CHADS2

Used to estimate baseline stroke risk in nonvalvular atrial fibrillation to guide need for anticoagulation. Best applied when rapid, simple stratification is required based on 5 factors (CHF, hypertension, age ≥75, diabetes, prior stroke/TIA). Predicts stroke with moderate discrimination (c≈0.68–0.70) 1 [Level I evidence]; higher scores correlate with rising annual stroke rates and increasing anticoagulation benefit 2 [Level I evidence].

Start at its limitations: only moderate discrimination (c≈0.66–0.68) and wide variability in stroke rates across cohorts, risking misclassification 1 [Level I evidence]. It omits vascular disease, age 65–74, and sex, leading to underestimation in lower‑risk groups and poor performance in CKD or diverse populations 2 [Level II evidence]. Updated tools (CHA₂DS₂‑VASc, ATRIA) consistently outperform it and are preferred.