A maturation model for midazolam clearance

Background: Physiological-based pharmacokinetic models have been used to describe midazolam clearance (CL) maturation. There are no maturation descriptors of CL from neonate to adulthood based on reported estimates at different ages. Methods: Published CL estimates after intravenous administration from time-concentration profiles were used to construct a maturation model based on size and age. Curve fitting was performed using nonlinear mixed effects models. Results: There were 16 publications reporting an estimate of CL after intravenous administration in children, although few estimates were available from 44-80 weeks postmenstrual age (PMA). CL maturation, standardized to a 70 -kg person was described using the Hill equation. Mature CL was 523 (CV 32%, 95% CI 469, 597) ml.min(-1).70 kg(-1). The maturation half-time was 73.6 (95% CI 59.4, 80.0) weeks PMA and the Hill coefficient 3 (95% CI 2.2, 4.1). Predicted CL changes with age based on this model were in close agreement with physiologically based pharmacokinetic (PBPK) models. A comparison with a published PBPK model predictions revealed a root mean squared prediction error (precision) of 4.0% (95% CI 1.1, 5.8) and bias was -0.9% (95% CI -4.3, 2.6). Conclusions: Previously published pharmacokinetic parameters can be used to develop maturation models that address gaps in current knowledge regarding the influence of age on a drug's disposition. If a midazolam sedation target concentration of 0.1 mg.l(-1), similar to that given to adults, is assumed, then we might anticipate steady-state infusion rates of 0.014 mg.kg(-1).h(-1) in neonates, 0.05 mg.kg(-1).h(-1) in a 1-year-old, 0.06 mg kg(-1).h(-1) in a 5-year-old and 0.05 mg.kg(-1).h(-1) in a 12-year-old child. Age-related pharmacodynamic differences that will affect dose and the impact of active metabolites on response are not yet quantified.