Valoración del valor predictivo de eventos adversos de seis biomarcadores en pacientes con fibrilación auricular y anticoagulación oral estable.

Résumé

ABSTRACT Background: The use of biomarkers as predictors of thromboembolic, hemorrhagic and mortality risk, together with clinical factors in atrial fibrillation (AF), is a strategy that has been developing in recent years in an attempt to improve available prognostic scales. (Most relevant citations) Knowledge of their predictive capacity in a real-life population and with long follow-up may be of great interest for the evaluation of their real utility. Objectives: to evaluate the predictive value for cardiovascular events, stroke, bleeding and death of six biomarkers in real life patients and investigate if a multi-biomarker strategy could improve the current predictive performance of CHA2DS2-VASc and HAS-BLED scores in the same cohort of patients. Methods: A prospective study with patients under antivitamin K (AVK) therapy and an stable INR (2-3) the six months prior to inclusion. Determination of troponin T (hsTnT), interleukin 6 (hsIL6), von Willebrand factor (VWF), amino terminal end of the natriuretic peptide precursor protein type B (NT-proBNP), glomerular filtration rate (GFR) and time in therapeutic range (TTR) at baseline. All adverse events were recorded. After 6 years of follow-up, the independent predictive value of the six biomarkers against the CHA2DS2-VASc and HAS-BLED scores was determined, as well as the clinical value of the implementation of a multi-biomarker strategy (six biomarkers associated with each one of the scores ) versus the classical scores, measured by AUC (area under the curve), IDI (integrated discrimination improvement), NRI (net reclassification improvement) and DCA (decision curve analysis) curves. Results: 1363 patients (49% male, mean age 75 ± 8.5 years) were evaluated with 6.5 years of median follow up. VWF was an independent risk predictor for all events in the multivariate analysis (cardiovascular event: p = 0.001, stroke: p = 0.015, bleeding: p <0.001 and all-cause mortality: p <0.001). HsTnT is significant for cardiovascular risk (p = 0.017) and mortality (p <0.001). NT-proBNP (p = 0.026) and hsIL6 (p = 0.009) are only independent predictors of mortality. GFR is a predictor of bleeding risk (p = 0.001). The multi-biomarker strategy designed associated to the CHA2DS2-VASc score had a significant improvement in the predictive value measured by AUC, IDI with a negative reclassification by the NRI for all the events evaluated. The multi-marker strategy associated with the HAS-BLED score had a significant improvement in the predictive capacity assessed by AUC and IDI, but not by NRI. The DCA curves found lower net benefit compared with the original scores for both multi-biomarker strategies. Conclusions: The usefulness of the six biomarkers as predictors of cardiovascular, hemorrhagic and mortality events in patients with anticoagulated AF with acenocoumarol is reduced with the long-term follow-up of patients. Only maintain their predictive value: VWF and hsTnT for cardiovascular events, VWF for stroke, VWF and GFR for bleeding and VWF, hsTnT, hsIL6, NT-proBNP for mortality.The use of multi-biomarker strategies for cardiovascular, hemorrhagic and mortality risk in the same cohort of patients does not provide an improvement of clinical utility in the characterization of patients respect to originals strategies based on clinical parameters when a long-term follow-up of the patients is performed. Key words: Atrial Fibrillation. Biomarkers. Thromboembolic risk. Cardiovascular risk. Haemorrhagic risk. Mortality risk. Oral anticoagulation.