Edited by: De-Pei Li, University of Missouri, United States
Reviewed by: Bruno Bonaz, Centre Hospitalier Universitaire de Grenoble, France; Jinwei Tian, The Second Affiliated Hospital of Harbin Medical University, China
†These authors have contributed equally to this work
This article was submitted to Autonomic Neuroscience, a section of the journal Frontiers in Physiology
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Previous studies suggest that coronavirus disease 2019 (COVID-19) is a systemic infection involving multiple systems, and may cause autonomic dysfunction.
To assess autonomic function and relate the findings to the severity and outcomes in COVID-19 patients.
We included consecutive patients with COVID-19 admitted to the 21st COVID-19 Department of the east campus of Renmin Hospital of Wuhan University from February 6 to March 7, 2020. Clinical data were collected. Heart rate variability (HRV), N-terminal pro-B-type natriuretic peptide (NT-proBNP),
All patients were divided into a mild group (13) and a severe group (21). The latter was further divided into two categories according to the trend of HRV. Severe patients had a significantly lower standard deviation of the RR intervals (SDNN) (
HRV was associated with the severity of COVID-19. The changing trend of HRV was related to the prognosis, indicating that HRV measurements can be used as a non-invasive predictor for clinical outcome.
Coronavirus disease 2019 (COVID-19) is a newly recognised infectious disease, and has rapidly spread worldwide. As of November 1, 2020, more than 40,000,000 cases have been confirmed throughout the world. The death toll has risen dramatically, with more than 1,000,000 deaths having been recorded. Early identification and intervention were crucial for reducing mortality.
The severity of COVID-19 is evaluated according to vital symptoms, physical signs, and physiological parameters. Although COVID-19 is primarily manifested as respiratory disease, many studies have shown that it is a systemic infection involving multiple systems. Previous studies have attempted to predict the severity and prognosis based on immune function, myocardial injury, and coagulation (
In this study, we enrolled 34 consecutive patients admitted to the 21st COVID-19 Department of the east campus of Renmin Hospital of Wuhan University, Wuhan, China, from February 6 to March 7, 2020. This designated hospital is responsible for the treatment of severe COVID-19 patients. The diagnosis of all patients was based on the Diagnosis and Treatment of New Coronavirus Pneumonia (trial version seven)
Detailed clinical data for all 34 patients were collected from electronic medical records by physicians who worked at Renmin Hospital of Wuhan University during the epidemic period, including demographic characteristics (sex and age), clinical symptoms and signs, comorbidities, laboratory examination results (routine blood parameters, coagulation function, myocardial injury markers, lymphocyte subpopulation, etc.), treatment and clinical outcome. Recovery was defined by discharge from the hospital. Viral RNA negative conversion was defined by negative results of at least two consecutive tests for SARS-CoV-2 nucleic acids during hospitalization. Severe patients also had a follow-up telephone call monthly after discharge by trained investigators. And psychological symptoms were assessed by the 2-item Generalized Anxiety Disorder Scale (GAD-2) and the Patient Health Questionnaire-2 (PHQ-2). The end of follow-up was July 26, 2020, 3 months from the last discharge. All the data were entered into the electronic database after deleting patients’ private information.
A Holter monitor (Synwing Tech, Chengdu, China) was used to record dynamic electrocardiogram (ECG) data over 24 consecutive hours for all patients. The data were collected into the electronic database, including the time, duration, subsection, count, and the kind of arrhythmia. Then, the HRV was analysed according to the 24 h dynamic ECG recordings. All the data calculated automatically by the computer were manually adjusted and verified by experienced physicians. The intervals with significant changes were excluded to prevent the effect of data mistakes. Time domains were calculated from 24 h dynamic ECG recordings including standard deviation of the RR intervals (SDNN, normal values: 141 ± 39 ms), standard deviation of the averages of NN intervals (SDANN, normal values: 127 ± 35 ms), square root of the mean of the sum of the squares of differences between adjacent NN interval (RMSSD, normal values: 27 ± 12 ms), and percent differences between adjacent NN intervals that are greater than 50 ms (pNN50, normal values: 16.7 ± 12.3%). Frequency-domain analysis included low-frequency power 0.04–0.15 Hz (LF, normal values: 300–1,750 ms2), high-frequency power 0.15–0.4 Hz (HF, normal values: 50–120 ms2), and the ratio of low- to high-frequency power (LF/HF, normal values: 1–3).
All categorical variables were compared by using the Pearson Chi-Square test or Fisher’s exact test, and all continuous variables were compared by using an unpaired t-test, a paired
The research population included consecutive hospitalised patients who were laboratory confirmed to be positive for COVID-19: 13 patients (38%) were classified as mild patients, and 21 patients (62%) were classified as severe. The mean age was 56.2 ± 16.0 years old, and 23 (68%) were female. Compared with mild patients, severe patients were older (
Clinical characteristics of all patients.
Variable | Total | Mild group | Severe group | |
( |
( |
( |
||
Sex (male), |
11 (32%) | 3 (23%) | 8 (38%) | 0.465 |
Age, median, year | 56.2 ± 16.0 | 47.5 ± 14.2 | 61.5 ± 15.0 | 0.011 |
Hospital length of stay | 24 (15.8–39.3) | 15 (13–16.5) | 36 (26.5–47) | <0.001 |
Time from onset to discharge | 46 (32.8–52.3) | 38 (22.5–45.5) | 50 (39–58) | 0.002 |
Days to initial HRV† | 8.5 (6.8–19) | 7 (5–7) | 17 (8.5–24.5) | <0.001 |
Days to nucleic acid test positive (corresponding to initial HRV) † | 9.5 (6.8–20) | 7 (5.5–8) | 18 (9.5–25.5) | <0.001 |
Days to follow-up HRV† | 18 (11.8–26.8) | 12 (11–14) | 26 (18–37) | <0.001 |
Days to nucleic acid test negative (corresponding to follow-up HRV) † | 17 (11–26.8) | 11 (10–12.5) | 25 (17.5–34) | <0.001 |
Days between initial and follow-up HRV | 7 (6–12.5) | 6 (4–7) | 10 (7–14) | 0.001 |
Fever | 26 (76 %) | 9 (69%) | 17 (81%) | 0.679 |
Cough | 15 (44%) | 7 (54%) | 8 (38%) | 0.484 |
Shortness of breath | 5 (15%) | 1 (8%) | 4 (19%) | 0.627 |
Fatigue | 7 (21%) | 2 (15%) | 5 (24%) | 0.682 |
Diarrhoea | 3 (9%) | 1 (8%) | 2 (10%) | 1.000 |
Chest pain | 5 (15%) | 2 (15%) | 3 (14%) | 1.000 |
Hypertension | 5 (14.71%) | 3 (23.08%) | 2 (9.52%) | 0.348 |
Diabetes | 1 (2.94%) | 1 (7.69%) | 0 | 0.382 |
Coronary heart disease | 2 (5.88%) | 1 (7.69%) | 1 (4.76%) | 1.000 |
Cerebrovascular disease | 1 (2.94%) | 0 | 1 (4.76%) | 1.000 |
Empirical antiviral therapy | 34 (100%) | 13 (100%) | 21 (100%) | 1.000 |
Empirical antimicrobial therapy | 23 (67.65%) | 8 (61.54%) | 15 (71.43%) | 0.709 |
Traditional Chinese medicine treatment | 31 (91.18%) | 13 (100%) | 18 (85.71%) | 1.000 |
Gut microbiota regulator | 28 (82.35%) | 11 (84.62%) | 17 (80.95%) | 1.000 |
The results of HRV indices were consistent with the severity of illness. The severe group had a significantly lower SDNN (
HRV measurements in mild and severe patients with COVID-19.
To assess the diagnostic value of the SDNN, SDANN, and LF/HF, ROC curve analysis was used. The ROC curve was above the diagonal, indicating good sensitivity and specificity. The areas under the curve (AUCs) were 0.767 (95% CI, 0.649–0.861), 0.772 (95% CI, 0.654–0.865), and 0.675 (95% CI, 0.550–0.784), respectively, in severe patients with COVID-19 (
ROC analysis for significant HRV variables. The sensitivity and specificity of SDNN, SDANN, and LF/HF for the severity of COVID-19.
Lymphocytes and lymphocyte subsets, NT-proBNP, and
The comparison of the
Our study revealed the relationship between HRV and
The correlation between HRV and NT-proBNP,
We analysed the changing trend of HRV between the two time points (nucleic acid test positive and negative) in severe patients. We further subdivided the patients into two groups according to the trend of LF/HF: Group A (LF/HF decreased,
The comparison of the HRV variables between the two time points (nucleic acid test positive and negative) in group A (blue) and group B (red). The
Furthermore, we analysed the trends of NT-proBNP, D-dimer, lymphocytes and lymphocyte subsets. In group A, NT-proBNP (
The comparison of the laboratory examination variables between the two time points (nucleic acid test positive and negative) in group A (blue) and group B (red). The
According to the Kaplan–Meier analysis, patients in group A had a shorter time to viral RNA negative conversion [median, 13 days, compared with 25 days; hazard ratio (HR), 2.36; 95% confidence interval (CI), 0.94–5.90;
Kaplan–Meier curve analysis for the proportion of viral RNA negative conversion
The severe patients were followed up for three months after discharge from the hospital. There was no positive RNA test after discharge in severe patients. Clinical features such as fever, cough, didn’t recur. However, 47, 21, and 26% of patients, respectively, still presented fatigue, shortness of breath, or chest stuffiness after exercise. About 32% of patients reported anxiety symptoms, and 11% reported depression symptoms (
In this study, we tried to predict severity and outcome with autonomic changes in patients with COVID-19. Besides, this study demonstrated novel evidence that HRV was associated with the severity of the disease. Indeed, severe patients tended to show more severe impaired HRV, which exhibited a linear correlation with NT-proBNP,
Alteration of autonomic nervous system is associated with severity and outcomes in patients with COVID-19. All patients were divided into a mild group and a severe group, and the severe group was then further divided into two categories according to the trend of HRV indices, which showed a consistent trend with immune function,
The autonomic nervous system plays a critical and complex role in maintaining the body’s balance (
Our research showed that severe patients had more severe autonomic dysfunction than mild patients, as indicated by the HRV analysis. Our results showed lower SDNN and SDANN and higher LF/HF in the severe group than in the mild group. The ROC curve illustrated the significant discriminatory power of these indices. This study was limited by a small sample size, which may explain the lack of statistical significance in other HRV indices, such as RMSSD and PNN50. Our results agreed with previous findings that COVID-19 patients with a more serious degree of disease or a poor outcome had more pronounced decreases in immune parameters, higher cardiac injury biomarkers, and more severe coagulation dysfunction than patients with milder disease (
The autonomic nervous system is closely related to many pathophysiological processes. Neurotransmitters produced by autonomic nerves interact with immune cells, including neutrophils, monocytes, and T cells, to regulate immunoreactions and inflammation (
Studies have shown that some patients with a detectable positive RNA test after discharge (
The cause for the more severe autonomic dysfunction in the severe group is not clear and might be multifactorial. Lung injury leading to hypoxemia, which can affect autonomic nerve activity (
Previous studies indicated that controlled cytokine release is imperatively linked to a well-balanced autonomic nervous system (
Evaluating the host response to SARS-Cov-2 infection as a complex system provides novel insights for predicting. Compared with other laboratory indicators, HRV is a widely accepted, noninvasive method for the evaluation of autonomic balance. Moreover, HRV-evaluation is cost-neutral and available for use under study- and clinical conditions. Our findings are preliminary because they are based on a single-center study with a small sample of patients with COVID-19. However, ECG recordings were carefully acquired and analysed to reduce errors introduced by the experimental method. Although confirming the predictive ability of HRV requires larger validation studies, this pilot study presents the first data to suggest that HRV may be a non-invasive marker for COVID-19.
This study suggests that patients with COVID-19 with autonomic dysfunction are more likely to have an increased severity of illness. The underlying mechanism of these findings for the prognosis of COVID-19 is unclear. However, this knowledge about HRV as a predictor of severity and outcome is important for monitoring disease progression and assessing treatment effects. Further study recruitment may shed more light on the predictive ability of HRV. There is no single tool for diagnosis or prediction in COVID-19. As a non-invasive modality, HRV biomarkers can be combined with other clinical predictors to monitor disease conditions and estimate prognosis, which may avoid alternative medical prescription and intervention. Besides, HRV may also guide the treatment of patients with autonomic dysfunction. Vagal tone restore therapy may help rebalance the autonomic nervous system of patients, which shows a potential application value in COVID-19 treatment.
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
The studies involving human participants were reviewed and approved by this study was reviewed and approved by the Medical Ethical Committee of Renmin Hospital of Wuhan University (approval number WDRY2020-K079). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.
HJ and LY conceptualised the study. YP, ZY, YY, and JH designed the study. YP, YY, ZhuW, HC, SW, ZheW, and HH recruited patients and collected and processed samples. ZY, JH, LZ, YL, ZZ, YW, and GM performed the statistical analyses. YP, ZY, YY, and JH co-wrote the manuscript. All authors read and approved the final version of the manuscript.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The Supplementary Material for this article can be found online at:
heart rate variability
N-terminal pro-B-type natriuretic peptide
standard deviation of the RR intervals
standard deviation of the averages of NN intervals
square root of the mean of the sum of the squares of differences between adjacent NN interval
percent differences between adjacent NN intervals that are greater than 50 ms
low-frequency power
high-frequency power
ratio of low- to high-frequency power
2-item Generalized Anxiety Disorder Scale
Patient Health Questionnaire-2
electrocardiogram
receiver operating characteristic.
The Diagnosis and Treatment of New Coronavirus Pneumonia (trial version seven), National Health Commission of China, New coronavirus pneumonia prevention and control program (7th edn).