Direct or Coincidental Elimination of Stable Rotors or Focal Sources May Explain Successful Atrial Fibrillation Ablation

On-Treatment Analysis of the CONFIRM Trial (Conventional Ablation for AF With or Without Focal Impulse and Rotor Modulation)

Study design and enrollment

CONFIRM (15) prospectively enrolled 92 subjects at 107 consecutive AF ablation procedures for standard-of-care indications. Subjects were ≥21 years of age with AF despite receiving one or more Class I or III anti-arrhythmic drugs. To study AF sources and their response to ablation across a broad spectrum of AF presentations, we included patients with paroxysmal, persistent, and longstanding persistent AF (8) and AF despite prior conventional ablation. The only exclusion criterion was an inability or refusal to provide specific written informed consent.

Consecutive cases were prospectively mapped and assigned in a 2:1 ratio to FIRM-blinded (conventional) ablation alone or, after real-time FIRM mapping had been developed, to FIRM-guided ablation comprising targeted source ablation (FIRM) followed by conventional ablation.

Electrophysiology study

Electrophysiology study was performed after patients discontinued antiarrhythmic medications for 5 half-lives (for amiodarone: >60 days [median: 230 days]) (Table 1). Intravenous heparin was infused to maintain an activated clotting time >350 s. A 64-pole basket catheter (48- or 60-mm diameter; 4 to 6-mm electrode spacing; Constellation, Boston Scientific, Natick, Massachusetts) was advanced trans-septally to map the left atrium in all patients and to map the right atrium in 73 patients (including all FIRM-guided cases).

Figure 1A shows a patient in whom simultaneous biatrial baskets were used (recent studies have used one basket for both atria [16,17]). AF was observed in 101 cases (including FIRM-guided cases) (15). When required, AF was induced (rapid pacing: n = 26; isoproterenol: n = 2) because several studies have shown similar frequency (18) and spatial activation (19) for induced versus spontaneous AF. Unipolar and bipolar AF electrograms were filtered at 0.05 to 500 Hz and exported at 1-kHz sampling frequency.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

FIRM of AF Sources

(A) Fluoroscopy of bi-atrial 128 electrode wide-field of view mapping in atrial fibrillation (AF) (right atrial [RA] oblique projection). An implantable continuous electrocardiography recorder, diagnostic catheter in the coronary sinus, ablation catheter in the left atrium and esophageal temperature probe are also seen. (B) Focal impulse and rotor mapping (FIRM) maps show two independent sources: an AF rotor in an unconventional lateral RA location and a concurrent left atrial focal source. Orientation right atrium: Top, superior vena cava; Left, lateral tricuspid. Left atrium: Top, superior mitral annulus. (C) AF rotor in lateral RA on patient-specific electroanatomic shell (same patient as in B).

FIRM mapping of AF sources

FIRM mapping for AF has been described previously (15,19). Briefly, AF electrograms were recorded in a wide field of view and analyzed using RhythmView (Topera, San Diego, California) to produce maps of AF propagation. Figure 1B shows a FIRM map of an AF rotor (i.e., red-to-blue, early-meets-late spiral wave). As shown in Figure 1C, this rotor lay at a lateral right atrial site that would not be ablated conventionally.

AF propagation (FIRM) maps were analyzed intraprocedurally in FIRM-guided patients to guide ablation, and post-procedurally in FIRM-blinded patients. Electrical rotors (Figs. 1B and 2) were defined as sustained clockwise or counterclockwise activation around a center of rotation, and focal impulses (Fig. 2) were defined by centrifugal activation from an origin. Rotors and focal impulses were considered AF sources only if stable in repeated samples over 30 to 120 min (i.e., thousands of cycles), distinct from transient fibrillatory activity (20,21). The AF focal source origin or rotor center of rotation was located by its electrode coordinates and “shadowed” digitally within each patient-specific anatomic shell (NavX, St. Jude Medical, Inc., St. Paul, Minnesota) at the time of each FIRM map to eliminate errors from subsequent possible movement. Each digital shell was created, in turn, using careful sampling of atrial and venous points with reference to (or fused with) pre-procedural patient-specific computed tomographic scans when available.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

AF Sources Remote From Conventional Ablation Sites

(A) AF rotor (clockwise, 3.2 cm²) in inferior left atrium in an 85-year-old with paroxysmal AF. (B) AF focal source (2.0 cm²) in mid-posterior left atrium in a 64-year-old with PAF. (C) AF focal source (1.8 cm²) in anterior left atrium (anterior to typical roof line) in a 48-year-old with persistent AF. All cases received FIRM-guided ablation, and are AF-free on implanted cardiac monitoring at >1 year. Orientations for FIRM maps as in Figure 1B. Abbreviations as in Figure 1.

Ablation approach

Radiofrequency energy was delivered with a 3.5-mm tip irrigated catheter (Thermocool, Biosense-Webster, Diamond Bar, California) at 25 to 35 W or, in heart failure subjects, an 8-mm tip catheter (Blazer, Boston Scientific) at 40 to 50 W, target 52°C. Each lesion was registered digitally on each patient's digital NavX anatomy. Ablation in FIRM-guided subjects always commenced with FIRM. Energy was applied at adjacent sites for 15 to 30 s to cover the AF source (rotational center or focal source origin) (15). The endpoint was AF termination, coverage of the typical ≈2-cm² source area (≈5-min ablation; <2% to 5% of atrial area [22]) or 10 min of energy application, whichever came first. If AF terminated, vigorous attempts were made to reinitiate AF. FIRM was repeated for no more than three sources (≤30 min permitted [15]), followed by conventional ablation.

Conventional ablation (8), performed after FIRM in FIRM-guided patients and as sole therapy in FIRM-blinded patients, comprised wide-area circumferential ablation (WACA) to isolate left and right PV pairs, with verification of PV entrance block using a circular mapping catheter. In persistent AF, a left atrial roof line with confirmation of block was also performed. Cardioversion was performed if necessary.

On-treatment analysis: did ablation pass through sources?

Electroanatomic and FIRM maps were analyzed per patient, blinded to demographics, outcomes, and assignment to FIRM-guided or conventional ablation limbs. Source ablation was assigned if 5-mm lesion markers on NavX (Figs. 1 to 4⇓⇓) lay within ±50% of the interelectrode spacing (on the digital shell) to electrodes recording the rotor core or focal origin. Two analyses were performed. First, any source ablation was assigned if ≥1 source was ablated, and no source ablation was assigned otherwise. Conventional ablation was left atrial, so that this analysis could be performed in all patients (FIRM-blinded and -guided) who all received left atrial FIRM maps. Second, we measured FIRM dose response by assigning all source ablation (e.g., 3 of 3 concurrent sources), some source ablation (e.g., 2 of 3 sources ablated), and no source ablation. Because of potential right atrial AF sources, this secondary analysis was performed only in patients with biatrial FIRM maps. Assignments were performed independently by 4 of the authors (S.M.N., D.E.K., K.S., and J.M.M.), and disputes were resolved using consensus.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

AF Focal Impulse and Rotor Locations in the CONFIRM Trial

Locations as a percentage of all sources for each atrial fibrillation (AF) type: (A) paroxysmal and (B) persistent. Focal impulses indicated by stars; rotors, by circles. CONFIRM = Conventional Ablation for AF With or Without Focal Impulse and Rotor Modulation; IVC = inferior vena cavae; LAA = left atrial appendage; LIPV = left inferior pulmonary vein; LSPV = left superior pulmonary vein; RAA = right atrial appendages; RIPV = right inferior pulmonary vein; RSPV = right superior pulmonary vein; SVC = inferior, superior venae cavae.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

AF Rotors Near Conventional Ablation Targets

(A) Left atrial AF rotor successfully ablated by FIRM, that was then incorporated into left wide-area circumferential PV ablation. (B) Two left atrial AF rotors, each successfully ablated by FIRM. The superior rotor was incorporated into the left wide-area circumferential pulmonary vein ablation, while the inferior rotor (isochrones shown) was nearby. (C) Left atrial roof AF rotor ablated coincidentally by empirical roof ablation in a man with persistent AF. AF terminated at the point indicated by a red dot. Offline FIRM analysis later revealed an AF rotor at the precise point of AF termination. All patients are AF-free. FIRM atrial orientations are as in Figure 1B. Abbreviations as in Figure 1.

Post-procedure clinical management

Follow-up for arrhythmia recurrence met or exceeded guidelines (23). In the first 3 months post-ablation, antiarrhythmic medications were continued and cardioversion performed if indicated, but neither repeat ablation nor cross-overs were permitted. Subjects were then evaluated quarterly for up to 24 months. Arrhythmia recurrence was detected using continuous implanted electrocardiographic (ECG) monitors when consent was obtained (Reveal XT, Medtronic, Minneapolis, Minnesota) after U.S. approval in 2009 (Fig. 1A), or clinical pacemaker/defibrillator. Remaining subjects received external Holter or event monitors each quarter and also at the time of any symptoms.

Study endpoints

The primary long-term efficacy endpoint was freedom from AF after the FIRM-mapping procedure, defined as <1% burden on continuous implanted ECG monitors (actual burden: 0.1 ± 0.2%, as presented in CONFIRM [15]), or AF <30 s on intermittent monitors (8). Secondary efficacy measures included freedom from all atrial arrhythmias, and outcomes in patients at first ablation. Four patients in the CONFIRM trial were lost to follow-up (15).

Statistical analysis

Continuous data are represented as mean ± SD or median (interquartile range [IQR]) as appropriate. Normality was evaluated using the Kolmogorov-Smirnov test. Comparisons between the two groups were made with the Student t test and summarized with mean ± SD for independent samples if normally distributed or, if not normally distributed, evaluated using the Mann-Whitney U test and summarized with median (IQR). Nominal values are expressed as n (%) and compared using the Fisher exact test. Associations between continuous variables were evaluated using Spearman correlation. Interobserver agreement in assigning source ablation (S.M.N., D.E.K., K.S., and J.M.M.) was measured using the kappa score. Long-term outcome was assessed and reported after the single FIRM-map procedure, raw event rates were compared using the chi-square test, and event-free survival plots were made using the Kaplan-Meier method and compared using the log rank test. A probability of <0.05 was considered statistically significant throughout. Statistical analyses were completed using SPSS version 19 (Systat Software, Inc., Chicago, Illinois).

Characteristics of AF rotors and focal sources

FIRM mapping revealed 2.3 ± 1.1 concurrent AF sources per patient (median: 2; IQR: 1 to 3) in diverse biatrial locations.

Figures 1B to 2C show stable AF sources at atrial sites not conventionally ablated. After FIRM-guided ablation, these patients remain AF-free at 15, 15, 15, and 14 months, respectively. Figure 3 summarizes source locations in the CONFIRM trial in patients with paroxysmal AF (Fig. 3A) and persistent AF (Fig. 3B) patients. Notably, no sources lay inside PV sleeves, as would have been indicated by activation emanating from electrodes spanning PV ostia. However, 50.9% of sources lay near the PVs and LA roof in paroxysmal AF, and 33.5% in persistent AF (p < 0.05).

Source-ablated and non–source-ablated groups

Ablation lesions passed through one or more stable AF sources in 64.9% (61 of 94) of cases, composing 45% (27 of 60) of FIRM-blinded patients and 100% (34 of 34) of FIRM-guided cases (Table 1). The interobserver kappa score for assignment of source- versus no-source lesion delivery was 0.898.

Sources comprised rotors in 79.1% and focal impulses in 20.9%; the difference between source-ablated and non–source-ablated groups was not significant. The ratio of right-to-left atrial sources trended higher in patients with no source ablation versus with source ablation (p = 0.13) (Table 2).

Coincidental ablation of AF sources

Figure 4 illustrates AF sources lying near conventional ablation sites. Figure 4A shows an AF rotor septal to the left superior PV ostium in persistent AF. Its FIRM lesion area (2.1 cm²) fell naturally along the subsequent left WACA, indicating that this AF rotor may have been ablated “coincidentally” by conventional ablation alone. Figure 4B shows 2 FIRM sites near the left PVs sustaining persistent AF, of which the superior FIRM set (2.2 cm²) also lay naturally along the subsequent left WACA. In this case, the inferior FIRM site (1.9 cm²) may have been missed by conventional ablation, although with FIRM-guidance it was ablated and joined to the left WACA.

Figure 4C illustrates an AF rotor in persistent AF that was coincidentally ablated during an empirical LA roof line. Bilateral PV isolation had been performed, with no measurable impact on AF. Shortly after starting the LA roof line, before it was completed, AF terminated abruptly to sinus rhythm by ablation at the red dot site. The FIRM map was generated after the procedure, and showed a rotor at this precise site. Each patient remains AF-free on follow-up.

AF source ablation and long-term efficacy

Over a median follow-up of 273 days (IQR: 138 to 636 days), freedom from AF was higher in the group in which ablation passed through sources versus no sources (49 of 61 [80.3%] vs. 6/33 [18.2%], respectively; p < 0.001). Results were unchanged when examining only first-ablation cases (33 of 42 [78.6%] vs. 4 of 22 [18.2%]; p < 0.001). Fig. 5A shows the Kaplan-Meier plot. Freedom from AF was similar whether sources were ablated directly (FIRM-guided) or coincidentally (FIRM-blinded) (Fig. 5B). Freedom from any atrial tachyarrhythmias was also higher when ablation did versus did not pass through sources (43 of 61 [70.5%] vs. 5 of 33 [15.2%]; p < 0.001). In FIRM-guided patients, these included one case of cavotricuspid isthmus–dependent flutter and two cases of left atrial tachycardia; in FIRM-blinded patients, this included four cases of left atrial tachycardia. Finally, Figure 5C presents a Kaplan-Meier plot showing that freedom from AF was highest if all sources were ablated (29 of 33 [87.9%]), intermediate if some sources were ablated (8 of 12 [66.7%]), and lowest if no sources were ablated (4 of 4 [16.7%]).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Cumulative Freedom From AF

Based on (A) whether ablation did (blue) or did not (red) pass through AF sources. Entire population (solid lines) and first ablation patients only (dashed lines). (B) Direct (FIRM-guided, blue) or coincidental (FIRM-blinded, red) source ablation. (C) elimination of all (green), some (yellow) or no (red) stable AF sources, in patients with bi-atrial baskets. The p values reflect the complete follow-up periods. Abbreviations as in Figure 1.

Overall, 44% of patients had implantable ECG monitors (88.2% FIRM-guided, 26.1% FIRM-blinded), comprising 33 of 61 (54.1%) in the source-ablated group versus 10 of 33 (30.3%) in the non–source-ablated group (p = 0.03).

Stable AF sources

Although direct evidence for stable sources for human AF is recent (15,16), a large indirect body of literature exists showing that localized ablation at sites of high dominant frequency, fractionated electrograms, and other “substrates” may eliminate AF (24–26); that persistent AF can terminate at the first ablation steps (25); and that AF patients show consistent gradients in frequency (27,28) and organization (29,30) over time.

Human AF rotors show many characteristics similar to those of rotors in some animal models (31–34), including limited precession (22,35) in conserved areas that define FIRM lesion sets. The role of stable sources in sustaining human AF is shown by the ability of FIRM ablation alone (35,36) to terminate and render AF noninducible (16,17,22), and by higher freedom from AF in patients receiving FIRM-guided versus conventional ablation (82.4 vs. 44.9% in the CONFIRM trial) (15). Other recent reports have used contact mapping to reveal stable human AF rotors by Shannon entropy (37), wavelet similarity (38), and isochronal (39) analyses. Collectively, these studies contradict earlier reports that human AF rotors did not exist (20) or were not stable (21).

The CONFIRM trial (15) and a preliminary 12-center experience in >200 patients (17) suggest that AF freedom may be substantially improved with FIRM-guided ablation compared with conventional ablation. Ongoing studies are investigating why coincidental source ablation terminates AF less often than does direct FIRM ablation (15). The most likely explanation is that FIRM ablation targets all of the approximately 2 or 3 concurrent rotors or focal sources (15), whereas this finding is inconsistent with conventional ablation. Elimination of all sources (Fig. 5C) was actually more effective than reported in the CONFIRM trial by intention-to-treat FIRM-guided ablation (15), which in some patients was not completed at all sources. Elimination of all sources (Fig. 5C) was more effective than was elimination of some or none, whereas ablation of some sources yielded similar results to those reported with conventional ablation (3–5).

Patient-specific distributions of stable AF sources

We show that stable AF sources lie in patient-specific locations with ≈40% to 50% near PVs and left atrial roof, and ≈20% to 30% in right atrium. AF source locations were more widely distributed for persistent than paroxysmal AF.

AF source distributions may reflect fiber architecture (40), fibrosis or scar (41), electrical remodeling (42), or altered innervation (43). The precise determinants of AF source location await definition, which will require detailed translational studies in model systems (32,34) that recapitulate human AF. However, the similar source numbers and FIRM-ablation success in patients with and without prior ablation (15) argue that FIRM-identified sources are not created by prior lesions and are thus distinct from those in studies of ablation line gaps (44).

Elimination of stable AF sources as a possible explanation of AF freedom after diverse ablation strategies

It is clear that AF can be triggered by ectopic beats from the PVs. However, although PV-isolation strategies can eliminate triggers acutely, it is often difficult to achieve robust PV isolation and additional triggers may lie at undefined sites outside the PVs (3–6,8).

The present study provides an alternative potential explanation of why ablation treats AF in some patients and not others. Elimination of AF sources may explain why WACA is more effective than is ostial PV isolation (45); why AF may not recur in patients whose PVs have reconnected (46); why non–PV-encircling lines (47) or fractionated electrogram ablation (12) may be effective; and, potentially, why ablation success correlates with the extent of ablated tissue (25) in persistent AF.

Study limitations

The CONFIRM study was limited by its nonrandomized design, although subjects were enrolled, mapped, and treated prospectively. Its strengths included an active-control group (conventional ablation) (rather than controls receiving previously ineffective drugs, as in prior studies), and the use of implanted monitors in 88% of FIRM-guided patients. We accept that the lower use of implanted monitors in patients receiving conventional ablation (and in prior AF trials) reduces the accuracy of follow-up by potentially missing intermittent arrhythmia recurrences. Future studies may improve on current definitions of source ablation by using contact force-sensing and/or magnetic resonance imaging to ensure and verify effective ablation at each site. The relatively small numbers of at-risk patients at longer follow-up periods is also a limitation. Multicenter studies will address the limitations introduced by this predominantly male Veterans Affairs population and will test the benefits of FIRM-guided ablation in a much larger population.