Although the benefit of antithrombotic therapy for stroke prevention in atrial fibrillation (AF) is well recognized, its potential effect on stroke severity and outcome is less well established. Our objective was to examine the effect of preadmission antithrombotic therapy on stroke severity and outcome in patients with AF within a large comprehensive nationwide stroke survey. The data from consecutive patients with AF admitted with acute ischemic stroke or transient ischemic attack during a 2-month period were collected. The patients were categorized into 4 groups according to the use of preadmission antithrombotic therapy: no antithrombotic therapy, antiplatelet therapy, warfarin with an admission international normalized ratio (INR) <2 and INR of ≥2. Of 1,938 patients presenting with acute brain ischemia, 329 (17%) had AF. The age-adjusted rate of more severe stroke (baseline National Institutes of Health stroke scale score >5) stratified by antithrombotic therapy use was 70% for no antithrombotic therapy use, 55% for antiplatelet therapy use, 59% for warfarin with an INR <2, and 38% for warfarin with an INR of ≥2 (p = 0.01). Compared to warfarin therapy with an admission INR of ≥2, the adjusted odds ratio for more severe strokes was 4.0 (95% confidence interval [CI] 1.7 to 10.0) for no antithrombotic therapy, 2.2 (95% CI 1.0 to 9.4) for antiplatelet therapy, and 2.7 (95% CI 1.1 to 6.7) for warfarin therapy with an INR of <2. Similarly, graded associations of antithrombotic medication were observed with severe disability (modified Rankin Scale score >3) or death at discharge, with corresponding adjusted odds ratios of 4.1 (95% CI 1.8 to 9.9), 2.1 (95% CI 1.0 to 4.6), and 1.5 (95% CI 0.6 to 3.5), and 1-year mortality, with corresponding adjusted ORs of 2.4 (95% CI 0.9 to 6.7), 1.9 (95% CI 0.8 to 5.0), and 2.2 (95% CI 0.8 to 6.2). In conclusion, in addition to its established benefit for stroke prevention, effective anticoagulation therapy is associated with decreased stroke severity and better functional outcome and survival in patients with AF presenting with acute brain ischemia.
Atrial fibrillation (AF) is the most common arrhythmia in older persons and a potent risk factor for stroke. AF is not only a frequent cause of stroke in the aging Western population, but also considerably increases stroke mortality and morbidity, leaving many patients permanently disabled. Although the benefit of antithrombotic therapy for stroke prevention in patients with AF is well established, its potential effects on stroke severity and outcome is less clear. Recently, reports have suggested that effective preadmission anticoagulation is associated with better outcomes in patients with AF and acute ischemic stroke. To test this hypothesis further, we examined the association between preadmission antithrombotic therapy and stroke severity and outcome in a large comprehensive nationwide study of unselected patients with AF and acute brain ischemia.
Methods
The National Acute Stroke Israeli Survey (NASIS) was a prospective nationwide survey conducted during a consecutive 2-month period (February and March 2004) to assess the incidence, characteristics, management, and outcome of hospitalized patients with acute cerebrovascular events (e.g., stroke or transient ischemic attack [TIA]). The survey was performed throughout all hospitals operating in Israel (n = 28). All patients with acute stroke or TIA, aged ≥18 years, nationwide, who had been hospitalized during this period were included. The research and ethics review boards approved the survey.
A study neurologist responsible for data collection throughout the hospital wards was designated for each hospital. The patients were identified by screening of all hospital departments for patients with stroke or TIA and by a review of medical records throughout the survey period. A standardized case-report form was used with uniform definitions for patient-related variables, clinical diagnoses, treatment modalities, and in-hospital complications. To ensure uniformity of data collection, the study investigators underwent training and received a detailed study manual. Stroke and TIA were reported in accordance with the medical report on discharge from the hospital. Whenever doubt existed regarding the diagnosis, a central adjudication committee, consisting of experienced stroke neurologists, made the final decision. Ischemic stroke and intracerebral hemorrhage were classified according to brain computed tomographic and/or magnetic resonance imaging findings. The few cases in which no brain imaging had been performed were counted as “undetermined stroke.”
Neurologic deficits were determined by a study neurologist within each center according to the National Institutes of Health stroke scale (NIHSS) score in categories of ≤5, 6 to 10, 11 to 15, 16 to 20, or >20. Functional disability was determined using the modified Rankin Scale (mRS) in categories of 0 to 1, 2 to 3, and 4 to 5 from clinical evaluation or a review of the medical charts. The clinical subtypes of stroke were classified according to the Oxfordshire Community Stroke Project (OCSP) classification and were categorized into lacunar versus all other clinical subtypes. Documentation of previous stroke or dementia was determined from the patient’s history and/or medical records.
Severe hemorrhage was defined as a decrease in hemoglobin by 5 mg/dl or the necessity for ≥2 blood transfusions. In all patients treated with warfarin, the blood samples for determining the international normalized ratio (INR) levels were taken on admission to the hospital and were recorded in the medical chart.
Data entry, editing, and analysis were performed at the coordinating center of the Israel Society for the Prevention of Heart Attacks. Data checks for completeness and consistency were done using the discharge reports attached to each patient form and computerized data queries. The data were verified using on-line checks incorporated into the data entry interface and by batch logical checks applied to the database following the data entry process. Study neurologists resolved all queries at the coordinating center, using the data from the discharge summaries. Mortality during the first year after hospitalization was assessed by matching the patients’ files with the national mortality registry.
For the present analyses, we included patients with acute brain ischemia (ischemic stroke or TIA) and AF (known previously or diagnosed during admission). Five patients for whom information about antithrombotic treatment before admission was not complete were excluded from the present analysis. The patients with AF were categorized into 4 groups according to their antithrombotic medication before admission: no antithrombotic therapy, antiplatelet therapy only (aspirin and/or clopidogrel, ticlodipine, or dipyridamole), warfarin with an admission INR <2, and warfarin with an admission INR of ≥2. Patients who were taking both warfarin and antiplatelet therapy were assigned to the warfarin group.
Univariate associations between the 4 groups and other baseline demographic and clinical variables were evaluated with analysis of variance and chi-square tests. The effect of the 4 groups on hospital duration was analyzed using a nonparametric Wilcoxon test. The survival rates after 1 year of follow-up are presented as Kaplan-Meier curves and were compared using the log-rank test. Multivariate analyses were applied to assess the effect of the antithrombotic medication group before admission on the severity of stroke measured according to the NIHSS score and on poor outcome and mortality. The antithrombotic medication category was entered into the logistic model as 3 dummy variables, with warfarin with an INR of ≥2 as the reference group. We included in the multivariate model, age as a continuous covariate and gender, hypertension, dyslipidemia, previous stroke, pre-event disability (per the mRS), dementia or malignancy, congestive heart failure or mechanical valve, and lacunar versus nonlacunar stroke as categorical variables. Additional sensitivity analyses were performed, excluding patients with lacunar strokes. The analyses were performed using the Statistical Analysis Systems statistical software, version 8.2 (SAS Institute, Cary, North Carolina).
Results
During the 2-month nationwide survey, 2,175 patients had an acute cerebrovascular event. Of 1,938 patients with acute brain ischemia (ischemic stroke or TIA), 329 (17%) had AF and 324 had complete data on preadmission antithrombotic management. Most patients had nonvalvular AF (95%; n = 307); only 5% (n = 17) had valvular AF. Of the patients with AF, 89 had received no antithrombotic therapy before the event, 124 had received antiplatelet therapy, and 111 had received warfarin. Of the patients receiving warfarin therapy, 40% were octogenarians compared to 54% octogenarians in the group receiving no antithrombotic or antiplatelet therapy (p = 0.02). Figure 1 shows the distribution of admission INR values for the patients admitted with warfarin therapy. Of those receiving warfarin therapy, 43% (n = 48) had an admission INR of ≥2% and 57% (n = 63) had an admission INR of <2. The demographic characteristics of the patients stratified by preadmission antithrombotic medication group are listed in Table 1 .
| Variable | None (n = 89) | Antiplatelet (n = 124) | Warfarin (n = 111) | p Value | |
|---|---|---|---|---|---|
| INR <2.0 (n = 63) | INR ≥2.0 (n = 48) | ||||
| Mean age (years) | 78 ± 12 | 79 ± 8 | 76 ± 9 | 75 ± 11 | 0.02 |
| Women | 63% | 60% | 62% | 48% | 0.36 |
| Previous severe disability (modified Ranking Scale 4-5) | 21% | 15% | 8% | 8% | <0.01 |
| Hypertension ⁎ | 65% | 84% | 79% | 75% | 0.02 |
| Diabetes | 28% | 39% | 31% | 31% | 0.45 |
| Dyslipidemia † | 26% | 46% | 31% | 48% | 0.01 |
| Current smoker | 6% | 5% | 3% | 7% | 0.88 |
| Renal failure | 22% | 18% | 16% | 19% | 0.85 |
| Past stroke | 16% | 35% | 31% | 38% | 0.01 |
| Past angina pectoris | 20% | 29% | 25% | 38% | 0.15 |
| Past myocardial infarction | 18% | 22% | 23% | 21% | 0.88 |
| Heart failure | 25% | 29% | 37% | 33% | 0.43 |
| Mechanical valve | 0% | 0% | 7% | 17% | <0.01 |
| Past coronary artery bypass graft/percutaneous coronary intervention | 9% | 15% | 10% | 13% | 0.60 |
| Peripheral arterial disease | 2% | 9% | 17% | 21% | <0.01 |
| Oxfordshire Community Stroke Project Stroke type | 0.03 | ||||
| Total anterior circulation infarct | 27% | 19% | 36% | 0% | |
| Partial anterior circulation infarct | 41% | 45% | 44% | 44% | |
| Posterior circulation infarct | 15% | 19% | 9% | 28% | |
| Lacunar infarct | 15% | 13% | 9% | 22% | |
| Unknown | 3% | 4% | 2% | 6% | |
⁎ Hypertension, defined by history or medical record.
The stroke severity on admission was associated with the type and intensity of preadmission antithrombotic therapy, as was the in-hospital mortality and functional disability at discharge. The age-adjusted rates of more severe stroke on admission (NIHSS score >5) and the rates of severe disability or mortality (mRS >3) at hospital discharge were greatest in patients who were receiving no antithrombotic therapy, intermediate in patients receiving antiplatelet therapy or subtherapeutic anticoagulation, and lowest in patients receiving warfarin with therapeutic INR levels ( Figure 2 , Table 2 ).
| Variable | No Antithrombotic Therapy (n = 89) | Antiplatelet Therapy (n = 124) | Warfarin (n = 111) | p Value | |
|---|---|---|---|---|---|
| INR <2.0 (n = 63) | INR ≥2.0 (n = 48) | ||||
| Baseline National Institute of Health Stroke Scale >5 | 70% | 54% | 59% | 38% | 0.01 |
| Discharge modified Rankin Scale 4-5 or deceased | 65% | 50% | 43% | 36% | <0.01 |
| Mortality rate | |||||
| At 1 month | 24% | 14% | 13% | 12% | 0.09 |
| At 3 months | 32% | 22% | 16% | 12% | 0.02 |
| At 6 months | 35% | 27% | 20% | 16% | 0.05 |
| At 1 year | 40% | 31% | 31% | 26% | 0.23 |
From the assessment according to the clinical subtype classification by the Oxfordshire Community Stroke Project, no patient with therapeutic INR levels who received warfarin had a total anterior circulation infarct compared to 27% of patients who received no antithrombotic medication, 19% who received antiplatelet therapy, and 36% who were treated with warfarin but were admitted with subtherapeutic INR levels. The proportion of patients with lacunar strokes was greatest in the patients with therapeutic INR levels (p = 0.03, Table 1 ). The association between the preadmission antithrombotic medication and stroke severity and outcome remained significant after excluding patients with lacunar strokes (data not shown).
Any antithrombotic medication status other than effective warfarin therapy (INR of ≥2) seemed to be associated with increased stroke severity and unfavorable outcomes after adjustment for potential confounders, including the increased proportion of lacunar strokes in patients with effective anticoagulation on admission. Comparable graded associations of antithrombotic medication were observed with severe disability or mortality (mRS >3) at hospital discharge ( Table 3 ). Additional adjustment for categories of NIHSS attenuated these associations.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
