An empirical comparison of several commercial enzyme immunoassays for the non-invasive assessment of adrenocortical and gonadal function in mountain gorillas.

Wildlife researchers seeking to non-invasively examine endocrine function in their study species are presented with a dense and technical 'garden of forking paths' to navigate between collecting a biological sample and obtaining a final measurement. In particular, the choice of which enzyme immunoassay (EIA) to use with collected fecal samples, out of the many options offered by different manufacturers and research laboratories, may be one of the most consequential for final results. However, guidance for making this decision is still emerging. With this gap in mind, we performed a head-to-head comparison of results obtained from four different EIAs for fecal glucocorticoid metabolites (FGCMs), and three different EIAs for fecal androgen metabolites (FAMs), applied to the same set of fecal samples collected from the mountain gorillas (Gorilla beringei beringei) monitored by the Dian Fossey Gorilla Fund in Volcanoes National Park, Rwanda. We provide a) an analytical validation of the different EIAs via tests of parallelism and linearity; b) an estimate of inter-assay correlation between EIA kits designed for the same metabolites; and c) a test of the kits' ecological validity, in which we examine how well each captures endocrine changes following events that theory predicts should result in elevated FGCM and/or FAM concentrations. Our results show that kits differ to some degree in their performance; at the same time, nearly all assays exhibited at least moderate evidence of validity and covariance with others for the same analyte. Our findings, which differ somewhat from similar comparisons performed in other species, demonstrate the need to directly assess assay performance in a species- and context-specific manner as part of efforts to develop the burgeoning discipline of wildlife endocrinology.