HEART RATE VARIABILITY AND DEPRESSION IN CHD
Some studies have assessed HRV and depression following acute MI,41–45 whereas others have focused on HRV in medically stable patients with CHD. 46–49 Most of the studies have used frequency domain indices to calculate HRV.
HRV in post-MI patients with depression
In the largest study of depressed post-MI patients published to date,41 24-hour HRV levels were compared between 380 patients with a recent MI who had either major or minor depression and 425 post-MI patients who were not depressed. In univariate analyses, the four frequency domain indices of HRV (ULF, VLF, LF, and HF) were significantly lower in the depressed than in the nondepressed patients. After adjustment for possible confounders, all the indices except HF remained significantly lower in depressed patients than in nondepressed patients.
HRV in depressed patients with stable coronary disease
Most46–48 but not all49 studies have also found HRV to be lower in depressed than in nondepressed patients with stable CHD. The one exception was reported by Gehi et al,49 who assessed participants from the Heart and Soul Study cohort who had stable CHD at the time HRV was determined. Of the 873 outpatients with stable CHD who received 24-hour ambulatory ECG monitoring, 195 were found to have major depression. No differences between depressed and nondepressed patients were found on any time domain or frequency domain measure of HRV. This is the largest study to date of medically stable CHD patients assessed for depression and HRV, but its results differ from those of most smaller studies. The authors noted that although there was no difference in HRV between depressed and nondepressed patients, HRV in the nondepressed patients was similar to that in depressed patients in other samples.50 They speculated that the participants in their study, who were largely recruited from a Veterans Affairs hospital, may have been sicker than most participants in other studies and that this might have obscured depression-related differences in HRV.50
What is the clinical significance of HRV differences?
When evaluating differences in HRV between depressed and nondepressed patients, it is important to look past statistical comparisons and consider the clinical significance of these differences—ie, whether they are large enough to affect clinical outcomes or to be responsible for the depressed patients’ increased risk of death.
In the Cardiac Arrhythmia Pilot Study, HRV was assessed 1 year after acute MI in 331 patients.51 All measured indices of HRV were strong predictors of mortality. Patients with VLF power of less than 600 ms2 (natural logarithm of VLF power [LnVLF] 51 In a study of a similar group of medically stable (ie, eventfree for ≥6 months) patients with CHD,48 47% of those who were moderately to severely depressed, 29% of those who were mildly depressed, and 13% of those who were not depressed had VLF power below this cutpoint.
In the Multicenter Post-Infarction Program study, which evaluated patients in the immediate post–acute MI period, an LnVLF less than 5.2 was associated with a relative risk of 4.7 for cardiac mortality over the next 2.5 years.37 In our own study of post-MI patients, 7% of the nondepressed participants and 16% of the depressed participants had VLF power below this value, a difference that was significant even after adjusting for covariates (P = .006).41 Thus, mean 24-hour HRV is low enough in depressed patients with medically stable CHD and in those with recent acute MI to have prognostic significance.
How much of depression’s effect is due to low HRV?
In an attempt to determine whether low HRV accounts for at least part of the effect of depression on mortality, a statistical mediation model was applied to data collected in a follow-up study of the 311 depressed patients with recent acute MI described above,41 who were enrolled in the Enhancing Recovery in Coronary Heart Disease (ENRICHD) trial,52 and 367 patients who met the ENRICHD medical inclusion criteria but were without depression.53 VLF was selected as the index of HRV for this study because of its prognostic importance in post-MI patients. As noted earlier, VLF was significantly lower in the depressed patients.41 During a median follow-up of 24 months, there were 47 deaths within the overall study population of 766 patients (6.1%).53 Consistent with earlier studies, the depressed patients were at higher risk for all-cause mortality, even after adjusting for potential confounders (hazard ratio = 2.8; 95% confidence interval [CI], 1.4 to 5.4; P P = .03), indicating that the LnVLF accounted for about one-quarter of the total mortality risk. Thus, the study results suggest that low HRV at least partially mediates the effect of depression on survival after acute MI.
A role for premature ventricular contractions
In one of the first prognostic studies of depression following acute MI, Frasure-Smith et al reported an interaction between depression and premature ventricular contractions (VPCs) on subsequent mortality.6 Specifically, they found that depressed patients who had 10 or more VPCs per hour after an MI were at considerably higher risk of death than were either depressed post-MI patients without VPCs or nondepressed post-MI patients with 10 or more VPCs per hour. One interpretation of these data is that depressed patients may be at greater risk for death due to an abnormal response to VPCs or other arrhythmias.
HR turbulence analysis is a method for quantifying HR response to VPCs. In most individuals, when a VPC occurs, HR first accelerates and then decelerates. HR responses that differ from this pattern have been found to be even better predictors of post-MI mortality than more traditional measures of HRV in these patients.54,55
A total of 498 patients from the study reported above53 were found to have VPCs during 24-hour ambulatory monitoring.34 Of these patients, 260 had normal HR turbulence, 152 had equivocal HR turbulence, and 86 had abnormal HR turbulence. The depressed patients were more likely than their nondepressed counterparts to have abnormal HR turbulence (risk factor–adjusted odds ratio [OR] = 1.8; 95% CI, 1.0 to 3.0; P = .03). The patients were followed for a median of 24 months. Consistent with earlier studies, depressed patients had worse survival (OR for death = 2.4; 95% CI, 1.2 to 4.6; P = .02) than the nondepressed patients. When HR turbulence was added to the statistical model, the adjusted hazard ratio for depression decreased to 1.9 (95% CI, 0.9 to 3.8; P = .08). When the LnVLF was added to this model, the adjusted hazard ratio decreased further, to 1.6 (95% CI, 0.8 to 3.4; P = .18). Thus, the combination of VLF and HR response to VPCs explained about half of the effect of depression on survival in these patients.
Causality not proven, but further study warranted
Obviously, these results do not prove that there is a causal relationship between depression, low HRV, and mortality. However, they are consistent with the interpretation that HRV, especially when combined with measures of HR response to VPCs, may account for a significant proportion of depression’s association with mortality following an MI. Future studies of these risk markers should explore their potential interrelationships to clarify how they may jointly contribute to the risk of death in patients with depression.