Objectives. This study sought to determine whether the clinical and electrophysiologic criteria developed in adults also identify children with Wolff-Parkinson-White syndrome at risk for sudden death. Background. In adults with Wolff-Parkinson-White syndrome, a shortest RR interval < 220 ms during atrial fibrillation is a sensitive marker for sudden death. However, because reliance on the shortest RR interval has a low positive predictive value, the clinical history has assumed a pivotal role in assessing risk. This approach has not been evaluated in children. Methods. We retrospectively evaluated 60 children ≤ 18 years old who underwent comprehensive electrophysiologic evaluation between 1979 and 1989 before undergoing operation for Wolff-Parkinson-White syndrome. Clinical and electrophysiologic data were analyzed after patients had been grouped by their clinical presentation: high risk (cardiac arrest), intermediate risk (syncope or atrial fibrillation) or low risk (orthodromic reciprocating tachycardia alone). Results. Ten children had a clinical cardiac arrest (high risk); only one had a prior history of syncope or atrial fibrillation. Compared with the intermediate (n = 19) and low risk groups (n = 31), there were no differences in age ([mean ± SD] 14.8 ± 0.6 vs. 14.7 ± 0.6 vs. 14.5 ± 1.7 years), duration of symptoms (1.9 ± 0.5 vs. 4.1 ± 1.1 vs. 5.2 ± 0.8 years), incidence of congenital heart disease (30% vs. 26% vs. 32%), presence of multiple pathways (20% vs. 16% vs. 16%) or accessory pathway location. A shortest pre-excited RR interval < 220 ms was found in 7 of 7 high risk patients (sensitivity 100%), 14 of 19 intermediate risk patients and 11 of 31 low risk patients (prevalence 35%). Conclusions. Cardiac arrest was the only distinguishing clinical feature between high and low risk groups and the first manifestation in 80% of the children of an accessory pathway that can precipitate a life-threatening arrhythmia. In this series, the largest reported to date of children with Wolff-Parkinson-White syndrome having a cardiac arrest, a shortest pre-excited RR interval < 220 ms was more sensitive than clinical history for identifying those at risk for sudden death.