CHADS and modified CHADS-VASc scores for the prediction of congestive heart failure in patients with nonvalvular atrial fibrillation.

Background & Purpose: We have conducted a retrospective observational study to analyze the correlation between the CHADS score, the modified CHADS-VASc (mCHADS-VASc) score, and the incidence of all-cause death and congestive heart failure (CHF).

Methods: The study cohort consisted of 292 consecutive patients with nonvalvular atrial fibrillation (NVAF) admitted to our hospital from 2012 to 2014. Electronic medical records were used to confirm medical history including prior heart failure, hypertension, diabetes, stroke, and coronary disease. A follow-up survey for all-cause deaths and incidence of CHF was carried out from the baseline data to May 2015. We analyzed the correlation between each score and the endpoints using the Kaplan-Meier method and the Cox proportional hazards model.

Result: During the follow up period (mean=1.6 years), 69 all-cause deaths and 58 CHF events occurred in the cohort. There was no significant association between these scores and all-cause death in our CHF cohort. The incidence of CHF significantly increased along with increased CHADS (=0.018) or mCHADS-VASc scores (=0.044). The hazard ratio (HR) for CHF after adjustment for drug treatment was obtained from a Cox proportional hazards model. The HRs for the CHADS and mCHADS-VASc scores were 1.38 (95% CI; 1.13-1.68) and 1.35 (95% CI; 1.24-1.59), respectively.

Conclusion: Calculation of the CHADS and mCHADS-VASc scores in order to evaluate the risk of systemic thromboembolism was useful to predict the onset of CHF, but not all-cause death, in patients with NVAF.