Meta-analysis of pharmacokinetic data of veterinary drugs using the Food Animal Residue Avoidance Databank: oxytetracycline and procaine penicillin G

Investigators frequently face the quandary of how to interpret the oftentimes disparate pharmacokinetic parameter values reported in the literature. Combining of data from multiple studies (meta-analysis) is a useful tool in pharmacokinetics. Few studies have explored the use of meta-analysis for veterinary species. Even fewer studies have explored the potential strengths and weaknesses of the various methods of performing a meta-analysis. Therefore, in this study we performed a meta-analysis for oxytetracycline (OTC) and procaine penicillin G (PPG) given intramuscularly to cattle. The analysis included 28 individual data sets from 18 published papers for PPG (288 data points), and 41 individual data sets from 25 published papers for OTC (489 data points). Three methods were used to calculate the parameters. The first was a simple statistical analysis of the parameter values reported in each paper. The second method was a standard Two-Stage Method (TSM) using the mean concentration vs. time data extracted from each paper. The third method was the use of nonlinear mixed effect modeling (NMEM) of the concentration vs. time data reported in the various papers, treating the mean data as if each set came from an individual animal. The results of this evaluation indicate that all three methods generate comparable mean parameter estimates for OTC and PPG. The only significant difference noted was for OTC absorption half-lives taken from the published literature, a difference attributable to the use of an alternative method of parameter calculation. The NMEM procedure offers the possibility of including covariates such as dose, age, and weight. In this study the covariates did not influence the derived parameters. A combination approach to meta-analysis of published mean data is recommended, where the TSM is the first step, followed by the NMEM approach.