14th International Congress
THE "NEW FRONTIERS"
OF ARRHYTHMIAS 2000

Jan. 29 - Feb. 5, 2000
Marilleva, Trento, Italy

RT-88

Impairment (remodeling?) of cardiac autonomic response after cardioversion of chronic atrial fibrillation

Antonio Michelucci, Chiara Lazzeri, Luigi Padeletti, Giuseppe Bagliani*, Andrea Colella, Alessandra Sabini, Renato Zipoli, Alessandro Costoli, Paolo Pieragnoli, Gianfranco Gensini, Franco Franchi.
Department of Internal Medicine and Cardiology, University of Florence, *Ospedale di Foligno, Italy

Methods

The study protocol was performed in 30 consecutive patients with atrial fibrillation (mean age 61±2 yrs, 19 men) lasting 1.8±0.2 months. All of them underwent successful electrical cardioversion which was transthoracic in 18 and internal in 12. Valvulopathies were present in eight patients, ischemic heart disease in seven and essential hypertension in fifteen. All patients were under anticoagulant therapy (warfarin) combines with amiodarone in 18, verapamil in 8 and propaphenon in 4.

After successful electrical cardioversion, a 24-hour Holter recording was performed in order to evaluate heart rate variability (HRV). Data were compared to a group of 30 sex and age-matched healthy subjects as controls.

ECG Holter recording of bipolar leads (CM1 and CM5) was scanned by a computer-based system (ELATEC 3.0, ELA medical, Segrate, Italy) with correction of beat morphology and timing by one of the authors, who did not know the patients’ background. HRV was evaluated in the time and frequency domains using a software provided by ELA medical (HRV module for ELATEC 1.0, ELA Medical, Segrate, Italy). From the surface ECG, the computer program calculated a series of 512 consecutive intervals as a function of beat numbers, thus obtaining the tachogram. The power spectra, that is the energy in the power spectrum between 0.01 and 0.40 Hz, were computed over a 256-second sampling period, with an overlapping of 128 seconds, using the Fast Fourier Transform mathematical function. In the time domain, we considered the 24-hour standard deviation of all RR intervals (24h-SDNN), an index of the overall variability, and the 24-hour square root of the mean of the sum of the squares of differences between adjacent normal RR intervals (24h-RMSSD), which is a measure of high-frequency, vagally mediated HRV². The following 24-hour frequency-domain indexes were determined: 1) the Low-Frequency (24h-LF) component, that is the value of the power (msec²) in the band from 0.04 to 0.15 Hz; 2) the High-Frequency (HF) component, that is the value of the power (msec²) in the band from 0.16 to 0.40 Hz; 3) the 24-hour (24h-) LF/HF ratio, an index of the sympatho-vagal interplay². The LF and HF bands were expressed in normalized units (nu). The use of the normalization procedure allows a better comparison among spectra with large differences in total variance². Moreover some studies suggest that, when expressed in normalized units, the LF component is a quantitative marker of sympathetic modulation². The LF/HF ratio was also calculated in a daytime (8.00-12.00 a.m.) and in a nighttime (0-4 a.m.) period.

Statistical analysis. Data are expressed by mean ± standard error (ES). Comparison between healthy subjects and patients after electrical cardioversion was performed using the t test.