October 2009
Reviewer: Jeff C. Huffman, MD
Generalized anxiety disorder, major depressive disorder, and their comorbidity as predictors of all-cause and cardiovascular mortality: the Vietnam Experience Study
Background: Depression has been independently associated with cardiovascular mortality. In contrast, there has been much less study of the impact of anxiety, especially formally-diagnosed generalized anxiety disorder (GAD), on cardiac outcomes.
Methods: Participants in this retrospective study were identified from data collected through the Vietnam Experience Study, a government-commissioned study of male Vietnam veterans (1965-1971). Subjects in the study had been randomly sampled from the population of Vietnam veterans during that time period, interviewed by phone (in 1985), and participated in a medical exam in 1986. During the medical examination, patients underwent the Diagnostic Interview Schedule (DIS) for DSM-III diagnoses, including evaluations for major depression (MDD) and GAD, for disorders over the preceding 12 months. Primary outcomes were all-cause and cardiovascular mortality by the end of the year 2000. Data analysis accounted for multiple demographic, psychosocial, and medical covariates, along with alcohol use.
Results: Of the 4256 participants (mean age 38.3 at time of medical exam), 6.5% met criteria for MDD and 9.7% for GAD; 3.6% of patients had both conditions. There were 236 deaths over the follow-up period (mean follow-up time= 14.6 years), 63 of which were due to ‘major cardiovascular diseases’.
Both MDD (hazard ratio [HR] 1.6; p=0.02) and GAD (HR 1.7; p=0.002) were both associated with all-cause mortality, independent of the other covariables. Regarding cardiovascular mortality, GAD and MDD were associated with mortality on univariable analysis, but when other variables were considered, the link between GAD and mortality was reduced to a trend (HR 1.8; p=0.06) and the MDD/mortality link was nonsignficant (p=0.30). However, among patients who had both MDD and GAD, having these comorbid conditions was strongly and independently associated with both all-cause and cardiovascular mortality (p<0.001 in both cases). When PTSD (present in 313 patients in the cohort) was also considered, comorbidity of all three conditions led to an even greater risk of death.
Discussion: This study is an important contribution, suggesting that GAD, both alone and comorbid with MDD, is a real risk factor for mortality, at least among male Vietnam veterans. The study was limited by several factors: it included a sample that represented a small proportion of the veteran population (and one willing to undergo phone interview and medical exam) and consisted of only men. Furthermore, there was a relatively limited number of cardiovascular deaths, considering the number of variables considered (risking overfitting of the statistical model). Finally, the prevalence of GAD in this study was substantially higher than that seen in the general population and higher than the prevalence of MDD; these are both unusual findings that may also being into question the complete generalizability of the findings. This is especially true given prior studies finding no GAD/mortality connection.
Overall, however, this is a powerful study of thousands of patients followed for almost 15 years, and it raises a substantial question: what if prior studies linking depression and negative cardiac outcomes had also assessed subjects for GAD? In this cohort, more than half of MDD patients had GAD, and so perhaps the prior studies linking MDD and negative cardiac outcomes were actually a result of the connection between GAD and the negative outcomes, at least in part. In any event, it appears that further study of the GAD/mortality question is warranted, and that patients with comorbid depression and anxiety may be at the greatest risk of all.