After obtaining familial informed consent, between January 1996 and July 1997, 173 children (5 to 15 years old) with sickle cell disease were enrolled in a prospective multicenter study using blood screening, transcranial Doppler ultrasonography (n = 143), cerebral magnetic resonance imaging (n = 144), and neuropsychologic performance evaluation (n = 156) (Wechsler Intelligence tests WISC-III, WIPPSI-R), which were also performed in 76 sibling controls (5 to 15 years old). Among the 173 patients with sickle cell disease (155 homozygous for hemoglobin SS, 8 sickle cell beta0 thalassemia, 3 sickle cell beta+ thalassemia, 7 sickle cell hemoglobin C disease SC), 12 (6.9%) had a history of overt stroke, and the incidence of abnormal transcranial Doppler ultrasonography (defined as mean middle cerebral artery velocity > 200 cm/sec or absent) was 8.4% in the overall study population and 9.6% in patients with homozygous sickle cell anemia The silent stroke rate was 15%. Significantly impaired cognitive functioning was observed in sickle cell disease patients with a history of stroke (Performance IQ and Full Scale IQ), but also in patients with silent strokes (Similarities, Vocabulary, and Verbal Comprehension). However, infarcts on magnetic resonance imaging were not the only factors of cognitive deficit: Verbal IQ, Performance IQ, and Full Scale IQ were strongly impaired in patients with severe chronic anemia (hematocrit < or = 20%) and in those with thrombocytosis (platelets > 500 x 10(9)/L). Multivariate logistic regression analysis showed that abnormal magnetic resonance imaging (odds ratio [OR] = 2.76) (P = .047), hematocrit < or =20% (OR = 5.85) (P = .005), and platelets > 500 x 10(9)/L (OR = 3.99) (P = .004) were independent factors of cognitive deficiency (Full Scale IQ < 75) in sickle cell disease patients. The unfavorable effect of low hematocrit has already been suggested, but this is the first report concerning an effect of thrombocytosis and showing that silent stroke alone is not a factor of cognitive deficit when not associated with low hematocrit or thrombocytosis. The effect of hydroxyurea, which is known to increase hematocrit and decrease platelet count, on cognitive functioning of sickle cell patients should be evaluated prospectively.