The association between novel ST2, BNP, atrial fibrillation and heart failure

Abstract

Abstract Background/Introduction ST2, a novel marker of fibrosis has been proposed as a novel biomarker for heart failure. There is paucity of data suggesting its association with Atrial Fibrillation (AF), in particular, with regards to the its independent effect after adjustment for clinical factors and traditional biomarkers such as BNP. There is increasing evidence that inflammation and fibrosis are important players in the pathogenesis of AF, the most common arrhythmia that is associated with considerable morbidity. Purpose We sought to examine the association of ST2 with AF and its possible incremental value in combination with traditional biomarkers already used in routine clinical practice such as BNP. Methods Unselected patients presenting to a Cardiology service who had BNP for clinical reasons, concomitantly had ST2 sent. These patients included those with heart failure, acute coronary syndromes and AF, as well as patients attending cardioversion and cardiac catheterisation. The association between ST2, BNP and clinical factors was examined. ST2 was examined in tertiles as its distribution was highly skewed. Raised BNP was defined as usual (≥100 pg/ml). Results Of the 619 patients, the mean age was 69 years and 66% were male. The prevalence of co-morbidities were: Coronary heart disease - CHD (41%), Atrial Fibrillation – AF (30%), Heart Failure - HF (20%), Chronic Kidney Disease – CKD (23%). The mean eGFR was 68 ml/h and the mean LV ejection fraction was 55%. Both ST2 and BNP levels were significantly higher in patients with heart failure, CKD and AF. Mean levels of ST2 were (in ng/ml): HF vs no HF (65 vs. 38, p=0.0001) and in CKD vs. no CKD (49 vs. 41, p=0.001), CAD vs no CAD (43 vs. 42, p=0.3). Mean levels of BNP were (in pg/ml): HF vs no HF (579 vs. 131, p=0.0001), CKD vs. no CKD (379 vs. 173, p=0.0001), CAD vs. no CAD (239 vs. 206, p=0.5). In patients with AF, levels of both ST2 and BNP were higher, as were their mean ages and prevalence of co-morbidities (see Table 1). Both ST2 and BNP correlated with AF (Pearson r=0.2, p<0.0001 for ST2 and r=0.5, p<0.0001 for BNP). There was also a graded association of ST2 with AF (15%, 33% and 40% across tertiles of ST2, p=0.0001). Both ST2 and raised BNP were significant associated with AF in both univariable and multi-variable regression models, and remained independent predictors when adjusted for HF, CHD, LV EF and eGFR (Table 2, Model 2 – Relative risk ratio of AF with ST2 across tertiles 1.5, p0.005, and Odds ratio of AF with raised BNP 7.4, p<0.001). However, ST2 loses significance when BNP is added to the model in combination (Table 2 – Model 3). Conclusions Both ST2 and BNP are independent predictors of AF even after adjustment for heart failure and other clinical factors. When ST2 is used in combination with BNP, raised BNP appears to be a stronger predictor of AF than ST2. Funding Acknowledgement Type of funding sources: None.