Insulin-like growth factor I (IGF-I) and IGF-binding protein-3 in benign prostatic hyperplasia and prostate cancer

In view of evidence indicating significant involvement of the insulin-like growth factor (IGF) system in the pathogenesis of prostate cancer, we measu red serum IGF-I and IGF-binding protein-3 (IGFBP-3) in men with benign pros tatic hyperplasia (BPH; n = 75) or prostatic carcinoma (CaP; n = 84). The a ge-matched patient populations were selected to have circulating prostate-s pecific antigen (PSA), the most reliable predictor of CaP, in the overlappi ng diagnostic gray zone range of approximately 4-10 mug/L. Of particular in terest was investigation of intact, fragment, and total IGFBP-3 levels in r elation to PSA, which is also a well established IGFBP-3 protease. Among th e key findings were significantly higher IGF-I and intact IGFBP-3 levels in CaP vs. BPH (P < 0.001), whereas changes in fragment and total IGFBP-3 wer e statistically insignificant. As expected, total PSA levels were similar i n the two groups of patients (P = 0.173), whereas free PSA levels were sign ificantly lower in those with CaP (P < 0.001). IGF-I and IGFBP-3 (intact an d total) correlated significantly (P = 0.024 to <0.001) and inversely (r = -0.26 to -0.35) with free PSA in BPH, but not in CaP, and no correlations w ere found in comparisons involving total PSA. Statistical analysis of the v arious markers and their combinations indicated enhanced performance of IGF -I/free PSA [receiver operating characteristics area under the curve (AUC) = 0.728] and intact IGFBP-3/free PSA (AUC = 0.737) ratios in discriminating between BPH and CaP compared with the currently used free/total PSA ratio (AUC = 0.689). Multivariate logistic regression models confirmed the observ ed relationships and identified IGF-I/free PSA and intact IGFBP-3/free PSA as independent factors in predicting the presence of CaP. We conclude that increases in IGF-I and intact IGFBP-3 levels are positively associated with the presence of CaP in this group of patients with low to moderately eleva ted PSA, and that their measurements in relation to PSA may help improve di agnostic discrimination between BPH and prostate cancer.