Prediction of morphine dose in humans.

BACKGROUND: Morphine is widely used throughout the human life span. Several pharmacokinetic models have been proposed to predict how morphine clearance changes with weight and age. This study uses a large external data set to evaluate the ability of pharmacokinetic models to predict morphine doses.
METHODS: A data set of morphine clearance estimates was created from published reports in premature neonates, full-term neonates, infants, children, and adults. This external data set was used to evaluate published models for morphine clearance as well as other models proposed for use in neonates and infants. Morphine clearance predictions were used to predict morphine dose rates to achieve similar target concentrations in all age groups.
RESULTS: An allometric ¾ power model using weight combined with a sigmoid maturation model using postmenstrual age successfully predicted the morphine dose rate (within 25% of target) in all age groups except infants [predicted dose 30% under target (95% CI, 7-46%)]. Other published models based on empirical allometric scaling all made unacceptable predictions (>100% of target) in at least one age group.
CONCLUSIONS: Clearance based on empirical allometric scaling predicted unacceptable doses. Theory-based allometric scaling combined with a maturation function has been confirmed by external evaluation to provide a sound basis for describing clearance and predicting morphine doses in humans of all ages.