Abstract
Background
Inflammatory biomarkers, including immune cells appear associated with reduced heart rate variability (HRV), an indicator of cardiac autonomic nervous dysfunction. However, the influence of leukocytes on HRV and their relation across sex and glucose metabolism status remains unclear. Hence, we sought to investigate the cross‐sectional association between leukocytes and HRV indices and their association across sex and glucose metabolism status in a general population.
Methods
We retrieved cross‐sectional data of 1888 participants of the Maastricht Study who completed the baseline survey between November 2010 and December 2017. Blood leukocyte numbers (including neutrophils, monocytes, basophils, eosinophils, and lymphocytes) were determined with an automated hematology analyzer. Qualitative changes within lymphocytes were assessed with flow cytometry. HRV indices were assessed using a 24‐hour ECG. Time‐domain and frequency‐domain HRV index average scores were calculated. We regressed HRV average scores on standardized leukocytes and adjusted the models for several covariates.
Results
In fully adjusted models, higher neutrophil (−0.07 [−0.11 to −0.03]; false discovery rate (FDR)‐corrected P value, P FDR=0.006) and monocyte (−0.05 [−0.09 to −0.02]; P FDR=0.020) counts were associated with a lower time‐domain average score. Higher neutrophil (−0.07 [−0.11 to −0.03]; P FDR=0.006) and monocyte (−0.05 [−0.09 to −0.01]; P FDR=0.020) counts were also associated with a lower frequency‐domain HRV index average score. No effect modification by either sex or glucose metabolism status was found (P intFDR>0.05). There were no other associations between immune cells and HRV indices.
Conclusions
Higher neutrophil and monocyte counts were associated with lower HRV, irrespective of sex and glucose metabolism status. This suggests a role of these cells reflecting innate immune processes and inflammatory response in cardiac autonomic nervous dysfunction.
Keywords: heart rate variability indices, immune cells, lymphocytes, monocytes, neutrophils
Subject Categories: Epidemiology
Nonstandard Abbreviations and Acronyms
- HRV
heart rate variability
- LF
power in the low frequency range
- NGM
normal glucose metabolism
- PSNS
vagal parasympathetic nervous system
- SDANN
SD of the average NN intervals in all 5‐min segments of the entire recording
- SDNN
SD of normal‐to‐normal (NN) intervals
- SNS
sympathetic nervous system
- TP
total power
- ULF
power in the ultra‐low frequency range
- VLF
power in the very‐low frequency range
Clinical Perspective.
What Is New?
The influence of leukocytes on heart rate variability and their relation across sex and glucose metabolism status remains unclear.
This study observed a relationship between systemic inflammation, driven by neutrophils and monocytes, and reduced cardiac autonomic function, regardless of sex or glucose metabolism status.
What Are the Clinical Implications?
These findings suggest the potential role of innate immune cells on the heart through autonomic pathways.
Heart rate variability (HRV), the variation of time intervals between heartbeats, 1 is a noninvasive measure of the interplay between the sympathetic (SNS) and vagal parasympathetic nervous system (PSNS) that innervate the sinoatrial node. 2 Higher HRV indicates a balanced interaction between these 2 autonomic systems and provides insight into cardiac autonomic flexibility and the heart’s responsiveness to environmental changes and demands. In contrast, reduced HRV indicates a decrease in PSNS or an increase in SNS activity, pointing toward a decreased ability of the heart to handle stress. Reduced HRV is a validated marker of cardiac autonomic dysfunction with established evidence from population‐based studies associating it with increased risk of incident cardiovascular diseases and cardiovascular mortality. 3
It is well known that the pathophysiological perturbation of HRV includes inflammatory processes and immune dysregulation. 4 Epidemiological evidence suggests that inflammatory biomarkers, including acute phase proteins such as C‐reactive protein, 5 cytokines and adipokines such as adiponectin, 6 and proteomics biomarkers such as C‐C motif chemokine 23 and macrophage colony‐stimulating factor 1 7 are independently associated with HRV indices. White blood cells (leukocytes) are integral to cell‐mediated inflammation and immune responses. 8 Neutrophils, monocytes, macrophages, eosinophils, basophils, and mast cells are involved in the nonspecific innate immune response, while lymphocytes are crucial to the exquisitely specific adaptive immune response. 9
Further, type 2 diabetes (T2D) is a determinant of reduced HRV in the general population. 10 , 11 Evidence from population‐based epidemiological studies such as the KORA (Cooperative Health Research in the Region of Augsburg) study and the Maastricht Study also implicated prediabetes as an important risk factor of reduced HRV. 12 , 13 In individuals with T2D, reduced HRV presents the earliest sign of subclinical cardiac autonomic nervous dysfunction. In addition, it has been reported that sex‐related differences exist in HRV and its various cardiovascular risk factors. 10 In fact, a systematic review and meta‐analysis reported greater vagal activity despite a higher heart rate in healthy women as compared with healthy men. 14 The population‐based MONICA (Monitoring of Trends and Determinants in Cardiovascular Disease) Augsburg survey also showed differential risk factors of reduced HRV between men and women. 10
Despite their pathophysiological relevance and routine measurement, the association of leukocytes and their subsets with HRV in population‐based epidemiological studies remains incompletely understood. Moreover, comparative investigation of the relation of leukocytes and HRV across sex and glucose metabolism status is scarce.
Therefore, we aimed to examine the associations of leukocyte subsets with HRV indices in a population‐based epidemiological study enriched with people with T2D. Secondly, we sought to determine whether these associations vary by either sex or glucose metabolism status.
METHODS
The data of this study derive from the Maastricht Study, but restrictions apply to the availability of these data, which were used under license for the current study. Data are, however, available from the authors upon reasonable request and with permission of the Maastricht Study management team.
Study Design and Study Participants
We used data from the Maastricht Study, an observational prospective population‐based cohort study. The rationale and methodology have been described previously. 15 In brief, the study focuses on the cause, pathophysiology, complications, and comorbidities of T2D and is characterized by an extensive phenotyping approach. Eligible for participation were all individuals aged between 40 and 75 years and living in the southern part of the Netherlands. Participants were recruited through mass media campaigns and from the municipal registries and the regional Diabetes Patient Registry via mailings. Recruitment was stratified according to known T2D status, with an oversampling of individuals with T2D, for reasons of efficiency. The examinations of each participant were performed within a time window of 3 months. The study was approved by the institutional medical ethical committee (NL31329.068.10) and the Minister of Health, Welfare and Sports of the Netherlands (Permit 131088‐105234‐PG) and was conducted in accordance with the Declarations of Helsinki. All participants gave written informed consent. The present analysis considered the cross‐sectional data from the 3807 participants, who completed the baseline survey between November 2010 and December 2017.
Assessment of Glucose Metabolism Status
In order to determine glucose metabolism status, all participants (except those using insulin) underwent a standardized 2‐hour 75‐g oral glucose tolerance test after an overnight fast. For safety reasons, participants with a fasting glucose level >11.0 mmol/L, as determined by finger prick, did not undergo the oral glucose tolerance test. For these individuals, fasting glucose level and information about diabetes medication were used to determine glucose metabolism status. 13 Glucose metabolism status was defined according to the World Health Organization 2006 criteria as normal glucose metabolism (NGM), prediabetes (impaired fasting glucose or impaired glucose tolerance), and T2D. 16
Study Population
A total of 2464 out the 3807 participants had HRV data, and 1888 out these 2464 participants had leukocyte data (Figure 1). Of this final study population, 408 had data on lymphocyte subsets (Figure 1). The lack of HRV data in the excluded participants was due to nonavailable 24‐hour ECG because of logistic issues, ECG recording <18 hour or arrhythmia. Lack of leukocyte data in the excluded participants was mainly caused by logistic issues that prevented the analysis of fresh blood, not by any selection criteria. Overall, the current study population comprised 1888 individuals.
Figure 1. Flow chart of the study population.
Measurement of Leukocytes
Automated leukocyte counts from whole fresh heparinized venous blood were performed on the Sysmex XE5000 (Sysmex, Kobe, Japan). Monocytes, basophils, eosinophils, neutrophils, and lymphocytes counts were reported as absolute numbers. In addition, flow cytometry analysis of lymphocyte subsets was performed in a subset of the study participants, as described previously. 17 Results were expressed as percentages based on the corresponding parent gate (total lymphocytes‐monocytes as parent gate for B and total T cells, total T cells as parent gate for CD4+ and CD8+ T cells, and CD4+ T cells as parent gate for regulatory T cells).
Measurement of Heart Rate Variability Indices
Linear HRV time‐ and frequency‐domain indices were derived from 24‐hour ECG recorded with 12‐lead Holter system (Fysiologic ECG Services, Amsterdam, the Netherlands) as previously described. 15 The minimum duration of ECG recording was 18 hours after exclusion of nonsinus cardiac cycles. The time‐domain and frequency‐domain indices were extracted. The selected time‐domain indices were SD of normal‐to‐normal (NN) intervals (SDNN, ms), SD of the average NN intervals in all 5‐minute segments of the entire recording (SDANN, ms), mean of the SD of all the NN intervals in all 5‐minute segments of the entire recording (SDNN index, ms), root mean square of successive NN interval differences, and percentage of successive NN intervals that differ by >50 ms (%), and the selected frequency‐domain indices were total power (TP, ms2), power in the ultra‐low‐frequency range (ULF, ms2), power in the very‐low‐frequency range (VLF, ms2), power in the low‐frequency range (LF, ms2), and power in the high‐frequency range (ms2).
Measurement of Covariates
Self‐administered questionnaires were used to assess sociodemographic and lifestyle factors such as age, sex, educational level (low, intermediate, and high), smoking behavior (never, former, and current), alcohol consumption (none: 0 glass/week; low: ≤7 glasses/week for women and ≤14 glasses/week for men, and high: >7 glasses/week for women and >14 glasses/week for men) and history of cardiovascular disease, cardiovascular disease (history of myocardial or cerebrovascular infarction or hemorrhage, or vascular surgery on the coronary, abdominal, peripheral, or carotid arteries). Physical activity (total activity hours per week) was assessed by a questionnaire and accelerometer (mean stepping time per hour). Information on use of medications such as blood pressure‐lowering drugs, lipid‐lowering drugs, and nonsteroidal anti‐inflammatory drugs (NSAIDs) was collected during an interview. Blood‐pressure lowering drugs include β‐blocking agents, calcium channel blockers, and agents acting on the renin angiotensin system and nonloop diuretics. Body mass index (BMI), glycated hemoglobin, systolic blood pressure (SBP), total cholesterol, high‐density lipoprotein cholesterol, and triglycerides were measured. Estimated glomerular filtration rate (in mL/min per 1.73 m2) was calculated with the Chronic Kidney Disease Epidemiology Collaboration equation based on both serum creatinine and serum cystatin C. 18
Statistical Analysis
Continuous and categorical basic characteristics were summarized as mean±SD and numbers (percentages), respectively. Differences between sexes were assessed with Kruskal–Wallis test (for continuous variables) or chi‐square test (for categorical variables).
For each participant, Z‐scores (mean of 0 and SD of 1) were calculated for each time‐ and frequency‐domain HRV index and combined into time‐domain HRV index average score ([SDNNz‐score+root mean square of successive NN interval differencesZ‐score+SDANNz‐score+SDNN indexZ‐score+percentage of successive NN intervals that differ by >50 ms Z‐score]/5) and frequency‐domain HRV index average score ([TPZ‐score+ULFZ‐score+VLFZ‐score+LFz‐score+power in the high‐frequency rangeZ‐score]/5). These average scores are not new indices, but composite of continuous HRV variables. We used this composite of continuous HRV variables, because assuming that the immune cells have similar associations with single indices within the same domain, these variables would increase statistical power and reduce multiple testing. The main predictors, that is, leukocytes, were also Z‐score standardized. Included in the main predictor variables was the neutrophil‐to‐lymphocyte ratio (NLR), which is the conjugate of the innate and adaptive immune responses. 19
The associations between each leukocyte subsets with time‐ and frequency‐domain HRV index average scores (outcomes) were assessed with linear regression models. We fitted 3 models with increasing complexity: model 1, crude. model 2 was additionally adjusted for age (years), sex (men versus women), and glucose metabolism status (NGM, prediabetes, and T2D). Model 3 was additionally adjusted for BMI (kg/m2), glycated hemoglobin (mmol/L), SBP (mm Hg), total cholesterol/high‐density lipoprotein cholesterol‐cholesterol, triglycerides (mmol/L), estimated glomerular filtration rate (mL/min per 1.73 m2), cardiovascular disease (yes/no), education (low, intermediate, high), physical activity (hours/week), smoking behavior (never, former, current), alcohol consumption (none, low, high), blood pressure‐lowering medication (yes/no), lipid‐lowering medication (yes/no), and nonsteroidal anti‐inflammatory drugs (yes/no). Missing values in covariates were imputed using R “mice” package using 20 imputed data sets. All main predictor variables and covariates were used for imputation. The predictive mean matching, proportional odds model, and polytomous logistic regression were used to impute the missing values of continuous covariates, ordered categorical covariates, and nominal categorical covariates, respectively. The pooled effect estimates of the main predictor variables were reported. The effect estimates are standardized beta coefficients (β) interpreted as the number of SD units change in the average scores per 1‐SD increase in the main predictor variables.
To assess the effect modification by sex and glucose metabolism status (NGM versus prediabetes/T2D [non‐NGM] as well as NGM/prediabetes [non‐T2D] versus T2D), we performed stratified analysis across sex and glucose metabolism status. We also formally tested the statistical significance of interaction by including 2‐way interaction terms between sex or glucose metabolism status and each main predictor variable in model 3.
In sensitivity analyses, we assessed the influence of HR (beats/min) on the results by additionally adjusting for HR in model 3. Furthermore, for main predictor variables that were significantly associated with the HRV average scores, we examined their associations with individual HRV index Z‐scores with covariate adjustment similar to the main analysis. We used the Benjamini–Hochberg procedure to obtain false discovery rate‐corrected P values, P (P false discovery rate). P<0.05 indicated statistical significance. All analyses were performed with R software (version: 4.2.2).
RESULTS
Characteristics of the Study Population
The characteristics of the study population (n=1888) are shown in Table 1. The mean age was 60 years and there were 49% women. Study participants had a mean BMI of 27 kg/m2, 13% were smokers, and 28% had T2D. Compared with women, men were older, had higher BMI and SBP, were more likely to have T2D, had lower HR, higher prevalence of cardiovascular disease, and were more likely to use blood pressure‐lowering drugs. Additionally, men had lower counts of monocytes, eosinophils, and neutrophils but higher counts of basophils and lymphocytes. They also had lower root mean square of successive NN interval differences, SDNN index, VLF, and LF, but higher percentage of successive NN intervals that differ by >50 ms than women (Table 2). The characteristics of the subsample with lymphocyte subsets (n=408) were similar to the overall study population (Table S1). Histogram of the HRV indices in the overall study population and subsample with lymphocyte subsets are displayed in Figures S1 through S4.
Table 1.
Basic Characteristics of the Study Population (n=1888)
| All (n=1888) | Men (n=964) | Women (n=924) | P value | |
|---|---|---|---|---|
| Age, y | 60.0±8.1 | 61.1±7.9 | 58.9±8.2 | <0.001 |
| Body mass index, kg/m2 | 26.9±4.4 | 27.6±4.1 | 26.1±4.7 | <0.001 |
| Waist circumference, cm | 95.4±13.6 | 101.1±11.9 | 89.5±12.7 | <0.001 |
| Glucose metabolism status | ||||
| Normal glucose metabolism | 1050 (55.6) | 442 (45.9) | 608 (65.8) | <0.001 |
| Prediabetes | 279 (14.8) | 150 (15.6) | 129 (14.0) | |
| Type 2 diabetes | 532 (28.2) | 359 (37.2) | 173 (18.7) | |
| Other diabetes types | 27 (1.4) | 13 (1.3) | 14 (1.5) | |
| Glycated hemoglobin, mmol/mol | 41.1±10.4 | 42.8±11.5 | 39.4±8.8 | <0.001 |
| Systolic blood pressure, mm Hg | 134.8±17.8 | 139.3±16.8 | 130.1±17.7 | <0.001 |
| Diastolic blood pressure, mm Hg | 76.2±9.6 | 78.4±9.4 | 73.9±9.2 | <0.001 |
| Total cholesterol, mmol/L | 5.2±1.2 | 4.9±1.1 | 5.5±1.1 | <0.001 |
| High‐density lipoprotein cholesterol, mmol/L | 1.6±0.5 | 1.4±0.4 | 1.8±0.5 | <0.001 |
| Low‐density lipoprotein cholesterol, mmol/L | 3.0±1.0 | 2.9±1.0 | 3.2±1.0 | <0.001 |
| Triglycerides, mmol/L | 1.4±0.9 | 1.5±0.9 | 1.3±0.8 | <0.001 |
| Estimated glomerular filtration rate, mL/min per 1.73 m2 | 84.3±13.9 | 84.2±13.9 | 84.4±13.8 | 0.691 |
| Heart rate (beats/min) | 67.9±11.0 | 65.7±11.3 | 70.1±10.3 | <0.001 |
| Education | 0.001 | |||
| Low | 635 (33.6) | 286 (29.7) | 349 (37.8) | |
| Intermediate | 520 (27.5) | 272 (28.2) | 248 (26.8) | |
| High | 733 (38.8) | 406 (42.1) | 327 (35.4) | |
| Smoking behavior | ||||
| Never | 664 (35.2) | 298 (30.9) | 366 (39.6) | 0.001 |
| Former | 986 (52.2) | 540 (56.0) | 446 (48.3) | |
| Current | 238 (12.6) | 126 (13.1) | 112 (12.1) | |
| Alcohol consumption | ||||
| None | 362 (19.2) | 133 (13.8) | 229 (24.8) | <0.001 |
| Low | 1050 (55.6) | 615 (63.8) | 435 (47.1) | |
| High | 476 (25.2) | 216 (22.4) | 260 (28.1) | |
| Physical activity, h/wk | 5.6±4.5 | 5.2±4.4 | 5.9±4.6 | 0.002 |
| Cardiovascular disease, yes | 318 (16.8) | 187 (19.4) | 131 (14.2) | 0.003 |
| Blood pressure‐lowering drugs, yes | 762 (40.4) | 462 (47.9) | 300 (32.5) | <0.001 |
| Lipid‐lowering drugs, yes | 681 (36.1) | 440 (45.6) | 241 (26.1) | <0.001 |
| Nonsteroidal anti‐inflammatory drugs, yes | 170 (9.0) | 75 (7.8) | 95 (10.3) | 0.069 |
Continuous variables are given as mean±SD or median (Q1, Q3). Categorical variables are given as numbers and percentages (%). Alcohol consumption (none: 0 glass/wk; low: ≤7 glasses/wk for women and ≤14 glasses/wk for men, and high: >7 glasses/wk for women and >14 glasses/wk for men) and history of cardiovascular disease (history of myocardial or cerebrovascular infarction or hemorrhage, or vascular surgery on the coronary, abdominal, peripheral, or carotid arteries). Physical activity (total activity hours per week) was assessed by a questionnaire and accelerometer (mean stepping time per hour).
Table 2.
Leukocyte and Heart Rate Variability Indices in the Study Population (n=1888)
| All (n=1888) | Men (n=964) | Women (n=924) | P value | |
|---|---|---|---|---|
| Leukocytes | ||||
| Total, 109/L | 5.80±1.74 | 5.96±1.84 | 5.63±1.61 | <0.001 |
| Neutrophils, 109/L | 3.28±1.21 | 3.39±1.18 | 3.16±1.24 | <0.001 |
| Lymphocytes, 109/L | 1.83±0.87 | 1.82±1.08 | 1.83±0.52 | <0.001 |
| Monocytes, 109/L | 0.49±0.16 | 0.53±0.17 | 0.45±0.14 | <0.001 |
| Basophils, 109/L | 0.04±0.03 | 0.03±0.03 | 0.04±0.03 | 0.013 |
| Eosinophils, 109/L | 0.18±0.12 | 0.20±0.13 | 0.17±0.11 | <0.001 |
| Lymphocyte subset* | ||||
| T cells, % | 59.7±9.0 | 57.7±9.1 | 61.7±8.5 | <0.001 |
| CD4+ T cells, % | 64.6±11.8 | 62.5±12.5 | 66.7±10.7 | <0.001 |
| CD8+ T cells, % | 30.2±10.6 | 31.8±11.1 | 28.7±9.8 | 0.003 |
| Regulatory T cells, % | 7.8±2.0 | 8.1±2.2 | 7.6±1.8 | 0.01 |
| B cells, % | 15.1±7.0 | 14.9±7.2 | 15.3±6.8 | 0.528 |
| Neutrophil‐to‐lymphocyte ratio | 1.9±0.9 | 2.1±1.0 | 1.8±0.8 | <0.001 |
| HRV, time domains | ||||
| Time domain average score | −0.13 (−0.56, 0.40) | −0.11 (−0.57, 0.45) | −0.13 (−0.56, 0.34) | 0.127 |
| SDNN, ms | 130.9 (107.5, 156.1) | 131.7 (107.3, 157.9) | 129.6 (107.8, 154.9) | 0.348 |
| SD of the averages of normal‐to‐normal intervals in all 5‐min segments of the entire recording, ms | 118.0 (94.5, 142.6) | 118.0 (92.7, 142.9) | 118.1 (97.3, 142.2) | 0.41 |
| Root mean square of successive normal‐to‐normal interval differences, ms | 25.2 (19.3, 34.1) | 25.5 (19.2, 34.6) | 25.0 (19.3, 33.7) | <0.001 |
| SDNN index, ms | 50.6 (41.4, 62.7) | 51.9 (42.2, 65.3) | 49.3 (40.1, 59.3) | <0.001 |
| Percentage of successive normal‐to‐normal intervals that differ by more than 50 ms, % | 6.1 (2.6, 12.1) | 5.9 (2.6, 11.7) | 6.2 (2.7, 12.5) | 0.039 |
| HRV, frequency domains | ||||
| Frequency domain average score | −0.18 (−0.54, 0.32) | −0.17 (−0.54, 0.39) | −0.20 (−0.55, 0.25) | 0.044 |
| Total power, ms2 | 11 348 (7650, 16 422) | 11 471 (7576, 16 770) | 11 231 (7809, 15 931) | 0.233 |
| Power in the ultra‐LF range, ms2 | 9632 (6268, 13 967) | 9622 (6039, 14 027) | 9642 (6616, 13 921) | 0.876 |
| Power in the very LF range, ms2 | 1054 (705, 1546) | 1156 (782, 1763) | 948 (653, 1388) | <0.001 |
| LF, ms2 | 337 (202, 575) | 375 (211, 664) | 308 (194, 502) | <0.001 |
| Power in the high‐frequency range, ms2 | 83.3 (47.0, 144.4) | 74.8 (43.4, 134.6) | 92.2 (50.4, 155.5) | 0.071 |
Continuous variables are given as mean±SD or median (Q1, Q3). Categorical variables are given as numbers and percentages (%).
HRV indicates heart rate variability; LF, power in the low frequency range; and SDNN, SD of all normal‐to‐normal intervals.
Available for 408 participants.
Associations of Leukocyte Subsets With HRV Average Scores
Table 3 shows the association of leukocyte subsets with time‐domain and frequency‐domain HRV index average scores. Every 1‐SD higher neutrophil and monocyte counts were associated with lower time‐ and frequency‐domain average scores. These inverse associations were attenuated by covariate adjustment, but remained significant across all models. After full covariate adjustment, the β for neutrophil count was −0.07 (95% CI, −0.11 to −0.03) for both time‐domain and frequency‐domain HRV index average scores, and that for monocyte count was −0.05 (95% CI, −0.09 to −0.02) for the time‐domain HRV index average score and −0.05 (95% CI, −0.09 to −0.01) for the frequency‐domain HRV index average score. Lymphocytes and NLR also showed inverse associations with both time‐domain and frequency‐domain HRV index average scores, but these associations were no longer statistically significant following covariate adjustment. There were no associations of basophil and eosinophil count with HRV index average scores in any model.
Table 3.
Association of Leukocyte With Heart Rate Variability Index Average Scores (n=1888)
| Leukocyte (absolute values, per 1‐SD) | Heart rate variability index | Model 1 | Model 2 | Model 3 | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| β (95% CI) | P value | P FDR | β (95% CI) | P value | P FDR | β (95% CI) | P | P FDR | ||
| Neutrophils |
Time‐domain average score |
−0.12 (−0.16 to −0.09) |
<0.001 | <0.001 |
−0.10 (−0.13 to −0.06) |
<0.001 | <0.001 |
−0.07 (−0.11 to −0.03) |
0.001 | 0.006 |
| Frequency‐domain average score |
−0.13 (−0.16 to −0.09) |
<0.001 | <0.001 |
−0.10 (−0.14 to −0.06) |
<0.001 | <0.001 |
−0.07 (−0.11 to −0.03) |
0.001 | 0.006 | |
| Lymphocytes |
Time‐domain average score |
−0.05 (−0.09 to −0.02) |
0.004 | 0.001 |
−0.05 (−0.08 to −0.01) |
0.015 | 0.030 |
−0.02 (−0.06 to 0.02) |
0.273 | 0.41 |
| Frequency‐domain average score |
−0.05 (−0.09 to −0.02) |
0.005 | 0.010 |
−0.04 (−0.08 to −0.01) |
0.015 | 0.03 |
−0.02 (−0.06 to 0.01) |
0.253 | 0.41 | |
| Monocytes |
Time‐domain average score |
−0.09 (−0.13 to −0.06) |
<0.001 | <0.001 |
−0.08 (−0.12 to −0.04) |
<0.001 | <0.001 |
−0.05 (−0.09 to −0.02) |
0.006 | 0.020 |
| Frequency‐domain average score |
−0.10 (−0.13 to −0.06) |
<0.001 | <0.001 |
−0.08 (−0.12 to −0.04) |
<0.001 | <0.001 |
−0.05 (−0.09 to −0.01) |
0.007 | 0.02 | |
| Basophils |
Time‐domain average score |
−0.01 (−0.05 to 0.03) |
0.549 | 0.599 |
−0.004 (−0.04 to 0.03) |
0.838 | 0.914 |
0.001 (−0.04 to 0.04) |
0.971 | 0.971 |
| Frequency‐domain average score |
−0.01 (−0.05 to 0.03) |
0.611 | 0.611 |
−0.001 (−0.04 to 0.03) |
0.943 | 0.943 |
0.005 (−0.03 to 0.04) |
0.795 | 0.867 | |
| Eosinophils |
Time‐domain average score |
−0.03 (−0.07 to 0.01) |
0.087 | 0.104 |
−0.02 (−0.06 to 0.02) |
0.250 | 0.311 |
−0.01 (−0.05 to 0.03) |
0.613 | 0.736 |
| Frequency‐domain average score |
−0.04 (−0.07 to 0.002) |
0.063 | 0.084 |
−0.02 (−0.06 to 0.01) |
0.198 | 0.297 |
−0.01 (−0.05 to 0.03) |
0.581 | 0.736 | |
| Neutrophil‐to‐lymphocyte ratio |
Time‐domain average score |
−0.04 (−0.08 to −0.003) |
0.034 | 0.051 |
−0.02 (−0.06 to 0.02) |
0.259 | 0.311 |
−0.02 (−0.06 to 0.01) |
0.219 | 0.410 |
| Frequency‐domain average score |
−0.05 (−0.09 to −0.01) |
0.007 | 0.012 |
−0.03 (−0.07 to 0.01) |
0.103 | 0.177 |
−0.03 (−0.06 to 0.01) |
0.112 | 0.269 | |
Model 1: crude. Model 2: Model 1 adjusted for age (y), sex (men vs women), and glucose metabolism status (normal glucose metabolism, prediabetes, and type 2 diabetes). Model 3: Model 2 adjusted for body mass index (kg/m2), glycated hemoglobin (mmol/mol), systolic blood pressure (mm Hg), total cholesterol/high‐density lipoprotein cholesterol, triglycerides (mmol/L), estimated glomerular filtration rate (mL/min per 1.73 m2), cardiovascular disease (yes/no), education (low, intermediate, high), physical activity (h/wk), smoking behavior (never, former, current), alcohol consumption (none, low, high), blood pressure‐lowering medication (yes/no), lipid‐lowering medication (yes/no), and nonsteroidal anti‐inflammatory drugs (yes/no).
P FDR indicates false discovery rate‐corrected P value using Benjamini and Hochberg.
Table 4 illustrates the association of lymphocyte subsets with time‐domain and frequency‐domain HRV index average scores. Neither the crude models nor the adjusted models showed associations of total T cells, T‐cell subpopulations (CD4+, CD8+, and regulatory T), and B cells with HRV index average scores.
Table 4.
Association of Lymphocyte Subsets With Heart Rate Variability Index Average Scores (n=408)
| Lymphocyte subsets (%, per 1‐SD) | Heart rate variability index | Model 1 | Model 2 | Model 3 | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| β (95% CI) | Pvalue | P FDR | β (95% CI) | P value | P FDR | β (95% CI) | P | P FDR | ||
| T cells |
Time‐domain average score |
0.03 (−0.05 to 0.10) |
0.483 | 0.531 |
0.03 (−0.04 to 0.11) |
0.391 | 0.877 |
0.05 (−0.03 to 0.13) |
0.22 | 0.663 |
|
Frequency‐domain average score |
0.04 (−0.04 to 0.11) |
0.338 | 0.531 |
0.04 (−0.03 to 0.11) |
0.288 | 0.877 |
0.05 (−0.02 to 0.12) |
0.188 | 0.663 | |
| CD4+ T cells |
Time‐domain average score |
−0.03 (−0.11 to 0.04) |
0.398 | 0.531 |
−0.01 (−0.08 to 0.07) |
0.866 | 0.911 |
0.01 (−0.06 to 0.09) |
0.746 | 0.933 |
|
Frequency‐domain average score |
−0.04 (−0.12 to 0.03) |
0.246 | 0.531 |
−0.01 (−0.09 to 0.06) |
0.689 | 0.911 |
0.001 (−0.07 to 0.07) |
0.982 | 0.982 | |
| CD8+ T cells |
Time‐domain average score |
0.03 (−0.05 to 0.10) |
0.502 | 0.531 |
−0.01 (−0.08 to 0.07) |
0.903 | 0.911 |
−0.02 (−0.09 to 0.06) |
0.681 | 0.933 |
|
Frequency‐domain average score |
0.04 (−0.04 to 0.11) |
0.307 | 0.531 |
0.004 (−0.07 to 0.08) |
0.911 | 0.911 |
−0.01 (−0.08 to 0.07) |
0.896 | 0.983 | |
| Regulatory T cells |
Time‐domain average score |
−0.03 (−0.11 to 0.05) |
0.431 | 0.531 |
−0.03 (−0.11 to 0.05) |
0.439 | 0.877 |
−0.05 (−0.12 to 0.03) |
0.236 | 0.633 |
|
Frequency‐domain average score |
−0.04 (−0.11 to 0.04) |
0.334 | 0.531 |
−0.03 (−0.10 to 0.04) |
0.364 | 0.877 |
−0.04 (−0.11 to 0.03) |
0.265 | 0.633 | |
| B cells |
Time‐domain average score |
−0.03 (−0.10 to 0.05) |
0.531 | 0.531 |
−0.02 (−0.10 to 0.05) |
0.526 | 0.877 |
−0.02 (−0.09 to 0.06) |
0.683 | 0.933 |
|
Frequency‐domain average score |
−0.03 (−0.10 to 0.04) |
0.404 | 0.531 |
−0.03 (−0.10 to 0.04) |
0.397 | 0.877 |
−0.02 (−0.09 to 0.05) |
0.504 | 0.933 | |
Model 1: crude. Model 2: Model 1 adjusted for age (y), sex (men vs women), and glucose metabolism status (normal glucose metabolism, prediabetes, and type 2 diabetes). Model 3: Model 2 adjusted for body mass index (kg/m2), glycated hemoglobin (mmol/mol), systolic blood pressure (mm Hg), total cholesterol/high‐density lipoprotein cholesterol, triglycerides (mmol/L), estimated glomerular filtration rate (mL/min per 1.73 m2), cardiovascular disesae (yes/no), education (low, intermediate, high), physical activity (h/wk), smoking behavior (never, former, current), alcohol consumption (none, low, high), blood pressure‐lowering medication (yes/no), lipid‐lowering medication (yes/no), and nonsteroidal anti‐inflammatory drugs (yes/no).
P FDR indicates false discovery rate‐corrected P value using Benjamini and Hochberg.
Table 5 shows that the association of monocytes with the time‐domain HRV index average scores and of NLR with both HRV index average scores may be modified by sex (P for interaction <0.05). The inverse associations seem more pronounced in women than in men, but these effect modifications were nonsignificant after correction for multiple testing (all P false discovery rate for interaction >0.05).
Table 5.
Sex‐Specific Associations of Leukocyte and Lymphocyte Subsets With Heart Rate Variability Index Average Scores
| Sex | Time‐domain average score | Frequency‐domain average score | ||||||
|---|---|---|---|---|---|---|---|---|
| β (95% CI) | P int | P intFDR | β (95% CI) | P int | P intFDR | |||
| Leukocytes (absolute values, per 1‐SD) | Neutrophils | Men (n=964) | −0.04 (−0.11 to 0.02) | 0.097 | 0.356 | −0.05 (−0.11 to 0.02) | 0.25 | 0.723 |
| Women (n=924) | −0.12 (−0.17 to −0.07) | −0.11 (−0.16 to −0.06) | ||||||
| Lymphocytes | Men (n=964) | −0.02 (−0.06 to 0.03) | 0.894 | 0.908 | −0.02 (−0.06 to 0.03) | 0.916 | 0.916 | |
| Women (n=924) | −0.03 (−0.10 to 0.04) | −0.03 (−0.11 to 0.04) | ||||||
| Monocytes | Men (n=964) | −0.03 (−0.08 to 0.03) | 0.041 | 0.226 | −0.03 (−0.08 to 0.03) | 0.141 | 0.723 | |
| Women (n=924) | −0.11 (−0.16 to −0.06) | −0.10 (−0.15 to −0.04) | ||||||
| Basophils | Men (n=964) | 0.0006 (−0.06 to 0.06) | 0.894 | 0.908 | −0.002 (−0.06 to 0.05) | 0.644 | 0.857 | |
| Women (n=924) | 0.01 (−0.04 to 0.05) | 0.02 (−0.03 to 0.06) | ||||||
| Eosinophils | Men (n=964) | 0.003 (−0.05 to 0.05) | 0.303 | 0.832 | 0.0002 (−0.05 to 0.05) | 0.488 | 0.857 | |
| Women (n=924) | −0.03 (−0.09 to 0.02) | −0.03 (−0.08 to 0.02) | ||||||
| Neutrophil‐to‐lymphocyte ratio | Men (n=964) | 0.01 (−0.05 to 0.06) | 0.006 | 0.066 | −0.004 (−0.06 to 0.05) | 0.035 | 0.385 | |
| Women (n=924) | −0.09 (−0.14 to −0.04) | −0.08 (−0.13 to −0.03) | ||||||
| Lymphocyte subsets (%, per 1‐SD) | T cells | Men (n=202) | 0.05 (−0.08 to 0.18) | 0.908 | 0.908 | 0.04 (−0.07 to 0.16) | 0.770 | 0.857 |
| Women (n=206) | 0.06 (−0.04 to 0.16) | 0.08 (−0.02 to 0.17) | ||||||
| CD4+ T cells | Men (n=202) | 0.02 (−0.10 to 0.13) | 0.76 | 0.908 | −0.01 (−0.12 to 0.10) | 0.779 | 0.857 | |
| Women (n=206) | −0.002 (−0.11 to 0.10) | −0.01 (−0.11 to 0.09) | ||||||
| CD8+ T cells | Men (n=202) | −0.02 (−0.14 to 0.10) | 0.701 | 0.908 | −0.004 (−0.12 to 0.11) | 0.685 | 0.857 | |
| Women (n=206) | 0.003 (−0.10 to 0.11) | 0.02 (−0.09 to 0.12) | ||||||
| Regulatory T cells | Men (n=202) | −0.03 (−0.15 to 0.08) | 0.378 | 0.832 | −0.01 (−0.12 to 0.10) | 0.263 | 0.723 | |
| Women (n=206) | −0.09 (−0.19 to 0.01) |
−0.10 (−0.19 to 0.0003) |
||||||
| B cells | Men (n=202) | −0.01 (−0.13 to 0.11) | 0.889 | 0.908 | −0.04 (−0.15 to 0.07) | 0.646 | 0.857 | |
| Women (n=206) | −0.02 (−0.11 to 0.07) | −0.02 (−0.11 to 0.07) | ||||||
Estimates are adjusted for age (y), glucose metabolism status (normal glucose metabolism, prediabetes, and type 2 diabetes), body mass index (kg/m2), glycated hemoglobin (mmol/mol), systolic blood pressure (mm Hg), total cholesterol/high‐density lipoprotein cholesterol, triglycerides (mmol/L), estimated glomerular filtration rate (mL/min per 1.73 m2), cardiovascular disease (yes/no), education (low, intermediate, high), physical activity (h/wk), smoking behavior (never, former, current), alcohol consumption (none, low, high), blood pressure‐lowering medication (yes/no), lipid‐lowering medication (yes/no), and nonsteroidal anti‐inflammatory drugs (yes/no).
P int indicates P value for interaction; and P intFDR, false discovery rate‐corrected P value using Benjamini and Hochberg.
There was also absence of evidence of effect modification by sex for neutrophil, basophil, eosinophil, or lymphocyte subsets. Finally, glucose metabolism status (NGM versus non‐NGM, Table 6 as well as non‐T2D versus T2D, Table 7) did not modify any association.
Table 6.
Glucose Metabolism Status‐Specific (NGM Versus Non‐NGM) Associations of Leukocytes and Lymphocyte Subsets With Heart Rate Variability Index Average Scores
| Glucose metabolism status | Time‐domain average score | Frequency‐domain average score | ||||||
|---|---|---|---|---|---|---|---|---|
| β (95% CI) | P int | P intFDR | β (95% CI) | P int | P intFDR | |||
| Leukocytes (absolute values, per 1‐SD) | Neutrophils | NGM (n=1050) | −0.04 (−0.10 to 0.02) | 0.533 | 0.944 | −0.04 (−0.10 to 0.02) | 0.293 | 0.796 |
| Prediabetes/T2D (n=811) | −0.09 (−0.15 to −0.01) | −0.1 (−0.16 to −0.04) | ||||||
| Lymphocytes | NGM (n=1050) | 0.10 (−0.04 to 0.24) | 0.656 | 0.944 | 0.09 (−0.06 to 0.23) | 0.229 | 0.796 | |
| Prediabetes/T2D (n=811) | 0.19 (0.05 to 0.33) | 0.24 (0.10 to 0.37) | ||||||
| Monocytes | NGM (n=1050) | −0.02 (−0.08 to 0.03) | 0.184 | 0.944 | −0.02 (−0.08 to 0.04) | 0.165 | 0.796 | |
| Prediabetes/T2D (n=811) | −0.01 (−0.06 to 0.09) | 0.02 (−0.04 to 0.09) | ||||||
| Basophils | NGM (n=1050) | 0.05 (−0.003 to 0.11) | 0.503 | 0.944 | 0.07 (0.01 to 0.13) | 0.359 | 0.796 | |
| Prediabetes/T2D (n=811) | 0.02 (−0.05 to 0.09) | 0.02 (−0.03 to 0.09) | ||||||
| Eosinophils | NGM (n=1050) | 0.01 (−0.05 to 0.06) | 0.858 | 0.944 | −0.004 (−0.06 to 0.06) | 0.74 | 0.904 | |
| Prediabetes/T2D (n=811) | −0.01 (−0.07 to 0.06) | −0.02 (−0.08 to 0.04) | ||||||
| Neutrophil‐to‐lymphocyte ratio | NGM (n=1050) | −0.05 (−0.12 to 0.10) | 0.945 | 0.945 | −0.07 (−0.13 to 0.001) | 0.996 | 0.996 | |
| Prediabetes/T2D (n=811) | −0.06 (−0.13 to 0.001) | −0.08 (−0.13 to −0.02) | ||||||
| Lymphocyte subsets (%, per 1‐SD) | T cells | NGM (n=213) | 0.11 (−0.04 to 0.27) | 0.143 | 0.944 | 0.06 (−0.09 to 0.22) | 0.633 | 0.87 |
| Prediabetes/T2D (n=194) | −0.03 (−0.16 to 0.10) | 0.001 (−0.12 to 0.14) | ||||||
| CD4+ T cells | NGM (n=213) | 0.02 (−0.13 to 0.17) | 0.847 | 0.944 | 0.07 (−0.09 to 0.23) | 0.585 | 0.87 | |
| Prediabetes/T2D (n=194) | −0.06 (−0.07 to 0.19) | −0.03 (−0.10 to 0.15) | ||||||
| CD8+ T cells | NGM (n=213) | −0.05 (−0.20 to 0.11) | 0.848 | 0.944 | −0.08 (−0.24 to 0.07) | 0.362 | 0.796 | |
| Prediabetes/T2D (n=194) | −0.05 (−0.18 to 0.09) | −0.004 (−0.13 to 0.13) | ||||||
| Regulatory T cells | NGM (n=213) | −0.02 (−0.17 to 0.13) | 0.656 | 0.944 | −0.02 (−0.17 to 0.13) | 0.567 | 0.87 | |
| Prediabetes/T2D (n=194) | −0.09 (−0.22 to 0.04) | −0.08 (−0.21 to 0.04) | ||||||
| B cells | NGM (n=213) | −0.03 (−0.17 to 0.1) | 0.726 | 0.944 | −0.01 (−0.15 to 0.13) | 0.856 | 0.942 | |
| Prediabetes/T2D (n=194) | 0.03 (−0.10 to 0.16) | 0.003 (−0.12 to 0.13) | ||||||
Estimates are adjusted for age (y), sex (men vs women), body mass index (kg/m2), glycated hemoglobin (mmol/mol), systolic blood pressure (mm Hg), total cholesterol/high‐density lipoprotein cholesterol, triglycerides (mmol/L), estimated glomerular filtration rate (mL/min per 1.73 m2), cardiovascular disease (yes/no), education (low, intermediate, high), physical activity (h/wk), smoking behavior (never, former, current), alcohol consumption (none, low, high), blood pressure‐lowering medication (yes/no), lipid‐lowering medication (yes/no), and nonsteroidal anti‐inflammatory drugs (yes/no).
NGM indicates normal glucose metabolism; P int, P‐value for interaction; P intFDR, false discovery rate‐corrected P‐value using Benjamini and Hochberg; T2D, type 2 diabetes.
Table 7.
Glucose Metabolism Status‐Specific (Non‐T2D Versus T2D) Associations of Leukocytes and Lymphocyte Subsets With Heart Rate Variability Index Average Scores
| Glucose metabolism | Time‐domain sum score | Frequency‐domain sum score | ||||||
|---|---|---|---|---|---|---|---|---|
| β (95% CI) | P int | P intFDR | β (95% CI) | P int | P intFDR | |||
| Leukocytes (absolute values, per 1‐SD) | Neutrophils |
NGM/prediabetes (n=1329) |
−0.05 (−0.10 to 0.01) | 0.628 | 0.891 | −0.05 (−0.10 to 0.01) | 0.433 | 0.741 |
| T2D (n=532) | −0.08 (−0.17 to −0.01) | −0.09 (−0.17 to −0.02) | ||||||
| Lymphocytes |
NGM/prediabetes (n=1329) |
0.10 (−0.02 to 0.23) | 0.420 | 0.880 | 0.09 (−0.04 to 0.22) | 0.076 | 0.418 | |
| T2D(n=532) | 0.23 (0.07 to 0.40) | 0.29 (0.13 to 0.44) | ||||||
| Monocytes |
NGM/prediabetes (n=1329) |
−0.01 (−0.06 to 0.04) | 0.419 | 0.880 | −0.01 (−0.06 to 0.05) | 0.674 | 0.741 | |
| T2D (n=532) | −0.01 (−0.10 to 0.09) | 0.01 (−0.08 to 0.07) | ||||||
| Basophils |
NGM/prediabetes (n=1329) |
0.05 (−0.003 to 0.10) | 0.371 | 0.880 | 0.07 (0.01 to 0.12) | 0.323 | 0.741 | |
| T2D (n=532) | 0.01 (−0.09 to 0.10) | 0.02 (−0.06 to 0.09) | ||||||
| Eosinophils |
NGM/prediabetes (n=1329) |
0.004 (−0.05 to 0.05) | 0.831 | 0.941 | −0.01 (−0.06 to 0.05) | 0.658 | 0.741 | |
| T2D (n=532) | −0.02 (−0.11 to 0.07) | −0.04 (−0.11 to 0.04) | ||||||
| Neutrophil‐to‐lymphocyte ratio |
NGM/prediabetes (n=1329) |
−0.05 (−0.11 to 0.10) | 0.976 | 0.976 | −0.07 (−0.13 to 0.01) | 0.812 | 0.812 | |
| T2D (n=532) | −0.06 (−0.14 to 0.01) | −0.06 (−0.13 to −0.0004) | ||||||
| Lymphocyte subsets (%, per 1‐SD) | T cells |
NGM/prediabetes (n=264) |
0.07 (−0.06 to 0.20) | 0.648 | 0.891 | 0.03 (−0.10 to 0.17) | 0.672 | 0.741 |
| T2D (n=143) | −0.02 (−0.14 to 0.17) | 0.05 (−0.10 to 0.20) | ||||||
| CD4+ T cells |
NGM/prediabetes (n=264) |
0.07 (−0.13 to 0.17) | 0.480 | 0.880 | 0.10 (−0.03 to 0.23) | 0.176 | 0.645 | |
| T2D (n=143) | −0.02 (−0.18 to 0.14) | −0.07 (−0.22 to 0.09) | ||||||
| CD8+ T cells |
NGM/prediabetes (n=264) |
−0.09 (−0.22 to 0.04) | 0.230 | 0.880 | −0.11 (−0.25 to 0.02) | 0.069 | 0.418 | |
| T2D (n=143) | −0.05 (−0.11 to 0.21) | −0.10 (−0.05 to 0.26) | ||||||
| Regulatory T cells |
NGM/prediabetes (n=264) |
−0.02 (−0.15 to 0.11) | 0.400 | 0.880 | −0.02 (−0.15 to 0.12) | 0.356 | 0.741 | |
| T2D (n=143) | −0.14 (−0.29 to 0.02) | −0.11 (−0.26 to 0.04) | ||||||
| B cells |
NGM/prediabetes (n=264) |
−0.001 (−0.13 to 0.13) | 0.855 | 0.941 | 0.01 (−0.12 to 0.15) | 0.585 | 0.741 | |
| T2D (n=143) | 0.003 (−0.14 to 0.15) | −0.03 (−0.16 to 0.11) | ||||||
Estimates are adjusted for age (y), sex (men vs women), body mass index (kg/m2), glycated hemoglobin (mmol/mol), systolic blood pressure (mm Hg), total cholesterol/high‐density lipoprotein cholesterol, triglycerides (mmol/L), estimated glomerular filtration rate (mL/min per 1.73 m2), cardiovascular disease (yes/no), education (low, intermediate, high), physical activity (h/wk), smoking behavior (never, former, current), alcohol consumption (none, low, high), blood pressure‐lowering medication (yes/no), lipid‐lowering medication (yes/no), and nonsteroidal anti‐inflammatory drugs (yes/no).
NGM indicates normal glucose metabolism; NLR, neutrophil‐to‐lymphocyte ratio; Pint, P‐value for interaction; PintFDR, false discovery rate‐corrected P value using Benjamini and Hochberg; and T2D, type 2 diabetes.
Sensitivity Analysis
Table S2 shows that the association of neutrophil and monocyte counts with HRV index average scores remained consistent after additional adjustment for HR. Table S3 shows the association of neutrophil and monocyte counts with individual HRV indices. Both higher neutrophil and monocyte counts were associated with lower SDANN and SDNN of the time domain as well as TP and ULF of the frequency domain. Neutrophil counts were additionally inversely associated with SDNN index of the time domain and VLF of the frequency domain.
DISCUSSION
In this large population‐based epidemiological study, we observed that higher circulating neutrophil and monocyte counts were associated with lower time‐domain and frequency‐domain HRV index average scores. These associations were not modified either by sex or by glucose metabolism status. Other leukocytes (basophils, eosinophils, lymphocytes) and lymphocyte subsets were not associated with either HRV average score. Furthermore, the association of neutrophil and monocyte counts with time‐domain HRV index average score was driven by their associations with SDNN and SDANN. Additionally, the association of monocyte and neutrophil counts with frequency‐domain HRV index average score was driven by their associations with TP and ULF.
In agreement with our findings, others have reported inverse associations of neutrophils with SDNN, 20 , 21 of neutrophils with TP, 20 and of monocytes with SDNN and TP. 20 The absence of association of lymphocyte subsets with individual HRV indices is in line with another study showing that most T‐cell subsets are not related to SDNN, LF, and power in the high‐frequency range. 22 In addition, leukocyte counts have been linked to SDNN, 23 , 24 SDANN 23 and one of the robust predictors of a time‐domain HRV index. 7 Thus far, no previous study has linked neutrophils and monocytes with any composite time‐ and frequency‐domain HRV indices, which means that a direct comparison with previous findings is not feasible. Similar to Aeschbacher et al, 20 the significant associations that we observed were robust to additional adjustment for HR. Interestingly we observed similar effect sizes for the associations of neutrophils and monocytes with SDNN, TP, ULF, and SDANN. This suggests that selection of a few of these linear HRV indices may offer sufficient insights for these research aims in study populations sharing similar characteristics to ours.
SNS and PSNS activity contribute to SDNN, 25 and SDNN, SDANN, TP, and ULF are under the combined effect of the SNS and PSNS. 26 It is also reported that SDNN and TP are highly correlated. 25 , 27 Therefore, the association of neutrophils and monocytes with these individual indices in the present study suggests that neutrophils and monocytes influence both SNS and PSNS such that their higher circulating levels affect sympathovagal imbalance.
After correction for multiple testing, sex did not modify the association of neutrophils and monocytes with HRV index average scores suggesting nondifferential associations of neutrophils and monocytes with HRV average scores in men and women. This finding holds true for the neutrophil‐ and monocyte‐related individual HRV indices, SDNN, SDANN, TP, and ULF as well. Contrastingly, 2 previous studies with similar proportions of sex (47% men and 53% women) observed sex‐specific associations. 20 , 28 Aeschbacher et al reported a stronger inverse association of neutrophil count with SDNN and TP among men when compared with women, 20 whereas Jensen‐Urstad et al reported that leukocyte count negatively correlated with TP, VLF, and LF only in men. 28 The younger age, 37 years 20 and 35 years, 28 in these studies as compared with 60 years of age in our study could be one explanation for the divergent findings. Further, because previous studies did not correct for multiple testing and adjust for covariates as extensively, our findings are likely to be more robust to chance and residual confounding.
Despite the well‐reported influence of glucose metabolism status on HRV, 10 , 11 , 12 , 13 it is important to note that glucose metabolism status did not modify the association of these innate immune cells with either the HRV average score or the individual indices. These findings are consistent with a prior report showing that glucose metabolism status did not influence the association of inflammatory and cardiovascular proteomics biomarkers with HRV indices. 7 These findings suggest that further investigation on how and when sex and glucose metabolism status modulate the relation between immune cells with HRV is warranted.
Neutrophils and monocytes are innate immune cells, which arise from common precursors and are produced and stored in large reserves in the bone marrow. They share complex relationships as well as similar roles that include coexpression of similar antigens and ready production of effector molecules such as granular proteins, oxidants, chemokines, and cytokines. 29 Because neutrophils and monocytes are activated by similar signals such as acute inflammation, infection, and tissue damage, 9 and the fact that both immune cells are associated with similar HRV indices, argues in favor of innate rather than adaptive signaling pathways that may be linked to HRV. This conclusion is supported by the absence of association of lymphocytes and their subsets with HRV. Moreover, the absence of association of NLR with HRV suggests that lymphocytes are the major driver of the impact of NLR on HRV in this population. With the extensive understanding of the immune system and its impact on the autonomic nervous system, this study underlines the influence of the systemic inflammatory state, especially the innate immune cells, on the vagus nerve function. 30
The main strength of this study is that it is a population‐based study. Additionally, the 24‐hour ECG helps capture a thorough picture of the HRV. Our study is also adequately powered to assess the association of the main predictor, leukocyte subsets with HRV, as well as effect modification by sex and glucose metabolism status. Other strengths of our study include the standardized assessment of all variables, the adjustment for multiple covariates and the ability to exclude the impact of HR, because change in HRV and altered morbidity could be attributable to the change in HR. 31 , 32 The immune cells were derived from 2 methods, automated Sysmex XE5000 and flow cytometry. The consistent absence of association between HRV indices and automated Sysmex XE5000‐derived lymphocyte counts and between HRV indices and the flow cytometry‐derived lymphocyte subset provides an internal validation of our findings. The limitations of this study include its cross‐sectional design precluding definitive conclusions about causal directions of the associations. We cannot exclude reverse association, such that dysregulation in the autonomic nervous system might have influenced immune cell counts. A bidirectional association between cardiac autonomic function and inflammation or the immune system has been reported. 33 We constructed composite scores because we expected similar pathophysiological effects of immune cells on the HRV indices within the same domain. Subsequent analysis of composite HRV score‐related immune cells with individual indices reduces the broader exploration of vagal versus sympathetic influence. Our indices were generated from 24‐hour ECG; therefore, we cannot fully exclude time‐of‐day‐dependent findings. However, as observed by another study, 28 it is likely that findings from daytime and nighttime HRV will be similar to the present 24‐hour findings. Additionally, there is potential selection bias because some measurements were in a subset of the study population. However, this bias is unlikely to have a substantial impact on our overall conclusion given the similar characteristics of the overall study population and the subset with lymphocyte subsets profiling. Our lymphocyte subsets were characterized as percentages of specific subset cells relative to their parent population. Different methodological approaches of characterizing the interrelationships among lymphocyte subsets might yield different associations with HRV. Finally, our study population is mainly of European ancestry, between 40 and 75 years old. Therefore, our findings may have limited generalizability to other ancestries and other age groups.
Conclusions
In conclusion, higher neutrophil and monocyte counts were associated with lower HRV regardless of sex and glucose metabolism status, as demonstrated by the absence of effect modification by either sex or glucose metabolism status. This study contributes novel hypotheses about the communication between the immune system and the autonomic nervous system and provides insights on potential mechanisms underlying cardiac autonomic dysfunction.
Sources of Funding
This study was supported by the European Regional Development Fund via OP‐Zuid, the Province of Limburg, the Dutch Ministry of Economic Affairs (grant 31O.041), Stichting De Weijerhorst (Maastricht, The Netherlands), the Pearl String Initiative Diabetes (Amsterdam, The Netherlands), the Cardiovascular Center (CVC, Maastricht, the Netherlands), CARIM School for Cardiovascular Diseases (Maastricht, The Netherlands), CAPHRI Care and Public Health Research Institute (Maastricht, The Netherlands), NUTRIM School for Nutrition and Translational Research in Metabolism (Maastricht, the Netherlands), Stichting Annadal (Maastricht, The Netherlands), Health Foundation Limburg (Maastricht, The Netherlands), and by unrestricted grants from Janssen‐Cilag B.V. (Tilburg, The Netherlands), Novo Nordisk Farma B.V. (Alphen aan den Rijn, the Netherlands), and Sanofi‐Aventis Netherlands B.V. (Gouda, the Netherlands). This study was also financed in part by The Netherlands Organization for Scientific Research (NWO) (Veni 916.12.056), The Netherlands Heart Foundation (2013T143), and a Seventh Framework Program (FP7) Grant (CIG 322070) to K. Wouters. The German Diabetes Center is funded by the German Federal Ministry of Health (Berlin, Germany) and the Ministry of Culture and Science of the state North Rhine‐Westphalia (Düsseldorf, Germany) and receives additional funding from the German Federal Ministry of Education and Research (BMBF) through the German Center for Diabetes Research (DZD e.V.). The funders had no role in study design or data collection, analysis, or interpretation.
Disclosures
None.
Supporting information
Tables S1–S3
Figures S1–S4
STROBE Checklist
Acknowledgments
The researchers are indebted to the participants for their willingness to participate in the study. Haifa Maalmi, Kristiaan Wouters, Christian Herder, and Coen D. A. Stehouwer designed the study. Haifa Maalmi and Kristiaan Wouters drafted the analysis plan. Haifa Maalmi, Anna Zhu, and Kolade Oluwagbemigun performed the statistical analysis. Kristiaan Wouters, Nicolaas C. Schaper, Jordi Heijman, Carla J. H. van der Kallen, Marleen M. J. van Greevenbroek, Pieter C. Dagnelie, Bastiaan E. de Galan, Anke Wesselius, and Coen D. A. Stehouwer contributed data. Haifa Maalmi, Christian Herder, Kolade Oluwagbemigun, and Coen D. A. Stehouwer interpreted data. Kristiaan Wouters, Nicolaas C. Schaper, Anna Zhu, Jordi Heijman, Carla J. H. van der Kallen, Marleen M. J. van Greevenbroek, Pieter C. Dagnelie, Bastiaan E. de Galan, Anke Wesselius, Dan Ziegler, and Michael Roden contributed to data interpretation. Haifa Maalmi and Kolade Oluwagbemigun wrote the article. Kristiaan Wouters, Christian Herder, and Coen D. A. Stehouwer critically reviewed and edited the article. All authors approved of its submission. Haifa Maalmi and Christian Herder are the guarantors of this work and, as such, had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
This article was sent to Olufunmilayo H. Obisesan, MD, MPH, Assistant Editor, for review by expert referees, editorial decision, and final disposition.
Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.125.043994
For Sources of Funding and Disclosures, see page 13.
References
- 1. Heart rate variability: standards of measurement, physiological interpretation and clinical use. Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Circulation. 1996;93:1043–1065. [PubMed] [Google Scholar]
- 2. Hayano J, Yuda E. Pitfalls of assessment of autonomic function by heart rate variability. J Physiol Anthropol. 2019;38:3. doi: 10.1186/s40101-019-0193-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Fang SC, Wu YL, Tsai PS. Heart rate variability and risk of all‐cause death and cardiovascular events in patients with cardiovascular disease: a meta‐analysis of cohort studies. Biol Res Nurs. 2020;22:45–56. doi: 10.1177/1099800419877442 [DOI] [PubMed] [Google Scholar]
- 4. Williams DP, Koenig J, Carnevali L, Sgoifo A, Jarczok MN, Sternberg EM, Thayer JF. Heart rate variability and inflammation: a meta‐analysis of human studies. Brain Behav Immun. 2019;80:219–226. doi: 10.1016/j.bbi.2019.03.009 [DOI] [PubMed] [Google Scholar]
- 5. Haensel A, Mills PJ, Nelesen RA, Ziegler MG, Dimsdale JE. The relationship between heart rate variability and inflammatory markers in cardiovascular diseases. Psychoneuroendocrinology. 2008;33:1305–1312. doi: 10.1016/j.psyneuen.2008.08.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Herder C, Schamarek I, Nowotny B, Carstensen‐Kirberg M, Straßburger K, Nowotny P, Kannenberg JM, Strom A, Püttgen S, Müssig K, et al. Inflammatory markers are associated with cardiac autonomic dysfunction in recent‐onset type 2 diabetes. Heart. 2017;103:63–70. doi: 10.1136/heartjnl-2015-309181 [DOI] [PubMed] [Google Scholar]
- 7. Oluwagbemigun K, Ziegler D, Strom A, Heier M, Bönhof GJ, Roden M, Rathmann W, Meisinger C, Peters A, Hauck SM, et al. Epidemiological investigation into inflammatory and cardiovascular proteomics biomarkers of heart rate variability indices of cardiac autonomic nervous dysfunction: results from the KORA S4/FF4 study. J Am Heart Assoc. 2025;14:e042144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Leick M, Azcutia V, Newton G, Luscinskas FW. Leucocyte recruitment in inflammation: basic concepts and new mechanistic insights based on new models and microscopic imaging technologies. Cell Tissue Res. 2014;355:647–656. doi: 10.1007/s00441-014-1809-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Chaplin DD. Overview of the immune response. J Allergy Clin Immunol. 2010;125:S3–S23. doi: 10.1016/j.jaci.2009.12.980 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Ziegler D, Zentai C, Perz S, Rathmann W, Haastert B, Meisinger C, Löwel H; KORA Study Group . Selective contribution of diabetes and other cardiovascular risk factors to cardiac autonomic dysfunction in the general population. Exp Clin Endocrinol Diabetes. 2006;114:153–159. doi: 10.1055/s-2006-924083 [DOI] [PubMed] [Google Scholar]
- 11. Rooney MR, Norby FL, Soliman EZ, Chen LY, Selvin E, Echouffo‐Tcheugui JB. Duration of diabetes, glycemic control, and low heart rate variability: the Atherosclerosis Risk in Communities (ARIC) study. J Diabetes Complications. 2024;38:108903. doi: 10.1016/j.jdiacomp.2024.108903 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Ziegler D, Voss A, Rathmann W, Strom A, Perz S, Roden M, Peters A, Meisinger C; KORA Study Group . Increased prevalence of cardiac autonomic dysfunction at different degrees of glucose intolerance in the general population: the KORA S4 survey. Diabetologia. 2015;58:1118–1128. doi: 10.1007/s00125-015-3534-7 [DOI] [PubMed] [Google Scholar]
- 13. Coopmans C, Zhou TL, Henry RMA, Heijman J, Schaper NC, Koster A, Schram MT, van der Kallen C, Wesselius A, den Engelsman R, et al. Both prediabetes and type 2 diabetes are associated with lower heart rate variability: the Maastricht study. Diabetes Care. 2020;43:1126–1133. doi: 10.2337/dc19-2367 [DOI] [PubMed] [Google Scholar]
- 14. Koenig J, Thayer JF. Sex differences in healthy human heart rate variability: a meta‐analysis. Neurosci Biobehav Rev. 2016;64:288–310. doi: 10.1016/j.neubiorev.2016.03.007 [DOI] [PubMed] [Google Scholar]
- 15. Schram MT, Sep SJ, van der Kallen CJ, Dagnelie PC, Koster A, Schaper N, Henry RM, Stehouwer CD. The Maastricht study: an extensive phenotyping study on determinants of type 2 diabetes, its complications and its comorbidities. Eur J Epidemiol. 2014;29:439–451. doi: 10.1007/s10654-014-9889-0 [DOI] [PubMed] [Google Scholar]
- 16. World Health Organization , International Diabetes Federation . Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia: report of a WHO/IDF consultation. Geneva: World Health Organization; 2006. [Google Scholar]
- 17. Maalmi H, Wouters K, Savelberg HHCM, van der Velde JHPM, Reulen JPH, Mess W, Schalkwijk CG, Stehouwer CDA, Roden M, Ziegler D, et al. Associations of cells from both innate and adaptive immunity with lower nerve conduction velocity: the Maastricht study. BMJ Open Diabetes Res Care. 2021;9:e001698. doi: 10.1136/bmjdrc-2020-001698 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Inker LA, Schmid CH, Tighiouart H, Eckfeldt JH, Feldman HI, Greene T, Kusek JW, Manzi J, Van Lente F, Zhang YL, et al. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367:20–29. doi: 10.1056/nejmoa1114248 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Buonacera A, Stancanelli B, Colaci M, Malatino L. Neutrophil to lymphocyte ratio: an emerging marker of the relationships between the immune system and diseases. Int J Mol Sci. 2022;23:3636. doi: 10.3390/ijms23073636 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Aeschbacher S, Schoen T, Dörig L, Kreuzmann R, Neuhauser C, Schmidt‐Trucksäss A, Probst‐Hensch NM, Risch M, Risch L, Conen D. Heart rate, heart rate variability and inflammatory biomarkers among young and healthy adults. Ann Med. 2017;49:32–41. doi: 10.1080/07853890.2016.1226512 [DOI] [PubMed] [Google Scholar]
- 21. Zhang Y, Han L, Shi L, Gao M, Chen J, Ding Y. Reduced heart rate variability is associated with altered clinical laboratory profile in people living with HIV. Brain Behav Immunity ‐ Health. 2024;41:100858. doi: 10.1016/j.bbih.2024.100858 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Schirato SR, El‐Dash I, El‐Dash V, Bizzarro B, Marroni A, Pieri M, Cialoni D, Chaui‐Berlinck JG. Association between heart rate variability and decompression‐induced physiological stress. Front Physiol. 2020;11:743. doi: 10.3389/fphys.2020.00743 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Sajadieh A, Nielsen OW, Rasmussen V, Hein HO, Abedini S, Hansen JF. Increased heart rate and reduced heart‐rate variability are associated with subclinical inflammation in middle‐aged and elderly subjects with no apparent heart disease. Eur Heart J. 2004;25:363–370. doi: 10.1016/j.ehj.2003.12.003 [DOI] [PubMed] [Google Scholar]
- 24. Anichkov DA, Shostak NA, Ivanov DS. Heart rate variability is related to disease activity and smoking in rheumatoid arthritis patients. Int J Clin Pract. 2007;61:777–783. doi: 10.1111/j.1742-1241.2006.01099.x [DOI] [PubMed] [Google Scholar]
- 25. Shaffer F, Ginsberg JP. An overview of heart rate variability metrics and norms. Front Public Health. 2017;5:258. doi: 10.3389/fpubh.2017.00258 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Assaad‐Khalil SH, Abdel Aaty T, El Feky M, Mohamed Abdel Naby H, Ramadan El Essawy N, Gaber Amin N. Neutrophil‐to‐lymphocyte ratio as a marker for cardiac autonomic neuropathy in Egyptian patients with type 2 diabetes: a cross‐sectional study. Cureus. 2024;16:e61634. doi: 10.7759/cureus.61634 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ernst G. Hidden signals‐the history and methods of heart rate variability. Front Public Health. 2017;5:265. doi: 10.3389/fpubh.2017.00265 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Jensen‐Urstad M, Jensen‐Urstad K, Ericson M, Johansson J. Heart rate variability is related to leucocyte count in men and to blood lipoproteins in women in a healthy population of 35‐year‐old subjects. J Intern Med. 1998;243:33–40. doi: 10.1046/joim.1998.243.1.33 [DOI] [PubMed] [Google Scholar]
- 29. Prame Kumar K, Nicholls AJ, Wong CHY. Partners in crime: neutrophils and monocytes/macrophages in inflammation and disease. Cell Tissue Res. 2018;371:551–565. doi: 10.1007/s00441-017-2753-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Pavlov VA, Tracey KJ. The vagus nerve and the inflammatory reflex—linking immunity and metabolism. Nat Rev Endocrinol. 2012;8:743–754. doi: 10.1038/nrendo.2012.189 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Monfredi O, Lyashkov AE, Johnsen AB, Inada S, Schneider H, Wang R, Nirmalan M, Wisloff U, Maltsev VA, Lakatta EG, et al. Biophysical characterization of the underappreciated and important relationship between heart rate variability and heart rate. 2014;64:1334–1343. doi: 10.1161/hypertensionaha.114.03782 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Stauss HM. Heart rate variability: just a surrogate for mean heart rate? Hypertension (Dallas, Tex: 1979). 2014;64:1184–1186. doi: 10.1161/HYPERTENSIONAHA.114.03949 [DOI] [PubMed] [Google Scholar]
- 33. Talbot S, Foster SL, Woolf CJ. Neuroimmunity: physiology and pathology. Annu Rev Immunol. 2016;34:421–447. doi: 10.1146/annurev-immunol-041015-055340 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Tables S1–S3
Figures S1–S4
STROBE Checklist