The Association Between Antidepressant Treatment and Heart Rate Deceleration Capacity in Patients With Mood disorders—A Potential New Predictor of Sudden Cardiac Death

Introduction

Antidepressant use has been associated with an increased risk of sudden cardiac death (SCD) in patients with mood disorders such as major depressive disorder (MDD) or bipolar disorder (BD).^1,2 Previously, antidepressants, especially tricyclic antidepressants (TCAs), have been reported to increase the incidence of SCD.^3–5 However, recent studies have reported that antidepressant use is associated with a lower risk of SCD and ventricular arrhythmias, indicating an inconsistency in the literature.^6–8

A prolonged corrected QT interval (QTc) observed via electrocardiography (ECG) is a risk factor for torsade de pointes, which can cause SCD.⁹ The use of TCAs in addition to some selective serotonin reuptake inhibitors (SSRIs) has previously been reported to prolong the QT interval.^10–13 Therefore, QT interval prolongation is routinely used as a surrogate marker of the risk of fatal arrhythmias in patients with mood disorders who are taking antidepressants. However, the QT interval is incomplete as a biomarker for proarrhythmic risk of inducing sudden cardiac death¹⁴ because it shows diurnal variation¹⁵ and has measurement errors even among cardiologists,¹⁶ and the algorithms and correction equations used in automated ECG analysis differ depending on the model used.¹⁷ It has even been reported that evidence for clinically meaningful QT prolongation in psychiatric drug treatment remains minimal.¹⁸ Therefore, to correctly assess the risk of SCD due to antidepressant drugs, other more accurate, simple, and noninvasive markers are needed.

Decreased activity of the autonomic nervous system (ANS) is associated with an increased risk of SCD.^19,20 Heart rate variability (HRV) is considered an indicator of cardiac and ANS function, and it represents beat-by-beat variation in heart rate.²¹ HRV decreases with increased sympathetic function and decreased parasympathetic function, and it has been associated with mortality in patients with heart failure, coronary artery disease, and acute myocardial infarction.^22,23 There are reports of decreased HRV in patients taking antidepressant medications.²⁴

Indices of HRVs, including frequency domain components, such as low-frequency (LF) and high-frequency (HF) components, and time domain components, such as the standard deviation of normal-to-normal (SDNN) intervals or root mean square successive difference (RMSSD), have been developed, each with their own measurement method and physiological interpretation.^25,26 Thus, conventional measures of HRV have been shown to provide prognostic information linking reduced HRV with poor prognosis; however, their predictive power is variable and only moderate overall, and their clinical usefulness is low.²⁷

A strong association between heart rate deceleration capacity (DC) and mortality has been reported in patients after myocardial infarction.²⁸ DC is believed to represent the parasympathetic nervous system’s ability to regulate cardiovascular modulation and serves as a functional biomarker for parasympathetic capacity by analyzing the deceleration of the instantaneous heart rate.²⁸ DC values less than 2.5 ms, 2.6 to 4.5 ms, and greater than 4.5 ms represent high, moderate, and low risks of mortality, respectively. DC is more accurate than conventional measures of HRV, such as the SDNN. A reduction in DC results in a high risk of mortality, regardless of ventricular premature beats.²⁸

Therefore, DC may represent a new and promising measure for noninvasive risk stratification to prevent cardiac death.²⁹ In the present study, we determined the DC of patients with mood disorders currently taking antidepressants and other psychotropic drugs to reevaluate their risk of SCD.

Material and Methods

Participants

Data from 112 Japanese patients with mood disorders such as MDD or bipolar disorder (BD) of mild to moderate severity (≦18 points on the Hamilton Depression Rating Scale (HAM-D)) treated with antidepressants or mood stabilizers at Dokkyo Medical University or Fudogaoka Hospital between 2019 and 2024 were collected. Psychiatric diagnoses were made on the basis of the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition. The medical history, pharmacotherapy status, and laboratory data of all patients were reviewed, ECGs were recorded, and blood pressure was measured to evaluate their physical condition. Among the 112 patients, the following patients were excluded because of factors known to affect heart rate and cardiac function: 2 patients with hypertension (systolic blood pressure at the time of measurement > 140 mmHg), 2 patients with hypotension (systolic blood pressure at the time of measurement < 100 mmHg) and 1 patient with hypokalemia (<3.5 mEq/L). Data from the remaining 107 patients (48 males; age range, 20–80 years; mean age, 53.6 ± 16.0 years) were analyzed in this study. Among the 107 patients, 94 were diagnosed with MDD, and 13 were diagnosed with BD. All 94 patients with MDD were taking one or more antidepressants. Additionally, 86 healthy volunteers among the staff at Dokkyo Medical University Hospital and Fudougaoka Hospital who provided consent were recruited. These healthy controls had no physical abnormalities or psychiatric abnormalities.

Measurement of DC

We used an instrument (Wireless Wearable Heart Rate Sensor^®: Union Tool Co., Tokyo, Japan) to measure heart rate in an automatic and noninvasive manner by simply taping electrodes to the chests of patients, and DC was analyzed from the heart rate measured via RRI Analyzer^® (Union Tool Co., Tokyo, Japan).

DC was calculated as described below. From the recorded waveforms of heart rate, the time series data of the R‒R interval were extracted. Points in this time series where the R-R interval was longer than the previous interval were selected as anchors. All segments of the same size around the anchors were selected, and all selected segments were translated in the horizontal axis direction and overlaid with the anchor in the segment as the origin (X0). Then, the signals X within the aligned segments are averaged. DC was quantified via the following equation: DC = (1/4)[X(0)+X(1) - X(−1) - X(−2)]. X(0) and X(1) are the averages of the anchor points and the following R‒R intervals, respectively, whereas X(−1) and X(−2) are the averages of the 2R‒R intervals preceding the anchor points. This method thus followed Bauer et al’s definition of phase-rectified signal averaging, which constitutes averaging the partial time series obtained under specific conditions such as increases or decreases (see also²⁸ ‘s report for a detailed description of the method).

All the measurements were obtained in the same order from 9:00 to 12:00 in a quiet, comfortable room. After appropriate skin preparation, the subjects were fitted with chest electrodes and allowed to rest for 5 minutes before the measurements were taken. After resting, DC and heart rate were recorded continuously for 10 minutes while the participants were in a seated position. During the measurement period, all the subjects were instructed to breathe 15 times per minute (cued with a metronome) to ensure that rhythmic variation in heart rate fluctuations was associated with respiration and not due to another source.

Determination of the QTc From ECG Data

All patients lay supine, and ECG data were recorded during the procedure. A standard 12-lead ECG device was used and recorded at a paper speed of 25 mm/s. The QT interval was measured manually following a previous report.³⁰ The end of the T wave was determined as the intersection between the tangent to the steepest downslope of the T wave and the isoelectric line. The QT interval was then corrected according to the heart rate. In the present study, we used Hodges’ formula (QTc = QT interval+1.75 (heart rate – 60)) because it is considered more appropriate than Bazett’s formula (QTc = QT interval/(RR interval)^1/2) in cases of tachycardia or bradycardia.^31–33

Medications

Information on the drugs taken by patients with mood disorders was obtained from their medical records. The administered drugs remained unchanged for at least 1 week. The doses of antidepressants prescribed to participants were converted into imipramine-equivalent doses.³⁴ The antipsychotic drugs and benzodiazepines prescribed to participants were converted into chlorpromazine- and diazepam-equivalent doses, respectively.³⁴ The distribution of medications and dosages administered are shown in Table 1.

Table 1 Demographic Data of Patients Diagnosed With Mood Disorders and the Distributions of Medication and Dosage

Statistical Analysis

For the first analysis, linear regression analysis was applied to examine risk factors for DC. Age, sex, body mass index (BMI), and TCA (imipramine equivalent), non-TCA (imipramine equivalent), antipsychotic (chlorpromazine equivalent), benzodiazepine (diazepam equivalent), and mood stabilizer (lithium, sodium valproate, and carbamazepine) doses were entered as independent variables in the stepwise regression model (Model 1). The TCAs included clomipramine, amitriptyline, and nortriptyline. Non-TCAs include fluvoxamine, paroxetine, sertraline, escitalopram, milnacipran, duloxetine, venlafaxine, and mirtazapine. In the second analysis, age, sex, BMI and individual antidepressant doses were entered as independent variables in the stepwise regression Model 2. In the third analysis of the 94 patients with MDD who were taking antidepressants and 86 healthy controls, the propensity score was adjusted for age and sex. The mean heart rate and DC values were compared in 68 patients with MDD taking antidepressants adjusted for age and sex (33 males; mean age, 47.9±14.1 years) and 68 healthy subjects adjusted for age and sex (33 males; mean age, 46.7±12.8 years) via independent-samples t tests. All the statistical analyses were performed via SPSS version 29.0 (IBM Japan, Tokyo, Japan). All reported P values were two-tailed, and statistical significance was set at p<0.05, p<0.01, and p<0.001.

Ethics Approval

The protocol for this study was approved by the Bioethics Committee of Dokkyo Medical University (approval number: R-17-3J). This study was conducted in accordance with the Declaration of Helsinki. All clinical records used in this study were anonymized by removing any identifying information (such as name and registration number). Furthermore, the clinical data were collated and managed by one person. The person who performed the statistical analysis could not access the original data of each participant. Healthy controls underwent the examination voluntarily. Our study required written informed consent from each subject and healthy control, and data were collected only from consenting participants.

Results

Demographic Data

Demographic data are presented in Table 1. Among the 107 patients with mood disorders, 27 (25.2%) had a DC of <2.5 ms (mean DC: 1.44±0.63), 25 (23.4%) had a DC between 2.5 ms and 4.5 ms (mean DC: 3.59±0.58), and 55 (51.4%) had a DC over 4.5 ms (mean DC: 7.23±2.18). There were 94 patients with mood disorders who were taking antidepressants, 29 of whom were taking more than two kinds of antidepressants.

Correlation Between DC and Antidepressant Dose

In Model 1, age (partial regression coefficient (PRC) = −0.062, 95% confidence interval (CI) = −0.094–-0.030, p<0.001), TCA dose (PRC=−3.62, 95% CI= −5.69–-1,55, p<0.001) and non-TCA dose (PRC= −0.69, 95% CI= −1.34–-0.042, p<0.05) were negatively correlated with DC in patients with mood disorders (Table 2). In Model 2, age (PRC=−0.070, 95% CI= −0.10–-0.037, p<0.001), clomipramine dose (PRC=−4.21, 95% CI= −8.24–-0.17, p<0.05) and amitriptyline dose (PRC=−5.65, 95% CI= −10.4–-0.94, p<0.05) were negatively correlated with DC in patients with mood disorders (Table 3).

Table 2 Effects of Distinct Psychotropic Drugs on DC

Table 3 Effects of Individual Antidepressants on the DC of Patients With Mood Disorders

Comparison Between Patients With MDD Prescribed Antidepressant Drugs and Healthy Controls

There was a significant difference in heart rate and DC between patients with MDD taking antidepressant medication and healthy controls (Figures 1 and 2). The mean heart rate was 80.2±12.8/min for patients with MDD in the antidepressant group, 75.6±10.9/min for those in the healthy control group (p<0.05) and the mean DC was 5.32±2.97 ms for the patients with MDD in the antidepressant group and 7.60±3.44 ms for those in the healthy control group (p<0.001) (Table 4).

Table 4 Comparison of Heart Rate and DC Between Patients Treated With Antidepressants and Healthy Controls Adjusted for Age and Sex With Propensity Scores

Figure 1 Comparison of pulse between patients treated with antidepressants and healthy controls. Abbreviations: bpm, beat per minute.

Figure 2 Comparison of DC between patients treated with antidepressants and healthy controls. Abbreviations: DC, deceleration capacity.

Discussion

In our study, DC was negatively correlated with the TCA dose, especially for doses of clomipramine and amitriptyline. Although non-TCAs as a whole reduced DC in a dose-dependent manner, no association with DC was demonstrated for individual non-TCAs. Compared with healthy controls, patients with MDD on antidepressant medication presented decreased DCs. To our knowledge, there have been no reports on the association between DC and patients with mood disorders who received antidepressants.

We confirmed that the use of TCAs, particularly clomipramine and amitriptyline, was significantly associated with lower DC in a dose-dependent manner. Compared with those who did not use antidepressants, those who used TCAs had a dose-related increased risk of sudden cardiac death.³⁵ The use of TCAs is associated with decreased parasympathetic activity.²⁴ Previous reports have suggested that TCAs are significantly associated with several markers of reduced parasympathetic function.³⁶

Our previous report revealed that the use of antipsychotics with wide-ranging receptor affinity profiles, such as olanzapine and clozapine, dose-dependently reduced DC in patients with schizophrenia.³⁷ TCAs such as clomipramine and amitriptyline exert relatively stronger blocking effects on muscarinic acetylcholine, adrenergic, and histamine receptors (such as olanzapine and clozapine) than do other types of antidepressants.^38,39 The effects of these drugs on DC may be due to their pharmacological properties, although more detailed validation is needed in the future.

Furthermore, TCAs have been identified as having arrhythmogenic effects, which are caused by strong blockade of cardiac sodium channels and arrhythmogenic activity in potassium channels.³⁸ The use of TCAs is also known to prolong the QT interval, which is considered a predictor of ventricular arrhythmias.¹⁰

Owing to their pharmacological profiles and arrhythmogenic activity, TCAs have potent and complex effects on the cardiovascular system, and these changes can occur at therapeutic levels or in overdose.⁴⁰

Compared with TCAs, non-TCAs have relatively weaker effects on the ANS.²⁴ In particular, SSRIs have been reported in meta-analyses to have no significant association with conventional measures of HRV.^36,41 There are a few reports suggesting that serotonin–noradrenaline reuptake inhibitors (SNRIs) have the second strongest association with decreased HRV after TCAs²⁴ and that venlafaxine also decreases HRV.⁴² The association between mirtazapine and HRV is not consistent across studies.^41,42 A recent systematic review revealed that no SSRIs had an adverse effect on mortality in heart failure patients.⁴³ Although an association with QT interval prolongation in escitalopram and citalopram has been reported,⁴⁴ there are reports of no excess risk of arrhythmia or sudden cardiac death,^11,45,46 and within a clinical dose range, most SSRIs have not been associated with a prolonged QT interval.¹² For SNRIs and mirtazapine, the risk of QT prolongation at therapeutic doses is considered lower.⁴⁷ In our present study, we found no association between individual non-TCAs and DC. When considered together with previous reports, we suggest that there is little cardiac risk associated with using clinical doses of individual non-TCAs.

In contrast, the present study revealed a dose-dependent decrease in DC for non-TCAs overall. When antidepressant monotherapy is considered unresponsive, treatment with two antidepressants in combination is frequently performed, and combination treatment is considered a significantly superior treatment to monotherapy.⁴⁸ However, antidepressant polytherapy has been reported to be associated with an increased risk of ventricular arrhythmias and SCD.⁴⁹ Even the use of antidepressants other than TCAs may have an effect on cardiac function when used in combination with multiple drugs or in excess doses due to overmedication.

Additionally, it has been suggested that the risk of arrhythmias is increased in elderly individuals treated with any antidepressant.⁵⁰ In this study, age was significantly associated with lower DC, suggesting that antidepressant therapy may increase the risk of cardiac function decline in elderly individuals.

In our previous reports, we reported that antipsychotic use decreased DC in a dose-dependent manner.³⁷ However, in the present study, we did not detect a significant association between antipsychotic use and DC reduction. This could be because the patients in that study were patients with schizophrenia, the number of antipsychotic users among patients with mood disorders was lower, and the total use of antipsychotics among patients with mood disorders was lower than that among patients with schizophrenia.

Sodium valproate and carbamazepine are typically used as antiepileptic drugs. There have been several studies on antiepileptic drugs and conventional HRV, but the results have been inconsistent.^51,52 To our knowledge, there are no reports of lithium carbonate and its effects on the ANS. In this study, no significant association with DC was found for mood stabilizers; however, the number of patients taking carbamazepine, sodium valproate or lithium in this study was very small.

There was no significant association between DC and the use of benzodiazepines in this study. Similarly, our previous study revealed no association between benzodiazepine use and DC.³⁷

In the present study, no significant association was detected between DC and QTc. These results were similar to those of our previous report.³⁷ Although it may not be directly comparable to the findings of previous studies of heart failure patients, DC has been reported to be an independent predictor of SCD.⁵³ Hence, DC may be considered an independent factor with respect to cardiac mortality, even in patients with mood disorders.

Limitations

This study has several limitations. First, given the variety of antidepressants used by the participants, the sample sizes of the groups were small. Second, because this study had a cross-sectional design, we could not evaluate changes in DC before and after medication use in the same patients. Third, the original method of Bauer et al involved calculating DC from a 24-hour ECG recording; in contrast, in the present study, the measurement time was only 10 minutes.²⁸ However, short ECG recordings have been reported to provide HRV data that are as useful as those obtained from an entire 24-hour period.^54,55 Finally, with respect to severity, all selected patients had a HAM-D of 18 or less at the time of DC measurement, but this is not a sufficient assessment of severity. Since there are reports that the severity of MDD and BD affects autonomic function,^56–58 detailed validation, taking severity into account, is needed in the future.

Conclusion

Although this research has several limitations, as mentioned above, we found that the use of antidepressants in patients with mood disorders was associated with decreased DC in a dose-dependent manner. Assessing DC might facilitate monitoring and identification of an increased risk of SCD in patients with mood disorders who are taking antidepressants. To increase the safety of antidepressant use, an evaluation of ANS function may be necessary. Future large-scale cohort studies or longitudinal studies are needed to clarify in detail the association between DC and SCD in patients with mood disorders receiving antidepressant treatment in the future.