LAUSR

Cerebrovascular complications are commonly observed in children with tuberculous meningitis. We aimed to determine which clinical factors were associated with stroke at admission in children with tuberculous meningitis and, in children stroke-free at admission, which factors were associated with development of stroke on treatment. We analysed a cohort of 474 children diagnosed with ‘definite’ and ‘probable’ tuberculous meningitis, with prospectively collected data, at Tygerberg Hospital, Cape Town, South Africa from 1985 to 2005. We considered either hemiparesis or radiological arterial ischemic infarction as evidence of stroke. At admission, 339 (71.5%) children presented with stroke. Features associated with stroke at admission included age <3 years (odds ratio (OR) 3.70; 95% confidence interval (CI): 2.44–5.63; p < 0.01), convulsions (OR: 2.25; 95% CI: 1.46–3.45; p < 0.01) and hydrocephalus (OR: 1.63; 95% CI: 1.05–2.53; p = 0.03). In the group of children without stroke at admission (n = 135), 33 (24.4%) developed stroke by 1 month. Similar factors predicted stroke and included age <3 years (OR: 2.60; 95% CI: 1.17–5.80; p = 0.02), convulsions (OR: 2.25; 95% CI: 1.46–3.45; p < 0.01), CSF cell count <10 or >500/L (OR: 3.12; 95% CI: 1.03–9.43; p = 0.04) and hydrocephalus (OR: 2.99; 95% CI: 1.30–6.89; p = 0.01). A large proportion of children with tuberculous meningitis present with stroke at admission. Of those with no evidence of stroke at admission, a quarter develop stroke by 1 month, suggesting that there could be a brief window in which to give preventive therapy.