Prenatal betamethasone (Celestene) therapy reduces the incidence of brain damage, whereas prenatal or neonatal dexamethasone (Soludecadron) increases the risk of brain lesions or neuromotor deficits. To determine whether this increase is ascribable to the sulfites used as preservatives in Soludecadron, we investigated the effects of 12 h of exposure to pure dexamethasone, Soludecadron, pure betamethasone, Celestene, and sulfites on in vitro and in vivo death of neurons cultured under basal conditions or with excitotoxic agents (N-methyl-D-aspartate or (S)-5-bromowillardiine) or hypoxia. Apoptotic features were quantitated using a fluorescent chromatin stain (Hoechst 33258). Neuronal viability was unaffected by pure dexamethasone, pure betamethasone, or Celestene. Soludecadron or sulfites significantly increased neuronal loss. Pure dexamethasone or pure betamethasone produced a 40-50% decrease in neuronal death induced by N-methyl-D-aspartate, (S)-5-bromowillardiine, or hypoxia, whereas Soludecadron had no effect and sulfites significantly increased the neurotoxicity of excitotoxic agents. In in vivo experiments involving terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling after several i.p. injections of fluorinated glucocorticoids, Soludecadron, but not pure dexamethasone, significantly increased the number of terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling-stained cells in neocortical layers and thalamus. These experimental findings suggest that injectable dexamethasone should be used with caution during the perinatal period.