OUP user menu

Efficacy of adjunctive ablation of complex fractionated atrial electrograms and pulmonary vein isolation for the treatment of atrial fibrillation: a meta-analysis of randomized controlled trials

Melissa H. Kong, Jonathan P. Piccini, Tristram D. Bahnson
DOI: http://dx.doi.org/10.1093/europace/euq384 193-204 First published online: 30 October 2010


Aims Although useful, percutaneous left atrial ablation for pulmonary vein isolation (PVI) does not eliminate atrial fibrillation (AF) in all patients. The ablation of complex fractionated atrial electrograms (CFAEs) has been proposed as an adjunctive strategy to improve the maintenance of sinus rhythm after PVI. Our objective was to analyse the efficacy of adjunctive CFAE ablation.

Methods and results We meta-analysed six randomized controlled trials (total, n= 538) using random-effects modelling to compare PVI (n= 291) with PVI plus CFAE ablation (PVI + CFAE) (n= 237). The primary outcome was freedom from AF or other atrial tachyarrhythmias (ATs) after a single ablation with or without antiarrhythmic drugs. Following a single ablation, PVI + CFAE improved the odds of freedom from any AF/AT compared with PVI alone (odds ratio 2.0, 95% confidence interval 1.04–3.8, P = 0.04) at ≥3-month follow-up. There was moderate heterogeneity among trials (I2= 63.0%). Complex fractionated atrial electrogram ablation significantly increased mean procedural (178.5 ± 66.9 vs. 331.5 ± 92.6 min, P< 0.001), mean fluoroscopy (59.5 ± 22.2 vs. 115.5 ± 35.3 min, P< 0.001), and mean radiofrequency (RF) energy application times (46.9 ± 36.6 vs. 74.4 ± 43.0 min, P< 0.001).

Conclusions Pulmonary vein isolation followed by adjunctive CFAE ablation is associated with increased freedom from AF after a single procedure. Adjunctive CFAE ablation increased procedural, fluoroscopy, and RF application times, and the risk/benefit profile of adjunctive CFAE ablation deserves further evaluation with additional studies and longer-term follow-up.

  • Atrial fibrillation
  • Catheter ablation
  • Meta-analysis
  • Complex fractionated electrograms
  • Pulmonary vein isolation


Catheter ablation of atrial fibrillation (AF) is an evolving treatment strategy for the maintenance of sinus rhythm, offering a reasonable therapy alternative for patients with symptomatic, drug-refractory AF. Recent reports suggest that 65–75% of patients are free of AF at 1 year after a single procedure.110 New catheter ablation strategies include ablation for atrial substrate modification in addition to targeting and eliminating AF triggers associated with the pulmonary veins.11 The ablation of complex fractionated atrial electrograms (CFAEs) in addition to PVI has been proposed as one such strategy to improve the maintenance of sinus rhythm.12 Whereas CFAE ablation as a sole primary therapy has not demonstrated reproducible efficacy,13,14 many experts advocate a combined approach that eliminates AF triggers associated with the pulmonary veins via pulmonary vein isolation (PVI) together with additional ablation targeting CFAEs to modify the atrial substrate for AF.

Several observational studies have demonstrated improved outcomes with adjunctive CFAE ablation, particularly for paroxysmal AF; however, reported successes with this approach are not uniform.1520 Specifically, several randomized clinical trials of PVI vs. PVI plus adjunctive CFAE ablation (PVI + CFAE) published in the last year report results with relatively small sample sizes consisting of different patient populations and using various laboratory techniques.2125 As such, the relative efficacy of PVI + CFAE ablation compared with PVI alone remains unclear. In addition to the limited availability of outcomes information on adjunctive CFAE ablation from prospective multicentre randomized trials, the strategy of targeting CFAEs during left atrial ablation has been reported in some studies to be associated with increased rates of post-ablation organized atrial tachyarrhythmias (ATs) ranging from 10 to 25%.13,18,24 The goal of this study is to synthesize the available prospective randomized data comparing PVI alone to PVI plus adjunctive CFAE ablation to help evaluate the risks and benefits of adjunctive CFAE ablation compared with stand-alone PVI for the treatment of drug-refractory AF.


Search strategy

A systematic search of MEDLINE was conducted with the following terms: ‘atrial fibrillation’ AND ‘fractionated electrograms'. The same query was conducted in (i) Cochrane database, (ii) FDA web-portal, and (iii) clinicaltrials.gov. The bibliographies of identified studies were reviewed for additional studies meeting inclusion criteria.

Eligibility and data abstraction

We included randomized controlled trials (RCTs) in which patients with AF referred for catheter ablation were randomized to undergo stand-alone PVI vs. PVI + CFAE ablation. Additional inclusion criteria included follow-up ≥3 months, full-length peer-reviewed publication in English, and a study population >18 years old.

All studies that met the inclusion and exclusion criteria were independently reviewed with standardized data abstraction by two investigators (M.H.K. and J.P.P.). A third investigator (T.D.B.) adjudicated any discrepancies. The results of the MEDLINE query included those reports identified by other search methods (www.fda.gov, clinicaltrials.gov, Cochrane database, and bibliographies). Abstracted data included eligibility criteria, baseline characteristics, study design (including treatment and control arms), follow-up, and outcomes.

The primary outcome of interest was freedom from AF or AT after a single ablation procedure with or without the ongoing use of antiarrhythmic drugs (AADs). When available, data were also abstracted on pre-specified secondary measures of the efficacy and safety of the catheter ablation procedure including: repeat ablation; incidence of post-procedure ATs; procedural times [total, fluoroscopy, and radiofrequency (RF) energy application]; complications (pericardial effusion, pericarditis, pulmonary vein stenosis, and vascular/bleeding complications); and post-ablation AAD use. Outcomes were analysed by the intention-to-treat principle.

Statistical analysis

Random-effects modelling according to the method of DerSimonian and Laird26 was used to evaluate the effects of CFAE ablation on the primary outcome. Sensitivity analyses were performed based on AF subgroups (paroxysmal vs. persistent) as well as serial exclusion of each individual trial. The patient was the individual unit of analysis (as opposed to person years) and the measure of treatment effect for the primary endpoint was reported by odds ratios (OR) with 95% confidence intervals (CI). Heterogeneity between studies was determined using the I2-statistic. All analyses were conducted using Comprehensive Meta-Analysis SoftwareTM (BIOSTAT, Englewood, NJ, USA).

Baseline characteristics and secondary outcomes of the stand-alone PVI group were compared with the PVI + CFAE group. Categorical variables are presented as frequencies and percentages and continuous variables are presented as means for central tendency measurement and standard deviation for variance reporting. Categorical measures were compared using Fisher's exact test and continuous data were compared using Student's t-test. Statistical testing was two-tailed and statistical significance was declared at P < 0.05.


Search results

After searching MEDLINE, the Cochrane database, fda.gov, clinicaltrials.gov, and the bibliographies of expert consensus documents,27,28 we identified 136 abstracts and 4 registered clinical trials, which were reviewed for inclusion and exclusion criteria (Figure 1). Among this group of abstracts, 129 were excluded for reasons listed in Figure 1. For the remaining seven studies, full manuscripts were reviewed and six trials were included in the analysis. One recently published trial was excluded because patients were not randomized to receive CFAE ablation, but were randomized to circumferential PVI (CPVA) with possible supplemental CFAE ablation vs. potential-guided antral PVI (APVI) and as such, there was no way to isolate the treatment effect of adjunctive CFAE ablation for evaluation.29

Figure 1

QUOROM flow diagram for the meta-analysis.

Trial characteristics and study quality

We identified six RCTs of PVI with and without adjunctive CFAE ablation for inclusion, which enrolled 538 patients (Table 1). Four of the studies were published since 2009.21,22,24,25 Two studies enrolled patients with long-standing persistent AF,23,24 three studies enrolled patients with paroxysmal AF,21,22,30 and the most recently published study enrolled both patients with paroxysmal or persistent AF. Of note, the Oral et al.30 study from 2004 only included the subgroup of paroxysmal AF patients who had induction of AF with rapid pacing after CPVA. Although all studies examined the primary endpoint of freedom from AF/AT, post-procedure follow-up periods (3 to 16.4 ± 1.3 months) and post-procedure AAD use varied widely (Table 2).

View this table:
Table 1

Summary of RCTs included in meta-analysis

StudyJournalYearCentreTotal (n)Inclusion criteriaExclusion criteriaControl armSpecified primary endpoint
Deisenhofer et al.21J Cardiovasc Electrophysiol2009Germany98Age 18–80 years; symptomatic PAF with episodes lasting up to 7 days and with ≥4 episodes per month; failed therapy with ≥1 Class I or III AADIntracardiac thrombi documented by TEE; LVEF <35%; history of MI or cardiac surgery in previous 3 months; prior ablation procedures for AFSegmental PVI aloneFreedom from AT >30 s duration on a 7-day Holter 3 months post-procedure in combination with freedom from symptomatic AF/AT 3-months post-ablation
Di Biase et al.22Circ ArrhythmElectrophysiol2009Multicentre103PAF 1 year; refractory to ≥2 AADs; presented to EP lab with spontaneous AFPrior ablationAPVI aloneFreedom from AF defined as no episodes of AF/AT with or without AADs that lasted more than 1 min at 1-year follow-up
Elayi et al.23Heart Rhythm2008Multicentre144Long-standing history of AF; failed to maintain SR with ≥2 AADs; AF duration >1 year with ≥1 failed DCCV after >1 week on AADs referred for first-time ablation from January 2005 to February 2006Paroxysmal or persistent AF; <18 years old; >80 years old; previous RF ablation in left atrium or surgical MazeAPVI or CPVAFreedom from AF/AT defined as no episodes of AF/AT >1 min during follow-up. AF/AT occurring during 2-month blanking period after first procedure while patient was on AADs were not considered recurrence
Oral et al.30Circulation2004Ann Arbor, MI, USA60AF inducible by atrial pacing after left atrial circumferential ablation for symptomatic, drug-resistant PAFNRCPVAFreedom from AF defined at 6 months without AAD
Oral et al.24J Am Coll Cardiol2009Ann Arbor, MI, USA100Long-lasting persistent AFPrior ablationAPVI aloneFreedom from symptomatic or asymptomatic AF and other ATs without AADs after single procedure
Verma et al.25Eur Heart J2010Multicentre100Drug-refractory, high-burden paroxysmal (episodes >6 h, >4 in 6 months) or persistent AF‘Permanent AF', AF secondary to an obvious reversible cause, inadequate anticoagulation, LA thrombus on TEE, contraindications to systemic anticoagulation with heparin or Coumadin, prior AF ablation, LA size >55 mm, pregnant or possibly pregnantCPVA aloneFreedom from AF recurrence from months 3 to 12 post-ablation after 1 or 2 procedures on and off AADs
  • DCCV, direct current cardioversion; EKG, electrocardiogram; LA, left atrium; LVEF, left ventricular ejection fraction; MI, myocardial infarction; NR, not reported; RF, radiofrequency; TEE, transoesophageal echocardiogram.

View this table:
Table 2

Study follow-up protocols

StudyDuration of anticoagulation before ablation (weeks)Duration of anticoagulation after ablation (months)AAD post-procedureFollow-up protocolMean duration of follow-up for primary endpoint (months)Mean duration of follow-up after last procedure (months)
Deisenhofer et al.216NRNo AADs until 3-months post-procedure1- and 3-month outpatient clinic visit post-procedure, then q3 months; intensive questions for arrhythmia symptoms; 24-h Holter 1-month post-ablation; 7-day Holter 3-months post-ablation with MRI or CT scan to rule out PV stenosis; TTE 1-month post-ablation3319 ± 819 ± 8
Di Biase et al.22NRWarfarin continued for minimum of 6 monthsPre-procedure AAD (except amiodarone) continued for 8 weeks post-procedure3-month outpatient clinic visit, then q3 months; event recorder for 5 months with 4 recordings per week even if asymptomatic and any time they felt symptoms; 48-h Holter obtained at 3-, 6-, 9-, 12-, and 15-month post-procedure13.7 ± 2.213.7 ± 2.29 ± 79 ± 7
Elayi et al.23NR12 months minimumDiscontinued after 2 monthsEvent recorders for the first 6 months with 4 recordings per week even if asymptomatic and any time with symptoms; outpatient visits with 12-lead EKG, 48-h Holter at 3, 6, 9, 12, and 15 months16.7 ± 1.6 (CPVI);16.4 ± 1.1 (APVI)16.2 ± 1.214.8 ± 3.6 (CPVI);14.6 ± 3.8 (APVI)14.9 ± 3.0
Oral et al.30NRHeparin for 5 days, warfarin for 3 monthsClass I or III AAD for 8–12 weeks post-procedure3-month outpatient clinic visit post-procedure, then q6 months; NP follow-up q4 weeks; patients instructed to call whenever symptoms in which case an event recorder was used to document rhythm at the time of symptoms6666
Oral et al.24NRWarfarin discontinued after 6 months if SRPre-procedure AAD continued for 8–12 weeks post-procedure3-month outpatient clinic visit post-procedure, then q3–6 months; 30-day auto-triggered event monitor 6-months post-procedure; patients instructed to call if symptoms10 ± 310 ± 39 ± 49 ± 4
Verma et al.2543 months minimumAADs continued for 2 months post-procedure, but discontinued in all patients after 2 months3-, 6-, and 12-month outpatient clinic visits; monthly telephone interviews; 12-lead EKG and 48-h Holter at 3-, 6-, and 12-months follow-up visits; external loop recorders (minimum 2 weeks) and or transtelephonic monitors for any patient-reported symptoms outside of routine follow-ups; interrogation of implanted devices when applicable; routine spiral CT or MRI of chest within 6 months to assess for PV stenosis12121212
  • CT, computed tomography; EKG, electrocardiogram; MRI, magnetic resonance imaging; NP, nurse practitioner; NR, not reported; PV, pulmonary vein; TTE, transthoracic echocardiogram.

Five trials randomized patients to some form of PVI (antral or circumferential) as the control group vs. PVI followed by additional CFAE ablation.21,22,24,25,30 Elayi et al.23 was the only study in which the patients randomized to a CFAE strategy underwent CFAE ablation first, followed by APVI. Patients randomized to a non-CFAE strategy were randomized to APVI alone or CPVA alone.23 For this analysis, both the CPVA and APVI groups were combined for comparison with the group that underwent CFAE ablation followed by APVI. Two trials randomized patients to one of three treatment groups: Di Biase et al.22 randomized patients to APVI alone, CFAE ablation alone, or APVI followed by CFAE ablation; and the Substrate and Trigger Ablation for Reduction of AF (STAR-AF) trial, published by Verma et al.,25 randomized patients to CPVA alone, CFAE ablation alone, or CPVA followed by CFAE ablation. As treatment efficacy of stand-alone CFAE ablation was not the focus of this analysis, for both studies we excluded the group that received CFAE ablation alone.

Baseline patient characteristics

Of the 538 patients included in this meta-analysis, 270 (50.3%) had paroxysmal AF and 267 (49.7%) had persistent AF (AF classification was not reported for 1 patient). Overall, baseline patient characteristics between patients who underwent PVI alone compared with those who underwent PVI + CFAE were not significantly different (Table 3). Patients’ mean age across all trials was 57.9 ± 9.8 years and over three-fourths were male (76.0%). Across the five studies reporting the mean duration of AF, it was 6.1 ± 5.0 years. The mean left atrial diameter was 44.4 ± 5.7 mm and the mean left ventricular ejection fraction was 55.7 ± 8.8.

View this table:
Table 3

Baseline characteristics of patients enrolled in RCTs

StudyPatients (n)Mean age (years ± SD)Males [n (%)]Duration of AF (years ± SD)LA diameter (mm ± SD)LVEF (% ± SD)
Deisenhofer et al.21485058 ± 1055 ± 1033 (69)41 (82)NRaNRa43 ± 644 ± 5NRNR
Di Biase et al.22353457 ± 8.158.4 ± 7.529 (83)30 (80)5.3 ± 5.75.3 ± 543 ± 0.644 ± 0.655 ± 854.6 ± 6
Elayi et al.23954959 ± 10.259.2 ± 11.563 (66)32 (65)6.1 ± 3.46.3 ± 2.545 ± 6.946.2 ± 6.45455
Oral et al.30303055 ± 1156 ± 924 (80)26 (87)7 ± 56 ± 443 ± 744 ± 659 ± 458 ± 7
Oral et al.24505058 ± 1062 ± 841 (82)41 (82)6 ± 55 ± 447 ± 646 ± 653 ± 1254 ± 9
Verma et al.25323455 ± 1159 ± 1024 (75)25 (74)6.4 ± 6.67.6 ± 9.443 ± 541 ± 662 ± 759 ± 12
Total29124757.6 ± 10.158.4 ± 9.5214 (74)195 (79)6.1 ± 4.86.0 ± 5.344.3 ± 6.044.4 ± 5.555.6 ± 10.055.8 ± 10.1
  • NR, not reported.

  • aReported as 38 ± 31 h compared with 36 ± 33 h on pre-procedural Holter monitor in the PVI group.

Catheter ablation techniques and complex fractionated atrial electrogram definitions

The catheter ablation techniques and ablation endpoints varied among the six studies (Table 4). For each study, the control groups underwent stand-alone PVI—APVI in two studies,22,24 segmental PVI in one,21 CPVA in two studies,25,30 and either CPVA or APVI in one.23 All studies used RF energy and five used electroanatomical mapping with CARTO (Biosense Webster Inc., Diamond Bar, CA, USA),2124,30 whereas one used Ensite NavX (St Jude Medical, Minneapolis, MN, USA).25 Maximum energy outputs and target temperatures during RF energy delivery differed for each study.

View this table:
Table 4

Catheter ablation techniques and CFAE definitions

StudyAblation techniquesEnergy sourceCatheter tip size (mm)Maximum energy output (W)Maximum temperature (°C)Definition of CFAEAblation endpoint
Deisenhofer et al.21Segmental PVI; APVI + CFAERFNRNR30354343(i) Atrial EGMs that have fractionated EGMs composed of ≥2 deflections, and/or perturbation of the baseline with continuous deflection of a prolonged activation complex over a 10s recording period; or (ii) atrial EGMs with a very short cycle length (≤120 ms) averaged over a 10 s recording periodPVI: electrical isolation of all PVs
CFAE: termination to sinus rhythm with subsequent non-inducibility using high-frequency burst pacing × 5
Di Biase et al.22APVI; CFAE; APVI + CFAERF3.53.54545<41<41(i) Atrial EGMs with ≥2 deflections or with fractionated baseline complexes with continuous activity over a 10 s recording time; or (ii) atrial EGMs with a cycle length ≤120 ms over a 10 s recording timePVI: local elimination of all PV potentials along antra or inside veins (entry and exit block)
CFAE: complete elimination of CFAE areas and electrical isolation of all PV antra defined by entrance and exit block
Elayi et al.23CPVA; APVI; CFAE + APVIRF3.53.550504141(i) Atrial EGMs with fractionation and composed of ≥2 deflections and/or with continuous activity of the baseline; or (ii) atrial EGMs with cycle length ≤120 msPVI: elimination of all PV potentials along antra or inside veins (entry block)
CFAE: complete elimination of CFAEs
Oral et al.30CPVA; CPVA + CFAERF8870705555Not describedAF termination with non-inducibility with atrial burst pacing × 5 at the shortest cycle length resulting in 1:1 atrial capture for ≥ 15 sa
Oral et al.24APVI; APVI + CFAERF3.53.535354545(i) Atrial EGMs with a cycle length ≤120 ms or shorter than the AF cycle length in the coronary sinus; or (ii) EGMs that were fractionated or displayed continuous electrical activityPVI: complete electrical isolation of all PVs
CFAE: complete electrical isolation of all PVs with additional CFAE ablation in the LA and CS for up to 2 additional hours or until AF terminated, whichever came first
Verma et al.25CPVA; CFAE; CPVA + CFAERF3.5 or 83.5 or 860605050CFAE sites defined by automated algorithm (CL <120 ms) (Ensite NavX, St Jude Medical)PVI: abolishment of all PV potentials within each antrum
CFAE: complete electrical isolation of all PVs followed by CFAE ablation if spontaneous or inducible AF
  • CS, coronary sinus; EGM, electrogram; LA, left atrium; PV, pulmonary vein; RF, radiofrequency.

  • aThe endpoint of ablation was not defined separately for those patients who underwent additional CFAE ablation.

All but one study outlined the investigators’ definition of a CFAE to be targeted for ablation (Table 4).30 For four of the five studies that defined CFAEs, definitions were based on Nademanee et al.'s12 description: atrial electrograms with (i) fractionation composed of two or more deflections or continuous electrical activity over a 10 s period; or (ii) cycle length ≤120 ms over a 10 s period. Verma et al.25 used an automated algorithm (Ensite NavX, St Jude Medical) to define CFAE sites with cycle lengths <120 ms and Elayi et al.23 used automated algorithms from either CARTO or NavX.

Efficacy of pulmonary vein isolation + adjunctive complex fractionated atrial electrogram ablation

Random-effects modelling was used to assess the efficacy of adjunctive CFAE ablation compared with stand-alone PVI. For the primary endpoint of freedom from any AF/AT at ≥3 months of follow-up following a single ablation with or without AAD therapy, adjunctive CFAE ablation improved the odds of maintaining sinus rhythm compared with PVI alone (OR 2.0, 95% CI 1.04–3.8, P = 0.04) (Figure 2). However, there was moderate heterogeneity among the six RCTs (I2 = 63.0%). Across all studies, the rate of repeat ablation was 25% among patients who underwent PVI alone and 21% among those who received adjunctive CFAE ablation (P = 0.27).

Figure 2

Freedom from AF/AT at post-procedure follow-up. Odds ratios for PVI vs. PVI + CFAE.

Paroxysmal atrial fibrillation vs. persistent atrial fibrillation

To evaluate possible sources of the moderate heterogeneity among the combined RCTs, we performed separate sensitivity analyses based on AF classification: paroxysmal vs. persistent. Combining the three trials enrolling patients with paroxysmal AF, heterogeneity was markedly diminished (I2 = 6.2%); however, in this patient subgroup, adjunctive CFAE ablation did not demonstrate a significant treatment effect (OR 1.4, 95% CI 0.74–2.7, P = 0.30). Between the two trials that enrolled patients with long-standing persistent AF, heterogeneity remained high (I2 = 88.2%) and again there was no evident treatment effect of adjunctive CFAE ablation among this patient subgroup (OR 2.1, 95% CI 0.42–10.3, P = 0.36). The results of the most recently published STAR-AF trial could not be included in sensitivity analyses as the primary endpoint data were not separately available for patients with paroxysmal vs. persistent AF.

Influence of individual trials

In order to assess the impact of individual studies on the overall treatment effect, we conducted serial sensitivity analyses in which each of the six RCTs was excluded in turn. The overall treatment effect with the serial exclusion of Oral et al.,24 Deisenhofer et al.,21 and Di Biase et al.22 each resulted in statistically significant increased odds of freedom from AF after a single procedure with or without AADs (OR 2.37, 95% CI 1.22–4.60, P = 0.01; OR 2.25, 95% CI 1.09–4.65, P = 0.03; and OR 2.19, 95% CI 1.05–4.54, P = 0.04, respectively). Serial exclusion of Elayi et al.,23 Oral et al.,30 or Verma et al.25 led to similar OR and associations, but without statistical significance (OR 1.56, 95% CI 0.87–2.79, P = 0.14; OR 1.83, 95% CI 0.88–3.79, P = 0.10; and OR 1.75, 95% CI 0.84–3.75, P = 0.13, respectively).

Secondary endpoints

Performing CFAE ablation significantly increased mean procedural times (178.5 ± 66.9 vs. 331.5 ± 92.6 min, P< 0.001), mean fluoroscopy (59.5 ± 22.2 vs. 115.5 ± 35.3 min, P< 0.001), and mean RF energy application times (46.9 ± 36.6 vs. 74.4 ± 43.0 min, P< 0.001) (Figure 3).

Figure 3

Mean procedural, fluoroscopy, and RF energy application times. Comparison of PVI vs. PVI + CFAE.

Pulmonary vein isolation + CFAE ablation organized AF into AT during the ablation in 32% of the patients across the five trials that reported this information, and of these patients, 40% were successfully converted to sinus rhythm by additional catheter ablation targeting these ATs. Reporting was asymmetric and incomplete for patients who underwent PVI alone; however, among the two trials that did report these numbers, 27% of the patients organized AF into AT and 40% of these patients were successfully converted to sinus rhythm after additional ablation targeting the resultant ATs (Table 5).

View this table:
Table 5

Secondary outcomes of RCTs

StudyPatients (n)No. of patients with post-procedure ATRepeat procedure for AFNo. of patients with AF organization to ATNo. of patients with AF termination to SRPersistence of AF requiring DCCV or ibutilide
Deisenhofer et al.214850NR2/30 (7%)15/46a (33%)17/48a (35%)NR10/30 (33%)NR17/30 (57%)NR3/30 (10%)
Di Biase et al.223534NRNR4/35 (11%)3/34 (9%)12/35 (34%)10/34 (29%)21/35 (60%)22/34 (65%)2/35 (6%)2/34 (6%)
Elayi et al.239549NRNR26/95 (27%)10/49 (20%)23/95 (24%)34/49 (70%)4/95 (4%)2/49 (4%)68/95 (72%)13/49 (27%)
Oral et al.303030NR8/30 (27%)0/30 (0%)0/30 (0%)NRNRNRNRNRNR
Oral et al.2450502/50 (4%)6/50 (12%)18/50 (36%)17/50 (34%)NR4/50 (8%)NR5/50 (10%)NR45/50 (90%)
Verma et al.253234NRNR10/33 (31%)5/34 (15%)NR8/3426/3311/346/337/34
Total2912472/50 (4%)16/110 (15%)73/289 (25%)52/245 (21%)35/130 (27%)66/197 (34%)51/163 (31%)46/197 (29%)76/163 (47%)70/197 (36%)
  • NR, not reported.

  • aTwo patients from each treatment group lost during long-term follow-up.

The risk of recurrence of organized ATs post-procedure was not consistently reported. One study reported rates for both treatment arms: 4% in the stand-alone PVI group and 12% in the PVI + CFAE group.24 In this study, one patient in the stand-alone PVI and three patients in the PVI + CFAE group underwent a separate ablation to target clinically significant AT (one patient in the PVI + CFAE group subsequently underwent a third ablation for recurrent AT after the second procedure).24 Two other studies reported the rate of post-procedure ATs in the PVI + CFAE groups only (27% in Oral et al.24 and 7% in Deisenhofer et al.21), but did not report whether subsequent ATs developed in the PVI only groups.


For most of these studies, complications were reported in aggregate from both initial and repeat procedures. There were very few procedural complications overall (n= 18/538 patients, 3.3%) among the five studies for which complications were reported.2125 The most common complication was bleeding or vascular complications (n= 6), followed by pericardial effusion (n= 5), of which one was due to cardiac perforation with tamponade, and at least three of these patients had undergone PVI + CFAE. Although one study reported 5 of the 18 total reported complications (1 pericardial effusion, 2 pericarditis, and 2 bleeding/vascular complications), this study did not specify whether the patients were in the PVI group vs. the PVI + CFAE group.24 Pulmonary vein stenosis was reported in three patients who received PVI alone and one patient who received PVI + CFAE. None of the studies reported deaths related to the catheter ablation procedures. One patient who underwent PVI + CFAE survived a prolonged asystolic arrest that occurred 3 h post-procedure during venous sheath removal.21 Overall, the small numbers of complications (4 with PVI vs. 9 with PVI + CFAE) precluded further analysis of the relative risk of complication with one ablation strategy vs. the other given the available data.


There are three main findings of this meta-analysis. First, the efficacy of CFAE ablation in addition to PVI for the maintenance of sinus rhythm after a single procedure with or without AADs is 66%, which is significantly improved compared with PVI alone. Second, procedural complications are minimal; however, adjunctive CFAE ablation is associated with increased procedural, fluoroscopy, and RF energy application times. Third, sensitivity analyses suggest that differences in ablation technique, patient selection, and other variables may be important determinants of outcome after adjunctive CFAE ablation and additional studies are still needed to assess the long-term efficacy and safety of adjunctive CFAE ablation for the maintenance of sinus rhythm.

The strategy of adjunctive CFAE ablation continues to evolve and there are a variety of factors that likely impact the identification of CFAEs and the potential for procedural efficacy when they are targeted for ablation. First, in the earlier five of the six RCTs analysed, CFAEs were identified by real-time visual inspection, which can be challenging and subject to inter-observer variability.31 With the increasing utilization of automated CFAE detection algorithms, CFAEs will become more reliably and objectively identified and targeted.32 However, the optimal criteria for the detection of CFAEs has yet to be determined.31 Secondly, the five RCTs performed before the most recently published STAR-AF trial limited CFAE ablation to the left atrium. The role of right atrial CFAEs remains incompletely elucidated, and thus, the clinical efficacy of adjunctive CFAE ablation may improve with biatrial CFAE ablation. Thirdly, understanding of both the genesis and clinical significance of CFAEs in different patient subgroups is unclear.33 Theoretically, CFAEs could be produced by multiple mechanisms and could thus represent a multitude of electroanatomic substrates not all of which are necessarily relevant to AF elimination.34 As such, it is possible that different CFAE subtypes exist and that techniques to identify and target CFAEs together with patient selection criteria will impact the efficacy of adjunctive CFAE ablation.

Safety considerations

Evaluation of the therapeutic value of a treatment strategy should include consideration of the risk–benefit relationship. These studies suggest that the strategy of CFAE ablation in addition to PVI results in lengthened procedural and fluoroscopy times, which could theoretically increase anaesthesia-related complications and would be expected to increase exposure to ionizing radiation.35,36 In this analysis, we found a nearly 50% increase in ionizing radiation burden with adjunctive CFAE ablation compared with PVI alone. In light of known associations between increased exposure to ionizing radiation and the development of solid tumours, leukaemia, and even cardiovascular disease,37,38 this procedural risk may be real, though warranted, if procedural efficacy is significantly improved. Adjunctive CFAE ablation also lengthens RF energy application time, which could potentially increase the risk of developing a pericardial effusion; however, given the small number of complications in these RCTs, there are not yet enough data available to make any conclusions. Furthermore, additional non-linear ablation could result in a greater propensity for atrial proarrhythmia. On the basis of these data, it is possible that CFAE ablation is proarrhythmic. Recurrent post-procedure AT was 4% with PVI vs. 15% with PVI + CFAE; however, adequately powered trials are still needed to answer this question. To optimally balance procedural risk and benefit, more data on complications and longer-term follow-up, as well as ablation methods and patient selection, are necessary.


As with any meta-analysis, the possibility of publication bias cannot be excluded. Likewise, the inclusion of RCTs performed at high-volume centres with experienced operators may not reflect the experience of patients treated in general clinical practice, particularly given the complex nature of CFAE identification and ablation. Our findings were sensitive to the inclusion of several trials, likely reflecting the relatively small sample size, different patient populations, varying ablation techniques, and/or modest treatment effect. These studies’ relatively short and highly variable follow-up times, different ablation endpoints, wide variation in post-procedure AAD use and reporting, and differences in the primary endpoints further limited our analysis. Procedural efficacy after a single procedure may also disproportionately reflect an inability to achieve permanent isolation of the pulmonary veins rather than a lack of benefit from CFAE ablation, thus underestimating the efficacy of adjunctive CFAE ablation. Furthermore, with the exception of the STAR-AF trial, patients who received repeat/multiple procedures did not follow the initial randomization assignment regarding ablation strategy. Owing to the relatively small sample sizes of the available RCTs, studies with distinctly different AF populations (paroxysmal vs. long-standing persistent) and varying PVI techniques were combined. Pooling paroxysmal and persistent AF patients, who likely have different underlying substrates, may underestimate the efficacy of CFAE ablation in one or the other group. As only two RCTs have yet been published evaluating adjunctive CFAE ablation in persistent AF patients, more studies are clearly needed in this patient population for whom CFAE ablation may have more incremental benefit. Finally, as we did not have access to the original data from each of the trials, extrapolation errors and the combining of data from multiple control groups into one group through weighting may have affected our results. Without access to patient-level data, meta-analytical methods could not be used to identify predictors of ablation outcomes and AF recurrence.

Clinical implications

With the ever-increasing prevalence of AF combined with the limited efficacy of AADs, PVI has become a commonly performed RF catheter ablation procedure. The current American College of Cardiology/American Heart Association/Heart Rhythm Society guidelines recommend catheter ablation as second-line therapy once a patient has failed one or more AADs (Class IIb).39 Strategies to improve the efficacy of AF ablation continue to evolve, such as the adjunctive ablation of CFAEs, which improves the maintenance of sinus rhythm after a single procedure compared with PVI alone. Although there may be increased risk associated with the increased procedural, fluoroscopy, and RF energy application times, additional studies are still needed to fully evaluate the risk–benefit profile of adjunctive CFAE ablation. Future randomized studies comparing stand-alone PVI to adjunctive CFAE ablation may provide additional information about clinical efficacy in the absence of AADs, and in specific subpopulations like persistent AF patients, as well as include longer-term follow-up to define the durability of treatment effect, and further elucidate the efficacy and safety profile of CFAE ablation.


The efficacy of adjunctive CFAE ablation is improved compared with stand-alone PVI for the maintenance of sinus rhythm after a single catheter ablation with or without AADs. The improved efficacy of PVI with adjunctive CFAE ablation should be weighed against possible added risk due to significantly increased procedural, fluoroscopy, and RF energy application times.

Conflict of interest: M.H.K. receives research funding from Biotronik and serves on an advisory board for Medtronic. J.P.P. receives research funding from Boston Scientific and serves on an advisory board for Medtronic. T.D.B. receives research funding from Medtronic, St Jude Medical, and Biosense Webster, Inc.; and participates as a consultant and/or speaker for Sequel Pharma, Sanofi-Aventis, Hansen Medical, and Philips Medical Systems.


M.H.K. was supported by the Ruth L. Kirschstein-National Research Service Award (Kirschstein-NRSA) National Institutes of Health (NIH) (grant number 5-T32-DK-007731-15). The contents of this work are solely the responsibility of the authors and do not necessarily represent the official views of the NIH.


View Abstract