Poster and platform presentations Conclusions: Cortical infarcls and PICH have significant differences in both absolute and cusum derived diurnal BP change compared to controls. This may reflect damage to the central component of the baroreceptor reflex arc. As yet the prognostic implications of these findings are unclear. A non-invasive examination of scapula motion in affected and unaffected shoulders after stroke : correlations with pain, muscle tone and subluxation. 96 C.Price 1 , P. Franklin 3 , R. Curiess 1 , G.Johnson 3 and H. Rodgers 2 1 Academic Department, North Tyneside General Hospital, department of Medicine (Geriatncs) & 3 Centre for Rehabilitation & Engineering Studies, Newcastle University. Introduction Painful restriction of shoulder movement after stroke is common, but its nature remains unclear. The relationship between scapula motion and pain has not previously been considered. Following the development of a scapula locator that utilises surface landmarks within an electromagnetic field (Johnson GR 1993 Clin. Biomech. 8:269-273), we have been able to record magnitude and patterns of scapula rotation (ScR) during humeral abduction. Methodology We examined the affected (AS) & unaffected (US) shoulders of 30 patients 6 months post-stroke (23 with first ever stroke) during humeral abduction from 0 to 50 degrees. The following correlations were considered : a history of shoulder problems pre- stroke, current pain, upper limb Motricity score, muscle tone, palpable subluxation & performance on the Frenchay Arm Test. Results Three patterns of ScR emerged when the AS & US were compared : 16 subjects had symmetrical scapula movement; 6 had quicker progression of ScR in the AS than US ("lead"); 8 had slower rate of ScR in the AS ("lag"). Using Fishers exact test & Mann-Whitney U test, lead was associated with pain (p=0.007). Lag was associated with subluxation (p<0.007), increased tone (p=0.02), lower Motricity score (p=0.0023) & reduced function (p<0.005), but not pain. There was no association between pain & subluxation. A history of shoulder problems pre-stroke showed no associations. These relationships also held for 1st ever stroke. Conclusions This new technique has provided information not available by clinical examination, and has demonstrated three clear ScR categories with different clinical associations. A prospective study is planned to investigate these findings in more detail THE DYNAMICS OF FUNCTIONAL RECOVERY FROM 97 STROKE IN SPECIALIST SETTINGS M PATEL, J POTTER, I. PEREZ, C SWIFT AND L KALRA Nunnery Fields Hospital, Canterbury and Clinical & Health Services Studies Unit, King's College School of Medicine & Dentistry, London Introduction The rate of functional recovery may vary according to initial stroke severity This study investigates the relationship between stroke seventy, functional recovery and hospital discharge in 300 stroke unit patients Methodology Eligible subjects (n=273) were stratified using the Barthel Index (BI) at 7 days post stroke into: mild(BI >10)<n°44), moderate(BI 6-10)(n = 106) and severe(BI 0-5X123) BI was recorded at weekly intervals for 20 weeks and plotted against rate of discharge in each patient group. Results The baseline demographic characteristics were comparable in all three groups. Median baseline and discharge BI differed between groups - mild-13/18, moderate 8/15, severe 4/11, p<0.001. Median lengths of stay were. 37, 57, 71 days (p<0.001). Changes in BI were similar. 5v7v7. All groups showed a rapid and linear increase in median BI after week 1 reaching a plateau at: mild - 4 weeks, moderate - 12 weeks, severe - 10 weeks. The rate of change in BI was similar in the moderate and severe groups (0.7/week and 0.6/week) but were slower than in the mild group (1.25/weekXp<0.01). Functional recovery was complete in mild strokes by week 4 compared with 80% of possible recovery in moderate stroke (p<0.0001) and 70% of possible recovery in severe strokes (p<0.0001) Functional recovery time accounted for 80% of hospital stay in all patient groups. 20% of hospital stay was accounted for by other factors and was not influenced by initial stroke severity Conclusions Functional recovery occurs in a constant definable pattern after stroke, the extent and rate of which depends upon initial severity Differences in hospitalisation between severe and other stroke patients appear to be due to slower recovery rather than delays in discharge on specialist units 98 CAN SIMPLE CLINICAL FEATURES BE USED TO IDENTIFY SYMPTOMATIC PATIENTS WITH SEVERE CAROTID STENOSIS? GE MEAD JM WARDLAW S LEWIS AND MS DENNIS Department of Cfinfcd Neuroaoiences, Western General Hospital, Edinburgh Introduction CaroSd endarterectomy reduces the risk of stroke in symptomatic patients with severe Ipelateral carotid stenosis. Symptomatic paSents should therefore undergo carotid Doppter Imaging, but in some centres access to imaging is Bmrted We herefore investigated whether simple cfinlcaf features alone or in combination could be used to identify patents with severe carotid stenosis, so that they could be referred preferenSaSyforcarotid maging. 1041 pafiente wfih acute stroke, transient techaemJc attacks (TIAs) or retinal artery occlusion, admitted to our hospital or seen in neurovasouiar cSnics were 48