266 Scientific Programme 22. European Stroke Conference © 2013 S. Karger AG, Basel 17 Management and economics Utility of CT brain-imaging in patients with symptoms typical of TIA: prospective popula-tion- based study NG Lovett, CE Lovelock, PM Rothwell for the Oxford Vascular Study N.G. Lovett1, C.E. Lovelock2, P.M. Rothwell3 Stroke Prevention Unit, John Radcliffe Hospital, Oxford, UNITED KINGDOM1, Stroke Pre-vention Unit, John Radcliffe Hospital, Oxford, UNITED KINGDOM2, Stroke Prevention Unit, John Radcliffe Hospital, Oxford, UNITED KINGDOM3 BACKGROUND: Most guidelines recommend brain imaging after TIA. MR imaging provides much useful clinical information (e.g. DWI, MRA, microbleeds etc), but in centres without access to routine MRI and in patients unable to undergo MRI, CT is recommended mainly to exclude intra-cerebral haemorrhage (ICH) and non-vascular pathology (e.g. tumour, subdural haematoma - SDH). Previous studies of such pathology in TIA and stroke report rates of 0.5%-15%, but were either highly selected cohorts and did not stratify rates by TIA vs stroke or typical/atypical symptoms. METHODS: In all patients referred to the Oxford Vascular Study TIA and stroke clinic with sus-pected TIA or non-disabling stroke (2002-2012), rates of major non-vascular pathology on initial CT or MRI brain-imaging were stratified by the nature of the presenting event (TIA vs stroke) and the certainty of the initial pre-scan clinical diagnosis (typical vs atypical symptoms). All patients were followed-up to identify intraparenchymal tumours (IT) misdiagnosed as infarction. RESULTS: Of 1811 patients with symptoms suggestive of cerebral TIA or minor stroke, 1768 (97.6%) agreed to brain imaging. In 728 patients with typical TIA, acute ICH was found in one case (0.1%, 95%CI 0-0.4) and major non-vascular pathology in 3 cases (0.4%, 0-0.9%; 1 small SDH, 2 IT). In contrast, major non-vascular pathology alone was found in 4/248 cases with atypical TIA (1.6%, 0.2-3.2%; 2 SDH, 1 IT, 1 meningioma), in 19/725 cases with typical stroke symptoms (2.6%, 1.5-3.8%; 16 IT, 1 demyelination, 1 arachnoid cyst, 1 subdural haematoma) and in 9/108 with atyp-ical stroke symptoms (8.3%, 5.1-30.5%; 9 IT). Five cases of IT presenting as stroke were misdiag-nosed as cerebral infarction on initial CT brain scan. CONCLUSION: In this large population-based study, the rate of acute ICH or serious non-vascular pathology in patients with typical TIA was about 0.5%, raising questions about cost-effectiveness of CT imaging if MRI is unavailable. 16 Management and economics Identification of strokes and stroke subtypes from hospital discharge diagnostic data: a popu-lation- based study S. Lourenco1, L.E. Silver2, P.M. Rothwell3 Stroke Prevention Research Unit, University of Oxford, Oxford, UNITED KINGDOM1, Stroke Prevention Research Unit, University of Oxford, Oxford, UNITED KINGDOM2, Stroke Prevention Research Unit, University of Oxford, Oxford, UNITED KINGDOM3 Background: Studies of trends in disease incidence, other epidemiology, screening for outcome events in clinical trials and health economic studies often use hospital diagnostic coding data, known as Hospital Episode Statistics (HES) in the UK. However, the validity of these data for stroke is uncertain, particularly for identifying pathological subtypes. We aimed to validate routinely collected ICD-10 HES diagnostic codes for hospitalised patients with stroke. Methods: Multiple sources of case-ascertainment were used, including daily identification and as-sessment of all potentially eligible hospital admissions, in a prospective population-based study of all incident and recurrent strokes (Oxford Vascular Study). All hospitalised strokes from 2002-2008 were prospectively categorised as ischaemic stroke, intracerebral haemorrhage (ICH) and subarach-noid haemorrhage (SAH) by study neurologists. These categories were subsequently compared with the ICD-10 primary and secondary codes given by HES for the same hospital admission. Results: Of 618 hospitalised strokes identified, only 157/506 (31%) ischaemic strokes had a HES dagnostic code for infarction or occlusion (ICD-10 I63-I66), 44/73 (60%) ICH had a HES code for intracerebral haemorrhage (ICD-10 I61-I62) and 26/39 (67%) SAH had a HES code for subarach-noid haemorrhage (ICD-10 I60). Of the remaining cases, 181 (29%) were coded as stroke not speci-fied (haemorrhage or infarction - ICD-10 I64), 26 as TIA and related syndromes (ICD-10 G45-G46), and there was no mention of cerebrovascular disease in 107 (17%) cases (ICD-10 I60-I69). There was no difference in the accuracy of the stroke subtype by age. Conclusion: The stroke subtype could be identified in less than half of hospitalised strokes and stroke was not recorded in a fifth of cases. In the UK, at least, discharge coding data alone are not sufficient to reliably identify or subtype strokes.
Karger_ESC London_2013
To see the actual publication please follow the link above