Purpose. The differences in free fraction serum valproic acid concentrations between inpatients and outpatients were compared, and factors associated with therapeutic discordance were evaluated. Methods. This retrospective, single-center, cross-sectional study included patients with both a free and a total valproic acid concentration drawn within 30 minutes of each other between January 7, 2003, and June 1, 2011. Serum valproic acid concentrations were stratified by admission status (inpatient versus outpatient). In the primary analysis, for patients who had multiple paired concentrations drawn, one free valproic acid level and one total valproic acid level (i.e., one pair) were chosen at random for each inpatient. Information regarding patient demographics, laboratory data, and concomitant medication therapy was collected. Single and multivariable logistic regressions were performed to determine the odds of therapeutic discordance. Results. During the study period, inpatient concentrations were measured in 220 patients, and outpatient concentrations were measured in 41 patients. The median total valproic acid concentration in the inpatient group was significantly lower than that in the outpatient group (54 μg/mL versus 83 μg/mL, p < 0.001). The median free fraction of inpatient levels was nearly twice as high as that of outpatient levels (28.8% versus 15.5%, p < 0.001). Levels drawn in inpatients were therapeutically discordant 63% of the time compared to only 19% in outpatients (p < 0.001). Multivariable logistic regression identified low albumin as an independent risk factor for therapeutic discordance. Conclusion. The median free fraction of valproic acid was significantly higher in inpatients than in outpatients. Low albumin concentration was a predictor of discordance between free and total valproic acid concentrations.
ASJC Scopus subject areas
- Health Policy