We conducted a substudy of the Rate Control Versus Electrical Cardioversion (RACE) study to evaluate cardiovascular morbidity, mortality, and the outcome of rate and rhythm control treatment in subjects with and without hypertension with persistent atrial fibrillation.
Hypertension is the most prevalent independent risk factor of atrial fibrillation.1-3 Although atrial fibrillation is not fatal, it is associated with thromboembolic complications, bleeding, and heart failure. Hypertension independently increases the risk of cardiovascular diseases.4 The risk of stroke is increased 2- to 3-fold in patients with hypertension and atrial fibrillation.5 Although the incidence of cardiovascular complications can be decreased with proper treatment, adequate control of blood pressure to target levels is frequently not accomplished.6-9
Studies comparing rate and rhythm control have shown these 2 therapies to be comparable in terms of cardiovascular morbidity and mortality among patients with persistent atrial fibrillation.10,11 We performed a substudy of the Rate Control Versus Electrical Cardioversion (RACE) study to evaluate the effect of hypertension on morbidity and mortality among subjects with persistent atrial fibrillation receiving rate and rhythm control therapy.12
Subjects and methods
The RACE study included 522 subjects with recurrent persistent atrial fibrillation, 256 of whom had hypertension, who were randomly assigned to receive rate or rhythm control treatment.10 The rate control group was treated with digitalis, a nondihydropyridine calcium channel blocker, or beta blockers, alone or in combination, to achieve a resting heart rate < 100 beats/minute. Subjects in the rhythm control group underwent serial electrical cardioversions and received serial antiarrhythmic drugs, including sotalol (Betapace, Sorine), class IC drugs, and amiodarone (Pacerone). Electrical cardioversion was performed under adequate anticoagulant therapy. Acenocoumarol or fenprocoumon was given to all subjects (target international normalized ratio [INR], 2.5-3.5). In subjects without stroke risk factors or for whom long-term sinus rhythm was achieved, aspirin (80-100 mg daily) was allowed. Subjects were considered to have a history of hypertension if the mean of at least 2 measured blood pressures was > 140/90 mm Hg before the start of the study. The primary study endpoint was the composite of cardiovascular mortality, heart failure, thromboembolic complications, bleeding, severe adverse effects of antiarrhythmic drugs, and pacemaker implantations.10
Subjects without hypertension were younger (mean age, 67 ± 9 years) than those with hypertension (mean age, 69 ± 8 years; = .01). They were also more often men (n = 72; 27%; < .001). Subjects with hypertension had higher systolic and diastolic blood pressures and thicker septums and posterior walls compared with subjects without hypertension. There were important differences in drug therapy between the 2 groups; those in the hypertension group were more often treated with angiotensin-converting enzyme (ACE) inhibitors or angiotensin receptor blockers, diuretics, and dihydropyridine calcium channel blockers than those in the normotensive group. There were no significant differences in baseline characteristics between hypertensive subjects randomly assigned to rate and rhythm control treatment.
Fifty-seven percent of hypertensive subjects and 45% of normotensive subjects were randomly assigned to receive rhythm control treatment. The mean follow-up period was 2.3 ± 0.6 years. Sinus rhythm was restored in 35% and 41% of the hypertensive and normotensive subjects, respectively ( = .4). Seventy-one percent of normotensive subjects and 75% of all hypertensive subjects received continuous oral anticoagulation therapy. The systolic blood pressure of hypertensive subjects treated with rhythm control was a mean of 9 mm Hg higher than normotensive subjects at 3 months to 2 years of follow-up. In addition, hypertensive subjects who received rhythm control therapy were less frequently treated with verapamil (Calan, Covera-HS, Isoptin, Verelan) or diltiazem (Cardizem, Dilacor, Tiazac).
The primary endpoint was reached more frequently among hypertensive subjects than normotensive subjects (25% vs 15%, = .01). There was no significant difference between the 2 groups in terms of cardiovascular mortality. Hypertensive subjects more often experienced severe adverse effects of antiarrhythmic drugs (5% vs 0.4%, < .001), thromboembolic complications (10% vs 4%, = .009), and hospitalization for heart failure (6% vs 2%, = 0.01). Most thromboembolic complications occurred while subjects were in atrial fibrillation and were being treated with oral anticoagulant therapy.
Two hypertensive subjects were not receiving anticoagulation therapy at the time of thromboembolic complication, and 18 received inadequate anticoagulant therapy (INR, < 2.0). Adverse effects of antiarrhythmic drugs included digitalis intoxication (n = 2), rapid atrioventricular conduction during atrial flutter (n = 1), ventricular fibrillation (n = 1), torsade de pointes (n = 2), and sick sinus syndrome or atrioventricular block (n = 7). While being treated with flecainide (Tambocor), 1 normotensive subject experienced drug-induced heart failure. In hypertensive subjects, implantation of a pacemaker was required for 4 subjects in whom sick sinus syndrome was unmasked by cardioversion, for 1 subject with bradycardia during atrial fibrillation, and for 2 subjects after atrioventricular node ablation. For nonhypertensive subjects, pacemaker implantation was required for 1 subject with sick sinus syndrome unmasked by cardioversion and for 3 subjects after atrioventricular node ablation.
The occurrence of the primary endpoint was significantly influenced by the randomization treatment in hypertensive subjects but not in normotensive subjects. An endpoint occurred in 31% of the hypertensive subjects randomly assigned to receive rhythm control therapy compared with only 17% of the hypertensive subjects in the rate control arm. This unequal distribution of endpoints was mainly the result of the occurrence of thromboembolic complications, severe adverse effects of antiarrhythmic drugs, and pacemaker implantations. Among subjects with hypertension, there was an independent association between rhythm control treatment and the occurrence of the primary endpoint over time (adjusted hazard ratio = 1.9; 95% confidence interval, 1.1-3.5; = .03), based on multivariate Cox proportional hazards regression analysis. In normotensive subjects, there was no relation between treatment and event-free survival.
The Atrial Fibrillation Follow-up Investigation of Rhythm (AFFIRM) and RACE studies have shown that rate control is not inferior to rhythm control with regard to cardiovascular morbidity and mortality in the general atrial fibrillation population.10,11 The current analysis of the RACE study suggests that this especially holds true for hypertensive subjects with recurrent persistent atrial fibrillation.12
In accordance with earlier studies, the characteristics of hypertensive subjects differed from normotensive subjects. Hypertensive subjects had higher systolic and diastolic blood pressures, were more often women, were older, and more often had diabetes mellitus.3 As expected, the cardiovascular morbidity and mortality were higher in hypertensive subjects than in normotensive subjects because of a higher incidence of thromboembolic complications, heart failure, and severe adverse effects of antiarrhythmic drugs. Despite the high use of oral anticoagulant therapy in the RACE study, the occurrence of more thromboembolic complications in hypertensive subjects can be explained by the potent stroke risk associated with hypertension. In addition, at the moment of a thromboembolic complication, either the oral anticoagulant therapy was not used at all because of long-term (> 1 month) maintenance of sinus rhythm, or it was instituted inadequately. Furthermore, both systolic and diastolic blood pressures were continuously higher in subjects with hypertension, despite the greater use of antihypertensive drugs, such as ACE inhibitors and angiotensin receptor blockers, diuretics, and calcium channel blockers, although we statistically corrected for those effects.
Hypertension is a major risk factor for heart failure, which explains the higher incidence of heart failure in hypertensive subjects.13,14 There were no differences in left ventricular function and presence and severity of heart failure between hypertensive and normotensive subjects. Hypertensive subjects experienced nearly all the severe adverse effects of antiarrhythmic drugs, most of which did not result from tachyarrhythmias but instead were caused by the unmasking of sick sinus syndrome, triggering symptomatic bradycardia. This was surprising because the hypertensive and nonhypertensive groups were similar in the use of negative chronotropic as well as rhythm control drugs. It is possible that hypertension might have led to diastolic heart failure with elevated filling pressures in both the ventricles and the atria.15 This may enhance atrial remodeling, that is, atrial dilatation and fibrosis of the atria including the sinus node and atrioventricular node. The absence of differences in echocardiographically measured atrial diameters and left atrial volume does not exclude more severely remodeled atria in hypertensive patients.
The most important result of the current study was the disproportionately high incidence of cardiovascular morbidity and mortality in hypertensive subjects randomly assigned to receive rhythm control therapy, which did not occur in the nonhypertensive group. The high incidence of endpoints was driven by thromboembolic complications and severe adverse effects of antiarrhythmic drugs in hypertensive subjects. This imbalance in occurrence of endpoints cannot be explained by differences in subject characteristics. The high rate of thromboembolic complications might have been caused by (asymptomatic) recurrences of atrial fibrillation after sinus rhythm was restored. In the hypertensive subjects receiving rhythm control treatment, asymptomatic recurrences of atrial fibrillation may have contributed to more thromboembolic events. Furthermore, hypertension itself is related to an increased risk of stroke after cardioversion.5 As mentioned previously, the potential role of interruption in or inadequately instituted oral anticoagulant therapy and the higher systolic blood pressure during follow-up might also be important factors responsible for an increased stroke risk. Throughout most of the follow-up period, the mean systolic blood pressure was 9 mm Hg higher in the hypertensive subjects randomly assigned to rhythm control therapy. This might have been caused by a lower prescription rate for verapamil and diltiazem in subjects randomly assigned to receive rhythm control therapy, possibly because verapamil and diltiazem were more often prescribed for rate control.
The results of our study indicate that hypertension is a major factor for increased cardiovascular risk in patients with persistent atrial fibrillation. This, together with low efficacy and a high adverse event rate of the present pharmacologic rhythm control strategy, especially in hypertensive patients, favors acceptance of atrial fibrillation in hypertensive patients earlier in the course of the disease. These findings suggest that treatment in atrial fibrillation may be guided by the underlying heart disease. Whether this also is the case in nonpharmacologic rhythm control strategies, in particular, curative catheter ablation procedures, remains to be seen.
This study was supported by grants from the Center of Health Care Insurance (OG96-047) and the Interuniversity Cardiology Institute, the Netherlands, and by an unrestricted grant from 3M Pharma, the Netherlands.