Improved prediction of all-cause mortality by a combination of serum total testosterone and insulin-like growth factor i in adult men

Objective: Lower levels of anabolic hormones in older age are well documented. Several studies suggested that low insulin-like growth factor I (IGF-I) or testosterone levels were related to increased mortality. The aim of the present study was to investigate the combined influence of low IGF-I and low testosterone on all-cause mortality in men. Methods and results: From two German prospective cohort studies, the DETECT study and SHIP, 3942 men were available for analyses. During 21,838 person-years of follow-up, 8.4% (n = 330) of men died. Cox model analyses with age as timescale and adjusted for potential confounders revealed that men with levels below the 10th percentile of at least one hormone [hazard ratio (HR) 1.38 (95% confidence-interval (CI) 1.06-1.78), p = 0.02] and two hormones [HR 2.88 (95% CI 1.32-6.29), p < 0.01] showed a higher risk of all-cause mortality compared to men with non-low hormones. The associations became non-significant by using the 20th percentile as cut-off showing that the specificity increased with lower cut-offs for decreased hormone levels. The inclusion of both IGF-I and total testosterone in a mortality prediction model with common risk factors resulted in a significant integrated discrimination improvement of 0.5% (95% CI 0.3-0.7%, p = 0.03). Conclusions: Our results prove that multiple anabolic deficiencies have a higher impact on mortality than a single anabolic deficiency and suggest that assessment of more than one anabolic hormone as a biomarker improve the prediction of all-cause mortality.