This study investigated the effects of shift work on abdominal obesity among young female nurses (20–29 years old vs. 30–39 years old) during a 5-year period. The results indicate that 20–29-year-old shift workers had a significant risk of abdominal obesity. Although various studies have reported an association between shift work and obesity, the underlying mechanisms remain unclear . For example, shift work might cause a disruption in circadian rhythms, which could result in lifestyle changes (e.g., eating and exercise habits) and weight gain [20, 21]. In addition, sleep disorders can result in appetite-regulating hormone imbalance (e.g., in leptin and ghrelin) [20, 22]. Furthermore, shift workers might have limited opportunities for physical activity, based on their work schedules and sleep deprivation . Moreover, age may act as an effect modifier in the relationship between shift work and obesity [9, 10], and we hypothesized that age might affect the strength and direction of the association between shift work and abdominal obesity. Unfortunately, there are few studies of people who are <30 years old, and we speculated that this might be related to the limited number of overweight or obese people in that age group, which would limit the ability to statistically evaluate the magnitude of changes in weight or WC.
Young female adults have a higher basal metabolic rate and are more concerned with esthetics and beauty, compared to older women. Therefore, the effects of shift work on abdominal obesity might be confounded by these characteristics, and repeated measures may help estimate the relevant risk. In the present study, we used repeated measures of WC to evaluate the effects of rotating shifts on WC among young female nurses. In the 20–29-year-old group, WC increased during the early study period, but returned to baseline by the fifth follow-up for both shift and day workers. Unfortunately, there is limited research regarding this association among young people, which limits our ability to compare these findings to previous findings. Nevertheless, we propose two explanations for this trend. First, the younger age group may experience a more powerful trend of regression to the mean, compared to the 30–39-year-old nurses. This theory is supported by the strong rebounds that we observed after significant increases in WC, which would imply that the changes were not solely related to the greater basal metabolic rate of the 20–29-year-old group. Thus, the younger age group may perform conscious interventions to control their increasing WC, and this hypothesis is supported by a study of involuntary weight change, which revealed greater involuntary weight gain among the groups that were in their 30s and 40s . A second possible explanation is that the 20–29-year-old age group has a lower prevalence of abdominal obesity, which makes it difficult to set a cut-off value for this group. Thus, a larger sample size may be needed to detect significant differences in subtle WC changes. Furthermore, previous studies have consistently revealed an association between WC change and shift work among 30–39-year-old individuals [12, 24, 25].
Among the 20–29-year-old nurses, the shift work-related OR for abdominal obesity was 3.21 (95 % CI: 1.29–7.98) in the multivariate analysis, which appears to conflict with the absence of a significant change in WC (from baseline) at all follow-ups in the univariate analysis. However, this discrepancy is likely related to the fact that the univariate and multivariate analyses used different dependent variables and the multivariate analyses were adjusted for various confounders. In the present study, abdominal obesity was defined as a WC of ≥80 cm, which reflects the World Health Organization’s Asia-West Pacific standard in 2000 . In contrast, abdominal obesity has been defined as a WC of ≥85 cm for Korean women, although it can be inappropriate to apply a general standard for the female population to an age-specific population, such as in the present study. For example, according to 2014 KNHANES data, 83.5 % of 20–29-year-old women and 53 % of 30–39-year-old women had WC measurements of <80 cm . Furthermore, the present study’s population exhibited a very high proportion of WCs that were <80 cm (20–29-year-olds: 97 %, 30–39-present study’s population exhibited a very high proportion of WCs that were <80 cm (20–29-year-olds: 97 %, 30–39-year-olds: 94 %), which indicates that these individuals had smaller WCs, compared to the general population. Moreover, only approximately 5 % of the included nurses had abdominal obesity (a WC of ≥80 cm), and the mean changes in WC and ORs for abdominal obesity only reflect changes in a small subset of the population. Therefore, we believe that these factors explain the discrepancy between the results of the univariate and multivariate analyses. In addition, the OR for abdominal obesity in the 30–39-year-old group was not statistically significant, although this may be related to the limited number of shift workers in that group.
During the follow-up period, the ORs for increased abdominal obesity over time were 1.40 (95 % CI: 1.23–1.58) among 20–29-year-old nurses and 1.60 (95 % CI: 1.36–1.87) among 30–39-year-old nurses. As all of these ORs are statistically significant, it suggests that the risk of abdominal obesity clearly and steadily increases over time. Moreover, these results are consistent with a previous study’s findings that the prevalence of abdominal obesity increases steadily over time .
Only a few participants had unhealthy alcohol consumption habits, and alcohol consumption did not significantly influence abdominal obesity. However, analysis of alcohol consumption may not identify significant effects on WC, as this factor is influenced to a greater extent by food intake, rather than alcohol consumption. Thus, it may be more appropriate to consider the total caloric intake in future studies.
The risk of abdominal obesity among the insufficient exercise group was lower than that among the basic and sufficient exercise groups. We assume that this paradox may be explained by the insufficient exercise group performing less exercise because they were sufficiently slender and did not have to lose weight. Furthermore, 20–39-year-old adults have a higher basal metabolic rate, compared to older individuals, and their weight and WC do not tend to change involuntarily . Moreover, total caloric intake may have a greater effect on abdominal obesity, compared to intense exercise. For example, one study found no statistically significant differences in WC and BMI changes among overweight and obese youth, despite high-intensity interval training .
The present study has several limitations. First, the study population does not reflect the general population, because it only included female nurses from a single hospital in a specific region. Second, our data were collected from health screening results and did not include basic information regarding the individual’s obstetric history, socio-economic status, or details regarding their diet and eating habits. We also could not obtain accurate data regarding the nurses’ reasons for retirement and physical workload. However, we assumed that our participants (young female nurses working in the same university hospital) were likely to have similar socioeconomic statuses, and we restricted the study population to only female nurses in order to simplify the analyses and adjustments. Third, our study may include selection bias, as the acute adverse effects of shift work (e.g., sleep disorders and gastrointestinal problems) may contribute to weight loss . Moreover, weight loss-prone individuals may have dropped out of the study before the one-year follow-up because they tend to quit their job more easily. Loss to follow-up may also have biased our findings, although the overall rate of loss to follow-up in the study population was only 24.8 %, and the specific rates for dependent and independent variables were generally <30 %.
The present study also has several strengths. First, we performed a retrospective longitudinal analysis based on data from a 5-year period. Second, we adjusted for unidentified intra- and inter-individual changes using the GEE for statistical and time series analyses. Furthermore, it was possible to specifically identify inter-participant variability, rather than error, based on the use of repeated measures . Third, the relatively large sample size allowed us to analyze the effects of shift work, despite the relatively low prevalence of abdominal obesity among 20–29-year-old women. These factors may explain the difference between our findings and those of previous cross-sectional studies. Third, we were able to evaluate the influence of age by stratifying the participants into 20–29-year-old and 30–39-year-old groups. Thus, we could identify the associations between shift work and increased WC, and also between shift work and risk of abdominal obesity.