Single-photon emission tomography (SPET) is widely used in the investigation of acute stroke. We investigated the relationship between SPET data and functional outcome in a large group of acute stroke patients. One hundred and eight patients underwent cerebral computed tomography (CT) and technetium-99m hexamethylpropylene amine oxime SPET after acute ischaemic stroke. We categorised the clinical presentation according to the Oxford classification of acute stroke. Outcome was measured 1 year after stroke using mortality and the Barthel Index for survivors. SPET scans were interpreted without reference to the clinical data using a semi-automatic technique. Three experienced observers determined the presence of luxury perfusion using suitably scaled SPET images in conjunction with the CT scan. Both SPET volume and severity of deficit were significantly negatively correlated with Barthel Index at 1 year (rs=-0.310, P<0.0001, and rs=-0.316, P<0.0001 respectively). In patients scanned with SPET within 16 h of stroke onset, the correlations were more strongly negative (rs=-0.606, P<0. 001, and rs=-0.492, P<0.005 respectively). Luxury perfusion was not associated (chi2=0.073, df=1, P=0.79) with good functional outcome (Barthel score >/=60). Stepwise logistic regression identified Oxford classification, total deficit volume and patient's age as significant predictors of functional outcome. Overall predictive accuracy was 72%. Predictive accuracy was better in patients who received SPET within 16 h of stroke onset. SPET provides useful information about the functional outcome of acute stroke at 1 year. However, the accuracy of prediction decreases the longer SPET is delayed. Prognostication using SPET in combination with clinical assessment and other investigations may also be considered.