Objectives: To assess changes induced by lithium maintenance therapy on the incidence of thyroid, parathyroid and ion alterations. These were evaluated with respect to the duration of lithium therapy, age, sex, and family history (whether or not the patient had a first-degree relative with thyroid disease). Design: Prospective study. Setting: Affective Disorders Clinic at St Mary's Hospital, Montreal. Patients: One hundred and one patients (28 men and 73 women) with bipolar disorder receiving lithium maintenance therapy ranging from 1 year's to 32 years' duration. The control group consisted of 82 patients with no psychiatric or endocrinological diagnoses from the hospital's out-patient clinics. Outcome measures: Laboratory analyses of calcium, magnesium and thyroid-stimulating hormone levels performed before beginning lithium therapy and at biannual follow-up. Results: Hypothyroidism developed in 40 patients, excluding 8 patients who were hypothyroid at baseline. All patients having first-degree relatives affected by thyroid illness had accelerated onset of hypothyroidism (3.7 years after onset of lithium therapy) compared with patients without a family history (8.6 years after onset of lithium therapy). Women over 60 years of age were more often affected by hypothyroidism than women under 60 years of age (34.6% versus 31.9%). Magnesium levels in patients on lithium treatment were unchanged from baseline levels. After lithium treatment, calcium levels were higher than either baseline levels or control levels. Thus, lithium treatment counteracted the decrease in plasma calcium levels associated with aging. Conclusions: Familiar thyroid illness is a risk factor for hypothyroidism and hypercalcemia during iithium therapy.