Abstract
Evidence from numerous animal models shows that vagal activity regulates inflammatory responses by decreasing cytokine release. Heart rate variability (HRV) is a reliable index of cardiac vagal regulation and should be inversely related to levels of inflammatory markers. Inflammation is also regulated by sympathetic inputs, but only one previous paper controlled for this. In a larger and more representative sample, we sought to replicate those results and examine potential sex differences in the relationship between HRV and inflammatory markers. Using data from the MIDUS II study, we analyzed the relationship between 6 inflammatory markers and both HF-HRV and LF-HRV. After controlling for sympathetic effects measured by urinary norepinephrine as well as a host of other factors, LF-HRV was found to be inversely associated with fibrinogen, CRP and IL-6, while HF-HRV was inversely associated with fibrinogen and CRP. We did not observe consistent sex differences. These results support the existence of the vagal anti-inflammatory pathway and suggest that it has similar effects in men and women.
Keywords: Vagal anti-inflammatory pathway, heart rate variability, inflammation, urinary norepinephrine
Introduction
The vagus nerve plays an important role in regulating inflammation and preventing tissue damage from excessive inflammatory responses. Vagal activity decreases production of pro-inflammatory cytokines such as TNF (Bernik 2002) and inhibits the migration of leukocytes to sites of inflammation (Saeed 2005), in part by its action on the reticuloendothelial system of the liver and spleen where cytokines are produced, and may function to dampen systemic inflammatory processes (Tracey 2007). Data from numerous animal studies support this anti-inflammatory pathway. For example, administration of endotoxin in mice following vagotomy or in mice possessing knockout of the α7 subunit of the nicotinic acetylcholine receptor (α7nAChR) expressed in macrophages causes an unrestrained cytokine response (Borovikova 2000, Wang 2003). On the other hand, stimulation of the vagus nerve or administration of α7nAChR agonists has been found to decrease cytokine release (Wang 2004).
Because heart rate variability (HRV) is a well-established and reliable index of cardiac vagal regulation, it should be inversely related to levels of inflammatory markers. Many studies show this predicted inverse relationship. For example, decreased low frequency HRV (LF-HRV) was found to be associated with increased levels of C-reactive protein (CRP) in a study of 1,601 healthy young people (Haarala 2011). A prospective cohort study of 188 middle-aged and older adults found an inverse relationship between high frequency HRV (HF-HRV) and CRP (p < 0.01) (Singh 2009). A study of 264 middle-aged male twins found that ultra low frequency HRV and very low frequency HRV were inversely related to CRP and IL-6 after controlling for a host of factors (p < 0.01) (Lampert 2008). IL-6 levels were shown to have an inverse relationship with HF-HRV and LF-HRV in a study of 682 patients after cardiac catheterization for acute myocardial infarction (MI) or unstable angina with elevated Troponin-T levels (Frasure-Smith 2009). Inverse relationships between IL-6 and HRV have also been observed in patients with sepsis, type 1 diabetes and type 2 diabetes (Tateishi 2007, González-Clemente 2007, Stuckey 2013).
Inflammatory processes are also influenced by the sympathetic nervous system (SNS), but its role is less well understood. The SNS possesses both pro- and anti-inflammatory properties and has been implicated in the production of cytokines (Koopman 2011). Adrenergic signaling may activate or suppress macrophages depending on the subtype of adrenergic receptor they express (Bellinger 2008). SNS activity can reduce Th1 response in favor of Th2 (Elenkov 2000). Sympathetic activity has also been found to enhance leukocyte attraction (Viswanathan 2005) and alter expression of cell adhesion markers (Redwine 2003).
A thorough examination of the inflammatory role of the autonomic nervous system thus requires consideration of both vagally-mediated and sympathetically-mediated effects. With only a single exception, studies linking HRV and inflammation fail to control for levels of SNS activity. In that study, Thayer and Fischer found that even after controlling for SNS effects, measured by urinary epinephrine, the inverse relationships between HRV and CRP and between HRV and WBC count remained significant (Thayer & Fischer 2009). In addition, they observed interesting sex differences in these relationships. For example, an increase of 1 SD in HRV measured as root mean square of successive interval differences was associated with a 48% decrease in CRP in men (p = 0.05), whereas in women, an increase of 1 SD in HRV was associated with a 104% decrease in CRP (p = 0.008). Larger differences in WBC count, another marker of inflammation, were also seen in women. This study suggests that there may be important sex differences in the relationship between parasympathetic activity and inflammatory markers. However, the study was limited by a small number of women (n = 66) relative to men (n = 545) and a relatively homogeneous sample of factory workers.
In the current study, we sought to replicate these findings on the relationship between HF-HRV and inflammatory markers using a larger, more diverse, and more representative sample. We tested the hypothesis that HF-HRV, as an index of cardiac vagal regulation, would be inversely related to inflammatory markers even after control for sympathetic effects. Because many studies also examine the relationship between LF-HRV and inflammatory markers, we also tested this relationship.
Methods
Participants
The data were collected from 1,255 participants in Midlife Development in the U.S. (MIDUS), a study of the behavioral, psychological and social factors accounting for age-related variation in health and well-being in a national sample of middle-aged and older Americans (Brim 2004). Data for the current study are from MIDUS II, a 9-year follow-up of the MIDUS I cohort, conducted between 2004 and 2006. MIDUS II consisted of five projects, including a self-administered survey of a wide array of behavioral, social and psychological factors and a Biomarker Project, with data collection conducted during a 1.5-day visit to a clinical research center (CRC) at the University of Wisconsin, UCLA, or Georgetown University. Biomarker data were collected from mid-2004 to mid-2009 (Ryff 2012). IRB approval was obtained for data collection at the three sites, and written consent was obtained from all study participants.
Physical Exam
Clinicians or trained staff evaluated vital signs, morphology, functional capacities, bone densitometry and medication usage and performed a physical exam. Medical history was obtained from participants.
Biomarker Data
Subjects underwent fasting blood draws prior to breakfast. Samples were sent to the MIDUS Biocore Lab for analysis. Additionally, patients had glycated hemoglobin and cholesterol panel assays analyzed at Meriter Labs (Madison, WI) using a Cobas Integra® analyzer (Roche Diagnostics, Indianapolis, IN). IL-6 was measured using Quantikine® High-sensitivity ELISA kit #HS600B (R&D Systems, Minneapolis, MN). Soluble IL-6 receptor levels were measured using Quantikine® ELISA kit #DR600 (R&D Systems, Minneapolis, MN). Human soluble intercellular adhesion molecule-1 was measured by Parameter Human sICAM-1 Immunoassay (R&D Systems, Minneapolis MN). Soluble E-selectin was measured by Parameter Human sE-selectin Immunoassay (R&D Systems, Minneapolis, MN). Fibrinogen and CRP were measured by BNII nephelometer (Dade Behring Inc., Deerfield, IL). 12-hour urine samples were collected overnight (7:00 PM-7:00 AM). Urinary catecholamine assays were performed using high-pressure liquid chromatography at the Mayo Medical Laboratory (Rochester, MN). Urinary norepinephrine levels were corrected for creatinine levels.
HRV Assessment
After an overnight stay at the CRC, participants were provided with a light breakfast, but no caffeine consumption was permitted. Following breakfast, they began the HRV psychophysiology protocol.
ECG electrodes were placed on the left and right shoulders as well as in the left lower quadrant. Respiration bands were placed around the chest and abdomen, and the finger cuff of a Finometer beat-to-beat blood pressure monitor was placed around the middle finger of the non-dominant hand. Respiration was calibrated using an 800 cc spirobag. While participants were in the seated position, data were recorded during an 11-minute baseline as part of a more extensive psychophysiology protocol with exposure to challenging stimuli and recovery periods. Here we report HRV data from this resting baseline.
Analog ECG signals were digitized at 500 Hz by a 16-bit A/D conversion board (National Instruments, Austin, TX) and passed to a microcomputer. The ECG waveform was submitted to an R-wave detection routine implemented by custom-written software, resulting in an RR interval series. Errors in marking R waves were corrected by visual inspection. Ectopic beats were corrected by interpolation.
HF-HRV (0.15–0.40 Hz) was computed based on 300-second epochs, using an interval method for computing Fourier transforms similar to that described by DeBoer, Karemaker and Strackee (DeBoer 1984). The mean value of HF-HRV from the two baseline 300-second epochs was computed. The process was repeated for LF-HRV (0.04–0.15 Hz).
Respiration
Respiratory rate was measured using an Inductotrace respiration monitor (Ambulatory Monitoring Systems, Ardsley, NY). Signals from thoracic and abdominal stretch bands were collected by the A/D board at 20 Hz and submitted to a custom-written program that computed respiratory rate on a minute-by-minute basis. The mean respiratory rate for the baseline period was computed.
Statistical Analysis
All analyses were carried out in SAS 9.3. The distributions of variables were examined and the right-skewed variables (HF- and LF-HRV, CRP, E-selectin, ICAM, IL-6, urinary norepinephrine) were natural log transformed prior to analysis. The associations of HF-HRV with each of the 6 inflammatory markers were separately tested within five hierarchical linear models, and the process was then repeated for LF-HRV. Individuals with data missing for a particular variable were removed from analyses involving that variable. Significance levels were corrected using the Bonferroni method to account for the 6 associations tested with each HRV variable, for each model.
In Model 1, each inflammatory marker was regressed on both HF- and LF-HRV without any covariates adjusted. Model 2 adjusted for urinary norepinephrine. Model 3 added sex, age, race, BMI, site of assessment, menstrual status, exercise and smoking status as covariates. Exercise was entered as a dichotomous variable indicating whether or not the subject engaged in at least 20 minutes of exercise at least 3 times a week. Smoking status was categorized into three components: current smoker, former smoker and never a smoker. In Model 4, data on use of statins, anti-inflammatory medications, and medications affecting parasympathetic activity were also adjusted for, as were heart problems and history of stroke, hypertension, diabetes, Parkinson’s disease, and any other neurological conditions. Finally, Model 5 utilized the same covariates as Model 4, but both HRV variables were residualized for respiratory rate.
To address missing covariates in the data, a multiple imputation analysis was also performed, for which missing data were imputed using all variables included in Model 5 regressions using PROC MI in SAS 9.3 (Little & Rubin 2002, Rubin 1976). This uses a Markov Chain Monte Carlo method (Shafer 1997) assuming arbitrary missing data patterns.
Results
Demographic Data
Analysis was carried out on participants with data for HF-HRV and LF-HRV as well as at least one inflammatory marker (n = 1,153). 91.3% of study participants were white. Demographic data are shown in Tables 1a and 1b. Men and women did not significantly differ in age, BMI, smoking status, history of hypertension, history of diabetes, LDL levels, ratio of total cholesterol to HDL, 12-hour urinary norepinephrine, or urinary norepinephrine adjusted for creatinine. Men had significantly higher levels of heart disease, blood pressure, triglycerides and urinary norepinephrine, and significantly lower levels of glycated hemoglobin, total cholesterol, HDL and CRP.
Table 1.
| a. Demographic information on subjects included in the analyses. | |||||
|---|---|---|---|---|---|
| Males (497) | Females (656) | ||||
| Label | N | % | N | % | P-Value |
| Current Smoker | 79 | 16 | 94 | 14 | 0.4522 |
| Ever Heart Disease | 66 | 13 | 46 | 7 | 0.0004 |
| Ever HTN | 166 | 34 | 236 | 36 | 0.3623 |
| Ever Diabetes | 64 | 13 | 79 | 12 | 0.6703 |
| b. Demographic information on subjects included in the analyses. | |||||||
|---|---|---|---|---|---|---|---|
| Males (497) | Females (656) | Difference | |||||
| Variable | Mean | SD | Mean | SD | Mean | SD | P- Value |
| Age (years) | 57.23 | 11.55 | 56.38 | 11.09 | 0.85 | 11.29 | 0.2060 |
| Glycated hemoglobin (%) | 6.39 | 7.36 | 7.59 | 12.64 | −1.21 | 10.69 | 0.0421 |
| SBP 1 (millimeters Hg) | 134.30 | 15.12 | 131.30 | 19.17 | 3.00 | 17.54 | 0.0030 |
| SBP 2 (millimeters Hg) | 132.85 | 15.98 | 129.43 | 19.81 | 3.42 | 18.26 | 0.0012 |
| SBP 3 (millimeters Hg) | 131.74 | 15.36 | 128.04 | 19.25 | 3.71 | 17.68 | 0.0003 |
| SBP Mean of All 3 (millimeters Hg) | 132.47 | 14.99 | 129.13 | 18.95 | 3.34 | 17.35 | 0.0009 |
| SBP Mean of 2 and 3 (millimeters Hg) | 132.52 | 15.14 | 128.99 | 19.07 | 3.53 | 17.49 | 0.0005 |
| DBP 1 (millimeters Hg) | 79.30 | 10.09 | 74.08 | 10.90 | 5.22 | 10.55 | <.0001 |
| DBP 2 (millimeters Hg) | 78.45 | 9.94 | 73.22 | 10.51 | 5.23 | 10.27 | <.0001 |
| DBP 3 (millimeters Hg) | 77.95 | 9.69 | 72.89 | 9.98 | 5.06 | 9.86 | <.0001 |
| DBP mean of all 3 (millimeters Hg) | 78.44 | 9.53 | 73.25 | 10.16 | 5.19 | 9.89 | <.0001 |
| DBP mean of 2 and 3 (millimeters Hg) | 78.43 | 9.52 | 73.29 | 9.89 | 5.14 | 9.73 | <.0001 |
| BMI (kilogram/meter^2) | 29.73 | 5.34 | 29.81 | 7.42 | −0.08 | 6.61 | 0.8302 |
| Total Cholesterol (milligrams/deciliter) | 183.07 | 41.03 | 190.31 | 38.85 | −7.24 | 39.81 | 0.0023 |
| HDL (milligrams/deciliter) | 47.73 | 15.32 | 60.85 | 17.46 | −13.12 | 16.57 | <.0001 |
| LDL (milligrams/deciliter) | 105.80 | 34.95 | 106.33 | 35.24 | −0.53 | 35.11 | 0.8005 |
| Ratio Total Cholesterol/HDL | 4.75 | 7.58 | 4.25 | 9.28 | 0.49 | 8.59 | 0.3208 |
| Triglycerides (milligrams/deciliter) | 156.86 | 189.05 | 116.22 | 68.13 | 40.64 | 134.53 | <.0001 |
| CRP (micrograms/milliliter) | 2.17 | 3.22 | 3.36 | 4.88 | −1.19 | 4.24 | <.0001 |
| Urinary NE (micrograms/deciliter) | 2.34 | 1.71 | 1.86 | 1.55 | 0.48 | 1.62 | <.0001 |
| 12-hour Urinary NE (micrograms/12 hours) | 402.01 | 1915.52 | 518.32 | 2183.90 | −116.30 | 2072.51 | 0.3369 |
| Urinary NE Adjusted for Creatinine (micrograms/gram) | 124.19 | 996.64 | 89.95 | 776.93 | 34.24 | 878.37 | 0.5263 |
Relationships between HRV and Inflammatory Markers
As Table 2 indicates, univariate analyses showed significant inverse relationships between HF-HRV and fibrinogen, soluble IL-6 receptor, ICAM and IL-6. After controlling for urinary norepinephrine, sex, age, race, BMI, study site, menstrual status, exercise, smoking, medications affecting parasympathetic activity, cardiac disease, stroke, hypertension, diabetes, Parkinson’s disease and other neurological condition, and respiratory rate, HF-HRV was significantly and inversely related to levels of fibrinogen and CRP.
Table 2.
Univariate and adjusted relationships between HRV and inflammatory markers in sequential models 1 through 5.
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Predictor | Response | N | Esti mate |
SE | P-Value | N | Esti mate |
SE | P-Value | N | Esti mate |
SE | P-Value | N | Esti mate |
SE | P-Value | N | Esti mate |
SE | P-Value |
| log HF HRV | Fibrinogen | 1138 | −6.68 | 1.943 | 0.0006^ | 1129 | −5.21 | 1.944 | 0.0074^ | 946 | −6.91 | 2.042 | 0.0007^ | 931 | −6.89 | 2.085 | 0.0010^ | 928 | −8.78 | 2.706 | 0.0012^ |
| Soluble IL6R | 1145 | −1005 | 233.4 | <.0001^ | 1136 | −985 | 236.2 | <.0001^ | 950 | −585 | 270.9 | 0.0311 | 935 | −505 | 278.1 | 0.0696 | 932 | −670 | 360.9 | 0.0637 | |
| log CRP | 1138 | −.053 | 0.026 | 0.0450 | 1129 | −.035 | 0.026 | 0.1863 | 946 | −.076 | 0.027 | 0.0045 | 931 | −.086 | 0.027 | 0.0016^ | 928 | −.110 | 0.035 | 0.0019^ | |
| log E-Selectin | 1153 | −.016 | 0.014 | 0.2338 | 1144 | −.019 | 0.014 | 0.1633 | 956 | −.019 | 0.015 | 0.2143 | 941 | −.013 | 0.016 | 0.3886 | 938 | −.016 | 0.020 | 0.4186 | |
| log ICAM | 1153 | −.034 | 0.011 | 0.0029^ | 1144 | −.029 | 0.012 | 0.0109 | 956 | −.004 | 0.011 | 0.7393 | 941 | 0.003 | 0.011 | 0.7518 | 938 | 0.003 | 0.014 | 0.8115 | |
| log IL6 | 1145 | −.050 | 0.017 | 0.0026^ | 1136 | −.039 | 0.017 | 0.0200 | 950 | −.037 | 0.017 | 0.0295 | 935 | −.041 | 0.017 | 0.0176 | 932 | −.051 | 0.022 | 0.0228 | |
| log LF HRV | Fibrinogen | 1138 | −15.6 | 2.115 | <.0001^ | 1129 | −13.8 | 2.142 | <.0001^ | 946 | −10.3 | 2.299 | <.0001^ | 931 | −10.6 | 2.350 | <.0001^ | 928 | −12.1 | 2.761 | <.0001^ |
| Soluble IL6R | 1145 | −801 | 259.7 | 0.0021^ | 1136 | −761 | 265.2 | 0.0042^ | 950 | −743 | 306.5 | 0.0155 | 935 | −701 | 315.1 | 0.0263 | 932 | −835 | 370.0 | 0.0242 | |
| log CRP | 1138 | −.182 | 0.029 | <.0001^ | 1129 | −.164 | 0.029 | <.0001^ | 946 | −.157 | 0.030 | <.0001^ | 931 | −.162 | 0.031 | <.0001^ | 928 | −.188 | 0.036 | <.0001^ | |
| log E-Selectin | 1153 | −.035 | 0.015 | 0.0216 | 1144 | −.041 | 0.015 | 0.0090^ | 956 | −.025 | 0.017 | 0.1480 | 941 | −.022 | 0.018 | 0.2192 | 938 | −.025 | 0.021 | 0.2380 | |
| log ICAM | 1153 | −.046 | 0.013 | 0.0003^ | 1144 | −.040 | 0.013 | 0.0021^ | 956 | −.012 | 0.012 | 0.3344 | 941 | −.006 | 0.013 | 0.6260 | 938 | −.008 | 0.015 | 0.6020 | |
| log IL6 | 1145 | −.154 | 0.018 | <.0001^ | 1136 | −.141 | 0.018 | <.0001^ | 950 | −.093 | 0.019 | <.0001^ | 935 | −.085 | 0.019 | <.0001^ | 932 | −.098 | 0.023 | <.0001^ | |
Significant after Bonferroni correction for 6 comparisons (p≤0.009)
Model 1: Unadjusted for any covariates
Model 2: Adjusted for log urinary norepinephrine (adjusted for creatinine)
Model 3: Adjusted for log urinary norepinephrine, sex, age, race, BMI, site, menstrual status, exercise, and smoking
Model 4: Adjusted for log urinary norepinephrine, sex, age, race, BMI, site, menstrual status, exercise, smoking, statins, anti-inflammatories, medications affecting parasympathetic activity positively or negatively, history of heart problems, history of stroke, hypertension, diabetes, Parkinson’s disease, and any other neurological condition
Model 5: Adjusted for log urinary norepinephrine, sex, age, race, BMI, site, menstrual status, exercise, smoking, statins, anti-inflammatories, medications affecting parasympathetic activity positively or negatively, history of heart problems, history of stroke, hypertension, diabetes, Parkinson’s disease, and any other neurological condition, corrected for respiratory rate
In univariate analyses, LF-HRV was significantly inversely related to all inflammatory markers except E-selectin. After controlling for all covariates, LF-HRV was significantly inversely related to fibrinogen, CRP and IL-6 but lost significance for ICAM and soluble IL-6 receptor.
As Table 2 indicates, the number of subjects available for each analysis varied, due to differences in information about the covariates available for each model. Missing data were imputed using all variables included in Model 5 regressions. At least one value was imputed for approximately 18% of the subjects. A total of 10 imputed datasets were created and analyzed, which is considered sufficient to yield relatively high efficiency (Graham et al. 2007). The significant findings were the same for the imputed and non-imputed analyses.
We found little evidence for sex differences in the HRV-inflammatory marker relationships for both HF-HRV and LF-HRV (Table 3).
Table 3.
| a. Comparison of the relationship and interaction between HF-HRV and inflammatory markers in men and women. | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Men | Women | Interaction | ||||||||
| Model | Response | Estimate | SE | P-Value | Estimate | SE | P-Value | Estimate | SE | P-Value |
| Model 1 | Fibrinogen | −9.66 | 2.799 | 0.0006^ | −5.79 | 2.611 | 0.0268 | 3.861 | 3.898 | 0.3221 |
| Sol. Il-6R | −587 | 333.1 | 0.0787 | −1329 | 322.6 | <.0001^ | −742 | 475.4 | 0.1187 | |
| logCRP | −.069 | 0.039 | 0.0737 | −.056 | 0.035 | 0.1090 | 0.014 | 0.053 | 0.7957 | |
| logE-Sel | 0.002 | 0.019 | 0.9260 | −.027 | 0.019 | 0.1701 | −.028 | 0.028 | 0.3120 | |
| logICAM | −.029 | 0.015 | 0.0600 | −.039 | 0.016 | 0.0175 | −.010 | 0.023 | 0.6794 | |
| logIl-6 | −.065 | 0.025 | 0.0088 | −.045 | 0.022 | 0.0415 | 0.019 | 0.034 | 0.5636 | |
| Model 2 | Fibrinogen | −8.21 | 2.791 | 0.0034^ | −4.77 | 2.650 | 0.0724 | 4.298 | 3.880 | 0.2682 |
| Sol. Il-6R | −591 | 338.4 | 0.0814 | −1307 | 327.7 | <.0001^ | −730 | 476.4 | 0.1257 | |
| logCRP | −.052 | 0.039 | 0.1820 | −.044 | 0.035 | 0.2119 | 0.018 | 0.053 | 0.7330 | |
| logE-Sel | 0.005 | 0.019 | 0.8097 | −.034 | 0.020 | 0.0791 | −.029 | 0.028 | 0.2945 | |
| logICAM | −.019 | 0.015 | 0.2033 | −.038 | 0.017 | 0.0218 | −.008 | 0.023 | 0.7195 | |
| logIl-6 | −.054 | 0.025 | 0.0280 | −.033 | 0.022 | 0.1363 | 0.021 | 0.033 | 0.5256 | |
| Model 3 | Fibrinogen | −7.25 | 2.804 | 0.0100 | −6.06 | 2.506 | 0.0159 | 1.716 | 3.654 | 0.6387 |
| Sol. Il-6R | −293 | 344.4 | 0.3958 | −1024 | 336.0 | 0.0024^ | −600 | 471.9 | 0.2035 | |
| logCRP | −.047 | 0.037 | 0.2033 | −.095 | 0.032 | 0.0031^ | −.043 | 0.047 | 0.3577 | |
| logE-Sel | −.006 | 0.019 | 0.7496 | −.061 | 0.020 | 0.0021^ | −.052 | 0.027 | 0.0551 | |
| logICAM | −.007 | 0.015 | 0.6629 | −.034 | 0.017 | 0.0483 | −.014 | 0.023 | 0.5337 | |
| logIl-6 | −.039 | 0.023 | 0.0907 | −.057 | 0.020 | 0.0048^ | −.021 | 0.030 | 0.4752 | |
| Model 4 | Fibrinogen | −6.86 | 3.111 | 0.0281 | −6.35 | 2.907 | 0.0293 | 0.178 | 4.048 | 0.9650 |
| Sol. Il-6R | −29.4 | 386.8 | 0.9394 | −1011 | 409.5 | 0.0139 | −822 | 540.3 | 0.1284 | |
| logCRP | −.047 | 0.041 | 0.2575 | −.123 | 0.037 | 0.0010^ | −.067 | 0.053 | 0.2020 | |
| logE-Sel | 0.004 | 0.020 | 0.8480 | −.027 | 0.023 | 0.2482 | −.033 | 0.030 | 0.2697 | |
| logICAM | 0.008 | 0.013 | 0.5233 | 0.002 | 0.018 | 0.9320 | −.003 | 0.022 | 0.8893 | |
| logIl-6 | −.025 | 0.025 | 0.3293 | −.063 | 0.024 | 0.0092^ | −.035 | 0.033 | 0.2960 | |
| Model 5 | Fibrinogen | −8.45 | 4.035 | 0.0368 | −8.32 | 3.774 | 0.0280 | −.059 | 5.278 | 0.9911 |
| Sol. Il-6R | −58.3 | 502.6 | 0.9077 | −1334 | 531.3 | 0.0124 | −1093 | 704.5 | 0.1211 | |
| logCRP | −.056 | 0.054 | 0.2946 | −.161 | 0.048 | 0.0009^ | −.093 | 0.069 | 0.1774 | |
| logE-Sel | 0.005 | 0.026 | 0.8538 | −.033 | 0.030 | 0.2714 | −.041 | 0.039 | 0.2934 | |
| logICAM | 0.011 | 0.017 | 0.5146 | −.001 | 0.023 | 0.9507 | −.008 | 0.028 | 0.7834 | |
| logIl-6 | −.031 | 0.033 | 0.3383 | −.080 | 0.031 | 0.0111 | −.042 | 0.044 | 0.3353 | |
| b. Comparison of the relationship and interaction between LF-HRV and inflammatory markers in men and women. | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Men | Women | Interaction | ||||||||
| Model | Response | Estimate | SE | P-Value | Estimate | SE | P-Value | Estimate | SE | P-Value |
| Model 1 | Fibrinogen | −17.9 | 2.946 | <.0001^ | −11.7 | 2.951 | <.0001^ | 6.181 | 4.241 | 0.1452 |
| Sol. Il-6R | −547 | 359.7 | 0.1287 | −950 | 370.0 | 0.0105 | −402 | 527.3 | 0.4455 | |
| logCRP | −.207 | 0.041 | <.0001^ | −.136 | 0.039 | 0.0006^ | 0.071 | 0.057 | 0.2151 | |
| logE-Sel | −.025 | 0.020 | 0.2214 | −.048 | 0.022 | 0.0284 | −.023 | 0.031 | 0.4513 | |
| logICAM | −.054 | 0.017 | 0.0012^ | −.039 | 0.018 | 0.0355 | 0.015 | 0.026 | 0.5626 | |
| logIl-6 | −.163 | 0.026 | <.0001^ | −.139 | 0.025 | <.0001^ | 0.024 | 0.036 | 0.5012 | |
| Model 2 | Fibrinogen | −16.1 | 2.985 | <.0001^ | −10.8 | 2.987 | 0.0003^ | 6.125 | 4.234 | 0.1483 |
| Sol. Il-6R | −557 | 369.8 | 0.1323 | −896 | 374.5 | 0.0170 | −383 | 529.2 | 0.4690 | |
| logCRP | −.187 | 0.042 | <.0001^ | −.129 | 0.040 | 0.0013^ | 0.068 | 0.057 | 0.2353 | |
| logE-Sel | −.021 | 0.021 | 0.3142 | −.056 | 0.022 | 0.0113 | −.023 | 0.031 | 0.4536 | |
| logICAM | −.039 | 0.017 | 0.0202 | −.038 | 0.019 | 0.0406 | 0.012 | 0.026 | 0.6283 | |
| logIl-6 | −.153 | 0.026 | <.0001^ | −.130 | 0.025 | <.0001^ | 0.022 | 0.036 | 0.5457 | |
| Model 3 | Fibrinogen | −13.5 | 3.172 | <.0001^ | −8.19 | 2.856 | 0.0043^ | 5.826 | 4.025 | 0.1480 |
| Sol. Il-6R | −364 | 394.5 | 0.3562 | −927 | 384.2 | 0.0161 | −339 | 523.6 | 0.5180 | |
| logCRP | −.154 | 0.042 | 0.0002^ | −.136 | 0.037 | 0.0002^ | 0.035 | 0.052 | 0.5021 | |
| logE-Sel | −.023 | 0.022 | 0.2995 | −.055 | 0.022 | 0.0140 | −.032 | 0.030 | 0.2896 | |
| logICAM | −.020 | 0.017 | 0.2584 | −.031 | 0.019 | 0.1144 | 0.006 | 0.025 | 0.8238 | |
| logIl-6 | −.095 | 0.026 | 0.0003^ | −.106 | 0.023 | <.0001^ | −.004 | 0.033 | 0.9082 | |
| Model 4 | Fibrinogen | −11.1 | 3.576 | 0.0020^ | −10.9 | 3.204 | 0.0007^ | 0.631 | 4.395 | 0.8858 |
| Sol. Il-6R | −261 | 447.5 | 0.5602 | −1002 | 455.3 | 0.0282 | −470 | 590.8 | 0.4263 | |
| logCRP | −.145 | 0.047 | 0.0023^ | −.191 | 0.041 | <.0001^ | −.024 | 0.057 | 0.6796 | |
| logE-Sel | −.031 | 0.023 | 0.1874 | −.022 | 0.026 | 0.3978 | −.007 | 0.033 | 0.8184 | |
| logICAM | −.008 | 0.015 | 0.5783 | 0.005 | 0.020 | 0.8022 | 0.016 | 0.024 | 0.5111 | |
| logIl-6 | −.066 | 0.029 | 0.0237 | −.116 | 0.027 | <.0001^ | −.024 | 0.036 | 0.5102 | |
| Model 5 | Fibrinogen | −12.2 | 4.209 | 0.0039^ | −12.7 | 3.759 | 0.0008^ | 0.238 | 5.187 | 0.9635 |
| Sol. Il-6R | −332 | 527.2 | 0.5297 | −1194 | 534.0 | 0.0258 | −564 | 697.2 | 0.4184 | |
| logCRP | −.164 | 0.056 | 0.0035^ | −.226 | 0.048 | <.0001^ | −.035 | 0.067 | 0.6019 | |
| logE-Sel | −.036 | 0.028 | 0.1919 | −.024 | 0.030 | 0.4158 | −.008 | 0.038 | 0.8340 | |
| logICAM | −.009 | 0.018 | 0.5932 | 0.003 | 0.023 | 0.9076 | 0.015 | 0.028 | 0.5946 | |
| logIl-6 | −.075 | 0.034 | 0.0286 | −.134 | 0.031 | <.0001^ | −.026 | 0.043 | 0.5494 | |
p≤0.0092
p≤0.0043
Discussion
This study provides further support for the anti-inflammatory role of the vagus nerve, even after controlling for SNS activity. Using a large, diverse and nationally representative sample, we found that HF-HRV and LF-HRV were significantly and inversely related to several inflammatory markers after controlling for relevant covariates. These results confirm and extend those of Thayer and Fischer 2009, which demonstrated an inverse relationship between an index of HF-HRV and both CRP and WBC count after controlling for sympathetic activity in a smaller and more homogeneous sample composed mostly of men. The inverse relationship between HRV and inflammatory markers supports the role of vagus nerve activity in limiting and preventing excessive inflammatory reactions.
Previous studies have tended to examine the relationship between HRV and relatively few inflammatory markers (Haarala 2011, Singh 2009, Lampert 2008). Unlike these studies, we examined a larger panel of inflammatory markers while adjusting for multiple comparisons. Even after adjustment, CRP and fibrinogen showed the predicted significant inverse relationships with both HF- and LF-HRV. Additionally, IL-6 showed a significant inverse relationship with LF-HRV. As von Kånel et al. point out, control for multiple comparisons may be excessively conservative (von Kånel 2008). Without this adjustment, soluble IL-6R and IL-6 also would have attained a significant or marginally significant inverse relationship to measures of HRV. Overall, the consistency of these findings across several inflammatory markers and with previous studies further strengthens the role of the vagal anti-inflammatory pathway as a regulator of systemic inflammation.
Cholinergic vagal input is transmitted to the celiac ganglia and thence to the splenic nerve and beta-receptors on memory T-lymphocytes via noradrenergic signaling. T-lymphocytes go on to stimulate macrophages via the α7nAChR. The activated macrophages then produce acetylcholine, which acts to decrease levels of inflammation (Rosas-Ballina 2011). Vagotomy or knockout of α7nAChR in animal models causes an unrestrained cytokine response (Borovikova 2000, Wang 2003), while vagal stimulation or administration of α7nAChR agonists decreases cytokine release (Wang 2004). In a study of humans who had experienced traumatic injury, 56 patients who underwent vagotomy were compared with 115 controls with similar injury severity. The vagotomy group showed increased levels of ulcer disease (71.43% vs. 2.61%; p < 0.001), septicemia (26.79% vs. 3.48%; p < 0.001), and mortality (27.27% vs. 9.57%; p = 0.003) (Peterson 2012). Taken together, these studies suggest a critical role for the vagus nerve in regulating inflammatory processes.
Thayer and Fischer reported significant sex differences in the relationship between HRV and inflammatory markers. However, this conclusion was limited by a sample in which only 10% of participants were women. In our sample composed of 656 women and 497 men, we did not observe significant sex differences in the relationship between HRV and inflammatory markers after controlling for covariates. These results suggest that the vagal anti-inflammatory pathway has similar effects on inflammation in both men and women.
While studies consistently have shown significant inverse relationships between HRV and a series of inflammatory markers (Haarala 2011, Lampert 2008, Frasure-Smith 2009), the magnitude of these effects generally has been small, as they were in this study. These small magnitude effects raise questions about the vagal pathways that regulate the heart and the inflammatory reflex. The vagus nerve contains A, B, and C fiber subtypes, only the latter two of which are involved in heart rate regulation (Jones 1995). This neuroanatomy suggests the possibility of a dissociation of efferent vagal regulation of the heart and inflammation. Consistent with this point, Huston et al. have shown that electrical stimulation of the distal end of the transected vagus nerve (1 V, 5 Hz, 2 ms), while sufficient to elicit anti-inflammatory effects, had no effect on heart rate (Huston 2007). More intense stimulation had the expected cardioinhibitory effect. This finding implicates vagal A fibers in anti-inflammatory signaling and suggests a lower activation threshold for this effect. Because inflammatory and cardiac effects appear to be activated by different levels of vagal stimulation via different fibers, the limited relationship between HRV and vagally-mediated anti-inflammatory effects may not be surprising.
Our study had the advantage of a large, representative sample that controlled for sympathetic effects in examining the relationship between HRV and inflammation. However, there were several limitations to our study. Participation required traveling to one of three testing centers and participating in research over an extended period of time, which may limit the generalizability of these results. In this respect, our study is similar to that of Thayer and Fischer, whose research participants all were currently employed. Thus, in both studies, participants were likely to be healthier than if they were drawn from community samples. However, studies from community samples also show inverse relationships between HRV and inflammatory markers (Sloan 2007). Additionally, given the cross-sectional nature of our study, we were unable to assess possible bidirectional relationships between HRV and inflammation. Overall, our study adds to the growing literature reporting significant inverse relationships between HRV and several inflammatory markers after controlling for numerous relevant covariates including urinary norepinephrine. Our results provide further support for the existence of the vagal anti-inflammatory pathway and suggest that it has similar effects in men and women. These findings on the role of the vagus nerve may be of clinical significance in the development of new therapies for inflammatory processes.
Acknowledgements
The MIDUS I study (Midlife in the U.S.) was supported by the John D. and Catherine T. MacArthur Foundation Research Network on Successful Midlife Development. The MIDUS II research was supported by a grant from the National Institute on Aging (P01-AG020166) to conduct a longitudinal follow-up of the MIDUS I investigation.
The research was further supported by the following grants M01-RR023942 (Georgetown), M01-RR00865 (UCLA) from the General Clinical Research Centers Program and UL1TR000427 (UW) from the National Center for Advancing Translational Sciences (NCATS), National Institutes of Health.
Contributor Information
Timothy M. Cooper, Email: tmc2161@columbia.edu.
Paula S. McKinley, Email: pm491@cumc.columbia.edu.
Teresa E. Seeman, Email: tseeman@mednet.ucla.edu.
Tse-Hwei Choo, Email: tjc2143@columbia.edu.
Seonjoo Lee, Email: sl3670@columbia.edu.
References
- 1.Bellinger DL, Millar BA, Perez S, Carter J, Wood C, ThyagaRajan S, et al. Sympathetic modulation of immunity: relevance to disease. Cellular immunology. 2008;252(1–2):27–56. doi: 10.1016/j.cellimm.2007.09.005. Epub 2008/03/01. doi: 10.1016/j.cellimm.2007.09.005. PubMed PMID: 18308299; PubMed Central PMCID: PMC3551630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bernik TR, Friedman SG, Ochani M, DiRaimo R, Susarla S, Czura CJ, et al. Cholinergic antiinflammatory pathway inhibition of tumor necrosis factor during ischemia reperfusion. Journal of vascular surgery. 2002;36(6):1231–1236. doi: 10.1067/mva.2002.129643. Epub 2002/12/07. doi: 10.1067/mva.2002.129643. PubMed PMID: 12469056. [DOI] [PubMed] [Google Scholar]
- 3.Borovikova LV, Ivanova S, Zhang M, Yang H, Botchkina GI, Watkins LR, et al. Vagus nerve stimulation attenuates the systemic inflammatory response to endotoxin. Nature. 2000;405(6785):458–462. doi: 10.1038/35013070. Epub 2000/06/06. doi: 10.1038/35013070. PubMed PMID: 10839541. [DOI] [PubMed] [Google Scholar]
- 4.Brim OG, Ryff CD, Kessler RC. The midus national survey: An overview. In: Brim OG, Ryff CD, Kessler RC, editors. How healthy are we? A national study of well-being at midlife. Chicago: University of Chicago Press; 2004. pp. 1–36. [Google Scholar]
- 5.DeBoer RW, Karemaker JM, Strackee J. Comparing spectra of a series of point events particularly for heart rate variability data. IEEE transactions on bio-medical engineering. 1984;31(4):384–387. doi: 10.1109/TBME.1984.325351. Epub 1984/04/01. doi: 10.1109/tbme.1984.325351. PubMed PMID: 6745974. [DOI] [PubMed] [Google Scholar]
- 6.Elenkov IJ, Wilder RL, Chrousos GP, Vizi ES. The sympathetic nerve--an integrative interface between two supersystems: the brain and the immune system. Pharmacological reviews. 2000;52(4):595–638. Epub 2000/12/21. PubMed PMID: 11121511. [PubMed] [Google Scholar]
- 7.Frasure-Smith N, Lesperance F, Irwin MR, Talajic M, Pollock BG. The relationships among heart rate variability, inflammatory markers and depression in coronary heart disease patients. Brain, behavior, and immunity. 2009;23(8):1140–1147. doi: 10.1016/j.bbi.2009.07.005. Epub 2009/07/29. doi: 10.1016/j.bbi.2009.07.005. PubMed PMID: 19635552. [DOI] [PubMed] [Google Scholar]
- 8.Gonzalez-Clemente JM, Vilardell C, Broch M, Megia A, Caixas A, Gimenez-Palop O, et al. Lower heart rate variability is associated with higher plasma concentrations of IL-6 in type 1 diabetes. European journal of endocrinology / European Federation of Endocrine Societies. 2007;157(1):31–38. doi: 10.1530/EJE-07-0090. Epub 2007/07/05. doi: 10.1530/eje-07-0090. PubMed PMID: 17609399. [DOI] [PubMed] [Google Scholar]
- 9.Graham JW, Olchowski AE, Gilreath TD. How many imputations are really needed? Some practical clarifications of multiple imputation theory. Prevention science : the official journal of the Society for Prevention Research. 2007;8(3):206–213. doi: 10.1007/s11121-007-0070-9. Epub 2007/06/06. doi: 10.1007/s11121-007-0070-9. PubMed PMID: 17549635. [DOI] [PubMed] [Google Scholar]
- 10.Haarala A, Kahonen M, Eklund C, Jylhava J, Koskinen T, Taittonen L, et al. Heart rate variability is independently associated with C-reactive protein but not with Serum amyloid A. The Cardiovascular Risk in Young Finns Study. European journal of clinical investigation. 2011;41(9):951–957. doi: 10.1111/j.1365-2362.2011.02485.x. Epub 2011/02/18. doi: 10.1111/j.1365-2362.2011.02485.x. PubMed PMID: 21323913. [DOI] [PubMed] [Google Scholar]
- 11.Huston JM, Gallowitsch-Puerta M, Ochani M, Ochani K, Yuan R, Rosas-Ballina M, et al. Transcutaneous vagus nerve stimulation reduces serum high mobility group box 1 levels and improves survival in murine sepsis. Critical care medicine. 2007;35(12):2762–2768. doi: 10.1097/01.CCM.0000288102.15975.BA. Epub 2007/09/29. doi: 10.1097/01.ccm.0000288102.15975.ba. PubMed PMID: 17901837. [DOI] [PubMed] [Google Scholar]
- 12.Jones JF, Wang Y, Jordan D. Heart rate responses to selective stimulation of cardiac vagal C fibres in anaesthetized cats, rats and rabbits. The Journal of physiology. 1995;489(Pt 1):203–214. doi: 10.1113/jphysiol.1995.sp021042. Epub 1995/11/15. PubMed PMID: 8583404; PubMed Central PMCID: PMC1156804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Koopman FA, Stoof SP, Straub RH, Van Maanen MA, Vervoordeldonk MJ, Tak PP. Restoring the balance of the autonomic nervous system as an innovative approach to the treatment of rheumatoid arthritis. Molecular medicine (Cambridge, Mass) 2011;17(9–10):937–948. doi: 10.2119/molmed.2011.00065. Epub 2011/05/25. doi: 10.2119/molmed.2011.00065. PubMed PMID: 21607292; PubMed Central PMCID: PMC3188868. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Lampert R, Bremner JD, Su S, Miller A, Lee F, Cheema F, et al. Decreased heart rate variability is associated with higher levels of inflammation in middle-aged men. American heart journal. 2008;156(4):759 e1–759 e7. doi: 10.1016/j.ahj.2008.07.009. Epub 2008/10/18. doi: 10.1016/j.ahj.2008.07.009. PubMed PMID: 18926158; PubMed Central PMCID: PMC2587932. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Little RJA, Rubin DB. Statistical analysis with missing data. 2nd ed. xv. Hoboken, N.J.: Wiley; 2002. p. 381. [Google Scholar]
- 16.Peterson CY, Krzyzaniak M, Coimbra R, Chang DC. Vagus nerve and postinjury inflammatory response. Archives of surgery (Chicago, Ill: 1960) 2012;147(1):76–80. doi: 10.1001/archsurg.2011.237. Epub 2012/02/23. PubMed PMID: 22355817. [DOI] [PubMed] [Google Scholar]
- 17.Redwine L, Snow S, Mills P, Irwin M. Acute psychological stress: effects on chemotaxis and cellular adhesion molecule expression. Psychosomatic medicine. 2003;65(4):598–603. doi: 10.1097/01.psy.0000079377.86193.a8. Epub 2003/07/29. PubMed PMID: 12883110. [DOI] [PubMed] [Google Scholar]
- 18.Rosas-Ballina M, Olofsson PS, Ochani M, Valdes-Ferrer SI, Levine YA, Reardon C, et al. Acetylcholine-synthesizing T cells relay neural signals in a vagus nerve circuit. Science (New York, NY) 2011;334(6052):98–101. doi: 10.1126/science.1209985. Epub 2011/09/17. doi: 10.1126/science.1209985. PubMed PMID: 21921156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Rubin DB. Inference and Missing Data. Biometrika. 1976;63(3):581–590. doi: Doi 10.1093/Biomet/63.3.581. PubMed PMID: ISI:A1976CP66700021. [Google Scholar]
- 20.Ryff C, Almeida DM, Ayanian JS, Carr DS, Cleary PD, Coe C, et al. National Survey of Midlife Development in the United States (MIDUS II), 2004–2006. ICPSR04652-v6 ed: Inter-university Consortium for Political and Social Research (ICPSR) 2012 [Google Scholar]
- 21.Saeed RW, Varma S, Peng-Nemeroff T, Sherry B, Balakhaneh D, Huston J, et al. Cholinergic stimulation blocks endothelial cell activation and leukocyte recruitment during inflammation. The Journal of experimental medicine. 2005;201(7):1113–1123. doi: 10.1084/jem.20040463. Epub 2005/04/06. doi: 10.1084/jem.20040463. PubMed PMID: 15809354; PubMed Central PMCID: PMC2213139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Schafer JL. Analysis of incomplete multivariate data. New York: Chapman & Hall; 1997. [Google Scholar]
- 23.Singh P, Hawkley LC, McDade TW, Cacioppo JT, Masi CM. Autonomic tone and C-reactive protein: a prospective population-based study. Clinical autonomic research : official journal of the Clinical Autonomic Research Society. 2009;19(6):367–374. doi: 10.1007/s10286-009-0019-0. Epub 2009/06/09. doi: 10.1007/s10286-009-0019-0. PubMed PMID: 19504232; PubMed Central PMCID: PMC2783459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Sloan RP, McCreath H, Tracey KJ, Sidney S, Liu K, Seeman T. RR interval variability is inversely related to inflammatory markers: the CARDIA study. Molecular medicine (Cambridge, Mass) 2007;13(3–4):178–184. doi: 10.2119/2006-00112.Sloan. Epub 2007/06/27. doi: 10.2119/2006–00112.Sloan. PubMed PMID: 17592552; PubMed Central PMCID: PMC1892756. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Stuckey MI, Petrella RJ. Heart rate variability in type 2 diabetes mellitus. Critical reviews in biomedical engineering. 2013;41(2):137–147. doi: 10.1615/critrevbiomedeng.2013008103. Epub 2013/01/01. PubMed PMID: 24580567. [DOI] [PubMed] [Google Scholar]
- 26.Tateishi Y, Oda S, Nakamura M, Watanabe K, Kuwaki T, Moriguchi T, et al. Depressed heart rate variability is associated with high IL-6 blood level and decline in the blood pressure in septic patients. Shock (Augusta, Ga) 2007;28(5):549–553. doi: 10.1097/shk.0b013e3180638d1. Epub 2007/12/14. doi: 10.1097/shk.0b013e3180638d1. PubMed PMID: 18075483. [DOI] [PubMed] [Google Scholar]
- 27.Thayer JF, Fischer JE. Heart rate variability, overnight urinary norepinephrine and C-reactive protein: evidence for the cholinergic anti-inflammatory pathway in healthy human adults. Journal of internal medicine. 2009;265(4):439–447. doi: 10.1111/j.1365-2796.2008.02023.x. Epub 2008/11/21. doi: 10.1111/j.1365-2796.2008.02023.x. PubMed PMID: 19019182. [DOI] [PubMed] [Google Scholar]
- 28.Tracey KJ. Physiology and immunology of the cholinergic antiinflammatory pathway. The Journal of clinical investigation. 2007;117(2):289–296. doi: 10.1172/JCI30555. Epub 2007/02/03. doi: 10.1172/jci30555. PubMed PMID: 17273548; PubMed Central PMCID: PMC1783813. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Viswanathan K, Dhabhar FS. Stress-induced enhancement of leukocyte trafficking into sites of surgery or immune activation. Proceedings of the National Academy of Sciences of the United States of America. 2005;102(16):5808–5813. doi: 10.1073/pnas.0501650102. Epub 2005/04/09. doi: 10.1073/pnas.0501650102. PubMed PMID: 15817686; PubMed Central PMCID: PMC556309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.von Kanel R, Nelesen RA, Mills PJ, Ziegler MG, Dimsdale JE. Relationship between heart rate variability, interleukin-6, and soluble tissue factor in healthy subjects. Brain, behavior, and immunity. 2008;22(4):461–468. doi: 10.1016/j.bbi.2007.09.009. Epub 2007/11/06. doi: 10.1016/j.bbi.2007.09.009. PubMed PMID: 17977694; PubMed Central PMCID: PMC2373608. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wang H, Liao H, Ochani M, Justiniani M, Lin X, Yang L, et al. Cholinergic agonists inhibit HMGB1 release and improve survival in experimental sepsis. Nature medicine. 2004;10(11):1216–1221. doi: 10.1038/nm1124. Epub 2004/10/27. doi: 10.1038/nm1124. PubMed PMID: 15502843. [DOI] [PubMed] [Google Scholar]
- 32.Wang H, Yu M, Ochani M, Amella CA, Tanovic M, Susarla S, et al. Nicotinic acetylcholine receptor alpha7 subunit is an essential regulator of inflammation. Nature. 2003;421(6921):384–388. doi: 10.1038/nature01339. Epub 2003/01/01. doi: 10.1038/nature01339. PubMed PMID: 12508119. [DOI] [PubMed] [Google Scholar]