Establishment of a population pharmacokinetic model to estimate valproic acid clearance in psychiatric patients

• Main Authors: Hui-Ching Huang, 黃輝慶
• Other Authors: Ying-Lung Tseng
• Format: Others
• Language: en_US
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/89873898457134153004

碩士 === 慈濟大學 === 藥理暨毒理學研究所 === 94 === Background: Valproic acid (VPA; valproate) is a drug that has been widely used as an anticonvulsant since the 1970s. VPA is currently approved by the FDA (USA) for the treatment of absence seizures, mixed seizures, and bipolar I disorder, manic phase. The efficacy and toxicity of VPA in patients with bipolar disorders are associated with serum VPA concentration. VPA serum levels between 45 and 100-125 mg/L are much more likely to have efficacious and well-tolerated response. Toxic effects are frequently associated with daily doses over 1800 mg or blood levels over 100 mg/L .The aim of this study is to construct a formula intended to estimate VPA clearance in routine clinical care of a population of psychiatric patients in Taiwan. Methods: Prospective steady state serum concentration data (n=534) collected from 109 adult psychiatric patients who have been receiving VPA were studied. Data were analyzed according to a one-compartment model using the NONMEM program. Several possible influential parameters, such as VPA daily dose, gender, age, height, total body weight, BUN, creatinine and albumin were investigated. Results: In this study, a wide inter-individual variability was found to be present in drug dose and serum concentration distribution. The final regression model established for VPA clearance (CL) was as follows: CL(L/h) = 0.185＋ (0.478 / 280.5) × 0.25 × Dose ＋ (0.0660 / 38.7) × AGE. Validation of the established model was carried out by 10 psychiatric patients. The results demonstrated that Equation Cpss pre = 0.0500 × Dose / CL showed the least biased and more accurate in predicting the steady state serum VPA concentration in psychiatric patients who ingested non-sustained-release VPA formulation four times a day. In addition, another group consisting of 20 patients who ingested sustained-release VPA was also tested using the established model. The results showed the best prediction power obtained was Equation Cpss pre = 0.0556 × Dose / CL. In the Equations, Dose indicates the total daily dose (mg) of VPA, Cpss pre represent prediction serum concentration at steady state. Conclusions: This model may be used for a priori recommendation and dose optimization of VPA administration in Taiwanese psychiatric patients. In contrast to traditional practice on determining a proper drug dose based on experience, this model not only reduces time on looking for a suitable drug dose for a patient, but also ensures treatment effects and avoids side effects, in particular to those patients who have difficulties on self-care.