Prediction of Transition to Chronic Atrial Fibrillation in Patients With Paroxysmal Atrial Fibrillation
To the Editor:
We read with great interest the recent report of Abe et al1 demonstrating that P-wave–triggered signal-averaged electrocardiography may be useful to predict the transition to chronic atrial fibrillation in patients with paroxysmal atrial fibrillation. In their study, although sex, age, and presence of organic heart disease were not associated with an increased risk for the transition to chronic atrial fibrillation, patients with the abnormality of a P-wave–triggered signal-averaged ECG had an 11-fold greater risk for chronic atrial fibrillation than those without the abnormality. Echocardiographically, left atrial dimension in patients with chronic atrial fibrillation was larger than that in patients with paroxysmal atrial fibrillation (40.5 versus 36.4 mm, P<0.05), but there were no significant differences in left ventricular dimensions measured at end diastole and end systole or in ejection fraction between the 2 groups.
In contrast to the findings by Abe et al, we2 previously reported that congestive heart failure and reduction in left ventricular ejection fraction were predictors of the transition to chronic atrial fibrillation in patients with new-onset atrial fibrillation. To identify predictors of the transition to chronic atrial fibrillation within the first year after onset, we retrospectively reviewed clinical records, standard 12-lead ECGs, and M-mode echocardiograms of 137 patients with new-onset, nonrheumatic atrial fibrillation. One year after onset, 30 (22%) of 137 patients showed a transition to chronic atrial fibrillation, and the other 107 continued to have paroxysmal atrial fibrillation. Compared with patients with paroxysmal atrial fibrillation, patients with chronic atrial fibrillation were older at the time of onset (70.1 versus 62.4 years, P<0.01), more frequently had diabetes mellitus (37% versus 19%, P<0.05), and more frequently had congestive heart failure (13% versus 3%, P<0.05). These patients also had higher cardiothoracic ratios on chest x-ray (52.0% versus 47.0%, P<0.01), greater f-wave amplitude in lead V1 on ECG (1.48 versus 1.06 mm, P<0.05), larger left atrial dimension measured by echocardiography (41.0 versus 34.2 mm, P<0.01), larger left ventricular end-systolic dimension (32.9 versus 29.7 mm, P<0.05), and lower ejection fraction (0.71 versus 0.76, P<0.05). Furthermore, the presence of any 1 of the following 7 factors was associated with an increased risk for the transition to chronic atrial fibrillation: age >65 years (32% versus 11%, P<0.01), diabetes mellitus (35% versus 18%, P<0.05), congestive heart failure (57% versus 20%, P<0.05), cardiothoracic ratio >50% (41% versus 11%, P<0.01), f-wave amplitude in lead V1 >2.0 mm (80% versus 20%, P<0.01), left atrial dimension >38 mm (34% versus 5%, P<0.01), and ejection fraction <0.76 (35% versus 4%, P<0.01). When each of these 7 significant predictors was assigned 1 point in risk score, the transition to chronic atrial fibrillation occurred in >88% of the patients with a risk score >4 (a high-risk group), in 22% of the patients with a risk score of 3 (an intermediate-risk group), and in <6% of the patients with a risk score <2 (a low-risk group).
The discrepancy between these 2 studies may be due to patient characteristics. Patients in the study by Abe et al were outpatients who had maintained cardiac function relatively well, whereas our study patients were recruited in part from an inpatient population with more congestive heart failure or moderate to severe cardiac dysfunction. This suggests that P-wave–triggered signal-averaged electrocardiography is useful to predict the transition to chronic atrial fibrillation, especially in patients with normal cardiac function. On the other hand, in the population including patients with impaired cardiac function, we propose that congestive heart failure and reduced left ventricular ejection fraction are important predictors of the transition to chronic atrial fibrillation. Furthermore, the major advantage of our proposed predictors and risk scoring system is that they allow physicians to identify patients with new-onset atrial fibrillation at high risk for the transition to chronic atrial fibrillation using routinely available clinical parameters.
- Copyright © 1998 by American Heart Association
We appreciate the interest shown by Dr Sakamoto and his colleagues in our articleR1 and their valuable comments using their results. They raised some important issues on prediction of transition to chronic atrial fibrillation. They proposed their risk-scoring system using 7 predictors from clinical parameters, including indexes of heart failure, which proved to be significant in their retrospective study.R2
As they pointed out, the discrepancy between their and our results may be derived mainly from the difference of study population; study design (prospective or retrospective) may be also involved. Unfortunately, we have not yet conducted studies on the usefulness of P-wave signal–averaged electrocardiography for prediction of transition in patients with reduced cardiac function. However, we recently found that filtered P-wave duration prolongation can be a direct predictor of hospitalization for worsening heart failure in a prospective heart failure studyR3 , in which Kaplan-Meier analysis revealed that patients with heart failure (ejection fraction <40%) who had an abnormally prolonged filtered P wave (≥145 ms) were more often (29% versus 4%, log-rank test P<0.05) hospitalized for worsening heart failure than those without it during the follow-up period of 1 to 37 months. In addition, atrial fibrillation was often observed on admission. This implies that filtered P-wave duration may be an early predictor of heart failure deterioration as well as the establishment of atrial fibrillation. It is also well known that atrial fibrillation itself causes a deterioration in cardiac function, probably due to the disappearance of atrial contraction and shortening of the diastolic phase, although heart failure itself enhances the accomplishment of atrial fibrillation. Consequently, the mechanism for establishment of atrial fibrillation may be complicated in heart failure patients, in whom changes in autonomic nerve system, electrolytes, humoral factors, and hemodynamics often occur. In general, most clinical parameters are redundant and depend in part on each other. That is why the analysis had to be complex, especially in a prospective study. Subsequently, to make it simpler, we investigated only filtered P-wave characteristics, such as electrophysiological arrhythmogenic substrate, in patients without heart failure. Although the left atrial dimension and the number of atrial premature contractions a day were weakly but significantly different between the 2 groups with and without the transition in our study, we think that this might be because they were closely related to the filtered P-wave duration.
Many people would like to know what type of paroxysmal atrial fibrillation will eventually change to the chronic form because the prognosis and incidence of thromboembolism are different. Thus, we would like to propose again that P wave signal–averaged electrocardiography be used in patients without heart failure. If additional criteria in P-wave signal–averaged electrocardiography for patients with heart failure are proposed, our method will be more applicable in clinics.
Abe Y, Fukunami M, Yamada T, Ohmori M, Shimonagata T, Kumagai K, Kim J, Sanada S, Hori M, Hoki N. Prediction of transition to chronic atrial fibrillation in patients with paroxysmal atrial fibrillation by signal-averaged electrocardiography: a prospective study. Circulation. 1997;96:2612–2616.
Sakamoto H, Okamoto E, Imataka K, Ieki K, Fujii J. Prediction of early development of chronic nonrheumatic atrial fibrillation. Jpn Heart J. 1995;36:191–199.
Asano Y, Fukunami M, Shimonagata T, Kumagai K, Yamada T, Sanada S, Ogita H, Hoki N. Signal-averaged P-wave duration: as a predictor for the deterioration of heart failure. J Cardiac Failure. In press.