Midlife resting heart rate, but not its visit-to-visit variability, is associated with late-life frailty status in men with coronary heart disease

Abstract

Background

Autonomic imbalance is linked with multiple health conditions, yet its associations with frailty were rarely studied. We assessed the relationship of resting heart rate (RHR) and visit-to-visit heart rate variability (HRV) with future frailty among elderly men with coronary heart disease (CHD).

Methods

Three-hundred-six community-dwelling men with CHD who participated in the Bezafibrate Infarction Prevention (BIP) trial (1990-1998; mean age 56.6 ± 6.5 years) underwent assessment of physical frailty in 2011–2013 (mean age 77.0 ± 6.4 years). Mean RHR and visit-to-visit HRV were calculated from electrocardiogram as indicators of autonomic imbalance. Nominal logistic and linear regression models were used to assess the relationships of RHR and HRV with frailty status and its components (i.e. gait speed, grip strength, weight loss, exhaustion and activity), respectively. Adjustments were made for various demographic, clinical and metabolic covariates.

Results

Of the 306 men, 81 (26%) were frail and 117 (38%) were prefrail. After controlling for potential confounders, RHR, but not visit-to-visit HRV, was associated with higher odds of being prefrail [OR = 1.44 (95%CI 1.15, 1.79)] and frail [OR = 1.35 (95%CI 1.03, 1.77)]. Each 5-bpm increase in RHR was associated with weaker grip (β= −1.12 ± 0.32 kg; p-value < .001) and slower gait speed (β = 0.19 ± 0.08s/m; p-value = .022).

Conclusions

Midlife RHR may be associated with late-life frailty in men with CHD.

Keywords:

• Autonomic imbalance
• physical frailty
• cardiovascular disease
• observational study

Introduction

Frailty is increasingly recognized as a problematic expression of aging, especially in light of the efforts to increase healthy life span and improve healthcare among elderly individuals [1]. While frailty is common among the elderly population [2], it is even more frequent among individuals with cardiovascular disease (CVD) [3,4]. Indeed, frailty and CVD are closely related, and share common pathophysiological mechanisms including inflammation and insulin resistance [5,6].

A broad range of sociodemographic, physical, biological, lifestyle, and psychological factors have been identified as risk factors for frailty in the general population [7]. Yet, the heterogeneity in frailty status among older-adults is not fully explained, particularly among the large sub-population of individuals with CVD. Investigation of early markers for frailty in this high-risk population is fundamental for risk stratification and prevention.

Resting heart rate (RHR) indicates the influence of the autonomic nervous system on the heart. Thus, as elevated RHR reflects autonomic imbalance, which refers to excessive sympathetic activity and too little parasympathetic activity [8]. Previous literature assessing the link between autonomic imbalance and frailty is scarce. In several studies, autonomic imbalance assessed using indices of heart rate variability has been associated with increased frailty [9,10]. Additionally, findings from the Study of Pravastatin in the Elderly at Risk (PROSPER) suggest that higher RHR is related to worse functional status and to a higher risk of future functional decline in older adults, independently of CVD [11]. Autonomic dysfunction has also been linked with multiple health conditions which in turn are closely related to frailty, including cognitive decline [12] coronary artery disease, stroke, hypertension, metabolic dysfunction and mortality [13]. The potential value of assessing visit-to-visit RHR in relation to frailty is unknown.

In the current study, we sought to assess the relationship between midlife RHR and visit-to-visit HRV with frailty ∼20 years later, in old-aged men with coronary heart disease (CHD).

Methods

Study sample

This study includes participants who previously participated in a clinical trial of lipid modifications. This trial, the Bezafibrate Infarction Prevention (BIP) study, was a large multi-center, placebo-controlled randomized clinical trial investigating the efficacy of bezafibrate 400 mg daily, a fibric derivative, in secondary prevention among participants with established stable CHD [14]. Participants were recruited from 18 medical centers in Israel between May 1990 and January 1993 and included men and women 45 to 74 years of age with a history of myocardial infarction at least 6 months and no longer than 5 years before enrollment and/or stable angina pectoris during the last 2 years confirmed by coronary angiography, and/or radionuclide studies or standard exercise tests. Individuals with diabetes, hepatic or renal failure, and disabling stroke were excluded.

The sample for the current study consists of former participants of the BIP study, recruited from 8 medical centers located at the center of Israel. These individuals underwent 2 follow-up evaluations, which included collection of data on health status, cognitive function and frailty measures: the first was done between 2004 and 2008, and the second between 2011 and 2013, 19.9 ± 1.0 years after recruitment. The present study is based on data from the latter evaluation. Participants were assessed at the Sagol Neuroscience center, Sheba medical center, or at their residence if they were unable or unwilling to arrive.

The first evaluation included 558 participants, and of them 444 survived for the second evaluation. Among them, 58 refused to participate, four could not be contacted and 19 were unable to participate because of vision or hearing defects, or a severe illness. Thus, at the total sample consisted of 363 individuals, a response rate of 80%. We excluded 43 persons with prevalent atrial fibrillation at the BIP study enrollment, 2 who were lacking information on ECG and 12 women, resulting in a total sample of 306 men.

Data were obtained under a protocol approved by the institutional review board of the Sheba Medical Center and written informed consent was obtained from all participants.

Measurement of resting heart rate and visit-to-visit heart rate variability

Resting heart rate was measured in each visit during the BIP trial follow-up, a mean time duration of 6.2 years (range, 4.7–7.6 years) [14]. Data on heart beat rate were obtained from 12-lead electrocardiogram (ECG) performed at each center according to standardized procedures. A resting ECG was recorded at baseline and in each follow-up visit. Tracings were transferred for central reading at the study coordinating center.

For each participant, we estimated RHR as the average of all RHR measurements during the BIP study period. A previous study demonstrated that heart rate measured repeatedly over time (in-trial heart rate) provides more precise results with respect to cardiovascular end-points [15]. On average, 7.0 ± 1.8 visits were available for the whole group (range 1–18). Visit-to-visit HRV were calculated as the coefficient of variation (CV), which is the ratio of between visits standard deviation to the mean heart rate (CV = SD/mean*100%).

Assessment of frailty

The Fried phenotypic definition of frailty was used [16]. Participants were classified as frail if they met 3 or more of the following criteria: poor grip strength, slow gait speed, low levels of physical activity, unintentional weight loss, and exhaustion. Handgrip strength was tested twice using a hydraulic Jamar dynamometer (Sammons & Preston, Bolingbrook, IL, USA), and the maximum score was recorded. Grip strength was considered low if was ≤29 kg in participants with body mass index (BMI) ≤24 kg/m², ≤30 kg in BMI 24.1–28.0 kg/m² and ≤32 kg in BMI >28.0 kg/m². For gait speed measurement, the participants were asked to walk at their usual pace along a 5-meter line on the floor and the time taken to complete the task was recorded. Slow gait was denoted as ≥6 s (for height ≤173 cm) and ≥5 s (for height >173 cm). Physical activity was assessed by the Physical Activity Scale for the Elderly (PASE) [17]. Participants who scored in the lowest quintile were categorized as having low levels of physical activity. Unintentional weight loss was defined as a self- report of ≥4.0 kg loss of weight unintentionally in prior year. Lastly, exhaustion was assessed using two items from the Center for Epidemiological Studies Depression Scale (CES-D) [18]: (1) “I felt everything I did was an effort” or (2) “I could not get going”. A response of “moderate amount of time (3–4 days)” or “most of the time (5–7days) last week” was classified as exhaustion. Participants were considered frail if they fulfilled three or more of these five criteria, and prefrail if one or two criteria were present.

Study covariates

Information on education, smoking and physical activity habits and history of hypertension, diabetes and heart failure were recorded based on self-reporting during an interview held by a study physician at the BIP study inception. Clinical variables (i.e. height, weight and blood pressure) were assessed by the interviewing physician. Blood glucose, c-reactive protein and creatinine were measured from blood samples drawn from each study participant at the baseline of the BIP study, before randomization. Information on use of medications including β-blockers was collected as part of the BIP trial. Samples were collected after at least 12 h of fasting, with the use of standardized equipment and procedures and transferred to a central study laboratory. All analyses were performed with a Boehringer-Hitachi 704 random access analyzer using Boehringer diagnostic kits. Serum creatinine levels were measured employing the Jaffe method without deproteinization. Glomerular filtration rate (GFR) was estimated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [19].

Statistical analysis

Data were analyzed using SAS version 9.4. Chi-square tests and one-way analysis of variance were performed to compare clinical and demographic characteristics between non-frail, prefrail and frail individuals. For non-normal distributed variables (i.e. c-reactive protein levels and eGFR), Kruskal–Wallis analysis was performed, and median and intra-quartile ranges were presented. These variables were log-transformed when entered the models as covariates.

Multinomial logistic regression was used to assess the relationship of RHR and visit-to-visit RHR variability with frailty status (i.e. no frailty, prefrailty and frailty). Odds ratios and 95% confidence intervals were calculated for change in 5 bpm and 5 units of heart rate variability.

Linear and logistic regression models were conducted to assess the relationship between heart rate measures and frailty components. Covariates were added to the regression models in 2 steps: first, age was entered to the model, then the following covariates were added: education, current smoking, body mass index, systolic blood pressure, prevalent hypertension, diabetes, heart failure, physical activity, BIP study arm, blood glucose, estimated glomerular filtration rate, c-reactive protein and use of β-blockers. Statistically significance was considered as p values below .05.

Results

The participants’ mean age was 56.6 ± 6.5 years at the BIP study inception and 77.0 ± 6.4 years at the late-life evaluation. Of the total sample of 306 men, 81 (26%) were frail and 117 (38%) were pre-frail. Pre-frail and frail vs. non-frail men were older, had a higher systolic blood pressure, a lower eGFR and were more likely to have heart failure, in a gradual manner (Table 1). Mean RHR among the BIP study participants (N = 3101) was 69.9 ± 13.6 bpm and 66.4 ± 7.8 bpm among the current study participants.

Table 1. Characteristics of study participants at BIP study inception (N = 306).

	Frailty status
	Non-frail	Pre-frail	Frail	p-value
N	108	117	81
Age, years	53.2 ± 5.2	56.5 ± 6.4	61.3 ± 5.0	<.001
Education, n (%)
Primary	13 (12)	23 (20)	21 (26)	.191
Secondary	56 (52)	57 (49)	35 (43)
Intermediary or University	39 (36)	37 (32)	25 (31)
Current smoking	9 (8)	8 (7)	6 (7)	.913
Prevalent diabetes	6 (6)	6 (5)	10 (12)	.111
History of myocardial infarction	24 (22)	27 (23)	18 (22)	.985
Physical activity	26 (25)	34 (30)	24 (30)	.645
Prevalent hypertension	70 (71)	81 (79)	65 (86)	.063
Prevalent heart failure	4 (3.7)	7 (6.0)	12 (14.8)	.012
Systolic blood pressure, mmHg	127 ± 16	126 ± 16	133 ± 18	.005
Diastolic blood pressure, mmHg	80 ± 10	80 ± 10	81 ± 9	.632
Body mass index, kg/m^2	25.9 ± 2.8	26.1 ± 2.8	26.6 ± 2.9	.222
C-reactive protein, mg/L	2.0 (1.1, 3.6)	2.4 (1.3, 4.6)	2.4 (1.6, 3.8)	.120
Total cholesterol, mg/dL	214.1 ± 17.5	214.3 ± 17.9	214.0 ± 18.5	.994
LDL-cholesterol, mg/dL	148.8 ± 16.6	150.0 ± 17.0	151.2 ± 17.8	.622
HDL-cholesterol, mg/dL	34.9 ± 5.3	34.8 ± 5.3	35.5 ± 4.7	.594
Blood glucose, mg/dL	96.3 ± 13.5	97.6 ± 16.0	99.1 ± 15.0	.445
eGFR	68 (62, 74)	65 (59, 71)	62 (58, 70)	.001
BIP Study Arm = Bezafibrate	50 (46.3)	53 (45.3)	47 (58.0)	.166
Use of β-blockers	44 (40.7)	41 (35.0)	26 (32.1)	.445

BIP: Bezafibrate infarction prevention; LDL: low density lipoprotein; HDL: high density lipoprotein.

Values are n (%), mean ± standard error or median (25^th percentile, 75^th percentile); Numbers in bold indicate p values of 0.05 or lower.

After controlling for age, each 5-bpm increase in RHR measured at midlife was significantly associated with 34% higher odds of being prefrail and frail in late-life (95% CI (1.11–1.60 and 1.07–1.67, respectively) (Table 2). Further adjustment for potential confounders including history of heart failure, use of β-blockers and physical activity, resulted in slightly increased odds, which also remained statistically significant (OR = 1.44 95% CI 1.15–1.79 and OR = 1.35 95% CI 1.03–1.77 for prefrailty and frailty, respectively). No statistically significant associations were observed between visit-to-visit HRV and frailty status (Table 2).

Table 2. Association of RHR and visit-to-visit HRV with late-life frailty status.

		RHR	Visit-to-visit HRV
Model	Frailty status	OR (95% CI)^a	OR (95% CI)^b
1	Non-frail	1.00 (ref.)	1.00 (ref.)
	Pre-frail	1.34 (1.11–1.60)	1.24 (0.68–2.26)
	Frail	1.34 (1.07–1.67)	1.19 (0.58–2.45)
2	Non-frail	1.00 (ref.)	1.00 (ref.)
	Pre-frail	1.44 (1.15–1.79)	1.22 (0.86–1.73)
	Frail	1.35 (1.03–1.77)	1.15 (0.75–1.78)

RHR: resting heart rate; HRV: heart rate variability.

Numbers in bold indicate p values of 0.05 or lower. Model 1: adjusted for age; Model 2: age, education, current smoking, body mass index, systolic blood pressure, prevalent hypertension, diabetes, heart failure, physical activity, BIP study arm, blood glucose, estimated glomerular filtration rate, c-reactive protein and use of β-blockers.

^a ORs are for increase in 5 bpm.

^b ORs are for increase in 5 units.

The associations between RHR and frailty components according to the Fried criteria are described in Table 3. Each 5-bpm increase in midlife RHR was associated with ∼1.12 kg lower grip strength and with additional ∼0.19 s to complete a 5-meter walk, independently of multiple potential confounders. No associations were observed between RHR and odds of having physical inactivity, unintentional weight loss or exhaustion (Table 3). Visit-to-visit HRV was not associated with any of the 5 frailty components (data not shown).

Table 3. Association of 5-bpm increase in RHR with frailty components (N = 306).

Measure	Model	β ± SE	p-value
Grip strength (kg)	1	−1.14 ± 0.30	<.001
	2	−1.12 ± 0.32	<.001
Gait speed (seconds)	1	0.19 ± 0.07	.007
	2	0.19 ± 0.08	.022
		OR (95% CI)
Physical inactivity	1	1.14 (0.97–1.34)	.104
	2	1.09 (0.91–1.31)	.338
Unintentional weight loss	1	1.00 (0.81–1.24)	.976
	2	1.11 (0.87–1.44)	.413
Exhaustion	1	1.13 (0.95–1.34)	.155
	2	1.21 (1.00–1.46)	.054

bpm: beats per minute; RHR: resting heart rate.

Numbers in bold indicate p values of 0.05 or lower. Model 1: adjusted for age; Model 2: age, education, current smoking, body mass index, systolic blood pressure, prevalent hypertension, diabetes, heart failure, physical activity, BIP study arm, blood glucose, estimated glomerular filtration rate, c-reactive protein and use of β-blockers.

Discussion

Our study demonstrates an association between increased resting heart rate in midlife and risk of pre-frailty and frailty ∼20 years later, in men with preexisting CHD. The associations were independent of cardiovascular risk factors, inflammatory markers and prevalent heart failure.

Increased heart rate, a marker of autonomic imbalance, is an established risk factor for cardiovascular and cerebrovascular morbidity and mortality [20] and has also been associated with other health conditions including cancer [21]. More recently, accumulating evidence demonstrated a link between RHR and incident diabetes, the metabolic syndrome and glucose dysregulation [22], and there is a resurrection of interest in RHR role in brain aging [23]. Together, these health conditions constitute disease burden which largely overlaps with frailty [16]. In addition to disease burden, some studies exist on the association between RHR and disability, another construct distinct from frailty yet strongly associated with it [16]. Findings from these studies show that increased heart rate is related to worse functional performance in older adults at high cardiovascular risk [11] and in patients with a recurrent stroke [24]. While some evidence exists on the association between indices of heart rate variability in the context of frailty [9,10], to our knowledge we are the first to assess the link between RHR and future frailty status.

The observed association between RHR and frailty in the current study was mainly driven by weak handgrip strength and slow gait speed, but not by other components of the Fried’s criteria. This finding explains the non-liner associations found in the current study between RHR and the degree of frailty (prefrail vs. frail), as prefrailty was defined as having at least two out of five frailty components. Weak handgrip strength and slow gait speed are indicative of muscle strength, a consequence of muscle loss and change in muscle composition in the upper and lower extremities, respectively [25]. In support of our findings, a recent cross-sectional study points to a link between autonomic dysregulation and loss of muscle mass in community-dwelling old-adults [26]. The authors hypothesized, that this association is driven by specific myokines, chemicals produced by muscle tissue, which affect biological processes in distant tissues, including on autonomic balance and heart rate [25]. Although such mechanism cannot be excluded as an explanation to our findings, it is more likely that an inverse direction exists, with autonomic imbalance occurring before frailty and not vice versa, as heart rate was measured in our sample in midlife, ∼2 decades before frailty was assessed. Increased RHR may be associated with frailty through the extent of atherosclerosis, as RHR does not only predict CVD, but may also indicate a more severe heart disease and a greater atherosclerotic burden [27]. In turn, atherosclerosis and its accompanied disturbances in the hematological and inflammatory systems are key pathologies underlying frailty [6]. Particularly, white matter disease and subclinical strokes have been related to disturbances in gait speed [28] and muscle strength [29]. These proposed mechanisms may be particularly relevant in men with preexisting CHD who have high vascular burden.

We found no association between visit-to-visit HRV and frailty. This may suggest that this measure is not a good indicator of HRV complexity. Yet, in contrast to our observation of no association with frailty, low visit-to-visit variation in RHR predicted adverse health outcomes among individuals with systolic heart failure in the SHIFT trial [30]. Moreover, temporal changes in RHR were linked with CVD morbidity and mortality in a community-based sample from the Atherosclerosis Risk in Communities Study [20]. However, the potential value and mechanisms of RHR variability across visits have been rarely studied in general, and particularly in the context of healthy aging.

Our analyses focused on men because men consisted the great majority of participants in the BIP study. While sex hormones may not be directly associated with frailty [31,32], sex differences in both autonomic regulation of the heart [33] and in degree of frailty[34] have been demonstrated. Thus, relationships between autonomic balance and frailty may differ between sexes and should be studied separately.

Our findings should be interpreted in light of several important limitations: first, analyses were cross-sectional, and a direct assessment of frailty risk was not feasible. Second, despite controlling for multiple potential confounders, we cannot rule out the possibility of residual confounding by yet unidentified factors. Specifically, we did not have information on measures of cardiovascular fitness, however, physical activity was controlled for in our models and have also shown no association with frailty in the current sample. Third, survival bias may affect our results, as participation in the late-life frailty assessment was limited to BIP study men who survived and were non-hospitalized.

In summary, this study supports the notion that elevated resting heart rate predicts adverse clinical events and expands previous literature to include implications on frailty outcomes. Additionally, our study supports the life-course approach to healthy aging [35], as RHR was assessed in midlife, many years before frailty was detected. Further research is needed to establish whether resting heart rate, a simple, noninvasive measure of autonomic control, can be useful in screening and identifying clinical vulnerability in high-risk CHD patients as well as in the general elderly population. Moreover, it remains to explore whether pharmacologic response to heart rate reduction can contribute to frailty risk stratification, and whether reduction of resting heart rate through either pharmacologic or behavioral interventions (e.g. physical activity) is beneficial in delaying frailty and prolonging health span.

Authors’ contributions

All authors contributed to study conception, data analysis and interpretation. All authors contributed to revising the manuscript for publication. All authors have approved the final version for publication.

Disclosure statement

No potential conflict of interest was reported by the authors.