Depression increases the risk of mortality in patients with heart failure: A meta-analysis

- An updated meta-analysis of depression and mortality in HF is presented.
- Depression is a predictor of all-cause mortality in HF.
- Stronger effects were found in samples of older adults and with shorter follow-up.
- Future work should continue to explore depression and cardiac mortality.

BackgroundDepression is a risk factor for mortality in cardiovascular diseases. Prior studies confirm that depression predicts adverse outcomes in patients with heart failure (HF). However, data were inconclusive regarding the effect of depression on mortality. This meta-analysis examines the relationship between depression and mortality in HF.MethodsProspective studies of depression and mortality in HF published between 1999 and April 2016 were located using PubMed, PsychINFO, and MEDLINE. Comprehensive Meta-Analysis software was used to compute an aggregated effect size estimates of hazard ratios and to conduct subgroup analyses.ResultsEighteen studies met inclusion criteria. For 8 aggregated univariate and 14 multivariate estimates, depressive symptoms were related to all-cause mortality. A pooled HR of 3 multivariate analyses indicated that depressive symptoms were not linked to cardiovascular mortality. In subgroup analyses, depression predicted all-cause mortality in samples with a mean age > 65. The impact of depression on all-cause mortality also differed by follow-up duration, with samples with shorter follow-up durations demonstrating a larger effect.ConclusionsIn HF, depression is related to increased all-cause mortality risk, with stronger effects in samples with shorter follow-up and in older adults. In older adults, depression may serve as a marker of more severe HF. However, this possibility is difficult to examine given inconsistent adjustment for HF severity. Additional studies may assist in determining the relationship between depression and cardiovascular mortality, as the low number of studies examining cardiovascular mortality may have precluded detection of an effect.