RESUMO
PURPOSE: The aim of this study was to develop a clinically applicable limited sampling strategy for ambulatory Caucasian kidney transplant patients to estimate area under the curve in a 24-h period (AUC0-24) of prolonged-release tacrolimus. METHODS: Twenty six kidney recipients, at least 6 months after transplantation, receiving prolonged-release tacrolimus, were enrolled. In each patient, seven blood samples were collected during a period of 24 h by use of the validated dried blood spot method. Best subset selection multiple linear regression was performed to derive limited sampling strategy (LSS). The equations were constrained to include a maximum of three samples collected within 4 h after the intake to maintain clinical applicability. To assess the predictive performance of LSS, residuals for each patient were calculated based on models fitted to a dataset where that patient was omitted. RESULTS: The prediction formula for the AUC(0-24) using the time points 0, 2, and 4 h after ingestion (C(0h)-C(2h)-C(4h)) provided the highest correlation with the AUC(0-24) (r(2) = 0.95): AUC0-24 = 44.9 + 8.9 × C(0h) + 2.1 × C(2h) + 7.6 × C(4h). Measures for bias and precision, i.e., median percentage prediction error (MPPE) and median absolute prediction error (MAPE), were 0.4 and 4.8%, respectively. For the same patients, the correlation between C(24h) and AUC0-24 was worse (r(2) = 0.77) while MPPE and MAPE were 6.2 and 7.2%, respectively. CONCLUSION: In the outpatient department, a LSS using C(0h)-C(2h)-C(4h) can be used for reliable estimation of the AUC(0-24) of prolonged-release tacrolimus.