https://orcid.org/0000-0001-6082-7542
Abstract
To evaluate the efficacy and safety of vernakalant for the cardioversion of atrial fibrillation (AF).
We reviewed the literature for randomized trials that compared vernakalant to another drug or placebo in patients with AF of onset ≤7 days. We used a random-effects model to combine quantitative data and rated the quality of evidence using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation). From 441 total citations in MEDLINE, EMBASE, and CENTRAL (December 2018), we identified nine trials evaluating 1358 participants. Six trials compared vernakalant to placebo, two trials compared vernakalant to ibutilide, and one trial compared vernakalant to amiodarone. We found significant methodological bias in four trials. For conversion within 90 min, vernakalant was superior to placebo [50% conversion, risk ratio (RR) 5.15; 95% confidence interval (CI); 2.24–11.84, I2 = 91%], whereas we found no significant difference in conversion when vernakalant was compared with an active drug (56% vs. 24% conversion, RR 2.40; 95% CI 0.76–7.58, I2 = 94). Sinus rhythm was maintained at 24 h in 85% (95% CI 80–88%) of patients who converted acutely with vernakalant. Overall, we judged the quality of evidence for efficacy to be low based on inconsistency and suspected publication bias. There was no significant difference in the risk of significant adverse events between vernakalant and comparator (RR 0.95; 95% CI 0.70–1.28, I2 = 0, moderate quality evidence). Vernakalant is safe and effective for rapid and durable restoration of sinus rhythm in patients with recent-onset AF.
Vernakalant should be a first line option for the pharmacological cardioversion of patients with haemodynamically stable recent-onset AF without severe structural heart disease.
This systematic review found six randomized trials that compared vernakalant to placebo, two that compared vernakalant to ibutilide and one that compared vernakalant to amiodarone.
For conversion within 90 min, vernakalant was superior to placebo (50% conversion, RR 5.15; 95% CI 2.24–11.84, I2 = 91%), and not significantly different from ibutilide/amiodarone (56% vs. 24% conversion, RR 2.40; 95% CI 0.76–7.58, I2 = 94). In total, 85% (95% CI 80–88%) of patients who converted acutely with vernakalant were in sinus rhythm at 24 h. There was no significant difference in the risk of significant adverse events between vernakalant and comparator (RR 0.95; 95% CI 0.70–1.28, I2 = 0%).
Vernakalant should be a first line option for the pharmacological cardioversion of patients with haemodynamically stable recent-onset AF without severe structural heart disease.
Introduction
Atrial fibrillation (AF) is the most common cardiac arrhythmia and is associated with significant morbidity and mortality.1 Current guidelines indicate that rhythm control is a reasonable initial strategy in the management of stable acute AF.2,3 While electrical cardioversion is generally more effective at restoring sinus rhythm than antiarrhythmic agents, sedation does carry some risk and can be resource intensive.2 Current pharmacological therapies have varying success rates for conversion and potential adverse effects include ventricular arrhythmias and hypotension.4–6 Vernakalant is a rapid-acting, relatively atrial-selective antiarrhythmic drug that is approved in Canada and Europe for the pharmacological cardioversion of recent-onset AF.7,8
Whether vernakalant is a safe and efficacious alternative to other treatments available for recent-onset AF is clinically important. We therefore performed a systematic review and meta-analysis aiming to summarize the safety and efficacy data for vernakalant used for the acute conversion of recent-onset AF.
Methods
The study protocol was registered with PROSPERO (2018 CRD42018085020). Differences between the registered protocol and final manuscript are described in the Supplementary material online, Appendix S1.
Objective
Our primary objective was to determine if the rates of conversion within 90 min were different in patients with recent-onset AF who were treated with vernakalant as compared to placebo or an active comparator. Additionally, we aimed to compare data on sustained conversion and safety.
Eligibility criteria
Randomized clinical trials were included, irrespective of publication status, date of publication, risk of bias, outcomes published, or language. Trials were included if they enrolled adults with recent-onset AF (within 7 days). Included studies had to compare the administration of vernakalant to the administration of placebo or another active drug.
Outcomes of interest were the proportion of patients who converted within 90 min, time to converted, proportion converted within 24 h, serious adverse events, both in aggregate and in isolation (stroke, heart failure, cardiogenic shock, death, ventricular arrhythmia, bradycardia, or sinus arrest requiring intervention). We accepted the outcomes as defined by study authors.
Search methods
We searched MEDLINE, EMBASE, and Cochrane CENTRAL for keywords describing the condition, intervention, or comparator from inception to 12 December 2018 (Supplementary material online, Appendix S2).
We also searched trial registries for ongoing and unpublished clinical trials via http://www.controlled-trials.com using the multiple database search option metaRegister of Controlled Trials and the WHO trial registry. We screened the references of eligible papers and consulted experts to identify additional trials.
Data collection and analysis
Selection of studies
Study selection was performed using Covidence Systematic review software (Veritas Health Innovation, Melbourne, Australia). Two reviewers independently screened studies’ titles and abstracts for eligibility. Full papers of the potentially eligible studies were retrieved. The same two reviewers then independently screened full texts in duplicate and recorded the main reason for exclusion. Disagreements were resolved through discussion.
Data extraction and management
Independently, two reviewers abstracted data on intervention and outcome. They also recorded study and participant characteristics including age, sex, duration of AF prior to treatment, and concomitant conditions (e.g. heart failure and ischaemic heart disease). Reviewers compared results and resolved disagreements by discussion with a third party. Authors were contacted to clarify ambiguities and to request data on outcomes missing in primary reports.
Assessment of risk of bias
In duplicate, we assessed risk of bias using the Cochrane tool.9 In each trial, reviewers evaluated the following domains: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessors, incomplete outcome data, and selective reporting. The results were compared and disagreements resolved by discussion. Performance and detection bias were assessed separately. A priori, we decided to classify open-label designs as ‘likely low risk of bias’ for detection bias when a clear electrocardiogram monitoring protocol was described. For analysis and presentation purposes, risk of bias was dichotomized as high (or likely high) or low (or likely low). If a study was at risk of selection, performance, detection, or reporting bias for that outcome, we categorized it as high risk of bias. For subgroup analyses, study-level risk of bias was assessed for each outcome.
Measures of association with treatment
The reported standard association measures for clinical outcomes were risk ratios (RRs) and mean differences (MD) for time to conversion. The absolute risk difference for clinical outcomes was obtained by applying the RR with 95% confidence interval (CI) to the baseline risk in the control group.
Clinical and methodological heterogeneity were assessed based on study characteristics. Statistical heterogeneity was measured using the I2 statistic. We considered an I2 greater than 50% as showing substantial heterogeneity.9
RevMan 5.3 (The Cochrane Collaboration, Denmark) was used to combine data quantitatively when clinical heterogeneity was non-substantial. We used a random-effects model with Mantel–Haenszel weighting because several comparisons were expected to show heterogeneity. In trials where participants crossed over to the other treatment, the analysis was according to their first assigned group (intention-to-treat principle). P-values <0.05 (two-sided) were considered statistically significant.
Subgroup and sensitivity analyses
Pre-specified subgroup analyses were performed hypothesizing that vernakalant would be superior in the following subgroups: (i) patients treated with placebo as compared to an active drug, (ii) patients who were in the post-operative period following cardiac surgery compared to those who were not, (iii) patients without prior AF as compared to those with a history of AF, (iv) patients without a history of ischaemic heart disease as compared to those with a history of ischaemic heart disease, (v) patients without as compared to those with a history of depressed left ventricular ejection fraction (LVEF) or heart failure. We evaluated for interaction between subgroups and treatment effect; we report P-values for interaction.
Assessment of the quality of the evidence
We evaluated the quality of evidence using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach.10 GRADE appraises the confidence in estimates of effect by considering within-study risk of bias, directness of the evidence, heterogeneity of the data, precision of effect estimates, and risk of publication bias. We inspected the funnel plots of standard errors vs. effect estimates for publication bias and small-study effects.
Results
Screening
The electronic search resulted in 441 unique citations (Figure 1). After reference and full-text screening, nine trials met eligibility criteria (Table 1). Additional details from included trials are presented in the Supplementary material online, Appendix S3.
Study selection diagram for trials comparing vernakalant to placebo or active drug in patients with recent-onset atrial fibrillation.
Characteristics of included randomized controlled trials comparing vernakalant to comparator in patients with recent-onset atrial fibrillation
| Study name (year), ‘Study ID’ | Characteristics | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Number of patients randomized | Comparator, initial dosing of vernakalanta | Study design | AF duration prior | Primary endpoint | Age, vernakalant/ comparator (years), mean (SD) | Gender, male/female | Number of CAD (%) | Number of CHF (%) | Length of follow-up | |
| Beatch and Mangal (2016), ‘ACT V’ | 217 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 63.7 (12.7)/60.8 (14.1) | 146/71 | 48 (22) | Excluded | 7 days |
| Beatch et al. (2017), ‘ASIA-PACIFIC’ | 122 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 60.7 (13.7)/59.2 (12.0) | 67/55 | 13 (11) | 8 (7) | 7 days |
| Camm et al. (2011), ‘AVRO’ | 232 | Amiodarone, 5 mg/kg | Randomized, double-blind, multicentre | 3–48 h | Proportion converting to NSR within 90 min | 63.1 (10.81)/62.2 (11.63) | 146/86 | 52 (22) | 46 (20) | 30 days |
| Kowey et al. (2009), ‘ACT II’ | 161 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3–72 h | Proportion converting to NSR within 90 min | 68.3 (7.7)/67.8 (6.4) | 121/40 | 129 (80) | 51 (32) | 30 days |
| Pratt et al. (2010), ‘ACT III’ | 172 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre. Cardiac Surgery | 3 h to 7 days | Proportion converting to NSR within 90 min | 60 (16)/60 (15) | 118/54 | 24 (14) | 23 (13) | 30 days |
| Roy et al. (2004), ‘CRAFT’ | 56 | Placebo, 0.5 mg/kg 2.0 mg/kg | Randomized, double-blind, multicentre | 3–72 h | Proportion converting to NSR within 30 min | 60.8/64 | 32/22 | – | Excluded | 7 days |
| Roy et al. (2008), ‘ACT I’ | 220 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 60.4 (14)/59.9 (11.8) | 150/70 | 51 (23) | 19 (8.6) | 30 days |
| Simon et al. (2017) | 100 | Ibutilide, 3 mg/kg | Randomized, open-label, single centre. Emergency Department | No longer than 48 h | Time to conversion of AF to NSR | 56.2 (14.32)/56.7 (15.77) | 68/32 | 7 (7) | 99 (99) | Prior to discharge |
| Vogiatzis et al. (2017) | 78 | Ibutilide, 3 mg/kg | Pseudo-randomized, open-label, single centre | No longer than 48 h | Proportion converting to NSR within 90 min | 62.44 (7.24)/60.8 (14.1) | 56/22 | 31 (40) | – | – |
| Study name (year), ‘Study ID’ | Characteristics | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Number of patients randomized | Comparator, initial dosing of vernakalanta | Study design | AF duration prior | Primary endpoint | Age, vernakalant/ comparator (years), mean (SD) | Gender, male/female | Number of CAD (%) | Number of CHF (%) | Length of follow-up | |
| Beatch and Mangal (2016), ‘ACT V’ | 217 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 63.7 (12.7)/60.8 (14.1) | 146/71 | 48 (22) | Excluded | 7 days |
| Beatch et al. (2017), ‘ASIA-PACIFIC’ | 122 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 60.7 (13.7)/59.2 (12.0) | 67/55 | 13 (11) | 8 (7) | 7 days |
| Camm et al. (2011), ‘AVRO’ | 232 | Amiodarone, 5 mg/kg | Randomized, double-blind, multicentre | 3–48 h | Proportion converting to NSR within 90 min | 63.1 (10.81)/62.2 (11.63) | 146/86 | 52 (22) | 46 (20) | 30 days |
| Kowey et al. (2009), ‘ACT II’ | 161 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3–72 h | Proportion converting to NSR within 90 min | 68.3 (7.7)/67.8 (6.4) | 121/40 | 129 (80) | 51 (32) | 30 days |
| Pratt et al. (2010), ‘ACT III’ | 172 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre. Cardiac Surgery | 3 h to 7 days | Proportion converting to NSR within 90 min | 60 (16)/60 (15) | 118/54 | 24 (14) | 23 (13) | 30 days |
| Roy et al. (2004), ‘CRAFT’ | 56 | Placebo, 0.5 mg/kg 2.0 mg/kg | Randomized, double-blind, multicentre | 3–72 h | Proportion converting to NSR within 30 min | 60.8/64 | 32/22 | – | Excluded | 7 days |
| Roy et al. (2008), ‘ACT I’ | 220 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 60.4 (14)/59.9 (11.8) | 150/70 | 51 (23) | 19 (8.6) | 30 days |
| Simon et al. (2017) | 100 | Ibutilide, 3 mg/kg | Randomized, open-label, single centre. Emergency Department | No longer than 48 h | Time to conversion of AF to NSR | 56.2 (14.32)/56.7 (15.77) | 68/32 | 7 (7) | 99 (99) | Prior to discharge |
| Vogiatzis et al. (2017) | 78 | Ibutilide, 3 mg/kg | Pseudo-randomized, open-label, single centre | No longer than 48 h | Proportion converting to NSR within 90 min | 62.44 (7.24)/60.8 (14.1) | 56/22 | 31 (40) | – | – |
AF, atrial fibrillation; CAD, coronary artery disease; CHF, congestive heart failure; NSR, normal sinus rhythm; SD, standard deviation.
Subsequent doses are described in the Supplementary material online, Appendix.
Characteristics of included randomized controlled trials comparing vernakalant to comparator in patients with recent-onset atrial fibrillation
| Study name (year), ‘Study ID’ | Characteristics | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Number of patients randomized | Comparator, initial dosing of vernakalanta | Study design | AF duration prior | Primary endpoint | Age, vernakalant/ comparator (years), mean (SD) | Gender, male/female | Number of CAD (%) | Number of CHF (%) | Length of follow-up | |
| Beatch and Mangal (2016), ‘ACT V’ | 217 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 63.7 (12.7)/60.8 (14.1) | 146/71 | 48 (22) | Excluded | 7 days |
| Beatch et al. (2017), ‘ASIA-PACIFIC’ | 122 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 60.7 (13.7)/59.2 (12.0) | 67/55 | 13 (11) | 8 (7) | 7 days |
| Camm et al. (2011), ‘AVRO’ | 232 | Amiodarone, 5 mg/kg | Randomized, double-blind, multicentre | 3–48 h | Proportion converting to NSR within 90 min | 63.1 (10.81)/62.2 (11.63) | 146/86 | 52 (22) | 46 (20) | 30 days |
| Kowey et al. (2009), ‘ACT II’ | 161 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3–72 h | Proportion converting to NSR within 90 min | 68.3 (7.7)/67.8 (6.4) | 121/40 | 129 (80) | 51 (32) | 30 days |
| Pratt et al. (2010), ‘ACT III’ | 172 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre. Cardiac Surgery | 3 h to 7 days | Proportion converting to NSR within 90 min | 60 (16)/60 (15) | 118/54 | 24 (14) | 23 (13) | 30 days |
| Roy et al. (2004), ‘CRAFT’ | 56 | Placebo, 0.5 mg/kg 2.0 mg/kg | Randomized, double-blind, multicentre | 3–72 h | Proportion converting to NSR within 30 min | 60.8/64 | 32/22 | – | Excluded | 7 days |
| Roy et al. (2008), ‘ACT I’ | 220 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 60.4 (14)/59.9 (11.8) | 150/70 | 51 (23) | 19 (8.6) | 30 days |
| Simon et al. (2017) | 100 | Ibutilide, 3 mg/kg | Randomized, open-label, single centre. Emergency Department | No longer than 48 h | Time to conversion of AF to NSR | 56.2 (14.32)/56.7 (15.77) | 68/32 | 7 (7) | 99 (99) | Prior to discharge |
| Vogiatzis et al. (2017) | 78 | Ibutilide, 3 mg/kg | Pseudo-randomized, open-label, single centre | No longer than 48 h | Proportion converting to NSR within 90 min | 62.44 (7.24)/60.8 (14.1) | 56/22 | 31 (40) | – | – |
| Study name (year), ‘Study ID’ | Characteristics | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Number of patients randomized | Comparator, initial dosing of vernakalanta | Study design | AF duration prior | Primary endpoint | Age, vernakalant/ comparator (years), mean (SD) | Gender, male/female | Number of CAD (%) | Number of CHF (%) | Length of follow-up | |
| Beatch and Mangal (2016), ‘ACT V’ | 217 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 63.7 (12.7)/60.8 (14.1) | 146/71 | 48 (22) | Excluded | 7 days |
| Beatch et al. (2017), ‘ASIA-PACIFIC’ | 122 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 60.7 (13.7)/59.2 (12.0) | 67/55 | 13 (11) | 8 (7) | 7 days |
| Camm et al. (2011), ‘AVRO’ | 232 | Amiodarone, 5 mg/kg | Randomized, double-blind, multicentre | 3–48 h | Proportion converting to NSR within 90 min | 63.1 (10.81)/62.2 (11.63) | 146/86 | 52 (22) | 46 (20) | 30 days |
| Kowey et al. (2009), ‘ACT II’ | 161 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3–72 h | Proportion converting to NSR within 90 min | 68.3 (7.7)/67.8 (6.4) | 121/40 | 129 (80) | 51 (32) | 30 days |
| Pratt et al. (2010), ‘ACT III’ | 172 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre. Cardiac Surgery | 3 h to 7 days | Proportion converting to NSR within 90 min | 60 (16)/60 (15) | 118/54 | 24 (14) | 23 (13) | 30 days |
| Roy et al. (2004), ‘CRAFT’ | 56 | Placebo, 0.5 mg/kg 2.0 mg/kg | Randomized, double-blind, multicentre | 3–72 h | Proportion converting to NSR within 30 min | 60.8/64 | 32/22 | – | Excluded | 7 days |
| Roy et al. (2008), ‘ACT I’ | 220 | Placebo, 3 mg/kg | Randomized, double-blind, multicentre | 3 h to 7 days | Proportion converting to NSR within 90 min | 60.4 (14)/59.9 (11.8) | 150/70 | 51 (23) | 19 (8.6) | 30 days |
| Simon et al. (2017) | 100 | Ibutilide, 3 mg/kg | Randomized, open-label, single centre. Emergency Department | No longer than 48 h | Time to conversion of AF to NSR | 56.2 (14.32)/56.7 (15.77) | 68/32 | 7 (7) | 99 (99) | Prior to discharge |
| Vogiatzis et al. (2017) | 78 | Ibutilide, 3 mg/kg | Pseudo-randomized, open-label, single centre | No longer than 48 h | Proportion converting to NSR within 90 min | 62.44 (7.24)/60.8 (14.1) | 56/22 | 31 (40) | – | – |
AF, atrial fibrillation; CAD, coronary artery disease; CHF, congestive heart failure; NSR, normal sinus rhythm; SD, standard deviation.
Subsequent doses are described in the Supplementary material online, Appendix.
Included studies
We identified nine trials including a total of 1358 patients (Table 1).11–19 Six trials compared vernakalant to placebo,11,12,14–17 two trials compared vernakalant to ibutilide,18,19 and one trial compared vernakalant to amiodarone.13 Seven trials were multicentre.11–17 One trial evaluated patients with AF in the perioperative period after cardiac surgery.14
Risk of bias
Patients and clinicians were blinded in seven of nine trials (Supplementary material online, Appendices S4 and S5).11–17 We found significant methodological bias in four of nine trials (one with pseudo-randomization,19 two with early termination,11,12 and two with significant missing outcome data).11,16 We did not detect reporting bias. We performed pre-specified sensitivity analyses to assess the robustness of estimates to risk of bias by dichotomizing trials according to their risk of bias.
Outcomes
Conversion within 90 min
Pooling data from nine trials (1325 participants, 454 events) (Figure 2A) demonstrated that vernakalant significantly increased the rate of conversion to sinus rhythm within 90 min (RR 5.15; 95% CI 2.24–11.84, I2 test for heterogeneity = 91%). Overall, 50% of vernakalant-treated patients converted within 90 min. Based on the GRADE framework, this was judged to be low quality evidence due to inconsistency and suspected publication bias (Supplementary material online, Appendices S7 and S8). In a sensitivity analysis including only the six trials that compared vernakalant to placebo, 48% of participants converted within 90 min and the effect estimate was RR 7.49; 95% CI 3.57–15.72, I2 = 63%, which was not significantly different than the comparison against active drug (Figure 2B, 56% vs. 24% conversion within 90 min; RR 2.40; 95% CI 0.76–7.58, I2 = 94%, P for interaction = 0.10). In a sensitivity analysis limited to the five studies at low risk of bias, the rate of conversion was significantly increased with vernakalant (Supplementary material online, Appendix S6, Figure A1b, RR 5.83; 95% CI 1.91–17.80, I2 = 91%). In another sensitivity analysis, vernakalant was less effective in the trial of patients with recent cardiac surgery (45% conversion, RR 3.03; 95% CI 1.54–5.94),14 as compared to the five trials that compared vernakalant to placebo and excluded patients after cardiac surgery (Supplementary material online, Appendix S6, Figure A1c, 49% conversion, RR 9.62; 95% CI 4.46–20.77, I2 = 48%, P for interaction = 0.03).11,12,15–17
(A) Relative risks of conversion within 90 min in trials comparing vernakalant to comparator in patients with recent-onset atrial fibrillation. (B) Relative risks of conversion within 90 min in trials comparing vernakalant to comparator in patients with recent-onset atrial fibrillation: subgrouped according to whether participants received placebo or active drug.
Conversion at 24 h
Six trials reported sustained sinus rhythm at 24 h, compared to placebo.11,12,14–17 However, three of these reported this outcome for the vernakalant group only.12,14,17 Among all participants who received vernakalant or placebo, vernakalant administration was associated with a higher rate of conversion at 24 h (Figure 3A, RR 7.47, 95% CI 1.13–49.55, I2 = 86%). Overall, 43% of patients who received vernakalant were in sinus rhythm at 24 h. Among patients who converted acutely, vernakalant administration was not associated with a higher rate of sustained conversion at 24 h (Figure 3B, RR 1.10, 95% CI 0.74–1.64, I2 = 26%) when compared with placebo or active comparator. Overall, 85% (95% CI 80–88) of vernakalant-treated patients who converted acutely remained in sinus rhythm at 24 h.
(A) Relative risks of sustained conversion at 24 h in trials comparing vernakalant to comparator in patients with recent-onset atrial fibrillation: all participants. (B) Relative risks of sustained conversion at 24 h in trials comparing vernakalant to comparator in patients with recent-onset atrial fibrillation: participants who converted acutely.
Serious adverse events
The definition of serious adverse events varied between studies but generally included tachyarrhythmia, bradycardia, heart failure, thromboembolic events, and death. After pooling data from the seven trials that reported on adverse events (1402 patients, 151 events), adverse events occurred in 11% of participants who received vernakalant and 11% of participants who received comparator. We found no significant difference between vernakalant and comparators in the rate of adverse events (Figure 4, RR 0.95; 95% CI 0.70–1.28; I2 = 0%). We judged the evidence to be of moderate quality due to imprecision (Supplementary material online, Appendix S8). In a sensitivity analysis limited to the five studies that compared vernakalant to placebo the risk was not significantly different from the comparison against active drugs (Supplementary material online, Appendix S6, Figure A2c, RR 0.90; 95% CI 0.66–1.22, I2 = 0% vs. RR 2.67; 95% CI 0.73–9.80, P for interaction = 0.11).
Relative risks of significant adverse events in trials comparing vernakalant to comparator in patients with recent-onset atrial fibrillation.
We found no significant difference between vernakalant and comparators in the risks of the following adverse events (Supplementary material online, Appendix S6, Figures A2e–A2g): ventricular arrhythmia (RR 0.61; 95% CI 0.31–1.22, P = 0.2), death (RR 1.89; 95% CI 0.42–8.48, P = 0.4), and heart failure (RR 0.94; 95% CI 0.15–5.93, P = 1.0). In a sensitivity analysis, the risk of adverse event was not significantly different for vernakalant as compared to placebo in patients undergoing cardiac surgery (9% with vernakalant, RR 0.84; 95% CI 0.32–2.19),14 as compared to other patients (13% with vernakalant, RR 0.90; 95% CI 0.65–1.25, I2 = 0%, P for interaction = 1.0).11,12,15,17
Additional subgroup analyses
We were unable to obtain study-level subgroup data on efficacy and safety in pre-specified subgroups of patients with a history of paroxysmal AF, ischaemic heart disease or with left ventricular dysfunction or heart failure.
Discussion
Summary
Key findings
In this systematic review and meta-analysis of randomized controlled trials, the administration of vernakalant was more effective for the rapid conversion of recent AF than placebo. Vernakalant was also more likely to maintain sinus rhythm at 24 h when compared with placebo. However, we found no significant difference in successful conversion when vernakalant was compared with active drugs (i.e. ibutilide/amiodarone). Finally, vernakalant carries a risk of serious adverse events that is similar to placebo, amiodarone, or ibutilide.
How do these findings fit in the literature?
Among major guidelines for AF, only the most recent European Society of Cardiology (ESC) guidelines make a recommendation regarding vernakalant. The most recent American College of Cardiology/American Heart Association/Heart Rhythm Society guidelines for the management of AF do not address the role of vernakalant as the drug is not approved for use in the USA.20 The most recent Canadian Cardiovascular Society guidelines list vernakalant among the possible options for pharmacological cardioversion of AF, but do not make a direct, evidence-based recommendation about its use.2 For patients and clinicians who prefer pharmacological cardioversion, the ESC guidelines assign vernakalant a Class I recommendation in patients with no relevant structural heart disease3; the findings of this review support this recommendation. These guidelines assign vernakalant a Class IIb recommendation in patients with coronary artery disease, moderate heart failure with reduced ejection fraction, heart failure with moderately reduced or preserved ejection fraction, and left ventricular hypertrophy.3 We were unable to perform pre-planned subgroup analyses for these subgroups due to the absence of study-level subgroup data. However, Torp-Pedersen et al.14,15,17 previously evaluated the ischaemic heart disease population using study level data from three trials included in this meta-analysis, one non-randomized open-label study,21 and a randomized trial in patients with atrial flutter known as ‘Scene 2’.22,23 This analysis included 274 adult patients (91 placebo, 183 vernakalant) with and 778 patients (224 placebo, 554 vernakalant) without known ischaemic heart disease. They found no difference in efficacy or risk of serious adverse events according to ischaemic heart disease status. There were no instances of torsades de pointes, ventricular fibrillation, or death in patients with ischaemic heart disease. Thus, there is no evidence to support withholding vernakalant in this population. Guideline recommendations appear to have not been based on the exclusion criteria of most trials. However, given that the absence of difference in safety in the trial that compared vernakalant to placebo following cardiac surgery that included patients after both coronary artery bypass and valvular surgery and where the majority of patients had an LVEF <50%, there is little basis for these reservations.14
Following the approval of vernakalant in Europe, the United States Food and Drug Administration (FDA) requested an additional Phase 3 study.11,24 This trial was suspended after a fatal case of cardiogenic shock in a participant receiving vernakalant. The trial’s independent data safety monitoring board assessed the case and recommended continuation of the study. However, the FDA requested that enrolment be paused while they reviewed the full data on this particular patient. Following a meeting with the FDA in December 2011, the sponsors and investigators deemed that the study would not meet regulatory expectations and it was terminated early.11
Interpretation of the results
For patients with recent-onset AF in whom rhythm-control is desired, options include pharmacological or electrical cardioversion. The latter is more resource intensive, usually requires the presence of two appropriately trained physicians and carries risks associated with procedural sedation.2 For clinicians and patients who prefer pharmacological conversion or for whom electrical cardioversion is not feasible, intravenous options include procainamide, ibutilide, amiodarone, and vernakalant.2 Upon review of the body of evidence, vernakalant has a favourable no significant difference in serious adverse events as compared to either placebo, ibutilide, or amiodarone. Given its ease of use and rapid time to conversion, vernakalant should be considered among the first line options for pharmacological conversion of recent-onset AF.
Strengths and limitations
We performed a rigorous and up-to-date systematic review and meta-analysis of randomized controlled trial evidence evaluating vernakalant for the conversion of short-term AF. Other systematic reviews and meta-analyses have been previously published on this or a similar question.25–27 Our review offers several important advantages over prior reports, including include pre-registration, inclusion of contemporary trials, and the exclusion of patients with atrial flutter or AF lasting more than 7 days. Additionally, we used the GRADE framework to evaluate the quality of the evidence. Thus, our results inform clinicians and guideline developers.
Our review also has several limitations. Although conversion within 90 min was the primary outcome of many of the vernakalant trials, it may not translate into patient-important outcomes such as maintenance of sinus rhythm or system-important outcomes such as shorter length of stay. Furthermore, differences in mean time to cardioversion could not be evaluated without individual patient-level data.
However, the durability of conversion (85% at 2 h) may be a reasonable surrogate for these outcomes. Our review was limited to the questions addressed in the published literature, therefore, we could not compare vernakalant with other commonly used cardioversion drugs, including Class IC antiarrhythmics (e.g. fleicainide). Because specific study-level subgroup data are lacking for patients with structural heart disease, we were unable to definitely address questions of efficacy or safety in this population. Finally, cost effectiveness was not evaluated in this analysis but may be affect choice of agent for pharmacologic cardioversion.
Conclusions
Vernakalant is effective for rapid and durable restoration of sinus rhythm in patients with recent-onset AF. Its efficacy appears similar to amiodarone and ibutilide. Vernakalant is likely safe with no significant difference in serious adverse events, including ventricular arrhythmias, as compared to either placebo, ibutilide, or amiodarone. Vernakalant should be a first line option for the pharmacological cardioversion of patients with haemodynamically stable recent-onset AF without severe structural heart disease.
Funding
W.F.M. holds personnel awards from the Canadian Institutes for Health Research (CIHR) and the Canadian Stroke Prevention Intervention Network (C-SPIN). R.P.W. holds a mid-career personnel award from the Heart and Stroke Foundation. É.P.B.-C. holds a personnel award from CIHR. W.F.M., B.D. and É.P.B.-C. are trainee members of the Cardiac Arrhythmia Network of Canada (CANet).
Conflict of interest: J.S.H. has received speaking fees from Cipher Pharmaceuticals Inc. All other authors have no conflicts to declare.
