Summary
Atrial fibrillation is the most frequently occurring sustained cardiac arrhythmia and is associated with a significantly increased risk of thromboembolic events and death. We sought to compare the clinical efficacy of rate and rhythm control strategies in patients with non-postoperative atrial fibrillation. We searched the PubMed database and the Cochrane Central Register of Controlled Trials for randomized controlled trials comparing rate versus rhythm control in patients with atrial fibrillation. Studies were retrieved and we analysed major clinical outcomes. Risk ratios (RRs) and 95% confidence intervals were calculated assuming random effects due to the clinical heterogeneity of the study populations. Eight randomized controlled trials were identified, with a total of 7499 patients with atrial fibrillation. There were no significant differences in the effects of rate and rhythm control on any outcome: all-cause mortality (RR: 0.95; CI: 0.86–1.05), cardiovascular mortality (RR: 0.99; CI: 0.87–1.13), arrhythmic/sudden death (RR: 1.12; CI: 0.91–1.38), ischaemic stroke (RR: 0.89; CI: 0.52–1.53), systemic embolism (RR: 0.89; CI: 0.69–1.14) and major bleeding (RR: 1.10; CI: 0.89–1.36). Updated data pooled from a large population of patients with atrial fibrillation suggests that rate and rhythm control strategies have similar effects on major clinical outcomes. Other factors, including individual preferences, comorbidities, drug tolerance and cost issues, should be considered when choosing the approach for these patients.
Résumé
La fibrillation atriale est l’arythmie cardiaque soutenue la plus fréquente et est associée avec une augmentation du risque de complication thromboembolique et de décès. Notre objectif a été de comparer l’efficacité clinique des stratégies contrôles du rythme versus contrôle de la fréquence cardiaque chez des patients en fibrillation atriale en excluant la phase postopératoire. Notre recherche a été centrée sur les bases de données PubMed et CENTRAL, avec les mots clés « essais randomisés contrôlés » comparant les stratégies de contrôle de la fréquence du rythme cardiaque chez les patients en fibrillation atriale. Les études ont été analysées pour ce qui concerne les évènements cliniques majeurs. Le risque relatif et les intervalles de confiance à 95 % ont été calculés, en retenant l’hypothèse d’effets au hasard dus à l’hétérogénéité clinique des populations incluses dans les différentes études. Huit essais randomisés contrôlés ont été identifiés, incluant un total de 7499 patients en fibrillation atriale. Il n’y avait pas de différence significative entre les deux stratégies pour ce qui concerne le taux de mortalité globale (RR : 0,95 ; IC 95 % : 0,86–1,05), la mortalité cardiovasculaire (RR : 0,99 ; IC 95 % : 0,87–1,13), la mort subite ou arythmique (RR : 1,12 ; IC 95 % : 0,91–1,38), le taux d’accident ischémique cérébral (RR : 0,89 ; IC 95 % : 0,52–1,53), d’embolie systémique (RR : 0,89 ; IC 95 % : 0,69–1,14) et le taux de saignement majeur (RR : 1,10 ; IC 95 % : 0,89–1,36). Cette méta-analyse, mise à jour des précédentes, basée sur une population importante de patients en fibrillation atriale suggère que les deux stratégies: contrôle du rythme ou de la fréquence cardiaque dans la fibrillation atriale, ont des effets similaires sur les évènements cardiovasculaires majeurs. D’autres facteurs prenant en considération les préférences individuelles, les comorbidités, la tolérance des médicaments ou encore les coûts devraient être pris en considération lorsqu’une des deux stratégies est retenue dans la prise de ces patients.
Introduction
AF is the most common sustained cardiac arrhythmia. Risk factors for this condition include male sex, smoking, heart failure, diabetes, arterial hypertension, left ventricular hypertrophy, valvular heart disease, myocardial infarction and advanced age. The lifetime risk of developing AF in subjects aged greater than 40 years is around 25% .
AF is associated with increased cardiovascular morbidity and mortality. In addition to the deleterious haemodynamic effects that can trigger or worsen heart failure, AF also predisposes to thromboembolic events; stroke incidence increases with age, with the 5-year risk rising from 1.5% in the 6th decade of life to 23.5% in those aged more than 80 years . AF is an independent risk factor for death .
The pharmacological treatment of AF follows one of two strategies: rate control (controlling the ventricular rate with beta-blockers, non-dihydropyridine calcium-channel blockers and/or digitalis); or rhythm control (restoring and maintaining sinus rhythm with electrical cardioversion and/or antiarrhythmic agents) . Rhythm control maintenance with antiarrhythmic drugs can improve symptoms and reduce the incidence of stroke but may lead to an increased risk of adverse events due to negative inotropic and proarrhythmic effects . The rate control strategy may have fewer adverse events . Antithrombotic agents are usually used to prevent thromboembolic events in association with rate control drugs, whereas their use after sinus rhythm is restored in rhythm control depends on the patient’s risk factors.
Our aim was to systematically review RCTs and compare the relative effects of rate and rhythm control in AF. We proposed to analyse clinically relevant outcomes: all-cause mortality, cardiovascular mortality, arrhythmic/sudden death, ischaemic stroke, systemic embolism and major bleeding.
Methods
Searching
A search strategy was developed in September 2011 using the PubMed and Cochrane Central Register of Controlled Trials (CENTRAL) databases for study identification. There were no language limitations. References of obtained studies were also comprehensively searched.
The electronic search for RCT publication type was undertaken using the following keywords: rate, beta-blocker, acebutolol, atenolol, bisoprolol, carvedilol, esmolol, metoprolol, nadolol, pindolol, propranolol, timolol, calcium-channel blocker, verapamil, diltiazem, digitalis, digoxin, rhythm, amiodarone, sotalol, propafenone, disopyramide, dofetilide, flecainide, ibutilide, dronedarone and azimilide, mortality, death, stroke, embolism, thromboembolic, thromboembolism, bleeding and atrial fibrillation. Boolean operators AND and OR were used to combine terms. Postoperative and postpercutaneous intervention studies were excluded from the search using the Boolean operator NOT.
Selection
We included RCTs comparing pharmacological approaches to maintaining rate or rhythm control in patients with AF. Quasi-randomized studies and prospective cohorts were not included. The patients’ mean age had to be greater than 55 years due to the low risk of death or stroke in patients with AF under this age . To determine the efficacy of the interventions as primary strategies, the trials had to have intention-to-treat analysis or provide data that allowed this type of calculation. The analysed outcomes were all-cause mortality, cardiovascular mortality, arrhythmic/sudden death, ischaemic stroke, systemic embolism (combining ischaemic stroke and other systemic embolic events) and major bleeding.
Validity assessment
The PEDro score was used to assess the methodological quality of data reporting in the studies . This score is based on the presence/absence of 11 items: eligibility criteria, random allocation, allocation concealment, similar baseline characteristics, blinding of all subjects, blinding of therapists, blinding of outcome assessors, crossover rate less than 15%, intention-to treat analysis, statistical comparisons between groups and measures of variability.
Data abstraction
Studies that met the criteria outlined above were assessed by one review author and checked by another. AF population, study size, follow-up, primary outcome, demographic characteristics and comorbidities were retrieved. Anticoagulation therapy, and rate and rhythm interventions were also sought. Data entry into software was also double checked. All disagreements were solved by consensus.
Quantitative data synthesis
The statistical analyses were performed using the RevMan software (version 5.1.4) provided by the Cochrane Collaboration when more than one trial had data for pooled analysis. Dichotomous outcomes were analysed by the Mantel-Haenszel method. RR and 95% CI were calculated. When zero cells were present in one arm, RevMan automatically added 0.5 to them to perform the calculations. The results estimates were based on a fixed-effects model or a random effects model, depending on heterogeneity. Statistical heterogeneity was assumed if I 2 was greater than 50%. Clinical heterogeneity was analysed by authors according to clinical differences between study patients. Publication bias assessment with a funnel plot was planned if more than 10 studies were retrieved.
Sensitivity analysis was planned according to baseline heart failure, age, anticoagulation treatment, mean follow-up and study size.
Results
One hundred and fifty-seven studies were found in the database search and 146 studies were excluded: 11 were not RCTs; 36 were not AF studies; 73 citations reported trials that did not compare pharmacological rate versus rhythm control in AF; and 26 records were substudies of eligible trials. Eleven RCTs remained and three were excluded due to lack of data necessary for analysis based on intention-to-treat.
Eight RCTs were included for meta-analysis . Data not provided in the main papers were sought from the RACE and AFFIRM post-hoc studies ( Fig. 1 ) .
Included studies enrolled a total of 7499 AF patients with a mean age of 68 years. In all these trials the majority of patients were men (63.4–82.0%). Prevalence of hypertension ranged from 42.8 to 64.3%, valvular disease from 4.9 to 17% and coronary disease from 7.4% to 43.5%. The AF-CHF study and the CAFÉ-II study included only heart failure patients. The PIAF study provided no heart failure data. In the other trials, the prevalence of heart failure ranged from 3.6 to 70%. Weighted mean follow-up was 2.9 years, ranging from 1 year (PIAF) to 3.5 years (AFFIRM).
The PEDro quality score (scale range 1–11) for the obtained trials varied between 6 and 7. None of the trials described allocation concealment methods and interventions were unblinded for patients and physicians. The RACE study was the only one that reported blinding of the researchers who recorded the outcomes. The PIAF and RACE studies had differences in relevant baseline characteristics. The AFFIRM and AF-CHF studies and the J-RHYTHM had crossover rates greater than 15%. Table 1 details the main characteristics of each study and Table 2 contains their conclusions.
| PIAF | RACE | AFFIRM | STAF | HOT CAFÉ | AF-CHF | J-RHYTHM | CAFÉ-II | |
|---|---|---|---|---|---|---|---|---|
| AF population | New-onset AF present for ≥ 7 days and < 1 year | Recurrent and persistent AF or flutter for < 1 year | Likely to be recurrent AF in pts aged > 65 years with risk factors for stroke or death | Recurrent and persistent AF present for ≥ 4 weeks and < 2 years with ≥ 1 previous cardioversion | AF present for ≥ 7 days and < 2 years | LVEF ≤ 35%, symptoms of CHF and history of paroxysmal or persistent AF for < 1 year | Paroxysmal and persistent AF for < 1 year | Persistent AF with chronic HF and NYHA ≥ II |
| No. of pts | 252 | 522 | 4060 | 200 | 205 | 1376 | 823 | 61 |
| Mean age (SD) | 61 (10) | 68 (9) | 70 (9) | 66 (9) | 61 (11) | 67 | 64.8 (11) | 72 (7) |
| Men (%) | 73 | 63.4 | 60.7 | 63.5 | 65.4 | 82 | 69.3 | 84 |
| Hypertension (%) | 48.8 | 50 | 50.8 | 62.5 | 64.3 | 48 | 42.8 | 70 |
| Valvular disease (%) | 16.2 | 17 | 4.9 | 13 | 15.1 | 5 | 5.6 | N/A |
| HF (%) | N/A | 50 | 23.1 | 55.5 (NYHA ≥ II) | 70 | 100 | 3.6 | 100 |
| CHD (%) | 23.4 | 27 | 26.1 | 43.5 | 43.9 | 48 | 7.4 | 50 |
| Recommended anticoagulation | All pts anticoagulated (INR 2–3) | Acenocoumarol or fenprocoumon 4 weeks before and after electrical cardioversion. Rate arm anticoagulated if age > 65 years or cardiac disease | Both arms anticoagulated; if sinus rhythm restored after 4 to 12 weeks with antiarrhythmic agents, anticoagulation could be stopped | ACCP guidelines | ACCP guidelines | Recommended for all pts: ACC/AHA/ESC 2006 AF guidelines | Modified AFFIRM protocol for non-valvular AF; Japanese guidelines on AF management for valvular AF | Recommended for all pts: warfarin (INR 2–3) |
| Anticoagulated pts: rate vs rhythm (%) | N/A | 96 vs 86 | > 85 vs 70 | N/A | 74.3 vs 15.6 | 90 pts received oral anticoagulant | 59.4 vs 60.1 | 98 |
| Rate intervention | Diltiazem as initial drug | Digitalis, non-dihydropyridine calcium-channel blocker and beta-blocker | Beta-blocker, non-dihydropyridine calcium-channel blocker and digoxin | Beta-blocker, digitalis and calcium-channel blocker; AV nodal ablation or modification ± pacemaker | Beta-blocker, non-dihydropyridine calcium-channel blocker and digoxin; cardioversion or AV nodal ablation and pacemaker | Beta-blocker and digoxin; AV nodal ablation and pacemaker | Beta-blocker, calcium-channel blocker or digitalis | Digoxin and beta-blocker |
| Beta-blocker in rate control (%) | N/A | 41 | 68.1 | N/A | 89.1 | N/A | 51.5 | 90 |
| Rhythm intervention | Amiodarone or electrical cardioversion followed by amiodarone | Electrical cardioversion and sotalol, flecainide, propafenone or amiodarone | Many antiarrhythmic agents and/or electrical cardioversion | Electrical cardioversion and class I antiarrhythmic agents or sotalol; CHD or LV dysfunction, amiodarone | Electrical cardioversion and disopyramide, propafenone, sotalol or amiodarone | Electrical cardioversion and amiodarone (or sotalol or dofetilide) | Electrical cardioversion and antiarrhythmic agents (pilsicainide, cibenzoline, propafenone, disopyramide) | Amiodarone and electrical cardioversion |
| Amiodarone in rhythm control (%) | 100 | N/A | 62.8 | 42 | 56.7 | 82 | 0.5 | 80 |
| Mean years of follow-up (SD) | 1 | 2.3 (0.6) | 3.5 | 1.6 (0.7) | 1.7 (0.4) | 3.1 (1.6) | 1.58 | 1.2 a |
| PEDro score | 6 | 7 | 6 | 7 | 7 | 6 | 6 | 7 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
