We investigated 1209 patients with valvular and nonvalvular atrial fibrillation who were treated with standard anticoagulant therapy, the cyclooxygenase inhibitor triflusal, or combined therapy with triflusal and an anticoagulant. Compared with patients receiving standard anticoagulation treatment, the addition of antiplatelet therapy to reduced-intensity anticoagulation in patients stratified for risk of stroke significantly reduced the vascular event rate.
The Framingham Study showed that a high prevalence of atrial fibrillation (AF) exists in elderly patients and that AF is an independent risk factor for stroke.1 Since then, warfarin (Coumadin) has been widely investigated as an anticoagulant, and aspirin has been investigated as an antiplatelet agent for the prevention of stroke in AF patients.2 The benefit of warfarin was superior to that seen with aspirin, and treatment with an anticoagulant to a target international normalized ratio (INR) of 2.0 to 3.0 has been considered the treatment of choice.3
Compared with adjusted-dose warfarin, the combination of low-dose warfarin (median INR, < 1.5) and aspirin was not effective, and its use has been disapproved.4,5 In the National Study for Prevention of Embolism in Atrial Fibrillation (NASPEAF), we compared a moderate-intensity anticoagulant in combination with the antiplatelet drug triflusal (Disgren) with standard anticoagulant monotherapy (target INR, 2.0—3.0). Triflusal is a cyclooxygenase in­hibitor–antiplatelet agent with an antithrombotic effect similar to aspirin but with a lower risk of bleeding.6
Patients and Methods
A total of 1209 patients with AF (nonvalvular AF, 898; valvular AF, 311) aged 60 years and older or at risk for stroke according to Stroke Prevention in Atrial Fibrillation (SPAF) III classification were included in the study.3 All patients were stratified for the existence of previous embolism and independently randomized. Risk factors for vascular events were recorded and analyzed. Patients with nonvalvular AF with a history of previous embolism and those with mitral stenosis were considered to be at high risk for embolism (high-risk group) and were assigned to receive standard anticoagulant treatment to a target INR range of 2.0 to 3.0 or combined therapy with 600 mg/day of triflusal and anticoagulant to a target INR range of 1.4 to 2.4. Patients with nonvalvular AF without a history of previous embolism (intermediate-risk group) were assigned to receive standard anticoagulant therapy, 600 mg/day of triflusal, or the combination of 600 mg/day of triflusal plus anticoagulant to achieve a target INR range of 1.25 to 2.0.
The primary end point was a composite of vascular death and nonfatal systemic embolism or stroke, which were evaluated by a blinded independent committee. The Kaplan-Meier method was used to calculate the 4-year event-free survival rate, and a Cox proportional hazards regression model was used to evaluate the efficacy and safety of the treatments.
The median follow-up period was 33 months. The median INR was 2.49 for patients receiving standard anticoagulant treatment and 1.93 and 2.17 for those receiving combined therapy in the intermediate-risk and high-risk groups, respectively. The anticoagulant arms compared with the combination arms had a larger number of INR values over 3.5 (P = .001).
Patients receiving combined therapy were less likely to have primary events in either the intermediate-risk (P = .02; Figure 1) or high-risk (P = .03; Figure 2) groups. Patients with nonvalvular AF without a history of previous embolism had a primary event rate and a composite of em­bo­lism—stroke rate of 2.70 and 1.26, respectively; these rates were 6.74 and 4.81, respectively, for those patients who had had a previous embolic event. The corresponding event rates for patients with mitral stenosis were 1.43 and 1.26, respectively, for those without previous embolism, and 5.77 and 4.81, respectively, for those who had had a previous embolic event. The event rate difference between patients with previous embolism and those without, who received an equal intensity of anticoagulant therapy, was significant (hazard ratio [HR], 3.845; 95% confidence interval [CI], 1.463-10.107; P = .006). There was no difference between patients with valvular and nonvalvular AF (Figure 3).
Intracranial bleeding with high mortality tended to be associated with anticoagulant therapy (9 of 12 events). Gastric bleeding, without mortality, tended to be associated with combination therapy (11 of 15 events). Gas­troscopy showed superficial gastric erosions in these patients. The composite of primary outcome and severe bleeding events in the intermediate-risk group was most likely to be associated with anticoagulant therapy (HR, 0.38; 95% CI, 0.17—0.87; P = .02).
The effect of combined antiplatelet and anticoagulant therapy was compared with anticoagulant monotherapy for the prevention of vascular events in AF patients in the NASPEAF study. In contrast to previous trials, NASPEAF provided several new ap­proaches regarding the treatment of AF with antithrombotic therapy.
Previous trials used the combination of 300 mg/day of aspirin and a very low intensity of antivitamin-K anticoagulant (mean INR, 1.3).4,5 The NASPEAF study used 600 mg/day of triflusal instead of as­pirin and a higher intensity of anticoagulant effect; the median INR was 1.93 and 2.17 in the intermediate-risk and high-risk groups, respectively. This higher intensity might explain the successful outcome of our trial. The Warfarin Aspirin Re-Infarction Study (WARIS) II used a similar combination therapy (75 mg/day of aspirin and an anticoagulant to a target INR range of 2.075—2.5) for ischemic heart disease patients and showed a favorable and safe antithrombotic effect.7 The optimal intensity of anticoagulant and the ideal antiplatelet dose to achieve the best balance between the prevention of embolism and the risk of bleeding in combined therapy remains unclear. A meta-analysis of antithrombotic therapy trials in prosthetic valve patients showed that aspirin increases the risk of severe bleeding.8 A dose of 75 to 100 mg/day is recommended in the current guidelines for these pa­tients9, influenced by the favorable re­sults shown by Turpie and colleagues.10 We used 600 mg/day of triflusal, and the risk of intracranial bleeding was low, although nonfatal gastric bleeding was more frequent.
Previous trials showed that low-intensity anticoagulant in combination with aspirin should not be used.3,4 According to our experience with combination therapy, the lower INR range limit of safety may be below 2 because more than 40% of INR tests were below this figure, and the therapy was effective. The anticoagulant treatment in combined therapy, nevertheless, needs laboratory control, so there may not be justification for an INR level below 1.8 or 1.9. Our trial also showed that INR levels greater than 2.5 in combined therapy are not needed, and this policy may reduce the risk of intracranial bleeding.
Rheumatic mitral stenosis patients
Previous nonrandomized trials did not allow the different committees to reach a consensus on the anticoagulant in­tensity to prevent stroke in valvular AF patients. Target INR ranges varying from 3.5 to 4.5,11 2.5 to 3.5,12 and 2.0 to 3.09 have been recommended in guidelines. Our trial confirms that a target INR between 2.0 and 3.0 is effective and safe in mitral stenosis patients, provided there is no history of previous embo­lism. We also showed that nonvalvular and valvular AF patients, stratified for the presence of embolism at baseline and receiving equal anticoagulant therapy, have a similar risk of vascular events.
Influence of previous embolism
Our study allowed us to analyze and compare the event-free survival in pa­tients with nonvalvular and valvular AF with previous embolism versus those without embolism at baseline, receiving the same anticoagulant therapy. The difference was highly significant. Patients without previous em­bo­lism receiving an anticoagulant in­tensity to maintain an INR between 2.0 and 3.0 were reasonably controlled, whereas those with a history of previous embolism had a very high level of vascular events.4,13 The composite of embolism—vascular death–severe bleeding was 7.3% per year. Some authors have proposed a higher intensity of anticoagulant therapy for these patients (target INR, 3.0).13 Our trial showed that combined antiplat­elet plus moderate anticoagulant therapy significantly reduced vascular events in these high-risk patients without increasing the risk of bleeding.
Combination therapy was also significantly effective in patients who had not had a previous embolism, but the absolute rate difference was small, and a large number of patients was needed to show significant difference. In addition, the composite of primary outcome and severe bleeding was lower in the combined therapy arm (P < .03), making this therapy an efficacious and safe treatment.
Theoretically, the combination of antiplatelet and anticoagulant therapy to prevent stroke should offer additional benefits, provided that the risk of major bleeding is not increased. Our trial showed that combined antiplatelet plus moderate anticoagulant therapy was safe and more efficacious than anticoagulant monotherapy in AF patients. No previous rand­omized trials have been performed in AF mitral stenosis patients, and it has been reasonably accepted that these patients, without anticoagulant therapy, are at high risk for vascular events. We have shown that adjusted-dose anticoagulant therapy aimed at maintaining an INR between 2.0 and 3.0 is equally effective in patients with val­vular and nonvalvular AF. Patients with a history of previous embolism are at very high risk for a new vascular event,4,6 even if they receive anticoagulation therapy to a target INR of 2.0 to 3.0. For these patients, combined antiplatelet plus moderate-intensity anticoagulant therapy was more effective than anticoagulation alone.