Objectives To study the effect of organ failure and inflammation on midazolam clearance in critically ill children, using population pharmacokinetic modeling.

Methods A total of 83 critically ill children (median age 5 months (range 1 day-17 years), n=523 samples) receiving intravenous midazolam for continuous sedation during mechanical ventilation were included. Disease severity was described using the validated and clinically used scores PELOD, PIM2 and PRISM II. Cytokines (IL-1, IL-2, IL-6, TNF-a) and C-reactive protein (CRP) were used as markers for inflammation. A population pharmacokinetic model for midazolam was developed using NONMEM 7.3. Body weight, age, severity of organ failure and inflammatory markers were considered as potential covariates.

Results In a two-compartmental PK model, body weight was found as most significant covariate for clearance and volume of distribution. Moreover, both severity of organ failure (PELOD) and inflammation (IL6 and CRP) were significant determinants of clearance (p<0.01), and either of these factors improved the model significantly. With increasing number of organ failures, midazolam clearance significantly reduced. CRP was linearly correlated with clearance (slope −0.095), with higher CRP levels resulting in lower clearances. Either one of the covariates could explain part of the variability in clearance.

Conclusion For midazolam clearance, apart from body weight, we found organ failure reflected by the PELOD score, and inflammation reflected by IL6 and CRP, as significant covariates. Most likely this effect is due to reduced activity of CYP3A in critically ill mechanically ventilated children. Both CRP concentration and organ failure should be considered when dosing midazolam and potentially other CYP3A substrates in critically ill children.