Association Between Heart Rate Variability and Home Blood Pressure: The Toon Health Study

Abstract

The autonomic nervous system plays an important role in the regulation of blood pressure (BP),¹ and autonomic dysfunction can cause hypertension that is associated with reduced baroreflex sensitivity.^2–4 Recently, the joint scientific statement on hypertension by several professional societies strongly recommended the use of home BP monitoring both to prevent cardiovascular outcomes and to control hypertension.^5,6 Given the pathological mechanisms that can raise BP, sympathovagal imbalance may have an important effect on home BP; however, there are few epidemiological studies on the association between sympathovagal imbalance and home BP in the general population.

Heart rate variability (HRV) is regulated predominantly by cardiac vagal tone, and lower HRV is associated with an increased risk of diabetes, hypertension, and cardiovascular disease (CVD) in epidemiologic studies.^7–10 We have reported that reduced HRV was associated with insulin resistance,¹¹ metabolic syndrome,¹² and elevated C-reactive protein concentration.¹³ In addition, we found that impaired parasympathetic function was associated with diastolic BP but not systolic BP at health checkups.¹⁴ This was explained to some extent by baroreflex sensitivity, which was inversely related to mean arterial pressure (MAP) and diastolic BP rather than systolic BP.²

Our study hypothesis was that autonomic dysfunction based on standardized indices of HRV¹⁵ obtained during health checkups may be associated with an increased MAP at home. HRV assessment may also help to manage BP at home.

METHODS

Study subjects

The Toon Health Study enrolled 2,032 men and women, 30–79 years of age between 2009 and 2012. The study included subjects who had HRV assessment (n = 2,030) and did not have atrial fibrillation (n = 15) on an electrocardiogram (ECG). Among 2,019 participants who had BP recorded at home, we included data from 1,905 who had morning and evening BP recorded for more than 3 days. Based on these criteria, 1,888 individuals were included in the analysis.

Written informed consent was obtained from all participants. The study protocol was approved by the Human Ethics Review Committees of Ehime University Graduate School of Medicine (no: 1705011).

Measurements

Overnight fasting blood samples were drawn from the antecubital vein into vacuum tubes containing a serum separator gel (for glucose and blood chemistry). The serum tube was centrifuged immediately at 3,000 rpm for 15 minutes, and the separated serum was sent to the laboratory for analysis.

BP was measured twice with individuals in the sitting position after a rest of at least 5 minutes using an automatic sphygmomanometer (BP-103iII; OMRON Colin Co., Ltd., Tokyo, Japan). The mean of the two measurements was used for analysis. All measurements were done from 8 to 11 am. Participants were asked to monitor their BP at home twice in the morning (within 1 hour after waking) and twice in the evening (before going to bed) using an automated device (UA-767; A&D Co., Ltd., Tokyo, Japan) that stored all BP measurements in the device memory. Subjects were required to measure their home BP for 1 week. Home BP was defined as the mean of the first two measurements of systolic and diastolic BP during each time period. Hypertension was defined as a systolic BP ≥140 mm Hg, and/or a diastolic BP ≥90 mm Hg, and/or the current use of any antihypertensive medication. Type 2 diabetes was defined as a fasting plasma glucose ≥7.0 mmol/l, a 2-hour-postload glucose ≥11.1 mmol/l, or the current use of antihyperglycemic agents. Body mass index (BMI) was calculated as weight (kg) divided by height (m) squared.

A self-administered questionnaire was used to assess medical history (presence of hypertension, dyslipidemia, or diabetes), smoking habits, and alcohol consumption. The amount of alcohol consumed each week was evaluated by measuring the weekly frequency of drinking and the type of alcoholic beverage consumed (beer, sake, whiskey, shochu, or wine). A regular alcohol drinker was defined as an individual with alcohol consumption ≥1 g/wk. Snoring was defined as a “Yes” response among three possible answers (“Yes,” “No,” or “Unknown”) to a question about snoring (“Do you snore?”). Self-reported snoring has been shown to have a 94% sensitivity and 58% specificity compared with recordings of snoring sounds as a gold standard.¹⁶ Furthermore, self-reported snoring has also been validated using pulse oximetry and cardiovascular disease outcomes.¹⁷ Physical activity levels were assessed using a validated questionnaire that consisted of 14 questions on occupation, locomotion, housework, sleep time, and leisure time physical activities.¹⁸ Responses for each physical activity category were converted to metabolic equivalents (METs), according to the Compendium by Ainsworth et al.,¹⁹ and expressed as METs·h/day. We identified mental health status using the Short Form-8 (SF-8) mental component score (MCS).²⁰

Assessment of autonomic function

HRV was measured with a fingertip pulse wave sensor (TAS9; YKC, Co., Ltd., Tokyo, Japan). The 5-minute pulse rate was recorded, and then the standard deviation of normal-to-normal RR intervals (SDNN) and root mean square of successive differences in RR intervals (RMSSD) were determined. Power spectral analysis of the pulse recordings was also used to obtain frequency-domain measures of HRV. The power spectrum was decomposed into its frequency components and quantified in terms of the relative intensity (power) of each component. The power spectrum was divided into frequency bands, and we determined the high frequency band (HF) (0.15–0.40 Hz) and low frequency band (LF) (0.04–0.15 Hz). The HF and LF power and the LF/HF ratio were used for further analysis.

We performed 24-hour Holter ECG monitoring (PMP400; Pacific Medico, Co., Ltd., Tokyo, Japan) in 140 volunteers among 557 study participants in 2015. To validate the 5-minute HRV measurement using the fingertip device, power spectral analysis of RR intervals from the ECG was performed every 5 minutes over 24 hours in 139 individuals who had complete records. For each 5-minute interval, we calculated LF, HF, and their ratio (LF/HF) in the same frequency bands.

To assess the reliability of the HRV parameters, each parameter was measured twice in the same individual at an interval of 2 months (n = 37). Pearson’s correlation coefficients between the two measurements within subjects were 0.67 for SDNN, 0.51 for RMSSD, 0.63 for LF, 0.54 for HF, and 0.21 for LF/HF. Coefficients of variation were 0.15, 0.20, 0.29, 0.27, and 0.20, respectively, which were similar to the values reported in a previously published validation study.²¹

Statistical analysis

Because of skewed distributions, SDNN, RMSSD, LF, and HF were log-transformed before analysis. The LF/HF ratio was calculated using log-transformed values of LF and HF. Sex- and age-adjusted means were calculated using analysis of covariance. The trend test was performed using log-transformed continuous values of SDNN, RMSSD, LF, HF, and LF/HF. Multivariable models were further adjusted for smoking, alcohol drinking, the use of hypertensive agents, diabetes, sleep duration, snoring, physical activity, and SF-8 MCS. A test for interaction of the stratified variables was done using the interaction terms that consisted of log-transformed values of SDNN, RMSSD, LF, HF, and LF/HF ratio as continuous variables adjusted for sex and age. Statistical significance was assumed at P <0.05. All statistical analyses were performed using SAS software, version 9.4 (SAS Institute, Inc., Cary, NC).

RESULTS

Table 1 shows the characteristics of all subjects included in the study. The mean values of systolic BP, diastolic BP, and MAP were lower when they were measured at home than in the health checkups.

Table 2 shows the correlation coefficients between HRV parameters estimated from 5-minute fingertip pulse recordings and the same HRV parameters obtained from 24-hour Holter ECG recordings. The mean HRV parameters from successive 5-minute ECG recordings were determined during five different time periods: morning, awake, evening, asleep, and all 24 hours. Heart rate, LF, HF, and LF/HF measured from pulse recordings were moderately correlated with the means of these same parameters measured from the ECG during each time period. The means of heart rate, ln LF, ln HF, and ln LF/HF in the 24-hour Holter ECG are shown by time period in Supplementary Table S1. Ln HF was higher and ln LF/HF was lower in the evening than in the morning.

Sex- and age-adjusted, and multivariable-adjusted means of casual and home MAP by quartile of HRV parameters are shown in Table 3. Sex- and age-adjusted means of casual, and home-measured morning and evening MAP were significantly different among quartiles of SDNN, RMSSD, and HF. When further adjusted for smoking, alcohol drinking, medication for hypertension, diabetes, sleeping hours, snoring, and SF-8 MCS, the associations were attenuated. Inverse relationships were found between the mean of morning MAP at home and RMSSD (P = 0.02) and HF (P = 0.051) after adjustment for confounders. The means of casual, and home systolic and diastolic BPs adjusted for covariates are shown by quartile of the HRV parameters in Supplementary Tables S2 and S3. Morning diastolic BP at home was strongly associated with both RMSSD and HF as well as MAP, and these associations were independent of covariates.

Regression analysis was performed using casual, home morning, or home evening MAP as dependent variables, and log-transformed values of RMSSD or HF as independent variables (Table 4). Partial regression coefficients were determined after subjects were stratified into two groups for each of the following six variables: sex, age, overweight, alcohol drinking, sleeping duration, and snoring. For the association of morning MAP at home with RMSSD or HF, there were no significant differences in the partial regression coefficients between the two groups (P > 0.05 for interaction) for any of the six variables included in the analyses. However, for the association of evening MAP at home with RMSSD or HF, there was a significant difference in the partial regression coefficients between the two age groups (P < 0.05 for interaction). A significant association between MAP and RMSSD or MAP and HF was evident in individuals <65 years old.

DISCUSSION

Low RMSSD and HF were significantly associated with an increased home MAP in the morning rather than in the evening, and these associations were independent of sex, age, BMI, smoking, alcohol drinking, use of hypertensive agents, diabetes, sleeping duration, snoring, physical activity, and SF-8 MCS. Our validation study suggested that 5-minute HRV parameters measured from fingertip pulse recordings at health checkups were moderately associated with HRV parameters from 24-hour Holter ECG recordings. Thus, HRV assessed from pulse recordings was fairly representative of HRV over a 24-hour period and influenced home MAP in the morning regardless of daily activities and sleep conditions.

Both home BP monitoring and ambulatory BP monitoring (ABPM) are thought to be superior to BP measurement in the clinic for the prediction of future CVD.²² In addition, day-to-day variability of home BP was shown to independently predict cardiovascular events and total mortality.²³ Furthermore, self-monitoring of BP in conjunction with other interventions (medication titration, patient education, and lifestyle counseling) in patients with hypertension led to a clinically significant BP reduction that persisted for at least 12 months.²⁴ Therefore, it is important to identify the factors influencing home BP to obtain better BP control throughout the day. Using ABPM, Ohira et al.²⁵ found that habitual alcohol intake raised morning BP, and this was linked with increased sympathetic activity during sleep.

HRV is a noninvasive tool to evaluate autonomic nervous system function.¹⁵ SDNN and LF are considered to represent overall sympathetic and parasympathetic activity. RMSSD and HF are indices of parasympathetic activity, and LF/HF is an indicator of sympathovagal balance. We confirmed that there are daily variations in autonomic nervous activity in the present study; there was a relative increase in parasympathetic activity in the evening and during sleep, which contributed to a lower evening BP at home. Consequently, a relative increase in parasympathetic tone and a decrease in sympathetic tone in the evening might weaken the associations between MAP and the HRV indices.

An association between autonomic nervous function and hypertension can be explained in part by the arterial baroreflex.¹ The baroreflex is believed to be a short-term controller of BP reflecting parasympathetic activation and sympathetic inhibition; however, Lohmeier et al.³ reported that the baroreflex has a long-term effect on BP. Hesse et al.² demonstrated that baroreflex sensitivity was inversely associated with the means of 24-hour or daytime MAP. Consistent with this, the Rotterdam Study showed that impaired baroreflex sensitivity reflected arterial stiffness and orthostatic BP changes in older adults.⁴

In the present study, we focused on MAP because of the role of baroreflex sensitivity in regulating MAP, as mentioned above. When systolic and diastolic BPs were analyzed separately, RMSSD and HF were inversely associated with morning home diastolic BP, rather than morning home systolic BP. This finding is consistent with a previous study that showed baroreflex sensitivity was highly correlated with diastolic but not systolic BP.²

The associations between RMSSD, HF, and MAP were obvious in individuals <65 years old. Since baroreflex sensitivity depends on age,²⁶ the impact of reduced parasympathetic activity on baroreflex sensitivity might be eliminated in the elderly.

The strength of the present study is that it included standardized HRV assessment and a large number of lifestyle variables. Nonetheless, several potential limitations should be noted. First, because this was a cross-sectional study, we could not confirm a causal relationship between decreased HRV and increased MAP. Second, we could not exclude the effects of agents that act on autonomic cardiac function, such as antihypertensive drugs. Our data on medication use were based on a self-administered questionnaire that did not involve information on specific classes of drugs. Furthermore, sensitivity analysis verified that the association between HRV and MAP did not change when patients on antihypertensive drugs were excluded. Third, because our subjects were volunteers from a single community, our findings may not be representative of the general Japanese population.

CONCLUSIONS

Low HF and RMSSD due to impaired parasympathetic nervous activity were associated with increased MAP levels in the morning rather than in the evening. Our findings suggest that factors which induce sympathovagal imbalance (physical inactivity, insomnia, and socioeconomic stress) should be eliminated or reduced to lower home BP in the morning.

SUPPLEMENTARY MATERIAL

Supplementary data are available at American Journal of Hypertension online.

ACKNOWLEDGMENTS

This study was supported, in part, by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (Grants-in-Aid for Research B, no. 22390134 in 2010–2012 and 25293142 from 2013, Grants-in-Aid for Young Scientists (B), no. 25860443 and 25860441 from 2013, and Grant-in-Aid for Research C, no. 26460767) and Health and Labor Sciences Research Grants from the Ministry of Health, Welfare, and Labor, Japan (Comprehensive Research on Life-Style Related Diseases including Cardiovascular Diseases and Diabetes Mellitus, no. 201021038A in 2010–2012). We thank Takuma Akita, other staff and participants of the Toon Health Study and the municipal authorities, officers, and health professionals of Toon City for their valuable contributions.

DISCLOSURE

The authors declared no conflict of interest.

REFERENCES