A linear mixed-effects statistical model for in-vivo evaluation of recombinant human growth hormone implants in hypophysectomized rats.

ABSTRACT

We have used a linear mixed-effects statistical model to evaluate previously published results of body weight evolution in hypophysectomized (Hpx) rats after administration of two different controlled-release formulations of recombinant human growth hormone (rhGH). Using the linear mixed-effects model, it was possible to distinguish between maximal pharmacological response with time in different subjects and relate it to the structure of the different formulations, the release of the hormone from them and the time necessary to obtain a quantitative result as a consequence of the hormone activity, contrary to the multivariate variance analysis model (MANOVA) used in our earlier work. These results confirmed that the maximum body weight gain obtained with the controlled-release implants is similar to that with subcutaneous rhGH, but with the advantage that laminar biodegradable implants need to be administered only once every 2 weeks.