Sex differences in heart rate responses to occupational stress

Abstract

Cardiovascular disease (CVD) remains the leading cause of disease burden globally and chronic stress is associated with increased risk of CVD. Recognition of chronic occupational stressors as a potential contributor to CVD highlights the need to recognize and prevent stress during work. The ubiquity of wearable technology devices to monitor health provides a new opportunity to noninvasively examine the cardiovascular system throughout a work shift. In the current study, we examined changes in heart rate (HR) during a work shift in a retail store setting using 23 healthy female and male subjects that differed in their physical fitness status. Subjects had their HR tracked via an Apple Watch during three typical work shifts. The results demonstrated an increase in HR during a work shift to a level observed during a moderate stressor (resting HR = 83.2 BPM ± 7.8; highest HR mean = 109.1 BPM ± 11.7; p < .0001). Female subjects demonstrated a significantly elevated maximum HR, a larger change in HR, and a larger percent change in HR compared with males (all p < .05). Physical activity status did not influence the observed changes in HR for females or males. Neither the time of day the work shift occurred nor the length of the shift modulated the observed pattern of HR changes. Collectively, our findings demonstrate the potential for wearables in biomedical research and personalized health.

Keywords:

• Stress
• sex differences
• cardiovascular system
• occupation
• wearable devices

Introduction

Homeostatic control of the cardiovascular system is critical for organisms in order to maintain optimal health (Ramsay & Woods, 2014). During times of stress, the cardiovascular system must respond appropriately in order to properly meet the demands of the stressor on the organism. For example, during an acute bout of physical activity, a number of physiologic adjustments must occur to provide the necessary supply of oxygen and nutrients to meet the demands of muscular work required by the activity (Rowell, 1993). A failure of the cardiovascular system to respond appropriately to this stressor can have detrimental impacts on the organism (e.g. the inability to escape a predator).

While turning on the cardiovascular system during a “fight-or-flight” situation via the autonomic nervous system (ANS) and hypothalamic–pituitary–adrenal axis (HPA) is an adaptive response in an organism (Dhabhar 2018; Fleshner and Crane 2017; McEwen 1998), chronic activation of the cardiovascular system can have deleterious consequences (Burford, Webster, & Cruz-Topete, 2017; DeMorrow, 2018; Sapolsky, 1999). For example, numerous studies indicate that exposure to chronic, daily stressors, and/or severe psychological trauma can also increase the risk of developing and dying from cardiovascular disease (CVD) (Cohen, Edmondson, & Kronish, 2015; Kivimäki & Steptoe, 2018). A recent meta-analyses of prospective observational studies found that social isolation and loneliness were associated with a 50% increased risk of incident CVD events, while the increased risk associated with work-related stress was similar at 40% (Steptoe & Kivimäki, 2013). Chronic stressors have also been linked to worse prognosis in patients with existing CVD (Carroll, Naylor, Marsden, & Dornan, 2003; Sabzmakan et al., 2014). Given that coronary heart disease and stroke remain the two leading causes of disease burden globally (GBD 2016 Causes of Death Collaborators, 2017), and chronic stress is associated with increased risk of CVD (Kivimäki & Steptoe, 2018), there is a need to identify sources of chronic stress as they might serve as complementary targets for disease prevention/therapeutic management.

Researchers have recently recognized that chronic occupational stressors such as adverse psychosocial working conditions, job strain, effort-reward imbalance, and long working hours are associated with a higher risk of CVD (Kivimäki et al., 2012), musculoskeletal disorders (Ariëns et al., 2001; Hoogendoorn, van Poppel, Bongers, Koes, & Bouter, 2000), in addition to depression and anxiety (Bernburg, Vitzthum, Groneberg, & Mache, 2016; Tanner, Bamberg, & Kozak, 2015). In addition to potential ill health effects, chronic occupation stress is also a cause for absenteeism from work resulting in financial implications (Houtman et al., 1999; Mauss, Li, Schmidt, Angerer, & Jarczok, 2015). Thus, it would be beneficial to be able to objectively monitor (and subsequently reduce) one’s physiological stress while working.

One increasingly popular tool to assess occupation stress is through monitoring changes in heart rate (HR) through wearable devices (Camm et al., 1996; Hernando, Roca, Sancho, Alesanco, & Bailón, 2018). In fact, altered HR variability has been reported in several diseases including cardiovascular diseases, such as ischemia, myocardial infarction and heart failure (Billman, 2011; Camm et al., 1996) and can indicate chronic stress (Hernando et al., 2016; Kim, Cheon, Bai, Lee, & Koo, 2018). Advances in the technology and accessibility of health related wearable devices has grown dramatically over the last decade (Canalys, 2018; Hernando et al., 2018) making it easier for nonpatient populations to monitor physiological measurements such as HR. Some of the most popular wearable devices are watch-like devices, which provide reliable HR estimates, derived from the pulse signal recorded at the wrist. The majority of these devices, such as the Apple Watch, are based on photophethysmography (PPG), which utilizes a light source that illuminates the skin and a photodetector that measures the intensity of the reflected light. HR estimates are based on the pulsatile changes in reflected light induced by fluctuations in blood flow every heartbeat (Hernando et al., 2018; Shcherbina et al., 2017).

Previous work has validated the use of HR assessment via the Apple Watch for stress monitoring (Bai, Hibbing, Mantis, & Welk, 2018; Hernando et al., 2018; Shcherbina et al., 2017) and demonstrated that wearable devices are effective in quantifying occupational stress in extreme environmental conditions (e.g. cold and heat stress situations) (Austad, Wiggen, Faerevik, & Seeberg, 2018). However, many of the studies involving examining changes in HR using wearable devices have been conducted in laboratory conditions with healthy subjects or in extreme conditions (Austad et al., 2018; Järvelin-Pasanen, Sinikallio, & Tarvanien, 2018). Less is known about the association between occupational stress and HR changes measured in real working life settings using wearable devices. In addition, a number of factors can influence HR responsiveness to stress, including age (Gerritsen et al., 2003; Koskinen et al., 2009), sex (Gerritsen et al., 2003, Koskinen et al., 2009), health and physical fitness status (Buchheit & Gindre, 2006; Dekker et al., 2000; Thayer, Åhs, Fredrikson, Sollers, & Wager, 2012), but few studies have examined these. Therefore, the aim of this study was to examine changes in HR during a work shift in an environmentally stable work environment (i.e. a retail store) using healthy female and male subjects that differed in their physical fitness status. The information obtained by this study may help in planning strategies for health assessment and stress reduction at work.

Materials and methods

Participants

Participants were employees of an Apple retail store (n = 23; 7 females and 16 males) aged 22-56 years who responded to in-store recruitment solicitations (Table 1). In order to assess physical activity status subjects were asked about their weekly physical activity levels. The American Heart Association suggests 150 minutes per week of moderate exercise or 75 minutes per week of vigorous exercise (or a combination of moderate and vigorous activity) for adults (AHA, 2016), thus subjects were categorized as either regularly meeting the physical activity guidelines (n = 9) or not regularly meeting the guidelines (n = 14) (Table 1). All subjects reported no chronic or acute illness, no regular medication regimen (with the exception of birth control), and good health prior to study onset. All procedures were approved by the University Institution Review Board (#1128772-3).

Table 1. Demographic characteristics of participants.

		Age
Characteristic	Number (%)	Range	Mean
Sex
Female	7 (30)	22–56	30.6
Male	16 (70)	23–42	30.1
Physically Active
Female	2 (29%)	22–26	24
Male	7 (44%)	22–34	27.5
Not Physically Active
Female	5 (71%)	22–56	33.2
Male	9 (56%)	22–43	32.1

Procedures

All participants completed an informed consent and regularly wore an Apple Watch, Series 1. Participants were asked to abstain from caffeine and smoking 45 minutes prior to beginning work. Heart rate was recorded throughout a normal work shift (mean shift = 7.5 h) during three separate work days at an Apple Store. Subject job titles and responsibilities included sales, support and leadership. The QS Access app was used to extract HR data from each subject. This app converts the XML format type used by the Apple Health app into a CSV file that can be exported. For each work shift, the lowest HR, highest HR, change in HR, and percent change in HR was analyzed for each subject.

Data analysis

In order to examine differences in HR initial analysis utilized a t-test, analysis of variance (ANOVA) or two-way (Physical activity status by Sex) ANOVA to determine if main effects or interactions were present between groups of participants. Alpha was set at 0.05. When statistical significance was determined with ANOVA, Fisher’s PLSD post-hoc analysis was conducted to examine which groups differed. All figures are shown with group means and standard deviation bars.

Results

Participant characteristics

Of the 23 subjects who completed the study, the majority were males, not physically active, with a mean age of 30 years (Table 1). Two-way ANOVA revealed no differences in age based on sex or physical activity status (Female, Physically Active (PA)=24, Female, Not PA = 33.2, Male, PA = 27.5, Male, Not PA = 32.1; all p > .05).

Resting HR

Resting HR was determined to be the lowest recorded heart rate at the start of the subjects’ shift. Two-way ANOVA indicated no significant differences in resting HR between any of the groups (Table 2 and Figure 1; all p > .05).

Figure 1. Lowest average observed heart rate (HR) during a work shift in subjects. PA = physically active; Not PA = not physically active. Bars are means ± standard deviations of the means. N = 23.

Table 2. Observed HR (BPM) throughout work shift.

	Lowest	Highest	Change	Percent change
Mean (SD) HR in all subjects	83.2 (7.8)	109.1 (11.7)*	25.9 (13.3)	32.1 (18.5)
Mean (SD) HR in females	82.5 (6.2)	114.0 (16.7)**	31.5 (20.1)**	39.5 (27.5)**
Mean (SD) HR in males	83.4 (8.4)	107.1 (8.4)*	23.6 (8.3)	29.1 (12.2)

* p < .05, lowest vs. highest.

** p < .05, lowest vs. highest and females vs. males.

HR changes during a work shift

During a work shift subjects demonstrated significant increases in HR (Table 2 and Figure 2). Unpaired t-test revealed that resting HR values (mean = 83.2 BPM ± 7.8) significantly increased (p < .0001) during the shift (highest HR mean = 109.1 BPM ± 11.7) in all subjects. Differences between the sexes were observed in both the highest HR observed (Figure 2) as well as the percent change in HR (Figure 3). While both female (p < .0001) and male (p < .0001) subjects demonstrated a significant increase in HR (Table 2), two-way ANOVA indicated that female subjects demonstrated a significantly elevated highest HR compared with males (Figure 2); F(1,64)=4.4, p = .03). Physical activity status, however, did not influence the highest HR observed (Figure 2); p = .8). Overall, subjects demonstrated a mean increase in 25.9 BPM in HR at some point during their shift which resulted in a mean percent change of 39.5 (Table 2). Two-way ANOVA revealed that females had a larger change in HR during a work shift compared with males (Table 2 and Figure 3(A), F(1,64)=3.8, p = .05); however, physical activity status did not modulate this change (p = .94). Similarly, ANOVA indicated that females had a larger percent change in HR during a work shift compared with males (Table 2 and Figure 3(B), F(1,66)=4.6, p = .03) while physical activity had no effect (p = .96). In the current study, work shift duration and time of day varied between and within subjects and, therefore, we also examined if temporal factors influenced the previously mentioned changes in HR. No differences were observed in HR changes when a more typical work shift time (9:00–17:00, e.g.) was compared with an atypical shift time (12:00–21:00) in the same subject nor in different subjects (data not shown). The length of the work shift also did not appear to modulate the observed changes in HR.

Figure 2. Highest average observed heart rate (HR) during a work shift in subjects. Unpaired t-test revealed that resting HR values (mean = 83.2 BPM ± 7.8) significantly increased (p < .0001) throughout the shift (highest HR mean = 109.1 BPM ± 11.7) in all subjects. Both female (n = 7, p < .0001) and male (n = 16, p < .0001) subjects demonstrated a significant increase in HR. Two-way ANOVA indicated that female subjects demonstrated a significantly elevated highest HR compared with males (p = .03). Physical activity status did not influence the highest HR observed (p = .8). PA = physically active; Not PA = not physically active. Bars are means ± standard deviations of the means.

Figure 3. Average change (A) and percent change (B) in heart rate (HR) during a work shift in subjects. Two-way ANOVA revealed that females had a larger change in HR throughout a work shift compared with males (Figure 3(A), p = .05), however, physical activity status did not modulate this change (p = .94). Females had a larger percent change in HR throughout a work shift compared with males (Figure 3(B), p = .03) while physical activity had no effect (p = .96). PA = physically active; Not PA = not physically active. Bars are means ± standard deviations of the means.

Discussion

Between 1990 and 2016, age-standardized life-years lived with ischemic heart disease declined by 47% and those of cerebrovascular disease by 17% (GBD 2016 Causes of Death Collaborators, 2016), in large part due to reductions in known cardiovascular disease factors such as smoking, blood pressure and serum cholesterol (Ezzati et al., 2015; Tzoulaki, Elliott, Kontis, & Ezzati, 2016). Despite these improvements, coronary artery disease and stroke remain the two leading causes of disease globally (GBD 2016 Causes of Death Collaborators, 2017) illustrating the need to identify new prevention targets. An emerging risk and prognostic factor for cardiovascular disease is chronic stress as mechanistic studies have corroborated previous observations demonstrating stress-related pathophysiological changes that underlie cardiovascular disease (Kivimäki & Steptoe, 2018; Steptoe & Kivimäki, 2012).

Researchers are increasingly recognizing occupational stress as one of the most commonly experienced stressors in adults (Kivimäki et al., 2012; Kivimäki & Steptoe, 2018). Fortunately, the popularity of wearable devices to monitor health has created a new noninvasive tool to assess occupational stress through monitoring changes in HR (Camm et al., 1996; Hernando et al., 2018). In the current study, we examined changes in HR measured via an Apple Watch during a work shift in an environmentally stable work environment using healthy female and male subjects that differed in their physical fitness status. Although previous work has validated the use of the HR assessment via the Apple Watch for stress monitoring (Bai et al., 2018; Hernando et al., 2018; Shcherbina et al., 2017) and for walking (Khushhal, Nicholas, Evans, 2017), this is the first study that we are aware of that examined occupational stress and HR changes measured in real working life settings using wearable devices.

Overall our results demonstrated that subjects experienced significant increases in HR during their work shift (Table 2, Figures 2, 3(A,B)). Heart rate values at rest (mean = 83.2 BPM ± 7.8) significantly increased (p < .0001) during the shift (highest HR mean = 109.1 BPM ± 11.7) at a level typically observed for this demographic during acute stress (Campisi, Bravo, Cole, & Gobeil, 2012; Hawley, Hargreaves, Joyner, & Zierath, 2014). Interestingly, although subjects’ lowest heart rates were not different among groups (Table 2, Figure 1), female subjects experienced the highest HR and the largest increase in HR (demonstrating a close to 40% change in HR) during t the work shift (Table 2, Figures 2, 3(A,B)). Numerous physiological differences exist between men and women including differences in height, weight, body composition (including fat and lean muscle mass), hormones (e.g. estrogen, progesterone, testosterone, etc.), and hemoglobin levels (Charkoudian, 2001; Charkoudian & Joyner, 2004). In addition, sex differences in response to physiological and psychological stressors have been reported in rodents and humans suggesting that females respond more robustly to an acute stressor than males (Heck & Handa, 2018; Oyola & Handa, 2017). The results from the current study are consistent with these findings but are the first that we are aware of that demonstrate a differential response to occupational stress (as indicated by changes in HR) in females during a normal work shift. It will be interesting to examine other indicators of stress (e.g. blood pressure, cortisol levels) in similar situations to further corroborate our findings.

Considerable data indicate that the physical fitness status of an organism modulates that organism’s response to a stressor (Campisi & Fleshner, 2003; Campisi et al., 2003; Fleshner et al., 2002; Fleshner, Greenwood, & Yirmiya, 2014; Lloyd et al., 2017). Therefore, we were surprised to observe in the current study that physical activity status did not influence the highest HR observed (Figure 2) nor the change in HR observed throughout the work shift (Figure 3(A,B)). One explanation for this finding was that we relied on subjects’ responses to a survey which might not have been reliable. Furthermore, it is possible that the use of AHA guidelines to classify subjects based on physical activity did not result in significant physiological differences in physical fitness between the subjects. A more direct, quantifiable measure of fitness status (e.g. VO₂ max levels) or body composition (e.g. BMI) might have resulted in different results. In addition, we did not collect HR data during a resting (stabilization) time which might have provided for interesting comparisons. Finally, it is possible that a larger sample size might uncover a significant finding.

Because work shift duration and time of day varied in the current study we also examined if temporal factors influenced the previously mentioned changes in HR. No differences were observed in HR changes when a more typical work shift time (9:00–17:00 pm) was compared with an atypical shift time (12:00–21:00) (data not shown). The length of the work shift also did not appear to modulate the observed changes in HR. Given that circadian changes in hormone levels differ in females in males and that these hormones can influence HR (Charkoudian, 2001; Charkoudian & Joyner, 2004) and stress responses (Heck & Handa, 2018; Oyola & Handa, 2017) these were surprising results. We did not monitor food, beverage, or stimulant intake throughout the work shift or measure the type of interactions subjects had with customers or employees. It is possible that these external factors might have influenced hormonally modified changes in HR.

The results from the current study add to a growing literature demonstrating occupational stress as a potential modulator of the cardiovascular system and indicate that females might respond to occupational stress differently compared with males. Our analysis indicates that wearable metrics can potentially help in planning strategies for health assessment and stress reduction at work. As public adoption of wearable activity trackers continues to increase (Chuah et al., 2016), our findings demonstrate the potential for wearables in biomedical research and personalized health.

Acknowledgements

The authors wish to thank our subjects for participating in the study as well as the Apple Store in Boulder, CO for allowing us to conduct our study.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Bethany Lucas

Bethany Lucas is an assistant professor.

Stella Grayson

Stella Grayson, Halimah Hamidu, Andrew Han, Sandra No, and Ajay Varghese are recent graduates of the Master of Science in Biomedical Sciences program at Regis University.

Halimah Hamidu

Stella Grayson, Halimah Hamidu, Andrew Han, Sandra No, and Ajay Varghese are recent graduates of the Master of Science in Biomedical Sciences program at Regis University.

Andrew Han

Stella Grayson, Halimah Hamidu, Andrew Han, Sandra No, and Ajay Varghese are recent graduates of the Master of Science in Biomedical Sciences program at Regis University.

Sandra No

Stella Grayson, Halimah Hamidu, Andrew Han, Sandra No, and Ajay Varghese are recent graduates of the Master of Science in Biomedical Sciences program at Regis University.

Ajay Varghese

Stella Grayson, Halimah Hamidu, Andrew Han, Sandra No, and Ajay Varghese are recent graduates of the Master of Science in Biomedical Sciences program at Regis University.

Jay Campisi

Jay Campisi is an associate professor at Regis University.