The authors' specific aim was to assess hypocholesterolemia in 203 patients hospitalized because of affective disorders (depression, bipolar disorder, and schizoaffective disorder) compared with 1,595 self-referred subjects in an urban supermarket screening and with 11,864 subjects in the National Health and Nutrition Examination Survey II, a national probability sample. Low plasma cholesterol concentrations (< 160 mg/dL) were much more common in patients with affective disorders (20%) than in urban supermarket screenees (4%, P < or = 0.001) or in the National Health and Nutrition Examination Survey II subjects (10%, P < or = 0.001). When paired with supermarket screenees by age and sex, patients with affective disorders had much lower plasma total cholesterol (P < or = 0.0002), low-density lipoprotein cholesterol (P < or = 0.001), and high-density lipoprotein cholesterol (P < or = 0.0001), and higher triglyceride concentrations (P < or = 0.03). Neither the severity of the affective disorders nor severity-age interactions were associated with plasma cholesterol concentrations (P > 0.1); age and plasma cholesterol were positively associated (P = 0.01). None of the psychoactive drugs had a significant independent association with the patients' low-density lipoprotein cholesterol. Plasma cholesterol in patients hospitalized with affective disorders is shifted markedly downward toward hypocholesterolemic concentrations (< 160 mg/dL). There is no evidence that low plasma cholesterol could cause or worsen affective disorders.