Associations of daytime, nighttime, and 24 h heart rate with four distinct markers of inflammation in hypertensive patients: the Styrian Hypertension Study

Abstract

The current study assessed which measure of heart rate (HR) is most associated with inflammatory activity. Among 368 hypertensive patients (mean age±SD, 60.6±10.8; 52.9% women), mean daytime (from 06:00–22:00 h), nighttime (from 22:00–06:00 h), and 24 h HR were recorded using a continuous 24 h ambulatory blood pressure monitoring portable device. Associations of daytime, nighttime, and 24 h HR with leukocytes, platelets, C-reactive protein [CRP], and 25-hydroxyvitamin D were calculated using multivariate linear regression, reporting unstandardized coefficients (B) with standard errors (SE). Mean daytime, nighttime, and 24 h HR were 73, 64, and 71 beats/min, respectively. All HR measures were positively associated with leukocytes after adjustment. Nighttime HR was additionally related with higher CRP. When all HR measures were simultaneously added to a single multivariate model, only the positive associations of nighttime HR with leukocytes (B [SE] = 0.06 [0.03], P =0.04), as well as with CRP (B [SE] = 0.20 [0.07], P =0.005) persisted. Nighttime HR was more closely associated with inflammatory activity. These observations lend some insight towards the pathophysiological mechanisms that implicate HR in cardiovascular risk, and afford valuable direction for forthcoming investigations.

INTRODUCTION

Prior epidemiological studies have documented a strong and independent relationship between high heart rate (HR) and adverse cardiovascular prognosis within a broad spectrum of populations.^1–5 Although the precise mechanisms have not been clarified thoroughly, convincing evidence indicates that this easy-to-measure and affordable parameter may heighten the inflammatory response, an essential component of the atherosclerotic process.^{6, 7} While the positive association between elevated HR and various markers of inflammation has been defined,^8–10 prior studies have evaluated only a single measure of daytime HR.

HR is a sensitive indicator of autonomic nervous system activity,¹¹ and is influenced by numerous factors including (but not limited to) environmental conditions,¹² psychological circumstances,¹³ physical fitness,¹⁴ medication therapy,^{15, 16} and circadian rhythm.¹⁷ To date, it remains uncertain whether the positive association between HR and inflammation can be more accurately predicted by other HR metrics determined over the course of 24 h, as compared with HR measured during daytime only.

Hence, the aim of this study was to assess the associations of HR obtained from daytime, nighttime, and 24 h recordings, with four distinct markers of inflammation (i.e., leukocytes, platelets, C-reactive protein [CRP], and 25-hydroxyvitamin D [25[OH]D]), in a population of hypertensive patients.

METHODS

Study patients and setting

In this investigation, cross-sectional data from the Styrian Hypertension Study (SHS) were used.¹⁸ Briefly, the SHS is an on-going single-centre cohort study designed to evaluate biomarkers in relation to arterial hypertension and cardiovascular risk in patients with a history of hypertension defined according to medical records or patient interview. Patients aged 18 years or older were recruited across 3 years (between August 2010 and July 2013) from the outpatient clinics at the Department of Cardiology and the Department of Internal Medicine, Division of Endocrinology and Metabolism, Medical University of Graz, Austria. The main exclusion criteria were stroke or myocardial infarction (MI) within the previous 4 weeks before study inclusion, pregnancy and lactating women, and an estimated life expectancy of less than one year. Patients presenting with any clinically significant acute disease such as infectious diseases were additionally excluded. Based on these criteria, the study sample included 383 patients, all of whom provided written informed consent. The appropriate Ethics Committee at the Medical University of Graz, Austria, approved the study, which was also designed to comply with the Good Clinical Practice (GCP) guidelines and the Declaration of Helsinki.

Assessment of ambulatory blood pressure and heart rate

A continuous 24 h ambulatory blood pressure monitoring (ABPM) portable device (Spacelabs 90207, OSI Systems Inc., Issaquah, WA, USA) was employed to determine patients’ mean 24 h systolic and diastolic blood pressure (BP). The circumference of the upper arm was measured in all patients to select the appropriate cuff for BP recordings. Mean daytime HR was acquired from 15 min ABPM sequences during the day from 06:00–22:00 h, and mean nighttime HR was determined from 30 min ABPM sequences during the night from 22:00–06:00 h. Mean 24 h HR was also derived from 24 h ABPM, similar to BP recordings as mentioned earlier. All measurements were recorded and analysed by trained personnel at the Medical University of Graz, Austria.

Definition of study variables

Weight and height were measured wearing no shoes and light clothes, and body mass index (BMI) was subsequently calculated using the formula: weight (kg)/height (m²). Smoking status, diabetes mellitus based on the American Diabetes Association (ADA) criteria, history of cardiovascular diseases (CVD) including MI and stroke or transient ischemic attack (TIA), and self-reported use of cardiovascular medication (i.e., beta-blockers, ACE-inhibitors, and calcium channel blockers) were recorded by questionnaire with patients answering “yes” or “no”, and these data were additionally validated by review of all available medical reports belonging to the study patients.

Laboratory assessment

Blood sampling was conducted during the morning between 07:00 and 10:00 h, with patients resting in the seated position for 10 min. On the morning of blood collection, patients were fasted for at least 12 h, and were instructed to refrain from tobacco consumption. Four distinct markers of inflammation were selected a priori for analyses. Total leukocyte count (expressed as the number of leukocytes • 10³/uL) was measured on a Sysmex® XE-5000 automated hematology analyser (Sysmex America, Inc., Mundelein, IL, USA), with the absolute count quantified by fluorescent flow cytometry methods. Platelet (number of platelets • 10³/uL) measurement was also performed using the same automated hematology analyser, with the total count later quantified using a DC voltage sheath flow detection method. Both hematological parameters were derived from blood samples anti-coagulated with K₂EDTA. Plasma CRP in mg/L was measured by the commercially available Cobas c system (Roche Diagnostics Corp., Indianapolis, IN, USA), and a CRP Gen.3 automated particle enhanced turbidimetric assay. The lower and upper detection limits were 0.3 mg/L and 350 mg/L, and the maximum intra- and inter-assay coefficients of variation were 3.7% and 4.0%, respectively. 25[OH]D was also used as a marker of inflammation as previous data strongly suggest a clinically significant inverse association between 25[OH]D and inflammation.¹⁹ 25[OH]D (ng/mL) was measured by means of a chemiluminescence assay (IDS, iSYS 25-hydroxyvitamin D; Immunodiagnostic systems Ltd., Boldon, UK) on an IDS-iSYS multi-discipline automated analyser. Within-day coefficients of variation were 5.5%–12.1%, and inter-day coefficients of variation were 8.9–16.9%, respectively. Standard analyses of fasting plasma glucose, total cholesterol, and triglyceride concentrations were performed according to routine laboratory procedures.

Statistical methods

Of the 383 patients enrolled, 15 (3.9%) were excluded due to missing information on the HR measures, leaving 368 patients in the analytic sample. For normally distributed variables, the mean±SD was reported; otherwise, the median with interquartile range was presented. Categorical variables were summarized as counts with proportions. The mean differences in the inflammatory markers across quartiles of daytime, nighttime, and 24 h HR were summarized using analysis of variance (ANOVA) with a Bonferroni post hoc test. Multivariate linear regression reporting unstandardized coefficients (B) with standard errors (SE) was employed to evaluate the associations of daytime, nighttime, and 24 h HR with the markers of inflammation. Three models were employed. The first model was unadjusted, the second model adjusted for age and sex, and the third model additionally adjusted for smoking status, BMI, blood pressure, fasting glucose, total cholesterol and triglyceride concentrations, a history of CVD, and cardiovascular medication (i.e., beta-blockers, ACE-inhibitors, and calcium channel blockers). Subsequently, daytime, nighttime, and 24 h HR were each added to a single multivariate model to determine which measure was most associated with each of the inflammatory markers. Because some patients may not have been asleep when the nighttime HR recording commenced, we tested the robustness and stability of this measure by substituting the current variable (i.e., from 22:00–06:00 h) with a stricter version of nocturnal HR (from 01:00–06:00 h). Further, as a sensitivity check, we repeated the analyses after omitting patients who reported the use of cardiovascular medications or who had a prior history of CVD, as these factors may have influenced the relationship between HR and inflammation. Results were considered statistically significant at a two-tailed P-value of <0.05. Statistical calculations were computed using STATA version 13.0 (StataCorp LP, College Station, TX, USA).

RESULTS

Baseline characteristics of the study cohort are provided in Table 1. The mean age was 61 years and over half (52.9%) of the study patients were women. According to BMI, the population was, on average, overweight (BMI = 29.5). The prevalence of smoking (15.0%) and type 2 diabetes mellitus (20.1%) were relatively low, as were prior history of MI (7.2%) and stroke or TIA (7.5%). Overall, more than half (51.3%) of the study patients reported using beta-blockers, while more than one third (35.3%) and one quarter (26.2%) reported taking ACE-inhibitors and calcium channel blockers, respectively. As expected, mean nighttime HR was lower compared with daytime and 24 h HR.

Table 1.

Baseline characteristics

No. of patients	368
Variables
Age, years, mean±SD	60.6±10.8
Women, n (%)	198 (52.9)
Current smoker, n (%)	56 (15.0)
Type 2 diabetes mellitus, n (%)	75 (20.1)
Objective measures
BMI, kg/m2, mean±SD	29.5±5.1
24 h ambulatory DBP, mmHg, mean±SD	77±9
24 h ambulatory SBP, mmHg, mean±SD	128±13
Fasting glucose, mg/dL, median (Q1-Q3)	95 (88–107)
Total cholesterol, mg/dL, median (Q1-Q3)	200 (167–227)
Triglycerides, mg/dL, median (Q1-Q3)	111 (77–150)
Markers of inflammation, median (Q1-Q3)
Leukocyte count • 103/μL	5.7 (4.8–6.8)
Platelet count • 103/μL	228 (196–275)
CRP, mg/L	1.7 (0.9–3.3)
25[OH]D, ng/mL	28.2 (21.5–36.2)
Prior history of CVD, n (%)
MI	27 (7.2)
Stroke or TIA	28 (7.5)
Atrial fibrillation	12 (3.2)
Medication use, n (%)
Beta-blockers	192 (51.3)
ACE-inhibitors	132 (35.3)
Calcium channel blockers	98 (26.2)
HR measures, beats/min, mean±SD
Daytime HR	73±10
Nighttime HR	64±9
24 h HR	71±10

Abbreviations: BMI = body mass index; DBP = diastolic blood pressure; SBP = systolic blood pressure; WBC = white blood cell; CRP = C-reactive protein; 25[OH]D = 25-hydroxyvitamin D; CVD = cardiovascular disease; MI = myocardial infarction; TIA = transient ischemic attack; ACE = angiotensin converting enzyme; HR = heart rate.

The HR variables demonstrated high collinearity with one another, as evidenced by Pearson correlation coefficients of 0.99 (between daytime and 24 h HR), 0.87 (between nighttime and 24 h HR) and 0.81 (between daytime and nighttime HR), all P <0.001.

Table 2 provides the mean differences of the inflammation markers according to quartiles of the HR variables. There was a significant increase in the number of leukocytes and platelets according to all of the HR measures. Also, there was a significant increase in concentrations of CRP, as well as a decline in 25[OH]D levels according to elevated nighttime HR only. On closer inspection, these observed differences seemed mainly driven by the fourth, and in some cases, the third quartile of the HR variables.

Table 2.

Mean differences in the inflammatory markers according to quartiles of daytime, nighttime and 24 h heart rate

	HR quartiles				P value#
	1st (n=96) Reference	2nd (n=90)	3rd (n=96)	4th (n=86)
Daytime HR, beats/min	61±5	70±2	77±2	86±5
Leukocyte count • 103/μL	5.6±0.14	0.1±0.02	0.4±0.06	0.7±0.07*	0.02
Platelet count • 103/μL	223±5.8	7±−0.7	28±0.1*	23±0.1*	0.003
CRP, mg/L	2.3±0.34	0.3±−0.02	1.4±0.17	0.6±0.03	0.07
25[OH]D, ng/mL	31.1±1.23	−1.9±−0.11	−2.7±−0.09	−1.7±−0.05	0.42
Nighttime HR, beats/min	53±4	61±2	67±2	76±5
Leukocyte count • 103/μL	5.5±0.14	0.3±0.02	0.3±0.01	1.2±0.07*†‡	<0.001
Platelet count • 103/μL	226±5.6	7±0.1	15±0.1	25±1.8*	0.03
CRP, mg/L	1.9±0.26	1.0±0.16	0.8±−0.01	1.9±0.29*	0.005
25[OH]D, ng/mL	32.4±1.19	−3.9±−0.02	−2.6±−0.01	−4.8±−0.04*	0.02
24 h HR, beats/min	60±4	68±2	74±2	84±5
Leukocyte count • 103/μL	5.5±0.13	0.3±0.03	0.5±0.03	0.8±0.08*	0.01
Platelet count • 103/μL	222±6.0	12±−0.7	27±−0.1*	24±1.3*	0.007
CRP, mg/L	2.3±0.35	0.2±−0.04	1.2±0.08	0.6±0.04	0.09
25[OH]D, ng/mL	31.8±1.15	−2.5±0.16	−3.4±−0.12	−2.8±0.07	0.16

Differences shown for inflammatory markers are versus the first quartile of heart rate measures (reference). Values are reported as mean±SD for heart rate measures and mean±SE for the inflammatory markers.

P values were determined using ANOVA.

^*P<0.05 versus the first quartile;

^†P<0.05 versus the second quartile;

^‡P<0.05 versus the third quartile.

Abbreviations as in Table 1.

Table 3 describes the relationship between the HR variables and inflammatory markers using multivariate linear regression. All three HR measures were associated with heightened inflammatory activity. In unadjusted analyses (Model 1), nighttime and 24 h HR were positively associated with all of the inflammatory markers, whereas daytime HR was associated with leukocytes and platelets only. Yet, following adjustment for age and sex (Model 2), each HR variable was significantly associated with all of the individual markers of inflammation. Further adjustment for a number of conventional risk factors (Model 3) attenuated the associations of daytime and 24 h HR with platelets, CRP, and 25[OH]D. Likewise, the relationship of nighttime HR with platelets and 25[OH]D disappeared after full adjustment; though, in addition to leukocytes, nighttime HR remained significantly associated with CRP.

Table 3.

Crude, partial, and fully adjusted regression coefficients for each marker of inflammation according to daytime, nighttime, and 24 h heart rate

	Model 1		Model 2		Model 3
	B	SE	B	SE	B	SE
Daytime HR
Leukocytes	0.03***	0.01	0.03***	0.01	0.02*	0.01
Platelets	1.01**	0.30	0.70*	0.30	0.54	0.34
CRP	0.03	0.02	0.04*	0.02	0.03	0.02
25[OH]D	−0.10	0.06	−0.12*	0.06	−0.03	0.07
Nighttime HR
Leukocytes	0.05***	0.01	0.05***	0.01	0.04***	0.01
Platelets	0.87**	0.34	0.71*	0.33	0.59	0.36
CRP	0.07**	0.02	0.08**	0.02	0.06**	0.02
25[OH]D	−0.18**	0.06	−0.20**	0.07	−0.09	0.07
24 h HR
Leukocytes	0.04***	0.01	0.04***	0.01	0.03**	0.01
Platelets	1.04**	0.31	0.73*	0.32	0.57	0.36
CRP	0.04*	0.02	0.05*	0.02	0.03	0.02
25[OH]D	−0.12*	0.06	−0.14*	0.06	−0.04	0.07

Values are unstandardized regression coefficients (B) with standard errors (SE).

Model 1 was unadjusted; Model 2 adjusted for age and sex; Model 3 also adjusted for smoking status, body mass index, 24 h ambulatory blood pressure, fasting glucose, total cholesterol and triglyceride concentrations, history of cardiovascular disease, and cardiovascular medication.

Abbreviations as in Table 1.

^*P <0.05.

^**P <0.01.

^***P <0.001.

In a multivariate model that included all three HR variables, only nighttime HR was retained as an important predictor of inflammation (Table 4). That is, each beat/min increment in nighttime HR increased leukocytes by 0.06 (SE = 0.03, P =0.04), as well as CRP by 0.20 (SE = 0.07, P =0.005) after adjustment.

Table 4.

Summary of regression coefficients for each marker of inflammation according to heart rate measures included in the same model

	Leukocytes		Platelets		CRP		25[OH]D
	B	SE	B	SE	B	SE	B	SE
Daytime HR	−0.01	0.10	3.28	3.70	0.39	0.24	0.08	0.73
Nighttime HR	0.06	0.03*	1.19	1.08	0.20	0.07**	−0.17	0.21
24 h HR	−0.01	0.12	−3.84	4.60	−0.53	0.30	0.03	0.90

Values reported are unstandardized regression coefficients (B) with standard errors (SE).

Model included daytime, nighttime and 24 h heart rate. Model further included age and sex, smoking status, body mass index, 24 h ambulatory blood pressure, fasting glucose, total cholesterol and triglyceride concentrations, history of cardiovascular disease, and cardiovascular medication.

Abbreviations as in Table 1.

^*P <0.05.

^**P <0.01.

When the analyses were repeated, substituting nighttime HR (i.e., from 22:00–06:00 h) with a more stringent measure of nocturnal HR (from 01:00–06:00 h), the results were virtually identical for the association between strict nighttime HR with leukocytes (B [SE] = 0.04 [0.01], P <0.001), as well as with CRP (B [SE] = 0.06 [0.02], P =0.01). Reanalysis of the data omitting those who reported using beta-blockers, ACE-inhibitors, and/or calcium channel blockers did not affect the results appreciably; nor did the findings differ after excluding those with a prior history of MI, stroke and TIA, or atrial fibrillation (data not reported). The results remained materially unchanged when men and women were analysed separately (data not shown).

DISCUSSION

The current investigation set out to determine the relationship of HR recorded during daytime, nighttime, and over 24 h with a set of inflammatory markers in hypertensive patients. All three measures of HR were associated with inflammatory activity. Specifically, each HR measure was positively related with leukocytes, whereas nighttime HR was additionally associated with higher concentrations of CRP.

The findings from this investigation are consistent with those of prior studies that assessed the relationship between daytime HR and inflammation. In a 3 year follow-up of persons with known cardiovascular risk, Nanchen and colleagues⁹ found elevated resting HR was highly correlated with CRP and numerous other markers of systemic inflammation. Results from the Multi-Ethnic Study of Atherosclerosis (MESA),²⁰ found that a high resting HR was significantly associated with CRP, interleukin-6, and fibrinogen. In a healthy adult population, Inoue and co-workers²¹ reported that for every 10 beat/min increment in the resting HR, the odds of having a high leukocyte count increased by approximately 30%, even in the absence of cigarette smoking.

To our knowledge, no prior study has evaluated the association between nighttime and 24 h HR with inflammatory markers. It is possible that these other HR metrics could provide better predictive value for identifying persons at-risk for subclinical inflammation, and consequently, for poor cardiovascular outcomes. A major finding of this study was that nighttime HR emerged as a robust predictor of inflammatory activity from among the three HR measures. Moreover, the association between nighttime HR and inflammation persisted even after adjusting for the other HR measures. In the same model, however, the association of either daytime or 24 h HR with the inflammation markers disappeared, which is fitting with the findings of Johansen et al.,²² where nighttime HR was the only HR measure associated with worse prognosis following multivariable adjustment. Seemingly, nighttime HR may provide a more superior indication of inflammatory activity as compared with other HR measures. Though, clearly, this deserves further investigation.

The actions of the autonomic nervous system follow a distinct circadian pattern, whereby sympathetic activity increases during daytime, and diminishes during the night.¹⁷ HR tends to follow a similar circadian pattern, characterized by a morning surge, followed by progressively higher values during daytime, and a clear nocturnal dip,^{17, 22} demonstrating a relative vagal dominance in the regulation of nighttime HR.²³ Nevertheless, impairment of autonomic neural activity during the night may negatively influence cardiovascular risk. In light of this conjecture, prior studies have reported a link between a blunted nocturnal decline in BP and adverse cardiovascular outcomes,²⁴ presumably owing to an altered circadian rhythm in the absence of sympathetic withdrawal during nighttime. Though the precise intrinsic and extrinsic mechanisms affecting the circadian profile of the autonomic nervous system are yet to be fully elucidated, it is biologically plausible that an exaggerated nocturnal rise in sympathetic overdrive may have accounted for the interplay observed between nighttime HR and inflammation in the current study. Indeed, others have documented a relationship between resting HR and systemic inflammation via mechanisms driven by autonomic nervous system dysfunction.^{10, 25} However, further examination of this hypothesis was beyond the scope of the current investigation. Additional studies are warranted to test this notion.

Study limitations

Our study has several limitations. The cross-sectional design precludes causal inference. The study was performed in a sample of Caucasians presenting with hypertension. Caution should be taken when extrapolating these findings to other populations. Patients enrolled in this study were not strictly and consecutively derived from our outpatient clinics, which may have introduced selection bias in the current investigation. The study sample was relatively small and each marker of inflammation was determined only once. Likewise, HR measures were recorded over the course of only one day. To further advance our understanding of the relationship between HR and inflammation, larger epidemiological studies with time-dependent measures are needed. Although four distinct markers of inflammation were employed, we cannot discount the possibility that the associations would have differed according to other unmeasured inflammatory markers. Moreover, HR is a sensitive indicator of autonomic nervous system function, and it is possible that other unmeasured factors (i.e., environmental conditions, psychological disturbances, and physical fitness) could have altered the association between HR and inflammation. The duration of time used for excluding persons who suffered a prior stroke or MI might not have been sufficient, and as a consequence, could have impacted upon the markers of inflammation. It is likely some individuals were awake during the early phases of nighttime recording (i.e., from 22:00–00:00 h), which could have influenced the current findings. Although, when nighttime HR was replaced with a more stringent nocturnal measure (from 01:00–06:00 h), when patients were expected to be sleeping, the results remained unchanged. Because internal (i.e. physiologic) and external (i.e. environmental) influences are minimized, nighttime HR may be a more reliable and accurate measure of resting HR, and a better predictor of inflammatory activity.

Conclusions

We examined the relationship between daytime, nighttime, and 24 h HR with four distinct markers of inflammation in a sample of hypertensive patients. Nighttime HR proved to be the most accurate predictor of inflammatory activity. These results suggest that it might be necessary to assess other measures of HR over the course of 24 h (e.g., nighttime HR), rather than just daytime alone, thereby permitting the opportunity to capture a better estimate of one’s habitual HR, along with a more reliable and better predictor of inflammation.