Validation and Refinement of the ABCD2 Score: A Population-Based Analysis

1. Fothergill A, Christianson TJ, Brown RD, Rabinstein AA, ,  Validation and refinement of the ABCD2 score: a population-based analysis.,  Stroke; a journal of cerebral circulation,  40 (8) 2669-73. View in PubMed

This study was done to assess the clinical value of the ABCD2 score in a population-based cohort known as the TIA and Stroke Registry of the Rochester Epidemiology Project. The ABCD score is a combination of prognostic indicators in a model to determine one's stroke risk after a transient ischemic attack (TIA). The ABCD2 model includes diabetes in it.

The Rochester Epidemiology Project Medical Record Linkage system provides resources to identify new cases of stroke and TIA in the Olmstead County area. The medical records of those individuals with a TIA from 1985-1994 were examined (n=284). Patients were evaluated on the ABCD and ABCD2 scales, and new scores were created by adding a history of hypertension and hyperglycemia.

The ABCD score combines the variables of age, blood pressure, clinical features, and duration of symptoms into a 6-point scale. The ABCD2 score includes a history of diabetes. Previously, an emergency evaluation and hospital admission was recommended for those with a score of at least 4. However, there had not been independent validation of these scoring systems.

Scores greater than 4 had a statistically significant predictive value for stroke. However, a substantial proportion of strokes within 7 days (9/36 cases, 25%) affected individuals with low or intermediate risk scores (≤4) on the ABCD2 scale. If a history of hyperglycemia and hypertension on presentation was added, it increased the sensitivity of the score to identify patients who had a stroke within 1 week.

The 7-day stroke rate was relatively high compared with other recent series of more contemporary populations. Greater vigilance with regard to possible TIA symptoms and more urgent aggressive management of risk factors and implementation of single or dual antiplatelet therapy may explain some of this difference.

Four patients classified as low risk by the ABCD2 score had a stroke within a week of the TIA (6% of those in the ABCD2 low-risk category). It is, thus, premature to use the ABCD2 score alone to define subjects at low risk for a stroke in the near future.

Clinical Implication/Application

The ABCD2 score is useful, but it may not identify patients who are destined to have a stroke within a week after a TIA. Adding a history of hypertension and hyperglycemia at the time of presentation with a TIA, as well as information from imaging studies such as a brain magnetic resonance imaging and carotid ultrasound, should be evaluated in the future to improve diagnostic accuracy.

The study had several limitations. More extensive scores could not be calculated for all subjects because of the retrospective nature of the investigation. Missing information hampered the categorization of some of the subjects who had a stroke within 1 week of the TIA when testing the scores with additional variables and this impaired the ability of the investigators to validate the scores more definitely. The predictive value of the scores could be regarded as hypothesis-generating observations.

Clinicians need to remember the ABC's of secondary prevention when dealing with TIA patients: antiplatelet therapy, blood pressure management, cholesterol and cigarette management, diet/weight instruction, diabetes prevention and/or management, exercise prescription, and family history implications. 

-- The opinions expressed in this commentary are not necessarily those of the editors or of the American Heart Association.