Prediction of flucytosine-induced thrombocytopenia using creatinine clearance

Aims: To evaluate and investigate the efficacy and accuracy of an earlier d erived relationship between thrombocyte nadir and mean creatinine clearance in order to make it possible to predict flucytosine (5-FC)-related thrombo cytopenia. Methods: The earlier derived relationship between thrombocyte na dir and mean creatinine clearance is validated using a patient population o f 35 intensive care unit (ICU) patients. Patients have to meet the inclusio n criteria as defined in the initial study (i.e., developing thrombocytopen ia during 5-FC treatment or remaining within the normal thrombocyte range; individual pharmacokinetic parameters within certain limits; a minimum of t hree measured 5-FC serum levels). Results: Of a total of 35 patients, 11 ar e eligible since they meet the inclusion criteria as defined in the initial study. A significant relationship between the observed and predicted throm bocyte nadir values can be derived: Y-pred = 0.608.X-obs + 24.330 (r = 0.61 5; t(obs) = 2.337; t at p < 0.05 and 9 degrees of freedom = 2.262). The pre dicted value of the thrombocyte nadir corresponds with a mean prediction er ror (bias) of -18.3 (95% CI -45.9; 9.4) and a root mean squared prediction error (precision) of 48.3 (95% CI 23.9-63.9). Conclusion: The earlier deriv ed relationship between thrombocyte nadir and mean creatinine clearance is accurate and precise. However, due to the strict inclusion criteria used to derive and validate this relationship, it cannot be applied to all 5-FC-tr eated patients. Copyright (C) 2000 S. Karger AG, Basel.