Aims: To examine if a simple biomarker can identify people with diabetes who are at high risk of atrial fibrillation. Methods: A retrospective cohort study was conducted at a single centre in people with Type 2 diabetes referred to our department between January 2000 and December 2007. In 517 consecutive people without any history, signs or symptoms of atrial fibrillation at baseline, the association between baseline B-type natriuretic peptide level and future atrial fibrillation incidence was examined, with adjustments for other potentially confounding factors. Results: A total of 28 people were diagnosed with new-onset atrial fibrillation during a median 6-year follow-up. When people were categorized into three groups according to B-type natriuretic peptide clinical thresholds (20 and 100 pg/ml), hazard ratios for the development of atrial fibrillation in the middle and highest B-type natriuretic peptide groups were 2.8 and 9.4, respectively, compared with the lowest B-type natriuretic peptide group. Time-dependent receiver-operating curve analysis identified a threshold for B-type natriuretic peptide to detect atrial fibrillation development of 52.8 pg/ml (sensitivity 75.2%, specificity 68.8%). The B-type natriuretic peptide predictive value was independent of and similar to that of left atrial size and ventricular dimension. Conclusion: In people with Type 2 diabetes, high baseline B-type natriuretic peptide levels were significantly associated with future atrial fibrillation development.
ASJC Scopus subject areas
- Internal Medicine
- Endocrinology, Diabetes and Metabolism