Limited sampling strategy to predict mycophenolic acid area under the curve in pediatric patients with nephrotic syndrome: a retrospective cohort study

• Published in: European journal of clinical pharmacology 2019-09, Vol.75 (9), p.1249-1259
• Main Authors: Sobiak, Joanna, Resztak, Matylda, Pawiński, Tomasz, Żero, Paweł, Ostalska-Nowicka, Danuta, Zachwieja, Jacek, Chrzanowska, Maria
• Format: Article
• Language: English
• Subjects: Adolescent; Algorithms; Area Under Curve; Bias; Biomedical and Life Sciences; Biomedicine; Child; Child, Preschool; Children; Cohort analysis; Computer programs; Female; High-performance liquid chromatography; Humans; Immunosuppressive Agents - blood; Immunosuppressive Agents - pharmacokinetics; Kidney diseases; Liquid chromatography; Male; Measurement methods; Mycophenolic acid; Mycophenolic Acid - blood; Mycophenolic Acid - pharmacokinetics; Nephrotic syndrome; Nephrotic Syndrome - blood; Patients; Pharmacokinetics and Disposition; Pharmacology/Toxicology; Retrospective Studies; Sampling; Software; Therapeutic drug monitoring
• ISSN: 0031-6970; 1432-1041
• DOI: 10.1007/s00228-019-02701-5

Purpose Limited sampling strategy (LSS) is a precise and relatively convenient therapeutic drug monitoring method. We evaluated LSSs for mycophenolic acid (MPA) in children with nephrotic syndrome treated with mycophenolic mofetil (MMF) and validated the LSSs using two different approaches. Methods We measured MPA plasma concentrations in 31 children using HPLC-UV method and received 37 MPA pharmacokinetic profiles (0–12 h). For six children, MPA profiles were estimated twice after two MMF doses. LSSs were developed using multilinear regression with STATISTICA and R software and validated using validation group and bootstrap method, respectively. Results The best three time point equations included C 1 , C 3 , C 6 (good guess 83%, bias − 2.78%; 95% confidence interval (CI) − 9.85–0.46); C 1 , C 2 , C 6 (good guess 72%, bias 0.72%; 95% CI − 5.33–7.69); and C 1 , C 2 , C 4 (good guess 72%, bias 2.05%; 95% CI − 4.92–13.01) for STATISTICA software. For R software, the best equations consisted of C 1 , C 3 , C 6 (good guess 92%, bias − 2.69%; 95% CI − 27.18–33.75); C 0 , C 1 , C 3 (good guess 84%, bias − 2.11%; 95% CI − 24.19–22.29); and C 0 , C 1 , C 2 (good guess 84%, bias − 0.48%; 95% CI − 30.77–54.07). During validation, better results were obtained for R evaluations, i.e., bootstrap method. Conclusions The most useful equations included C 0 , C 1 , C 3 and C 0 , C 1 , C 2 time points; however, the most precise included C 1 , C 3 , C 6 time points because of MPA enterohepatic recirculation. Better results were obtained for bootstrap validation due to greater number of patients. Validated LSS should be used only in the population for which it was developed. As there is growing evidence that underexposure of MPA is associated with insufficient treatment response, we recommend the introduction of therapeutic drug monitoring for MPA in children with nephrotic syndrome.