admitted to an intensive care unit (ICU) who were receiving valproate and had concurrent total and free valproate concentrations measured. We examined whether valproate concentrations were independently associated with adverse drug effects (ADEs) including thrombocytopenia, hepatotoxicity, hyperammonemia, and pancreatic injury. Determinants of occult toxicity were also identified using logistic mixed-effects models, adjusting for age, weight, albumin, propofol and aspirin use, and blood urea nitrogen (BUN). Occult toxicity was defined as a free valproate concentration that was discordant with total concentration (e.g., supratherapeutic free concentration associated with therapeutic total concentration).
311 unique patients (mean age 58 [±17] years, 36% female, 31% non-white, and 29% on valproate prior to admission) with 550 concurrent free and total valproate concentration pairs met inclusion criteria. The median (IQR) total valproate concentration was 46 mcg/mL (34-63) and the median free valproate concentration was 17 mcg/ml (11-23); median free fraction was 35% (25-63%). Eighty-four percent of total valproate concentrations represented occult free toxicity; a therapeutic total with a supratherapeutic free valproate concentration was the most common pattern (32% of concentration pairs). Each 2.5 mcg/mL increase in free valproate concentration was associated with thrombocytopenia (adjusted unit odds 1.15, 95% CI 1.05-1.26) and hepatotoxicity (adjusted unit odds 1.11, 95% CI 1.05-1.18). Albumin concentration (adjusted odds; aOR 0.17, 95% CI 0.08-0.36), BUN (aOR 1.36, 95% CI 1.09-1.70), and propofol exposure (aOR 3.06, 95% CI 1.38-6.79) were associated with occult toxicity.
Free valproate concentrations should be measured in critically ill patients because it is associated with ADEs and is often underrepresented by total concentrations. Most critically ill patients are at risk, especially those with hypoalbuminemia, uremia, and lipid exposure.