Delta wave morphology correlates with the site of ventricular insertion of accessory AV pathways. Because lesions due to radiofrequency (RF) current are small and well defined, it may allow precise localization of accessory pathways. The purpose of this study was to use RF catheter ablation to develop an ECG algorithm to predict accessory pathway location. An algorithm was developed by correlating a resting 12-lead ECG with the successful RF ablation site in 135 consecutive patients with a single, anterogradely conducting accessory pathway (Retrospective phase). This algorithm was subsequently tested prospectively in 121 consecutive patients (Prospective phase). The ECG findings included the initial 20 msec of the delta wave in leads I, II, aVF, and V1 [classified as positive (+), negative (-), or isoelectric (+/-)] and the ratio of R and S wave amplitudes in leads III and V1 (classified as R > or = S or R < S). When tested prospectively, the ECG algorithm accurately localized the accessory pathway to 1 of 10 sites around the tricuspid and mitral annuli or at subepicardial locations within the venous system of the heart. Overall sensitivity was 90% and specificity was 99%. The algorithm was particularly useful in correctly localizing anteroseptal (sensitivity 75%, specificity 99%), and mid-septal (sensitivity 100%, specificity 98%) accessory pathways as well as pathways requiring ablation from within ventricular venous branches or anomalies of the coronary sinus (sensitivity 100%, specificity 100%). A simple ECG algorithm identifies accessory pathway ablation site in Wolff-Parkinson-White syndrome. A truly negative delta wave in lead II predicts ablation within the coronary venous system.