Pitfalls of immunoassay and sample for IGF-I: comparison of different assay methodologies using various fresh and stored serum samples.

Objective: Determination of insulin-like growth factor (IGF)-I is now a routine adjunct to multiple research and clinical investigations. Evidence has associated higher IGF-I levels with various human pathologies, but the reported associations have not been invariably confirmed. We examined the potential for post-sampling proteolysis and evaluated the impact of such events on IGF-I immunoassays.

Design And Methods: We compared IGF-I in different sets of fresh and frozen old samples using four different and commonly used immunoassays. The potential for post-sampling proteolysis was further examined by assaying fresh samples stored for 4 weeks at various temperatures in the absence or presence of protease inhibitors.

Results: IGF-I levels in fresh serum samples from adult males, females, and pregnant subjects by all methods were similar and were highly correlated (r=0.85-0.97). The same was true for levels in frozen ( approximately 2 years at --80 degrees C) samples from diabetic patients, which are reportedly associated with enhanced proteolytic activity. In contrast, in another set of frozen adult male and female samples ( approximately 8 years at --20 degrees C), the inter-method median IGF-I levels varied by approximately 3- to 4-fold and the values poorly correlated. Similar variability in the inter-method response was also observed when IGF-I in the replicates of fresh samples stored at 4 degrees C for 4 weeks was measured. However, the 4 degrees C storage effect could be completely blocked by the addition of protease inhibitors, allowing for all assays to detect 92--101% of the expected mean levels.

Conclusions: The data indicate susceptibility of IGF-I to significant post-sampling proteolysis and suggest the importance of immunoassays for the intact molecule. Immunoassays that lack specificity for intact IGF-I may mask the potential pathophysiological effects of proteolysis and generate misleading results, particularly in studies involving inappropriately stored and/or proteolyzed samples. In such cases, underestimation of the in vivo levels by the intact assays would occur, but the findings of low IGF-I levels may be indicative of questionable sample quality.