Repeated measurement of the novel atrial biomarkerBMP10improves risk stratification in anticoagulated patients with atrial fibrillation: Insights from the ARISTOTLE trial

Abstract Background Atrial fibrillation (AF) increases the risk of stroke, heart failure, and death. Biomarkers provide important prognostic information about these complications in AF. Recently, bone morphogenic protein 10 (BMP10), a protein mainly expressed in atrial tissue, was identified as a novel marker in AF for the risk of ischemic stroke. Purpose To assess the association between baseline characteristics with BMP10 concentrations, and if repeated BMP10 measurements at 2 months improve the prognostication of ischemic stroke or systemic embolism and other cardiovascular events in patients with AF. Methods BMP10 was measured in plasma samples collected at randomization and at 2 months in patients with AF who were randomized to apixaban or warfarin in the ARISTOTLE trial (n = 2878) with a median follow-up time of 1.8 years. BMP10 was analyzed with a prototype Elecsys immunoassay. The association of baseline variables with BMP10 at 2 months was evaluated with linear regression analysis. Cox-regression models adjusted for baseline characteristics and comorbidities including kidney function and BMP10-level at baseline were used in order to evaluate associations with outcomes. Results Median age was 70 (63.0-76.0) and 63% were male. The events during the follow up period were 34 ischemic strokes, 155 deaths and 99 hospitalizations for heart failure. Median concentration of BMP10 was 2.4 ng/mL at inclusion and 2.6 ng/mL at 2 months (median increase 7.8%, p<0.001). The Spearman correlation for BMP10 at baseline and 2 months was 0.71. The baseline variables most strongly associated with higher BMP10 levels at 2 months, in descending order of importance, were: higher BMI, female sex, worse kidney function, older age, AF-rhythm, diabetes, and randomized treatment. BMP10 levels at 2 months improved the risk stratification in models adjusted for clinical variables including baseline levels of BMP10. For ischemic stroke and death, the estimated associations were nonlinear (Figure) with increased risk with higher BMP10 levels (p=0.037 and p<0.001, respectively), whereas the association with HF hospitalization was linear with HR of 1.91 (p=0.012) comparing the third and first sample quartiles. Patients randomized to apixaban had approximately 2.2% lower BMP10 levels at 2 months compared to those randomized to warfarin (p<0.001). Conclusions In patients with AF, the level of the atrial biomarker BMP10 is consistently associated with the risk of stroke at baseline and after 2 months of treatment with warfarin or apixaban. Repeated compared to only one measurement of BMP10 further improves risk stratification. Apixaban as compared to warfarin is associated with a reduction in the BMP10 level, which is in accordance with the lower risk of stroke with apixaban as compared to warfarin.