EBM Conference
Friday May 14, 2010
Andrew Smitherman
Background:
- Atrial fibrillation is the most common sustained cardiac arrhythmia.
- AFFIRM Trial (12/2002): No increased overall mortality with rate control and
anticoagulation when compared to rhythm control.
- With equivalent mortality between rate control and rhythm control, rate control may
be considered the first-line, more easily achieved goal. The optimal level of rate control has not been determined – whether strict or lenient heart rate control is associated with fewer adverse effects.
Study:
Lenient versus Strict Rate Control in Patients with Atrial Fibrillation. Van Gelder et al. NEJM 2010; 362 (15): 1363-1373.
Hypothesis:
Lenient rate control is not inferior to strict rate control in patients with permanent atrial fibrillation.
Design:
- Prospective, multi-center (33 centers), randomized, open-label non-inferiority trial funded by the
Netherlands Heart Foundation and unnamed pharmaceutical and device companies.
- Enrollment from Jan 2005 – June 2007 and data collected until June 2009
- Independent study monitoring, data management and validation
Population:
- Patients in the Netherlands with permanent atrial fibrillation for 12 months or less
- Inclusion Criteria:
- ≤ 80 years-old
- mean resting HR > 80
- current anticoagulation or aspirin use
- Exclusion Criteria:
- paroxysmal afib - known contraindication to rate control agents
- history of stroke - cardiac surgery in < 3 months
- unable to exercise - current or foreseen pacemaker, ICD
- NYHA Class IV or admission for heart failure exacerbation in < 3 months
- untreated hyperthyroidism or euthyroid for < 3 months
- evidence of sick sinus, AV conduction abnormality, symptomatic bradycardia,
pauses > 3secs, escape rate < 40bpm
- The typical participant by percentages: Low-risk, male, who had undergone prior cardioversion, who had
HTN, BMI 29, NYHA Class I heart failure w/ EF of 45%, and on vitamin k antagonist. Very few patients had CAD or diabetes (10% in each group). Most had CHADS2 of 0 or 1.
Assignment:
- 614 total patients enrolled who were randomly but openly assigned
- Two study groups:
- Lenient Rate Control: target resting HR < 110bpm
304/311 met goal (97.7%)
- Strict Rate Control: target resting HR < 80 and <110 for moderate exertion
203/303 met goal (67%)
- Initially 500 patients enrolled (250 per group) with 2.5 years of follow up planned in order to obtain
necessary power. Adverse event rate was actually lower than estimated based on previous studies and sample sizes were increased with an extension of follow up to 3 years.
- Patients were randomized with similar baseline characteristics except increased
prevalence of CAD, dBP and statin use in the lenient rate group.
Treatment:
- Heart rate determined by EKG initially
- Follow up was scheduled every two weeks until target heart rate achieved
- Follow up also scheduled at 1, 2 and 3 years
- No set protocol for rate control (no protocol for agent selection, dose selection, proceeding to ablation and
pacemaker placement), although physicians were encouraged to maximize dosing of current agents prior to initiating new agent
- If medication adjustment made, 24 holter monitoring performed after change in strict group only
- If patient remained symptomatic, even if goal HR achieved, protocol permitted further adjustment of
medications, electrical cardioversion or ablation
Measured Outcome:
- Primary: Composite outcome of:
- death due to cardiovascular cause - placement of pacer / ICD
- hospitalization for hear failure - major bleeding
- stroke - systolic embolization
- arrhythmic event – syncope, sustained VT, cardiac arrest
- Secondary:
- individual components of primary outcome
- death from any cause
- symptoms
- functional status
- Outcomes assessed by blinded committee
Statistical Analysis:
- Kaplan-Meier curves were generated using time to first event of primary outcome
- These curves estimate the cumulative incidence of the primary outcome for each group at various times
during the study. The estimated incidences at three years were used. But with Kaplan-Meier curves the confidence intervals widen as time passes. This is a function of having fewer participants remaining in the study as time progresses (N is decreasing). Could this contribute to a conclusion that the groups are not different?
- Absolute difference between the two cumulative incidences was calculated with a 90% confidence
interval using the standard errors from the curves. In order to satisfy non-inferiority criteria, the upper limit of the 90% confidence interval for the absolute difference could be no more than 10% (determined at the outset).
- The Hazard Ratio was a second measure of non-inferiority. The Hazard Ratio could be interpreted as the
probability that a participant in the lenient-control group experiences a primary outcome before a participant in the strict-control group does. Once again, to meet non-inferiority criteria, the hazard ratio would need to be < 1.4 which is the upper bound of the 90% confidence interval. The Hazard Ratio provides no indication as to when each group will experience the event.
Results:
- Mean resting HR(±SD) after rate adjustment period: 93±9 in lenient group and 76±12 in strict group,
p<0.001
Mean Resting HR / Strict / Lenient1 year f/u / 75±12 / 86±15
2 year f/u / 75±12 / 84±14
3 year f/u / 76±14 / 85±14
- During follow-up 18 in lenient group and 22 in strict group converted to sinus rhythm
- Nine in each group in sinus rhythm at end of follow-up (p=0.96)
- 81 patients reached the primary outcome
- 38 in lenient group and 43 in the strict group
- Three year cumulative incidence estimated by Kaplan-Meier Curves
- 12.9% in the lenient group
- 14.9% in strict group
- Absolute difference is -2.0%, 90% CI: -7.6 to 3.5%, p<0.001
- Hazard ratio 0.84, 90% CI: 0.58 to 1.21, p=0.001
- After adjustment for dBP, CAD and statins Hazard Ratio was 0.8, CI 0.55 to 1.17
- How this adjustment is made is not disclosed
- Death from any cause at 3 yrs: Lenient 17(5.6%) Strict 18(6.6%), HR 0.91, CI 0.52-1.59
- Symptoms: Lenient 45.6% v. Strict 46%
- Dyspnea: 30% v. 29.6%, p=0.90
- Fatigue: 24.4% v. 22.6%, p=0.63
- Palpitations: 10.6% v. 9.5%, p=0.66
- NYHA Classes: Lenient v. Strict
- Class I: 70% v. 70.4%, p=0.74
- Class II: 23.3% v. 23.4%, p=0.74
- Class III: 6.7 v. 6.2%, p=0.74
- Pacemaker Implantation: Lenient (0.8%) v. Strict (1.4%) no p value provided
Authors’ Conclusions:
- Lenient rate control is noninferior to strict rate control in the prevention of major cardiovascular events in
patients with permanent atrial fibrillation.
- Death rate from cardiovascular causes was similar between groups.
- Adverse effects of rate control agents, symptomatic atrial fibrillation
(syncope/dyspnea/fatigue/palpitations) and pacemaker implantation rate was similar in both groups
Authors’ Stated Limitations:
- Low-risk study population: no prior strokes, few diabetics, most with NYHA class I heart failure, most
CHADS2 0 or 1 and all able to exercise.
- Low incidence of primary outcome
- Short follow-up period
- Inability to ensure that all patients in a group are able to meet HR goals
My Assessment:
- Overall this is a well designed study with good internal validity in spite of a few inequalities in randomization. The main problem is with external validity. The patient population was generally healthy and low-risk – not many participants with diabetes or CAD, only mild heart failure, low CHADS2 scores. This is a good start in determining that lenient rate control in atrial fibrillation is not inferior to strict control, at least in a low-risk population. I will be cautious in applying these results to my clinic patient with diabetes and diastolic heart failure who is 85 years-old. Further evaluation of lenient rate control in patients with high-risk characteristics such as heart failure or history of stroke is still needed.
- Initially I thought that the study lost significance as only 67% of the patients in the strict control group met their heart rate goal. The importance is in the fact that patients were assigned to a strategy of care and this plan was carried out. In our clinic we are not always successful in reaching a certain HR goal with patients. In the study these patients were aggressively treated with rate-control agents with the intention of strict rate control even if the goal itself was not achieved. The difference between the two groups is in the strategy of lenient versus strict rate control, not necessarily the actual heart rate.
- Having said that, I would like to be certain that a difference existed between the groups in respect to heart rate. Heart rate in the lenient group is not so lenient. Initially following rate adjustment period, a nice difference is seen between mean heart rates in the two groups. In the 1, 2 and 3 year follow up visits some of this distinction between the two groups had disappeared. It would have been informative to see a more detailed plot of HR v. time so that one could see a distinct difference between the groups throughout the study.
-For a low-risk patient a heart rate of 100 is not inferior to a heart rate of 75.