Little information exists on the epidemiology of central nervous system side effects in patients taking antihypertensive medications. We examined prevalence rates of tricyclic antidepressant (TCA) use among a random sample (N=143,253) of Medicaid recipients. The TCA use was compared for patients taking any of seven antihypertensive agents and for those prescribed insulin or oral hypoglycemic agents. Use of TCA was significantly higher in patients taking β-blockers (23% over two years) than for patients taking hydralazine or hypoglycemics (both 15%) or methyldopa or reserpine (both 10%). Prevalence rate ratios revealed a risk of being prescribed a TCA of 1.5 (95% confidence interval, 1.4 to 1.7) for patients receiving β-blockers relative to patients receiving hydralazine or hypoglycemics. β-Blocker use may be an important cause of iatrogenic depression among hypertensive patients.