Neuroactive steroids may play a role in the pathophysiology of psychotic disorders, but few studies examined this issue. We compared serum levels of cortisol, testosterone, dehydroepiandrosterone, and progesterone between a representative sample of first-episode psychosis (FEP) patients and age- and gender-matched healthy subjects. Furthermore, we analyzed the associations between neuroactive steroids levels and the severity of psychotic symptom dimensions. Male patients had lower levels of progesterone than controls (p=0.03). Progesterone levels were inversely associated with the severity of positive symptoms (p=0.007). Consistent with preclinical findings, results suggest that progesterone might have a role in the pathophysiology of psychotic disorders.