Eighty-five consecutive patients (age 53 ± 10; 76% male) with paroxysmal AF arriving to the electrophysiology laboratory in sinus rhythm were included in the study. In 69 patients, AF either occurred spontaneously or was induced, whereas in 16 patients AF was not inducible according to the required criteria (Figure ). As shown in Table , the only independent predictors of AF inducibility were longer history of AF [odds ratio (OR): 1.25, 95% confidence interval (CI): 1.02–1.54; P= 0.03] and larger LA diameter (OR: 1.12, 95% CI: 1.02–1.23; P= 0.02).
Patient inclusion flowchart. AF, atrial fibrillation; PV-LAJ, the pulmonary vein to left atrial junctions; CS, coronary sinus; HRA, high right atrium.
In 12 patients, AF episodes lasted for <5 min and/or the catheters moved during the recording time, and these patients were not considered for spectral analysis. Overall, we analysed 68 AF episodes in 57 patients. In 41 patients, 45 (66%) AF episodes were induced by high-rate stimulation, 30 of them pacing from the PVs, 11 from the CS and 4 from the HRA. Twenty-three patients presented 23 (34%) spontaneous AF episodes while the catheters were in stable position. The origin of the spontaneous ectopic discharges initiating AF were the PVs in 9 cases (39%), the LA in 11 (48%), and the RA in 1 patient (4%); it could not be documented in 2 patients (8%). In seven patients, both spontaneous and induced episodes were recorded. Table shows the clinical characteristics of patients in whom spectral analyses were performed. Eight patients received isoproterenol during the procedure, but no differences were observed according to inducibility or AF spectral characteristics (data not shown).
Clinical characteristics of patients with spectral analysis (n= 57)
Atrial activation frequency in induced vs. spontaneous atrial fibrillation
The proportions of the total number of recordings excluded due to RI < 0.2 (low regularity of activity) were similar in patients with spontaneous (LA 34%, CS 45%, and HRA 25%) and induced (LA 30%, CS 33%, and HRA 30%) AF episodes for all locations (P= ns). Dominant frequencies were determined for the AF episodes with RI ≥ 0.2. Figure shows representative examples of 5 s recordings from the LA, CS, and HRA and corresponding power spectra from one spontaneous and one induced AF episode. On comparison, DF values in the three locations differ by <0.5 Hz. On average, our analysis demonstrates no significant differences in regional atrial activation frequencies between spontaneous and induced AF episodes at the HRA (5.18 ± 0.69 vs. 5.06 ± 0.91 Hz; P= 0.64), the CS (5.27 ± 0.69 vs. 5.36 ± 0.76 Hz; P= 0.69), and the LA (5.72 ± 0.88 vs. 5.64 ± 0.75 Hz; P= 0.7). Similarly, a separate analysis of AF induced by pacing at different sites showed also consistent regional activation frequencies: Compared with spontaneous AF episodes, regional DFs during AF induced from the PV (HRA, 5.11 ± 1 Hz; P= 0.8; CS, 5.48 ± 0.73 Hz; P= 0.37; and LA, 5.71 ± 0.71 Hz; P= 0.76) or AF induced from the CS/HRA (HRA, 4.99 ± 0.74 Hz; P= 0.46; CS, 5.16 ± 0.81 Hz; P= 0.69; and LA, 5.51 ± 0.84 Hz; P= 0.47) showed no differences. These results were not influenced by prior AF ablations (data not shown). As shown in Figure , a left-to-right DF gradient was observed in both spontaneous (PV-LAJ 5.71 ± 0.81 vs. HRA 5.18 ± 0.69 Hz; P= 0.005) and induced AF episodes (PV-LAJ 5.62 ± 0.72 vs. HRA 5.07 ± 0.91 Hz; P= 0.002), with no differences between them (P= not specific). Consistent with these results, paired analysis in a subset of seven patients presenting with both spontaneous and induced AF sustained episodes, showed no regional DFs differences, respectively, at the HRA (4.88 ± 0.75 vs. 4.9 ± 1.32 Hz; P= 0.92), the CS (4.95 ± 0.83 vs. 5.4 ± 0.99 Hz; P= 0.16), and the LA (5.49 ± 0.78 vs. 5.23 ± 0.79 Hz; P= 0.35).
Figure 2 Spectral analyses of 5 s recordings obtained from spontaneous (left panel) and induced (right panel) atrial fibrillation episodes in the same patient. Each panel shows a 5 s electrogram with its corresponding power spectrum. DF, dominant frequency; RI, (more ...)
Figure 3 Paired comparisons between the pulmonary vein-left atrial junctions and high right atrium dominant frequencies in spontaneous (A) vs. induced (B) atrial fibrillation episodes, demonstrating the presence of a left to right frequency gradients. DF, dominant (more ...)
Spatial distribution of activation frequencies
In a subset of 18 patients (8 spontaneous and 10 induced), real-time DF maps of the LA were obtained by acquiring a mean of 81 ± 25 points. There was no difference in the number of maximal DF (DFmax) sites (2.1±0.84 vs. 1.8±0.92; P= 0.45) or the DFmax values (7.46±1.37 vs. 7.93±1.42 Hz; P= 0.5) between spontaneous and induced AF maps, respectively. Maximal DF sites, up to three in both circuntancies, were preferentially located either at the PV antrum or a specific PV ostium, without significant differences (P= 0.88) in the percentage of DFmax sites according to atrial chamber location (Figure ). As shown in Figure , there were no significant differences in DFmax sites distribution at the LA in spontaneous vs. induced AF episodes.
Figure 4 Spatial distribution of maximal dominant frequency sites in spontaneous vs. induced atrial fibrillation. LSPV: left superior pulmonary vein; LIPV, left inferior pulmonary vein; RSPV, right superior pulmonary vein; RIPV, right inferior pulmonary vein; (more ...)
Figure 5 Representative examples of left atrium high-density dominant frequency maps (upper panels) together with the corresponding 1 s intracardiac electrograms time-series (lower panels) from paroxysmal atrial fibrillation patients with spontaneous (A) and pacing-induced (more ...)