Author + information
- Received June 4, 2012
- Revision received November 15, 2012
- Accepted November 20, 2012
- Published online April 23, 2013.
- Douglas L. Packer, MD⁎,⁎ (, )
- Robert C. Kowal, MD†,
- Kevin R. Wheelan, MD†,
- James M. Irwin, MD‡,
- Jean Champagne, MD§,
- Peter G. Guerra, MD∥,
- Marc Dubuc, MD∥,
- Vivek Reddy, MD¶,
- Linda Nelson, RN#,
- Richard G. Holcomb, PhD⁎⁎,
- John W. Lehmann, MD, MPH††,
- Jeremy N. Ruskin, MD‡‡,
- STOP AF Cryoablation Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Douglas L. Packer, 2-416 Alfred Building, Saint Mary's Hospital, Mayo Clinic, 1216 Second Street, SW, Rochester, Minnesota 55905
Objectives This study sought to assess the safety and effectiveness of a novel cryoballoon ablation technology designed to achieve single-delivery pulmonary vein (PV) isolation.
Background Standard radiofrequency ablation is effective in eliminating atrial fibrillation (AF) but requires multiple lesion delivery at the risk of significant complications.
Methods Patients with documented symptomatic paroxysmal AF and previously failed therapy with ≥1 membrane active antiarrhythmic drug underwent 2:1 randomization to either cryoballoon ablation (n = 163) or drug therapy (n = 82). A 90-day blanking period allowed for optimization of antiarrhythmic drug therapy and reablation if necessary. Effectiveness of the cryoablation procedure versus drug therapy was determined at 12 months.
Results Patients had highly symptomatic AF (78% paroxysmal, 22% early persistent) and experienced failure of at least one antiarrhythmic drug. Cryoablation produced acute isolation of three or more PVs in 98.2% and all four PVs in 97.6% of patients. PVs isolation was achieved with the balloon catheter alone in 83%. At 12 months, treatment success was 69.9% (114 of 163) of cryoblation patients compared with 7.3% of antiarrhythmic drug patients (absolute difference, 62.6% [p < 0.001]). Sixty-five (79%) drug-treated patients crossed over to cryoablation during 12 months of study follow-up due to recurrent, symptomatic AF, constituting drug treatment failure. There were 7 of the resulting 228 cryoablated patients (3.1%) with a >75% reduction in PV area during 12 months of follow-up. Twenty-nine of 259 procedures (11.2%) were associated with phrenic nerve palsy as determined by radiographic screening; 25 of these had resolved by 12 months. Cryoablation patients had significantly improved symptoms at 12 months.
Conclusions The STOP AF trial demonstrated that cryoballoon ablation is a safe and effective alternative to antiarrhythmic medication for the treatment of patients with symptomatic paroxysmal AF, for whom at least one antiarrhythmic drug has failed, with risks within accepted standards for ablation therapy. (A Clinical Study of the Arctic Front Cryoablation Balloon for the Treatment of Paroxysmal Atrial Fibrillation [Stop AF]; NCT00523978)
Pulmonary vein (PV) isolation is an effective treatment for patients with symptomatic paroxysmal atrial fibrillation (PAF) (1–2). Recent clinical trials have demonstrated that ablation success rates at 12 months are generally in the 60% to 75% range (3–6). Nevertheless, the radiofrequency (RF) ablation used in these trials requires tedious point-to-point delivery of multiple applications to isolate PVs (1,3–7) and is associated with complications such as PV stenosis and atrial-esophageal fistula (8).
Cryoablation balloon therapy has been developed and tested as an alternative, single-delivery approach to isolation of PVs (9). Acceptable success rates with low adverse event rates have been reported in large European observational studies (10–14). Nevertheless, the safety and effectiveness of cryoballoon ablation have not been tested in large randomized clinical trials. The STOP AF trial was undertaken to test the hypothesis that cryoballoon ablation would produce a significantly greater treatment success than antiarrhythmic drug therapy in achieving freedom from AF with an acceptable safety profile.
The STOP AF trial was a prospective, multicenter, randomized, controlled study designed to compare outcomes of cryoablation and antiarrhythmic drug therapies in patients with PAF and was designed in conformance with U.S. Food and Drug Administration (FDA) guidance documents. The study design (Fig. 1) was approved by the institutional review board at each center and the Office of Device Evaluation, FDA. Patients with more than 2 episodes of PAF in the 2 months prior to randomization, for whom at least one membrane-active drug had failed, were eligible for enrollment. Patients were excluded for a left atrium (LA) ≥5.0 cm, a left ventricular ejection fraction <40%, New York Heart Association functional class III or IV congestive heart failure (CHF), coronary artery disease warranting intervention, stroke or transient ischemic attack (TIA) within 6 months, previous LA ablation or surgery for AF, a prosthetic heart valve, amiodarone therapy in the preceding 3 months, more than 2 cardioversions within 2 years, or an implantable rhythm device. After providing written informed consent, patients were randomized 2:1 to receive cryoablation or antiarrhythmic drug therapy with flecainide, propafenone, or sotalol.
The goal of cryoablation was electrical PV isolation of the four major PVs (left inferior [LI], left superior [LS], right inferior [RI], right superior [RS]) as confirmed by entrance and/or exit block. Combined common left PVs were likewise ablated to achieve similar blockage within the constituent PVs. Transseptal catheterization was performed under fluoroscopic or intracardiac echocardiography (ICE) guidance using a standard transseptal sheath, which was increased to a 15-F (Medtronic, Inc., Minneapolis, Minnesota) FlexCath steerable sheath. Thereafter, a 23- or 28-mm (Medtronic, Inc.) Arctic Front cryoablation balloon catheter was advanced over a guidewire to the orifice of the PV. Balloon selection was based on the PV size as established by preprocedure computed tomography (CT) or ICE. Cryoablation was performed in 240-s deliveries, with assessment of entrance and/or exit blockage by using pacing and a circular mapping catheter after a 30-min waiting period. Additional focal cryoablation deliveries were made at investigator discretion, using an 8-mm tipped cryoablation catheter (Freezor MAX cryoablation catheter; Medtronic) if isolation was not achieved by one or more cryoballoon ablations or if required for the ablation of focal triggers. Cavotricuspid isthmus ablation using this catheter was undertaken in patients with prior clinical or inducible cavotricuspid isthmus-dependent atrial flutter, with bidirectional block required. At the completion of the procedure, pacing was used to demonstrate entrance and, where assessable, exit block; isoproterenol use was optional, and adenosine was not used to confirm the block.
Ablation was undertaken with the patient receiving heparin, with an activated clotting time ≥300 s. Sedation and/or anesthesia and reversal of anticoagulation were performed according to the site's standard practices. Prior to discharge, cryoablation patients underwent repeated inspiratory and expiratory chest radiographic examinations to screen for phrenic nerve palsy. Anticoagulation with warfarin to a target international normalization ratio (INR) of 2.0 to 3.0 was required for the first 3 months after ablation. Subsequently, warfarin was discontinued at the investigator's discretion, as guided by published clinical guidelines (15). During the 90-day blanking period, patients could be empirically treated with an FDA-allowed antiarrhythmic study drug (flecainide, propafenone, or sotalol), after which the drug was discontinued. One repeat cryoablation procedure was permitted during this blanking period, with the recommendation that all repeated ablations use the same cryoablation method. Nevertheless, the final decision about the mode of reablation was left to the investigator. Use of RF energy for reablation was taken as a treatment failure.
Patients randomized to receive drug therapy received flecainide, propafenone, or sotalol, if they had not previously experienced failure with these drugs. Drug dosages were optimized within approved labeling according to the 2006 AHA/ACC guidelines for treatment of AF (15). If necessary, a change to one of the other three drugs was allowed. Once the patient was stabilized, the drug therapy was maintained throughout the study. AF drugs such as dofetilide and amiodarone were precluded by the FDA, as they had not previously received approval for treatment of PAF. Patients were allowed to cross over to cryoablation treatment only after meeting protocol-defined effectiveness failure endpoints. AF documentation during the blanking period was required prior to crossing over to an alternate treatment strategy.
All patients, including those drug patients who crossed over to ablation, were followed for 12 months after their therapy initiation date. Each patient underwent repeated assessments at 1, 3, 6, 9, and 12 months. Occurrences of AF during the blanking period were not counted against the primary objective, nor considered to be chronic treatment failures. Personal trans-telephonic monitoring (TTM) systems were provided for weekly scheduled transmissions and recordings at the occurrence of arrhythmia symptoms. Twenty-four-hour Holter monitoring was required at 6 and 12 months. Repeated CT or magnetic resonance (MR) imaging studies of PV anatomy were conducted at baseline, 6, and 12 months and read, as were electrocardiography (ECG) strips, by independent ECG and imaging core laboratories. All clinical adverse events and trial endpoints were reviewed and adjudicated by an independent clinical events committee. Screening for National Institutes of Health Stroke Score was performed at each visit; with potential strokes/TIAs reviewed by the study neurologist. Compliance was high for follow-up visits (>95%), weekly TTMs (>90%) and Holters (>90%), postprocedural chest radiography (>95%), and CT/MRI studies (>95%).
The primary effectiveness endpoint for the trial was freedom from chronic treatment failure, as defined by the absence of: 1) any detectable AF after the blanking period; 2) use of a nonstudy, antiarrhythmic drug; or 3) any nonprotocol intervention for AF (i.e., RF ablation). Freedom from AF after ablation while being treated with a previously ineffective antiarrhythmic drug at the same or a lower dose was considered a treatment success if patients remained in sinus rhythm.
Two inferential coprimary safety endpoints were evaluated for study success: 1) the proportion of intent-to-treat-ablated patients with more than 1 cryoablation procedure-related event (CPE); and 2) an intent-to-treat comparison of the freedom from major AF events (MAFE) between groups over 12 months of follow-up. CPEs were defined as device- and procedure-related serious adverse events in the following categories: access site complications, cardiac damage (including myocardial infarction [MI]), embolic complications (including stroke), arrhythmias, persistent phrenic nerve palsy, PV stenosis, and death. MAFEs were defined as nonprocedure-related serious adverse events including: cardiovascular death, hospitalization for AF recurrence or ablation, atypical atrial flutter ablation, systemic embolization, CHF, nonstroke hemorrhagic events, MI, stroke, or antiarrhythmic drug initiation, adjustment or complications requiring hospitalization. In addition, the overall safety profile of the cryoablation procedure was reported as event rates for all ablated patients (“on treatment”) including those crossing over from drug therapy.
The results of continuous variables are given as mean ± SD. Comparisons were made using paired t or Kruskal-Wallis testing as appropriate for the distribution of values. Categorical variables were compared using exact binomial or chi-square analyses. Long-term outcomes were expressed using Kaplan-Meier statistics, with the significance of differences in values indicated by log-rank testing. Differences between groups were also established using proportional hazards models. The primary effectiveness endpoint (proportion of patients with treatment success at 12 months) was assessed using a 2-sided Fisher's exact test of binomial proportions. The proportion of cryoablation patients with freedom from cryoablation procedure events was compared to a fixed proportion of 14.8%, using an exact 1-sided binomial test, and the proportion of patients with MAFEs in the cryoablation group was compared with those in the drug-treated group, using a noninferiority margin of 10%. A p value of <0.05 was viewed as significant.
A total of 245 patients were randomized and enrolled over 21 months; 163 patients were assigned to cryoablation and 82 to antiarrhythmic drugs. Thirty-one cryoablation patients underwent a repeat cryoablation during the blanking period. Sixty-five drug-treated patients crossed over to cryoablation after recurrent AF, and 3 drug-treated patients were lost to follow-up.
General patient characteristics were similar in both treatment arms (Table 1). Randomized patients were highly symptomatic with 23.2 and 21.2 self-reported AF episodes and 14.4 and 11.1 episodes of rapid heartbeat in the ablated and drug arms, respectively, in the 2 months prior to enrollment. Prior cardioversion and a history of atrial flutter in the ablated and drug arms were similar. The median number for whom antiarrhythmic drugs failed prior to enrollment was 1.3 in the cryoballoon-treated arm and 1.4 in those receiving antiarrhythmic drug therapy (p = NS). These patients were also at low risk of stroke as judged by congestive heart failure (C), hypertension (H), age 75 or older (A), diabetes (D), and previous stroke (S2) (CHADS2) score of 0.6 to 0.7 (Table 1), without differences between groups.
Effectiveness outcome: acute cryoablation success
In the 163 intent-to-treat patients randomized to ablation, 655 PVs were cryoablated at the initial procedure. In 160 (98.2%) patients, 3 or more PVs were isolated as manifested by entrance and/or exit the block. All 4 major PVs were isolated in 97.6% of patients, as were 21 of 21 left common (LC) PVs and 10 of 13 right middle (RM) PVs. In 83% of patients, cryoballoon intervention alone was sufficient for PV-isolating ablative therapy, with a total of 12.5 ± 0.31 applications required for all PVs. The average cryoballoon application time was 214.4 ± 1.5 s/application. Cryoballoon temperature averaged −51.0°C ± 0.4°C.
Twenty-seven patients required additional catheter-based focal cryoablation in one or more PVs and at non-PV sites in 8 (4.9%) patients; the need for focal cryoablation was not significantly associated with the use of a particular size cryoballoon. A 23-mm balloon was used in 61.4% of the veins, while a 28-mm balloon was applied in 58.5%. Both 23- and 28-mm balloons were used in 20.8% of PVs. In the 19 PVs where isolation could not be achieved by cryoballoon ablation alone, an average of 2.2 cryoablations were attempted, including in one LC, 3 RMs, 4 LI (mean diameter 12.3 mm), 5 LS (mean diameter 14.4 mm), 3 RI (mean diameter 16.4 mm), and 3 RS (mean diameter 21.5 mm) PVs. Twelve of the nonisolated PVs were initially treated with a 23-mm cryoballoon and seven with a 28-mm cryoballoon. In each case where the PV could not be isolated, the cryoballoon used was at least 2 mm larger than the diameter of the PV.
The mean procedure duration including all repeat PV assessment was 371 min, fluoroscopy exposure averaged 63 min, and total cryoablation time averaged 66 min. Over the course of the trial, the percentage of patients undergoing ablation with a 28-mm balloon remained consistent. The cavotricuspid isthmus was cryoablated in 66 (40.5%) patients, with bi-directional block achieved in 64 (97.0%).
Long-term success: cryoablation group by intention to treat
After 12 months of follow-up, freedom from chronic treatment failure was seen in 114 of 163 (69.9%) cryoballoon-ablated patients (Fig. 2). This included 13 (8.0%) patients for whom previous drug treatment was ineffective at the same or received a lower dose after the blanking period. Four patients were receiving flecainide, 200 mg daily, 4 were receiving propafenone, 450 to 675 mg daily, and 5 patients were receiving sotalol, 30 to 240 mg daily. The single procedure success rate was 57.7%, including the 13 patients (8%) receiving previously ineffective drugs as permitted by the protocol (9), in whom ultimate success was seen in 69.9%, including the 31 patients (19%) who underwent repeated cryoablation during the blanking period. Only 6 (7.3%) patients randomized to the antiarrhythmic drug treatment arm remained free from chronic treatment failure. The difference between ablation- and drug-treated patients was highly significant (p < 0.001) (Fig. 3A).
Of 163 patients, 121 (74.2%) received either flecainide, propafenone, or sotalol immediately after cryoablation, but these drugs were discontinued by the end of the blanking period, as stipulated by the protocol. At the end of 12 months, only 26% of ablation patients were taking antiarrhythmic drugs; 95% of patients were treated with warfarin upon enrollment, with only 24% of patients continuing this medication at 12 months of follow-up. The CHADS2 score was not different between those who stopped and those who remained taking warfarin.
Symptomatic AF occurrence fell from 100% at baseline to 19.0% at 12 months. Arrhythmia-related symptoms were dramatically reduced in ablation patients by 12 months of follow-up: AF symptoms (100% to 20%), dizziness (48% to 9%), palpitations (86% to 25%), and fatigue (76% to 13%). This symptomatic improvement was confirmed by improved SF-36 quality of life subscores (16).
Drug treatment control group by intention to treat
Drug-treated patients were tried on one or more drug therapies during the blanking period, using flecainide (65%), propafenone (57%), and/or sotalol (44%). Over the course of the study, 65 of the 82 drug-treated patients (79%) crossed over to ablative intervention. Eight (10%) highly symptomatic drug-treated patients crossed over within the blanking period. Seventeen patients crossed over at between 3 and 6 months of follow-up, and 40 additional patients crossed over after the 6-month follow-up visit. All drug-treated patients who crossed over during follow-up experienced confirmed AF drug treatment failure requiring ablative intervention. The median time to cross over was 186 (range 29 to 364) days. Only 6 (7.3%) patients randomized to the antiarrhythmic drug treatment arm remained free from chronic treatment failure.
Long-term outcomes: On-treatment analysis
Including the original randomized group of 163 patients and the 65 crossover patients for whom early drug treatment failed, 228 patients underwent cryoballoon ablation, providing a more complete “on-treatment” analysis. The drug-treated, cross-over patients were then followed for an additional 12 months following cryoablation. A post-hoc analysis of freedom from AF beginning 90 days after cryoablation allowed comparison of the intent-to-treat ablation patients with the cryoablated drug patients as seen in Figure 3B.
Crossover drug-treated patients (n = 65) had an estimated rate ± standard error of the mean (SEM) of freedom from any postblanking AF following cryoablation of 61.6 ± 7.3%. Intent-to-treat ablation patients (n = 163) had an estimated rate of freedom from any postblanking AF following first cryoablation of 63.7 ± 4.1%. This difference was not significant (p = 0.9107, log-rank test).
Safety outcomes: safety endpoints by intention to treat
Both prespecified primary safety endpoints were met. Serious CPE occurred in 5 (3.1%) patients undergoing cryoablation, with an upper confidence bound (UCB) of 6.3%, which was well below the pre-specified UCB of 14.8% (p < 0.001). One patient with pre-existing coronary artery disease (CAD) and coronary artery bypass graft sustained a periprocedural non–Q-wave MI, which was attributed to anesthesia-induced hypotension; 1 patient developed tamponade; 2 patients developed symptomatic PV stenosis where intervention was recommended; and 1 patient had atrial flutter requiring prolonged hospitalization.
MAFEs were seen in 5 (3.1%, UCB 7.0%) cryoablation patients and 7 (8.5%, UCB 16.8%) drug-treated patients (noninferiority, p < 0.001). Five cryoablation patients experienced MAFEs: 1 (0.6%) patient sustained an unrelated fatal MI at 10 months; 1 patient had Wegener's-related hemoptysis, AF recurrence and hospitalization for antiarrhythmic drug adjustment; 1 patient had a subarachnoid hemorrhage; 1 patient had intestinal bleeding accompanying an elevated INR; and 1 patient was hospitalized with AF-related CHF. There were 7 drug-treated patients who experienced 12 MAFEs, including hospitalizations for AF recurrence or ablation (5 of 82, 6.1%), antiarrhythmic therapy adjustment (4 of 82, 4.9%), or treatment for atrial flutter (1 of 82, 1.2%). Two (2.4%) drug-treated patients had gastrointestinal or subdural bleeding complications.
The combined rate of cryoablation procedure events and MAFEs in the ablation arm was lower than the MAFE rate in drug-treated patients: 6.1 versus 8.5%, respectively (p < 0.001) (Fig. 4). The overall serious adverse event rates over 12 months were not different in the two treatment arms: 12.3% in cryoablation and 14.6% in drug-treated patients (p = 0.688).
PV stenosis, defined as a reduction of >75% in cross-sectional area (approximately a 50% reduction in diameter), occurred in 5 of 163 (3.1%) cryoablated patients in the original randomized treatment arm. In subsequent on-treatment analysis including the 65 patients who crossed over from drug to ablation treatment, 7 (3.1%) of the 228 cryoablated patients (10 PVs) developed stenosis. Stenotic veins included 5 LIPVs, 4 LSPVs, and 1 RIPV. A 23-mm cryoballoon alone was used exclusively in 4 PVs, and both 23- and 28-mm cryoballoons were used in 6 PVs. PV stenosis occurred in 2.1% (4 of 186) of patients with only a single cryoablation procedure and in 6.5% (2 of 31) of patients with 2 cryoablation procedures; 1 additional patient had PV stenosis after 2 procedures: the first with cryoablation and the second with RF ablation. Five patients were asymptomatic throughout follow-up, and 2 patients developed increasing dyspnea, which resolved in 1 patient after PV stenting. A second patient, refused intervention and was asymptomatic at 12 months of follow-up.
Phrenic nerve paralysis
Phrenic nerve paralysis was assessed by inspiration/expiration chest radiography after 259 cryoablation procedures in 228 patients. Paralysis was documented after 29 of 259 (11.2%) procedures. Slightly more than half the affected patients were asymptomatic; 1 subject had paralysis following both the initial and repeat cryoablations. Of the 28 patients with phrenic nerve injury, 25 showed complete radiographic resolution over a mean 144 ± 27 days. Four patients had persistent radiographic abnormalities at 12 months of follow-up, 1 of whom had ongoing symptoms of mild cough and dyspnea. Of the 29 procedures resulting in phrenic nerve paralysis, the injury was related to a 23-mm cryoballoon in 17 cases, a 28-mm balloon in 9 cases, with both sizes in 2 cases, and with only the Freezor MAX catheter in one reablation case.
Stroke occurrence/neurologic events
Stroke occurred in 5 of 228 (2.2%) randomized and crossover cryoablated patients. One procedure-related stroke occurred on the day of intervention in a patient crossing over from drug to cryoablation therapy, which was the only procedure- related stroke event (0.4%, 1 of 228). An additional 4 patients sustained nonprocedure-related cerebral vascular events over follow-up. In 1 postablation patient, a small hemorrhagic stroke occurred on day 183 of follow-up, 5 days after RF ablation for atrial flutter. A second patient demonstrated a lacunar infarct of indeterminate age on CT examination on day 51. An additional patient had transient bilateral visual disturbance, occurring one month after cryoablation, with no abnormalities by CT scan. Subarachnoid hemorrhage occurred on day 260 in an additional patient on aspirin. Symptom severity was mild in 3 and severe in 2 at the onset of the event. All patients recovered without residual neurological deficits.
Four additional patients sustained apparent TIAs. One occurred 3 days after ablation. Dizziness occurred in a second patient on the day of ablation, with full resolution of symptoms within several minutes. Transient left lower arm and hand numbness was noted in a fourth patient 19 days after ablation. Temporary amaurosis fugax was experienced 9 days after ablation in a crossover patient.
Other adverse events
No patient developed an atrial-esophageal fistula. Myocardial perforation and vascular injury were uncommon. A femoral arteriovenous fistula occurred in 2 patients, and 2 others developed pseudo-aneurysms. Postprocedure cough developed following 44 of 259 (17%) procedures, which resolved completely in 91% of patients by the end of the study. The median time to recovery in 33 patients was 39 (range 0 to 272) days. Pericardial chest pain following ablation was uncommon. Serious adverse events occurred in 34 of 228 (14.9%) cryoablated patients: 11 (4.8%) patients had one or more serious adverse events that were either device- or procedure-related. The safety outcomes of all 228 patients treated with cryoballoon ablation are provided in Table 2.
The Arctic Front STOP AF trial demonstrates a persistent benefit of cryoballoon-based ablation for PAF, with an effectiveness rate that was significantly greater than that seen with antiarrhythmic drug therapy over 12 months of follow-up in highly symptomatic patients with AF, failing treatment with at least one antiarrhythmic drug. This study also demonstrates a low serious adverse event rate, although minor adverse events did occur. Nevertheless, PV stenosis, phrenic nerve injury, and stroke remain potential complications, as detected by complete screening in this trial.
Effectiveness of cryoablation
The 69.9% treatment success rate was higher than that predicted by a recent meta-analyses of randomized and observational studies (17) of standard RF ablation, suggesting the potential for simplification of the ablation process by eliminating the need for point-to-point manipulation of the catheter. Overall outcome is similar to that suggested by a recent meta-analysis of cryoballoon studies (14), as accomplished with acceptably low redo rates. Randomized cryoablation versus RF ablation studies will be required to more directly establish the comparative impact of energy type, investigator experience, underlying disease, AF type, and the comparative requirement of additional LA or RA lesions beyond PV isolation alone (18–20).
Procedure and cryoablation delivery times were longer than seen in some prior studies of RF or cryoablation (7,10–13), but are similar to the average times seen in earlier, observational, cryoablation studies (14). This reflects the inclusion of all patients from the beginning of an individual operator's experience with cryoballoon technology. Protocol-mandated study components, particularly the 30-min assessment period at the end of the ablation, also contributed to lengthy procedure times. With incremental experience, the procedure and cryoablation delivery times decreased, and single procedure success rates increased, reflecting an expected learning curve.
The efficacy of cryoablation was also demonstrated by the ability to discontinue antiarrhythmic drug and warfarin therapy in three quarters of these patients over follow-up. Nevertheless, these data from low-risk patients do not necessarily apply to patients with more persistent AF or higher disease-dependent stroke risk. In these patients AHA/ACC guidelines continue to provide a reasonable basis for guiding ongoing antithrombotic therapy after ablative intervention (15).
Antiarrhythmic drug failure
Patients treated with antiarrhythmic drugs had high early AF recurrence rates, which is not surprising given the frequency of baseline AF. Aggressive follow-up monitoring may have detected more recurrences than seen in other trials. Furthermore, almost all of these patients were treated with alternative antiarrhythmic drug therapy during a comparable blanking period, and 89% of drug-treated patients completed the blanking period without crossing over. This allows a reasonable comparison of antiarrhythmic drug and cryoablation effectiveness based on equivalent blanking periods, unlike other past trials using different ablation and drug treatment arm blanking periods (7).
Safety outcomes of cryoballoon and drug therapy for AF
The prespecified safety endpoints in STOP AF, defined as low rates of cryoablation procedure events and major AF events, were both met. The presentation of all adverse events in Table 2 occurring over a total of 12 months follow-up provides a more comprehensive assessment of the safety profile of this therapy than previously available. When viewed within this context, the event rates in this trial are as low as seen in recently reported AF ablation studies (7,10–14).
Pulmonary vein stenosis
Although other studies of cryoballoon catheter ablation have suggested little to no PV stenosis risk (9), this study demonstrates the possibility of this adverse event, as does an additional case of cryoablation-related PV stenosis reported by others (21). The use of PV cross-sectional area in the primary analysis is a sensitive measure, which overstates the degree of clinically relevant PV stenosis. Other clinical reports (2) have defined PV stenosis as a >70% to 75% reduction in vein diameter, rather than the cross-sectional area used in this trial. The occurrence of stenosis appears related to ablation within the vein, or due to repeated interventions at that site.
Atrial-esophageal fistula formation was not seen in any patient in this study, possibly reflecting differences in the histopathology of cryothermy with infrastructural sparing, compared with that seen with RF energy (22–23). Esophageal ulcers with cryoablation energy delivery have been reported (22, 24), which resolved spontaneously or with protein pump inhibitor therapy, and recently a case of atrial/esophageal fistula was reported after pulmonary vein cryoablation using a 23-mm balloon (25). There does not appear to be a relationship between ulcer formation and a true atrial-esophageal fistula.
Stroke may occur after cryoballoon ablation as evidenced by the single procedure-related stroke in a patient crossing over from drug to ablative therapy. The other 4 neurological events occurred between 1 and 8 months following cryoablation, suggesting a closer relationship with the natural history of AF and anticoagulant therapy than the ablation procedure itself, although the CHADS2 scores in these patients were <1.
Phrenic nerve injury
Phrenic nerve injury has been previously described in association with cryoballoon ablation (19,20). This has been observed in 11.2% of original and cross-over ablation procedures in this study but was largely reversible (19). Similar transient phrenic nerve injury rates from 8% to 11% have been reported in other early European studies (13,19,26). Because the protocol mandated follow-up inspiratory and expiratory chest radiographs in all patients, detection of phrenic nerve injury was maximized in this study. The mechanism of the occurrence of phrenic nerve injury is probably related to positioning the balloon further into the RS or RI PVs, as the phrenic nerve may run behind the right atrium. Undersizing of the balloon to a 23-mm versus a 28-mm balloon, relative to the size of the vein could contribute to this complication (14,19), as could PV distortion with balloon inflation (27). It remains unclear whether phrenic nerve pacing would be effective in preventing this complication (19,27–29), although early detection with phrenic nerve pacing may be expected to accelerate recovery of the nerve, if injured.
Other cardiac and vascular events
Other anticipated adverse events as shown in Table 2 were similar to those expected with other catheter-based ablative interventions (30–32). Pericardial pain typically observed after RF ablation was very uncommon in these patients, suggesting less pericardial injury with cryoablation. The mechanism of postprocedure cough is unclear but suggests airway irritation that is reversible over time. Whether other mediators are involved is unclear. It is impossible to know whether this is related to cryothermy itself, although direct cryo bronchial injury has been observed on bronchoscopy in a single patient undergoing AF ablation (33).
Although this is the largest randomized evaluation of cryoballoon ablation reported to date, there are several important limitations to be considered. First, the trial is limited by the frequency and timing of crossovers from drug treatment to cryoablation. While this does not negate the effectiveness of cryoablation, it renders it more difficult to directly compare ablation and drug therapy beyond the median 186 days to cross-over. Nevertheless, the number of drugs previously failed and the median doses taken were equivalent in both groups. All other demographics were also similar, suggesting the appropriateness of a comparison of these patients. The limited choice of antiarrhythmic drugs may also have contributed to the high AF recurrence rate observed in the drug-treated patients, although agents used were similar to those in other recent clinical trials. The use of dofetilide, dronedarone, or amiodarone, not permitted by the FDA in this study, might have increased the success rate of drug therapy.
The STOP AF trial demonstrates that cryoballoon ablation is effective in preventing recurrent, symptomatic, paroxysmal AF in patients who are resistant to at least one antiarrhythmic drug. PV isolation can be achieved in most patients using the cryoballoon alone. Nevertheless, like other treatment modalities, complications affecting PVs and the phrenic nerve are possible.
For a list of study participants, please see the online version of this article.
The STOP AF study was funded by Medtronic, Inc. (which purchased Cryocath over the course of the study). Dr. Packer receives research funding from Biosense Webster, Boston Scientific/EPT, Endosense, EpiEP, EP Advocate, Medtronic CryoCath LP, Minnesota Partnership for Biotechnology and Medical Genomics/University of Minnesota, National Institutes of Health, St. Jude Medical, Siemens AcuNav, and Thermedical (EP Ltd.); is a consultant without personal compensation for Abiomed, Biosense Webster, Inc., CardioFocus, Cardiomedics, Cyberheart, Endosense, Johnson & Johnson Healthcare Systems, Medtronic/CryoCath, OrthoMcNeill, Sanofi-Aventis, St. Jude Medical, Siemens AG, and Valencia Technologies; and receives royalties from St. Jude Medical and Blackwell Publishing. Dr. Kowal is a Medtronic advisor; and teaches at Arctic Front Programs. Dr. Wheelan receives honoraria as a Medtronic consultant and has stock ownership. Dr. Irwin contracts with Medtronic to train other EPs in the use of Arctic Front. Dr. Guerra is a Sanofi-Aventis board member and receives payment for lectures, including service on speakers bureaus at St. Jude Medical and Medtronic. Dr. Dubuc received a research grant, consulting fees and/or honoraria, and payment for lectures, including service on speakers' bureau, from Medtronic/CryoCath. Dr. Reddy is a consultant for Medtronic. Dr. Lehmann is an hourly consultant with Medtronic for STOP AF clinical trial design and implementation and holds patents assigned to Medtronics, although no royalties or benefits are received. Dr. Ruskin receives research grants from Medtronic/CryoCath, Biosense Webster, Boston Scientific, and St. Jude Medical; is an advisory board member for Medtronic/CryoCath and CardioFocus (no personal compensation) and a board member for Pfizer, Biosense Webster, CardioInsight Scientific, and Sequel (no personal compensation); a compensated consultant with Astellas/Cardiome, GE Healthcare, Sanofi-Aventis, Medtronic, Portola, MedIQ, and Third Rock Ventures; and receives honoraria for lectures from Med-IQ and holds stock/options with Portola. Dr. Nelson is an employee of Medtronic, Inc. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- American Heart Association/American College of Cardiology
- coronary artery disease
- congestive heart failure, hypertension, age, diabetes and stroke scale
- congestive heart failure
- cryoablation procedure event
- computed tomography
- Food and Drug Administration
- intracardiac ultrasound
- international normalization ratio
- left atrium
- left common
- left inferior
- left superior
- major atrial fibrillation event
- myocardial infarction
- magnetic resonance
- paroxysmal atrial fibrillation
- pulmonary vein
- right inferior
- right middle
- right superior
- Short Form-36
- transient ischemic attack
- trans-telephonic monitoring
- upper confidence bound
- Received June 4, 2012.
- Revision received November 15, 2012.
- Accepted November 20, 2012.
- American College of Cardiology Foundation
- Haissaguerre M.,
- Jais P.,
- Shah D.,
- et al.
- Packer D.
- Stabile G.,
- Bertaglia E.,
- Senatore G.,
- et al.
- Jais P.,
- Cauchemez B.,
- Macle K.,
- et al.
- Pappone C.,
- Giuseppe A.,
- Simone S.,
- et al.
- Pappone C.,
- Oral H.,
- Santinelli V.,
- et al.
- Sarabanda A.,
- Bunch T.,
- Packer D.L.,
- et al.
- Neumann T.,
- Vogt J.,
- Schumacher B.,
- et al.
- Van Belle Y.,
- Janse P.,
- Theuns D.,
- et al.
- Malmborg H.,
- Lonnerholm S.,
- Blomstrom-Lundquist C.
- Fuster V.,
- Ryden L.,
- Cannom D.,
- et al.
- Linhart M.,
- Bellmann B.,
- Mittmann-Braun E.,
- et al.
- Chun K.,
- Schmidt B.,
- Metzner A.,
- et al.
- Ahmed H.,
- Neuzil P.,
- d'Avila A.,
- et al.
- Evonich R.,
- Nori D.,
- Haines D.
- Ripley K.,
- Gage A.,
- Olsen D.,
- et al.
- Mahapatra S.,
- Peterson L.,
- Packer D.,
- et al.
- Calkins H.,
- Brugada J.,
- Packer D.,
- et al.