Current issues of ACP Journal Club are published in Annals of Internal Medicine


Warfarin plus aspirin reduced mortality and stroke after heart-valve replacement

ACP J Club. 1994 Jan-Feb;120:3. doi:10.7326/ACPJC-1994-120-1-003

Source Citation

Turpie AG, Gent M, Laupacis A, et al. A comparison of aspirin with placebo in patients treated with warfarin after heart-valve replacement. N Engl J Med. 1993 Aug 19;329:524-9.



To study the efficacy and safety of aspirin added to warfarin for high-risk patients who received either mechanical heart valves or tissue valves and had associated atrial fibrillation or previous thromboembolism.


Randomized, double-blind, placebo-controlled trial with a mean follow-up of 2.5 years.


3 Canadian university hospitals.


370 consecutive patients (mean age 58 y, 51% men) who received mechanical replacement heart valves or tissue valves and had preoperative atrial fibrillation or a history of thromboembolism. Heart valves were aortic, mitral, or tricuspid, singly or in combination. Exclusion criteria were allergy to or an indication for aspirin, contraindications to anticoagulant or antiplatelet therapy, patient inaccessibility because of geography, or refusal. Follow-up was 100%.


Warfarin was started as soon as patients could take oral medication. The initial dose was 10 mg/d and was adjusted to obtain an international normalized ratio (INR) of 3.0 to 4.5. 186 patients were allocated to receive aspirin, 100 mg/d (delayed-release, enteric-coated capsules), and 184 patients were allocated to receive placebo. Patients could receive low-dose heparin for prophylaxis against venous thrombosis for 3 days after warfarin was started.

Main outcome measures

Death from vascular causes, major systemic embolism, valve thrombosis, and clinically important hemorrhage.

Main results

Compared with patients receiving placebo, patients receiving aspirin had fewer deaths { P = 0.01}*, fewer deaths from vascular causes { P = 0.003}*, fewer strokes { P = 0.04}*, and more bleeding { P = 0.02}* (Table). A trend towards fewer major systemic embolisms existed in the aspirin group { P = 0.06}* (Table). The groups did not differ for anticoagulant control (mean INR of 3.0 for aspirin vs 3.1 for placebo).


For patients who received a mechanical valve or tissue valve replacement and had preoperative atrial fibrillation or a history of thromboembolism, low-dose aspirin added to warfarin reduced total mortality, vascular mortality, and stroke, although it increased bleeding complications. Benefits were greater than risks.

Source of funding: Heart and Stroke Foundation of Ontario.

For article reprint: Dr. A.G. Turpie, HGH-McMaster Clinic, Hamilton Civic Hospitals, General Division, 237 Barton Street East, Hamilton, Ontario L8L 2X2, Canada. FAX 905-521-1551.

*Numbers calculated from data in article.

Table. Aspirin vs placebo after heart-valve replacement†

Outcomes at mean follow-up of 2.5 y Aspirin Placebo RRR (95% CI) NNT (CI)
Death (any cause) 5% 12% 60% (16 to 81) 15 (8 to 66)
Vascular death 1% 7% 85% (41 to 96) 17 (10 to 45)
Major systemic embolism 3% 7% 62% (0 to 86) Not significant
Stroke 2% 7% 67% (5 to 89) 23 (11 to 407)
Bleeding 38% 27% 43% (6 to 94) 9 (5 to 51)

†Abbreviations defined in Glossary; RRR, RRI, NNT, NNH, and CI calculated from data in article.


The study by Turpie and colleagues is important because it suggests how to reduce the thromboembolic complications of prosthetic heart valves. Adding low-dose aspirin to warfarin decreased the rates of vascular complications and of death. Using low-dose aspirin rather than dipyridamole or high-dose aspirin simplifies therapy and should allow more clinicians to alter their practice.

The efficacy of warfarin plus low-dose aspirin was tested using the best method—a randomized, placebo-controlled trial with clear-cut end points and sufficient patient numbers to provide clear results. After randomization, the treatment groups were similar in demographic terms and in characteristics (type and location of valve, atrial fibrillation, ventricular function, previous systemic thromboembolism, and coronary disease) that might influence the outcomes. Cointerventions that might affect embolism or bleeding were probably equally distributed in the 2 treatment groups. The results appear to be clinically and statistically significant.

The results can be applied to patients with prosthetic heart valves without causing inordinate bleeding risks. To quell fears about the additional anticoagulant treatment, the authors included a composite outcome measure. Major systemic embolism, nonfatal intracranial bleeding, death from bleeding, or vascular death occurred in 12 patients receiving warfarin and aspirin (3.9%/y) compared with 28 patients receiving warfarin and placebo (9.9%/y), with an overall risk reduction of 61%. Most of the difference in bleeding was caused by minor episodes. Because the average level of warfarin-induced anticoagulation was relatively low (INR of 3.0 or 3.1 at an average warfarin dose of 5.8 or 5.5 mg/d in the aspirin and placebo arms, respectively), other studies should examine the efficacy of low-dose warfarin therapy (INR of 2.0) with low-dose aspirin. Using low-dose warfarin with low-dose aspirin may also prevent systemic embolization from atrial fibrillation or acute myocardial infarction.

Daniel M. Becker, MD
University of VirginiaCharlottetown, Virginia, USA