The postulated therapeutic activity of nortriptyline metabolites has prompted investigation of dosage adjustments based on plasma levels of nortriptyline (NT) and its metabolites. The method assumes that metabolite concentrations vary independently of nortriptyline concentrations among patients. This study tests that assumption and investigates different means of obtaining metabolite concentrations. Forty-two psychiatric inpatients were divided into three maintenance dose groups: 50, 100, and 150 mg NT/day. After a 28-day study for each inpatient, steady-state plasma concentrations for days 14, 18, 21, 25, and 28 were determined. Concentrations were averaged for each patient. Nortriptyline concentrations did not correlate well with corresponding 10-OH(E)nortriptyline (p greater than 0.05) or 10-OH(Z)nortriptyline concentrations (r2 = 0.31, p less than 0.05). Concentrations of 10-OH(E)nortriptyline did not correlate well with corresponding 10-OH(Z)nortriptyline concentrations (r2 = 0.236, p less than 0.05). Neither dose/body weight nor obesity were good predictors of individual concentrations of nortriptyline, 10-OH(E)nortriptyline, or 10-OH(Z)nortriptyline or even the sum of the three. In conclusion, optimal drug therapy may involve dosage adjustments according to the combined plasma levels of nortriptyline and metabolites. Assurance of obtaining certain plasma concentrations requires plasma level monitoring.