Natural history of complications after intracerebral haemorrhage

<b>Background and purpose</b>: Numerous trials of haemostatic and neuroprotective agents for intracerebral haemorrhage (ICH) have failed. We characterized the risk of complications after ICH in a trial-eligible patient population, to inform safety in future trials. <b>Methods</b>: We used the Virtual International Stroke Trials Archive database to identify placebo-treated patients with spontaneous ICH, who were not comatose at admission, where randomization took place within 4 h of symptom onset, and where serious complication and outcome data were available. We described the complications encountered and assessed whether the absence of common complications influenced attainment of good functional outcome (mRS ≤4) at 90 days using logistic regression. <b>Results</b>: Of 201 patients examined, 70.2% experienced at least one serious complication. Neurological complications occurred in 21%, infections amongst 11%, and thromboembolic complications in 2%. Extension of the haemorrhage occurred most frequently: its absence was a significant predictor of good functional outcome (P < 0.0001, adjusted OR for good functional outcome = 21.9, 95% CI: [5.5, 88.3]). Neither infection, nor cardiac, nor thromboembolic complications influenced functional outcome at 90 days. <b>Conclusions</b>: Three month outcome in ICH patients depends on initial stroke severity and on enlargement of the haemorrhage. Our results should inform safety in future clinical trials of putative ICH therapies