In order to evaluate the association of long-term mortality with antidepressant use and severity of depression, assessed using the Beck Depression Inventory (BDI), O’Connor and colleagues studied 1,006 patients, aged 18 years or older who were admitted to hospital with clinical (NYHA ≥ II) and/or echocardiographic evidence of systolic heart failure (left ventricular ejection fraction (LVEF) of ≤ 35%).
On admission severity of depression was assessed using the BDI questionnaire. Use of antidepressant pharmacotherapy was determined from inpatient pharmacy records. If patients were found to be depressed, and were not already prescribed antidepressant therapy, intervention for their depressive symptoms was deferred to the primary care provider. Patients were divided in four groups; (i) no antidepressant, (ii) SSRI-only, (iii) TCA ± SSRI, (iv) other antidepressants ± TCA ± SSRI, and followed up for vital status over a mean of 971 (± 730) days.
During the index hospital admission, 162 patients (16.1%) were taking antidepressant therapy, whereas 302 patients (30%) were considered to have significant depression. Of these depressed patients, 24.5% were taking antidepressant drugs in comparison to 12.5% of non-depressed patients.
A total of 429 patients died; 161 (53.3%) with depression versus 268 (38.1%) without, p < 0.001. Higher rates of death were observed in patients taking antidepressants (76 deaths among 162 patients, 46.9%) compared to those who were not (353 deaths among 843 patients, 41.9%). However, after adjustment for depression neither antidepressant therapy nor use of SSRIs only was associated with reduced survival (antidepressants: HR, 1.19; 95% CI, 0.84-1.71; p = ns, SSRIs only: HR, 1.06; 95% CI, 0.69-1.62; p = ns), whereas depression remained associated with reduced survival (HR, 1.39; 95% CI, 1.12-1.74; p = 0.003).
These data suggest that depression, rather than antidepressant use, is independently associated with increased long-term mortality. However, this study has several limitations. Depression assessment and information regarding antidepressant use was only available during the index hospital admission. No data are available regarding compliance, nor the prescription of antidepressant drugs following hospital discharge. Finally, this study was of observational design and therefore inferences regarding causality are difficult. Although further randomised controlled trials are required to better evaluate the safety of antidepressant drugs the present study contributes to the growing body of data supporting the hypothesis that depression remains a significant predictor of mortality among patients with heart disease.
- O’Connor C, Jiang W, Kuchibhatla M et al. Antidepressant use, depression, and survival in patients with heart failure. Arch Intern Med. 2008;168(20):2232-2237