The predictive value of casual blood pressure (BP) levels following acute stroke on outcome is currently unclear. This may in part reflect the observer bias and variability of casual recordings, which are reduced with 24-hour recordings. We therefore proposed to assess the prognostic significance of 24-hour compared to casual BP in predicting 30-day mortality, dependency and neurological outcome. A total of 136 consecutive patients were assessed within 24 h of ictus by one observer, with casual and 24-hour BP recording, and National Institutes of Health Stroke Scale and Modified Rankin Scale scores. Repeat assessments were made at 7 and 30 days. Admission casual and 24-hour systolic BP (SBP) and diastolic BP levels were significantly higher in patients with poor outcome at 1 month following acute stroke, whether expressed in terms of mortality, dependency or neurological deterioration, on single-variable logistic regression analysis. However, of these variables, only admission 24-hour (not casual) SBP remained a significant outcome predictor in a multiple model containing factors with an established association with poor prognosis. The odds ratio for outcome of death or dependency associated with each 10-mm-Hg increase in 24-hour SBP at admission was 1.88 (95% confidence interval: 1.27-2.78). For an outcome of death or high dependency, the model had a specificity of 75% and sensitivity of 76% when tested by the jackknife technique. Therefore, increasing 24-hour BP levels following acute stroke predict poor outcome. Whether BP should be reduced pharmacologically in the acute stroke period now warrants a suitable prospective intervention trial.