Before predicting survival in children with pulmonary hypertension and congenital heart disease...

drug therapy, but not so after adjustment for left ventricular ejection fraction, left atrial diameter or duration of atrial fibrillation. We feel that our results justify the conclusion that “patients who undergo electrical cardioversion display a greater degree and longer duration of mechanical atrial dysfunction than those who convert pharmacologically or spontaneously” (1). We agree with Falk et al. that some of the patients who underwent electrical cardioversion may have had attempts at pharmacologic cardioversion in the past. By the same token, some of the patients who had pharmacologic cardioversion in our study had also undergone previous attempts at cardioversion. Previous history of cardioversion attempts was not specifically monitored in our study; hence, we are unable to comment on the actual number of study patients who had had previous attempts at cardioversion. To say that “the study is predominantly a comparison of patients who responded to pharmacologic agents compared with those in whom pharmacologic cardioversion failed” is therefore erroneous. To the best of our knowledge, there is only one previous study that showed that duration of atrial fibrillation ,14 days (vs. .6 weeks) is associated with faster recovery of atrial function (2). In that particular study, no multivariate analysis was performed to account for differences between groups of patients who had atrial fibrillation ,14 days versus .6 weeks. For this reason, we are not convinced that atrial fibrillation duration ,14 days should be a standard reference division point for all future studies with regard to atrial fibrillation. Until more definitive information is available on the impact of duration of atrial fibrillation on atrial function recovery, any dichotomization of this variable will have to be “arbitrary.” The division of atrial fibrillation duration in our study is identical to the inclusion criterion used by Falk et al. in their study addressing postcardioversion atrial function (3). Due to uncertainty about the exact duration of atrial fibrillation in many cases, it is not feasible to use duration of atrial fibrillation as a continuous variable.