The effects of intravenous glycerol in elderly patients with recent onset of acute ischaemic stroke were evaluated in a double-blind randomised controlled trial. 173 patients received either 500 ml of a 10% solution of glycerol in physiological saline or 500 ml of physiological saline administered intravenously over 4 h daily for 6 consecutive days. The number of deaths within the first week was 10 (12%) in the glycerol group versus 26 (30%) in the controls. Subsequent mortality up to 12 months was similar in the two groups and a survival analysis confirmed a beneficial effect of treatment (p less than 0.02). The neurological and functional recovery of survivors, their length of hospital stay, and the proportion able to return to live in their own home were similar in the two groups. The improvement in survival time with glycerol was achieved without serious adverse effects and without an increase in the proportion of survivors with severe residual disability.