Abstract 16157: Single Measurement of GDF-15 Provides Reliable Prognostication of Stroke and Major Bleeding in Atrial Fibrillation
BACKGROUND: GDF-15 level was recently shown to predict stroke, mortality and bleeding in anticoagulated patients with atrial fibrillation (AF). However, the variability in GDF-15 levels over time and the robustness of the risk prediction in a single individual have not been investigated in this population. We investigated the value of repeated measurements of GDF-15 concerning prognostication of new events in patients with AF.
METHODS: GDF-15 levels were determined with the Elecsys GDF-15 pre-commercial assay (Roche Diagnostics) in plasma samples obtained in 4,548 pts both at randomization and at 2 months follow-up in pts with AF included in the ARISTOTLE trial. The reliability of the two measurements was investigated by intraclass correlation obtained from repeated measures ANOVA. The associations between GDF-15 levels at baseline, at 2 months and the changes between the two measurements and the occurrence of cardiovascular events after 2 months until the end of follow-up were evaluated by Cox models.
RESULTS: Median level at baseline was 1361 ng/L (Q1, Q3: 973, 2015) and increased by 2.8% reaching 1410 ng/L (1003, 2069) (p<0.0001) at 2 months. The median change was 1.4% (Q1, Q3: -12.2, 16.6). The intraclass correlation was 0.86 between the two measurements and the within subject coefficient of variance was 3%. GDF-15 levels at baseline and at 2 months showed similar significant (p<0.0001) relations to stroke, total mortality and major bleeding occurring during the follow-up period after the 2 months visit. The c-statistics at baseline and after adding the 2 months results to the Cox model were for stroke 0.652 and 0.660 (p=0.2490), death 0.688 and 0.699 (p<.0001) and major bleeding 0.627 and 0.633 (p=0.0654). Mortality increased by an increasing GDF-15 level at 2 months corresponding to a HR of 1.60 (95% CI 1.40-1.83) per 50% increase of GDF-15 (adjusted for the level at baseline).
CONCLUSIONS: A single measurement of the level of GDF-15 provides a reliable estimate of the risk of stroke, death and major bleeding in patients with AF during anticoagulant treatment. Repeated measurement after 2 months provides no significant incremental information concerning the risk of stroke and major bleeding but a somewhat better prognostic information on the risk of death.
Author Disclosures: The Disclosure Block has exceeded its maximum limit. Please call Tech support at (217) 398-1792 for more information.
- © 2014 by American Heart Association, Inc.