Introduction

Heart rate variability (HRV) is a physiological phenomenon indicating the variation in the time intervals between consecutive heartbeats referred to as beat-to-beat intervals. HRV measurement is a non-invasive procedure commonly used to evaluate the effects of the autonomic nervous system (ANS) on heart rate (HR).^[1] Non-invasive HRV measurement methods are divided into four main categories: time domain, spectral or frequency domain, geometric, and non-linear methods. Other methods of HRV measurement are baroreflex sensitivity and HR turbulence. The measurements can be calculated over either a short time as a 5-min electrocardiogram (ECG) recording or as a 24-h long-term ECG recording based on the chosen parameter.^[2]

The stability of the selected time domain analysis (such as SDNN, SDNN index, and rMSSD) parameters over time in healthy individuals makes them preferable for the evaluation of ANS function.^[3] Increased sympathetic nervous system (SNS) or decreased parasympathetic nervous system (PNS) activity decreases HRV, whereas increased PNS or decreased SNS activity increases HRV.^[4]

Moreover, low HRV values have also been used as a risk marker for many cardiac or non-cardiac conditions.^{[5],[6],[7],[8]} Aging and psychiatric disorders can lead to reduced HRV, and gender variations have also been shown in HRV.^[9],[10]

The severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), which is the agent virus of coronavirus disease-2019 (COVID-19), has spread worldwide in the past 2 years causing serious morbidity and mortality in millions of patients.^[11],[12] Two recent clinical trials have suggested that COVID-19 may be associated with autonomic symptoms due to ANS dysfunction.^[13],[14]

In a recently published case report, reduced HRV was observed in a patient during COVID-19 infection.^[15] To the best of our knowledge, no clinical study has yet investigated the effect of COVID-19 on HRV changes. Therefore, this study aimed to determine whether COVID-19 infection decreased HRV acutely due to ANS involvement.

Materials and Methods

This cross-sectional study was conducted in the cardiology department of a high-volume training and research hospital. The study protocol was approved by the local ethics committee of the hospital in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines. Written informed consent was obtained from all the study participants.

Patient selection

From the registry of the COVID-19 outpatient clinics between July 2020 and October 2021, 6127 patients with polymerase chain reaction (PCR) results positive for COVID-19 in real-time PCR test of upper respiratory specimens were obtained. All were aged >18 years. Of those patients, 304 had a 24-h Holter ECG recording (Pathfinder, Spacelabs Healthcare, Snoqualmie, WA, USA) at least 6 months before their first positive PCR results were determined from hospital records. The Holter ECG recordings had been ordered due to symptoms of heart palpitation. Exclusion criteria were defined as: a rhythm other than sinus rhythm, diagnosis of sinoatrial atrioventricular conduction defect, use of beta-blockers, diltiazem, digoxin, or other anti-arrhythmic drugs that could cause significant changes in HRV values, any infectious disease diagnosed during Holter ECG recording, use of any medication for a psychiatric disorder, active smoking, a history of thyroid disease, cancer, autoimmune disease, hypertension, diabetes, peripheral neuropathy, metabolic disorders, chronic bronchitis, or asthma (documented or newly diagnosed after referral to internal medicine or neurology outpatient clinics),^{[16],[17],[18],[19]} computed tomography (CT) findings of pulmonary involvement in COVID-19, the clinical presentation with at least one symptom (eg, breathlessness, ankle swelling, or fatigue) and/or at least one sign (eg, elevated jugular venous pressure, pulmonary crackles, or peripheral edema) of heart failure.^[20] After implementation of the exclusion criteria, 151 patients were eligible for the study.

The 24-h Holter ECG recordings of those 151 patients' data were collected. Only 87 patients with at least 24 h of Holter ECG recording with at least 90% normal–normal interval analysis referred to as the first Holter ECG analysis were accepted as suitable for HRV measurement. Those 87 patients were invited to participate in the study within 3 months following the positive PCR (after completing 14 days in quarantine) and included for analysis [Figure 1].

Time and frequency domain measures of heart rate variability in 24-h Holter electrocardiogram monitoring

The 24-h Holter ECG monitoring of all 87 patients was made using the same recording devices. These 87 ECG recordings, referred to as the second Holter ECG analysis, were found to be sufficient for HRV measurement. The mean HR, three main time-domain standard HRV parameters (standard deviation [SD] of all normal RR [NN] intervals [SDNN], mean of the SDs of all normal sinus RR intervals for all 5-min segments [SDNN index], and the root-mean-square of successive differences between NN intervals [rMSSD]), and three frequency-domain HRV parameters (low-frequency [LF] band, high-frequency [HF] band, and LF/HF ratio) were automatically calculated by the recording devices. All the parameters from the first Holter ECG analysis and the second Holter ECG analysis were obtained by a clinician blinded to the characteristics of the patients. A third Holter ECG analysis was planned if any statistically significant differences were observed among the first and the second Holter ECGs.

Each patient's demographic characteristics, clinical characteristics, chronic medications, treatments for COVID-19 (favipiravir, oseltamivir, remdesivir, antibiotics, heparin, and hydroxychloroquine), and laboratory measurements for COVID-19 (creatinine level, white blood cell count, D-dimer level, ferritin level, BNP/proBNP [N-terminal prohormone of brain natriuretic peptide], troponin level, and C-reactive protein level) were collected. In addition, all the patients included in the study were assessed for orthostatic hypotension (OH) using a manual blood pressure cuff. OH was defined as a fall of >20 mmHg systolic and >10 mmHg diastolic after standing for 3 min.^[21]

Statistical analysis

All statistical analyses were performed using IBM (SPSS Inc., Chicago, IL, USA) Statistics 25 software. The variables were stated as mean ± SD values when the distribution was normal according to the Kolmogorov–Smirnov test, otherwise as median and minimum–maximum values. Fisher's exact test was used to compare the categorical variables. For comparisons of the repeated measurements (HRV parameters, mean HR, and Holter ECG recording durations), the paired samples t-test was applied. P < 0.05 was considered statistically significant.

Ethical statement

This study was approved by the University of Health Sciences Turkey, Bursa Yuksek Ihtisas Training and Research Hospital Non-Interventional Health Sciences Research Ethics Committee with the decision number 2011-KAEK-25 2021/04-11, dated 28/04/2021.

Results

The patients comprised 26 (29.9%) males and 61 (70.1%) females, with a mean age of 43.39 ± 14.12 years (minimum: 19 and maximum: 73). None of the patients had syncopal episodes, whereas 11 (%12.6) of the patients stated at least an episode of dizziness that occurs after COVID-19 positivity. The demographic characteristics and chronic medications of the patients are shown in [Table 1].

The median time interval between the date of the positive PCR test and the date of the second Holter ECG analysis was 36 days (minimum: 15 and maximum: 83). The laboratory measurements and treatments for COVID-19 are shown in [Table 2].

The mean edited Holter ECG recording duration was similar in the first Holter ECG analysis and the second Holter ECG analysis (23.45 ± 0.91 h vs. 23.46 ± 0.87 h, P = 0.17). The second Holter ECG analysis showed that there was a significant decrease in SDNN and SDNN index, and an increase in LF/HF ratio (P < 0.05). The HF band was decreased but not at a significant level (P = 0.092) [Table 3].

To determine whether the changes in the HRV parameters recovered with time, 24-h Holter monitoring was repeated 3 months after the second Holter ECG analysis (third Holter ECG analysis). After implementation of the exclusion criteria and manual removal of artifacts and extra beats, 48 of the third Holter ECG analysis were eligible for comparison with the second Holter ECG analysis. In the comparison, a significant increase was observed in SDNN and SDNN index, and a significant decrease in the LF band (P < 0.005). The decrease in LF/HF ratio was not found to be statistically significant (P = 0.052) [Table 4].

Discussion

The study revealed that a significant decrease was determined in the mean SDNN and SDNN index, a significant increase in LF/HF ratio, and a non-significant decrease in HF band in the patients who had a positive PCR result for COVID-19 within the past 3 months. The third Holter ECG analysis showed that the decrease in SDNN and SDNN index and the increase in LF/HF ratio were reversed 6 months after the second Holter ECG analysis. Moreover, a significant decrease in the LF band was observed.

SDNN, which is the SD of all normal RR (NN) intervals, is the most widely used time-domain HRV parameter. When calculated over 24 h, it indicates all the cyclic components which contribute to HRV.^{[22],[23],[24]} A reduced SDNN value mostly refers to sympathetic overactivity or reduced vagal tonus or both.^[25] LF band refers to oscillations of HR in the 0.04–0.15 Hz frequency range. It denotes both parasympathetic and sympathetic activities. The LF/HF ratio is another parameter indicating the sympathovagal balance.^[26] It expresses the ratio of LF band to HF band which includes 0.16–0.4 Hz oscillations of HR. An increased LF/HF ratio indicates depressed vagal activity.^[27],[28] The current study findings that the decrease in SDNN and SDNN index and the increase in LF/HF ratio were reversible and may be associated with ANS dysfunction in COVID-19 favoring sympathetic overactivity or reduced vagal tonus for a short period. Similarly, Mittal et al.^[29] showed that all the HRV parameters were decreased in the early stages of human immunodeficiency virus infection, in which autonomic dysfunction is relatively common.^[30] In another clinical trial, the onset of autonomic symptoms occurred within a short interval after COVID-19 symptoms (median: 7 days).^[14] In parallel with the results of the third Holter ECG analysis in this study, Asarcikli et al.^[31] reported that post-COVID patients were more likely to have SDNN >60 msn, RMSSD >40 msn, and low LF/HF ratio compared to healthy individuals, indicating parasympathetic overtones in the post-COVID period.

Previous reports have shown that COVID-19 may cause Guillain–Barré syndrome or other types of peripheral neuropathy.^[32],[33] Neurological symptoms have been mostly observed in cases with severe COVID-19 infection. Direct viral invasion of the peripheral nervous system by the SARS-Cov-2 virus and/or immune-mediated damage to peripheral nerves have been considered two possible mechanisms of peripheral neuropathy in patients with COVID-19.^[34]

The release of cytokines, which are tiny circulating peptides that act as mediators of the inflammatory response, is the primary host reaction in sepsis.^[35] The pathophysiology of the systemic inflammatory response syndrome is thought to be an imbalance between the pro- and anti-inflammatory effects of cytokines.^[36] Cytokines have wide-ranging impacts on signal transduction mechanisms and can hamper the sympathetic and PNSs on the regulation of HR.^[35] For instance, the cytokines of tumor necrosis factor-alpha, interleukin (IL)-1b, and IL-6 increase HR. However, they can also blunt HR responses to beta-adrenergic agonists.^[35]

IL-6 levels in the blood have been found to significantly associate with measures of reduced HRV in a variety of clinical situations, according to epidemiological research.^[36] The indices of reduced HRV in systemic inflammation showed the highest connection with IL-6 among the cytokines. Gholami et al.^[36] showed that on BALB/c mice, IL-6 receptor (gp130) is expressed in mouse atria, and incubation of isolated atria with recombinant IL-6 impaired the negative chronotropic response to cholinergic stimulation. Because cytokines could potentially blunt beta-adrenergic signaling, it has been proposed that cytokine overexpression and subsequent loss of beta-adrenergic responsiveness may contribute to the decrease in HRV during inflammation.^[37] As a result, there is insufficient evidence to support the hypothesis that impaired responsiveness to the beta-adrenergic system contributes to changes in HRV indexes during systemic inflammation.^[38] We did not measure plasma concentrations of catecholamines and cytokines in the study population. Further, studies are needed to elucidate whether COVID-19 affects the ANS in the same way as it affects the peripheral nervous system.

Study limitations

This study had several limitations, primarily that it was conducted in a single center, it was retrospective in design, and the female predominance may have affected the results. Second, anxiety disorders due to COVID-19 disease were not evaluated in the study and these may have contributed to the decrease in HRV values by increasing SNS activity. Third, we did not measure plasma concentrations of cytokines in the study population. Finally, coincident infectious diseases may have affected the results, so it is not possible to state definitively that the findings were specific to COVID-19.

Conclusion

The reversal in the changes of HRV parameters that occurred within the first 3 months following COVID-19 diagnosis may be an indicator of acute autonomic dysfunction due to COVID-19 infection.

Declaration of patient consent

The authors certify that they have obtained all appropriate patient consent forms. In the form, the patients have given their consent for images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.