Background: Eculizumab is effective in treating paroxysmal nocturnal hemoglobinuria and atypical hemolytic uremic syndrome. Given its high cost and the clinical risks associated with undertreatment or overtreatment, dose individualization is essential. Model-informed precision dosing can optimize therapy using pharmacokinetic (PK) model predictions. This study conducted a primarily internal validation to evaluate the predictive performance of a previously published PK model of eculizumab for model-informed precision dosing. Methods: The model's ability to predict future free trough concentrations was assessed using retrospective data from therapeutic drug monitoring. Acceptable bias and precision were defined as mean prediction error and normalized root mean square error less than 25%. Dose adjustments were evaluated based on target trough concentrations of 50 µg/mL for dose escalation and 200 µg/mL for interval extension. Results: The model adequately predicted future trough concentrations. In patients with paroxysmal nocturnal hemoglobinuria, bias and precision were −3.0% and 10.3%, respectively, when using one trough concentration; −8.4% and 24.9%, respectively, when using a trough and a peak concentration; and −4.7% and 16.3%, respectively, when using two trough concentrations. In patients with atypical hemolytic uremic syndrome, bias and precision were 6.1% and 35.5%, respectively, when using one trough concentration; 15.1% and 25.7%, respectively, when using a trough and a peak concentration; and 11.2% and 23.2%, respectively, when using two trough concentrations. The model's predictions led to correct dose recommendations for >80% of patients in most analyses. Conclusions: These results demonstrate that the eculizumab PK model can predict future trough concentrations with acceptable bias and precision.
Heida et al. (Fri,) studied this question.