RACE was a prospective, randomized trial of rate control (n=256 patients) versus electrical cardioversion and rhythm control (n=266 patients) for persistent AF. Total mortality was significantly lower in the rate control arm (17.2%) than in the rhythm control arm (22.6%), especially in patients with hypertension. In those patients, total mortality, emboli, or other severe complications occurred in 19% versus 31%.

AFFIRM was the largest of the five trials, randomizing 4,060 patients to a rate control versus rhythm control strategy.Anticoagulation was strongly encouraged in both arms. Patients had to be at least 65 years old or have risk markers for adverse outcome in AF so that endpoints were likely to be seen.Total mortality was the primary outcome. The AF could be persistent or paroxysmal. Mean follow-up was 3.5 years. Most patients were male (61%), had underlying structural heart disease, and had had recurrent AF before enrollment. Overall, 24% had a reduced left ventricular ejection fraction. In the rate control arm, digitalis was used in 51% of patients, ßblockers in 49%, and calcium channel blockers in 41%. In the rhythm control arm, amiodarone was used in 39%, sotalol was used in 33%, and 29% received a class I AAD (class I drugs were dropped from assignment partway through the trial). Ultimately, amiodarone was used in over 60% of the rhythm control patients. AV node ablation and pacing were used in about 5% of the rate control patients.The probability of being in normal sinus rhythm at the one- and five-year visits, respectively, was 81% and >60% in the rhythm control arm and 42% and 38% in the rate control arm. Successful rate control (meeting rest and ambulatory criteria) was 63% at one year and 80% at five years.Therefore, success with NSR decreased over time and success with rate control increased.Cross-over was greater in the rhythm control to the rate control arm (20% at two years, 37% at five years) than in the rate control to the rhythm control arm (9% at two years, 12% at five years). Analysis was by the intention-to-treat (ITT) approach. Total mortality was higher in the rhythm control arm than in the rate control arm (356 versus 306 patients; P=0.058). Moreover, there were more ischemic strokes in the rhythm control arm (7.3% versus 5.7%), probably because warfarin was often discontinued after attainment of persistent sinus rhythm according to clinical judgment. Most of the embolic strokes (65/84) were with an INR <2.0 or off warfarin. The implications, as stated earlier, are generally not to discontinue anticoagulation. There were more hospitalizations, bradycardic arrests, and torsades de pointes in the rhythm control arm, while QOL scores and major bleeds were similar in the two arms. An on-therapy analysis of events has not yet been presented, but would be of interest given the STAF observations.

If these five trials are examined in composite, they appear to have taught us that a strategy of rhythm control, in contrast to rate control, is associated with more hospitalizations, greater cost, minimally better symptom relief, and a tendency toward more strokes and greater total mortality, using current therapies. Hence, rate control is an acceptable primary therapy when patients are rendered comfortable with this approach, and rhythm control should be pursued when symptoms are sufficient to dictate pursuit of NSR despite the risks now demonstrated to be present with this strategy. None of the studies revealed greater survival or lower embolic risk with NSR, despite teleologic assumptions.

Whether or not these trials would have had the same results had the AADs, now under investigation, been used instead of those now available, or had curative ablative procedures been widely applied in place of AAD therapy, warfarin anticoagulation been appropriately continued, or an oral thrombin inhibitor been available, cannot be known. What is certain is that in these trials, using current treatments, no demonstration of a survival benefit from a strategy of sinus rhythm was apparent; rather, a risk may exist. Consequently, one may interpret them as indicating that a strategy of allowing AF to continue so long as it is ratecontrolled, anticoagulated, and well tolerated is a reasonable and medically acceptable strategy. When symptom relief demands, NSR may be pursued, but it does, in fact, require close attention to the approaches used, as the international guidelines now dictate. Alternatively, one may interpret the trial outcomes as a failure of the strategy to maintain NSR, rather than of sinus rhythm itself.This interpretation is supported by the fact that the trial results were interpreted mainly by an intention-to-treat approach. Also, many patients failed to maintain NSR and many events in the patients assigned to the rhythm control arms actually occurred during a period of AF. The hope remains that the development of newer drugs and the maturation of ablational approaches as a cure of AF will provide the additional benefit we continue to seek.