BACKGROUND An increasing number of epidemiologic studies are investigating the relationship between serum levels of insulin-like growth factor-I (IGF-I) and IGF binding proteins (IGFBPs) and risk of cancer, cardiovascular disease, and other diseases. However, little is known regarding the effects of blood specimen processing time on measured levels of total and free IGF-I, and on IGFBP-3, the major binding protein. METHODS Two serum separation tubes were collected from each of 12 subjects. One tube was centrifuged as soon as possible following blood collection (a mean of 47 min; range=30-80 min), and serum aliquots were placed into -70 degrees C storage either shortly after centrifugation, or following 2, 4, 10, or 24 h at ambient temperature (measured from the time of blood draw). The second serum separation tube was maintained at ambient temperature for 24h before centrifugation and freezing. Total IGF-I, free IGF-I, and IGFBP-3 levels were determined using commercial enzyme linked immunosorbent assays (ELISAs) commonly employed in epidemiologic studies. The effects of time until centrifugation and freezing on seroassay results were evaluated using generalized estimating equation (GEE) linear regression models and Spearman correlation. RESULTS Total IGF-I and IGFBP-3 levels did not vary significantly with the amount of time at ambient temperature following centrifugation, even up to 24 h, in blood specimens that were centrifuged soon after collection (all pchi2). However, free IGF-I levels increased significantly with increasing time intervals between centrifugation and freezing in these same specimens (ptrend <0.001). Total IGF-I/IGFBP-3 molar ratio, a crude measure of free IGF-I levels, showed no clear association. In blood specimens that were not centrifuged for 24h, total IGF-I, free IGF-I, and IGFBP-3 were each significantly elevated (each pchi2) compared with results in blood specimens that were centrifuged and frozen soon after collection, whereas the total IGF-I/IGFBP-3 molar ratio was decreased pchi2. Nonetheless, all total IGF-I, free IGF-I, IGFBP-3, and total IGF-I/IGBFBP-3 molar ratio values altered by delays in processing were highly correlated with the values in specimens processed as soon as possible (all Spearman rank correlation coefficients 0.84). CONCLUSIONS Total IGF-I and IGFBP-3 can be fairly stably measured in serum with commonly used commercial assays regardless of the interval between blood collection and freezing, up to at least 24 h, as long as centrifugation and serum aliquoting take place shortly after blood collection. Free IGF-I levels, however, increase steadily with the time interval until freezing, even if serum separation has been completed soon after blood collection. Because the altered serum values are highly correlated with the referent values, analysis of total IGF-I, free IGF-I, IGFBP-3, and total IGF-I/IGFBP-3 molar ratio data by quartile might help mitigate concerns regarding the effects of delays in processing time.