Rates of Ischemic Stroke During Warfarin Treatment for Atrial Fibrillation
Background and Purpose—Recent evidence suggests that there may be an increased risk of ischemic stroke immediately after warfarin initiation. We examined the rate of ischemic stroke among patients with atrial fibrillation newly started on warfarin therapy.
Methods—We conducted a population-based cohort study among Ontario residents aged ≥66 years with atrial fibrillation who received warfarin between April 1, 1997, and March 31, 2010. Each patient was followed up for ≤5 years in 30-day intervals. For each interval, we determined the rate of ischemic stroke.
Results—After 5 years, the cumulative incidence of ischemic stroke among new users of warfarin (n=148 446) was 4.0% (n=6006). The risk was highest during the first 30 days after initiation (6.0% per person-year; 95% confidence interval, 5.5%–6.4%) compared with the remainder of follow-up (1.6% per person-year; 95% confidence interval, 1.5%–1.6%), and increased with higher baseline CHADS2 (congestive heart failure, hypertension, age ≥75 years, diabetes, previous stroke) scores. Less frequent monitoring may have contributed.
Conclusions—In a large cohort of older patients with atrial fibrillation, we observed the highest rate of ischemic stroke in the first 30 days after warfarin initiation. Although causation cannot be established given the observational nature of this study, our findings highlight the need for future research in this population.
Warfarin is effective for the prevention of ischemic stroke and systemic embolism in patients with atrial fibrillation.1 However, the inconvenience of laboratory monitoring and risk of hemorrhage can complicate therapy. The risk of major hemorrhage is greatest within the first 30 days of therapy,2 and a recent case–control study in atrial fibrillation patients found a 71% increased risk of ischemic stroke in the first 30 days of warfarin therapy.3 Remaining unclear is the absolute risk of ischemic stroke throughout the course of warfarin therapy and whether it is influenced by patient characteristics. The objective of this study was to examine whether an acute risk of ischemic stroke exists among patients with atrial fibrillation newly initiated on warfarin.
We conducted a population-based cohort study of Ontario residents aged ≥66 years who received warfarin between April 1, 1997, and March 31, 2010. The primary outcome was a hospital visit (inpatient or emergency department admission) for ischemic stroke. We assessed various baseline characteristics (Table 1; Table I in the online-only Data Supplement) and followed up each patient from warfarin initiation until the first of ischemic stroke, death, warfarin discontinuation, end of the maximum 5-year follow-up, or the end of the study period (March 31, 2012; online-only Data Supplement). The Research Ethics Board of Sunnybrook Health Sciences Center, Toronto approved this study.
Crude rates were calculated in 30-day intervals as the total number of participants with a hospital visit for ischemic stroke divided by the number of event-free warfarin users at the beginning of each interval and expressed as percentage per person-year (PPY). We stratified all analyses by CHADS2 (congestive heart failure, hypertension, age ≥75 years, diabetes, previous stroke) score and stroke history,4 and used the Cochrane–Armitage test for trend and χ2 tests to identify differences in crude stroke rates. We performed a sensitivity analysis for timing of atrial fibrillation diagnosis.
We identified 148 446 patients aged ≥66 years with atrial fibrillation who began warfarin therapy over the study period, and of these patients, 6006 experienced an ischemic stroke. Of the patients who experienced an event, 27.3% (n=1639) died in hospital or within 7 days of discharge.
The 30-day cumulative incidence of ischemic stroke among new users of warfarin was 0.5% (n=721), increasing to 2.0% (n=2,917) after 1 year, 2.7% (n=4067) after 2 years, and 4.0% (n=6006) after 5 years. The risk was highest during the first 30 days of warfarin therapy: 6.0% PPY (95% confidence interval, 5.5%–6.4%) versus 1.6% PPY (95% confidence interval, 1.5%–1.6%) during the remainder of the 5-year follow-up (Table 2; Figure).
Crude 5-year rates of ischemic stroke increased significantly as CHADS2 scores increased (P<0.0001; Table 2; Figure I in the online-only Data Supplement) and among those with a history of stroke (P<0.001; Table 2; Figure II in the online-only Data Supplement). A sensitivity analysis demonstrated no difference in stroke rates according to timing of warfarin initiation after atrial fibrillation diagnosis (Table 2). Relative to patients who did not experience a stroke, those who did were less likely to have received an international normalized ratio test in the preceding 14 days (Table II in the online-only Data Supplement). Few patients received heparin or antiplatelet therapy (Table 1).
In this population-based study, the rate of ischemic stroke in older patients with atrial fibrillation starting warfarin was significantly elevated during the first month of treatment. This early elevated risk increased with previous stroke history and higher baseline CHADS2 score.
The overall 5-year risk of stroke observed in our study (1.8% PPY) is comparable with the annual incidence rate of 1.66% (95% confidence interval, 1.41%–1.91%) reported in a meta-analysis of 8 randomized controlled trials and an observational study by Azoulay et al,3,5 but higher than the rate of 1.17% PPY reported in another observational study of younger patients with lower CHADS2 scores.1 The explanation for our findings is likely multifactorial. The results may reflect the high-risk period after a transient ischemic attack,6 and the time required to achieve therapeutic anticoagulation. Less frequent international normalized ratio testing may have been a contributor. More research is needed to elucidate whether a transient hypercoagulable state seen immediately after warfarin initiation explains our findings.7 Our sensitivity analysis stratifying patients by timing of warfarin initiation after atrial fibrillation diagnosis found similar stroke rates in both groups. This suggests our findings are not simply influenced by elevated rates of stroke after atrial fibrillation diagnosis, as has been previously reported.8 Moreover, few patients received heparins or antiplatelet agents before initiating warfarin. Studies are needed to determine whether bridging therapy, antiplatelet therapy, or use of direct-acting anticoagulants mitigate this early increased stroke risk.
Strengths of this study include its large, population-based sample and long follow-up. Limitations include our inability to assess the quality of anticoagulation. We also could not capture stroke events that did not result in hospitalization or some important risk factors for stroke, such as hypertension.
In this large, population-based cohort study of older patients with atrial fibrillation, we observed the highest rate of ischemic stroke in the first 30 days after warfarin initiation. Although this does not provide evidence of a causal link between warfarin initiation and stroke, the persistence of thromboembolic risk despite anticoagulation highlights the need for future research in this area.
We thank Brogan Inc, Ottawa for use of their Drug Product and Therapeutic Class Database.
Sources of Funding
This study was supported by the Ontario Drug Policy Research Network which is funded by grants from the Ontario Ministry of Health and Long-Term Care (MOHLTC), and supported by the Institute for Clinical Evaluative Sciences (ICES). Opinions, results, and conclusions reported are those of the authors and independent from funding sources. No endorsement by ICES or the MOHLTC is intended or should be inferred. Dr Kapral holds a career investigator award from the Heart and Stroke Foundation of Ontario.
T. Gomes, M.M. Mamdani, J.M. Paterson, and D.N. Juurlink have received grant funding from the Ontario Ministry of Health and Long-Term Care (MOHLTC). J.M. Paterson is an employee of Institute for Clinical Evaluative Sciences, which is funded by the MOHLTC. M.M. Mamdani is a consultant for AstraZeneca, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, Hoffmann-La Roche, Novartis, Novo Nordisk and Pfizer. The other authors report no conflicts.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.114.007852/-/DC1.
- Received October 18, 2014.
- Revision received January 25, 2015.
- Accepted January 28, 2015.
- © 2015 American Heart Association, Inc.
- Gomes T,
- Mamdani MM,
- Holbrook AM,
- Paterson JM,
- Hellings C,
- Juurlink DN
- Azoulay L,
- Dell’Aniello S,
- Simon TA,
- Renoux C,
- Suissa S
- Lehto M,
- Snapinn S,
- Dickstein K,
- Swedberg K,
- Nieminen MS