Objectives Atrial fibrillation is the most common sustained cardiac rhythm disorder, which is associated with a substantial risk of stroke and mortality. Contemporary clinical risk stratification schemata for predicting stroke and thromboembolism in patients with atrial fibrillation are largely derived from western cohorts. The purpose of the present study is to assess the risk factors of stroke and MACE in a group of large-scale Chinese AF patients.
Methods We enrolled patients who present to an emergency department (ED) with atrial fibrillation or atrial flutter from November 2008 to October 2011, either as the primary or secondary diagnosis. A follow-up was performed to assess stroke and incidence of major adverse cardiac events during 1 year. The major adverse cardiac events (MACE) included all cause mortality, stroke, non-central nervous systemic embolism and major bleeds.
Results A total of 2016 AF patients (1104 women) were included in the final analysis. 33.7% of the subjects aged < 65 years, 28.3% aged 65-74 years and 38.0% aged over 75 years. Hypertension was the most prevalent comorbidity (55.5%), followed by coronary artery disease (41.8%). Cox regression analysis showed that the risk factors for ischaemic stroke were gender (HR1.470, 95% CI 1.048-2.063, p = 0.026), age over 75 years (HR2.717, 95% CI 1.690-4.367, p < 0.001), prior stroke/TIA (HR2.021, 95% CI 1.408-2.900, p < 0.001), left ventricular systolic dysfunction (LVSD) (HR1.701, 95% CI 1.024-2.827, p = 0.040), prior major bleeding (HR2.506, 95% CI 1.162-5.406, p = 0.019), untreated hypertension (HR1.948, 95% CI 1.101-3.446, p = 0.022). For MACE, age over 75 years (HR3.451, 95% CI 2.623-4.540, p < 0.001), heart failure (HR1.356, 95% CI 1.078-1.706, p = 0.009), prior stroke/TIA (HR1.531, 95% CI 1.221-1.918, p < 0.001), LVSD (HR1.444, 95% CI 1.101-1.893, p = 0.008), untreated hypertension (HR1.752, 95% CI 1.228-2.500, p = 0.002) were the independent predictors. The c-statistics for predicting stroke was 0.671 (95%CI: 0.625-0.716) and for MACE was 0.703 (0.675-0.730), respectively.
Conclusions The Cox regression prediction models have favorable predictive capability for estimating the risk in these Chinese AF patients. Besides the general risks, clinicians should pay more attention to patients with left ventricular systolic dysfunction and hypertension without medically treated.