In three children undergoing bone marrow transplantation, the cyclosporine (CyA) regimen was designed by the Bayesian method, and its utility was evaluated.CyA was intravenously infused for the first few weeks after transplantation and thereafter was orally administered. The initial intravenous dose was 3 mg/kg/day, but pharmacokinetic analysis showed that this dose failed to achieve the target through concentration of 100 to 150 ng/ml, and the dose was increased in all three patients.The whole blood concentrations later increased and the dose was decreased again in all three patients.The dose was modified according to the Bayesian feedback method during treatment.To evaluate predictability by the Bayesian method, the mean prediction error (ME) was used as an index of bias, the mean absolute prediction error (MAE) for precision, and the root mean squared error (RMSE) for indicating the discrepancy of the prediction.The ME value was negative because the whole blood levels tended to increase, but the MAE and RMSE values were 22.5 ng/ml and 29.5 ng/ml, respectively, which were considered to be in the permissible range with regard to the target concentration.
To examine the cause of the increase in the CyA concentration, the correlations between trough levels and laboratory findings were studied.A significant relation was noted between ALP and SCr in multiple stepwise regression analysis with the trough levels of CyA as the response variable and GOT, LAP, ALP, BUN, SCr, Hct and TT levels as regressor variables. Since a correlation was recognized between CyA undergoing biliary excretion and ALP an indicator of cholestasis, the possibility of a cholestatic liver disorder was suggested as a cause of the elevation of the CyA concentration.
