Predicting stroke in Asian patients with atrial fibrillation using machine learning: A report from the KERALA-AF registry, with external validation in the APHRS-AF registry.

Atrial fibrillation (AF) is a significant risk factor for stroke. Based on the higher stroke associated with AF in the South Asian population, we constructed a one-year stroke prediction model using machine learning (ML) methods in KERALA-AF South Asian cohort. External validation was performed in the prospective APHRS-AF registry. We studied 2101 patients and 83 were to patients with stroke in KERALA-AF registry. The random forest showed the best predictive performance in the internal validation with receiver operator characteristic curve (AUC) and G-mean of 0.821 and 0.427, respectively. In the external validation, the light gradient boosting machine showed the best predictive performance with AUC and G-mean of 0.670 and 0.083, respectively. We report the first demonstration of ML's applicability in an Indian prospective cohort, although the more modest prediction on external validation in a separate multinational Asian registry suggests the need for ethnic-specific ML models.

Kerala Atrial Fibrillation Registry: a prospective observational study on clinical characteristics, treatment pattern and outcome of atrial fibrillation in Kerala, India, cohort profile.

PURPOSE: Limited published data exist on the clinical epidemiology of atrial fibrillation (AF) in South Asia including India. Most of the published data are from the Western countries and the Far East. The Kerala AF registry was initiated to collect systematic, prospective data on clinical characteristics, risk factors, treatment pattern and outcomes of consecutive AF patients who consulted cardiologists across the state of Kerala, India. PARTICIPANTS: All newly diagnosed and previously reported patients aged ≥18 years with documented evidence of AF on ECG were included. Patients with transient AF due to infection, acute myocardial infarction, alcohol intoxication, metabolic abnormalities and AF seen in postoperative cases and critically ill patients with life expectancy less than 30 days were excluded. FINDINGS TO DATE: A total of 3421 patients were recruited from 53 hospitals across Kerala from April 2016 to April 2017. There were 51% (n=1744) women. The median age of the cohort was 65 (IQR 56-74) years. Hypertension, diabetes mellitus and dyslipidaemia were present in 53.8%, 34.5% and 42.2% patients, respectively. Chronic kidney disease was observed in 46.6%, coronary artery disease in 34.8% and heart failure (HF) in 26.5% of patients. Mean CHA2DS2-VASc score of the cohort was 2.9, and HAS-BLED score was 1.7. Detailed information of antithrombotic and antiarrhythmic drugs was collected at baseline and on follow-up. During 1-year follow-up, 443 deaths (12.9%) occurred of which 332 (9.7%) were cardiac death and 63 (1.8%) were due to stroke. There were 578 (16.8%) hospitalisations mainly due to acute coronary syndrome, arrythmias and HF. FUTURE PLANS: Currently, this is the largest prospective study on AF patients from India, and the cohort will be followed for 5 years to observe the treatment patterns and clinical outcomes. The investigators encourage collaborations with national and international AF researchers. TRIAL REGISTRATION NUMBER: CTRI/2017/10/010097.