Exploratory Study of Heart Rate Variability and Sleep among Emergency Medical Services Shift Workers

Abstract

Objectives: To characterize the continuity and duration of sleep, and to measure nocturnal cardiac autonomic balance via heart rate variability (HRV) in a group of emergency medical technicians (EMTs) on and off duty. Methods: Fourteen EMTs completed an online, daily sleep log that recorded total sleep duration, bedtime, rise time, and the number of alarms that caused awakening. HRV was captured using a physiological status monitor (PSM) affixed to a chest strap during sleep. Results: For the 7-day trial, each of the 14 EMTs logged three work days (WDs) and four non-work days (NWDs). They reported sleeping significantly fewer hours per night on WDs (6.4 ± 2.1) than on NWDs (7.9 ± 0.5; P < 0.05), and experienced more sleep disruptions on WDs (4.4 ± 2.8) than on NWDs (1.3 ± 2.2; P < 0.001) as measured by the number of alarms. Global and vagal indices of HRV during sleep were significantly reduced during WDs (Standard Deviation of Normal R-R Intervals (SDNN) = 43.4 ± 2.0 ms and High Frequency (HF) = 24.3 ± 1.2 ms²) when compared to NWDs (SDNN = 61.1 ± 1.0 ms and HF = 42.7 ± 1.5 ms²; P < 0.001). Conclusion: EMTs who worked 24-hour shifts had shorter, more fragmented sleep associated with greater cumulative exposure to increased sympathetic and decreased parasympathetic activity as measured via sleep HRV. These changes in cardiac autonomic tone constitute one plausible pathway through which sleep deprivation may increase risk for cardiovascular disease.

Key words::

• EMT
• sleep
• heart rate variability

Introduction

Emergency medical technicians (EMTs) are responsible, on very short notice, for assessing, treating, and transporting the injured and critically ill from hazardous or unpredictable situations. In addition to their care-giving role, EMTs must be capable of responding in new and unknown locations at any time of day, dealing with potentially challenging patients and bystanders, all while working long shifts that can last more than 24 hours. Despite these stressful conditions, EMTs must remain alert and vigilant to thwart potential errors in protocol and subsequent injury to citizens or each other. Apart from occupational health risks, such as vehicle collisions, EMTs appear to be particularly susceptible to cardiovascular disease and other significant comorbidities of chronic stress such as obesity and diabetes.¹ Several studies have suggested that these occupational health risks are significantly influenced by lifestyle behaviors.^1-3 A 2010 survey revealed that 26% of EMTs are obese, 17% are smokers, and 75% did not meet the level of daily physical activity recommended by the Center for Disease Control.¹ A study by Patterson et al.², found that 59% of EMTs reported having some sort of significant health problem such as diabetes or hypertension. Additionally, EMTs report that their work is more emotionally demanding than other health occupation personnel.¹

Problems sleeping, including both a lack of sleep and low-quality sleep, constitute additional factors in the lifestyle of EMTs that can be detrimental to their health—both on and off duty.² One study that analyzed ambulance crash statistics showed that sleep problems serve as one of the most significant predictors of whether an EMT would be involved in a vehicle collision while driving an ambulance.⁴ Furthermore, there is a significant association between EMTs who suffer from fatigue and the occurrence of medical errors, injuries, and other adverse events.⁵ Seeing how EMTs often work extremely long, uninterrupted shifts, sleep deprivation is a possible outcome resulting from work-related fatigue or other sleep problems. These, in turn, can manifest into other more serious health conditions. Multiple clinical studies indicate that sleep deprivation, fragmented sleep, and decreased sleep efficiency play key roles in the development of heath conditions such as diabetes, metabolic syndrome, and cardiovascular disease.^6-9 While cardiometabolic disorders are extremely complicated and often have a plethora of intertwining causes, recent research has suggested a strong connection between the development of major risk factors and poor sleep. For example, Altman et al.¹⁰ and Grandner et al.¹¹ examined both poor sleep duration and sleep insufficiency. They found that individuals whose sleep was regularly characterized by one or both sleep problems were at an increased risk of both hypertension and hypercholesterolemia, chief contributors to cardiovascular disease.

One plausible mechanism by which both acute total and chronic partial sleep deprivation can be linked to cardiovascular disease among EMTs and other first responders is via its effects on cardiovascular reactivity and autonomic nervous system (ANS) balance.^7,8 The ANS exerts control over the heart through the dynamic interplay of sympathetic and parasympathetic neural activity at the sinus node and this can be measured by heart rate variability (HRV). HRV reflects this complex interaction through variations between heartbeats (R-R intervals) over time.¹² Healthy ANS function is reliant on pronounced vagal modulation.¹² Population studies such as the Framingham Heart Study have found that ANS dysfunction, consisting of hyperactive sympathetic and/or diminished parasympathetic modulation, results in low HRV and is more strongly associated with an increased risk of cardiovascular pathology than any other cardiovascular disease marker.^6,7,13,14 Moreover, recent evidence suggests that HRV may be able to predict cardiovascular events in individuals before symptoms of cardiovascular diseases appear.⁷ These studies underscore the significance of HRV in assessing cardiac health in at-risk populations such as EMT personnel.

HRV, stress, and sleep deprivation have been shown to influence one another.^15,16 Given the stressful lifestyle led by EMTs, compounded by recent evidence of poor nocturnal sleep,^2,4 it is plausible that this pathway may be a significant contributor to the prevalence of cardiovascular disease observed in this population. The purpose of this study was to characterize the duration of sleep, the amount to which sleep was interrupted, and to measure nocturnal cardiac autonomic nervous system balance via HRV in a group of EMTs both while on and off duty.

Methods

Participants

Fourteen male EMTs (age 27 ± 7 years; BMI 25.9 ± 3.0 kg•m²) participated in this study. Eight were recruited from the University of California, Los Angeles (UCLA) Emergency Medical Services, three from a metropolitan city fire department in Arizona, and three from a metropolitan city fire department in Nevada. We advertised this study to these two fire departments because we had conducted previous research with firefighters based there (see Batalin et al.).¹⁷ All participants were recruited on a volunteer basis through an e-mail advertisement. The study was approved by the UCLA Institutional Review Board and all participants gave written informed consent.

Remote Data Capture

Prior to study initiation, each participant completed a detailed, step-by-step instructional tutorial on how to use the UCLA Exercise Physiology Research Laboratory Digital Health Network (DHN) architecture including measurement of HRV. Using a secure networked system for remote data capture, developed by the UCLA team and previously described by Batalin et al.,¹⁷ participants were able to capture HRV and sleep biometric data and transfer it to an encrypted web portal from their home or firehouse (Figure 1).

Figure 1.  The DHN architecture used to capture HRV and sleep biometric data from the EMT's home or work. Prior to sleep, participants fitted a PSM affixed to a chest strap and uploaded the subsequent data via Bluetooth to a smartphone then to an encrypted web portal that enabled researchers to analyze the HRV and sleep data.

Heart Rate Variability

The participants fitted themselves with a physiological status monitor (PSM) affixed to a chest strap (BioHarness-3; Zephyr Technologies, Annapolis, MD), which they wore during sleep for seven consecutive nights that included both work days (WD) and non-work days (NWD). Over the course of this week, every EMT's schedule consisted of three WDs—each WD consisting of a 24-hour shift—and four NWDs. The three WDs were not necessarily back-to-back-to-back; some participants worked three days in a row while others worked two in a row, took a day off, and then worked a third day. WD sleep is defined as any amount of sleep the participants received while on duty. NWD sleep, on the other hand, is defined as any period of sleep the participants received while off duty. While on shift, the EMTs were permitted to sleep between the hours of 11 pm–7 am. The sleeping quarters consisted of two beds, a television, and a landline telephone that rung when the EMTs needed to respond to an emergency. On WDs, this telephone served as an alarm; on NWDs, a personal alarm, if any, set by the participants (e.g., alarm clock, cell phone) fulfilled this role.

Data capture was initiated as soon as the participant was lying in bed with the room darkened and continued until they awoke. This process occurred for every episode of sleep regardless of its duration. The PSM included a single-channel electrocardiogram (EKG) sensor sampling at 250 Hz with the R-R intervals (ms) being calculated on a beat-to-beat basis using proprietary PC-based application. The data were exported as a text file to the HRV analysis software (Kubios Heart Rate Variability Software Version 2.0; Biosignal Analysis and Medical Imaging Group, Department of Physics, University of Kuopio, Kuopio, Finland). Data processing followed standard procedures described in the Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology.¹⁸ Raw R-R intervals were edited so that artifacts and non-sinus beats could be replaced by interpolation from adjacent normal R-R intervals. The spectrum for these R-R intervals was calculated with Welch's periodogram method (fast Fourier transform spectrum) with a window width of 256 s and overlap of 50%. The cleaned signal was then used to provide normal-to-normal (N-N) intervals to compute time and frequency domain HRV parameters. Time domain analyses included standard deviation of normal-to-normal intervals (SDNN) expressed in milliseconds (ms), which is a global index that reflects long-term measures of HRV.¹⁸ Frequency domain analysis included the high frequency (HF) component (frequency 0.15-0.4 Hz) in absolute units (ms²), generally defined as a marker of vagal/parasympathetic modulation.¹⁸

Sleep Log

Each participant completed an online, daily sleep log that documented total sleep duration, bedtime, rise time, and the number of alarms that caused awakening. All data was self-reported immediately after waking from a period of sleep.

Statistical Analysis

For analysis, data from the DHN were exported to a statistical software packages (Excel; Microsoft Corporation, Redmond, WA; JMP, SAS Institute, Inc., Cary, NC). Baseline characteristics for the EMT's were described as mean (±SD). WD and NWD data for HRV indices and sleep metrics were compared using paired t-tests. Statistical significance was set to p < 0.05.

Results

For the 7-day trial, three WDs and four NWDs were logged equally among the 14 EMTs. As shown in Table 1, participants reported sleeping significantly fewer hours per night on WDs (6.4 ± 2.1) than on NWDs (7.9 ± 0.5; P < 0.05). They also experienced more sleep disruptions on WDs (4.4 ± 2.8) than on NWDs (1.3 ± 2.2; P < 0.001) as measured by the number of alarms that caused awakening. Participants that were awoken by multiple alarms on NWDs initially ignored (i.e., hit the “sleep” button) their original, single alarm. As displayed in Table 2, global and vagal indices of HRV during sleep were significantly reduced during WDs (SDNN = 43.4 ± 2.0 ms and HF = 24.3 ± 1.2 ms) when compared to NWDs (SDNN = 61.1 ± 1.0 ms and HF = 42.7 ± 1.5 ms; P < 0.001) (Figure 2 and Figure 3). All subjects were 100% compliant in completing all sleep logs and wearing the PSM during sleep episodes.

Figure 2.  This graph displays each EMT's mean SDNN value over the course of one week, which was comprised of three WDs and four NWDs. The mean of all participants’ SDNN values for both WDs and NWDs are shown by their respective horizontal lines across the data. The difference between the means for all EMTs on a WD (43.4 ± 2.0) vs. a NWD (61.1 ± 1.0) is statistically significant (p < 0.001).

Figure 3.  This graph displays each EMT's mean HF value over the course of one week, which was comprised of three WDs and four NWDs. The mean of all participants’ HF values for both WDs and NWDs are shown by their respective horizontal lines across the data. The difference between the means for all EMTs on a WD (24.3 ± 1.0) vs. a NWD (42.7 ± 1.0) is statistically significant (p < 0.001).

Table 1.   Sleep characteristics of EMTs (n = 14) on WDs and NWDs

	Work Days	Nonwork Days
	Mean (±SD)	Mean (±SD)	p-value
Bedtime (military hour)	23.8 ± 0.4	22.4 ± 0.5	p = 0.035
Rise time (military hour)	7.5 ± 0.6	7.4 ± 0.7	p = 0.450
Total sleep time (hours)	6.4 ± 2.1	7.9 ± 0.5	p = 0.046
Number of alarms*	4.4 ± 2.8	1.3 ± 2.2	p < 0.001

*On WDs, the alarms woke participants from sleep between 11 pm–7am (night and early morning). On NWDs, the alarms woke participants from sleep no earlier than 6 am (morning and late morning).

Table 2.   Sleep HRV indices of EMTs (n = 14) on work days (WDs) and non-work days (NWDs)

	SDNN (WD)	SDNN (NWD)	HF (WD)	HF (NWD)
EMT	(ms)*	(ms)	(ms^2)	(ms^2)
1	43.7	62.7	22.1	42.4
2	42.2	60.0	22.5	41.9
3	42.0	62.6	22.8	44.7
4	42.7	60.3	24.5	41.6
5	43.5	61.6	23.6	42.9
6	42.9	61.9	25.2	42.5
7	43.5	61.6	22.6	42.5
8	43.5	60.0	23.6	42.4
9	42.2	62.2	24.4	44.7
10	42.7	61.1	24.5	41.4
11	41.7	61.5	24.0	42.8
12	43.3	62.1	22.3	42.3
13	43.1	61.8	24.6	42.5
14	43.5	60.7	23.9	41.9
Mean ± SD	43.4 ± 2.0	61.1 ± 1.0	24.3 ± 1.0	42.7 ± 1.0

*Each EMT's standard deviation of normal R-R intervals (SDNN) and high frequency (HF) value over the course of one week was comprised of three WDs and four NWDs.

Discussion

Our study found that EMTs’ on-shift sleep was significantly shorter and more fragmented than off-shift sleep. Furthermore, WD sleep was marked by increased sympathetic and decreased parasympathetic activity as indexed by HRV when compared to NWD sleep. The sleep HRV data captured during NWDs fell within the normal expected range of values when compared to age- and gender-matched norms obtained from Dolezal et al.¹⁹ Our data suggest that shorter, more frequently interrupted sleep is linked to reduced HRV, which may serve as a possible mechanism between sleep deprivation and cardiovascular disease. However, alternative explanations cannot be fully disregarded. For example, EMTs know that they will likely be woken up from sleep during WDs—this increased level of vigilance may cause increased sympathetic activity and decreased vagal tone, which may lead to reduced sleep HRV. Conversely, EMTs may be more relaxed during NWD sleep, exhibiting decreased sympathetic and increased parasympathetic activity, which may account for the higher sleep HRV than what was observed on a WD. Future studies ought to explore the deeper connection between how a greater number of sleep disruptions may negatively impact sleep HRV.

EMTs and other first responders participate in a variety of stressful situations ranging from critically ill patients requiring immediate treatment to environmental hazards that create unsafe conditions for workers and the general public. For the safety of EMTs themselves and the victims they serve, being physically and mentally fit for duty is essential. Cardiovascular reactivity, which can be measured by HRV, plays a key role in stress responses,²⁰ risk for a cardiovascular event^6,7,14,15 and, more recently understood, sleep quality.¹⁶ The ability of the body to perceive and react rapidly to certain stimuli is essential for good health and survival. Equally important, however, is the ability to reduce sympathetic input once life-threatening situations have subsided. Failure to do so has been shown to significantly increase the risk of chronic cardiovascular diseases (e.g., hypertension, atherosclerosis) and cardiovascular events.^6,7,13-15 Given the high incidence of cardiovascular events that occur among EMTs and other emergency responders,^1,21,22 as well as our finding of impaired HRV in EMTs during a shift, it is of great importance that the health implications of cardiovascular reactivity, stress, and sleep quality be better understood in this population.

Although there is a paucity of research on heart rate variability and cardiovascular reactivity in emergency responders, our results are concordant with several other studies that investigated EMTs and similar shift work occupations. Aasa et al.²³ revealed that among ambulance workers in Sweden, there was a significant difference in HRV on work days and non-work days, especially in individuals who reported having several medical problems. An important distinction between our study and Aasa et al.,²³ however, is the amount of uninterrupted sleep each study's population of ambulance workers received. The ambulance workers studied by Aasa et al.²³ exceeded 5 hours of uninterrupted sleep each night while on shift; those in our study frequently did not. Although the EMTs in our study averaged 6.4 hours of sleep on shift, this was frequently (73% of the time) interrupted by a call. A different study examined the paired effects of sleep deprivation with a bedtime that differed by 8.5 hours each day.²⁴ After eight days, participants exhibited higher nocturnal heart rates and decreased vagal indices of sleep HRV, suggesting that circadian misalignment may enhance the negative impact sleep deprivation has on cardiovascular health. Similar to our study, Borchini et al. monitored HRV in a group of nurses during WDs and NWDs.²⁵ These investigators found that HRV, measured via a 24-hour EKG, was impaired (as defined by significantly lower SDNN) on WDs for all participants. Additionally, it was found that nurses with high job strain, characterized by a survey, had continuing impairment of HRV on NWDs when compared to their coworkers with lower job strain. It is relevant to note that HRV was monitored for 24 hours as opposed to only during sleep, indicating that, on average, impaired HRV can be detected at all times of the day. Given the similar lifestyle and occupational stress between nurses and EMTs, it is certainly possible that HRV is impaired during the day in EMT populations as well following a shift.²⁶ Factory workers on a 24-hour rotating shift schedule have shown a similar pattern of HRV deterioration following a nightshift.^26,27 Additionally, it was found that blood pressure did not return to resting levels for more than 12 hours, likely due to sympathovagal imbalance.²⁷ These results are consistent with our finding of sympathovagal imbalance measured via sleep HRV during a 24-hour ambulance shift.

When comparing our sleep duration outcomes with other studies, agreement was equivocal. Similar to the findings of this study, Chung et al.²⁸ analyzed a group of nurses with rotating schedules (consecutive night shifts followed by consecutive day shifts with break days in between) and reported that total sleep time was significantly decreased during daytime sleep in nurses coming off of a night shift. Additionally, it was found that LF/HF was significantly lower during nocturnal sleep after NWDs. Chung et al.²⁸ theorized this was an indication of the body's attempt to recover from occupational stress, and suggested that two days off between shifts may be sufficient for full recovery. A study by Patterson et al.²⁹ found that improved recovery between work shifts (self-reported based on questionnaires) was associated with shifts lasting longer than 12 hours, regardless of whether they were during the day or the night. It is likely that individuals who work shifts longer than 12 hours need a longer recovery time between shifts than those who work shorter shifts in order to return to work less stressed, more rested, and revitalized. With these results in mind, an area of future study could focus on the difference in fatigue, sympathovagal imbalance, and sleep duration in EMTs that work a small number of shifts per week lasting longer than 12 hours versus EMTs that work significantly more shifts per week but of a shorter duration.

Limitations of this study include a small sample size and short time interval over which data was captured. Participants were recruited on a volunteer basis; therefore, there is a potential for sampling bias. Additionally, we did not objectively measure sleep time, and while self-reported sleep duration and the number of times awoken during sleep are important metrics, they may not fully capture the extent of sleep fragmentation. How precisely this might impact the performance of the self-report items is not clear, however, and additional validation work in this area would enhance interpretation of our findings. The population of EMTs studied was young men, which should be considered when interpreting the results. However, one might expect a younger population to be more resilient to stress and sleep deprivation. With this in mind, our results further emphasize the need for further study in this area as the true risk to this population may be greater than that indicated by this study.

Heart rate variability, a strong indicator of cardiovascular health, may be an important factor in the ability of EMTs to adapt to occupational stress. Adopting wellness initiatives including exercise training, stress management, and sleep hygiene may be good proactive strategies for EMTs to reduce the occupational stress incurred from long work shifts. Our results are consistent with several other studies that indicate sleep deprivation and stress play a key role in impairment of heart rate variability and risk for cardiovascular disease.^21-23,26,27 Further research, however, ought to be conducted in order to better understand the roles sleep and HRV play in assessing cardiovascular risk in emergency responders.

Conclusion

This study showed that EMTs reported shorter, more frequently interrupted sleep on WDs compared to NWDs. Furthermore, these reports were associated with greater cumulative exposure to increased sympathetic and decreased parasympathetic activity while on-shift as measured via sleep HRV. These changes in cardiac autonomic tone constitute one plausible pathway through which sleep deprivation may increase risk for cardiovascular disease.