Thirty patients suffering from arterial occlusive disease (stages II, III or IV of Fontaine's classification) were treated by isovolaemic haemodilution (IDH). Most of these patients received conventional medical treatment without success (negative selection). Angiography was performed on all patients prior to treatment. 500 ml of whole blood was withdrawn from each patient. 250 ml of dextran 40 mixed with 250 ml of 5% albumin solution was injected intravenously to avoid hypovolaemia. This procedure, carried out three or four times, was repeated until a haematocrit of 0.30 was reached. This haematocrit level of 0.30-0.35 was maintained for 60 days. The measurement of ankle systolic pressure, systolic index at rest, treadmill performance and lower limb thallium-201 muscular scintigraphy were used to assess laboratory efficiency. On the sixtieth day, 14 patients (46%) showed clinical improvement, nine were stable and arterial occlusive disease worsened in seven patients. Clinical improvement persisted for several months (15 months in one case). Clinically, it seemed that patients treated by IHD showed improvement of their arterial occlusive disease. However, the prediction of the chances of success of IHD, as well as the quantification of results by common laboratory examination, were difficult. Muscular scintigraphy with thallium-201 gave a good assessment as well as a quantifiable analysis of results.