Response to Letter by Ray et al
We read with interest the report by Ray et al outlining their experience of using the ABCD2 scoring system to triage referrals to an outpatient transient ischemic attack (TIA) clinic. Their study of 75 patients supports the generalizability of our findings, which were based on a large (n=3646) but retrospective (1992–2005) sample of patients, in whom acute neuroimaging was infrequently available.1 Ray et al describe a prospective study with universal access to imaging, conducted with a geographic and demographically similar population. It is reassuring that these 2 studies reach the same conclusion: referrals with noncerebrovascular diagnoses have lower ABCD2 scores.
However, as we discuss in our original article, we should be cautious about direct use of ABCD2 as a diagnostic tool in clinical care. TIA is a medical emergency, necessitating prompt assessment and treatment.2 Efforts to reduce potential barriers to rapid management are to be applauded. It remains true that services designed to deliver evidence-based therapies to TIA patients are populated by a large proportion of noncerebrovascular diagnosis—43% in Ray’s contemporary study and 48% in our own. The use of triage tools to help target services to an exclusively cerebrovascular population makes intuitive sense, and ABCD2 would seem initially attractive for this purpose. However, based on the data presented, ABCD2 will misclassify around one fifth of strokes. The default must always be to assess as many “true” strokes as possible, at the cost of contamination by nonstroke diagnosis. In its current form we cannot recommend routine use of stroke triage based on ABCD2 only. Other scoring systems exist and we have recently developed a novel scoring system, purpose made for TIA diagnosis. We have demonstrated use of this tool in our out patient department, using statistics weighted to take account of the importance of missing a true cerebrovascular diagnosis.3
Resources for stroke care are finite, and as incidence of stroke and TIA increase we will likely see greater use of triage systems. Use of triage systems to rationalize patient care opens up new clinical and economic questions. For example, optimal management of those patients referred to TIA services with noncerebrovascular diagnosis is unknown. Recent reports have suggested that this group cannot be considered to have a “benign” prognosis.4 In our cohort of TIA referrals we have demonstrated substantial cardiovascular risk in noncerebrovascular patients,5 and we suggest that this group may also benefit from assessment by vascular disease specialists. We encourage further large scale prospective studies on TIA triage and management to help inform future service delivery.
Quinn TJ, Cameron AC, Dawson J, Lees KR, Walters MR. ABCD2 scores and prediction of non-cerebrovascular diagnoses in an outpatient population: a case-control study. Stroke. 2009; 40: 749–753.
Dawson J, Lamb KE, Quinn TJ, Lees KR, Horvers M, Verrijth MR, Walters MR. A recognition tool for transient ischaemic attack. QJM. 2009; 102: 43–49.
Cameron AC, Dawson J, Quinn TJ, Walters MR, Sloan B, Harrison J, Dreijer F, Lees KR. Long-term outcome following attendance at a transient ischaemic attack clinic. International Journal of Stroke. 2008; 3: s404–405.