Background and purpose: Previous research has attempted to analyze the relationship between post-stroke hyperthermia and prognosis. These analyses have been hindered by a lack of information about the time course and determinants of temperature change after stroke.
Methods: Serial temperatures were measured until 48 h after ischaemic stroke in a prospectively recruited cohort. Potential determinants of temperature, including stroke severity [measured using the National Institutes of Health Stroke Scale (NIHSS)], infection and paracetamol use were recorded. Mixed-effects models were used to model serial temperature measurements over time, adjusted for significant determinants.
Results: In 155 patients the mean temperature rose from 36.5 degrees C at the time of stroke to 36.7 degrees C approximately 36 h later. The factors with significant multivariable associations with serial temperatures were: first- and second-order time components, infection, paracetamol administration and the interaction between stroke severity (NIHSS > or =6) and time (all p < 0.1). Patients with admission NIHSS > or =6 had a mean temperature rise of 0.35 degrees C during the first 36 h after stroke, compared with a rise of 0.17 degrees C in those with NIHSS < or =5.
Conclusions: Temperature spontaneously rises during the first 36 h after stroke, particularly after severer stroke and in the presence of infection.