Corresponding Author:
Daniel Wetherbee Nelson, MD
Cardiology Fellow
Tulane University – Cardiovascular Division
1430 Tulane Ave, New Orleans, LA 70112
Office: 504-988-5492
ORCID iD: 0000-0003-0493-6650
Email: wexbee09@gmail.com
Abstract:
Introduction: Pulmonary vein isolation (PVI) using
radiofrequency (RF) and cryoballoon (Cryo) ablation are standard
approaches for rhythm control of symptomatic atrial fibrillation. Both
strategies involve scar formation of the left atrium (LA). There have
been few studies investigating the differences in residual fibrosis and
scar formation in patients undergoing RF and Cryo using cardiac magnetic
resonance imaging (CMR).
Methods: The current study is a sub-analysis of the control arm
of the Delayed-Enhancement MRI Determinant of Successful Catheter
Ablation of Atrial Fibrillation study (DECAAF II). The study was a
multicenter, randomized, controlled, single blinded trial that evaluated
atrial arrhythmia recurrence (AAR) between PVI alone and PVI plus CMR
atrial fibrosis guided ablation. Pre-ablation CMR and 3–6-month post
ablation CMR were obtained to assess baseline LA fibrosis and scar
formation respectively.
Results: Of the 843 patients randomized in the DECAAF II trial,
we analyzed the 408 patients in the primary analysis control arm that
received standard PVI. Five patients received combined RF and Cryo
ablations so were excluded from this sub-analysis. Of the 403 patients
analyzed, 345 underwent RF and 58 Cryo. The average procedure duration
was 146 minutes for RF and 103 minutes for Cryo (p = 0.001). The rate of
AAR at ~15 months occurred in 151 (43.8%) patients in
the RF group and 28 (48.3%) patients in the Cryo group (p = 0.62). On
3-month post CMR the RF arm had significantly more covered fibrosis
(3.6% vs. 3.0%, p = 0.04) and scar (8.8% vs. 6.4%, p = 0.001)
compared to Cryo. Patients with ≥ 6.5% LA scar on 3-month post CMR had
less AAR independent of ablation technique (RF p = 0.009, Cryo p =
0.02). Cryo caused a greater percentage of right and left pulmonary vein
(PV) scar (p = 0.04, p = 0.02) and less non-PV scar (p = 0.009) compared
to RF. On Cox regression Cryo patients free of AAR had a greater
percentage of left PV scar (p = 0.01) and less non-PV scar (p = 0.004)
compared to RF free of AAR.
Conclusion: In this sub-analysis of the control arm of the
DECAAF II trial, there was no significant difference in the rate of AAR
in patients undergoing PVI alone between RF vs. Cryo. Post ablation LA
scar ≥ 6.5% predicted freedom from AAR, independent of ablation
technique. Cryo formed a greater percentage of PV scar and less non-PV
scar compared to RF, which may have prognostic implications.
Keywords: atrial arrhythmia recurrence, PVI, radiofrequency,
cryoballoon, CMR, fibrosis, scar
Introduction:
Catheter ablation in patients with symptomatic persistent atrial
fibrillation (AF) is effective in maintaining sinus rhythm and is a
class I indication in symptomatic patients who have failed
antiarrhythmic drug therapy.1 Standard pulmonary vein
isolation (PVI) remains the main-stay treatment while additional
ablations such as posterior wall isolation, left atrial roof line
ablation, targeting atrial rotors, complex fractionated electrograms,
low voltage/complex electrograms, or baseline left atrial (LA) fibrosis
have not shown added benefit.2-7 The two most common
ablation techniques for PVI include radiofrequency (RF) ablation using
thermal tissue heating, and cryoballoon (Cryo) using cryogenic freezing
to cause cellular necrosis.8
Cryo is gaining popularity as a treatment modality for AF, given its
relatively simpler technique, shorter procedure duration, and promising
clinical results.9-11 In patients with symptomatic
paroxysmal AF, Cryo has been shown to be noninferior to RF and superior
to drug therapy at maintaining sinus rhythm, reducing healthcare
utilization, and delaying progression to persistent
AF.9,12-14 Most prospective comparisons between RF and
Cryo have included mostly paroxysmal AF patients so further evaluation
in the persistent AF population is warranted.9,15
The difference in ablation induced scar formation by the two techniques
is also not well studied. Cardiac magnetic resonance (CMR) imaging has
been shown to reliably assess LA fibrosis, with post-ablation late
gadolinium enhancement (LGE) showing predictiveness for atrial
arrhythmia recurrence (AAR) or even sites of breakthrough conduction at
repeat ablation.16-19 Hence it would be interesting to
see if the two techniques have different effects on the LA post
ablation.
The current sub-analysis from the control arm of the DECAAF II trial was
designed to compare 1) the rate of AAR after standard RF and Cryo PVI
ablation in patients with persistent AF 2) the effect standard RF and
Cryo PVI ablation have on post ablation scar using CMR 3) if scar on
post ablation CMR can be used to predict AAR.
Materials & Methods:
Study Design: This study is a sub-analysis of the control arm
of the DECAAF II trial evaluating the rate of AAR and CMR results
between patients with persistent AF who underwent standard RF and Cryo
PVI ablation. A full discussion of the DECAAF II methods can be found at
DOI: 10.1111/jce.14957. In brief, the DECAAF II trial was a prospective,
randomized, investigator initiated, industry sponsored, multicenter,
international, controlled, single blinded trial that randomized
persistent AF patients to standard PVI ablation or PVI plus CMR guided
atrial fibrosis ablation. Of the 843 patients who were randomized, 417
patients were included in the control arm. Nine patients were lost to
follow up and 5 patients received both RF and Cryo ablations so were
excluded, resulting in 403 patients which were included in this as
treated sub-analysis. Of these 403 patients, 345 underwent RF and 58
Cryo ablation (Figure 1). All patients had LGE-CMR before and 3-6 months
post ablation. Patients were followed for 12-18 months using daily
smartphone electrocardiogram (ECG) recordings to assess for AAR. A Data
Coordinating Center at the University of Utah was responsible for
recording and analysis. Written informed consent was obtained from all
participants prior to study inclusion. The study was approved by the
ethics review committee at all sites.
Patient Population: To enroll, patients had to be at least 18
years old and have documented persistent AF without prior ablation.
Persistent AF was defined as lasting 7 days or more on rhythm strip or
chart review. Exclusion criteria included patients with prior ablation,
contraindication for CMR or beta blockers, inability to be positioned in
CMR scanner, pregnant women, terminally ill, or those without smartphone
capability.20
Imaging Protocol: All patients underwent LGE-CMR within 30 days
prior to ablation and between 90-180 days post. Merisight delayed CMR
protocol (MARREK Inc) and image processing software were used. If the
patient’s heart rate was greater than 90 beats per minute, he or she was
pre-medicated with a beta blocker. Images were transmitted to MARREK Inc
where they were reviewed, processed, and quantified with respect to LA
fibrosis (pre-ablation) and LA scar formation (post-ablation). LA
fibrosis was defined using the previously described Marrek Inc, LGE-CMR
protocol (MRI and image processing software).21,22Post-ablation CMR findings for the LA were classified as post ablation
scar, covered fibrosis, and non-covered fibrosis. Post ablation scar was
the percentage of the entire LA which enhanced at a higher LGE threshold
and constituted iatrogenic scar induced by the PVI. Of note, post
ablation scar had higher intensity signals that effectively masked the
fibrosis pattern noted on pre-ablation scans. The pre-ablation CMR was
overlaid on the 3–6-month post-ablation CMR and areas of baseline
fibrosis which overlapped with areas of scar were termed covered
fibrosis, whereas areas of baseline fibrosis not covered by scar were
termed non-covered fibrosis (Figure 2). Suboptimal images were not
processed. The site physician could opt to reimage patients who had
images below quality standards.
PVI: In this sub-analysis all patients underwent standard PVI
ablation with either RF or Cryo. The operator created lesions around the
pulmonary vein (PV) antra as described by the HRS consensus
statement.23 Standard techniques were used to confirm
entrance, and at the discretion of the operator, exit block. If normal
sinus rhythm was not achieved post ablation, patients were cardioverted.
Additional non-PV trigger ablation lesions could be performed at the
discretion of the operator if normal sinus rhythm could not be restored
post cardioversion.20
Cryoballoon ablation: Cryo ablation was chosen at the
operator’s discretion. A 28-mm Arctic Front Advance balloon was
recommended to maximize antral ablation. Termination of energy was
recommended if Cryo temperatures were below -55 degrees °C with 28-mm
Arctic Front and -60 degrees °C with 23-mm Arctic
Front.20 If Cryo was unsuccessful after 540 seconds of
therapy, the operator was encouraged to cannulate a different vein
branch, use a different size cryoballoon, try focal cryocatheter, or
employ RF.
Primary and Secondary Outcomes: The primary outcome was first
AAR (AF, atrial flutter, atrial tachycardia) after the 90-day blanking
period confirmed with at least one valid ECG tracing, 2 consecutive
smartphone ECG tracings more than 6 hours apart, or another continuous
monitoring device (Holter). Patients were told to submit ECGs daily and
these transmissions were reviewed at the central location. Secondary
outcomes after the 90-day blanking period included the rates of
antiarrhythmic use, repeat ablation, and cardioversion.
Safety Outcome: The primary safety composite outcome was the
occurrence of at least one of the following within the 30-day period
after the ablation procedure: stroke, heart failure, bleeding requiring
transfusion, PV stenosis, or death from any cause. Additional safety
outcomes include the individual components of the primary safety
composite, occurrence of esophageal injury, perforation, or tamponade,
and serious adverse events including cardiovascular, cerebrovascular,
and gastrointestinal events. The safety end points were adjudicated by a
3-member outcomes committee based on the 2017 HRS Consensus
Statement.23
Statistical Analysis: Statistical analysis was performed in the
as treated control arm of the DECAAF II study in patients who received
PVI ablation alone. The group was analyzed between those who received RF
and Cryo. Categorical data, expressed as a number and percent, were
compared using the chi-squared test. For data with sample size
< 5, the Fischer exact test was used. Continuous data,
expressed as mean +/- SD, were compared using the independent Studentt test. For the survival analysis, follow-up time began 90-days
after the date of ablation and ended at the diagnosis of an atrial
arrhythmia event. For the Kaplan-Meier curves, the log-rank test was
used for significance. The multivariate analysis used a Cox proportional
hazards model to obtain the appropriate estimates and confidence
intervals of hazard ratios. Proportionality assumptions were confirmed
by graphical methods. The primary analysis was performed with a
two-sided significance level of 0.05. Analysis was performed using R
version 3.3.2 GUI Mavericks build.
Results:
Baseline Results: Enrollment started in July 2016 and was
completed in February 2021. In the RF arm, all 345 patients underwent
the baseline CMR and 313 (90.7%) underwent the follow up CMR. In the
Cryo arm, all 58 patients underwent the baseline CMR and 51 (87.9%)
underwent the follow-up CMR. In the RF group, 341 patients (98.8%)
completed the 3-month follow up visit and 335 (97.1%) completed the
12-month follow up visit. In the Cryo group, 55 patients (94.8%) and 53
patients (91.4%) completed the 3 and 12-month follow up respectively.
The average number of days with transmitted ECGs was 193.9 +/- 139.5
days in the RF group and 134.4 +/- 136.5 days in the Cryo group (p =
0.003). The characteristics of the patients at baseline were balanced
between the two groups, except for the number of patients on
antiarrhythmic and anticoagulation therapy was higher in the RF group
and the baseline brain natriuretic peptide level was higher in the Cryo
group (Table 1A).
Baseline CMR Results: The pre-ablation CMR left atrial volume
index (LAVI) was comparable between RF and Cryo (p = 0.57) (Table 2A).
Additionally, the baseline fibrosis on the pre-ablation CMR was similar
between the two groups (Table 2A). The RF group had an average of 18.5
+/- 7.2% LA fibrosis while the Cryo group had an average of 20.8 +/-
8.4% LA fibrosis (p = 0.08). The percentage of atrial fibrosis around
the right and left PVs as well as fibrosis not located around the PVs
was similar between RF and Cryo groups (Table 2A). The percentage of
patients across Utah Stages 1, 2, 3, and 4 was also balanced between RF
and Cryo (Supplemental Table 1).
3-Month Post Ablation CMR Results: We observed a significant
reduction in LAVI in all patients undergoing ablation regardless of
technique (Table 2A). Reduction in LAVI was significantly greater in the
RF group compared to Cryo (10.1 +/- 11.6 cc vs. 5.7 +/- 11.4 cc, p =
0.01). Covered LA fibrosis was higher in RF compared to Cryo (3.6 +/-
2.2% vs. 3.0 +/- 1.8%, p = 0.04) while non-covered LA fibrosis was
lower in RF compared to Cryo (14.9 +/- 6.3% vs. 17.8 +/- 7.8%, p =
0.01). The follow up LA scar was higher in the RF group compared to Cryo
(8.8 +/- 4.2% vs. 6.4 +/- 3.6%, p = 0.001). Similar results were seen
after stratifying by Utah Stage (Table 2B).
When comparing the regionality of LA scar, the Cryo group had a greater
percentage of scar formed around the right and left PVs while the RF
group had a greater percentage of non-PV scar (Table 2A). We also noted
a significantly greater percentage of scar formed around the left PVs
when compared to the right PVs. This finding was significant for the
total cohort (37.7 +/- 16.3% vs. 5.2 +/- 8.6%, p < 0.001) as
well as the Cryo (43.7 +/- 20.0% vs. 8.5 +/- 12.9%, p <
0.001) and RF (36.7 +/- 15.4% vs. 4.6 +/- 7.6%, p < 0.001)
subgroups.
Primary End Points: The primary efficacy endpoint, AAR after
the blanking period, was similar between groups, occurring in 151 RF
patients and 28 Cryo patients (43.8% vs. 48.3%, p = 0.52) (Table 3A)
(Figure 3). The use of antiarrhythmics 90 days post blanking period was
similar, with 45 patients (13.0%) in the RF arm and 10 patients
(17.2%) in the Cryo arm receiving antiarrhythmics (p = 0.51) (Table
3A). Likewise, the groups were comparable with respect to the number
undergoing repeat ablation and repeat cardioversion (Table 3A).
Stratified by Utah Stages 1, 2 and 3, 4 showed comparable rates of AAR
between RF and Cryo (Table 3B).
Predictive Value of Scar Formation:
Total Scar:
For the entire cohort, we noted a trend towards a greater percentage of
LA scar in patients free of AAR compared to those with AAR (8.8 +/-
4.2% vs. 8.0 +/- 4.2%, p = 0.09) (Table 4A). Cryo patients free of AAR
had a significantly higher percentage of LA scar than those with AAR
(7.4 +/- 4.2% vs. 5.4 +/- 2.5%, p = 0.04) (Table 4A). This difference
was not seen in the RF subgroup (9.0 +/- 4.1% vs. 8.5 +/- 4.3%, p =
0.34) (Table 4A).
Univariate analysis showed that post ablation LA scar ≥ 6.5% predicted
freedom from AAR (p = 0.005) (Figure 4). Multivariate Cox regression
analysis accounting for CHA2DS2VASC,
age, sex, antiarrhythmic medications, and baseline fibrosis reaffirmed
the prior finding that post ablation LA scar ≥ 6.5% predicted freedom
from AAR in the entire cohort (HR 0.59 [95% CI, 0.43-0.81]; p =
0.001). This threshold of ≥ 6.5% LA scar was predictive of freedom from
AAR independently in both RF and Cryo groups (p = 0.003) (Figure 5). RF
patients with ≥ 6.5% LA scar had less left PV scar and more non-PV scar
compared to Cryo patients with ≥ 6.5% LA scar (Table 4B).
Regional Scar :
Scar regionality was predictive of AAR as patients with a greater
percentage of LA scar around the PVs had less AAR. Univariate (Figure 6)
and multivariate Cox regression analysis accounting for
CHA2DS2VASC, age, sex, antiarrhythmic
medications, and baseline fibrosis showed that in the entire cohort,
patients with ≥ 2.3% absolute LA scar around the PVs (PV scar) had less
AAR (HR 0.663 [95% CI, 0.48-91]; p = 0.01). This inverse
relationship between PV scar and AAR was statistically significant in
the RF group (p = 0.01) and showed a trend towards significance in the
Cryo group (p = 0.29) (Supplemental Figures 1 and 2).
Analyzing regional scar further we divided total LA scar into PV scar
(left PV, right PV) and non-PV scar (Figure 7). Cryo patients free of
AAR had a greater percent left PV scar (45.5 +/- 18.7% vs. 37.5 +/-
15.3%, p = 0.01) and less non-PV scar compared to RF patients free of
AAR (48.2 +/- 12.9% vs. 57.9 +/- 11.6%, p = 0.004) (Table 4C) (Figure
8). There was no difference in right PV scar between RF patients free of
AAR and Cryo patients free of AAR (p = 0.38) (Figure 8).
Safety End Points: The procedural characteristics showed
similar fluoroscopy time between the groups but increased total sheath
and total transseptal time in RF compared to Cryo (Table 1B). The
primary composite safety outcome (stroke, heart failure, bleeding
requiring transfusion, PV stenosis, and death within 30 days) did not
occur in either group (0 RF vs. 0 Cryo, p = 1.0) (Supplemental Table 2).
The total number of deaths from any cause between 31 – 451 days post
ablation was equivalent with both arms having 1 death (p = 0.27)
(Supplemental Table 2). With respect to esophageal injury and
perforation or tamponade there was no difference between RF and Cryo
(Supplemental Table 2). The number of serious adverse events was
balanced, with 14 RF patients (4.1%) and 4 Cryo patients (6.9%) having
a serious adverse event (p = 0.31) (Supplemental Table 2).
Discussion:
In this sub-analysis of the DECAAF II trial of patients with persistent
AF who underwent routine PVI, there was no significant difference in AAR
between RF and Cryo. Patients with ≥ 6.5% LA scar on post ablation CMR
had less AAR. Cryo patients had a greater percentage of scar formation
around the PVs and less non-PV scar compared to RF.
Cryo ablation is gaining popularity as the preferred method of ablation
given the shorter procedure duration, ease of ablation delivery, and
equivalent safety profile. When comparing RF to Cryo for paroxysmal AF,
both techniques appear to be equally efficacious with respect to
AAR.9,24 However, there is a paucity of data comparing
RF to Cryo in patients with persistent AF. In the prospective cluster
FREEZE Cohort study evaluating the effectiveness and safety of Cryo
compared to RF in patients with paroxysmal and persistent AF, subgroup
analysis of patients with persistent AF showed no difference in AAR
after 3 months between Cryo and RF (Hazard ratio 1.01, CI 0.70-1.46, p =
0.95).25 Moreover, a pooled registry trial studying
Cryo vs. RF as an initial AF ablation procedure showed that in patients
with persistent AF, Cryo was as effective as RF with respect to the rate
of repeat AF ablation at 12-month follow up (Hazard ratio 0.92,
confidence interval 0.631-1.343).26 The present
sub-analysis adds to the growing evidence that Cryo is noninferior to RF
with respect to AAR in patients with persistent AF undergoing their
first ablation procedure (p = 0.62).
This DECAAF II sub-analysis is one of the first studies to observe the
effects RF and Cryo have on the LA by CMR.17,27 It
strengthens the evidence that greater extent of LA scarring after PVI
ablation may be associated with less AAR.18,19Univariate (Figure 4) and multivariate analysis showed that patients
with ≥ 6.5% LA scar had less AAR, and this trend was seen in both RF
and Cryo groups (Figure 5). In addition to finding predictive value in
the percentage of total LA scar, we found that the location of scar
around the PVs has prognostic value (Figure 6). These findings suggest
that patients with a higher percentage of LA scar, specifically around
the PVs, may have more durable and continuous lesions that result in
better PVI, thus improving freedom from AAR. Scar formation is not
without risk however, and clinicians must balance the benefits of
ablation with the risk of stiff left atrial syndrome, esophageal injury,
and stroke.28-30 Overall, the percentage and location
of post ablation LA scar may help clinicians predict which patients
recur after PVI and consider additional strategies for rhythm control.
Interestingly in our cohort, RF and Cryo had similar rates of AAR
despite RF having a greater percentage of LA scar. This may be explained
by the regionality of LA scar formation. Cryo patients had more PV scar
and less non-PV scar compared to RF patients (Table 2A). Furthermore,
Cryo patients free of AAR had a greater percentage of left PV scar and
less non-PV scar than RF patients free of AAR (Figure 8, Table 4C). This
same trend of Cryo patients having more left PV scar and less non-PV
scar was also seen in the subset of patients with ≥ 6.5% total LA scar
(Table 4B). This may allude to a difference in the efficacy and
durability of scar formation between RF and Cryo. A study by Wieczorek
et al. examining early pulmonary vein reconnection (PVR) showed that
Cryo had less PVR compared to RF ablation 30 minutes post PVI (7% vs.
18%, p < 0.0001).31 Kurose et al. showed
that Cryo resulted in wider and more continuous ablation lesions with
less ablation gaps on 1–3-month post ablation CMR compared to
RF.32 The more extensive and continuous ablation
lesions formed by Cryo may explain the lower rate of PVR observed post
ablation. This study supports the above findings because Cryo had a
greater percentage of PV scar compared to RF, suggesting that Cryo
ablation may more efficiently and uniformly target the PVs while causing
less non-PV scar. The more targeted PV scar formed via Cryo helps
explain the similar rate of AAR between modalities despite RF having a
greater total LA scar percentage.
Initial CMR studies in patients with AF showed that baseline and
residual fibrosis were strong predictors of AAR.16,17However, PVI plus CMR guided fibrosis ablation compared to routine PVI
ablation, evaluated in the ALICIA and DECAAF II trials, did not lead to
improved freedom from AAR.7,33 In this sub-analysis,
despite RF covering a greater percentage of LA fibrosis compared to Cryo
(3.6 +/- 2.2 vs. 3.0 +/- 1.8, p = 0.04), the AAR rates were similar
between groups. Although there appears to be a link between fibrosis and
AF, the underlying pathophysiology of AF is not fully understood. The
arrhythmogenic nature of fibrosis is variable and depends on many
factors. Current imaging software is unable to effectively differentiate
arrhythmogenic from non-arrhythmogenic fibrosis, limiting the utility of
targeting fibrosis with ablation. Moreover, targeting fibrosis with
thermal injury might not be an effective strategy to nullify its
arrhythmogenicity, especially in patients with more advanced/transmural
fibrosis that is difficult to reach. Operators also utilize different RF
ablation techniques even when performing standard PVI, resulting in
variable recurrence rates. On the contrary, Cryo ablation is more
standardized which decreases inter-operator variability.