External validation of the Ramathibodi paroxysmal supraventricular tachycardia recurrence score and development of a refined prediction model in a Hong Kong emergency department adult cohort: A retrospective study

Abstract Background Paroxysmal supraventricular tachycardia (PSVT) recurrence after adenosine treatment is common in emergency departments (EDs), but predictive tools require validation. Objectives To externally validate the Ramathibodi PSVT recurrence score in a Hong Kong ED cohort and develop a refined local model for predicting 24‐h recurrence. Methods This retrospective cohort study included adults (≥18 years) with PSVT treated with ≤3 adenosine doses at Tseung Kwan O Hospital ED (July 2020–June 2025). Data were extracted from electronic records. The original score's performance was assessed via area under the receiver operating characteristic curve (AUC). Univariable and multivariable logistic regression identified predictors for a new score. Results Of 278 patients, 42 (15.1%) had recurrence. The original score had AUC 0.782 (95% CI 0.706–0.858), with good calibration ( p = 0.539). A new score (postconversion heart rate >100 bpm: 1 point; ≥2 adenosine doses: 2 points; hypertension: 1 point; and ischemic heart disease: 1 point; total 0–5) achieved AUC 0.827 (95% CI 0.764–0.891). Observed recurrence rates aligned with predictions (e.g., score 0: 1.1% and score 5: 80.0%). Conclusions The Ramathibodi score performed moderately in this cohort. The refined score offers improved discrimination, aiding ED risk stratification.