Growth differentiation factor 15 in a community-based sample: age-dependent reference limits and prognostic impact.

Background: Despite the growing body of evidence on growth differentiation factor 15 (GDF-15) reference values for patients with existing cardiovascular disease, limited investigation has been dedicated to characterizing the distribution and prognostic impact of GDF-15 in predominantly healthy populations. Furthermore, current cutoff values for GDF-15 fail to account for the well-documented age-dependence of circulating GDF-15. Methods: From 810 community-dwelling older adults, we selected a group of apparently healthy participants (n = 268). From this sample, circulating GDF-15 was modeled using the generalized additive models for location scale and shape (GAMLSS) to develop age-dependent centile values. Unadjusted and adjusted Cox proportional hazards models were used to assess the association between the derived GDF-15 reference values (expressed as centiles) and all-cause mortality. Results: Smoothed centile curves showed increasing GDF-15 with age in the apparently healthy participants. An approximately three-fold difference was observed between the 95th and 5th GDF-15 centiles across ages. In a median 8.0 years of follow-up, 97 all-cause deaths were observed in 806 participants with eligible values. In unadjusted Cox regression analyses, the hazard ratio (95% CI) for all-cause mortality per 25-unit increase in GDF-15 centile was 1.80 (1.48-2.20) and dichotomized at the 95th centile, >= 95th versus <95th, was 3.04 (1.99-4.65). Age-dependent GDF-15 centiles remained a significant predictor of all-cause mortality in all subsequent adjusted models. Conclusions: Age-dependent GDF-15 centile values developed from a population of apparently healthy older adults are independently predictive of all-cause mortality. Therefore, GDF-15 reference values could be a useful tool for risk-stratification in a clinical setting. ClinicalTrials.gov Identifier: NCT01452178.