Correspondence
Oral Anticoagulation and Stroke in Atrial Fibrillation
N Engl J Med 2003; 349:2360-2361December 11, 2003
- Article
To the Editor:
Hylek and colleagues (Sept. 11 issue)1 report that in patients with nonvalvular atrial fibrillation who were admitted with stroke, an international normalized ratio (INR) of less than 2.0 was associated with a greater severity of stroke and higher mortality than an INR of 2.0 or greater. The median of all INR values during the six months before the stroke was the same in the two INR groups. Taken together, these data indicate that thrombus formation is triggered even if anticoagulation is adequate and that thrombi grow larger if the INR is less than 2.0. These notions suggest that thrombi in atrial fibrillation arise directly before the stroke. Alternatively, thrombi may develop more slowly as a result of INR oscillations with rebound hypercoagulation.2 It has been shown that greater INR instability is associated with a higher incidence of stroke.3 We wonder whether the authors observed INR instability before the stroke. If so, such a finding — as well as the authors' finding that if stroke occurs at a low INR, it is more severe than it would be at a high INR — underscores the need for drugs that provide a constant level of anticoagulation.
3 ReferencesRichard J. Folkeringa, M.D.
Trang N.H. Dinh, M.D.
Harry J.G.M. Crijns, M.D., Ph.D.
University Hospital Maastricht, 6202 AZ Maastricht, the Netherlands
rfol@cardio.azm. nl 1
Hylek EM ,Go AS ,Chang Y , et al. Effect of intensity of oral anticoagulation on stroke severity and mortality in atrial fibrillation. N Engl J Med 2003;349:1019-1026
Full Text | Web of Science | Medline2
Raskob GE ,Durica SS ,Morrissey JH ,Owen WL ,Comp PC . Effect of treatment with low-dose warfarin-aspirin on activated factor VII. Blood 1995;85:3034-3039
Web of Science | Medline3
Nozawa T ,Asanoi H ,Inoue H , et al. Instability of anticoagulation intensity contributes to occurrence of ischemic stroke in patients with non-rheumatic atrial fibrillation. Jpn Circ J 2001;65:404-408
CrossRef | Medline
To the Editor:
In an observational study, Hylek et al. analyzed a cohort of patients taking warfarin for nonvalvular atrial fibrillation and reported a lower frequency of major ischemic stroke if the INR was 2.0 or greater than if the INR was less than 2.0. Among the reasons for subtherapeutic anticoagulation during warfarin therapy is poor compliance, and if compliance was poor among the patients in this study, it is feasible that many of them also complied inadequately with cholesterol-lowering treatment, antihypertensive treatment, and lifestyle modifications. The study did not take into account these factors, which may increase the risk of stroke; hence, the reported relation between subtherapeutic anticoagulation and an increased risk of major stroke is presumptive. Hylek et al. also fail to report on the full range of complications that may be more likely to occur with a higher INR, such as upper gastrointestinal hemorrhage,1 which may exert an influence on morbidity and mortality.
1 ReferencesElliot F. Epstein, M.B., Ch.B.
Indira Natarajan, M.B., B.S.
University Hospital of North Staffordshire, Stoke-on-Trent ST4 6QG, United Kingdom1
The Stroke in Atrial Fibrillation Investigators
CrossRef | Web of Science | Medline
To the Editor:
Hylek et al. suggest that among patients admitted to the hospital with stroke and nonvalvular atrial fibrillation, INR values of 2.0 or greater are associated with a significantly lower 30-day mortality rate than INR values of less than 2.0. By contrast, there is no absolute increase in the rate of intracranial hemorrhage at INR values of less than 4.0. These data are supported by the results of the recently reported Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM), in which a rate-control strategy was compared with a rhythm-control strategy in patients with a history of nonvalvular atrial fibrillation.1 The AFFIRM investigators noted that most ischemic strokes occurred in patients who were not taking warfarin or whose INR was less than 2.0; cerebral hemorrhage was uncommon and was associated with a mean INR of 4.3. I believe that the article by Hylek et al. greatly clarifies the current debate about whether certain groups of patients gain particular benefit from lower-level warfarin therapy with a target INR of less than 2.0.
1 ReferencesChristopher J. Boos, M.B., B.S.
Portsmouth Hospitals NHS Trust, Portsmouth PO3 6AD, United Kingdom
christopherboos@hotmail.com 1
The Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) Investigators
Full Text | Web of Science | Medline
Author/Editor Response
We thank Dr. Boos for his positive comments. Drs. Epstein and Natarajan are correct that our observational study may be subject to residual unmeasured confounding. However, the beneficial effect of INR values of 2.0 or greater on the outcome of stroke in patients with atrial fibrillation was very large, it persisted after adjustment for validated risk factors for stroke in patients with atrial fibrillation, and it is consistent with a substantial body of data from prior research, including randomized trials that demonstrated that anticoagulation is particularly more effective than aspirin in preventing severe stroke.1 We doubt that differences in cholesterol-lowering or antihypertensive treatment or in lifestyle, if there were any such differences between patients presenting with higher values and those presenting with lower INR values, could account for much of the effect that we observed.
The main goal of our article was to highlight the striking effect of the INR level on the severity of stroke. We addressed the effect of INR on the incidence of intracranial hemorrhage because such events have outcomes as devastating as ischemic stroke. We note that gastrointestinal hemorrhage, though important, is much less likely to result in death or long-term disability.2
Dr. Folkeringa and colleagues suggest that INR oscillations and possible rebound hypercoagulation account for some of our findings. Our data indicate that the INR at the time of the stroke was a much stronger predictor of the outcome than the variability in the recorded INR during the six months before the event. We would need serial INR levels in the days immediately before the stroke to address this issue more meaningfully; these values were not obtained as part of usual-care monitoring, since accurate prediction of the timing of stroke is not possible. Oral vitamin K antagonists are very effective in preventing ischemic stroke and severe stroke in patients with atrial fibrillation,3 despite the challenges involved in using such agents. We continue to encourage the use of warfarin for patients with atrial fibrillation who have risk factors for stroke. Our article should simply discourage the use of suboptimal levels of anticoagulation in these patients.
3 ReferencesElaine M. Hylek, M.D., M.P.H.
Massachusetts General Hospital, Boston, MA 02114Alan S. Go, M.D.
Kaiser Permanente of Northern California, Oakland, CA 94612Daniel E. Singer, M.D.
Massachusetts General Hospital, Boston, MA 021141
van Walraven C ,Hart RG ,Singer DE , et al. Oral anticoagulants vs aspirin in nonvalvular atrial fibrillation: an individual patient meta-analysis. JAMA 2002;288:2441-2448
CrossRef | Web of Science | Medline2
Fihn SD ,McDonell M ,Martin D , et al. Risk factors for complications of chronic anticoagulation: a multicenter study. Ann Intern Med 1993;118:511-520
Web of Science | Medline3
Albers GW ,Dalen JE ,Laupacis A ,Manning WJ ,Petersen P ,Singer DE . Antithrombotic therapy in atrial fibrillation. Chest 2001;119:Suppl:194S-206S
CrossRef | Web of Science | Medline
- Citing Articles (1)
Citing Articles
1
R.L. Jeffree, D.H. Gordon, R. Sivasubramaniam, A. Chapman. (2009) Warfarin related intracranial haemorrhage: A case-controlled study of anticoagulation monitoring prior to spontaneous subdural or intracerebral haemorrhage. Journal of Clinical Neuroscience 16:7, 882-885
CrossRef
