The study sample included 408 police officers (300 men and 108 women) with an average age of 42.8 years (Table 1). The prevalence of obesity (BMI ≥ 30 kg/m²) was 40.0% overall, 48.3% in male officers, and 16.7% in female officers. The prevalence of current smoking among women (25.7%) was double that among men (12.7%), whereas alcohol consumption among men (6.2 drinks/week) was about twice that among women (3.4 drinks/week). The average work duration (including a second job) was 37.6 hours/week; women worked an average of 6 hours per week less than men. Nearly 40% of men had a second job compared with 15% of women (P < 0.001). Male officers were almost evenly distributed by shiftwork status: day shift (35.3%), afternoon shift (38.3%), and night shift (26.3%); the majority female officers worked on day shift (69.4%). As expected, most adiposity measures were significantly higher in men than in women with the exception of percent body fat. Forty percent of male officers worked more than 40 hours per week whereas only 10% of women did. Approximately half of women spent their time caring for children or helping family members after regular work, but only 23% of men did.

Table 2 shows the unadjusted associations for selected covariates with adiposity measures and hours of work. Age was positively correlated with waist circumference (r = 0.181; P < 0.001), abdominal height (r = 0.177; P < 0.001), and BMI (r = 0.100; P = 0.044), but was not significantly correlated with work hours. Results for years of service and energy intake were similar to those for age. Waist circumference, abdominal height, and BMI were strongly correlated with each other (r > 0.8; P < 0.001), and each of these parameters was less strongly correlated with percent body fat (r values ranged from 0.214 to 0.357; P < 0.001) (data not shown). Smoking status was strongly associated with waist circumference (P < 0.001), abdominal height (P < 0.001), and BMI (P < 0.001); never and former smokers tended to have larger adiposity levels. The officers who worked afternoon and night shifts had larger waist circumference (P = 0.020), whereas day-shift officers had higher percent body fat. Afternoon shift officers had the longest work hours (40.6 hours/week; P < 0.001). Activities after work were significantly associated with most adiposity measures and work hours; the officers who took care of children or family members had smaller waist circumference and BMI, larger percent body fat levels, and fewer work hours.

Table 3 shows the association between work hours and adiposity measures stratified by sex. As mentioned earlier, only 10% of women officers worked 40 or more hours per week. We categorized women’s work hours differently from men’s (fewer than 30, 30 to 34.9, and 35 or more hours/week) to allow for a more equal distribution in the three categories. In the unadjusted models, work hours were inversely associated with waist circumference among women (mean waist circumference measurements were 84.1 ± 12.7 cm for fewer than 30 hours/week, 80.8 ± 13.2 cm for 30 to 34 hours/week, and 76.2 ± 7.9 cm for 35 or more hours/week; P = 0.020). Nevertheless, no significant association was observed between work hours and adiposity measures after adjustment for all covariates.

The associations between work hours and adiposity measures stratified by shiftwork are shown in Table 4 for male officers; the number of female officers in these stratified analyses was limited. Among male officers who worked the midnight shift, mean values of waist circumference significantly increased with longer hours of work (90.4 ± 4.7 cm for fewer than 30 hours/week, 96.8 ± 3.5 cm for 30 to 39 hours/week, and 102.1 ± 3.9 cm for 40 or more hours/week; P = 0.006) after adjusting for age, race, physical activity, food intake, sleep duration, smoking status, activities after work, police rank, and household income; whereas the associations were not statistically significant for day- and afternoon-shift officers. Similar results were observed for BMI. Among male officers who worked the midnight shift, mean values of BMI significantly increased with longer work hours (28.1 ± 1.3 kg/m² for fewer than 30 hours/week, 30.5 ± 1.1 kg/m² for 30 to 39 hours/week, and 32.4 ± 1.3 kg/m² for 40 or more hours/week; P = 0.001) after risk factor adjustment. Figure 1 shows the relationship for work hours with waist circumference and BMI stratified by shiftwork. Abdominal height and percent body fat were positively associated with work hours among male officers who worked the midnight shift after adjustment for potential confounding from all covariates, but the results did not reach statistical significance (Table 4). There was no association between work hours and adiposity for male officers who worked the day and afternoon shifts. Among male officers, shiftwork significantly modified the associations between work hours and waist circumference (P value for interaction = 0.017) and work hours and BMI (P value for interaction = 0.006).

The results of this study indicate that among male police officers who worked the midnight shift, working longer hours was significantly associated with larger waist circumferences and higher BMI. These associations were observed after adjustment for demographic and lifestyle factors such as age, household income, physical activity, energy intake, sleep duration, smoking status, activities after work, and household income. These findings are somewhat consistent with previous studies. Several investigators^17,28,29 have reported that male workers with long work hours had higher BMI. In this study, we did not observe a statistically significant association between shiftwork and adiposity. This is in contrast with several studies^6,13,14 that have found significantly higher levels of adiposity among night-shift workers compared with those working other shifts.

One plausible explanation for the positive association between long work hours and obesity is unhealthy behaviors such as regularly consuming an unhealthy diet. Police officers, especially those working on midnight shifts, may only have available less-than-optimal eating environments such as convenience stores or fast-food shops where high-fat, high-carbohydrate, and high-caloric foods are served.³⁰ During the past 30 years, the number of fast-food restaurants has increased, and many are open 24 hours a day, 7 days a week. Therefore, police patrol officers on night shifts have easy access to fast foods and fewer alternatives at night. Nocturnal eating may elevate serum total cholesterol levels.³¹ In some studies,^15,18,32 low physical activity was associated with overweight and obesity in individuals who worked long hours; however, we adjusted for physical activity. Having a less-than-optimal dietary intake could have contributed to night-shift officers being overweight or obese.

Another possible explanation is dysregulation of the circadian rhythm, which would not be uncommon in police officers who are exposed to a prolonged period of sleep deprivation. Changes in the natural rhythms of the body may reduce the metabolic rate. Changes in the sleep–wake pattern and circadian rhythm disruption could be responsible for the increased adiposity among night workers. Magee et al²⁸ reported that short sleep mediated the association between long work hours and BMI in male workers. In several recent studies, abnormal sleep patterns and abnormal circadian rhythms were significantly associated with changes in metabolism.^29,33–35

In other studies, night-shift workers’ sleep times were not synchronized to melatonin rhythms, and they had more disruptive sleep and hormonal changes that lead to increased risk of diabetes, obesity, and heart disease.^36,37 This study found that the mean of hours of sleep per day for night-shift officers was significantly lower than for that on the day or afternoon shift. Sleep deprivation could change endocrine hormones that control the hypothalamus and lead to weight gain and obesity. One study³⁸ revealed that men who restricted night sleep had lower leptin levels, a hormone that suppresses appetite, and had higher ghrelin levels, a hormone that promotes eating. A study by Schmid et al³⁹ showed similar results as reported in the previous study,³⁸ indicating that those with short sleep had higher ghrelin levels. A study conducted on the same cohort of police officers by Charles et al⁴⁰ observed a U-shaped pattern of leptin levels across sleep duration categories with officers who slept 5.0 to 5.9 hours having the lowest leptin levels. A Canadian study⁴¹ revealed that there was a U-shaped relation between sleep hours and obesity. It is possible that the officers who chronically work on midnight shifts may have a collection of problems: shorter sleep, abnormal leptin and ghrelin levels, increased hunger, less energy expenditure, and lower metabolic rate, which may contribute to overweight and obesity.

Other studies reported that workers exposed to long hours were more stressed,^42–44 and people tend to eat as a way to handle stress.^45–48 In a review article Torres and Nowson⁴⁸ concluded that lifetime stress was probably associated with a food consumption behavior oriented toward high calories. Wardle et al⁴⁹ reported that workers with high stress and long work hours consumed more energy-dense foods than workers with low stress. Police duty is one of the most stressful of all occupations. Common stressors among police are shiftwork, working overtime, overabundance of paperwork, effects on human relationships involving coworkers and family, danger associated with enforcing the law, and exposure to people in distress and pain.^50–52 Emotional and physical stressors are associated with elevated cortisol levels. Police officers with high levels of post-traumatic stress disorder had higher mean awakening cortisol levels,⁵³ and cortisol deregulation was found in male officers with metabolic syndrome.⁵⁴ Previous studies^55–57 have shown that chronic stress was positively associated with abdominal fat. Thus it is possible that officers who work long hours and also work the night shift may have more psychological stress, which is associated with greater obesity.

About 40% of male officers on midnight shift had a second job. The duration of the second job directly contributes to longer total work hours, which could have adverse consequences on their working environment and health. A study⁵⁸ revealed that moonlighting among police officers was associated with higher levels of stress.

Female officers had significantly fewer work hours than male officers. Most female officers (almost 70%) worked on the day shift, and 90% of female officers on day shift worked 40 or fewer hours per week. Previous studies^17,59,60 have shown a U-shaped relationship between work hours and obesity among women where women with full-time employment and longer work hours had increased BMI or weight gain, but women with less than full-time work and fewer work hours also had increased BMI or weight gain. In the present study, no significant associations were observed among women. Schulte et al⁶¹ revealed that long work hours were associated with obesity in men but not in women.

This study has several strengths and limitations. Because it is cross-sectional, we cannot infer causality. In addition, the estimate of total work hours used as our primary predictor may have been inaccurate because it included work hours from a second job, which was self-reported. The small number of female officers in afternoon and night shift may have limited the power to detect the shift-stratified association. The data of this study included some detectives and executive officers (15.1%) who normally work during the day and their experiences might not reflect the experience of solely patrol police officers.

Officers may have been exposed to job stress through lack of support from supervisors, excessive paperwork, interpersonal confrontations, and conflict involving family or friends.^62,63 Future investigations should include measurement of work-related factors and family concerns. Additional research is needed to investigate nutrient factors that could be involved in the elevated adiposity levels of officers, such as meal time and length, frequency of fast-food intake, use of vending machines, and snack or beverage preference. More than 90% of male officers were overweight (BMI ≥ 25 kg/m²) or obese (BMI ≥ 30 kg/m²), and further study should focus on identifying successful weight-control programs in this population.

One of the strengths of this study is that both the actual hours worked and the shift worked were obtained on a daily basis over a minimum of 10 years. It is likely that this provided more objective and accurate measures than would be the case for self-reported measures that are subject to recall. A second strength of the study is that we used four adiposity measurements as outcomes: waist circumference, BMI, abdominal height, and percent body fat. These data confirmed the results of a previous study⁶⁴ that reported that BMI, waist circumference, and abdominal height are more highly correlated with each other than with percent body fat (Table 2). Percent body fat provides a more accurate assessment of fat content than the other three measurements. Nevertheless, it has been infrequently used in research studies because of the high cost and time involved in measurement and the need for highly trained personnel. A recent study⁶⁵ used percent body fat as a predictor of obesity-related cardiovascular disease risk factors in Korean adults. Using data from the third National Health and Nutrition Examination Survey, Menke and colleagues⁶⁶ reported that waist circumference showed the strongest association with cardiovascular disease risk factors among the five adiposity measurements (BMI, total body fat, percent body fat, skin-fold thickness, waist circumference). All adiposity measurements in this study were obtained by trained staff and, therefore, less measurement error would be expected. If information bias in adiposity measures does exist in this study, it would be nondifferential. Most research conducted on police officers has focused on job stress and health outcomes such as depression, anxiety, hypertension, and cardiovascular disease. To our knowledge, this is the first study to focus on the association of work hours with multiple measures of adiposity in police.

Longer work hours were associated with larger waist circumference and higher BMI among male officers working the midnight shift, but not in male officers who worked the day or afternoon shift. Among female officers, there was no association between long work hours and adiposity for any shift. The results may implicate factors linked with night-shift work such as poor dietary intake, disrupted circadian rhythm, and increased work stress as potential causes of the undesirable levels of adiposity, which in turn might increase the risk of chronic disorders such as hypertension, diabetes, metabolic syndrome, stroke, and coronary heart disease. The positive association of work hours and adiposity in male officers should be confirmed in longitudinal studies, studies with larger sample sizes, and investigations of other populations.