ABCD2 Score as a Predictor of Three-Year Stroke Risk
February 15, 2025
By Alan Z. Segal, MD
Synopsis: This retrospective cohort study using the National Patient Registry in Denmark analyzed a three-year follow-up period following a transient ischemic attack, looking at the rate of recurrent stroke and mortality. The patients were divided into low risk (ABCD2 score 0-3) and high risk (ABCD2 score ≥ 4). The stroke rate was 6% in the high-risk group and 4% in the low-risk group.
Source: Al-Chaer K, Alhakak A, Emborg Vinding N, et al. Incident stroke after first-time TIA according to ABCD2 score — a nationwide cohort study. Neurology. 2024;103(12):e210053.
The ABCD2 score is a method of risk stratification to predict the likelihood of stroke in a patient presenting with transient ischemic attack (TIA). The score is higher in patients older than age 60 years (A = age), who have hypertension (B = blood pressure), clinical features such as unilateral weakness or speech disturbance (C = clinical), duration of symptoms > 1 hour (D1 = duration), and presence of diabetes (D2 = diabetes). Since this score is purely clinical, the presence or absence of a corresponding infarction on imaging (optimally, magnetic resonance imaging [MRI] with diffusion-weighted imaging [DWI], rather than a computed tomography [CT] scan) cannot be factored in. It is well recognized that a subset of patients with a very high ABCD2 score (6-7), with manifestations such as hemiparesis lasting more than an hour, actually are having a stroke rather than a TIA, if DWI were performed.
Although it may inappropriately be used as such, the ABCD2 score is not a substitute for clinical judgment or determination of whether a transient neurological event actually is a TIA or something else (such as acephalgic migraine or any other stroke mimic). Patients presenting with fleeting vague sensory symptoms most likely are not actually having a cerebrovascular event. Therefore, a low ABCD2 score and the absence of subsequent stroke would not be reflective of whether a patient’s TIA was high-risk, but rather that they never had a TIA at all.
Previously published data on the ABCD2 score have demonstrated that stroke risk following TIA can be as low as 1% at 48 hours and still only 3% at 90 days for scores of 0-3. By contrast, stroke risk in patients with scores of 6-7 are 8% at 48 hours and 18% at 90 days. However, of perhaps greater importance than risk assessment is the question of whether a change in management will produce a meaningful improvement of the natural history of TIA/stroke in any given patient. Detection of atrial fibrillation (either immediately or on long-term monitoring) is important, since oral anticoagulants are superior to antiplatelet therapy to prevent embolism.
“Vindicating the carotid” (ruling out severe stenosis) is important, especially in the situation of “crescendo TIAs,” which can progress rapidly into a major infarct. Carotid disease easily can be assessed by Doppler, CT angiography, or magnetic resonance angiography, but severe stenosis actually is only found in a minority of TIA patients. In the absence of these specific scenarios, any TIA patient, regardless of etiology, should be treated with antiplatelet therapy (aspirin alone or dual antiplatelet therapy [DAPT, aspirin + clopidogrel]) and aggressive risk factor management, particularly with antihypertensives and statins. Hospital admission for TIA often is necessary, given that the 48-hour stroke risk is as high as 8%, unless an expedited outpatient workup can be rapidly accomplished. Of note, immediate recurrent symptoms may relate to a specific small vessel stroke mechanism (known as a “stuttering lacune”). With this, fluctuating symptoms (even back to normal) may be mislabeled as a TIA followed by a stroke, when in fact there was only one stroke event evolving over time. The data suggest (but do not prove) that DAPT is superior to aspirin alone in this setting.
In the present study, the three-year stroke risk following TIA was analyzed in a large cohort (n = 21,433) derived from the National Patient Registry in Denmark. Patients were dichotomized according to ABCD2 score as low-risk (score 0-3, n = 20,153) and high-risk (score ≥ 4, n = 1,280). A modified version of the ABCD2 score was used because the registry lacked data on blood pressure and symptom duration. Imaging (CT or MRI scan but not clear which) was performed in 99% of patients.
The overall three-year event rate was very low, 6% in the high-risk group compared to 4.2% in the low-risk group. This 1.8% absolute difference translated into a statistically significant hazard ratio of 1.56. Stroke incidence (high vs. low risk) also was reported at 90 days (2.2% compared to 1.2%) and at one year (3.2% compared to 2.4%).
Factors associated with increased stroke incidence (which are not all components of the ABCD2 score) included age ≥ 60 years, current smoking, unilateral weakness, peripheral artery disease, and chronic kidney disease. The most profound of these was age, since the high-risk patients were older (98% older than 60 years of age), compared to 75% older than 60 years of age in the low-risk population. This produced a hazard ratio of 2.21, a far more robust effect than for any other factor (which were all less than 1.5).
Multiple other characteristics distinguished the high-risk group but were nonsignificant in multivariate analysis. These included hypertension, diabetes, atrial fibrillation, and ischemic heart disease/heart failure. For hypertension, 87% of patients in the high-risk group required two or more drugs, compared with 28% in low-risk patients, which is emblematic of the overall burden of vascular disease in the high-risk category.
The overall mortality rates were much higher than the stroke rates, 29% in the high-risk group and 10% in the low-risk group. The Kaplan-Meier curves for stroke separated at 90 days, with absolute differences in event rates relatively static over time, while the curves for mortality continued to dramatically separate at two and three years. Treatment was with anticoagulants or antiplatelet medications, which were given approximately 93% of the time in both groups. Antiplatelet agents included aspirin, adenosine diphosphate inhibitors (e.g., clopidogrel), or a phosphodiesterase inhibitor (dipyridamole). The cases were collected between 2014 and 2020, reflecting standard of care at that time, when Aggrenox (acetylsalicylic acid/dipyridamole) was a common agent used in secondary prevention.
Regarding stroke mechanism, only 1% of the population had carotid artery disease, although it is unclear at what rate this was investigated. The authors noted that there were no data on the severity of the carotid stenosis or the presence of intracranial atherosclerotic disease (ICAD).
Commentary
The ABCD2 score is a useful construct for rapid evaluation of patients in the emergency room, but it should not guide long-term treatment strategies. Although the vast majority of patients in this cohort were imaged, it is not known how many underwent MRI-DWI and had underlying infarcts, even if their symptoms were completely reversible.
Definitions of TIA vs. stroke remain muddled, since a TIA can be diagnosed for a syndrome lasting up to 24 hours, but the decision to treat a patient with thrombolysis for stroke must be made in the first 4.5 hours. It is impossible to know how many patients treated for stroke with thrombolysis could have resolved within 24 hours and been labeled as a TIA.
Studies such as this one, which do not include analysis of stroke subtypes (with detailed vascular imaging), can only take a general look at recurrence risk for cerebrovascular events. Without such data, it is not surprising that recurrent stroke was driven by risk factors, the most prominent being advancing age. This likely explains why nearly a third of high-risk patients were dead within the three years, an outcome that far outweighs the stroke rate of 6%. As the authors noted, this cohort of patients was well treated, with emphasis of control of risk factors and medications for secondary stroke prevention. Of course, none of this can combat Father Time.
These data are further of limited use in informing the choice of treatment in TIA patients. Currently, the use of DAPT in patients with minor stroke or TIA is limited to 21 days (according to the CHANCE and POINT trials) or 90 days (in patients with ICAD, according to the SAMMPRIS trial). This study does serve as a reminder that patients with high-risk features have a heightened likelihood of stroke extending well beyond these short-term targets. As the authors observed, the safety and efficacy of ongoing DAPT beyond 90 days in high-risk patients should be explored in future clinical trials.
Alan Z. Segal, MD, is Associate Professor of Neurology, Weill Cornell Medicine.
This retrospective cohort study using the National Patient Registry in Denmark analyzed a three-year follow-up period following a transient ischemic attack, looking at the rate of recurrent stroke and mortality. The patients were divided into low risk (ABCD2 score 0-3) and high risk (ABCD2 score ≥ 4). The stroke rate was 6% in the high-risk group and 4% in the low-risk group.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content