Author + information
- Peter de Jonge, PhD⁎ (, )
- Rob H.S. van den Brink, PhD,
- Titia A. Spijkerman, MD and
- Johan Ormel, PhD
- ↵⁎Department of Psychiatry, University Medical Center Groningen, P.O. Box 30.001, 9700 RB Groningen, the Netherlands
In reply to the letter by Dr. Ziegelstein and colleagues, we found that incident but not nonincident postmyocardial infarction (MI) depression was associated with new cardiovascular events and that these effects were not confounded by MI severity (1). Dr. Ziegelstein and colleagues suggested that perhaps an interaction between depression and MI severity might explain our findings. This is an intriguing question that we are happy to consider.
Using 2 indicators for MI severity, left ventricular ejection fraction (LVEF) <40% and Killip class >1, we assessed whether combinations of dysfunctions of the brain and the heart predict new cardiovascular events. We found that higher Killip class and lesser LVEF alone were associated with an increased risk of cardiovascular events, relative to patients with unimpaired cardiac status and no depression (hazard ratio [HR]Killip >1= 2.19; 95% confidence interval [CI] 1.29 to 3.73; HRLVEF <40%= 1.73; 95% CI 1.07 to 2.81). Also, depression alone resulted in a somewhat increased risk (HRDepression= 1.53; 95% CI 0.96 to 2.42). The highest cardiac risk appeared when both were present versus none: HRKillip >1 + depression= 2.41, 95% CI 1.10 to 5.31; HRLVEF <40% + depression= 2.36, 95% CI 1.19 to 4.67).
We then assessed whether these combined effects were restricted to incident post-MI depressions. A combination of impaired cardiac status and nonincidentdepression was not associated with an increased risk (HRLVEF <40% + nonincident= 0.76, 95% CI 0.10 to 5.33; HRKillip >1 + nonincident= 0.94, 95% CI 0.13 to 6.83). However, the combination of incidentpost-MI depression and impaired cardiac status was cardiotoxic (HRLVEF <40% + incident= 3.13, 95% CI 1.54 to 6.39; HRKillip >1 + incident= 3.26, 95% CI 1.40 to 7.59). Formal tests of interaction effects in Cox regression analyses resulted in no significant interactions.
We thus replicated and extended the findings presented by Dr. Ziegelstein and colleagues. What do they mean in terms of interactions between the heart and the brain? We believe they suggest the brain and the heart exert additiveeffects on cardiovascular prognosis in patients following an MI, suggesting biologic parallelism rather than synergism or interaction (2). An interaction between the heart and the brain would mean that, apart from the effects of the heart and the brain themselves, the combination of depression and cardiac dysfunction would produce an extra effect. This is not the case, however. As the HRKillip >1= 2.2 and the HRincident= 1.7 (1), the expected HRKillip >1 + incidentgiven no interaction effect would be (2.2 + 1.7–1) = 2.9. The observed HRKillip >1 + incident= 3.3 and its 95% CI includes 2.9. The same pattern emerges for all other interaction effects.
Thus, we believe the heart and the brain exert additive effects on cardiovascular prognosis, and these effects appear to be restricted to incident post-MI depression. Dysfunctions of the heart and the brain both have a negative impact on cardiovascular prognosis; as a result, their combination is notably cardiotoxic.
- American College of Cardiology Foundation