The aim of this study was to evaluate limited sampling designs to estimate the maximal concentration (C(max)) and area under the curve (AUC) of mizoribine in pediatric patients with renal disease. We utilized 48 serum mizoribine concentration profiles obtained from the full (6-point) sampling pharmacokinetic test, and estimated 48 individual C(max) and AUC values accurately with Bayesian analysis using the full sampling data. We then developed limited sampling models (LSM) for C(max) and AUC using 1-4 serum mizoribine concentration data points. The C(max) and AUC estimation performance of the Bayesian and LSM analysis was fairly good in the 3-point (2, 3, and 6 hr after the dose) sampling design. In addition, the C(max) estimation performance of the Bayesian and LSM analysis deteriorated only marginally even in the 1-point (3 hr) sampling design. On the other hand, the AUC estimation performance seemed to be inadequate in the 1-point (3 hr) sampling design; however, it improved markedly in the 2-point (3 and 6 hr) sampling design. These findings suggested that the 1-point (3 hr) sampling design is promising for approximate C(max) estimation, but that the 2-point (3 and 6 hr) sampling design is preferable to estimate the AUC of mizoribine.