Rhythm Versus Rate Control for Atrial Fibrillation : A Meta-analysis of Randomized Controlled Trials

Published by Oriental Scientific Publishing Company © 2018 This is an Open Access article licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License (https://creativecommons.org/licenses/by-nc-sa/4.0/ ), which permits unrestricted Non Commercial use, distribution and reproduction in any medium, provided the original work is properly cited. Rhythm Versus Rate Control for Atrial Fibrillation: A Meta-analysis of Randomized Controlled Trials

Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia, affecting more than 5% of the worldwide population 1 .It is associated with a high risk of thromboembolic events, including stroke, which occurs in about 23% of AF patients, older than 80 years 2,3 .Over the last decade, it accounted for about one third of hospital admissions for cardiac arrhythmias 4,5  with an increasing prevalence in patients with cardiovascular problems, such as valvular heart disease, heart failure (HF), and coronary artery disease (CAD) 6,7 .
The pharmacological management of AF targets either rate control (maintaining the heart rate at normal levels, using pharmacological agents, such as beta-blockers, non-dihydropyridine calcium-channel blocker, and cardiac glycosides) or rhythm control (restoration of sinus rhythm, using electrical cardioversion and/or antiarrhythmic agents, such as sodium channel blockers and potassium channel blockers) 8 .In the past few years, several randomized controlled trials (RCTs) have investigated whether rhythm control is superior to rate control with respect to mortality and cerebrovascular accidents 9-22 .
Besides the controversial results of these trials, former meta-analyses showed conflicting results, suggesting that rate control is either similar or superior to rhythm control in terms of mortality and stroke rates 23,24 .Moreover, recent trials have compared both strategies in different groups of AF patients, including younger and those with concomitant HF 9,12 .Therefore, we conducted this systematic review and meta-analysis to update the evidence regarding the optimal control approach for AF.

Methods
This study was conducted following the guidelines of the Cochrane handbook of systematic reviews of interventions and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement 25,26 .All steps have been prespecified in a published protocol on the PROSPERO register of systematic reviews (CRD42016049648).

Literature search strategy
We performed a comprehensive search of PubMed, Scopus, Web of Science (ISI), Embase, OVID and Cochrane Central register of controlled clinical trials (CENTRAL), during September 2016, to identify relevant studies.We developed the search strategy for each database using the following terms: "Atrial fibrillation", "Rate control", "Beta blockers", "Calcium channel blockers", "Antiarrhythmic", "Cardioversion", and "Rhythm control" (Supplementary file 1).No publication period or language restrictions were applied during literature search.We also checked the bibliography of included studies and searched the clinical trials registry (Clinicaltrials.gov)for any ongoing trials.

eligibility Criteria and study selection
We included all RCTs that compared the efficacy of rate control versus rhythm control strategies, including non-invasive procedures of electrical cardioversion, in AF patients.We excluded trials on other types of atrial arrhythmia, such as atrial flutter, reviews, non-randomized trials, observational, and studies from which data could not be reliably extracted.
Three reviewers independently screened the retrieved titles and abstracts for matching our criteria.Then, the eligible abstracts underwent further full-text screening for eligibility to meta-analysis.Unrelated or duplicate reports were removed and multiple reports for the same trial were linked together as one study.All disagreements were solved by discussion between the reviewers.

data extraction
Data was extracted from included studies by one reviewer and checked by another one.The extracted data included the following: a) baseline characteristics of enrolled patients, b) risk of bias assessment domains, and c) main outcomes including the incidence of all-cause mortality, cardiovascular mortality, arrhythmic mortality, stoke or transient ischemic attack (TIA), systemic embolism, HF or worsening of HF, major or life threatening bleeding, re-hospitalization and subsequent myocardial infarction (MI).

Risk of Bias Assessment
Two independent reviewers assessed the risk of bias in included trials, using the Cochrane risk of bias (ROB) assessment tool 25 .This tool is designed to detect six types of bias: selection bias, performance bias, detection bias, attrition bias, reporting bias, and other potential sources of bias.The authors classified the included studies in each domain as of low, high, or unclear risk of bias.The risk of publication bias was assessed, using funnel plot-based methods, whenever 10 or more studies reported on the same outcome 27 .

data synthesis
The statistical analyses were performed using the RevMan software (version 5.3 for windows), provided by the Cochrane Collaboration.
Under the fixed-effect model, dichotomous data were pooled as risk ratios (RRs) with their 95% confidence interval (CI), using the Mantel-Haenszel (M-H) method.The existence of heterogeneity was assessed by Chi-square test and its extent was measured by I-square test.In case of significant heterogeneity (Chi-square p < 0.1), the analysis was conducted under the randomeffects model.Sensitivity analysis was performed to resolve significant heterogeneity and to ensure that our results were not affected by the weights of individual studies.We also conducted subgroup analyses to compare both strategies in patients with HF or under 65 years of age (by pooling studies in which the mean age was less than 65 years).

Literature search and screening process
A comprehensive database search retrieved 3879 unique records.Following title and abstract screening, 74 full-text articles were retrieved for assessment of eligibility to metaanalysis.Finally, we included 12 RCTs (14 full text articles: 8451 patients) that compared rate control to rhythm control in AF patients (9-22) (Figure 1).Table 1 displays a summary of the used drugs and main findings of included studies and Table 2 shows baseline characteristics of enrolled patients in these studies.

Risk of Bias Assessment
All included studies had a low risk of selection (random sequence generation), attrition, and reporting biases, except for the AFFIRM trial 20,21 (which did not clarify their randomization method), as well as CAFÉ II ( 16) and RACE 15,22  trials (which had an unclear risk of reporting bias).None of the included studies achieved or reported on blinding of patients and personnel or the blinding of outcome assessors.Figure 2 shows a summary of the results of ROB assessment for each included study and the details of authors' judgements are illustrated in Supplementary file 2.

subgroup analysis heart failure patients (Grade II -IV)
Subgroup analysis of data from three trials (1637 patients) (9,12,16), collected from patients with grade II to IV HF, showed no significant difference between both groups in terms of

sensitivity analysis
All the effect-estimates remained robust when we removed the two largest studies (AF CHF and AFFIRM), except for all-cause mortality.Upon removal of AF CHF and AFFIRM trials, which reported a non-significant increase in allcause mortality in the rhythm control group, the Fig. 3. Forest plots of risk ratios for A) All-cause mortality, B) Cardiovascular mortality, and C) Arrhythmic mortality effect estimate favored rate control over rhythm control (RR = 1.66, 95% CI [1.15, 2.39], p = 0.006) regarding this particular outcome.The detailed results of sensitivity analysis are shown in Table 3.

dIsCussIon
Our meta-analysis of 12 studies showed no significant difference between rate and rhythm control groups in terms of mortality rates and other major clinical outcomes (including bleeding and thromboembolic events), except for rehospitalization rate, which was significantly higher in the rhythm control group.Across the five studies (9,12,15,19,20) that investigated arrhythmic, cardiovascular, and all-cause mortality, cardiovascular mortality represented 63.8% of allcause mortality, while arrhythmic death represented 45.5% of cardiovascular and 29% of all-cause mortality.
Although restoring a physiological cardiac rhythm is hypothesized to lower the risk of   mortality and embolic events, our analysis shows that neither strategy was superior to the other in these regards.A possible explanation is that the survival benefit of rhythm control is likely to be negated by the non-cardiac side-effects of antiarrhythmic drugs (28,29).The higher frequency of rehospitalization in the rhythm control group may be explained by the possible occurrence of dysrhythmias, as a complication of anti-arrhythmic drugs and the need to perform cardioversion in a monitored environment (20,30).This higher rate of rehospitalization can be translated to a higher treatment cost in the rhythm control group, which is confirmed by real-world data from observational studies (31)(32)(33).
We performed a subgroup analysis for AF patients with concomitant CHF because this comorbidity affects more than 50% of AF patients and its interaction with AF means that none of them can be optimally managed without treating the other (9,12).Except for rehospitalization rate, mortality and clinical outcomes' results were similar to those of the main analysis.This may be explained by the fact that these patients require frequent hospitalization for management of CHF, regardless the method of AF control.
Interestingly, rhythm control strategy was associated with lower rates of mortality, HF, and major bleeding than rate control in younger patients (mean age below 65 years), probably by delaying the progression to permanent AF, which has a higher rate of complications (23).This finding is supported by real life data from the RECORDAF registry (Registry on Cardiac Rhythm Disorders AF), established following the AFFIRM trial (34).
Postoperative AF occurs in 20% to 50% of patients following cardiac surgery (35,36).An included study by Gillinov et al. showed no significant difference between rate and rhythm control strategies in terms of mortality and complication rates in postoperative patients (10).Additionally, about 17 to 18% of rheumatic patients develop AF (8).The included CRAFT trial showed that rhythm control was superior to rate control in rheumatic heart patients in terms of reducing mortality and improving quality of life and exercise capacity (17).

strength points
Compared to the formerly mentioned meta-analyses (23,24), our analysis included a larger number of trials, with a fairly higher sample size.We performed subgroup analyses for younger patients and those with HF and conducted a sensitivity analysis to ensure that our results were not affected by the weights of individual studies.Unlike previous meta-analyses, we performed a thorough risk of bias assessment and investigated for publication bias, whenever appropriate.

Limitations
All included trials were open-label studies because the nature of electric cardioversion in the rhythm control group prevents proper blinding and applying a fake electrical cardioversion protocol is ethically controversial and would interfere with the results of other outcomes, such as rehospitalization rate.The main weight of our analysis comes from the two largest trials (AF CHF and AFFIRM); therefore, we performed a sensitivity analysis by excluding these trials to overcome this limitation.We did not assess the impact of either strategy on quality of life outcomes because these data were poorly reported in included studies.Future trials should further investigate the effect of other comorbidities, such as stroke and left ventricular dysfunction on the treatment outcomes.We are aware of few ongoing studies, comparing both strategies, in different categories of AF patients, such as AFARC-LVF trial (NCT02509754) and RAFT-AF trial (NCT01420393).

ConCLusIon
In older AF patients and those with concomitant CHF, both rate and rhythm control strategies have similar rates of mortality and major clinical outcomes; therefore, choosing an appropriate therapeutic strategy should consider individual variations such as patient preferences, comorbidities, and treatment cost.Future trials should compare both strategies in younger patients and those with other comorbidities such as stroke and left ventricular dysfunction.

Fig. 4 .
Fig. 4. Forest plot of risk ratios for A) Stroke/Transient Ischemic Attack, B) Systemic embolism, and C) Development or worsening of heart failure

table 1 .
Shows a summary of the used drugs and main findings of included trials Atrial Fibrillation Follow-up Investigation of Rhythm Management study, AF CHF: Atrial Fibrillation and Congestive Heart Failure study, BB: Beta-Blockers, CAFÉ-II:controlled study of rate versus rhythm control in patients with chronic AF and HF, CHF: congestive heart failure, HF: Heart failure, HOT CAFÉ: How to Treat Chronic Atrial Fibrillation study, J RHYTHM: Japanese Rhythm Management Trial for Atrial Fibrillation, NYHA: New York Heart Association, PIAF: Pharmacological Intervention in Atrial Fibrillation study, QoL: Quality of life, RACE: Rate Control versus Electrical Cardioversion for Persistent Atrial Fibrillation study, RHD: rheumatic heart disease, STAF: Strategies of Treatment of Atrial Fibrillation study.

table 2 .
Shows baseline characteristics of enrolled patients in included studies Atrial Fibrillation Follow-up Investigation of Rhythm Management study, AF CHF: Atrial Fibrillation and Congestive Heart Failure study, CAFÉ-II:controlled study of rate versus rhythm control in patients with chronic AF and HF, HOT CAFÉ: How to Treat Chronic Atrial Fibrillation study, J RHYTHM: Japanese Rhythm Management Trial for Atrial Fibrillation, PIAF: Pharmacological Intervention in Atrial Fibrillation study, RACE: Rate Control versus Electrical Cardioversion for Persistent Atrial Fibrillation study, STAF: Strategies of Treatment of Atrial Fibrillation study.

table 3 .
Results of sensitivity analysis after excluding the largest two studies (AF CHF and AFFIRM)