Altered estrogen-induced neuroprotection has been implicated in the etiology of psychotic disorders. Using bone mineral density as a marker of lifetime estrogen exposure, a longitudinal family study was conducted to discriminate between etiological mechanisms and secondary effects of disease and treatment. Dual X-ray absorptiometry scans were acquired twice, with an interval of 3 years, in 30 patients with psychotic disorder (male (M)/female (F): 24/6, mean age of 32 years at second measurement), 44 non-psychotic siblings of patients with a psychotic disorder (M/F: 26/18, mean age 32) and 27 controls (M/F: 7/20, mean age 35). Total bone mineral density, Z-scores and T-scores were measured in the lumbar spine and proximal femur. Associations between group and bone mineral density changes were investigated with multilevel random regression analyses. The effect of prolactin-raising antipsychotic medication was evaluated. (Increased risk of) psychotic disorder was not associated with disproportionate bone mineral density loss over a three year period. Instead, femoral bone mineral density measures appeared to decrease less in the patient versus control comparison (total BMD: B = 0.026, 95% CI 0.002 to 0.050, p = 0.037; Z-score: B = 0.224, 95% CI 0.035 to 0.412, p = 0.020; and T-score: B = 0.193, 95% CI 0.003 to 0.382, p = 0.046). Current or past use of a prolactin-raising antipsychotic medication was not associated with bone mineral density changes. In this small longitudinal study, there was no evidence of ongoing estrogen deficiency in psychotic disorder as there was no excessive loss of bone mineral density over a 3-year period in patients using antipsychotic medication.