Background: Depressive symptoms are common in hospitalized older persons. However, their relation to longterm mortality is unclear because few studies have rigorously considered potential confounders of the relation between depression and mortality, such as comorbid illness, functional impairment, and cognitive impairment. Objective: To measure the association between depressive symptoms and long-term mortality in hospitalized older persons. Design: Prospective cohort study. Setting: General medical service of a teaching hospital. Patients: 573 patients 70 years of age or older. Measurements: Depressive symptoms (Geriatric Depression Scale score), severity of acute illness (Acute Physiology and Chronic Health Evaluation II score), burden of comorbid illness (Charlson comorbidity index score), physical function (a nurse assessed dependence in six activities of daily living), and cognitive function (modified Mini-Mental State Examination) were measured at hospital admission. Mortality over the 3 years after admission was determined from the National Death Index. Mortality rates among patients with six or more depressive symptoms were compared with those among patients with five or fewer symptoms. Results: The mean age of the patients was 80 years; 68% of patients were women. Patients with six or more depressive symptoms had greater comorbid illness, functional impairment, and cognitive impairment at admission than patients with fewer depressive symptoms. Three- year mortality was higher in patients with six or more depressive symptoms (56% compared with 40%; hazard ratio, 1.56 [95% Cl, 1.22 to 2.00]; P < 0.001). After adjustment for age, acute illness severity, comorbid illness, functional impairment, and cognitive impairment at the time of admission, patients with six or more depressive symptoms continued to have a higher mortality rate during the 3 years after admission (hazard ratio, 1.34 [Cl, 1.03 to 1.73]). Although depressive symptoms contributed less to the mortality rate than did the total burden of comorbid medical illnesses, the excess mortality rate associated with depressive symptoms was greater than that conferred by one additional comorbid medical condition. Conclusions: Depressive symptoms are associated with long-term mortality in older patients hospitalized with medical illnesses. This association is not fully explained by greater levels of comorbid illness, functional impairment, and cognitive impairment in patients with more depressive symptoms.