Effects of shift work on sleep quality and cardiovascular function in Taiwanese police officers

ABSTRACT This study aimed to investigate the effects of shift work on sleep quality, cardiovascular function, and physical activity (PA) levels in Taiwanese police officers. Twenty-one male police officers aged 26.9 ± 4.1 years old located in Taipei voluntarily participated in this study. The participants completed the resting heart rate (HR) and hemodynamic variables (e.g. blood pressure, BP) before and after day-time (DTW) and night-time (NTW) shift work phases (5 working days and 2 resting days for each phase). Additionally, an actigraphy was administered to measure PA and sleep patterns in the last 3 working days. The average total sleep time and sleep efficiency were 278.5 ± 79. 6 min and 72.9 ± 10%, respectively, in the NTW phases, which were significantly lower than that in the DTW phases. A comparison of the PA characteristics between the two phases revealed that a lower proportion of moderate-vigorous PA (1.2 ± 0.8%) and a greater proportion of sedentary behaviour PA (74.8 ± 6.4%) was found in the NTW phases. The results of hemodynamic measures demonstrated that the police officers have significantly elevated systolic BP by 3.3% and diastolic BP by 3.9% after the NTW phases. Furthermore, the NTW phases exhibited a significantly higher percentage change ratio of systolic BP and diastolic BP compared to the DTW phases. Compared with the DTW phases, the NTW phase was significantly more likely to report higher decreasing parasympathetic-related HR variability with a range of −5.9% to −7.8%. In conclusion, night-time shift work resulted in negative physiological changes leading to adverse effects on the health and well-being of Taiwanese police officers.


Introduction
In many occupations, workers are required to undertake regular/irregular shift work, which has been shown to worsen their health status significantly (Ohayon et al. 2002).Regular shift work was reported to influence the circadian rhythm and sleep patterns (i.e., 24-h biological clock) of workers due to disturbances in neural-hormonal control axes (e.g., melatonin and cortisol) (Silva and Costa 2023).Subsequently, shift work leads to deterioration in cardiovascular function, potentially increasing the risk for metabolic syndrome (Chellappa et al. 2019;Sooriyaarachchi et al. 2022), which may be further exacerbated by poor sleep quality (Jeong et al. 2019;Ramey et al. 2012).
Among shift workers, the pivotal role of police enforcers in safekeeping has been documented to influence sleep hygiene (Neylan et al. 2002).Ramey et al. (2012) reported that night time shift work (NTW) contributed to shorter sleep duration (<6 hours/day) in male police officers.Similarly, Chang et al. (2015) recorded shorter sleep duration and poor sleep quality among Taiwanese police officers during NTW period.Long-term sleep disturbances among police officers may raise health concerns in the workforce (Grandner et al. 2016;Ma et al. 2013).
Shift work also affects physical activity (PA) behaviours in various occupations (Loprinzi 2015).For example, Khan et al. (2021) reported greater step counts during NTW than that during day-time shift work (DTW) in Australian paramedics.Conversely, high PA level were observed during DTW in vehicle manufacturing workers (Kawada et al. 2008).Furthermore, Ma et al. (2011) reported hard-intensity PA (measuring by self-report) during the afternoon shift among male police officers.Discrepancies in work activity between DTW, afternoon shift, and NTW may contribute to different occupational PA levels and sedentary behaviours (Hulsegge et al. 2017), influencing sleep and cardiovascular health (Park and Suh 2020;Wahid et al. 2016).Alruwaili et al. (2023) indicated positive relationship between physical activity (PA) on sleep quality among adults.The monitoring days of estimated physical activity using an accelerometer among adults were required at least 3 to 5 days (Trost et al. 2005).Owing to a lower percentage of missing records and a higher correlation among activity levels on weekdays, imputation may work better on weekdays than weekend days (Ward et al. 2005).To our knowledge, research on the potential association between shift work and PA among police officers is still limited.Hence, we collected the PA data and evaluated the effect of shift work on PA among police officers.
Indeed, shift work has been documented to attenuate cardiovascular indices.In vascular health, it has been reported that NTW results in blunted blood pressure (BP) (Patterson et al. 2021).In term of autonomic functions, Lee et al. (2015) reported that male night shift automobile manufacturing workers reduced time and frequency domain indices of heart rate variability (HRV), compared to the DTW.The impact of shift work on cardiovascular health among police officers has been received a great concern in occupational and work health (Holst et al. 2019).
To the best of our knowledge, there has been a scarcity of literature that employs objective assessment of physical activity and sleep among police officers.This information may provide valuable insights, facilitating informed decisions for health promotion in police officers due to the specificity of operational performance (Waggoner et al. 2012).Therefore, this study aimed to investigate the effects of weekly shift work on sleep, physical activity, and cardiovascular health among metropolitan police officers.A pre-and-post repeated measure design was implemented to examine the effects of shift work on measuring variables.It was hypothesized that DTW would demonstrate better sleep, PA, and cardiovascular indices compared to NTW.

Study question
This study protocol employed a within-subject comparison with an experimental period of 2 weeks.The police officers were exposed to two different work schedules: one DTW phase followed by one NTW phase.Each phase was designed 5 working days and 2 resting days.The participants completed the 10-min resting HR, and hemodynamic measures (e.g., BP) before and after the DTW and NTW phases [i.e., 5 working days (Day 1-5 and Day 8-12) and 2 resting days (Day 6-7 and Day 13-14)].Additionally, an ActiGraph accelerometer was worn on the wrist of the non-dominant hand for measuring individual PA and sleep patterns in the last 3 working days of each phase.Due to the varied work schedules of all participants, data collection commenced with the DTW, followed by the NTW phases for all participants.

Participants
Initially, 25 police officers, serving at 5 different regional police stations in Taipei, voluntarily participated in this study.Due to conflicts in personal schedules during the study period, four participants dropped out, with the remaining 21 participants completing all study aspects.The inclusion criteria for participants were: 1) field police officers in Taipei; and 2) field shift-work experience of more than 6 months.The exclusion criteria for participants were: 1) the presence of apnea, periodic limb movement disorder, or CVD; 2) taking BP or diuretics drugs; and 3) female police officers (to prevent the potential influence of the menstrual cycle on biological/physiological measures in this study).All participants provided informed written consent before the study.This study was approved by the Human Research Ethics Committee of the University (approval number TMU-JIRB-N201811044) and was conducted by the Declaration of Helsinki and its later amendment.

Experimental procedures
Participants initially completed a questionnaire to provide their demographic information, including educational background, weekly exercise levels, and smoking and alcohol consumption habits, followed by a familiarization session at their relevant police stations.Approximately 2-3 days later, participants completed a comprehensive 1-hour assessment in a private room of the police station before and after the DTW phase as well as the NTW phase (i.e., pre-DTW, post-DTW, pre-NTW, and post-NTW).For each session, participants completed a 1-min recording of resting hemodynamics, a 10-min seated HR recording, and setup of the accelerometer.To monitor sleep patterns and PA levels during Days 3-5 and Day 10-12, participants wore an ActiGraph accelerometer on the wrist (i.e., 72 recording hours).All sessions commenced at the same time of day (08:00) in March-June 2022 (see Figure 1).

Objective sleep quality and daily physical activity
A 3-axis accelerometer (ActiGraph LLC wGT3X-BT, Pensacola, US) was used to evaluate participants' sleep patterns and daily PA (Martin and Hakim 2011).The participants wore the accelerometer on their nondominant hand for consistency of data collection (Dieu et al. 2017).The participants could remove the wrist-worn accelerometer while showering only during each phase.A customized software package (ActiLife LLC Pro version 6.13.4,Pensacola, US) was used to process the accelerometer's raw data for subsequent analysis.The sample rate and the epoch of activity counts were set at 50 Hz and 1 min, respectively.The sleep parameters analysed by the ActiLife software were processed using the Sadeh algorithm for healthy young adults (Sadeh et al. 1994).The examined sleep parameters included: total sleep time, sleep latency, sleep efficiency, and wake after sleep onset.Additionally, PA epochs were categorized as sedentary behaviour, light intensity PA, moderate intensity PA, or moderatevigorous intensity PA (Sasaki et al. 2011).The intraclass correlation coefficient (ICC) of the ActiGraph accelerometer was reported to be excellent for total PA (ICC = 0.97).Conversely, its reliability was reported to be unaffected by the frequency of dynamic movements (Santos-Lozano et al. 2012).All sleep and PA results were analysed and expressed as the average of each phase's last 3 working days.

Hemodynamic measures
A portable MTX device (MTX, Cnoga Medical Ltd., Israel) was used to assess hemodynamic parameters.The MTX consisted of a color image sensor (spectral range of ~ 380 to ~1000 nm), light emitting diodes (four monochromatic light sources in the visual to infrared spectrum, ~600 to ~1000 nm), and a digital signal processor for data acquisition of the hemodynamic values (Segman and Sheiman 2018).While seated, participants inserted their left middle finger into a finger compartment of the MTX device, which was placed on a table at the level of the heart.The hemodynamic parameters measured included systolic BP, diastolic BP, blood oxygen saturation (O 2 sat%), partial pressure of oxygen (pO 2 ), partial pressure of carbon dioxide (pCO 2 ), hematocrit (Hct), red blood cell count (RBC), and cardiac output (CO).The MTX device was previously reported as a valid device for hematological analyses (Segman and Sheiman 2018).

Heart rate variability
A telemetric chest HR sensor and strap (Polar H7, Polar Electro, Kempele, Finland) was used to record resting HR, which was later analysed for HRV, an indicator of cardiac autonomic modulations that have been associated with cardiovascular health (Task Force 1996).The participants maintained a comfortable sitting position for 10-min with the last 5-min period used for HRV data analysis (Young and Leicht 2011).All raw HR data was exported and processed via Kubios HRV analysis software (Premium version 3.3.1,Kubios, Kuopio, Finland).A prior setting of moderate level for artefact correction, window width of 300 s, window overlap of 50%, and smoothing at 500 Lambda were used to process the data (Tarvainen et al. 2014).The HRV indices utilized in this study included mean HR, standard deviation of normal intervals between heart beats or R waves (SDNN), root-mean-square of successive differences between heart beats or R waves (RMSSD), low-frequency power spectrum (LF, 0.04-0.15Hz), high-frequency power spectrum (HF, 0.15-0.40Hz), standard deviation of the points perpendicular to the line of identity (SD1), and standard deviation along the line of identity (SD2) (Task Force 1996).All HRV data was transformed using a natural logarithm (Ln) for statistical analysis.

Statistical analyses
Descriptive data were presented as mean and standard deviation or median (interquartile range, IQR).An SPSS statistical package for Windows was used for all comparisons (version 25.0, IBM, New York, USA).The Shapiro-Wilk test assessed the normal distribution of each variable.One-way repeated measures of analysis of variance (ANOVA) or Friedman's test was used for statistical comparison of data across multiple time points, where appropriate.A post-hoc analysis with Bonferroni correction was used to identify the differences between mean values following identification of a main effect.The magnitude of difference or effect size (ES) was examined via partial eta squared (η 2 ) with thresholds of 0.01, 0.06, and 0.14 corresponding to small, medium, and large ES, respectively (Cohen 1988).Furthermore, and where appropriate, a paired sample t-test or Wilcoxon signed-rank test was used to compare sleep and PA variables and the change of variables between the DTW and NTW phases.The percentage change of variables in each phase was calculated as (post-measure -premeasure)/pre-measure × 100%.An alpha value of p ≤ 0.05 was set as the level of significance for all analyses.

Physical characteristics and demographic information
The physical characteristics and demographic information of participants are presented in Table 1.All participants were young and educated, with approximately one-quarter of participants being regular smokers (28.6%) and one-fifth regular consumers of alcohol (19.0%).

Objective sleep and working-day physical activity characteristics
For objective sleep measure, the average total sleep time and sleep efficiency were 278.5 ± 79. 6 min and 72.9 ± 10%, respectively, in the NTW phases, which were significantly lower than that in the DTW phases.There were no between-phase differences in sleep latency and wake-after-sleep onset.A comparison of the PA characteristics between the two phases revealed that a lower proportion of moderate-vigorous PA (1.2 ± 0.8%) and a greater proportion of sedentary behaviour PA (74.8 ± 6.4%) was found in the NTW phases.Levels of light intensity PA and moderate intensity PA were similar between these two phases (Figure 2).Data for objective sleep and working-day PA characteristics are shown in the supplementary material (Table S1).

Hemodynamic measures
The results of hemodynamic measures showed in Table 2.All hemodynamic measures were constant across the phases except for systolic BP (F 3,60 = 4.098, p = 0.01, η 2 = 0.170, large ES) and diastolic BP (F 3,60 = 3.735, p = 0.03, η 2 = 0.157, large ES).The post hoc analysis revealed a significantly greater systolic BP at post-NTW compared to pre-NTW (p = 0.004) and significantly greater diastolic BP at post-NTW compared to post-DTW (p = 0.04) and pre-NTW (p = 0.004).The police officers have significantly elevated systolic BP by 3.3% and diastolic BP by 3.9% after the NTW phases.Furthermore, the NTW phases exhibited a significantly higher percentage change ratio of systolic BP (p < 0.05) and diastolic BP (p < 0.05) compared to the DTW phases.

Discussion
This study aimed to investigate the effects of a-serial shift work on sleep quality, cardiovascular function, and PA levels among metropolitan police officers in Taiwan.
The participants completed the resting HR and hemodynamic variables before and after the DTW phase as well as the NTW phase.Additionally, an actigraphy was applied to evaluate PA and sleep patterns in the last 3 working days.The findings of this study were that the NTW resulted in significantly poorer sleep quality (total sleep time and sleep efficiency), greater sedentary behaviour, reduced moderate-vigorous intensity PA moderate-vigorous intensity PA, greater systolic BP and diastolic BP (large ES), and reductions in HRV (large ES).A significant finding of the current study was that total sleep time and sleep efficiency were greater during the DTW compared to the NTW phase.The values of total sleep time exhibited during both phases were less than 6 hours and below the recommendations of 7-9 hours of sleep duration for healthy adults (Hirshkowitz et al. 2015;Ohayon et al. 2017).This finding is consistent with previous reports showing poor sleep characteristics in Taiwanese (Chang et al. 2015;Ramey et al. 2012) and British (Taylor et al. 2019) police officers.Collectively, insufficient sleep duration highlights a serious concern about the sleep health of law enforcement officers.Additionally, the sleep efficiency experienced during the NTW phase was lower than that of the DTW phase, indicating inappropriate sleep quality (Ohayon et al. 2017).Others have reported low sleep efficiency values for firefighters (Jeong et al. 2019) and Hispanic shift workers (Reid et al. 2018).The prevalence of poor sleep patterns in shift workers, especially during the NTW, may be related to a disruption in melatonin release (Jensen et al. 2016) and autonomic function (Hulsegge et al. 2018).Consequently, poor sleep health, especially during NTW, may increase the risk of future CVD (Ramey et al. 2012) and metabolic disorders (e.g., obesity and diabetes) (Charles et al. 2011) in law enforcement officers.Garde et al. (2020) showed that the increasing number of consecutive night shifts did not effect sleep duration and quality.The finding also provided that the poorest sleep duration and sleep quality were found in the last night shift in a series of night shifts.We revealed lower total sleep time and efficiency after the night-time shift work in the last 3 working days in night-time shift work.Hence, our findings add to the growing body of studies related to the impact of the night-time shift work on sleep duration and quality.Though sleep disturbances were observed for the police officers during the DTW to NTW transition, most hemodynamic measures (i.e., O 2 sat%, pO 2 , pCO 2 , Hct, CO, and RBC) were unchanged by the shift work.Conversely, the NTW significantly increased systolic BP and diastolic BP, with the intraphase change greater than that during the DTW.Similarly, The increased BP in NTW may be related to increased vascular sympathetic activity and/or reduced parasympathetic modulations (Cherubini et al. 2021).Additionally, we noted significant reductions in LnSDNN, LnRMSSD, LnLF, LnHF, LnSD1, and LnSD2 were observed after the NTW phase.This finding was supported by the evidence showing imbalance of sympathetic and/or parasympathetic modulations of HR control among male night shift blue collar workers (Hulsegge et al. 2018).Impaired functioning of the autonomic nervous system after the NTW has been reported inconsistently, most likely due to differences in gender and occupational categories (Hulsegge et al. 2018;Jensen et al. 2022;Su et al. 2008).Interestingly, we observed lower LnRMSSD and LnSD1 at pre-DTW compared to pre-NTW, indicating a delay of parasympathetic recovery after the prior shift (i.e., 5-days of NTW and 2 days of rest).This lower HRV level may reflect a shift work maladaptation of police officers that may impair their immediate and potentially long-term health status.Previously, police officers who were maladapted to NTW experienced significantly higher levels of inflammatory biomarkers compared to their adapted counterparts (Nevels et al. 2021).The combination of altered sleep patterns and cardiac autonomic dysfunction from NTW may produce a physiological state that significantly impairs police officers' health.Future studies examining sleep health and cardiac function of adapted and maladapted police officers who engage in regular shift work will help organisations to manage work schedules for best practices to maintain and/or enhance the health of police officers.

Limitations
The present study was conducted with several limitations.Firstly, the study period was limited to two weeks (one week of DTW and one week of NTW), with the order of phases set due to logistical reasons.Long-term examination of shift work in this population over multiple DTW and NTW phases may extend the current results and confirm the significant impact of shift work on sleep patterns and cardiovascular risk in police officers.Secondly, the number of days for accelerometer recordings was limited to the last 3 working days, which may not completely describe the total PA levels of officers.Monitoring sleep patterns and PA across a regular operational week (i.e., 5 working days and 2 rest days) may elucidate further the effects of the circadian cycle experienced during shift work.Also, we did not assess the relationship between PA and sleep quality.Thirdly, the psychometric profile of participants needed to be examined with exposure to a stressful working environment critical for sleep disturbance in police officers (Neylan et al. 2002).Fourthly, some influential factors such as the general PA level, lifestyle, and dietary patterns (Alruwaili et al. 2023) were not included in our analysis, which limited the valid results of interest.Finally, the proportion of regular smokers in our results was 28.6% which was higher than the prevalence of tobacco smoking among general Taiwanese male adults in 2022 (24.4%) (HPA 2023).Also, all police officers who participated in this study were healthy male adults with no to low risk of cardiovascular diseases and sleep disorders.These phenomena may have limited the generalizability of findings.Further studies are needed to comprehensively assess potential factors on sleep quality.Furthermore, expanding the investigations to older and female field police officers increases the application to the broader profession.

Conclusions
Taiwanese police officers experienced lower sleep duration and sleep quality.Further, night shift work also reduced moderate to vigorous physical activity and greater sedentary behaviours.Increased systolic and diastolic blood pressure and reduced heart rate variability also resulted in night shift work.Therefore, night-time shift among police officers led to adverse physiological changes that increased the risk for metabolic disease.

Figure 1 .
Figure 1.The flowchart of the experimental design of study.Note: DTW = day time shift work; NTW = night time shift work; HRV = heart rate variability.

Figure 2 .
Figure 2. The average daily sleep (a) and physical activity (b) characteristics of Taiwanese police officers during day-time (DTW) and nigh-time (NTW) phases (n = 21).* indicates statistically significant between the phases (p < 0.05).

Table 1 .
Physical characteristics of the participants at prior to daytime working shift (n = 21).

Table 3 .
The heart rate variability measured during day-time and night-time shirtworks in Taiwan police officers (n = 21).