Abstract
Several techniques have been utilized for the ablation of persistent (P) and long-standing persistent (LsP) atrial fibrillation (AF); however, the best approach of substrate ablation remains poorly defined. This study aims to examine the impact of ablation approach on outcomes associated with P or LsP AF ablation by conducting a meta-analysis and regression on contemporary literature.
A systematic literature review was conducted up to 29 July 2015 for scientific literature reporting on outcomes associated with P or LsP AF ablation. One hundred and thirteen studies reported outcomes in a total of 18 657 patients undergoing various ablation approaches for the treatment of P–LsP AF between 2001 and 2015. The point efficacy estimate of a single-AF ablation procedure without the use of anti-arrhythmic drugs was 43% (95% CI; 39–47%). Multiple procedures and/or the use of anti-arrhythmic drugs increase success to 69% (95% CI; 66–71%). Meta-regression revealed that ablation technique (P < 0.001) and left atrial size (P = 0.02) were predictive of single procedure, drug-free success. The addition of extra-pulmonary substrate approaches was associated with declining efficacy when compared to a pulmonary vein ablation alone.
The efficacy of a single-AF ablation procedure for P or LsP AF is 43%; however, can be increased to 69% with the use of multiple procedures and/or anti-arrhythmic drugs. Current literature supports the finding that pulmonary vein antrum ablation/isolation is at least equivalently efficacious to other contemporary P–LsP ablation strategies.
Introduction
The genesis and maintenance of atrial fibrillation (AF) requires an interaction between a trigger, perpetuators, and substrate. The pulmonary veins play an important role in the initiation and maintenance of AF, with pulmonary vein electrical isolation proving to be an effective treatment of paroxysmal arrhythmia.1 However, when this approach alone is applied to the remodelled atrial substrate in persistent or long-standing persistent (P–LsP) AF, the clinical success has been thought to be limited.2–4 Ablation approaches for P–LsP AF have therefore moved from the pulmonary veins to the atrial substrate, commonly in the forms of pulmonary vein antrum ablation/isolation, linear transection of the left and right atrium and more recently, signal guided, or complex fractionated atrial electrogram (CFAE) ablation. Although the specific ‘extra’ pulmonary vein ablation approach that produces the maximum efficacy gains remains a point of debate, it has generally been accepted that substrate modification yields improved outcomes in P–LsP AF patients. However, the STAR AF II study—a prospective, multicentre, randomized controlled trial—provides a higher level of evidence, and demonstrated equivalent efficacy (single procedure, drug free ∼32%) between pulmonary vein antrum isolation (PVAI) and combined substrate (CFAE or linear) and PVAI approaches.4
A randomized controlled study would be prohibitive based on the required sample-size to assess the relative clinical efficacy of all ablation strategies present in the literature. Hence, this meta-analysis and regression aims to present the current ‘state of play’ across all reported ablation strategies for persistent AF.
Methods
Search strategy/sources
This systematic review was performed in accordance with MOOSE guidelines.5 Pubmed and Embase were searched between 1 January 1990 and 29 July 2015 using the search terms of: ‘(atrial fibrillation[MESH] AND ‘atrial fibrillation’ AND (success or outcome*) AND (persistent OR permanent OR chronic OR long-term OR long-standing) AND ablation)’ in Pubmed and ‘(“atrial fibrillation” AND ablation AND (persistent OR permanent OR chronic OR “long term” OR “long standing”) AND (success OR outcome*))’ in Embase. These searches were performed by the investigators (J.C., A.B.) and returned 1704 and 3838 results, respectively. All references (n = 5542) were imported into Endnote X7 referencing software; 1491 duplicates were removed, leaving 4051 references for review. Non-English language studies, abstracts, and un-published data were excluded. Bibliographies of included studies were reviewed to retrieve additional relevant information; however, only one study was retrieved via this pathway (see Supplementary material online, Reference 44). Figure 1 outlines the search strategy and pathway to the final included studies.
Search criteria and flow diagram of the reasons and number of studies excluded from the retrieved abstracts and full-text articles.
Inclusion and exclusion criteria
This review included studies that recruited patients with P–LsP AF undergoing an index ablation procedure. All cohorts therefore consisted of patients with at least 7 days of continuous AF, in accordance with the current HRS Consensus Statement.1 The definition of long-standing persistent AF was defined as continuous AF >12 months in accordance with the HRS guidelines. Studies that used the terms persistent and/or chronic/long-standing persistent AF (without specific definition of AF duration) were assumed to have met the aforementioned criteria. We segregated studies according to whether their sample included persistent, long-standing persistent, or mixed persistent/long-standing persistent AF sufferers. A minimum of 60 index P–LsP patients (per defined ablation strategy) with a defined follow-up of at least 12 months was required to be included in this review.
Exclusion criteria included review articles, patient cohorts, or specific patient subsets that were not representative of consecutive ablation series (i.e. patients only with heart failure, specified age ranges, and/or other demographic/clinical characteristics). In addition, cryoablation strategies and surgical/hybrid approaches to AF ablation were excluded. Studies that segmented outcomes according to another interest variable or had an indeterminate ablation protocol were also excluded.
Outcomes
We sought to determine (i) single procedure, drug-free as well as multiple procedure, drug-assisted clinical success for P–LsP AF and (ii) assess the association between ablation technique and other clinical/procedural parameters with clinical outcomes.
Data extraction
Study characteristics, ‘single procedure, drug free’ and ‘multiple procedure, drug assisted’ success rates and follow-up were extracted from the included studies. The term ‘single procedure’ refers to outcomes of cohorts that had single-ablation procedure performed and no use of antiarrhythmic drugs. The term ‘multiple procedure’ refers to outcomes of cohorts that had multiple ablation procedures performed and/or used antiarrhythmic drugs for greater than 3 months after the index procedure. The highest success of either multiple procedure drug-free, single procedure drug-assisted, or multiple procedure drug-assisted was recorded as the ‘multiple procedure’ success. Follow-up was categorized as either objective (scheduled Holter/loop monitoring for at least 12 months) or subjective (symptoms preceding investigation or isolated electrocardiograms). The follow-up duration was assumed to be the same for single-procedure and multiple-procedure success within a single cohort, unless specifically stated.
The term ‘study’ refers to the publication (n = 113 included; see Supplementary material online, Item S1), whereas the term ‘cohort’ refers to a group of patients who have undergone a particular ablation strategy between and potentially within (in comparative or randomized design) studies (n = 123). A meta-regression was performed only for single-procedure outcomes. Multiple-procedure clinical outcomes were collated given that patients in persistent, sometimes non-cardiovertable AF, are still likely to demonstrate quality of life improvements on restoration of sinus rhythm via multiple procedures and/or previously ineffective drugs. A meta-regression on multiple-procedure success was not conducted because of sparse reporting on repeat ablation strategies (which may have differed from the index strategy) and the timing of (and blanking period) associated with the repeat procedure in relation to the end of follow-up.
Ablation approaches
The various approaches reported in the literature were categorized as follows:
Pulmonary vein approaches including ostial PV isolation (PVI);
Pulmonary vein antrum ablation (without specific testing of electrical isolation; PVA) and PVAI;
Pulmonary vein (any PV approach) and linear (including roofline, mitral isthmus line, tricuspid isthmus line, inferior left atrial line; LIN);
Pulmonary vein (any PV approach) and complex fractionated atrial electrograms (either subjectively or objectively assessed; CFAE);
Pulmonary vein (any PV approach) and LIN + CFAE
Other (any non-PV approach)
Complex fractionated atrial electrogram only,
Left atrial line only,
Ganglionated plexi ablation (with high frequency stimulation to define regions; GP) only,
Posterior wall box isolation.
Statistical analysis
Single-procedure and multiple-procedure outcomes associated with P–LsP AF ablation were pooled across 74 and 99 cohorts, respectively using random effects meta-analysis models. For the purposes of analysis, cohorts within studies were assumed to have the same independence as those between studies. Publication bias was assessed by visual inspection of funnel plots by two independent investigators.
Where there was evidence of heterogeneity (P < 0.05) in single-procedure outcomes between cohorts, meta-regression models were utilized to identify predictors of success. Single-procedure meta-regression models were re-run on studies between 2010 and 2015 to examine the impact of inclusion year on results. Variables entered into univariate models included: ablation approach, age, gender, type of AF (persistent, long-standing persistent or mixed cohort), LA size, LVEF, procedure time, RF ablation time, duration of follow-up, type of follow-up, and year of publication. Where appropriate, the number of cohorts contributing to each point estimate or association is noted due to the highly variant reporting. Meta-regression was restricted to univariate analyses due to insufficient power to resolve the independence of multiple predictors of outcome and/or their interactions. Calculations were performed using Stats Direct and the R statistical package.
Results
Cohort and procedural characteristics
A total of 18 657 patients undergoing well-defined AF ablation procedures were reviewed in 123 cohorts spanning publication years 2001–2015 (81% were from 2010 onwards; see Supplementary material online, Item S2). The cohorts were extracted from four (three single-centre3,6,7 and one multi-centre4) randomized controlled trials (contributing to nine cohorts) and 109 observational studies (contributing to 114 cohorts and including five multi-centre studies (see Supplementary material online, References 5, 6, 23, 62, 64). There were 34 persistent AF, 31 long-standing persistent AF and 58 mixed cohorts. Table 1 presents the cohort descriptive data and the number of cohorts that reported each parameter. The P–LsP cohorts undergoing ablation had a mean age of 59 ± 3 years and by HRS guideline definition1 had suffered from a bout of continuous AF for at least 1 week. Patients undergoing ablation were predominantly men (78%) with enlarged left atria (45 ± 3 mm). On average, each patient underwent a procedure of approximately 3.6 h in length (range 1.3–6.1 h), with 82 min of radiofrequency energy delivery (range 33–160 min). The mean follow-up was 25 months (range 12–59 months) with 95 cohorts assessing outcomes objectively (see Supplementary material online, Item S3).
Descriptive characteristics of the typical P–LsP AF ablation cohort included in this review
| Estimate | Range | No. of cohorts | |
|---|---|---|---|
| Cohort size | 152 (89–166) | 60–1057 | 123 |
| Age (years) | 59 ± 3 | 50–67 | 123 |
| % men | 78 ± 8 | 56–94 | 123 |
| AF duration (months) | 59 ± 21 | 17–101 | 59 |
| LA size (mm) | 45 ± 3 | 41–51 | 89 |
| LVEF (%) | 58 ± 4 | 48–65 | 82 |
| Procedure time (min) | 215 ± 56 | 80–365 | 68 |
| RF ablation time (min) | 82 ± 28 | 33–160 | 38 |
| % Persistent in mixed cohorts | 52 ± 22 | 8–97 | 43 |
| No. of procedures per patient in multiple-procedure cohorts | 1.37 ± 0.3 | 1–2.3 | 79 |
| Mean follow-up (months) | 25 ± 12 | 12–59 | 102 |
| Estimate | Range | No. of cohorts | |
|---|---|---|---|
| Cohort size | 152 (89–166) | 60–1057 | 123 |
| Age (years) | 59 ± 3 | 50–67 | 123 |
| % men | 78 ± 8 | 56–94 | 123 |
| AF duration (months) | 59 ± 21 | 17–101 | 59 |
| LA size (mm) | 45 ± 3 | 41–51 | 89 |
| LVEF (%) | 58 ± 4 | 48–65 | 82 |
| Procedure time (min) | 215 ± 56 | 80–365 | 68 |
| RF ablation time (min) | 82 ± 28 | 33–160 | 38 |
| % Persistent in mixed cohorts | 52 ± 22 | 8–97 | 43 |
| No. of procedures per patient in multiple-procedure cohorts | 1.37 ± 0.3 | 1–2.3 | 79 |
| Mean follow-up (months) | 25 ± 12 | 12–59 | 102 |
LA, left atrial; LVEF, left ventricular ejection fraction; RF, radiofrequency.
Descriptive characteristics of the typical P–LsP AF ablation cohort included in this review
| Estimate | Range | No. of cohorts | |
|---|---|---|---|
| Cohort size | 152 (89–166) | 60–1057 | 123 |
| Age (years) | 59 ± 3 | 50–67 | 123 |
| % men | 78 ± 8 | 56–94 | 123 |
| AF duration (months) | 59 ± 21 | 17–101 | 59 |
| LA size (mm) | 45 ± 3 | 41–51 | 89 |
| LVEF (%) | 58 ± 4 | 48–65 | 82 |
| Procedure time (min) | 215 ± 56 | 80–365 | 68 |
| RF ablation time (min) | 82 ± 28 | 33–160 | 38 |
| % Persistent in mixed cohorts | 52 ± 22 | 8–97 | 43 |
| No. of procedures per patient in multiple-procedure cohorts | 1.37 ± 0.3 | 1–2.3 | 79 |
| Mean follow-up (months) | 25 ± 12 | 12–59 | 102 |
| Estimate | Range | No. of cohorts | |
|---|---|---|---|
| Cohort size | 152 (89–166) | 60–1057 | 123 |
| Age (years) | 59 ± 3 | 50–67 | 123 |
| % men | 78 ± 8 | 56–94 | 123 |
| AF duration (months) | 59 ± 21 | 17–101 | 59 |
| LA size (mm) | 45 ± 3 | 41–51 | 89 |
| LVEF (%) | 58 ± 4 | 48–65 | 82 |
| Procedure time (min) | 215 ± 56 | 80–365 | 68 |
| RF ablation time (min) | 82 ± 28 | 33–160 | 38 |
| % Persistent in mixed cohorts | 52 ± 22 | 8–97 | 43 |
| No. of procedures per patient in multiple-procedure cohorts | 1.37 ± 0.3 | 1–2.3 | 79 |
| Mean follow-up (months) | 25 ± 12 | 12–59 | 102 |
LA, left atrial; LVEF, left ventricular ejection fraction; RF, radiofrequency.
Outcomes: maintenance of sinus rhythm
Overall
The single-procedure success was reported in 74 cohorts with an overall clinical success rate of 43% (95% CI; 39–47%). Multiple procedures (mean of 1.37 procedures/patient) and/or anti-arrhythmic drugs increased the point estimate to 69% (95% CI; 66–71%; 99 cohorts). Significant heterogeneity in outcomes was present in both single-procedure (I2 = 95%; P < 0.001) and multiple-procedure (I2 = 92%; P < 0.001) estimates. Figure 2 demonstrates the combined point estimates and 95% CI for single-procedure and multiple-procedure cohorts across all six ablation approaches.
Combined point estimates and 95% CI for single-procedure and multiple-procedure cohorts across all six ablation approaches.
Ostial pulmonary vein isolation
Pulmonary vein isolation was used in seven cohorts with a mean follow-up of 2.2 years (range 1.3–4.1 years). All studies specifically used the term ‘ostial’ when describing their pulmonary vein approach. Single-procedure success was reported in four cohorts with an estimate of 46% (95% CI; 26–67%; Figure 3). Multiple-procedure success was reported in five cohorts and the point estimate increased to 69% (95% CI; 59–79%; Figure 3).
Forest plots for the single-procedure and multiple-procedure efficacies for PVI and PVA/PVAI cohorts. LSP, cohort consists of long-standing persistent patients only; M, cohort consistent of P and LsP patients; N, cohort size; P, cohort consists of persistent patients only.
Pulmonary vein antrum ablation ± isolation
Pulmonary vein antrum ablation ± isolation was used in 14 cohorts with a mean follow-up of 2.7 years (range 1.0–4.8 years). Only a single early (2003) study8 utilized a PVA approach that did not specifically set out to achieve electrical isolation of the vein and/or antrum. Of these cohorts, five used ablation of the superior vena cava (SVC) in addition to the PV technique in all patients. Single-procedure success was reported in 10 cohorts with an estimate of 57% (95% CI; 48–66%; Figure 3). Multiple-procedure success was reported in 10 cohorts and the point estimate increased to 71% (95% CI; 63–79%; Figure 3).
Pulmonary vein approach and linear
PV + LIN was used in 16 cohorts (11 cohorts used linear ablation for all patients, the remainder used linear ablation as part of a step-wise approach) with a mean follow-up of 1.9 years (range 1.0–4.1 years). Linear ablation consisted primarily of a roofline or mitral isthmus line. Single-procedure success was reported in 12 cohorts with an estimate of 46% (95% CI; 38–55%; Figure 4). Multiple-procedure success was reported in 12 cohorts and the point estimate increased to 75% (95% CI; 65–85%; Figure 4).
Forest plots for the single-procedure and multiple-procedure efficacies for PV + LIN cohorts.
Pulmonary vein approach and complex fractionated atrial electrograms
PV + CFAE was used in 26 cohorts (11 cohorts used CFAE ablation for all patients; the remainder used CFAE ablation as part of a step-wise approach) with a mean follow-up of 1.9 years (range 1.0–3.8 years). Single-procedure success was reported in 20 cohorts with an estimate of 46% (95% CI; 41–52%; Figure 5). Multiple-procedure success was reported in 21 cohorts and the point estimate increased to 67% (95% CI; 61–73%; Figure 5).
Forest plots for the single-procedure and multiple-procedure efficacies for PV + CFAE cohorts.
Pulmonary vein approach, linear, and complex fractionated atrial electrograms
PV + LIN + CFAE was used in 55 cohorts with a mean follow-up of 2.0 years (range 1.0–4.9 years). The components of ablation (PV, linear, and CFAE) were normally applied in a stepwise sequence until termination to sinus rhythm. Hence, only five of these cohorts used linear ablation in all patients and only four cohorts used CFAE ablation in all patients. Single-procedure success was reported in 24 cohorts with an estimate of 33% (95% CI; 28–38%; Figure 6). Multiple-procedure success was reported in 47 cohorts and the point estimate increased to 67% (95% CI; 63–70%; Figure 6).
Forest plots for the single-procedure and multiple-procedure efficacies for PV + LIN + CFAE cohorts.
Approaches without pulmonary vein ablation
A non-PV approach was used in five cohorts, with a mean follow-up of 1.2 years (range 1.0–1.4 years). Two cohorts used CFAE ablation only (see Supplementary material online, References 66, 72), two used linear ablation only (see Supplementary material online, References 44, 105, both using a contiguous ablation line to essentially isolate the posterior wall), and one GP ablation only (see Supplementary material online, Reference 79). Single-procedure success was reported in four cohorts with an estimate of 44% (95% CI; 34–55%; see Supplementary material online, Item S4). Multiple-procedure success was reported in four cohorts and the point estimate increased to 74% (95% CI; 57–88%; see Supplementary material online, Item S4).
Predictors of single-procedure success
The ablation strategy utilized was a predictor of clinical success among single-procedure cohorts (P < 0.001). Using PVA/PVAI as the reference strategy (57% single-procedure success), the most complex PV + LIN + CFAE approach was associated with significantly poorer outcomes (single-procedure success = 33%; P < 0.001), whilst PV + LIN (46%; P = 0.071) and PV + CFAE (46%; P = 0.051) both trended towards inferiority.
A decrease in single-procedure efficacy of 1.6% (95% CI; 0.2–3.0%) was seen for each millimetre increase in the average cohort left atrial size (P = 0.02; see Supplementary material online, Item S8). In addition, efficacy trended towards inferiority with an increase in publication year (P = 0.07), indicating an overall decline in success between 2001 and 2015. All remaining sample and procedural characteristics assessed (Table 2) were not significantly related to single-procedure efficacy (P ≥ 0.15).
Meta-regression—univariate predictors of single-procedure clinical success
| Variable | Estimate (95% CI) | No. of cohorts | P-value |
|---|---|---|---|
| Ablation strategy | 74 | <0.001 | |
| PVA/PVAI | 57 (48, 66) | 10 | Reference |
| PVI | 46 (26, 67) | 4 | 0.215 |
| PV + LIN | 46 (38, 55) | 12 | 0.071 |
| PV + CFAE | 46 (41, 52) | 20 | 0.051 |
| PV + LIN + CFAE | 33 (28, 38) | 24 | <0.001 |
| Other | 44 (34, 55) | 4 | 0.13 |
| Age | 0.0 (−1.2, 1.2) | 74 | 0.98 |
| Gender | −0.8 (−6.2, 4.7) | 74 | 0.78 |
| AF type | 74 | 0.44 | |
| Persistent | 47 (40, 54) | 18 | |
| LS persistent | 49 (37, 61) | 21 | |
| Mixed | 41 (36, 47) | 35 | |
| LA size | −1.6 (−3.0, −0.2) | 59 | 0.02 |
| LVEF | 0.2 (−0.9, 1.2) | 51 | 0.76 |
| Procedure time | 0.0 (−0.1, 0.0) | 46 | 0.22 |
| RF ablation time | −0.1 (−0.3, 0.0) | 22 | 0.15 |
| Follow-up duration | −0.1 (−0.4, 0.1) | 70 | 0.27 |
| Follow-up method | 74 | 0.80 | |
| Objective | 43 (39, 47) | 60 | |
| Subjective | 42 (33, 51) | 14 | |
| Year | −1.1 (−2.4, 0.1) | 74 | 0.07 |
| Variable | Estimate (95% CI) | No. of cohorts | P-value |
|---|---|---|---|
| Ablation strategy | 74 | <0.001 | |
| PVA/PVAI | 57 (48, 66) | 10 | Reference |
| PVI | 46 (26, 67) | 4 | 0.215 |
| PV + LIN | 46 (38, 55) | 12 | 0.071 |
| PV + CFAE | 46 (41, 52) | 20 | 0.051 |
| PV + LIN + CFAE | 33 (28, 38) | 24 | <0.001 |
| Other | 44 (34, 55) | 4 | 0.13 |
| Age | 0.0 (−1.2, 1.2) | 74 | 0.98 |
| Gender | −0.8 (−6.2, 4.7) | 74 | 0.78 |
| AF type | 74 | 0.44 | |
| Persistent | 47 (40, 54) | 18 | |
| LS persistent | 49 (37, 61) | 21 | |
| Mixed | 41 (36, 47) | 35 | |
| LA size | −1.6 (−3.0, −0.2) | 59 | 0.02 |
| LVEF | 0.2 (−0.9, 1.2) | 51 | 0.76 |
| Procedure time | 0.0 (−0.1, 0.0) | 46 | 0.22 |
| RF ablation time | −0.1 (−0.3, 0.0) | 22 | 0.15 |
| Follow-up duration | −0.1 (−0.4, 0.1) | 70 | 0.27 |
| Follow-up method | 74 | 0.80 | |
| Objective | 43 (39, 47) | 60 | |
| Subjective | 42 (33, 51) | 14 | |
| Year | −1.1 (−2.4, 0.1) | 74 | 0.07 |
Meta-regression—univariate predictors of single-procedure clinical success
| Variable | Estimate (95% CI) | No. of cohorts | P-value |
|---|---|---|---|
| Ablation strategy | 74 | <0.001 | |
| PVA/PVAI | 57 (48, 66) | 10 | Reference |
| PVI | 46 (26, 67) | 4 | 0.215 |
| PV + LIN | 46 (38, 55) | 12 | 0.071 |
| PV + CFAE | 46 (41, 52) | 20 | 0.051 |
| PV + LIN + CFAE | 33 (28, 38) | 24 | <0.001 |
| Other | 44 (34, 55) | 4 | 0.13 |
| Age | 0.0 (−1.2, 1.2) | 74 | 0.98 |
| Gender | −0.8 (−6.2, 4.7) | 74 | 0.78 |
| AF type | 74 | 0.44 | |
| Persistent | 47 (40, 54) | 18 | |
| LS persistent | 49 (37, 61) | 21 | |
| Mixed | 41 (36, 47) | 35 | |
| LA size | −1.6 (−3.0, −0.2) | 59 | 0.02 |
| LVEF | 0.2 (−0.9, 1.2) | 51 | 0.76 |
| Procedure time | 0.0 (−0.1, 0.0) | 46 | 0.22 |
| RF ablation time | −0.1 (−0.3, 0.0) | 22 | 0.15 |
| Follow-up duration | −0.1 (−0.4, 0.1) | 70 | 0.27 |
| Follow-up method | 74 | 0.80 | |
| Objective | 43 (39, 47) | 60 | |
| Subjective | 42 (33, 51) | 14 | |
| Year | −1.1 (−2.4, 0.1) | 74 | 0.07 |
| Variable | Estimate (95% CI) | No. of cohorts | P-value |
|---|---|---|---|
| Ablation strategy | 74 | <0.001 | |
| PVA/PVAI | 57 (48, 66) | 10 | Reference |
| PVI | 46 (26, 67) | 4 | 0.215 |
| PV + LIN | 46 (38, 55) | 12 | 0.071 |
| PV + CFAE | 46 (41, 52) | 20 | 0.051 |
| PV + LIN + CFAE | 33 (28, 38) | 24 | <0.001 |
| Other | 44 (34, 55) | 4 | 0.13 |
| Age | 0.0 (−1.2, 1.2) | 74 | 0.98 |
| Gender | −0.8 (−6.2, 4.7) | 74 | 0.78 |
| AF type | 74 | 0.44 | |
| Persistent | 47 (40, 54) | 18 | |
| LS persistent | 49 (37, 61) | 21 | |
| Mixed | 41 (36, 47) | 35 | |
| LA size | −1.6 (−3.0, −0.2) | 59 | 0.02 |
| LVEF | 0.2 (−0.9, 1.2) | 51 | 0.76 |
| Procedure time | 0.0 (−0.1, 0.0) | 46 | 0.22 |
| RF ablation time | −0.1 (−0.3, 0.0) | 22 | 0.15 |
| Follow-up duration | −0.1 (−0.4, 0.1) | 70 | 0.27 |
| Follow-up method | 74 | 0.80 | |
| Objective | 43 (39, 47) | 60 | |
| Subjective | 42 (33, 51) | 14 | |
| Year | −1.1 (−2.4, 0.1) | 74 | 0.07 |
When the meta-regression was repeated on studies published between 2010 and 2015 (see Supplementary material online, Item S5) the point estimate of PVA/PVAI was reduced from 57 to 47%, but PVA/PVAI results remained statistically superior to 32% single-procedure success associated with the PV + LIN + CFAE approach. PVA/PVAI success (47%; single procedure) was equivalent to the 44%-point estimate associated with PV + LIN and PV + CFAE (47%; single procedure) ablation. Left atrial size remained a predictor of outcome, whilst follow-up duration showed a negative association with single-procedure success (P = 0.004), that was not present in the overall analysis.
Discussion
Although P–LsP AF ablation outcomes are extensively reported in the scientific literature, techniques are rarely compared in moderate sized (>60 patients per strategy) randomized trials.3,4,6,7,9 Indeed, the most recent STAR AF II randomized controlled trial4 has observed outcomes distinct to conventional thinking indicating that ‘less may be more’ when treating P–LsP AF via ablation, although these findings are still subject to significant debate.10 We aimed to assess whether the literature on the whole supports the findings of STAR AF II, to extend efficacy comparisons to ablation strategies not assessed in randomized trials and to provide an overall current ‘state of efficacy’ of contemporary AF ablation.
Our main findings are:
Contemporary radiofrequency ablation techniques for P–LsP AF yield a single-procedure drug free success rate of 43% at ∼25 months follow-up. Repeat procedures and the use of previously ineffective anti-arrhythmic drugs increase the clinical success to 69%.
PVA/PVAI was the P–LsP ablation strategy statistically associated with the greatest single-procedure clinical success (57%) compared with PV + LIN, PV + CFAE, or PV + LIN + CFAE. When restricting the analysis to the most contemporary studies (2010–2015), PVA/PVAI was equivalent to PV + LIN and PV + CFAE, but remained statistically superior to PV + LIN + CFAE.
The current ‘state of efficacy’ for ablation of persistent and long-standing persistent atrial fibrillation
On average, a P–LsP ablation patient undergoing any one of the ablation approaches collated in this review has a 43% chance of maintaining sinus rhythm off anti-arrhythmic drugs with a single procedure at 25 months’ follow-up.
Reassuringly, it would seem that maintaining sinus rhythm in a significant proportion of P–LsP parents is possible, albeit after multiple procedures and potential application of anti-arrhythmic agents. The implementation of repeat procedures (∼1.37 procedures per patient) and/or the addition of previously ineffective anti-arrhythmic drugs in an indeterminate number of patients improves the outcomes of P–LsP AF ablation to 69% at 25 months.
The forest plots demonstrate large heterogeneity in success rates within and between ablation strategies, which may be a reflection of the variance in ‘real’ success in addition to patient, study design and monitoring variance (see Supplementary material online, Item S3) associated with observational studies. Hunter et al. ( see Supplementary material online, Reference 32) reported the lowest procedural success of 16% with a PV + LIN + CFAE approach, in contrast to the most efficacious reporting of 76% in the observational study that utilized a PVAI approach alone.( see Supplementary material online, Reference 6).
Ablation strategy as a predictor of atrial fibrillation ablation success
The STAR AF II randomized controlled trial,4 demonstrated equivalent single-procedure success in persistent ablation strategies of PVAI (41%), PVAI + LIN (29%), and PVAI + CFAE (33%) in 549 patients. The STAR AF II trial was powered toward a comparison between two substrate ablation strategies [PVAI + LIN (n = 244) and PVAI + CFAE (n = 244)], not PVAI ± substrate ablation. Hence, the numerical superiority of PVAI did not reach statistical significance, predominately because of the smaller sample size in the reference PVAI group (n = 61). Three additional randomized controlled trials were included in this review3,6,7; however, no two ablation strategies overlapped such that they could be combined (see Supplementary material online, Item S6).
We found that ablation approach predicted single-procedure success rate (P < 0.001), which was due primarily to the statistical superiority of the PVA/PVAI approach (57%) when compared with the PV + LIN + CFAE technique (33%; P < 0.001). Reflective of the aforementioned randomized results, PVAI + LIN (46%) and PVAI + CFAE (46%) strategies also trended toward (P ≤ 0.071) lower success rates compared with the PVAI approach alone. When we restricted our single-procedure meta-regression to contemporary studies (see Supplementary material online, Item S5) the over-arching patterns of PVA/PVAI equivalence (compared with PV + LIN and PV + CFAE) and superiority compared with PV + LIN + CFAE remained, which established that our overall findings were not an artifact of the broad timeline of study inclusion.
While our efficacy rates are consistently higher than STAR AF II, the relative comparison between the three ablation strategies is remarkably consistent and demonstrates that the addition of extra-pulmonary substrate approaches does not improve outcomes and may even be detrimental to long-term success. The association between more complex ablation approaches and declining single-procedure clinical success may in part be attributed to additional ablation strategies creating new arrhythmias (i.e. atrial tachycardia/flutter) in contrast to the index arrhythmia recurring. However, STAR AF II demonstrated no significant difference in atrial tachycardia/flutter rates between PVAI (11%), PVAI + LIN (14%), and PVAI + CFAE (11%) arms, even in the presence of a non-significant numerical decrease in success (as defined by freedom from all atrial arrhythmia) from 41% to 29% and 33%, respectively. These data clearly suggest, at least in this trial, additional ablation complexity did not create the substrate for new arrhythmia beyond that of AF.
The PVA/PVAI approach yielded the highest single-procedure clinical success (57%) in this review. Of note, two of the studies (see Supplementary material online, References 6, 95), that comprised 4/10 cohorts utilizing this approach also isolated the SVC in all patients. These studies reported an average single-procedure success of 69%, compared to 47% for the remaining PVAI approaches, which suggests that this ablation feature may play a role in increasing efficacy. However, these data are in direct conflict to randomized evidence suggesting equivalence9; hence, the improved clinical outcomes associated with SVC isolation in this review may be an artefact of the study design/monitoring associated with the common centre reporting these results.
Our overall PVA/PVAI 57% single-procedure success rate for P–LsP AF is slightly lower than 67% reported in a recent meta-analysis on PVI for P–LsP arrhythmia.11 However, there are important differences in inclusion criteria between the two reviews; namely (i) they allowed use of anti-arrhythmic drugs in their estimate (∼20% of their combined sample), (ii) compared with our 2.7 years for PVA/PVAI, their included studies reported on shorter mean follow-ups (1.2 years), (iii) 50% of their sample underwent cryo-balloon PVI that was excluded in our review, and (iv) six of their 14 studies had samples <60 that were excluded in our review. Indeed, after acknowledging the differences between the two reviews, it is reassuring that the point estimates remain this similar for a pure pulmonary vein focused strategy for P–LsP AF.
Other predictors of atrial fibrillation ablation success
A 1 mm increase in LA size was associated with a 1.6% (95% CI; 0.2–3.0%; P = 0.02) decrease in single-procedure clinical success, which is a relationship supported by other meta-analyses.12 It is particularly interesting that LA size within the P–LsP AF subgroup remains a predictor. This may indicate that within P–LsP AF, clinical categorization through echocardiography still plays a role in disease severity classification and hence, remains a predictor of outcome. In contrast, the segregation of persistent and long-standing persistent AF according to the HRS guidelines,1 did not predict outcomes in our meta-regression (P = 0.44) suggesting that these clinical categories do not have the resolution to discriminate disease severity and outcomes.
Interestingly, year of publication was negatively associated with single-procedural success with a 1.1% (95% CI; −0.1 to 2.5%; P = 0.07) decrease per increasing year of publication, showing a trend towards inferiority. The continued limited understanding of the underlying mechanisms of P–LsP AF may still remain the largest contributing factor to our ability to effectively treat sustained arrhythmia. The borderline decrease in efficacy may have also been due to a greater adoption of more complex and less efficacious ablation strategies over time (38 of the 55 PV + LIN + CFAE cohorts came from studies published since 2012) or potentially an artefact of more inexperienced centres publishing their first endeavours of P–LsP AF ablation. It is also plausible that more robust study designs with more frequent objective clinical monitoring has also led to a decline in published efficacy.
In contrast to several studies13–15 that have shown the potential for increased detection of occult AF with increased monitoring, the objectivity of follow-up did not predict outcomes in our meta-regression (P = 0.80), which may represent our ‘blunt’ study level classification of objective and subjective follow-up (see Supplementary material online, Item 3), rather than a genuine lack of relationship.
Comparison with previous review
We have previously presented a systematic literature review on the ablation success for long-standing persistent AF.16 Eighty-nine of the 113 papers included in this review were published after our 2010 publication,16 yet the overall efficacy point estimate has not increased. In addition, important methodological differences between the two reviews (including studies reporting ablation outcomes for persistent AF alone and restricting inclusion to 60 patients per strategy with >12 months of follow-up) has not impacted on the overall single-procedure efficacy reported in 2010 (44%) compared to now (43%).
Although our overall efficacy for single procedure and multiple-procedure clinical success is almost identical, the strategy specific single-procedure results are contrasting. Previously, we reported highest single-procedure efficacy in the more complex ablation strategies of PV + CFAE (54%) and the stepwise (PV + LIN + CFAE; 51%) technique, which is in direct conflict with our lowest efficacy reported in the most complex PV + LIN + CFAE (33%) in our current review. However, the two reviews had only seven studies in common (see Supplementary material online, Item S7), due to the fact that an increased sample size (n = 60 vs. n = 10) and follow-up restriction (12 months vs. 6 months) were applied. We feel that restricting entry to larger studies with greater follow-up provides a more real-world estimate of overall and technique specific P–LsP AF ablation success.
Limitations
Although the collation of all available published information on P–LsP AF ablation provides the best overall ‘state-of-play’ estimate on efficacy, the ‘success’ of AF ablation procedure integrates ablation technique, disease severity, intensity, and objectivity of follow-up as well as other unknown and unreported prognostic factors. Multiple-procedure success further confounds estimates by additionally introducing repeat ablation strategies (that would most often be different to the index approach) and/or variant use and blanking periods associated with anti-arrhythmic pharmaceuticals. Meta-regression of single-procedure outcomes is therefore a hypothesis generating methodology for assessing efficacy differences between ablation strategies. A randomized controlled trial still remains the gold standard. It is advised that the findings reported in this review are interpreted with the aforementioned factors in mind.
Conclusions
This review and meta-analysis demonstrates that radiofrequency ablation techniques for patients with P–LsP AF yield an overall success of 43% at 25 months and that multiple procedures and/or anti-arrhythmic drugs improve outcomes to 69%. The published literature supports the findings of the recent STAR AF II trial, with collated results indicating that a simpler PVAI approach (57% success) yields at least equivalent single-procedure results (or potentially better) compared with more complex substrate ablation techniques including PVAI + LIN (46%), PVAI + CFAE (46%), and PVAI + LIN + CFAE (33%) as currently performed and reported. The conduct of prospectively registered, collaborative RCTs with standardized ablation strategies, blinded follow-up, objective arrhythmia detection methodology, and reporting of recurrences according to the type of arrhythmia (i.e. atrial flutter, atrial tachycardia, or AF) is necessary to further our understanding of the true AF ablation efficacy. New ablation strategies that target the potential mechanism of AF should also continue to be assessed for efficacy with adjunctive conventional PV approaches.
Funding
Mahajan is supported by a Postdoctoral Fellowship jointly funded by the National Heart Foundation of Australia and the National Health and Medical Research Council of Australia, and by the Leo J. Mahar Lectureship from the University of Adelaide. Twomey is supported by Leo J. Mahar Electrophysiology Scholarships from the University of Adelaide. Pathak is supported by a Postdoctoral Fellowship from the National Health and Medical Research Council of Australia. Kumar and Munawar are supported by the Robert J. Craig Scholarship from the University of Adelaide. Munawar is supported by the Indonesia Endowment Fund for Education, Ministry of Finance, The Republic of Indonesia. Kalman and Sanders are supported by Practitioner Fellowships from the National Health and Medical Research Council of Australia. Lau is supported by the Robert J. Craig Lectureship from the University of Adelaide. Sanders is supported by the National Heart Foundation of Australia.
Conflict of interest: Brooks is an employee of LivaNova. The University of Adelaide has received lecture fees and research funding on behalf of Mahajan from Medtronic and St Jude Medical. Kalman reports having received research funding from St Jude Medical, Biosense-Webster, Medtronic and Boston Scientific. Sanders reports having served on the advisory board of Biosense-Webster, Medtronic, St Jude Medical, Boston Scientific and CathRx. The University of Adelaide has received lecture and/or consulting fees on behalf of Sanders from Biosense-Webster, Medtronic, St Jude Medical, and Boston Scientific. Sanders reports having received research funding from Medtronic, St Jude Medical, Boston Scientific, Biotronik and LivaNova.
