Assessing estradiol in biobehavioral studies using saliva and blood spots:Simple radioimmunoassay protocols, reliability, and comparative validity

We developed simple, reliable, and highly sensitive assay modifications of commercially available radioimmunoassay kits to measure estradiol in saliva and blood spot specimens. The saliva assay has average intra- and interass ay coefficients of variation (CV) of 6.45 and 9.01%, with average analytica l and serial dilution recoveries 100.65 and 89.25%. The blood spot assay ha s average intra- and interassay CVs of 7.57 and 8.22%, with analytical and serial dilution recoveries of 80.50 and 108.50%. The analytical sensitivity ranges of the saliva (0.25-7.50 pg/ml) and blood spot (2.00-375 pg/ml) ass ays are sufficient to determine levels in the majority of pre- and postpube rtal males and females. Blood spot assay results are correlated with serum estradiol levels for adult males, r (17) = 0.73, and females, r (18) = 0.96 . In contrast, the serum-saliva correlation is only modest for adult female s, r(14) = 0.60, and not significant for adult males. Substitution of blood spot assay results for serum values underestimates the known serum estradi ol-behavior correlation by only 3.45%, whereas substitution of saliva assay results for serum values underestimates the association by 37.55%. The fin dings have important implications for the use and potential misuse of nonin vasive measures of estradiol in studies of health and human development. (C ) 2000 Academic Press.