There was no difference in mortality when rate control versus
rhythm control was used in the management of atrial fibrillation
Clinical Problem: A 66 year old man, who has had a
prior stroke, has an episode of atrial fibrillation, recorded on EKG, lasting
more than six hours.
Clinical Question: Does
rhythm control decrease mortality relative
to rate control in the treatment of atrial fibrillation?
Search Strategy:
Keywords: (MeSH terms) “atrial fibrillation” AND
“heart rate” AND “cardioversion”
Limits: RCT or Meta-analysis, Human, English, 5
years
Sources: PubMed, DARE, SumSearch, Cochrane
Results: 21 articles, three of which were
relevant.RCTs: The HOT CAFE trial (ref. 1) examined management of persistent atrial
fibrillation (>7d); the RACE trial
(ref. 2) examined management of recurrent persistent atrial fibrillation or
flutter (>10d); the AFFIRM trial
(ref. 3, 4) was selected for review as it was broader, allowing inclusion of
patients with new-onset or episodic atrial fibrillation and examining
management after as few as six hours of atrial fibrillation. No relevant meta-analyses were found. One set of guidelines from the ACC/AHA/ESC
was found (ref. 5), but these guidelines were published before the three
relevant RCTs and were therefore not considered up-to-date.
Clinical
Bottom Lines:
1.
A rhythm control strategy was neither superior nor inferior to a rate
control strategy in preventing mortality in patients with atrial fibrillation.
(RR=1.15; 95%CI from 0.99 to 1.34)
2.
Comparison of the incidence of pre-specified adverse events between
rate- and rhythm-controlled groups showed either nonsignificant differences or
a benefit with rate control.
3.
The risk of ischemic stroke in a patient with atrial fibrillation is
6.3% overall (at five years) with no significant difference between management
strategies.
The
Evidence: The AFFIRM study is an unblinded, randomized,
multicenter treatment study involving 4060 patients, which compares a rhythm
control strategy to a rate control strategy for the management of atrial
fibrillation. All patients were aged 65
or older, or had at least one risk factor for stroke. The primary endpoint was mortality; a
composite secondary endpoint comprised death, disabling stroke, disabling
anoxic encephalopathy, major bleeding, and cardiac arrest.
Data:
Number of patients enrolled:
4060
Study period: November 1995 –
October 1999, at 213 centres
Average duration
of follow-up: 3.5 years (range 2 years to 6 years)
Rate control Rhythm control Total
Number dead:
310 356 666
Number alive:
1717 1677 3394
Total: 2027 2033 4060
RR = (356/2033) /
(310/2027) = 1.15; 95%CI = 0.99 to 1.34; P = 0.08
ARR = N/A RRR = N/A
Comments:
1. Well-designed trial: Randomized, ITT analysis, low dropout rate,
subgroup analysis pre-specified
2.
Not a blinded trial.
3.
This trial compared strategies for management of atrial fibrillation, not specific
drugs.
4.
Kaplan-Meier curves suggest a trend toward increased
mortality with time in the rhythm-controlled arm, but the effect does not
achieve significance.
5.
The authors stress that anticoagulation should be
continued in all patients with atrial fibrillation.
6.
It is not clear where the authors enroll their
patients from – clinics, emergency rooms, etc.
7.
The authors do not describe how many patients in each
group have had prior strokes.
8.
The HOT CAFÉ and RACE trials show similar findings to
the AFFIRM trial.
9.
The methodology and study design of the AFFIRM trial
are described in much more detail in a previously published paper (see
Reference 4, below).
References:
5. http://www.acc.org/clinical/guidelines/atrial_fib/af_index.htm
Key
Words: atrial fibrillation, arrhythmia, rate control, rhythm control,
cardioversion
Appraiser: Alex Fraser and the UWO Evidence Based Neurology Group
Date
Appraised: May 2004
Copyright 2002-2004
Evidence Based Neurology Group
University of Western Ontario