It is currently being debated whether pharmacokinetic monitoring of mycophenolic acid (MPA), the active constituent of mycophenolate mofetil (MMF), can optimize MMF therapy after organ transplantation. This open-label longitudinal study in pediatric renal transplant recipients was designed to investigate the pharmacokinetic (PK)/pharmacodynamic relationship of total and free MPA and to establish PK values for the assessment of an individual's MPA PK parameters. Fifty-four children, aged 2.2 to 17.8 yr, on an immunosuppressive triple regimen consisting of cyclosporin A, prednisone, and MMF (600 mg/m(2) body surface area twice daily) were investigated 1 wk and 3 wk (initial phase) and 3 mo and 6 mo (stable phase) after renal transplantation. MPA was measured by reverse phase HPLC, free MPA by HPLC after separation by ultrafiltration. There was an association between the risk of acute rejection episodes and MPA-AUC(0-12) values or MPA predose levels; by receiver operating characteristic analysis, an AUC(0-12) of 33.8 mg x h/L in the initial phase posttransplant had a diagnostic sensitivity of 75% and a diagnostic specificity of 64% for discrimination of patients with acute rejections. The respective discrimination threshold for the MPA predose concentration was 1.2 mg/L with a sensitivity of 83% and a specificity of 64%. In contrast, high free, but not total, MPA-AUC(0-12) values were associated with an increased risk of the MMF-related side effects leukopenia and/or infections. These data indicate that therapeutic drug monitoring of MPA has the potential for optimization of MMF efficacy in this patient population by steering patients away from the low values of MPA PK variables that are associated with an increased rejection risk. For the assessment of the toxic risk of MMF regarding leukopenia and/or infections, measurement of free MPA appears to be more appropriate.