Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
Contents lists available at ScienceDirect
Progress in Neuro-Psychopharmacology & Biological Psychiatry
Reactive heart rate variability in male patients with first-episode major depressive disorder☆ Chih-Sung Liang a,b, Jia-Fu Lee c,d,⁎, Chia-Chi Chen e, Yue-Cune Chang f a
Graduate Institute of Medical Sciences, National Defense Medical Center, Taipei, Taiwan, ROC Department of Psychiatry, Beitou Branch, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan, ROC Department of Psychiatry, Taipei Tzu Chi General Hospital, Buddhist Tzu Chi Medical Foundation, Taipei, Taiwan, ROC d School of Medicine, Tzu Chi University, Hualien, Taiwan, ROC e Department of Nursing, Kang-Ning Junior College of Medical Care and Management, Taipei, Taiwan, ROC f Department of Mathematics, Tamkang University, Taipei, Taiwan, ROC b c
a r t i c l e
i n f o
Article history: Received 22 June 2014 Received in revised form 26 July 2014 Accepted 8 August 2014 Available online 19 August 2014 Keywords: Cardiovascular reactivity First episode Heart rate variability Major depressive disorder Reverse trends Stress
a b s t r a c t Objective: The association between cardiovascular reactivity and major depressive disorder (MDD) remains unclear. This study aimed to examine this association via reactive heart rate variability (HRV) in a welldiagnosed first-episode MDD group and a control group. Methods: A total of 160 physically healthy, drug-naive patients presenting with their first-episode MDD and 50 healthy controls were recruited. All participants underwent a 5-min electrocardiography at rest and during a mental arithmetic task. Depression severity was assessed using the Beck Depression Inventory II (BDI). Results: HRV measures that showed between-group differences at rest did not reached significance during mental stress. In contrast, HRV measures that revealed between-group differences during stress did not reach significance at rest. In response to mental stress, HRV measures did not significantly change in both group. However, LF and HF in response to stress were different between groups. Patients with MDD revealed an increasing trend in HF and a decreasing trend in LF; conversely, healthy controls had a decreasing trend in HF and an increasing trend in LF. BDI scores correlated with changes in heart rate in the control group. Conclusions: The fundamental change to reactive HRV in patients with first-episode MDD appears qualitative, not quantitative. A distinctly reverse trend in reactive HRV measures were evident between these two groups. Moreover, patients with MDD showed entirely distinct changes in reactive HRV from those in resting HRV. We suggest that in patients with MDD, autonomic system shifts to sympathetic dominance at rest but toward parasympathetic dominance in response to stress. © 2014 Elsevier Inc. All rights reserved.
1. Introduction Understanding the cardiovagal function in depression is of great significance to public health since a large body of evidence has indicated
Abbreviations: MDD, major depressive disorder; HRV, heart rate variability; CVD, cardiovascular diseases; CVR, cardiovascular reactivity; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders Fourth Edition; BDI, Beck Depression Inventory II; ECG, electrocardiogram; LF, low frequency; HF, high frequency; TP, total power; VAR, variances of the RR interval; SDNN, the standard deviation of all normal-to-normal interbeat interval; MHR, the mean value of heart rate; MRR, the mean value of the RR interval; SD, standard deviation; SE, standard error; CI, confidence interval. ☆ The article has not been published previously, that it is not under consideration for publication elsewhere, that its publication is approved by all authors and tacitly or explicitly by the responsible authorities where the work was carried out, and that, if accepted, it will not be published elsewhere including electronically in the same form, in English or in any other language, without the written consent of the copyright-holder. ⁎ Corresponding author at: Department of Psychiatry, Taipei Tzu General Chi Hospital, Buddhist Tzu Chi Medical Foundation, Taipei, Taiwan, ROC, No.289, Jianguo Rd., Xindian Dist., New Taipei City 23142, Taiwan, ROC. Tel.: +886 2 6628 9779; fax: +886 2 2794 9366.
http://dx.doi.org/10.1016/j.pnpbp.2014.08.004 0278-5846/© 2014 Elsevier Inc. All rights reserved.
that depression increases the risk of cardiovascular diseases (CVD) and its related mortality (Frasure-Smith et al., 2009; Van der Kooy et al., 2007). Cardiovascular reactivity (CVR) is the responsiveness of the cardiovascular system to stress, and reactivity is derived as changes from baseline levels not absolute levels during stress. CVR reflects autonomic regulatory capacity, and evidence suggests that CVR predicts the development of preclinical and clinical CVD states (Treiber et al., 2003). Research on the association between CVR and depression has yield inconsistent results in recent years. A meta-analysis published in 2004, which consisted of 11 relevant studies, suggested depression involving enhanced CVR (Kibler and Ma, 2004); however, the negative association between attenuated CVR and depressive symptoms has been repeatedly observed in non-clinical participants (Schwerdtfeger and Rosenkaimer, 2011), in people with coronary artery disease (York et al., 2007), in large-scale community samples (Carroll et al., 2008; de Rooij et al., 2010; Phillips, 2011), and in patients with major depressive disorder (MDD) (Rottenberg et al., 2007; Salomon et al., 2009, 2013).
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
Heart rate variability (HRV) is a widely used tool to assess cardiac autonomic activity, and it can reflect the balance between the sympathetic and the parasympathetic regulatory control of heartbeat. HRV has been proposed as a more appropriate measure to examine the association between CVR and depression (Schwerdtfeger and Rosenkaimer, 2011). Although strong evidence from the meta-analysis suggests that depression without CVD is associated with attenuated HRV (Kemp et al., 2010), few studies have investigated HRV in response to stress (also known as reactive HRV or HRV reactivity) in depressed individuals. A study, which enrolled a small sample with non-clinical depression and coronary artery disease, showed that during mental stress, the increases in low-to-high frequency (LF/HF) ratio and heart rate were more pronounced in participants with higher depression scores (Sheffield et al., 1998). Another study with 53 otherwise healthy participants reported greater reduction in high frequency (HF) in the highly depressed mood group (Hughes and Stoney, 2000). The authors suggest that the magnitude of decrease in parasympathetic cardiac control during stress was associated with depressed mood (Hughes and Stoney, 2000). Nevertheless, a recent study including 22 drug-naive, clinical depressed patients produced different results (Shinba, 2013). In this study, patients with clinical depression revealed an increase in reactive HF and decreases in reactive LF/HF ratio and heart rate. Given that during the task performance, reactive HRV for clinical depressed patients were different from those of healthy controls, the authors suggest an altered state of autonomic reactivity in depression (Shinba, 2013). The methodological differences, clinical heterogeneity of subjects, and differences in medication use may have led to the inconsistent findings. Moreover, the sample size in prior studies was relatively small, probably skewing the results. The aim of the study was to explore the association between CVR and MDD by analyzing reactive HRV. We carefully enrolled only drug-naive, CVD-free, male participants diagnosed with their first episode of MDD. The study design could provide a better window into CVR in the early stage of MDD. We also examined whether changes in reactive HRV were different from changes in resting HRV and how depression severity was associated with changes in reactive HRV.
2. Methods 2.1. Participants Participants were recruited from the inpatient division of Beitou Branch, Tri-Service General Hospital, National Defense Medical Center, a psychiatric teaching hospital in Taiwan. Written informed consent was obtained in accordance with the National Health and Medical Research Council guidelines. All subjects were fully informed regarding the aims and details of the study and were free to withdraw their consent at any time. Between 2009 and 2012, a total of 160 individuals with a first episode of MDD were judged eligible. To avoid the potential confounders, such as age, gender, ethnicity, use of psychotropics, and CVD (Liao et al., 1995), individuals were required to be unrelated Han Chinese males, between the ages of 20 and 40 years, drug-naive, and in good health. The diagnosis of MDD—meeting the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria—was made by two certified psychiatrists by diagnostic interview. The severity of depression was assessed with the Chinese version of the Beck Depression Inventory II (BDI) (Beck et al., 1961). Individuals were excluded from the study for any of the following reasons: systolic blood pressure exceeding 180 mmHg or diastolic blood pressure exceeding 110 mmHg; histories of or concurrent medical illnesses such as heart disease, diabetes, CVD, liver or renal disease, endocrinopathies, neurological disorders, or cancer; meeting the DSM-IV criteria for substance-related disorders, psychotic disorders,
53
or bipolar disorder; use of any drugs or over-the-counter preparations; and being shift workers. A series of examinations were performed, including a complete physical examination, a routine biochemical panel, a complete blood count, a urine test, a stool study, a chest X-ray, and an electrocardiogram (ECG). Height (cm) and weight (kg) of each participant were measured by a standard balance beam scale, and body mass index (BMI) was calculated. The medical charts were systematically reviewed in order to confirm that all participants met the above operational criteria. The control group consisted of 50 healthy, never-depressed volunteers matched for sex, age, smoking status, and time of day of ECG recordings. Controls also signed an informed consent form, were interviewed, and completed a BDI rating. Two certified psychiatrists confirmed that none of the healthy controls had a history of any psychiatric illness or were taking any medications. The Institutional Review Board for the Protection of Human Subjects at the Tri-Service General Hospital, a medical teaching hospital within the National Defense Medical Center in Taiwan, approved the protocol. 2.2. HRV recording procedures Detailed procedures for HRV analysis have been previously reported and are only briefly summarized here (Kuo et al., 1999). Prior to the measurements, participant were instructed to refrain from alcohol and caffeine consumption for 24 h and from smoking, exercising, and heavy eating for at least 8 h. All participants had a usual breakfast on the study day to ensure standardization of ECG recordings. After the participants had rested for 15 m in a quiet, air-conditioned room, the ECG for the analysis of beat-to-beat HRV was recorded under standardized conditions. They were asked to relax, breath naturally, and move as little as possible. Experimental procedures were performed between 08.30 h and 11.30 h in a fixed order of rest followed by an arithmetic task. The mental arithmetic task consisted of a serial subtraction by 7, beginning with the number 2193. During the arithmetic task, participants were prompted to perform faster because they were going “slowly,” and this feedback was for all participants. Mental arithmetic tasks have been recognized as active, self-relevant stressors (Schwerdtfeger and Rosenkaimer, 2011). A 288-s signal sequence of the telemetrically transmitted lead I ECG was recorded. The raw ECG signals were amplified with a gain of 1000 and band-pass filtered (0.68–16 Hz). Signals were then recorded with an 8-bit analog-to-digital converter with a sampling rate of 256 Hz. The digitized ECG signals were analyzed online and simultaneously stored on removable hard disks for offline verification. Signal acquisition and data storage and processing were performed via a generalpurpose personal computer. Our computer algorithm identified each QRS complex and rejected each ventricular premature complex or noise according to its likelihood in a standard QRS template. Stationary RR values were re-sampled and interpolated at a rate of 7.11 Hz to produce continuity in the time domain. Frequency-domain analysis was performed by a nonparametric method of fast Fourier transformation. The direct current component was deleted, and a Hamming window was used to attenuate the leakage effect. For each time segment (288-second; 2048 data points), our algorithm estimated the power spectrum density based on the Fourier transformation. The resulting power spectrum was corrected for attenuation resulting from the sampling and the Hamming window. The power spectrum was subsequently quantified into standard frequency-domain measurements including low frequency (LF) (0.04–0.15 Hz), HF (0.15–0.40 Hz), and total power (TP). The timedomain parameters were the standard deviation of all normal-tonormal interbeat interval (SDNN) and the variance of RR interval (VAR). Both parameters were measured in milliseconds (ms). Although mean value of HR (MHR) and mean value of RR interval (MRR) during the 5 min did not belong to HRV components, these two measure were also calculated under both conditions. The spectral components
54
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
of HRV were analyzed as absolute units (ms2). All HRV analyses were performed by a trained research nurse blinded to individual depression severity and the protocol. Individuals were excluded from further analysis if they were not in a predominantly regular sinus rhythm or if they had sustained atrial arrhythmias, such as atrial fibrillation or greater than 5% ectopic complexes. 2.3. Statistical analysis Independent-sample t-tests examined group differences in clinical and demographic characteristics. A multivariate general linear model was performed to compare group differences in resting HRV measures, MHR, and MRR, with age and BMI as covariates. Logarithmic transformations were applied to correct the skewed distributions of LF, HF, TP, LF/HF ratio, VAR, SDNN, MHR, and MRR. Group differences in reactive HRV measures, MHR, and MRR were analyzed with a multivariate general linear model for repeatedmeasures data. Time was entered as a 2-level within-subject factor, and group was entered as a between-subject factor. The model used a full factorial design and included covariates of BMI and age. Healthy controls was the reference group. The model can examine the interaction effect of group × time. Within-group reactivity in HRV measures, MHR, and MRR during mental stress were examined as well. Spearman correlations examined whether BDI scores correlated with reactivity in HRV measures, MHR, and MRR. Reactivity measures were calculated by subtracting baseline values from during-task values. All analyses were performed using the Statistical Package for the Social Sciences for Windows, Version 19.0 (SPSS Inc., Chicago, IL, USA), and effect sizes are reported as partial eta-squared (η2p). Analyses were considered to be statistically significant if its p-values were less than or equal to 0.05 (two-tailed). 3. Results Data analysis consisted of 156 patients with first episode of MDD and 49 healthy controls. One healthy control and four participants with MDD had to be excluded because they failed to complete HRV protocol. The mean age in all participants were 24.4 ± 2.7 years. Table 1 includes data on demographic and clinical characteristics. The MDD group did not differ from the control group in age, height, and BMI, although they were shorter (t = −2.823, p = 0.006). Moreover, the MDD group had higher BDI scores (t = − 2.823, p b 0.001) and faster heart rate (t = 5.094, p b 0.001). Group differences in resting HRV measures, MHR, and MRR are shown in Table 2. Under resting condition, the MDD group showed lower values in TP (p = 0.002), VAR (p b 0.001), SDNN (p b 0.001), and MRR (p b 0.001), but they had higher LF/HF ratio (p = 0.024) and faster MHR (p b 0.001). Their resting LF and HF did not differ from those of the control group. Fig. 1 illustrates group differences in reactive HRV measures, MHR, and MRR. The interaction effect of group × time reached statistical Table 1 Demographic comparisons between patients with major depressive disorder and healthy controls. MDD, N = 156
Age (year)⁎ BDI score⁎ Height (cm)⁎ Body weight (kg) Body mass index (kg/m2)⁎ Heart rate (beat/min)⁎
Controls, N = 49
t test
Mean
SD
Mean
SD
t
p
24.4 34.1 171.4 66.5 22.6 75.3
3 12.3 5.8 10.7 3.4 12.7
24.7 4.9 173.7 69 22.8 66.4
1.5 3.5 4.6 8.8 2.3 10
−0.95 26.428 −2.823 −1.453 −0.425 5.094
0.344 b0.001 0.006 0.148 0.672 b0.001
Note. ⁎The Levene test for equality of variances was statistically significant. SD = standard deviation; MDD = major depressive disorder; BDI = Beck Depression Inventory II.
Table 2 Comparisons between patients with major depressive disorder (N = 156) and healthy controls (N = 49) in resting heart rate variability. Rest
B
SE
t
p
95% CI
LF (ms2) HF (ms2) LF/HF ratio TP (ms2) VAR (ms) SDNN (ms) MRR (ms) MHR (beat/min)
0.04 −0.13 0.161 −0.423 −0.499 −0.251 −0.123 0.122
0.051 0.86 0.132 0.138 0.136 0.068 0.027 0.027
0.792 −1.524 1.219 −3.064 −3.668 −3.684 −4.611 4.594
0.429 0.129 0.024 0.002 b0.001 b0.001 b0.001 b0.001
−0.06 −0.299 −0.1 −0.696 −0.768 −0.385 −0.175 0.07
η2p 0.141 0.038 0.422 −0.151 −0.231 −0.117 −0.07 0.175
0.003 0.011 0.007 0.045 0.063 0.063 0.096 0.095
Note. A multivariate general linear model was performed with covariates of age and body mass index. Healthy controls were the reference group. SE = standard error; CI = confidence interval; LF = low frequency; HF = high frequency; TP = total power; SDNN = the standard deviation of all normal-to-normal interbeat interval; VAR = variances of the RR interval; MRR = the mean value of the RR interval; MHR = the mean value of heart rate.
significance in LF (F = 4.068, p = 0.045, η2p = 0.02), HF (F = 5.425, p = 0.021, η2p = 0.026), LF/HF ratio (F = 5.018, p = 0.026, η2p = 0.024), MRR (F = 81.803, p b 0.001, η2p = 0.289), and MHR (F = 80.752, p b 0.001, η2p = 0.287). During the arithmetic task, both groups showed similar trends in TP, VAR, and SDNN, and the interaction effects in these measures were not significant. Notably, neither the MDD group (F = 0.578, p = 0.717, η2p = 0.019) nor the control group (F = 0.547, p = 0.769, η2p = 0.074) reached significance on analysis of within-group reactivity in HRV measures. The control group showed significant reactivity in MHR and MRR (both: F = 4.601, p = 0.015, η2p = 0.17). Nevertheless, within-group reactivity in MHR and MRR were not significant for the MDD group (both: F = 1.561, p = 0.213, η2p = 0.01). Because our data clearly showed bimodal distribution of scores on BDI, only within-group correlations were analyzed. BDI scores did not correlate with reactive HRV measures in both groups. In healthy controls, BDI scores negatively correlated with MHR (r = − 0.401, p = 0.004) but positively correlated with MRR (r = 0.313, p = 0.028); however, these two measures did not correlate with BDI scores in patients with MDD. Fig. 2 displays the correlation between BDI scores and reactive heart rate. 4. Discussion 4.1. Reactive HRV in first-episode MDD: neither enhanced nor attenuated Few studies have explored CVR in MDD via reactive HRV analysis. The importance of our study is that we compared reactive HRV between a well-diagnosed, first-episode MDD group and a healthy control group, and all participants were male, drug-naive, and screened for the absence of CVD. We found that both reactive HRV and resting HRV were different for patients with MDD and healthy controls, and patients with MDD showed entirely distinct changes in reactive HRV from those in resting HRV. Although at rest, the MDD group revealed lower HRV in TP, VAR, and SDNN, the interaction effects of group × time on these measure did not reach significance. Besides, under resting condition, the MDD group did not differ from the control group in LF and HF, but the interaction effects on these two measures reached significance. That is, patients with MDD revealed an increasing trend in HF and decreasing trends in LF and LF/HF ratio; conversely, healthy controls had a decreasing trend in HF and increasing trends in LF and LF/HF ratio. Considering the within-group reactivity, neither the MDD group nor the control group reached significance in HRV measures; nevertheless, the MDD group clearly showed similar but significantly attenuated trends in MHR and MRR compared with those of the control group. On the other hand, both groups did not have any association between reactive HRV and depression severity; nevertheless, in the
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
3.50
High Frequency
4.25
Low Frequency
Control Group
Control Group
MDD Group
MDD Group
3.25
4.00
3.00
3.75
Baseline
Stress
Baseline
Low Frequency / 1.25
55
8.25
Control Group MDD Group
1.00
Total Power
8.50
High Frequency Ratio
Stress
Control Group MDD Group
8.00 7.75 7.50
0.75
7.25 7.00
0.50
Baseline
7.00
Stress
Baseline
Stress
Standard Deviation of
Mean Value of RR Interval
RR Interval
4.50
Control Group MDD Group 6.75
Control Group
4.25
MDD Group 4.00
3.75
6.50
3.50
Baseline
8.50
Stress
Baseline
Variance of RR Interval
8.25
4.50
Stress
Mean Value of Heart Rate Control Group
Control Group
MDD Group
MDD Group
8.00
4.25
7.75
7.50
4.00
7.25
Baseline
Stress
Baseline
Stress
Fig. 1. Comparisons between patients with MDD and healthy controls in reactive HRV. The interaction effect of group × time reached significance in LF (F = 4.068, p = 0.045, η2p = 0.02), HF (F = 5.425, p = 0.021, η2p = 0.026), LF/HF ratio (F = 5.018, p = 0.026, η2p = 0.024), MRR (F = 81.803, p b 0.001, η2p = 0.289), and MHR (F = 80.752, p b 0.001, η2p = 0.287). The slope of the line representing reactive HF is positive in the MDD group, whereas it is negative in the control group. Conversely, the slope of the line representing LF and LF/HF ratio are negative in the MDD group, whereas they are positive in the control group. The slope of the line representing reactive MHR and MRR in the control group are steeper than those in the MDD group. MDD = major depressive disorder; HRV = heart rate variability; LF = low frequency; HF = high frequency; MRR = the mean value of the RR interval; MHR = the mean value of heart rate.
control group, BDI scores positively correlated with MRR and negatively correlated with MHR. In summary, in patients with first episode of MDD in response to mental stress, reactivity in HF and LF measures were
neither attenuated nor enhanced but found to show reverse trends from those of healthy controls. Furthermore, patients with MDD showed entirely distinct changes in reactive HRV from those in resting HRV.
56
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
in heart period are largely dependent on parasympathetic activity (Schwerdtfeger and Rosenkaimer, 2011). Our MDD group at rest revealed lower HRV, suggesting a shift of autonomic balance away from parasympathetic side. A previous study investigating HRV, QT variability, and baroreflex sensitivity reported that under resting condition, patients with MDD showed an overall shift of autonomic balance toward sympathetic dominance (Koschke et al., 2009). Indeed, faster heart rate was evident in our participants with MDD under resting condition. Taken these findings together, we suggest that in patients with MDD, autonomic system shifts to sympathetic dominance at rest but toward parasympathetic dominance in response to stress. 4.4. Within-group changes in HRV measures and heart rate
Fig. 2. The correlation between reactive heart rate (beat/min) and depression severity. In patients with major depressive disorder, BDI scores did not correlate with changes in heart rate (r = −0.056, p = 0.484). In healthy controls, BDI scores negatively correlated with changes in heart rate (r = −0.401, p = 0.004). BDI = Beck Depression Inventory II.
4.2. Non-clinical depression and dichotomization We provide evidence that the fundamental change to reactive HRV in first-episode MDD is qualitative, not quantitative. At rest, our MDD group had lower HRV, a finding that agreed with the well-established evidence from the meta-analysis (Kemp et al., 2010). During mental stress, their HRV relative to that of healthy controls is neither attenuated nor enhanced, but revealed totally distinct trends in LF and HF. Our results, in part, agreed with a prior studies using patients with clinical depression (Shinba, 2013). The author reported attenuated reactivity in heart rate and LF/HF ratio and increased reactivity in HF. However, two prior studies recruiting non-clinical subjects with depressed mood reported inconsistent findings on reactive HRV (Hughes and Stoney, 2000; Sheffield et al., 1998). The heterogeneity of subjects —patients with MDD versus non-clinical participants with depressed mood—may lead to the inconsistent findings. Moreover, the previous two studies categorized participants according to the median splits of depression scores (Hughes and Stoney, 2000; Sheffield et al., 1998). During data analysis, dichotomization of continuous variables have been strongly discouraged since it may increase the risk of a false positive result and reduce the power (Altman and Royston, 2006). Also, the categorized groups usually indicated that the depression psychopathology of these two groups were quite distinct from each other. In this regard, in non-clinical sample with depressed mood, dichotomizing participants according to depression scores might not be appropriate. Taken together, the qualitative, not quantitative, changes in LF and HF suggest that patients with first episode of MDD had qualitatively distinct changes in reactive HRV. 4.3. Toward parasympathetic dominance during mental stress Broad evidence supports that HF is a reliable measure of cardiac parasympathetic tone (Reyes del Paso et al., 2013). When confronted with a psychological stressor, our MDD group revealed an increase trend in parasympathetic activity, while the control group showed a reverse trend. A reciprocal relationship in LF and HF were evident for both groups. Therefore, mental stress appear to shift autonomic balance toward parasympathetic side in patients with MDD but away from parasympathetic dominance in healthy controls. Attention should be directed toward another issue—to which side autonomic balance tends to shift in patient with MDD under resting condition. It is conceivable that at rest, vagal tone prevails and variations
It is important to note that in response to mental stress, neither the MDD group nor the control group revealed significant reactivity in HRV measures. Moreover, reactivity in MHR and MRR were not significant for the MDD group but for the control group. This implies no quantitative change in HRV between groups when confronted with our arithmetic task. The attenuated reactivity in heart rate in the MDD group might be related to their autonomic shift toward parasympathetic dominance in response to stress. Moreover, in the MDD group, there might be a ceiling effect that could limit the extent of increase in heart rate, because evidence suggests that under resting condition, patients with MDD have shifted their autonomic balance toward sympathetic dominance (Koschke et al., 2009). From another physiological perspective, the attenuated reactivity in heart rate may be a compensatory mechanism from other stress systems (McEwen, 1998). For example, the hypothalamic–pituitary–adrenocortical (HPA) axis, a key component of the neuroendocrine system that controls reactions to stress, is well documented to be activated in MDD (Hatzinger et al., 2002). Thus, the attenuated reactivity in heart rate may simply reflect a compensation for the activation of the HPA axis. 4.5. Effect sizes According to the interpretive guideline for effect sizes by Cohen (Cohen, 1992), our study produced a small effect size (η2p b 0.04). This indicates that although the depressive symptoms in our patients with MDD were severe, as evidenced by a mean score of 34.1 on BDI, MDD state only explain a small variance of between-group differences in reactive HRV measures. Someone may treat the effects of MDD on reactive HRV as trivial, for effect sizes represents an index of clinical importance. It is, however, important to consider that our depressed patients were drug-naive and experiencing their first episode of MDD. Thus, the long-term effects of MDD on HRV may become more pronounced. It has been suggested that small effects in ongoing processes can accumulate to larger ones (Prentice and Miller, 1992). Thus, we expect that the effect sizes will be increased if these patients with MDD are followed prospectively. Because CVR can predict the development of preclinical and clinical CVD states (Treiber et al., 2003), and depression may increase the risk of CVD and its related mortality (Frasure-Smith et al., 2009; Van der Kooy et al., 2007), the clinical importance of an effect size can be a function of how mild an experimental manipulation is to still produce a difference. As a consequence, a small effect size still suggests a compromised autonomic regulation in patients with first episode of MDD. 4.6. Limitations First, available data, however, have challenged the view that LF component indicates sympathetic cardiac control (Reyes del Paso et al., 2013). It is suggested that both LF and HF components are mainly determined by the parasympathetic system but provide different information about autonomic regulation. That is, HF relates to respiratory influences
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
and LF reflects the modulation of vasomotor tone. Therefore, caution must be exercised when interpreting our results since we did not assess respiration and blood pressure. However, it is clear that distinctly different trends in LF and HF exist between the MDD group and the control group, suggesting a qualitative change in autonomic regulation in MDD. Second, our main purpose was to compare between-group differences in reactive HRV measures, not to assess how the magnitude of the differences can be; therefore, we used a serial 7 s test to mildly stress participants. Of course, a more strenuous test may provoke greater responsiveness, and if so, patients with first episode of MDD may show both qualitative and quantitative changes in reactive HRV. Third, the generalizability of our study is limited given the narrow inclusion criteria. For example, our patients with MDD were recruited in an inpatient unit. It is possible that inpatients presented with more severe forms of depression than those of outpatients. Fourth, it is unclear whether laboratory-induced stress mirrors real-life stress, which can be chronic, intermittent, and psychosocial in nature, although preliminary data suggested that shifts in cardiovascular responses are stable across contexts (Frankish and Linden, 1991). 5. Conclusions We provide evidence that in patients with their first episode of MDD, there have been qualitative, not quantitative, changes in reactive HRV measures. Unlike healthy controls, patients with MDD revealed an overall shift of autonomic balance toward parasympathetic activity in response to stress. Above all, even quite mild stress exposure in patients with first episode of MDD can still precipitate compromised autonomic regulation in HRV. These findings may signal the accumulative effects of MDD on cardiovascular system in real life stress conditions, which can be chronic, intermittent, tremendous, and psychosocial in nature. Further research with a prospective follow-up design to explore the long-term effects of MDD on reactive HRV is strongly encouraged. Contributors Chih-Sung Liang contributed to study design, literature review, and data collection and wrote the manuscript. Jia-Fu Lee contributed to study design, wrote the protocol, and provided overall scientific supervision. Chia-Chi Chen supervised the assessment of the patients and managed the literature searches. Yue-Cune Chang undertook the statistical analysis, interpreted the results, and edited the manuscript. All authors contributed to and approved the final manuscript. Conflict of interest Some of the raw data had been presented in World Psychiatric Association Regional Meeting, 3–5 November 2011, Kaohsiung, Taiwan (ROC). The authors have no disclaimer statements and no conflicts of interest to declare. Acknowledgements Funding for this study was provided by Civilian Administration Division of Beitou Branch, Tri-Service General Hospital, National Defense Medical Center (DOD98-71 and BAFH-99-10). The Civilian
57
Administration Division had no further role in study design; in the collection, analysis, or interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. References Altman DG, Royston P. The cost of dichotomising continuous variables. BMJ 2006;332: 1080. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4:561–71. Carroll D, Phillips AC, Der G. Body mass index, abdominal adiposity, obesity, and cardiovascular reactions to psychological stress in a large community sample. Psychosom Med 2008;70:653–60. Cohen J. A power primer. Psychol Bull 1992;112:155–9. de Rooij SR, Schene AH, Phillips DI, Roseboom TJ. Depression and anxiety: associations with biological and perceived stress reactivity to a psychological stress protocol in a middle-aged population. Psychoneuroendocrinology 2010;35:866–77. Frankish J, Linden W. Is response adaptation a threat to the high-low reactor distinction among female college students? Health Psychol 1991;10:224–7. Frasure-Smith N, Lesperance F, Habra M, Talajic M, Khairy P, Dorian P, et al. Elevated depression symptoms predict long-term cardiovascular mortality in patients with atrial fibrillation and heart failure. Circulation 2009;120:134–40. Hatzinger M, Hemmeter UM, Baumann K, Brand S, Holsboer-Trachsler E. The combined DEX-CRH test in treatment course and long-term outcome of major depression. J Psychiatr Res 2002;36:287–97. Hughes JW, Stoney CM. Depressed mood is related to high-frequency heart rate variability during stressors. Psychosom Med 2000;62:796–803. Kemp AH, Quintana DS, Gray MA, Felmingham KL, Brown K, Gatt JM. Impact of depression and antidepressant treatment on heart rate variability: a review and meta-analysis. Biol Psychiatry 2010;67:1067–74. Kibler JL, Ma M. Depressive symptoms and cardiovascular reactivity to laboratory behavioral stress. Int J Behav Med 2004;11:81–7. Koschke M, Boettger MK, Schulz S, Berger S, Terhaar J, Voss A, et al. Autonomy of autonomic dysfunction in major depression. Psychosom Med 2009;71:852–60. Kuo TB, Lin T, Yang CC, Li CL, Chen CF, Chou P. Effect of aging on gender differences in neural control of heart rate. Am J Physiol 1999;277:H2233–9. Liao D, Barnes RW, Chambless LE, Simpson Jr RJ, Sorlie P, Heiss G. Age, race, and sex differences in autonomic cardiac function measured by spectral analysis of heart rate variability—the ARIC study. Atherosclerosis Risk in Communities. Am J Cardiol 1995;76:906–12. McEwen BS. Stress, adaptation, and disease. Allostasis and allostatic load. Ann N Y Acad Sci 1998;840:33–44. Phillips AC. Blunted cardiovascular reactivity relates to depression, obesity, and selfreported health. Biol Psychol 2011;86:106–13. Prentice DA, Miller DT. When small effects are impressive. Psychol Bull 1992;112:160. Reyes del Paso GA, Langewitz W, Mulder LJ, van Roon A, Duschek S. The utility of low frequency heart rate variability as an index of sympathetic cardiac tone: a review with emphasis on a reanalysis of previous studies. Psychophysiology 2013;50:477–87. Rottenberg J, Clift A, Bolden S, Salomon K. RSA fluctuation in major depressive disorder. Psychophysiology 2007;44:450–8. Salomon K, Clift A, Karlsdottir M, Rottenberg J. Major depressive disorder is associated with attenuated cardiovascular reactivity and impaired recovery among those free of cardiovascular disease. Health Psychol 2009;28:157–65. Salomon K, Bylsma LM, White KE, Panaite V, Rottenberg J. Is blunted cardiovascular reactivity in depression mood-state dependent? A comparison of major depressive disorder remitted depression and healthy controls. Int J Psychophysiol 2013;90:50–7. Schwerdtfeger A, Rosenkaimer AK. Depressive symptoms and attenuated physiological reactivity to laboratory stressors. Biol Psychol 2011;87:430–8. Sheffield D, Krittayaphong R, Cascio WE, Light KC, Golden RN, Finkel JB, et al. Heart rate variability at rest and during mental stress in patients with coronary artery disease: differences in patients with high and low depression scores. Int J Behav Med 1998; 5:31–47. Shinba T. Altered autonomic activity and reactivity in depression revealed by heart-rate variability measurement during rest and task conditions. Psychiatry Clin Neurosci 2013;68:225–33. Treiber FA, Kamarck T, Schneiderman N, Sheffield D, Kapuku G, Taylor T. Cardiovascular reactivity and development of preclinical and clinical disease states. Psychosom Med 2003;65:46–62. Van der Kooy K, van Hout H, Marwijk H, Marten H, Stehouwer C, Beekman A. Depression and the risk for cardiovascular diseases: systematic review and meta analysis. Int J Geriatr Psychiatry 2007;22:613–26. York KM, Hassan M, Li Q, Li H, Fillingim RB, Sheps DS. Coronary artery disease and depression: patients with more depressive symptoms have lower cardiovascular reactivity during laboratory-induced mental stress. Psychosom Med 2007;69:521–8.
Contents lists available at ScienceDirect
Progress in Neuro-Psychopharmacology & Biological Psychiatry
Reactive heart rate variability in male patients with first-episode major depressive disorder☆ Chih-Sung Liang a,b, Jia-Fu Lee c,d,⁎, Chia-Chi Chen e, Yue-Cune Chang f a
Graduate Institute of Medical Sciences, National Defense Medical Center, Taipei, Taiwan, ROC Department of Psychiatry, Beitou Branch, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan, ROC Department of Psychiatry, Taipei Tzu Chi General Hospital, Buddhist Tzu Chi Medical Foundation, Taipei, Taiwan, ROC d School of Medicine, Tzu Chi University, Hualien, Taiwan, ROC e Department of Nursing, Kang-Ning Junior College of Medical Care and Management, Taipei, Taiwan, ROC f Department of Mathematics, Tamkang University, Taipei, Taiwan, ROC b c
a r t i c l e
i n f o
Article history: Received 22 June 2014 Received in revised form 26 July 2014 Accepted 8 August 2014 Available online 19 August 2014 Keywords: Cardiovascular reactivity First episode Heart rate variability Major depressive disorder Reverse trends Stress
a b s t r a c t Objective: The association between cardiovascular reactivity and major depressive disorder (MDD) remains unclear. This study aimed to examine this association via reactive heart rate variability (HRV) in a welldiagnosed first-episode MDD group and a control group. Methods: A total of 160 physically healthy, drug-naive patients presenting with their first-episode MDD and 50 healthy controls were recruited. All participants underwent a 5-min electrocardiography at rest and during a mental arithmetic task. Depression severity was assessed using the Beck Depression Inventory II (BDI). Results: HRV measures that showed between-group differences at rest did not reached significance during mental stress. In contrast, HRV measures that revealed between-group differences during stress did not reach significance at rest. In response to mental stress, HRV measures did not significantly change in both group. However, LF and HF in response to stress were different between groups. Patients with MDD revealed an increasing trend in HF and a decreasing trend in LF; conversely, healthy controls had a decreasing trend in HF and an increasing trend in LF. BDI scores correlated with changes in heart rate in the control group. Conclusions: The fundamental change to reactive HRV in patients with first-episode MDD appears qualitative, not quantitative. A distinctly reverse trend in reactive HRV measures were evident between these two groups. Moreover, patients with MDD showed entirely distinct changes in reactive HRV from those in resting HRV. We suggest that in patients with MDD, autonomic system shifts to sympathetic dominance at rest but toward parasympathetic dominance in response to stress. © 2014 Elsevier Inc. All rights reserved.
1. Introduction Understanding the cardiovagal function in depression is of great significance to public health since a large body of evidence has indicated
Abbreviations: MDD, major depressive disorder; HRV, heart rate variability; CVD, cardiovascular diseases; CVR, cardiovascular reactivity; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders Fourth Edition; BDI, Beck Depression Inventory II; ECG, electrocardiogram; LF, low frequency; HF, high frequency; TP, total power; VAR, variances of the RR interval; SDNN, the standard deviation of all normal-to-normal interbeat interval; MHR, the mean value of heart rate; MRR, the mean value of the RR interval; SD, standard deviation; SE, standard error; CI, confidence interval. ☆ The article has not been published previously, that it is not under consideration for publication elsewhere, that its publication is approved by all authors and tacitly or explicitly by the responsible authorities where the work was carried out, and that, if accepted, it will not be published elsewhere including electronically in the same form, in English or in any other language, without the written consent of the copyright-holder. ⁎ Corresponding author at: Department of Psychiatry, Taipei Tzu General Chi Hospital, Buddhist Tzu Chi Medical Foundation, Taipei, Taiwan, ROC, No.289, Jianguo Rd., Xindian Dist., New Taipei City 23142, Taiwan, ROC. Tel.: +886 2 6628 9779; fax: +886 2 2794 9366.
http://dx.doi.org/10.1016/j.pnpbp.2014.08.004 0278-5846/© 2014 Elsevier Inc. All rights reserved.
that depression increases the risk of cardiovascular diseases (CVD) and its related mortality (Frasure-Smith et al., 2009; Van der Kooy et al., 2007). Cardiovascular reactivity (CVR) is the responsiveness of the cardiovascular system to stress, and reactivity is derived as changes from baseline levels not absolute levels during stress. CVR reflects autonomic regulatory capacity, and evidence suggests that CVR predicts the development of preclinical and clinical CVD states (Treiber et al., 2003). Research on the association between CVR and depression has yield inconsistent results in recent years. A meta-analysis published in 2004, which consisted of 11 relevant studies, suggested depression involving enhanced CVR (Kibler and Ma, 2004); however, the negative association between attenuated CVR and depressive symptoms has been repeatedly observed in non-clinical participants (Schwerdtfeger and Rosenkaimer, 2011), in people with coronary artery disease (York et al., 2007), in large-scale community samples (Carroll et al., 2008; de Rooij et al., 2010; Phillips, 2011), and in patients with major depressive disorder (MDD) (Rottenberg et al., 2007; Salomon et al., 2009, 2013).
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
Heart rate variability (HRV) is a widely used tool to assess cardiac autonomic activity, and it can reflect the balance between the sympathetic and the parasympathetic regulatory control of heartbeat. HRV has been proposed as a more appropriate measure to examine the association between CVR and depression (Schwerdtfeger and Rosenkaimer, 2011). Although strong evidence from the meta-analysis suggests that depression without CVD is associated with attenuated HRV (Kemp et al., 2010), few studies have investigated HRV in response to stress (also known as reactive HRV or HRV reactivity) in depressed individuals. A study, which enrolled a small sample with non-clinical depression and coronary artery disease, showed that during mental stress, the increases in low-to-high frequency (LF/HF) ratio and heart rate were more pronounced in participants with higher depression scores (Sheffield et al., 1998). Another study with 53 otherwise healthy participants reported greater reduction in high frequency (HF) in the highly depressed mood group (Hughes and Stoney, 2000). The authors suggest that the magnitude of decrease in parasympathetic cardiac control during stress was associated with depressed mood (Hughes and Stoney, 2000). Nevertheless, a recent study including 22 drug-naive, clinical depressed patients produced different results (Shinba, 2013). In this study, patients with clinical depression revealed an increase in reactive HF and decreases in reactive LF/HF ratio and heart rate. Given that during the task performance, reactive HRV for clinical depressed patients were different from those of healthy controls, the authors suggest an altered state of autonomic reactivity in depression (Shinba, 2013). The methodological differences, clinical heterogeneity of subjects, and differences in medication use may have led to the inconsistent findings. Moreover, the sample size in prior studies was relatively small, probably skewing the results. The aim of the study was to explore the association between CVR and MDD by analyzing reactive HRV. We carefully enrolled only drug-naive, CVD-free, male participants diagnosed with their first episode of MDD. The study design could provide a better window into CVR in the early stage of MDD. We also examined whether changes in reactive HRV were different from changes in resting HRV and how depression severity was associated with changes in reactive HRV.
2. Methods 2.1. Participants Participants were recruited from the inpatient division of Beitou Branch, Tri-Service General Hospital, National Defense Medical Center, a psychiatric teaching hospital in Taiwan. Written informed consent was obtained in accordance with the National Health and Medical Research Council guidelines. All subjects were fully informed regarding the aims and details of the study and were free to withdraw their consent at any time. Between 2009 and 2012, a total of 160 individuals with a first episode of MDD were judged eligible. To avoid the potential confounders, such as age, gender, ethnicity, use of psychotropics, and CVD (Liao et al., 1995), individuals were required to be unrelated Han Chinese males, between the ages of 20 and 40 years, drug-naive, and in good health. The diagnosis of MDD—meeting the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria—was made by two certified psychiatrists by diagnostic interview. The severity of depression was assessed with the Chinese version of the Beck Depression Inventory II (BDI) (Beck et al., 1961). Individuals were excluded from the study for any of the following reasons: systolic blood pressure exceeding 180 mmHg or diastolic blood pressure exceeding 110 mmHg; histories of or concurrent medical illnesses such as heart disease, diabetes, CVD, liver or renal disease, endocrinopathies, neurological disorders, or cancer; meeting the DSM-IV criteria for substance-related disorders, psychotic disorders,
53
or bipolar disorder; use of any drugs or over-the-counter preparations; and being shift workers. A series of examinations were performed, including a complete physical examination, a routine biochemical panel, a complete blood count, a urine test, a stool study, a chest X-ray, and an electrocardiogram (ECG). Height (cm) and weight (kg) of each participant were measured by a standard balance beam scale, and body mass index (BMI) was calculated. The medical charts were systematically reviewed in order to confirm that all participants met the above operational criteria. The control group consisted of 50 healthy, never-depressed volunteers matched for sex, age, smoking status, and time of day of ECG recordings. Controls also signed an informed consent form, were interviewed, and completed a BDI rating. Two certified psychiatrists confirmed that none of the healthy controls had a history of any psychiatric illness or were taking any medications. The Institutional Review Board for the Protection of Human Subjects at the Tri-Service General Hospital, a medical teaching hospital within the National Defense Medical Center in Taiwan, approved the protocol. 2.2. HRV recording procedures Detailed procedures for HRV analysis have been previously reported and are only briefly summarized here (Kuo et al., 1999). Prior to the measurements, participant were instructed to refrain from alcohol and caffeine consumption for 24 h and from smoking, exercising, and heavy eating for at least 8 h. All participants had a usual breakfast on the study day to ensure standardization of ECG recordings. After the participants had rested for 15 m in a quiet, air-conditioned room, the ECG for the analysis of beat-to-beat HRV was recorded under standardized conditions. They were asked to relax, breath naturally, and move as little as possible. Experimental procedures were performed between 08.30 h and 11.30 h in a fixed order of rest followed by an arithmetic task. The mental arithmetic task consisted of a serial subtraction by 7, beginning with the number 2193. During the arithmetic task, participants were prompted to perform faster because they were going “slowly,” and this feedback was for all participants. Mental arithmetic tasks have been recognized as active, self-relevant stressors (Schwerdtfeger and Rosenkaimer, 2011). A 288-s signal sequence of the telemetrically transmitted lead I ECG was recorded. The raw ECG signals were amplified with a gain of 1000 and band-pass filtered (0.68–16 Hz). Signals were then recorded with an 8-bit analog-to-digital converter with a sampling rate of 256 Hz. The digitized ECG signals were analyzed online and simultaneously stored on removable hard disks for offline verification. Signal acquisition and data storage and processing were performed via a generalpurpose personal computer. Our computer algorithm identified each QRS complex and rejected each ventricular premature complex or noise according to its likelihood in a standard QRS template. Stationary RR values were re-sampled and interpolated at a rate of 7.11 Hz to produce continuity in the time domain. Frequency-domain analysis was performed by a nonparametric method of fast Fourier transformation. The direct current component was deleted, and a Hamming window was used to attenuate the leakage effect. For each time segment (288-second; 2048 data points), our algorithm estimated the power spectrum density based on the Fourier transformation. The resulting power spectrum was corrected for attenuation resulting from the sampling and the Hamming window. The power spectrum was subsequently quantified into standard frequency-domain measurements including low frequency (LF) (0.04–0.15 Hz), HF (0.15–0.40 Hz), and total power (TP). The timedomain parameters were the standard deviation of all normal-tonormal interbeat interval (SDNN) and the variance of RR interval (VAR). Both parameters were measured in milliseconds (ms). Although mean value of HR (MHR) and mean value of RR interval (MRR) during the 5 min did not belong to HRV components, these two measure were also calculated under both conditions. The spectral components
54
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
of HRV were analyzed as absolute units (ms2). All HRV analyses were performed by a trained research nurse blinded to individual depression severity and the protocol. Individuals were excluded from further analysis if they were not in a predominantly regular sinus rhythm or if they had sustained atrial arrhythmias, such as atrial fibrillation or greater than 5% ectopic complexes. 2.3. Statistical analysis Independent-sample t-tests examined group differences in clinical and demographic characteristics. A multivariate general linear model was performed to compare group differences in resting HRV measures, MHR, and MRR, with age and BMI as covariates. Logarithmic transformations were applied to correct the skewed distributions of LF, HF, TP, LF/HF ratio, VAR, SDNN, MHR, and MRR. Group differences in reactive HRV measures, MHR, and MRR were analyzed with a multivariate general linear model for repeatedmeasures data. Time was entered as a 2-level within-subject factor, and group was entered as a between-subject factor. The model used a full factorial design and included covariates of BMI and age. Healthy controls was the reference group. The model can examine the interaction effect of group × time. Within-group reactivity in HRV measures, MHR, and MRR during mental stress were examined as well. Spearman correlations examined whether BDI scores correlated with reactivity in HRV measures, MHR, and MRR. Reactivity measures were calculated by subtracting baseline values from during-task values. All analyses were performed using the Statistical Package for the Social Sciences for Windows, Version 19.0 (SPSS Inc., Chicago, IL, USA), and effect sizes are reported as partial eta-squared (η2p). Analyses were considered to be statistically significant if its p-values were less than or equal to 0.05 (two-tailed). 3. Results Data analysis consisted of 156 patients with first episode of MDD and 49 healthy controls. One healthy control and four participants with MDD had to be excluded because they failed to complete HRV protocol. The mean age in all participants were 24.4 ± 2.7 years. Table 1 includes data on demographic and clinical characteristics. The MDD group did not differ from the control group in age, height, and BMI, although they were shorter (t = −2.823, p = 0.006). Moreover, the MDD group had higher BDI scores (t = − 2.823, p b 0.001) and faster heart rate (t = 5.094, p b 0.001). Group differences in resting HRV measures, MHR, and MRR are shown in Table 2. Under resting condition, the MDD group showed lower values in TP (p = 0.002), VAR (p b 0.001), SDNN (p b 0.001), and MRR (p b 0.001), but they had higher LF/HF ratio (p = 0.024) and faster MHR (p b 0.001). Their resting LF and HF did not differ from those of the control group. Fig. 1 illustrates group differences in reactive HRV measures, MHR, and MRR. The interaction effect of group × time reached statistical Table 1 Demographic comparisons between patients with major depressive disorder and healthy controls. MDD, N = 156
Age (year)⁎ BDI score⁎ Height (cm)⁎ Body weight (kg) Body mass index (kg/m2)⁎ Heart rate (beat/min)⁎
Controls, N = 49
t test
Mean
SD
Mean
SD
t
p
24.4 34.1 171.4 66.5 22.6 75.3
3 12.3 5.8 10.7 3.4 12.7
24.7 4.9 173.7 69 22.8 66.4
1.5 3.5 4.6 8.8 2.3 10
−0.95 26.428 −2.823 −1.453 −0.425 5.094
0.344 b0.001 0.006 0.148 0.672 b0.001
Note. ⁎The Levene test for equality of variances was statistically significant. SD = standard deviation; MDD = major depressive disorder; BDI = Beck Depression Inventory II.
Table 2 Comparisons between patients with major depressive disorder (N = 156) and healthy controls (N = 49) in resting heart rate variability. Rest
B
SE
t
p
95% CI
LF (ms2) HF (ms2) LF/HF ratio TP (ms2) VAR (ms) SDNN (ms) MRR (ms) MHR (beat/min)
0.04 −0.13 0.161 −0.423 −0.499 −0.251 −0.123 0.122
0.051 0.86 0.132 0.138 0.136 0.068 0.027 0.027
0.792 −1.524 1.219 −3.064 −3.668 −3.684 −4.611 4.594
0.429 0.129 0.024 0.002 b0.001 b0.001 b0.001 b0.001
−0.06 −0.299 −0.1 −0.696 −0.768 −0.385 −0.175 0.07
η2p 0.141 0.038 0.422 −0.151 −0.231 −0.117 −0.07 0.175
0.003 0.011 0.007 0.045 0.063 0.063 0.096 0.095
Note. A multivariate general linear model was performed with covariates of age and body mass index. Healthy controls were the reference group. SE = standard error; CI = confidence interval; LF = low frequency; HF = high frequency; TP = total power; SDNN = the standard deviation of all normal-to-normal interbeat interval; VAR = variances of the RR interval; MRR = the mean value of the RR interval; MHR = the mean value of heart rate.
significance in LF (F = 4.068, p = 0.045, η2p = 0.02), HF (F = 5.425, p = 0.021, η2p = 0.026), LF/HF ratio (F = 5.018, p = 0.026, η2p = 0.024), MRR (F = 81.803, p b 0.001, η2p = 0.289), and MHR (F = 80.752, p b 0.001, η2p = 0.287). During the arithmetic task, both groups showed similar trends in TP, VAR, and SDNN, and the interaction effects in these measures were not significant. Notably, neither the MDD group (F = 0.578, p = 0.717, η2p = 0.019) nor the control group (F = 0.547, p = 0.769, η2p = 0.074) reached significance on analysis of within-group reactivity in HRV measures. The control group showed significant reactivity in MHR and MRR (both: F = 4.601, p = 0.015, η2p = 0.17). Nevertheless, within-group reactivity in MHR and MRR were not significant for the MDD group (both: F = 1.561, p = 0.213, η2p = 0.01). Because our data clearly showed bimodal distribution of scores on BDI, only within-group correlations were analyzed. BDI scores did not correlate with reactive HRV measures in both groups. In healthy controls, BDI scores negatively correlated with MHR (r = − 0.401, p = 0.004) but positively correlated with MRR (r = 0.313, p = 0.028); however, these two measures did not correlate with BDI scores in patients with MDD. Fig. 2 displays the correlation between BDI scores and reactive heart rate. 4. Discussion 4.1. Reactive HRV in first-episode MDD: neither enhanced nor attenuated Few studies have explored CVR in MDD via reactive HRV analysis. The importance of our study is that we compared reactive HRV between a well-diagnosed, first-episode MDD group and a healthy control group, and all participants were male, drug-naive, and screened for the absence of CVD. We found that both reactive HRV and resting HRV were different for patients with MDD and healthy controls, and patients with MDD showed entirely distinct changes in reactive HRV from those in resting HRV. Although at rest, the MDD group revealed lower HRV in TP, VAR, and SDNN, the interaction effects of group × time on these measure did not reach significance. Besides, under resting condition, the MDD group did not differ from the control group in LF and HF, but the interaction effects on these two measures reached significance. That is, patients with MDD revealed an increasing trend in HF and decreasing trends in LF and LF/HF ratio; conversely, healthy controls had a decreasing trend in HF and increasing trends in LF and LF/HF ratio. Considering the within-group reactivity, neither the MDD group nor the control group reached significance in HRV measures; nevertheless, the MDD group clearly showed similar but significantly attenuated trends in MHR and MRR compared with those of the control group. On the other hand, both groups did not have any association between reactive HRV and depression severity; nevertheless, in the
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
3.50
High Frequency
4.25
Low Frequency
Control Group
Control Group
MDD Group
MDD Group
3.25
4.00
3.00
3.75
Baseline
Stress
Baseline
Low Frequency / 1.25
55
8.25
Control Group MDD Group
1.00
Total Power
8.50
High Frequency Ratio
Stress
Control Group MDD Group
8.00 7.75 7.50
0.75
7.25 7.00
0.50
Baseline
7.00
Stress
Baseline
Stress
Standard Deviation of
Mean Value of RR Interval
RR Interval
4.50
Control Group MDD Group 6.75
Control Group
4.25
MDD Group 4.00
3.75
6.50
3.50
Baseline
8.50
Stress
Baseline
Variance of RR Interval
8.25
4.50
Stress
Mean Value of Heart Rate Control Group
Control Group
MDD Group
MDD Group
8.00
4.25
7.75
7.50
4.00
7.25
Baseline
Stress
Baseline
Stress
Fig. 1. Comparisons between patients with MDD and healthy controls in reactive HRV. The interaction effect of group × time reached significance in LF (F = 4.068, p = 0.045, η2p = 0.02), HF (F = 5.425, p = 0.021, η2p = 0.026), LF/HF ratio (F = 5.018, p = 0.026, η2p = 0.024), MRR (F = 81.803, p b 0.001, η2p = 0.289), and MHR (F = 80.752, p b 0.001, η2p = 0.287). The slope of the line representing reactive HF is positive in the MDD group, whereas it is negative in the control group. Conversely, the slope of the line representing LF and LF/HF ratio are negative in the MDD group, whereas they are positive in the control group. The slope of the line representing reactive MHR and MRR in the control group are steeper than those in the MDD group. MDD = major depressive disorder; HRV = heart rate variability; LF = low frequency; HF = high frequency; MRR = the mean value of the RR interval; MHR = the mean value of heart rate.
control group, BDI scores positively correlated with MRR and negatively correlated with MHR. In summary, in patients with first episode of MDD in response to mental stress, reactivity in HF and LF measures were
neither attenuated nor enhanced but found to show reverse trends from those of healthy controls. Furthermore, patients with MDD showed entirely distinct changes in reactive HRV from those in resting HRV.
56
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
in heart period are largely dependent on parasympathetic activity (Schwerdtfeger and Rosenkaimer, 2011). Our MDD group at rest revealed lower HRV, suggesting a shift of autonomic balance away from parasympathetic side. A previous study investigating HRV, QT variability, and baroreflex sensitivity reported that under resting condition, patients with MDD showed an overall shift of autonomic balance toward sympathetic dominance (Koschke et al., 2009). Indeed, faster heart rate was evident in our participants with MDD under resting condition. Taken these findings together, we suggest that in patients with MDD, autonomic system shifts to sympathetic dominance at rest but toward parasympathetic dominance in response to stress. 4.4. Within-group changes in HRV measures and heart rate
Fig. 2. The correlation between reactive heart rate (beat/min) and depression severity. In patients with major depressive disorder, BDI scores did not correlate with changes in heart rate (r = −0.056, p = 0.484). In healthy controls, BDI scores negatively correlated with changes in heart rate (r = −0.401, p = 0.004). BDI = Beck Depression Inventory II.
4.2. Non-clinical depression and dichotomization We provide evidence that the fundamental change to reactive HRV in first-episode MDD is qualitative, not quantitative. At rest, our MDD group had lower HRV, a finding that agreed with the well-established evidence from the meta-analysis (Kemp et al., 2010). During mental stress, their HRV relative to that of healthy controls is neither attenuated nor enhanced, but revealed totally distinct trends in LF and HF. Our results, in part, agreed with a prior studies using patients with clinical depression (Shinba, 2013). The author reported attenuated reactivity in heart rate and LF/HF ratio and increased reactivity in HF. However, two prior studies recruiting non-clinical subjects with depressed mood reported inconsistent findings on reactive HRV (Hughes and Stoney, 2000; Sheffield et al., 1998). The heterogeneity of subjects —patients with MDD versus non-clinical participants with depressed mood—may lead to the inconsistent findings. Moreover, the previous two studies categorized participants according to the median splits of depression scores (Hughes and Stoney, 2000; Sheffield et al., 1998). During data analysis, dichotomization of continuous variables have been strongly discouraged since it may increase the risk of a false positive result and reduce the power (Altman and Royston, 2006). Also, the categorized groups usually indicated that the depression psychopathology of these two groups were quite distinct from each other. In this regard, in non-clinical sample with depressed mood, dichotomizing participants according to depression scores might not be appropriate. Taken together, the qualitative, not quantitative, changes in LF and HF suggest that patients with first episode of MDD had qualitatively distinct changes in reactive HRV. 4.3. Toward parasympathetic dominance during mental stress Broad evidence supports that HF is a reliable measure of cardiac parasympathetic tone (Reyes del Paso et al., 2013). When confronted with a psychological stressor, our MDD group revealed an increase trend in parasympathetic activity, while the control group showed a reverse trend. A reciprocal relationship in LF and HF were evident for both groups. Therefore, mental stress appear to shift autonomic balance toward parasympathetic side in patients with MDD but away from parasympathetic dominance in healthy controls. Attention should be directed toward another issue—to which side autonomic balance tends to shift in patient with MDD under resting condition. It is conceivable that at rest, vagal tone prevails and variations
It is important to note that in response to mental stress, neither the MDD group nor the control group revealed significant reactivity in HRV measures. Moreover, reactivity in MHR and MRR were not significant for the MDD group but for the control group. This implies no quantitative change in HRV between groups when confronted with our arithmetic task. The attenuated reactivity in heart rate in the MDD group might be related to their autonomic shift toward parasympathetic dominance in response to stress. Moreover, in the MDD group, there might be a ceiling effect that could limit the extent of increase in heart rate, because evidence suggests that under resting condition, patients with MDD have shifted their autonomic balance toward sympathetic dominance (Koschke et al., 2009). From another physiological perspective, the attenuated reactivity in heart rate may be a compensatory mechanism from other stress systems (McEwen, 1998). For example, the hypothalamic–pituitary–adrenocortical (HPA) axis, a key component of the neuroendocrine system that controls reactions to stress, is well documented to be activated in MDD (Hatzinger et al., 2002). Thus, the attenuated reactivity in heart rate may simply reflect a compensation for the activation of the HPA axis. 4.5. Effect sizes According to the interpretive guideline for effect sizes by Cohen (Cohen, 1992), our study produced a small effect size (η2p b 0.04). This indicates that although the depressive symptoms in our patients with MDD were severe, as evidenced by a mean score of 34.1 on BDI, MDD state only explain a small variance of between-group differences in reactive HRV measures. Someone may treat the effects of MDD on reactive HRV as trivial, for effect sizes represents an index of clinical importance. It is, however, important to consider that our depressed patients were drug-naive and experiencing their first episode of MDD. Thus, the long-term effects of MDD on HRV may become more pronounced. It has been suggested that small effects in ongoing processes can accumulate to larger ones (Prentice and Miller, 1992). Thus, we expect that the effect sizes will be increased if these patients with MDD are followed prospectively. Because CVR can predict the development of preclinical and clinical CVD states (Treiber et al., 2003), and depression may increase the risk of CVD and its related mortality (Frasure-Smith et al., 2009; Van der Kooy et al., 2007), the clinical importance of an effect size can be a function of how mild an experimental manipulation is to still produce a difference. As a consequence, a small effect size still suggests a compromised autonomic regulation in patients with first episode of MDD. 4.6. Limitations First, available data, however, have challenged the view that LF component indicates sympathetic cardiac control (Reyes del Paso et al., 2013). It is suggested that both LF and HF components are mainly determined by the parasympathetic system but provide different information about autonomic regulation. That is, HF relates to respiratory influences
C.-S. Liang et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 56 (2015) 52–57
and LF reflects the modulation of vasomotor tone. Therefore, caution must be exercised when interpreting our results since we did not assess respiration and blood pressure. However, it is clear that distinctly different trends in LF and HF exist between the MDD group and the control group, suggesting a qualitative change in autonomic regulation in MDD. Second, our main purpose was to compare between-group differences in reactive HRV measures, not to assess how the magnitude of the differences can be; therefore, we used a serial 7 s test to mildly stress participants. Of course, a more strenuous test may provoke greater responsiveness, and if so, patients with first episode of MDD may show both qualitative and quantitative changes in reactive HRV. Third, the generalizability of our study is limited given the narrow inclusion criteria. For example, our patients with MDD were recruited in an inpatient unit. It is possible that inpatients presented with more severe forms of depression than those of outpatients. Fourth, it is unclear whether laboratory-induced stress mirrors real-life stress, which can be chronic, intermittent, and psychosocial in nature, although preliminary data suggested that shifts in cardiovascular responses are stable across contexts (Frankish and Linden, 1991). 5. Conclusions We provide evidence that in patients with their first episode of MDD, there have been qualitative, not quantitative, changes in reactive HRV measures. Unlike healthy controls, patients with MDD revealed an overall shift of autonomic balance toward parasympathetic activity in response to stress. Above all, even quite mild stress exposure in patients with first episode of MDD can still precipitate compromised autonomic regulation in HRV. These findings may signal the accumulative effects of MDD on cardiovascular system in real life stress conditions, which can be chronic, intermittent, tremendous, and psychosocial in nature. Further research with a prospective follow-up design to explore the long-term effects of MDD on reactive HRV is strongly encouraged. Contributors Chih-Sung Liang contributed to study design, literature review, and data collection and wrote the manuscript. Jia-Fu Lee contributed to study design, wrote the protocol, and provided overall scientific supervision. Chia-Chi Chen supervised the assessment of the patients and managed the literature searches. Yue-Cune Chang undertook the statistical analysis, interpreted the results, and edited the manuscript. All authors contributed to and approved the final manuscript. Conflict of interest Some of the raw data had been presented in World Psychiatric Association Regional Meeting, 3–5 November 2011, Kaohsiung, Taiwan (ROC). The authors have no disclaimer statements and no conflicts of interest to declare. Acknowledgements Funding for this study was provided by Civilian Administration Division of Beitou Branch, Tri-Service General Hospital, National Defense Medical Center (DOD98-71 and BAFH-99-10). The Civilian
57
Administration Division had no further role in study design; in the collection, analysis, or interpretation of data; in the writing of the report; and in the decision to submit the paper for publication. References Altman DG, Royston P. The cost of dichotomising continuous variables. BMJ 2006;332: 1080. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. An inventory for measuring depression. Arch Gen Psychiatry 1961;4:561–71. Carroll D, Phillips AC, Der G. Body mass index, abdominal adiposity, obesity, and cardiovascular reactions to psychological stress in a large community sample. Psychosom Med 2008;70:653–60. Cohen J. A power primer. Psychol Bull 1992;112:155–9. de Rooij SR, Schene AH, Phillips DI, Roseboom TJ. Depression and anxiety: associations with biological and perceived stress reactivity to a psychological stress protocol in a middle-aged population. Psychoneuroendocrinology 2010;35:866–77. Frankish J, Linden W. Is response adaptation a threat to the high-low reactor distinction among female college students? Health Psychol 1991;10:224–7. Frasure-Smith N, Lesperance F, Habra M, Talajic M, Khairy P, Dorian P, et al. Elevated depression symptoms predict long-term cardiovascular mortality in patients with atrial fibrillation and heart failure. Circulation 2009;120:134–40. Hatzinger M, Hemmeter UM, Baumann K, Brand S, Holsboer-Trachsler E. The combined DEX-CRH test in treatment course and long-term outcome of major depression. J Psychiatr Res 2002;36:287–97. Hughes JW, Stoney CM. Depressed mood is related to high-frequency heart rate variability during stressors. Psychosom Med 2000;62:796–803. Kemp AH, Quintana DS, Gray MA, Felmingham KL, Brown K, Gatt JM. Impact of depression and antidepressant treatment on heart rate variability: a review and meta-analysis. Biol Psychiatry 2010;67:1067–74. Kibler JL, Ma M. Depressive symptoms and cardiovascular reactivity to laboratory behavioral stress. Int J Behav Med 2004;11:81–7. Koschke M, Boettger MK, Schulz S, Berger S, Terhaar J, Voss A, et al. Autonomy of autonomic dysfunction in major depression. Psychosom Med 2009;71:852–60. Kuo TB, Lin T, Yang CC, Li CL, Chen CF, Chou P. Effect of aging on gender differences in neural control of heart rate. Am J Physiol 1999;277:H2233–9. Liao D, Barnes RW, Chambless LE, Simpson Jr RJ, Sorlie P, Heiss G. Age, race, and sex differences in autonomic cardiac function measured by spectral analysis of heart rate variability—the ARIC study. Atherosclerosis Risk in Communities. Am J Cardiol 1995;76:906–12. McEwen BS. Stress, adaptation, and disease. Allostasis and allostatic load. Ann N Y Acad Sci 1998;840:33–44. Phillips AC. Blunted cardiovascular reactivity relates to depression, obesity, and selfreported health. Biol Psychol 2011;86:106–13. Prentice DA, Miller DT. When small effects are impressive. Psychol Bull 1992;112:160. Reyes del Paso GA, Langewitz W, Mulder LJ, van Roon A, Duschek S. The utility of low frequency heart rate variability as an index of sympathetic cardiac tone: a review with emphasis on a reanalysis of previous studies. Psychophysiology 2013;50:477–87. Rottenberg J, Clift A, Bolden S, Salomon K. RSA fluctuation in major depressive disorder. Psychophysiology 2007;44:450–8. Salomon K, Clift A, Karlsdottir M, Rottenberg J. Major depressive disorder is associated with attenuated cardiovascular reactivity and impaired recovery among those free of cardiovascular disease. Health Psychol 2009;28:157–65. Salomon K, Bylsma LM, White KE, Panaite V, Rottenberg J. Is blunted cardiovascular reactivity in depression mood-state dependent? A comparison of major depressive disorder remitted depression and healthy controls. Int J Psychophysiol 2013;90:50–7. Schwerdtfeger A, Rosenkaimer AK. Depressive symptoms and attenuated physiological reactivity to laboratory stressors. Biol Psychol 2011;87:430–8. Sheffield D, Krittayaphong R, Cascio WE, Light KC, Golden RN, Finkel JB, et al. Heart rate variability at rest and during mental stress in patients with coronary artery disease: differences in patients with high and low depression scores. Int J Behav Med 1998; 5:31–47. Shinba T. Altered autonomic activity and reactivity in depression revealed by heart-rate variability measurement during rest and task conditions. Psychiatry Clin Neurosci 2013;68:225–33. Treiber FA, Kamarck T, Schneiderman N, Sheffield D, Kapuku G, Taylor T. Cardiovascular reactivity and development of preclinical and clinical disease states. Psychosom Med 2003;65:46–62. Van der Kooy K, van Hout H, Marwijk H, Marten H, Stehouwer C, Beekman A. Depression and the risk for cardiovascular diseases: systematic review and meta analysis. Int J Geriatr Psychiatry 2007;22:613–26. York KM, Hassan M, Li Q, Li H, Fillingim RB, Sheps DS. Coronary artery disease and depression: patients with more depressive symptoms have lower cardiovascular reactivity during laboratory-induced mental stress. Psychosom Med 2007;69:521–8.
