Validation assessment of risk scores to predict postthrombolysis intracerebral haemorrhage

Abstract

Background: Two clinical risk scores, the Haemorrhage After Thrombolysis and Multicentre Stroke Survey scores, have been proposed to predict the risk of intracerebral haemorrhage following thrombolysis in acute ischaemic stroke.

Aims: To validate Haemorrhage After Thrombolysis and Multicentre Stroke Survey scores as predictors of post-tissue plasminogen activator symptomatic intracerebral haemorrhage and asymptomatic intracerebral haemorrhage in an independent cohort.

Methods: Haemorrhage After Thrombolysis and Multicentre Stroke Survey scores were calculated for the cohort of tissue plasminogen activator-treated patients enrolled in the placebo arms of the SAINT-I and SAINT-II trials. The absolute risk of symptomatic intracerebral haemorrhage and asymptomatic intracerebral haemorrhage associated with each scoring system was determined. The overall predictive value was assessed using c-statistics.

Results: Symptomatic intracerebral haemorrhage occurred in 5.6% of 965 patients treated with tissue plasminogen activator in the SAINT cohorts. The risk of symptomatic intracerebral haemorrhage was modestly greater, with higher Haemorrhage After Thrombolysis scores (0: 4.1%, 1: 4.1%, 2: 8.8%, 3: 12.5%, 4: 0%, 5: no subjects). Similar results were seen with the Multicentre Stroke Survey score (0: 0%, 1: 4.8%, 2: 2.3%, 3: 7.3%, 4: 6.3%). In logistic regression, the Haemorrhage After Thrombolysis score was associated with the risk of symptomatic intracerebral haemorrhage (odds ratio = 1.41 per point, 95% confidence interval: 1.05-1.89, P = 0.021) and asymptomatic intracerebral haemorrhage (odds ratio = 1.59 per point, 95% confidence interval: 1.33-1.92, P< 0.001). The Multicentre Stroke Survey score was modestly associated with the risk of symptomatic intracerebral haemorrhage (odds ratio = 1.43 per point, 95% confidence interval: 0.95-2.15, P = 0.084) and asymptomatic intracerebral haemorrhage (odds ratio = 1.63 per point, 95% confidence interval: 1.27-2.08, P < 0.001). The c-statistic was 0.59 for predicting symptomatic intracerebral haemorrhage and 0.61 for asymptomatic intracerebral haemorrhage for both the Haemorrhage After Thrombolysis and the Multicentre Stroke Survey scores.

Conclusions: While both the Haemorrhage After Thrombolysis and Multicentre Stroke Survey scores were associated with a risk of symptomatic intracerebral haemorrhage, discriminatory ability was limited.