A systematic review of long-term studies of drug treated and non-drug treated depression.

BACKGROUND The evidence base for long term antidepressant pharmacotherapy is limited. This review describes long-term outcomes of antidepressant-treated depression during the modern pharmacological era (post-1988). METHODS Review of Medline, PsycInfo, EMBASE, and the Cochrane Library using keywords and controlled vocabulary terms for long-term outcome studies (10 years or more) of antidepressant-treated cohorts (with at least 1 follow-up assessment post 1988) and non-drug treated cohorts (no time frame). Manual search of references of included articles. RESULTS Searches resulted in 24 publications on 12 naturalistic studies of antidepressant-treated cohorts (n=3901 at final follow-up). Most participants were White females with one inpatient stay. Recurrence of depression ranged from 40% to 85%. About one-quarter of individuals achieved a rating of well or improved, as many as experienced systematically poor outcomes. Frequency, duration, and severity of episodes varied substantially. Functional and social outcomes were infrequently measured. Few studies adequately assessed treatments. Outcomes for 3 non-drug treated samples with 1160 patients were no worse than outcomes achieved in antidepressant-treated cohorts. LIMITATIONS Heterogeneity of study designs and outcome definitions, and crude measures of pharmacotherapy precluded a statistical analysis. CONCLUSIONS Long-term outcomes in depression appear generally poor. No clear relationship emerged between drug treatment and positive outcomes. A lack of evidence from these long-term naturalistic studies of depression precludes any recommendation for maintenance antidepressant pharmacotherapy. Given its public health significance, more data is needed to address this area.