A framework for meta-analysis of veterinary drug pharmacokinetic data using mixed effect modeling.

Combining data from available studies is a useful approach to interpret the overwhelming amount of data generated in medical research from multiple studies. Paradoxically, in veterinary medicine, lack of data requires integrating available data to make meaningful population inferences. Nonlinear mixed-effects modeling is a useful tool to apply meta-analysis to diverse pharmacokinetic (PK) studies of veterinary drugs. This review provides a summary of the characteristics of PK data of veterinary drugs and how integration of these data may differ from human PK studies. The limits of meta-analysis include the sophistication of data mining, and generation of misleading results caused by biased or poor quality data. The overriding strength of meta-analysis applied to this field is that robust statistical analysis of the diverse sparse data sets inherent to veterinary medicine applications can be accomplished, thereby allowing population inferences to be made.