Objective: The effect of castration and subsequent replacement of dehydroepiandrosterone (DHEA) or estradiol on parasitemia, mortality and incidence of cerebral malaria (CM) was evaluated in CBA mice infected with Plasmodium berghei ANKA. Methods: Female mice were castrated, and groups of 12–15 animals received daily injections of DHEA, estradiol or saline. Four days after the start of treatment, mice were inoculated with 1 x 10⁶ P. berghei ANKA-parasitized erythrocytes. DHEA treatment was continued during the 5 days after infection, and estradiol was administered during the follow-up. Parasitemia was evaluated daily in Giemsastained blood smears. Signs of CM were determined by the manifestation of coma, limb paralysis and/or convulsions. Plasma TNF- α levels were evaluated by sandwich ELISA. Nitric oxide synthase (NOS) activity in the brain of moribund mice was measured by the method of Bredt and Snyder. Results: In non-castrated infected mice, the incidence of CM was 50%, and plasma TNF- α increased and brain NOS activity decreased compared to non-infected controls. Castration had no major effect on the parameters analyzed (parasitemia, mortality, CM incidence, TNF- α levels or NOS activity). Estradiol replacement caused a decrease in parasitemia but resulted in higher CM incidence and faster mortality, with an increase in NOS activity. Conclusions: Estradiol modulated the immune response of P. berghei ANKA-infected CBA mice, decreasing parasitemia and increasing NOS activity, and impacted negatively on survival and CM incidence, showing that neuroimmunoendocrine interactions are important in the physiopathogenesis of malaria infections.