The website domain reflects its role for my own purpose only. Viewing of the webpages by others is not approved.
Main Index Site Developments
Patient Information Conditions of Use



---
Atrial Fibrillation

Non-Rheumatic Atrial Fibrillation

Should we accept chronic atrial fibrillation or attempt to maintain sinus rhythm?

Rationale for the AFFIRM trial

Between 1989 and 1992 publication of several well designed clinical trials in patients with non-rheumatic atrial fibrillation clearly showed that warfarin was superior to aspirin or placebo, in terms of reducing risk of stroke or transient ischaemic attacks.

Subsequently, the SPAF III trial enrolled high-risk patients and showed that conventional anticoagulation with warfarin was much better than the combination of low-dose warfarin and low-dose aspirin. In SPAF III the absolute annual reduction in thrombo-embolism was 5.2% indicating that only 19 high-risk patients need to be treated each year to prevent one event.

Epidemiologic studies have shown that patients with atrial fibrillation have a two-fold increase in mortality rate. The Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) study attempted to determine whether maintenance of sinus rhythm (with DC cardioversion and use of anti-arrhythmic agents) would reduce mortality compared with a strategy of accepting chronic atrial fibrillation and controlling the heart rate as well as continuing anticoagulation with warfarin.

Patients enrolled in the AFFIRM trial had high-risk features comparable with the SPAF III trial

Over four thousand patients with either paroxysmal or persistent atrial fibrillation were enrolled in AFFIRM. Patients over the age of 65 years and who had one other risk factor for stroke or death were eligible. Hypertension was present in 71% and congestive heart failure in 23%.

The tables below list the inclusion criteria for AFFIRM and SPAF III trials- at least one risk factor had to be present to be eligible for entry into the trials.

Comparison of AFFIRM and SPAF III entry criteria
AFFIRM (all at least 65 years of age (average 70 years) and with at least one of the following: SPAF III (average age~71 years) with at least one of the following:
Hypertension Systolic pressure greater than 160mmHg
Congestive heart failure Congestive heart failure
Ejection fraction less than 40 percent Poor LV systolic function
LV fractional shortening less than 25% LV fractional shortening less than 25%
Prior TIA or stroke Prior thromboembolism
LA greater than 5.0cm
Females aged over 75 years

Attempts to maintain sinus-rhythm in the AFFIRM trial did not reduce mortality

The mean follow-up was 3.5 years, with a maximum of 6 years. At three years, 13% of patients in the rhythm-control arm had died compared with 11% in the rate-control arm. The mortality rate at five years was 24% and 21%, respectively. However, this slightly higher mortality in the rhythm-control group was not statistically significant (hazard ratio 1.15, p=0.08).

Thus, although epidemiologic studies show a higher mortality in patients with atrial fibrillation, intention to treat analysis of the AFFIRM trial indicates that attempts to maintain sinus rhythm in this study did not reduce mortality. However, "post-hoc" analysis of the data did show that those that remained in sinus rhythm did have a lower mortality.

Not surprisingly, episodes of torsades de pointes were more frequent in patients in the rhythm-control arm who were treated with anti-arrhythmic agents (0.2% vs 0.8% in the rate-control and rhythm-control arms respectively). Rates of sustained ventricular tachycardia and cardiac arrest were low (0.5 to 0.8%) and not significantly different between the two groups.

Further analysis of the AFFIRM trial's mortality data

A further analysis of the AFFIRM trial data confirmed the increased risk of death in those with diabetes, coronary artery disease, congestive heart failure, history of stroke or TIA, current smoking and use of digoxin. Each of these factors increased the hazard ratio about 1.5 fold.

This analysis also showed reduced risk of death with use of warfarin- it is not possible to determine the mechanism of this benefit (in addition to presumed reduced risk of death due to major stroke). However, this reduced mortality risk is postulated to possibly relate to a reduced risk of ischaemic cardiac events.

Sinus rhythm was associated with reduced risk of death- the mechanism of this benefit is also not certain. It is not certain if this reduced risk relates simply to maintenance of sinus rhythm or other associated factors which might increase chances of maintenance of sinus rhythm. It was also not clear in this analysis whether anti-arrhythmic drugs were definitely harmful if these agents helped maintain sinus rhythm

Cardiac and non-cardiac death in the AFFIRM trial

The trend towards a slightly higher mortality in the rhythm control arm might have been assumed to relate to a slightly higher cardiac mortality related to antiarrhythmic drug therapy.

However, an interesting further analysis of this trial showed similar cardiac mortality in both groups. There was an unexplained higher non-cardiac mortality in the rhythm control arm, mainly related to pulmonary disease and cancer. The article reporting this findings states that amiodarone and sotalol were the two main antiarrhythmic agents used. Some, but not all, other trials that used amiodarone also suggested there was an increased risk of non-cardiac death with this agent. Undoubtedly, further study is required to confirm or refute this possiblity.

What were the chances of maintaining sinus rhythm in the AFFIRM trial?

Forty percent of patients in the rate-control arm (those in whom no attempt was made to maintain sinus rhythm) were in sinus rhythm at the end of the study after a mean followup of 3.5 years. One assumes this might relate to the enrollment of patients with paroxymal atrial fibrillation into the trial.

About 60% in the rhythm-control arm were in sinus rhythm at the end of the study, that is with a mean followup of 3.5 years.

The probability of remaining in sinus rhythm in the rhythm-control arm reduced with time. At one year about 82% were in sinus rhythm, at three years 73% were in sinus rhythm and at five years 62% were in sinus rhythm.

Eighteen percent of patients in the rhythm-control arm had electrical cardioversion once during follow-up. Eleven percent had electrical cardioversion twice, and nine percent had electrical cardioversion three or more times.

A sub-study of the AFFIRM trial showed that amiodarone was markedly superior to sotalol and to class I agents for maintenance of sinus rhythm. However, given this agent’s side-effect profile it is difficult to recommend amiodarone should be used as the first line agent in all patients.

Not surprisingly there was a slightly increased risk of torsades de pointes in patients in the rhythm-control arm. However, there were no differences in the rates of ventricular tachycardia or fibrillation, and no differences in the occurrence of cardiac arrest.

What was the stroke rate in the AFFIRM trial?

Ischaemic stroke rate was only about 1% per year in both arms. Of note is that there was a high rate of anticoagulation in both arms, with over 90% in the rate-control arm and about 70% in the rhythm-control arm remaining on warfarin. This study was not designed to determine the duration of anticoagulant therapy in those patients who appear to be maintained in sinus rhythm.

This relatively low stroke rate compares with an annual ischaemic stroke rate of 7.8% in the low-dose warfarin arm of SPAF III. It is reasonable to conclude that both strategies used in the AFFIRM trial did reduce risk of stroke, recognising the high rate of anticoagulation in patients that appear to be maintained in sinus rhythm.

Despite apparent maintenance of sinus rhythm there may be risk of cardiac thromboembolism related to episodes of prolonged asymptomatic paroxysmal atrial fibrillation. If patients are likely to have episodes of asymptomatic atrial fibrillation, and are in the high-risk group, I tend to recommend anticoagulation for longer periods after cardioversion to sinus rhythm.

This indicates the need for us to continue anticoagulation with warfarin for longer than the “conventional” three week period after electrical cardioversion in these patients. The decision to stop anticoagulation will remain an “inexact science” but may be influenced by an individual patient’s risk status and whether the patient is likely to have asymptomatic episodes of atrial fibrillation.

Do echocardiographic parameters help predict risk of recurrent atrial fibrillation?

Many of us tend to assume that those patients with enlarged atria or with heart failure are unlikely to remain in sinus rhythm after electrical cardioversion. However, the AFFIRM trial’s echo sub-study is consistent with other published observations and suggests that there is no echo parameter that helps predict risk of recurrence of atrial fibrillation, including size of the left atrium.

The only predictors of increased risk of recurrent atrial fibrillation in this study were a longer duration of atrial fibrillation and occurrence of more than one episode of atrial fibrillation at entry into the study.

Nevertheless, many of us are likely to assume that those patients with more extreme degrees of left ventricular dysfunction or atrial enlargement may be less likely to be maintained in sinus rhythm, or that in such patients we can have a lower threshold to use amiodarone as the first line anti-arrhythmic agent.

Other trials that have compared rate-control vs rhythm-control

The RACE trial also showed no mortality benefit with maintenance of sinus rhythm. It was a much smaller trial enrolling just over 550 patients compared to the AFFRIM trial that enrolled over four thousand patients.

In the RACE trial women and hypertensive patients had a worse outcome with the rhythm-control strategy. As with all studies, subgroup analysis should always be interpreted with caution. In the AFFIRM trial no such difference was seen in hypertensives or women.

As is well known, randomised clinical trials often leave some questions unanswered or raise other questions. Only 23% of patients in the AFFIRM trial had congestive heart failure. The AF-CHF trial is comparing the rate-control and rhythm control strategies in patients with congestive heart failure and with an ejection fraction less than 35%.

Concluding Comments:

  • In patients with chronic or paroxysmal atrial fibrillation, a strategy of anticoagulation and rate-control does not expose the patient to a higher mortality risk.
  • If attempts are made to maintain sinus rhythm, then for "high-risk" patients such as those enrolled in the AFFIRM trial, there is a need to consider longer periods of anticoagulation. We might be more comfortable stopping oral anticoagulant therapy sooner after cardioversion if the patient is in the "moderate-risk" or "low-risk" category for thromboembolism. I doubt there is any randomised trial data on risk of thromboembolism with a rhythm-control strategy and short periods of anticoagulation vs rhythm-control strategy and long periods of anticoagulation.
  • Further analysis of the AFFIRM trial data suggests a lower mortality risk if anticoagulation is continued. It is possible this benefit relates to improved secondary prevention in those with ischaemic heart disease- this is speculative. However, if this is the mechanism of benefit than those with known ischaemic heart disease and a history of atrial fibrillation would be better treated with long-term anticoagulation.
  • The findings of the AFFIRM trial may not be applicable to all patients and do not suggest we stop attempting to maintain sinus rhythm in all patients. I think we should always assess a patient with a view to proceeding to cardioversion if there has been new-onset atrial fibrillation. However, if, after cardioversion, sinus rhythm is not "easily" maintained we will be more comfortable switching to a rate-control strategy.
  • Some patients remain asymptomatic when atrial fibrillation develops- for example
    • those who have conducting system disease or are on drugs such as betablockers which prevent tachycardia with atrial fibrillation;
    • those without severe diastolic or systolic dysfunction of the left ventricle who do not develop heart failure or effort intolerance.
    Such patients may have been in atrial fibrillation for many weeks or months before the arrhythmia is detected at routine "check-ups". It is possible that many of these patients will be in the group of patients one does not even attempt DC cardioversion on the basis that it is less likely that we can maintain sinus rhythm long-term, we might more readily accept the need for long-term anticoagulation and rate control.
  • Since there is no reduction in mortality with attempts to maintain sinus rhythm, we might question the value of repeated DC cardioversion in patients that tolerate atrial fibrillation well. In such patients we need to try to determine if the benefits of maintenance of sinus rhythm exceed the risk of potential side-effects of antiarrhythmic agents that are used to try to maintain sinus rhythm.
  • Some antiarrhythmic agents are known to increase risk of torsades de pointes and there needs to be heightened awareness of this risk, including the need to ensure great care is taken when using other drugs that might prolong the QT interval and avoiding hypokalaemia in patients on diuretic therapy. Care should also be taken before high dose sotalol therapy is prescribed to women.
  • Patients whose heart rate response to atrial fibrillation can not be controlled easily or who develop heart failure with atrial fibrillation are often best treated with attempt to maintain sinus rhythm.
  • The rates of maintenance of sinus rhythm demonstrated in the AFFIRM trial are only applicable to the patient population enrolled in this study

Hitesh Patel, Cardiologist
5th September, 2004

References
  • SPAF III trial. The Lancet 1996; 348:633-38
  • AFFIRM trial. N Engl J Med 2002;347:1825-33
  • RACE trial. N Engl J Med 2002;347:1834-40.)
  • Relationships Between Sinus Rhythm, Treatment, and Survival in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) Study. Circulation 2004;109:1509-1513
  • Analysis of Cause-Specific Mortality in the Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) Study. Circulation 2004;109:1973-1980
Users should read this document on "copyright" and "conditions of use".