14.1 Atrial Fibrillation
In five prospective, randomized, controlled clinical trials involving 3711 patients with non-rheumatic AF, warfarin significantly reduced the risk of systemic thromboembolism including stroke (see Table 4). The risk reduction ranged from 60% to 86% in all except one trial (CAFA: 45%), which was stopped early due to published positive results from two of these trials. The incidence of major bleeding in these trials ranged from 0.6% to 2.7% (see Table 4).
Table 4: Clinical Studies of Warfarin in Non-Rheumatic AF Patients*
| n | | | Thromboembolism | % Major Bleeding |
| Study | Warfarin-Treated Patients | Control Patients | PT Ratio | INR | % Risk Reduction | p-value | Warfarin-Treated Patients | Control Patients |
| AFASAK | 335 | 336 | 1.5 to 2 | 2.8 to 4.2 | 60 | 0.027 | 0.6 | 0 |
| SPAF | 210 | 211 | 1.3 to 1.8 | 2 to 4.5 | 67 | 0.01 | 1.9 | 1.9 |
| BAATAF | 212 | 208 | 1.2 to 1.5 | 1.5 to 2.7 | 86 | < 0.05 | 0.9 | 0.5 |
| CAFA | 187 | 191 | 1.3 to 1.6 | 2 to 3 | 45 | 0.25 | 2.7 | 0.5 |
| SPINAF | 260 | 265 | 1.2 to 1.5 | 1.4 to 2.8 | 79 | 0.001 | 2.3 | 1.5 |
*All study results of warfarin vs. control are based on intention-to-treat analysis and include ischemic stroke and systemic thromboembolism, excluding hemorrhagic stroke and transient ischemic attacks.
Trials in patients with both AF and mitral stenosis suggest a benefit from anticoagulation with warfarin sodium [seeDosage and Administration (2.2)].
14.2 Mechanical and Bioprosthetic Heart Valves
In a prospective, randomized, open-label, positive-controlled study in 254 patients with mechanical prosthetic heart valves, the thromboembolic-free interval was found to be significantly greater in patients treated with warfarin alone compared with dipyridamole/aspirin-treated patients (p < 0.005) and pentoxifylline/aspirin-treated patients (p < 0.05). The results of this study are presented in Table 5.
Table 5: Prospective, Randomized, Open-Label, Positive-Controlled Clinical
Study of Warfarin in Patients with Mechanical Prosthetic Heart
Valves
| Patients Treated With |
| Warfarin | Dipyridamole/Aspirin | Pentoxifylline/Aspirin |
| Event | | | |
| Thromboembolism | 2.2/100 py | 8.6/100 py | 7.9/100 py |
| Major bleeding | 2.5/100 py | 0/100 py | 0.9/100 py |
py = patient years
In a prospective, open-label, clinical study comparing moderate (INR 2.65) vs. high intensity (INR 9) warfarin therapies in 258 patients with mechanical prosthetic heart valves, thromboembolism occurred with similar frequency in the two groups (4 and 3.7 events per 100 patient years, respectively). Major bleeding was more common in the high intensity group. The results of this study are presented in Table 6.
Table 6: Prospective, Open-Label Clinical Study of Warfarin in Patients with Mechanical Prosthetic Heart Valves
| Event | Moderate Warfarin Therapy INR 2.65 | High Intensity Warfarin Therapy INR 9 |
| Thromboembolism | 4/100 py | 3.7/100 py |
| Major bleeding | 0.95/100 py | 2.1/100 py |
py = patient years
In a randomized trial in 210 patients comparing two intensities of warfarin therapy (INR 2 to 2.25 vs. INR 2.5 to 4) for a three month period following tissue heart valve replacement, thromboembolism occurred with similar frequency in the two groups (major embolic events 2% vs. 1.9%, respectively, and minor embolic events 10.8% vs. 10.2%, respectively). Major hemorrhages occurred in 4.6% of patients in the higher intensity INR group compared to zero in the lower intensity INR group.
14.3 Myocardial Infarction
WARIS (The Warfarin Re-Infarction Study) was a double-blind, randomized study of 1214 patients 2 to 4 weeks post-infarction treated with warfarin to a target INR of 2.8 to 4.8. The primary endpoint was a composite of total mortality and recurrent infarction. A secondary endpoint of cerebrovascular events was assessed. Mean follow-up of the patients was 37 months. The results for each endpoint separately, including an analysis of vascular death, are provided in Table 7:
Table 7: WARIS – Endpoint Analysis of Separate Events
| | | % Risk |
| Warfarin | Placebo | | Reduction |
| Event | (n = 607) | (n = 607) | RR (95% CI) | (p-value) |
| Total Patient Years of Follow-up | 2018 | 1944 | | |
| Total mortality | 94 (4.7/100 py) | 123 (6.3/100 py) | 0.76 (0.6, 0.97) | 24 (p = 0.03) |
| Vascular death | 82 (4.1/100 py) | 105 (5.4/100 py) | 0.78 (0.6, 1.02) | 22 (p = 0.068) |
| Recurrent MI | 82 (4.1/100 py) | 124 (6.4/100 py) | 0.66 (0.51, 0.85) | 34 (p = 0.001) |
| Cerebrovascular event | 20 (1/100 py) | 44 (2.3/100 py) | 0.46 (0.28, 0.75) | 54 (p = 0.002) |
RR = Relative risk; Risk reduction = (1 - RR); CI = Confidence interval; MI = Myocardial infarction; py = patient years
WARIS II (The Warfarin, Aspirin, Re-Infarction Study) was an open-label, randomized study of 3630 patients hospitalized for acute myocardial infarction treated with warfarin to a target INR 2.8 to 4.2, aspirin 160 mg per day, or warfarin to a target INR 2 to 2.5 plus aspirin 75 mg per day prior to hospital discharge. The primary endpoint was a composite of death, nonfatal reinfarction, or thromboembolic stroke. The mean duration of observation was approximately 4 years. The results for WARIS II are provided in the Table 8.
Table 8: WARIS II – Distribution of Events According to Treatment Group
| Event | Aspirin (n = 1206) | Warfarin (n = 1216) | Aspirin plus Warfarin (n = 1208) | Rate Ratio (95% CI) | p-value |
| No. of Events | |
| Major bleedinga | 8 | 33 | 28 | 3.35b (ND) | ND |
| | | 4c (ND) | ND |
| Minor bleedingd | 39 | 103 | 133 | 3.21b (ND) | ND |
| | | 2.55c (ND) | ND |
| Composite endpointse | 241 | 203 | 181 | 0.81 (0.69 to 0.95)b | 0.03 |
| | | 0.71 (0.6 to 0.83)c | 0.001 |
| Reinfarction | 117 | 90 | 69 | 0.56 (0.41 to 0.78)b | < 0.001 |
| | | 0.74 (0.55 to 0.98)c | 0.03 |
Thromboembolic stroke | 32 | 17 | 17 | 0.52 (0.28 to 0.98)b | 0.03 |
| | | 0.52 (0.28 to 0.97)c | 0.03 |
| Death | 92 | 96 | 95 | | 0.82 |
a Major bleeding episodes were defined as nonfatal cerebral hemorrhage or bleeding necessitating surgical intervention or blood transfusion.
b The rate ratio is for aspirin plus warfarin as compared with aspirin.
c The rate ratio is for warfarin as compared with aspirin.
d Minor bleeding episodes were defined as non-cerebral hemorrhage not necessitating surgical intervention or blood transfusion.
e Includes death, nonfatal reinfarction, and thromboembolic cerebral stroke.
CI = confidence interval
ND = not determined
There were approximately four times as many major bleeding episodes in the two groups receiving warfarin than in the group receiving aspirin alone. Major bleeding episodes were not more frequent among patients receiving aspirin plus warfarin than among those receiving warfarin alone, but the incidence of minor bleeding episodes was higher in the combined therapy group.