Magnetic resonance-guided re-ablation for atrial fibrillation is associated with a lower recurrence rate: a case–control study

Abstract

Aims

Our aim was to analyse whether using delayed enhancement cardiac magnetic resonance imaging (DE-CMR) to localize veno-atrial gaps in atrial fibrillation (AF) redo ablation procedures improves outcomes during follow-up.

Methods and results

We conducted a case–control study with 35 consecutive patients undergoing a DE-CMR-guided Repeat-pulmonary vein isolation (Re-PVI) procedure. Those with more extensive ablations (e.g. roof lines, box) were excluded. Patients were matched for age, sex, AF pattern, and left atrial dimension with 35 patients who had undergone a conventional Re-PVI procedure guided with a three dimensional (3D)-navigation system. Procedural characteristics were recorded, and patients were followed for 24 months in a specialized outpatient clinic. The primary endpoint was freedom from recurrent AF, atrial tachycardia, or flutter. The duration of CMR-guided procedures was shorter compared to the conventional group (161 ± 52 vs. 195 ± 72 min, respectively, P = 0.049), with no significant differences in fluoroscopy or total radiofrequency time. At the 2-year follow-up, more patients in the DE-CMR-guided group remained free from recurrences compared with the conventional group (70% vs. 39%, respectively, P = 0.007). In univariate Cox-regression analyses, AF pattern [persistent AF, hazard ratio (HR) 2.66 (1.27–5.46), P = 0.006] and the use of DE-CMR [HR 0.36 (0.17–0.79), P = 0.009] predicted recurrences during follow-up; both factors remained independent predictors in multivariate analyses.

Conclusion

The substrate characterization provided by DE-CMR facilitates the identification of anatomical veno-atrial gaps and associates with shorter procedures and better clinical outcomes in repeated AF ablation procedures.

Introduction

Ablation procedures evolve as a cornerstone of the therapeutic arsenal for patients with atrial fibrillation (AF). Pulmonary vein isolation (PVI) has proven superior to drug therapy in maintaining sinus rhythm in the long term.¹ However, a significant percentage of patients present recurrences, and ∼20% require a second ablation procedure.² The underlying mechanisms behind recurrences are diverse, but discontinuities in the ablation lines (gaps) seem to be important contributors.^3,4 Redo procedures commonly involve a detailed mapping of the veno-atrial junction with a circular catheter placed inside the PV to localize veno-atrial reconnection sites.⁵ Nevertheless, technical (e.g. inaccurate electrogram assessment) and anatomical (e.g. electrical activation patterns) limitations often result in the need for multiple radiofrequency (RF) applications.

Delayed enhancement cardiac magnetic resonance imaging (DE-CMR) has been demonstrated to non-invasively characterize atrial myocardial fibrosis.⁶ Histological studies in biopsies of patients undergoing cardiac surgery have demonstrated that areas with more intense delayed enhancement correspond with fibrotic patches.⁷ The extension of atrial fibrosis by DE-CMR is a strong independent predictor of recurrence after ablation.⁸ Ongoing clinical trials [ALICIA (NCT02698631), DECAAF2 (NCT02529319)] are testing whether targeting native fibrotic patches during ablation could improve outcomes and be useful to further personalize AF treatment. On the other hand, DE-CMR also allows visualization of left atrial (LA) lesions (scarring) created by previous ablations. This enables the identification of areas without delayed enhancement within the ablation lines (gaps).^9,10 However, whether lesion discontinuities represent true electrical gaps that could serve to guide redo procedures remains controversial.^11,12

The objective of this study was to test whether preprocedural identification of gaps in a DE-CMR might serve to improve the results of Re-PVI procedures.

Methods

Population

We conducted a case–control study, including 35 consecutive patients undergoing a Re-PVI procedure guided by DE-CMR at our centre from June 2012 to November 2014. The exclusion criteria were previous recurrence as atrial flutter or atrial tachycardia, gadolinium allergy, claustrophobia, severe renal failure (glomerular filtration rate < 30 mL/min), or the presence of cardiac electronic implantable devices. To reduce their potential confounding effect on ablation outcomes, we focused on those patients undergoing isolated Re-PVI: patients in whom additional ablation lines (e.g. roof line, posterior box, or substrate homogenization) had been performed were excluded from the study. A control group of 35 patients who underwent Re-PVI following the conventional circular catheter-guided approach without magnetic resonance imaging (Conventional group) during the same period were retrospectively selected and matched with the study group for age (± 5 years), sex, recurrent AF pattern (paroxysmal/persistent), and LA dimensions (± 7 mm).

Patients in both groups were prospectively followed at 3, 6, 12, and 24 months after the re-ablation procedure with an electrocardiogram (ECG) and 24-h Holter monitoring at a specialized outpatient clinic. Patients presenting symptoms at any time between the pre-established follow-up points were encouraged to obtain an ECG. Recurrence was defined as any documented episode of AF, atrial tachycardia, or atrial flutter lasting more than 30 s after a 3-month blanking period. Clinical, echocardiographic, and follow-up data were collected in both groups from the prospective AF ablation registry at our institution.

The study was approved by the ethics committee of our institution, and informed consent was obtained from all the participants. The data that support the findings of this study are available from the corresponding authors (E.G. and L.M.) upon reasonable request.

Delayed enhancement cardiac magnetic resonance imaging image acquisition and post-processing

Patients undergoing a DE-CMR-guided Re-PVI had a DE-CMR obtained prior to the procedure to identify possible PV gaps. A gap was defined as any discontinuity in the circumferential scar around the PVs. Image acquisition and post-processing have been described previously.¹² In brief, the CMR was performed in sinus rhythm; if necessary, external cardioversion was previously performed. A 3T scanner (Magneton Trio-Tim^®, Siemens Healthcare, Erlangen, Germany) using a 32-channel cardiac coil was used for all patients. After 20 min of intravenous 0.2 mmol/kg gadobutrol administration (Gadovist, Bayer Pharma, Germany), a free-breathing three-dimensional (3D) navigator and electrocardiogram-gated inversion-recovery gradient-echo sequence were applied in axial projection. The acquired voxel size was 1.25 × 1.25 × 2.5 mm. Other typical sequence parameters were as follows: repetition time/echo time, 2.3/1.4 ms; flip angle, 11°; bandwidth, 460 Hz/pixel; inversion time (TI), 280–380 ms; and parallel imaging with GRAPPA technique, with reference lines of R = 2 and 72. A TI scout sequence was used to nullify the left ventricular myocardial signal and determine optimal TI. Patients were instructed to maintain steady, shallow breathing during image acquisition. The typical scan time for a DE-CMR sequence was 15 min (11–18 min) depending on heart rate and breathing patterns. Then, all DE-CMR images were segmented by experienced observers with ADAS3D software (ADAS3D Medical, Barcelona, Catalonia, Spain). The atrial wall was manually traced in axial slices, and the blood pool was automatically calculated. Colour-coded pixel signal intensity maps were projected to the 3D shell of the atrium. Pixel signal intensity maps were normalized to the mean intensity of the blood pool signal, and the resulting value (IIR, image intensity ratio) was plotted. Previously validated fibrosis thresholds were used¹³: an IIR >1.32 defines dense scar areas, and an IIR between 1.2 and 1.32 determines interstitial fibrosis.¹³ A gap was defined as a discontinuity in previous ablation lesions (Figure 1). Post-processing reproducibility has been established in previous studies.¹⁴

The resulting processed 3D model of the LA DE-CMR was integrated with the electroanatomical map of the navigation system (CARTO 3, Biosense Webster, Diamond Bar, CA, USA) to guide the ablation procedure.

Ablation procedure

All Re-PVI procedures were performed with a double transseptal puncture to access the LA. Anticoagulation was not interrupted in patients taking vitamin K antagonists, and a dose was skipped if direct oral anticoagulants were used. A bolus of heparin was administered, followed by additional boluses to maintain an activated clotting time >300 s. A 3D map of the LA and PV was reconstructed using a multipole catheter (Lasso, Biosense Webster) and an electroanatomic mapping system (CARTO 3, Biosense Webster, Diamond Bar, CA, USA). Radiofrequency ablation was performed using an irrigated-tip catheter in temperature-control mode 45°C/40 W at the anterior wall and 45°/30 W at the posterior wall. The procedural endpoint was both entrance (absence or dissociation of a local electrogram inside the entire encircled region) and exit block (by pacing at the PV ostia along the circular catheter) in all four pulmonary veins. No additional ablation lines or substrate modification were added during the procedures.

In the Conventional group, PV reconnection sites were identified as the earliest PV electrogram in the activation sequence of a circular catheter at the PV antrum.

In the DE-CMR group, ablation was performed as described previously.¹² The 3D electroanatomical reconstruction was merged with the model obtained from DE-CMR. Ablation was directed towards areas of the vein ostium identified as gaps by the DE-CMR. If isolation had not been achieved by DE-CMR-guided ablation, further RF applications were applied after mapping with the circular catheter. Figure 2 summarizes an example DE-CMR-guided procedure.

A contact-force catheter (Smart-Touch, Biosense Webster, Diamond Bar, CA, USA) was used in 12 procedures (17%), in a similar proportion between the two groups (seven in the DE-CMR group and five in the Conventional group, P = 0.75).

Statistical analyses

Continuous variables are presented as mean ± standard deviation. A t-test was used to compare the means of two variables. Categorical variables are expressed as total number (%) and compared using the χ² test. Recurrence-free survival over time was calculated by the Kaplan–Meier method, and comparison between groups was performed with the log-rank test. Univariate recurrence predictors were identified with Cox-regression models; the use of DE-CMR in the ablation procedure and other potential predictors (age, sex, hypertension, diabetes, obstructive sleep apnoea, left ventricular ejection fraction, LA size, and AF pattern) were tested. Proportionality of hazards was confirmed with a Schoenfeld test. Those significant predictors in univariate analyses were included in multivariate analyses carried out with a forward stepwise approach (P for entrance < 0.05, P to remain < 0.1). The hazard ratio (HR) and 95% confidence interval (CI) of AF recurrence are reported for each predictor. All analyses were performed with SPSS v18.0 (SPSS, Chicago, IL, USA). All statistical tests were two-sided, and a P-value < 0.05 was considered statistically significant.

Results

The baseline characteristics of patients in the DE-CMR and Conventional groups are shown in Table 1. Both groups were well matched for all parameters and mainly included middle-aged men with mildly dilated atria. Procedural characteristics are summarized in Table 2. A common trunk was found in three patients in the Conventional group (left common trunk in two patients, right common trunk in one patient) and four patients in the DE-CMR group (left in three patients, right in one patient). Acute electrical isolation was achieved in all cases. The ablation procedure was 32 min [95% CI (2–63); P = 0.049] shorter in the DE-CMR compared with the Conventional group. Although fluoroscopy and RF time were shorter in the DE-CMR compared to the Conventional group, these did not reach significance. Major procedural complications occurred in one patient in the DE-CMR (cardiac tamponade requiring drainage). Minor complications were similar amongst groups (two groin haematoma and two patients with fever in the Conventional group; three groin haematoma and one pericarditis in the DE-CMR group, P = 1.0).

Follow-up

At the 24-month follow-up, in the whole cohort, 55% of patients remained free from AF/atrial flutter/tachycardia. Kaplan–Meier survival analysis showed a lower probability of recurrence in the DE-CMR than in the Conventional group (P = 0.007; Figure 3). At 24 months of follow-up, 20 (70%) patients in the DE-CMR Re-PVI group and 10 (39%) patients in the Conventional group remained free of AF/atrial flutter/tachycardia. Recurrence as AF and atrial flutter are displayed in Figure 4. There were no recurrences reported as atrial tachycardia. More patients in the Conventional group than in the DE-CMR group recurred as atrial flutter [six patients in the Conventional group vs. one patient, respectively; HR for DE-CMR-guided ablation 0.1 (95% CI 0.012–0.86)]. At the 24-month follow-up, antiarrhythmics were more often prescribed in the Conventional than in the DE-CMR group [18/31 vs. 9/35, odds ratio 4.54 (95% CI 1.49–20)], likely due to the higher recurrence rate.

Recurrence predictors

Recurrence predictors were initially analysed in Cox-univariate analyses (Table 3). Amongst all tested predictors, only DE-CMR-guided procedures [HR 0.36 (95% CI 0.17–0.78), P = 0.009] and persistent AF before the Re-PVI procedure [HR 2.72 (95% CI 1.32–5.61), P = 0.006] predicted atrial tachyarrhythmias recurrence after a 24-month follow-up. Both factors remained significantly associated with ablation outcomes in the multivariate analysis [Table 3; HR 0.36 (95% CI 0.16–0.77) and HR 2.79 (95% CI 1.35–5.75) for DE-CMR use, and persistent AF, respectively]. In a sensitivity analysis, results remained unaltered after adjustment for ablation year (results not shown).

Discussion

We report the results of the first head-to-head comparison between a DE-CMR-guided and an electroanatomical approach in patients exclusively undergoing Re-PVI for recurrent AF. In this case–control study, we show that the use of DE-CMR associates with shorter procedures and a lower recurrence risk during follow-up in comparison to a conventional approach.

The optimal procedural endpoint in patients undergoing a redo AF ablation remains challenging, and several approaches have been proposed. In general, incomplete PVI or PVI reconnection underlie recurrences after a first AF ablation procedure,¹⁵ partially due to veno-atrial conduction recovery.¹⁶ Cryoballoon and RF ablation yield limited ability to create durable complete circular scars around the PV, which is achieved in <10% of patients.¹⁰ Underlining the importance of lesion completeness, the relative length of gaps surrounding pulmonary veins has been shown to predict recurrence risk after a first AF ablation procedure.⁹ Therefore, it seems reasonable in repeated AF ablation procedures to ensure that PVI has been achieved before considering a more extensive ablation strategy.^5,17 To specifically test the role of DE-CMR for this purpose, our work included patients undergoing Re-PVI only, ruling out the confounding effect of additional ablation lines.

Delayed-enhancement CMR evolves as a powerful tool to identify and target ablation lesion gaps. Previous studies have demonstrated that DE-CMR can predict electrical reconnection sites by identifying gaps in ablation lesions,¹² although conflicting data have been published.¹¹ More recently, Fochler et al.¹⁸ reported that targeting DE-CMR detected scar isthmuses and LA scar is feasible, and proposed their use to treat recurrent arrhythmias post-AF ablation. We extend these insights to patients recurring as AF and report the first study to assess the potential benefit of using pre-procedural DE-CMR over a conventional approach in comparable groups of patients at baseline. Our results demonstrate not only that DE-CMR-guided Re-PVI is feasible, but suggest that it outperforms a conventional, circular catheter-based approach.

Multiple reasons might account for the better results with the use of DE-CMR, including limitations of bipolar voltage mapping to localize gaps. Technical limitations (e.g. inappropriate contact, low-density maps, far-field of nearby areas) might jeopardize their accuracy to identify small gaps. Using a standard circular catheter inside the PV might fail to identify the exact location of the gap, as the dipole with the earliest activation time might not be located exactly at the functional gap site. Furthermore, the activation of myocardial bundles inside the PV may not follow a straight vector, as previously observed in an anatomic study.¹⁹ In contrast, DE-CMR displays a series of potential advantages. A preprocedural DE-CMR might aid in pre-procedural planning. During the procedure, the localization of ablation gaps could result in more accurate and precise lesions connecting adjacent scar lesions. In this regard, the Conventional ablation group showed greater recurrence of atrial flutter, which might be due to more extensive RF ablation deliveries, potentially facilitating re-entry circuits that could sustain atrial flutter.

Limitations

Some limitations of our work should be acknowledged. This was an observational study, and its non-randomized design might account for potentially undetected biases. Although characteristics were balanced between groups at baseline, randomized clinical trials are warranted to establish the role of DE-CMR in AF ablation procedures. Contact-force ablation catheters were used in less than one-fifth of cases, which might explain the high recurrence rate in the control group. The Conventional group aimed at Re-PVI using a circular mapping catheter; high-density electroanatomical mapping might improve gap identification and yield better long-term outcomes.²⁰ Last, while our results suggest that DE-CMR could aid in re-isolating PV, its role in more extensive ablation procedures including atrial lines or targeting fibrosis patches remains unclear.

Conclusions

The use of DE-CMR to guide PVI in redo procedures was associated with a shorter procedure time and a lower risk of recurrence. Randomized studies are warranted to confirm our results and establish the utility of this technique.

Funding

This work has been partially funded by research grants from the European Commission (CATCHME project, contract agreement 633196); Instituto de Salud Carlos III (PI16/00435; PI19/00443; PI19/00573); Fundació la Marató TV3 (reference 20152730); and CERCA Programme/Generalitat de Catalunya.

Conflict of interest: L.M. reports activities as a consultant, lecturer, and advisory board member for Abbott Medical, Boston Scientific, Biosense Webster, Medtronic, and Biotronik. L.M., J.B., and A.B. are shareholders of Galgo Medical and S.L. J.M.T. declares consulting services, and advisory boards for Abbott Medical, Medtronic, Biotronik and Boston Scientific. I.R.-L. declares consulting services for Abbott Medical and Boston Scientific. The remaining authors report no conflicts of interest.

References

Author notes