Prediction of atrial fibrillation in patients with cardiac dysfunctions†

Abstract

Atrial fibrillation (AF) is a common arrhythmia in advanced heart failure. The occurrence of AF increases the risk of death and hospitalization for patients with heart failure. The results of different studies indicated that patients with paroxysmal AF have a longer filtered P wave duration (FPD), a lower root mean square voltage of the last 20 ms of the P wave (RMS 20), and a lower chemoreflexsensitivity (CHRS). Our study bases on these observations in order to examine the methods for predicting AF in patients with a left ventricular ejection fraction below 40% without a prior documentation of AF. The ratio between the difference of RR intervals in ECG and venous pO(2) before and after 5-min oxygen inhalation was measured (ms/mmHg) in order to determine the CHRS. A P wave signal-averaged ECG was performed for the measurement of FPD and RMS 20. The measurements were only performed in 94 patients with sinus rhythm. AF occurred during the mean follow-up of 39.9 months in 24 patients (26%). There were no significant differences concerning age, heart diseases, sex, ejection fraction, heart rate, or the use of drugs. The FPD (130.3 +/- 4.2 vs. 118.9 +/- 12.4 ms, P < 0.0001) was significantly longer and the RMS 20 (3.03 +/- 0.95 vs. 3.83 +/- 1.58 microV, P = 0.02) was significantly lower in patients with AF than in sinus rhythm. The CHRS did not differ significantly between both groups (3.57 +/- 1.49 vs. 3.48 +/- 1.62 ms/mmHg, P = 0.81). The chi(2) test showed that the threshold of FPD>or=125 ms and RMS 20 <or=3.3 microV revealed the best predictive value for AF. A stepwise logistic regression analysis of all variables identified the threshold of FPD>or=125 ms and RMS 20 <or=3.3 microV (OR 18.71; 95% CI, 4.85-72.16, P < 0.0001) as independent predictors for AF. In summary, our data show that the results of a P wave signal-averaged ECG can predict the risk for new onset of AF in patients with heart failure. The value of signal-averaged FPD is probably the result of reflecting the intra-atrial conduction delay, which is a pathophysiological condition for AF. The CHRS is not a suitable method for predicting AF.