Commentary on selected evidence

With advice from topic experts we selected 3 studies for further commentary.

Rate and rhythm control

We selected the individual patient data (IPD) meta-analysis by Kotecha et al. (2014) for a full commentary because it provides stronger evidence than the single randomised controlled trial (RCT) reviewed by the guideline committee for this question in NICE's guideline on atrial fibrillation (AF), and has a potential impact on recommendations.

What the guideline recommends

The guideline recommends either a standard beta‑blocker (that is, a beta‑blocker other than sotalol) or a rate‑limiting calcium‑channel blocker as initial monotherapy to people with atrial fibrillation who need drug treatment as part of a rate control strategy. It advises that the choice of drug should be based on the person's symptoms, heart rate, comorbidities and preferences when considering drug treatment.


Kotecha et al. (2014) conducted an IPD meta-analysis to assess the efficacy and safety of beta-blockers in patients with heart failure and concomitant atrial fibrillation. Trials which were not head to head trials were excluded, as were those with less than 300 participants or a planned follow up of 6 months or less. Data were obtained from 11 studies and included data from a total of 18,254 patients. Risk of bias was assessed for the included studies according to the Cochrane handbook. The primary outcome was all-cause mortality. Where deaths occurred after early study termination or following a fixed censor point, these were included in the analysis.


In total, 13,946 (76%) participants had sinus rhythm and 3,066 (17%) had AF at baseline. All-cause mortality was significantly reduced in patients with sinus rhythm receiving beta-blocker therapy (hazard ratio [HR] 0.73, 95% confidence interval [CI] 0.67 to 0.80, p<0.001). However, all-cause mortality was not significantly reduced in patients with AF (HR 0.97, 95% CI 0.83 to 1.14, p=0.73).

Subgroup analysis of patients with AF confirmed this non-significant result across the following variables:

  • age

  • sex

  • left ventricular ejection fraction

  • New York Heart Association class

  • heart rate

  • baseline medical therapy.

Furthermore, there were no significant benefits of beta-blockers in any of the secondary outcomes of cardiovascular death, composite clinical outcomes of death and cardiovascular hospitalisation, or non-fatal stroke.

Strengths and limitations

  • The study only included data from RCTs with a low risk of bias.

  • The study included a very large overall sample size.

  • Published and unpublished trials were identified through a comprehensive search of multiple sources.

  • The authors provided a detailed description of how IPD were requested, collected, checked and managed.

  • Patients with AF and atrial flutter were not analysed separately, but the impact of this was lowered due to only a small percentage of patients having atrial flutter.

  • The lower than expected rate of incident AF, particularly paroxysmal AF, indicates potential under-reporting.

  • The authors were unable to specify the type, duration and persistence of AF, which may have yielded additional subgroup data.

  • The authors did not report any quality assurance of study selection or risk of bias assessment.

Impact on guideline

This new evidence indicates that in patients with comorbid heart failure and AF, beta-blocker therapy may not reduce all-cause mortality. This is broadly consistent with the guideline advice to base the choice of treatment on comorbidities, amongst other factors. However, it may need to omit beta-blockers as one of the stated options for initial monotherapy in patients with comorbid AF and heart failure. The new evidence is considerably stronger than the single trial included in the evidence review for the guideline relating to beta-blockers in this context.

Topic experts also noted the uncertainty about the value ofbeta-blockers in people with both conditions, and considered the new evidence to have an impact on the recommendations.

Prevention and management of postoperative atrial fibrillation

We selected the RCT by Gillinov et al. (2016) for a full commentary because it had a large sample size, is directly relevant to the guideline review question on the best treatment strategy for postoperative AF (POAF), and has a potential impact. It was also highlighted by topic expert feedback as a pivotal study in reviewing the recommendation on rhythm control.

What the guideline recommends

NICE's guideline on AF recommends that, unless contraindicated, a rhythm‑control strategy should be offered as the initial management option for the treatment of postoperative atrial fibrillation following cardiothoracic surgery.


The RCT by Gillinov et al. (2016) compared rate control versus rhythm control strategies among 523 haemodynamically stable adults with new-onset AF or atrial flutter after cardiac surgery. The exclusion criteria were:

  • patients whose POAF did not persist for more than 60 minutes or

  • patients who did not have recurrent episodes of AF during the index hospitalisation (7 days or less after surgery) or

  • patients with a history of AF.

The primary outcome was the total number of days in the hospital (including emergency department visits) within 60 days following randomisation.

Secondary outcomes included:

  • the duration of the hospital stay from randomisation to the time of eligibility for discharge on the basis of criteria regarding AF

  • the length of the index hospitalisation

  • the need for readmission

  • heart rhythm

  • time to conversion to a sustained stable rhythm without AF

  • the need for permanent placement of a pacemaker

  • the rates of death and adverse events.


Approximately 76% of patients in the rhythm-control group and 73% of patients in the rate control group completed the full course of assigned treatment. The reasons why the majority of patients discontinued treatment were reportedly accounted for in the trial protocol.

There was no significant difference between the rate control and rhythm control groups in terms of the number of hospital days within the first 60 days of randomisation (mean 6.4 days and 7.0 days, respectively, p=0.76; median 5.1 days and 5.0 days, respectively).

For the secondary outcomes, there were no significant differences between the groups in terms of:

  • The rates of death (3 in the rate control group, 2 in the rhythm control group p=0.64).

  • Overall serious adverse events, including thromboembolic and bleeding events (24.8 per 100 patient-months in the rate-control group and 26.4 per 100 patient-months in the rhythm-control group, p=0.61).

  • Overall rates of cerebrovascular thromboembolism (0.8 per 100 patient-months in the rate-control group and 0.4 per 100 patient-months in the rhythm control group, p=0.40).

  • Non-cerebral thromboembolism (0.6 per 100 patient-months and 0.2 per 100 patient-months, p=0.31).

  • Serious bleeding rates (2.2 per 100 patient-months in the rate-control group and 1.2 per 100 patient-months in the rhythm-control group, p=0.21).

  • Major infections (9.3 per 100 patient-months in the rate-control group and 6.6 per 100 patient-months in the rhythm-control group, p=0.13).

  • Cardiac arrhythmias (4.7 and 6.2 per 100 patient-months, for rate-control and rhythm-control groups, respectively, p=0.30).

  • Pleural effusions (3.0 and 4.8 per 100 patient-months, for rate-control and rhythm-control groups, respectively, p=0.16).

Strengths and limitations

  • The RCT had a larger sample size and duration than the studies reviewed for this area in the guideline.

  • Intention to treat analysis was conducted to preserve the randomisation and to give a measure of effectiveness under realistic conditions.

  • The authors acknowledged that the primary outcome of total number of days of hospitalisation within 60 days after randomisation was only an indicator of important clinical outcomes, such as stroke and serious bleeding. The RCT was not sufficiently powered to detect differences in these outcomes.

  • The results only related to patients with new-onset POAF, and did not consider those with previous or existing POAF. The authors did not include information regarding the 1,414 patients who remained free of POAF, particularly preoperative medication. A secondary analysis of this data could inform the use of preoperative preventative therapy.

  • Treatment discontinuation was high. Intention to treat analysis partially addressed this but the high rate of switching reduced any real difference between the groups.

  • The duration of the trial was too short to include important outcomes of quality of life and longer term POAF occurrence.

Impact on guideline

The RCT by Gillonov et al. provides stronger evidence than the smaller studies reviewed for the guideline, which also included relatively short follow-up periods (ranging from less than 1 hour to 30 days). The new evidence indicates that initial strategies for rate control and rhythm control in treating POAF do not differ significantly and that neither treatment strategy has superiority over the other. There is a potential impact on NICE's guideline on AF to review the advice to offer a rhythm‑control strategy as the initial management option for the treatment of POAF following cardiothoracic surgery, unless contraindicated. Topic experts also highlighted the importance of the study and its potential impact in common clinical practice.

Short-term use of antiarrhythmic drugs following ablation

We selected the systematic review and meta-analysis by Chen et al. (2016) for a full commentary because it has a potential impact on the guideline to introduce new recommendations for the short term use of antiarrhythmic drugs (AADs) post ablation. The review also includes a sequential analysis of RCTs to assess whether further studies are required to demonstrate a sufficient effect size.

What the guideline recommends

NICE's guideline on AF advises against the use of amiodarone for long term rate control but does not make specific recommendations relating to AADs post ablation.


The systematic review and meta-analysis by Chen et al. (2016) assessed the efficacy of short term AADs use compared with no-AADs after catheter ablation of AF in preventing atrial arrhythmia recurrence.

Inclusion criteria were published RCTs that included:

  • AF patient undergoing catheter ablation with pulmonary vein isolation (PVI)-based strategy.

  • The intervention of short-term administration of AADs within 3 months following ablation of AF.

  • The comparison intervention of no-AADs prescription after ablation procedure.

  • The primary outcomes of early recurrence of atrial arrhythmia lasting more than 30 days within the first 3 months after ablation.

  • The secondary outcome of late recurrence of atrial arrhythmia lasting more than 30 days post-3 months after ablation.


The 7 studies included in the meta-analysis provided data from 6 RCTs and a total of 2,667 patients, with sample sizes ranging from 74 to 2,038.

Application of GRADE categorised evidence for short term AADs versus no AADs post ablation as very low for the primary outcome, and as moderate for the secondary outcome. The evidence was partly downgraded because of the high risk of bias in the majority of trials. The lack of blinding of participants and personnel was the major contributor to the high risk of bias.

For the primary outcome, short-term use of AADs after AF ablation, compared to no AADs, significantly reduced the risk of early recurrence of atrial arrhythmia (Relative Risk [RR] 0.68, 95% CI 0.52 to 0.87, p=0.003).

A sensitivity analysis was performed to address moderate heterogeneity. This comprised the consecutive exclusion of one single trial each time, but did not affect the direction or significance of the results.

A further trial sequential analysis was conducted, in order to assess whether sufficient cumulative evidence was included to demonstrate a conclusive effect size. The results indicated that an adequate number of studies were synthesised to demonstrate a 25% reduction in the relative risk of early AF recurrence with short-term AADs after ablation.

For the secondary outcome of late recurrence of atrial arrhythmia, short-term use of AADs after AF ablation, compared to no AADs, did not significantly reduce the risk of late recurrence of atrial arrhythmia (RR 0.92, 95% CI 0.83 to 1.03; p=0.15).

Strengths and limitations

  • Only randomised controlled trials were included, which reduced the risk of bias among included studies.

  • Quality assurance was undertaken for study selection, data extraction and risk of bias assessment, involving independent reviewers. The quality of evidence for primary and secondary outcomes was independently evaluated by two reviewers according to the GRADE methodology, and the review reporting complied with established guidelines.

  • A trial sequential analysis was conducted to assess whether the included evidence was conclusive.

  • The authors acknowledged that heterogeneity of ablation procedures, follow-up periods, physician experience, and antiarrhythmic drugs used among the included studies was not fully addressed in the systematic review.

  • Most included trials were not blinded, which may have introduced performance and detection bias. However, this risk was mitigated by the predefined objective outcomes.

  • Only published studies were included. Harbord and Peters tests were used to analyse publication bias, and this was detected. It is possible that if unpublished studies were included in the review they may have influenced the results.

Impact on guideline

The new evidence indicates that the use of AADs following catheter ablation may reduce the incidence of early recurrent AF within 3 months, but did not find evidence of preventing late recurrence of AF beyond 3 months of ablation. The guideline did not review evidence for AADs post ablation and does not include specific recommendations in this area. There is therefore a potential impact to consider a new recommendation for the short term use of AADs following ablation. This was confirmed by the majority of topic experts, who also indicated the need to standardise the approach in clinical practice. It was further noted by topic experts that clinicians are likely to use AADs in the early post ablation period regardless of whether or not they reduce long-term recurrence of AF.

This page was last updated: 12 September 2017