Objective: To find limited sampling strategies (LSS) for prediction of the real AUC using the RIA analytical method.
Method: Blood samples of 40 male renal transplant patients taken pre-dose and after 0.5, 1, 1.5, 2, 3, 5, 8, and 12 h in the steady-state were analyzed with HPLC and the specific RIA method. I. Eight equations for AUC0-12 and one for AUC0-8 obtained from the literature, that produced the mean percentage prediction error (%PE) < ± 15% and absolute %PE < 30% in 95% of predictions, were analyzed for possibility to predict the real AUC of CsA. II. Multiple regression analysis (MRA) was provided for the AUC equation proposal. Patients were divided into two groups according to the AUC0-12. Group I was used for LSS : s proposals while Group II for validation. The bias and precision were expressed as %PE, r2 and RMSE. The relationship of %PE interassay and with LSS:s was expressed as Pearson correlation r. GraphPad InStatt Software was used for MRA and Pearson r calculation.
Results: None of the equations described in the literature predicts AUC of CsA proprietarily. Seven equations for AUC0-12 and five for AUC0-8 were proposed with MRA for prediction of real AUC from RIA values.
Conclusions: LSS:s can moderate the interassay %PE in AUC of CsA. New patients should be tested with both RIA and HPLC for the level of overestimation. The conversion factors should be calculated for patients with an overestimation higher than 90%. Our equation 251.09 + 0.5195 × C1h + 4.926 × C3h or 196.13 + 4.526 Â× C0h + 2.089 × C1.5h for AUC0-12, and 171.80 + 0.4759 × C1h + 4.132 × C3h for AUC0-8 may be used in patients with medium or low RIA and HPLC differences. Repeated analysis with HPLC is thus suggested in cases with AUC:s results close to the lower or upper margin of the therapeutic window.