Background Biomarkers might add to clinical measures of neurological impairment and age to predict poor outcome after stroke.
Methods We recruited symptomatic patients <24 h after stroke or TIA from an emergency department. We drew blood for markers of inflammation, thrombosis, thrombolysis, cardiac dysfunction and cerebral damage. We measured outcome at 3 months with the modified Rankin Scale (mRS). We used logistic regression, calculated measures of discrimination and reclassification and adjusted for multiple comparisons.
Results We drew blood from 285 patients (230 ischaemic strokes, 40 TIA, 15 haemorrhagic strokes) at a median of 7 h (IQR 3–19 h) after symptom onset and contacted 99% at 3 months: 160 (57%) had mRS 0–2 and 123 (43%) had mRS 3–6. After adjusting for neurological impairment and age, only IL-6 (OR 2.4 (95% CI 1.4 to 4.2), 75th to 25th centile) and Ln NT pro BNP (OR 2.2 (1.2 to 4.0)) were significantly associated with poor outcome. A clinical model containing the NIHSS score and age had an area under a receiver operator curve of 0.84 (95% CI 0.79 to 0.88); neither IL-6 nor Ln NT pro-BNP improved this significantly. There was no net reclassification improvement across clinically important boundaries after adding IL-6 or Ln NT pro-BNP to the clinical model.
Conclusion It is unlikely that the biomarkers measured in this study improve the clinical prediction of death or disability when measured at a uniform and early point after stroke or TIA.