The natural history of depression and differential patterns of development among subgroups of stroke patients: The South London Stroke Register

Abstract

Introduction: Carotid endarterectomy for symptomatic moderate carotid stenosis is still debated with variation in surgical practice. We demonstrated that MRI-detected plaque haemorrhage (PH) is associated with an increased risk of stroke in severe carotid stenosis and with all recurrent events in moderate stenosis. The aim of this study was to determine whether MRI-PH can predict future stroke in patients with mild/ moderate symptomatic carotid stenosis. Method: As part of the ICAD study, 135 patients with recent TIA/stroke, 30–69% carotid stenosis and no planned carotid intervention were recruited. Participants had a carotid MRI to determine PH status based on signal hyperintensity >150% of adjacent muscle. All patients were followed-up prospectively with stroke and diffusion positive TIA as the primary, and stroke as the secondary endpoints. Results: 132 patients completed the study (77 males, 55 females). 46 (34.8%) showed ipsilateral MRI-PH (35 males). We observed 12 strokes (8 MRI-PH and 4 without) and 13 infarctions during follow-up (median = 654 days, IQR = 643). Cox regression showed a HR of 4.51 (95% CI 1.38–14.68, P = 0.012) for infarction and 4.08 (95% CI 1.23– 13.58, P = 0.02) for stroke. Controlling for stenosis (>/<50%), HR for stroke was 4.69 (95% CI 1.40–15.74, P = 0.012). Strokes in patients without MRI-PH occurred late (after 3 years) in 3 out of 4. Conclusion: We show that MRI-PH is a significant predictor of future cerebral infarctions and stroke in recently symptomatic patients with mild/moderate carotid stenosis. MRI-PH status can thus offer decision support when there is clinical uncertainty regarding the benefit of carotid intervention.