Amphetamine enhances recovery after experimental ischaemia and has shown promise in small clinical trials when combined with motor or sensory stimulation. Amphetamine, a sympathomimetic, might have haemodynamic effects in stroke patients, although limited data have been published. Subjects were recruited 3-30 days post-ischaemic stroke into a phase II randomized (1:1), double-blind, placebo-controlled trial. Subjects received dexamphetamine (5 mg initially, then 10 mg for 10 subsequent doses with 3- or 4-day separations) or placebo in addition to inpatient physiotherapy. Recovery was assessed by motor scales (Fugl-Meyer (FM)), and functional scales (Barthel index (BI) and modified Rankin score (mRS)). Peripheral blood pressure (BP), central haemodynamics and middle cerebral artery blood flow velocity were assessed before, and 90 min after, the first two doses. Thirty-three subjects were recruited, aged 33-88 (mean 71) years, males 52%, 4-30 (median 15) days post stroke to inclusion. Sixteen patients were randomized to placebo and seventeen to amphetamine. Amphetamine did not improve motor function at 90 days; mean (s.d.) FM 37.6 (27.6) vs control 35.2 (27.8) (P=0.81). Functional outcome (BI, mRS) did not differ between treatment groups. Peripheral and central systolic BP, and heart rate (HR), were 11.2 mm Hg (P=0.03), 9.5 mm Hg (P=0.04) and 7 beats per minute (P=0.02) higher, respectively, with amphetamine, compared with control. A nonsignificant reduction in myocardial perfusion (BUI) was seen with amphetamine. Other cardiac and cerebral haemodynamics were unaffected. Amphetamine did not improve motor impairment or function after ischaemic stroke but did significantly increase BP and HR without altering cerebral haemodynamics.