Lower heart rate variability predicts increased level of C-reactive protein 4 years later in healthy, nonsmoking adults
Abstract
Background
Inflammation and vagally mediated heart rate variability (vmHRV) have been implicated in a number of conditions including diabetes and cardiovascular disease. Consistent with the inflammatory reflex termed the ‘cholinergic anti-inflammatory pathway’, numerous cross-sectional studies have demonstrated negative associations between vmHRV and inflammatory markers such as C-reactive protein (CRP). The only prospective study, however, showed the opposite: higher CRP at baseline predicted higher high-frequency heart rate variability (HF-HRV) at follow-up. Thus, additional studies are needed to examine the prospective association between vmHRV and CRP.
Methods
Healthy employees participated in a voluntary on-site health assessment. Blood samples and ambulatory heart rate recordings were obtained, and night-time HF-HRV was calculated. Useable heart rate data were available in 2007 for 106 nonsmoking employees (9% women; age 44.4 ± 8 years), all of whom returned for an identical follow-up health assessment in 2011. Bootstrapped (500 replications) bivariate (r) and partial Pearson's correlations (ppc) adjusting for sex, age and body mass index at baseline (2007) were calculated.
Results
Zero-order correlations indicated that higher HF-HRV was associated with lower levels of CRP at both time-points (2007: r = −0.19, P < 0.05; 2011: r = −0.34, P < 0.001). After adjustment, HF-HRV remained a significant predictor of CRP (ppc = −0.20, P < 0.05).
Conclusion
In this study, we have provided in vivo support for the cholinergic anti-inflammatory pathway in humans. Cardiac vagal modulation at baseline predicts level of CRP 4 years later. Our findings have important implications for the role of vmHRV as a risk factor for cardiovascular disease morbidity and mortality. Interventions targeted at vmHRV might be useful in the prevention of diseases associated with elevated systemic inflammation.
Introduction
Inflammation and vagally mediated heart rate variability (vmHRV) have been implicated in a multitude of disorders including the metabolic syndrome and cardiovascular disease 1, 2. Currently, it is generally accepted that the autonomic nervous system plays an important role in immune function 3, 4. For example, the inflammatory reflex is a physiological mechanism through which the vagus nerve regulates immune function. Accordingly, efferent vagal activity inhibits the release of pro-inflammatory cytokines via the release of acetylcholine and this physiological mechanism has been termed the cholinergic anti-inflammatory pathway 4-6. Moreover, the release of IL-6 and other cytokines triggers the hepatic synthesis of C-reactive protein (CRP) 7. In addition, the vagus nerve is known to relay information about the peripheral immune status to the brain via IL-1 receptors conveyed by paraganglion cells situated in parasympathetic ganglia 8. Thus, both vagal efferent and afferent pathways seem to have an important role in immune regulation.
As expected, based on the role of the cholinergic anti-inflammatory pathway, numerous cross-sectional studies have demonstrated that vmHRV is inversely associated with inflammatory markers such as CRP (e.g. 9; for a review of studies up to 2008, see 10).
However, to date, few studies have investigated the prospective relationship between vmHRV and CRP in humans. The results by Singh et al. 10, from the one published prospective study, are counterintuitive; the authors reported that higher CRP at baseline predicted increased high-frequency heart rate variability (HF-HRV) at follow-up after 2 years. Thus, additional studies are needed to examine the prospective association between vmHRV and CRP.
Based on the expectations derived from the cholinergic anti-inflammatory pathway, we tested the following hypotheses in this study: (i) lower vmHRV is associated with higher CRP at baseline, and (ii) lower vmHRV predicts higher CRP levels after 4 years of follow-up.
Materials and methods
Study population
Healthy employees of an industrial company in Southern Germany participated in a voluntary on-site health assessment during regular work hours in 2007. The age range of the sample was 18–65 years (mean age 48 ± 8 years), and the cohort spanned all levels of socio-economic status. A total of 657 individuals underwent medical examination, including a full day recording of heart rate. The study was approved by the Medical Ethic Committee II of the Medical Faculty Mannheim of Heidelberg University (approval numbers 2007-009E-MA and 2010-296E-MA). All participants gave written informed consent prior to examination.
Of the original 657 individuals, 259 underwent an identical follow-up health assessment in 2011. Of these participants, 153 were excluded due to electrocardiogram (ECG) recording failure, insufficient duration of night-time recording, smoking at follow-up in 2011, the presence of acute inflammation in 2011 (CRP > 10 mg L−1) or other missing data. Thus, data were available for 106 employees [9% women; age 44.4 ± 8 years; body mass index (BMI) in 2007 24.6 ± 3.4 kg m−2; CRP in 2007 1.77 ± 1.81 mg L−1; CRP in 2011 1.24 ± 1.34 mg L−1; Table 1].
| 2007 | 2011 | |
|---|---|---|
| BMI, kg m−2 | 24.55 (3.42) | 25.34 (4.64) |
| SBP, mmHg | 126.38 (13.25) | 136.86 (13.26) |
| DBP, mmHg | 80.47 (8.99) | 80.53 (10.71) |
| CRP, mg L−1 | 1.77 (1.81) | 1.24 (1.34) |
| vmHRV, ms2 | 235.86 (202.98) | – |
- All data are presented as mean (SD). BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; CRP, C-reactive protein; vmHRV, vagally mediated heart rate variability.
Measurements
Following a comprehensive health assessment, participants were connected to a five-lead CardioScout Holter ECG system (SR-Medizinelektronik, Stuttgart, Germany; sampling rate 500 Hz) using wet electrodes (Ambu BluSensor, Ølstykke, Denmark). Participants were instructed to continue with their normal work, leisure and sleep routine for a minimum recording length of 20 h. Noninvasively measured HF-HRV reflects the degree of vmHRV. The raw ECG recordings were analysed by researchers at the Center for Neuropsychological Research (University of Trier, Trier, Germany) according to the guidelines of the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology 11. The 24-h interbeat interval data were decomposed into 5.35-min blocks and averaged over self-reported sleep periods. ANS-Explorer software 12 was used to calculate night-time HF-HRV (0.15–0.4 Hz) by spectral analysis (Fast Fourier transformation).
The next morning between 06.30 and 08.30, fasting blood samples were collected and immediately transported to a commercial laboratory (Synlab, Augsburg, Germany). CRP was determined using a high-sensitivity assay (Dade Behring, Schwalbach, Germany).
The overall measurement protocol was repeated in 2011; however, the health risk assessment and data collection were conducted by an agent (HealthVision Ltd, Berlingen, Switzerland). CRP was determined using another high-sensitivity assay (OSR6199; Beckman Coulter GmbH, Krefeld, Germany).
Statistical analyses
Statistical analyses were performed using bivariate Pearson's correlation (r) and hierarchical multiple regression reported as unstandardized beta values (β). Results are also presented as partial Pearson's correlation (ppc), which reflects the association between two variables after controlling for the effects of other variables in the model, and does not assume a directional relationship. Four models were considered and compared. Model 1 determined the bivariate association between HF-HRV and CRP at baseline. Model 2 determined the bivariate association between HF-HRV at baseline and CRP at follow-up. Model 3 was adjusted for BMI and age at baseline and sex. Model 4 was additionally adjusted for CRP at baseline. In all models, bootstrapping with 500 replications was performed. HF-HRV and CRP were log-transformed according to the ladder of powers method 13 prior to analysis to approximate a normal distribution. Analyses were performed using spss version 21.0 (SPSS Inc., Chicago, IL, USA).
Results
Model 1 was used to test the first hypothesis and revealed that lower baseline HF-HRV was associated with higher levels of CRP at baseline (r = −0.19; β = −0.24, 95% confidence interval (CI) −0.50 to −0.01, P < 0.05; R2=4%).
Model 2 was used to test the second hypothesis and revealed that lower baseline HF-HRV predicted higher levels of CRP at follow-up (r = −0.34; β = −0.41, 95% CI −0.63 to −0.18, P < 0.001; R2 = 12%; see Fig. 1).
In addition, higher BMI at baseline and older age were associated with higher CRP at follow-up (BMI: r = 0.32, P = 0.001; age: r = 0.18, P < 0.05). After adjusting for age and BMI at baseline and sex (Model 3), higher baseline HF-HRV remained a significant predictor of CRP at follow-up (ppc = −0.20; β = −0.27, 95% CI −0.52 to −0.01, P < 0.05; R2 = 21%).
Additional adjustment for baseline CRP (Model 4) slightly attenuated the relationship between baseline HF-HRV and CRP at follow-up (pcc = −0.17; β = −0.21, 95% CI −0.47 to 0.03, P < 0.1; R2 = 28%). Of note, the 95% CI values for Models 3 and 4 almost completely overlapped. However, inclusion of baseline CRP may lead to overcorrection of the regression model and may introduce multicollinearity due to the high correlation between CRP levels at baseline and at follow-up (r = 0.39, P < 0.01).
Discussion
Here, we present some of the first prospective data indicating a significant role of vmHRV in predicting CRP levels in healthy adults. We were able to confirm both the first hypothesis (cross-sectional observation) and the second hypothesis (longitudinal observation). The present results provide in vivo support for the importance of the cholinergic anti-inflammatory pathway: cardiac vagal modulation at baseline predicts CRP levels 4 years later.
There is growing recognition of the importance of vago-vagal circuits in the regulation of physiological function 14, 15. In the context of neuroimmunomodulation, the afferent and efferent vagal pathways involved in the regulation of human immune function have been described in detail in several recent reviews 16, 17. With regard to the afferent pathway, IL-1 receptors on the vagus nerve rapidly transmit location-specific information about levels of peripheral inflammatory cytokines to the brain, particularly the nucleus tractus solitarious (NTS), where they are integrated and modulated by higher structures 17. On the efferent side, these higher brain structures influence vagal efferents from the nucleus ambiguous and the dorsal motor nucleus of the vagus via the NTS to release acetylcholine, which modulates peripheral immune activity at least in part via alpha-7 nicotinic receptors 16. Of importance for the present study, the cytokine IL-6 is a potent stimulus for the hepatic synthesis of CRP 7.
Consistent with previous work from our group and by others, we show here that these vago-vagal reflexes are functional in vivo at physiological levels of vagal activity and, more importantly, that vagus nerve activity, as indexed by vmHRV, is predictive of the degree of systemic inflammation 4 years later. Thus, our results add to the data from a small number of prospective studies linking vmHRV to inflammation. The associations found in the present study for both the adjusted cross-sectional and prospective analyses are similar in magnitude to those found in previous cross-sectional studies by our group as well as by others 9, 18. We have previously reported that this magnitude of association between vmHRV and CRP is comparable to the effect of smoking on CRP 9. Given that vmHRV may be enhanced by such behavioural factors as diet 19 and exercise 17, 18, our results have important clinical implications in the search for pathways to increase cardiovascular health and reduce inflammation-related disease.
Limitations
In addition to the small number of women in the present study, there were several other limitations. First, we did not obtain concurrent measurements of respiration or physical activity. We used vmHRV values at night-time which should be a period of low physical activity. In addition, this represents the best condition to evaluate the resting sympathovagal balance without the confounding effect of activity and mental stress (e.g. due to the laboratory environment) 20. However, future studies should include measurements of respiration and physical activity to further investigate their potential roles in the association between vmHRV and inflammatory markers.
Secondly, data on only a limited number of inflammatory markers were available at follow-up. Future studies are needed to ascertain whether our findings can be generalized to other markers such as IL-6 or tumour necrosis factor-alpha.
Conclusion
Our findings provide in vivo support for the importance of the cholinergic anti-inflammatory pathway, as cardiac vagal modulation at baseline predicts the level of CRP 4 years later. The present findings have important implications for the role of vmHRV as a risk factor for cardiovascular disease morbidity and mortality. Interventions targeted at vmHRV, given that it is a modifiable risk factor, might be useful in the prevention of diseases associated with elevated systemic inflammation.
Conflict of interest statement
Until December 2012, JEF was the major shareholder of Health Vision Ltd, the company that organized the data collection in 2011. None of the other authors has any conflicts of interest to declare.
Acknowledgement
This work was supported by internal funds from the Mannheim Institute of Public Health.
