Multicenter trial of hemodilution in acute ischemic stroke. Results of subgroup analyses. Scandinavian Stroke Study Group.

In a multicenter trial, 183 patients with acute ischemic stroke of less than 48 hours' duration and hematocrits of 38-50% were randomized to standardized hemodilution treatment (venesection and dextran 40 administration) and 190 to a control group. We have previously reported that there were no beneficial effects of hemodilution in the total patient population. In this report, the case fatality rates and neurologic outcome in survivors (3 months' follow-up) in subsets of patients have been analyzed. The patients were subgrouped by sex, age, medical history, smoking habits, delay from the onset of symptoms to the start of treatment, hematocrit at entry, venesection volume, neurologic score at entry into the study, blood pressure changes in the acute phase, presence of atrial fibrillation, location of brain lesion by computed tomography, type of diagnostic procedures, and hospital setting. No subset in which hemodilution reduced mortality or improved neurologic outcome could be identified. Case fatality rate was apparently higher in hemodiluted patients with infarction affecting deep brain structures than in control patients with such lesions. By simple clinical criteria, we have been unable to define subsets of stroke patients who benefit from the present standardized regimen of moderate hemodilution. The sample sizes are, however, too small to refute the possibility that a modest clinical effect of hemodilution may be present in some patients with stroke.