obesity reviews
doi: 10.1111/obr.12194
Obesity Comorbidity/Etiology and Pathophysiology
Meta-analysis on night shift work and risk of metabolic syndrome F. Wang1,2, L. Zhang3, Y. Zhang3, B. Zhang1, Y. He1, S. Xie1, M. Li1, X. Miao4, E. Y. Y. Chan1, J. L. Tang1,2,5, M. C. S. Wong1, Z. Li3, I. T. S. Yu1 and L. A. Tse1,2,5
1
JC School of Public Health and Primary
Care, The Chinese University of Hong Kong, Hong Kong SAR, China; 2CUHK Centre for Public Health and Primary Care (Shenzhen), Shenzhen Research Institute of the Chinese University of Hong Kong, Shenzhen, China; 3
Shenzhen Prevention and Treatment Center
for Occupational Diseases, Shenzhen, China; 4
Department of Epidemiology and
Biostatistics, Tongji School of Public Health, Huazhong University of Science and Technology, Wuhan, China; 5Shenzhen Municipal Key Laboratory for Health Risk Analysis, Shenzhen Research Institute of the Chinese University of Hong Kong, Shenzhen, China
Received 12 March 2014; revised 30 April 2014; accepted 30 April 2014
Address for correspondence: LA Tse, JC
Summary This study aims to quantitatively summarize the association between night shift work and the risk of metabolic syndrome (MetS), with special reference to the dose–response relationship with years of night shift work. We systematically searched all observational studies published in English on PubMed and Embase from 1971 to 2013. We extracted effect measures (relative risk, RR; or odd ratio, OR) with 95% confidence interval (CI) from individual studies to generate pooled results using meta-analysis approach. Pooled RR was calculated using random- or fixed-effect model. Downs and Black scale was applied to assess the methodological quality of included studies. A total of 13 studies were included. The pooled RR for the association between ‘ever exposed to night shift work’ and MetS risk was 1.57 (95% CI = 1.24–1.98, pheterogeneity = 0.001), while a higher risk was indicated in workers with longer exposure to night shifts (RR = 1.77, 95% CI = 1.32–2.36, pheterogeneity = 0.936). Further stratification analysis demonstrated a higher pooled effect of 1.84 (95% CI = 1.45–2.34) for studies using the NCEP-ATPIII criteria, among female workers (RR = 1.61, 95% CI = 1.10–2.34) and the countries other than Asia (RR = 1.65, 95% CI = 1.39–1.95). Sensitivity analysis confirmed the robustness of the results. No evidence of publication bias was detected. The present meta-analysis suggested that night shift work is significantly associated with the risk of MetS, and a positive dose–response relationship with duration of exposure was indicated.
School of Public Health and Primary Care, The Chinese University of Hong Kong. 4/F
Keywords: Metabolic syndrome, night work, shift work.
School of Public Health and Primary Care, Prince of Wales Hospital, Sha Tin, N.T., Hong Kong SAR, China.
Abbreviations: CI, confidence interval; IDF, International Diabetes Federation; MetS, metabolic syndrome; OR, odd ratio; RR, relative risk.
E-mail: [email protected]
obesity reviews (2014)
Introduction Shift work that involves night work is a popular job nature in around 20% of the work force in industrialized areas, with high prevalence in service industries, particularly for the sectors of health care, retail and transportation (1). Decrease of night sleep duration and desynchronization of circadian rhythm are the two major aspects of night shift work (2) that were found to disturb metabolism (3–5) and
© 2014 The Authors obesity reviews © 2014 World Obesity
impair components of metabolic syndrome (MetS) (6,7), which may consequently result in the occurrence of MetS among night shift workers. Sufficient evidence has shown that the presence of MetS was linked to an increased risk of type 2 diabetes, cardiovascular diseases and all-cause mortality (8). Growing attentions have been raised by the public about the impact of night shift work on the risk of MetS, while the findings tended to be inconsistent or even 1
2
obesity reviews
Night shift work and metabolic syndrome F. Wang et al.
conflicted. Canuto et al. published the first systematic review (without meta-analysis) on night shift work and MetS risk by including 10 pertinent studies (9); eight of them showed a positive association with a relative risk (RR) ranging from 1.51 to 5.10. Canuto et al. interpreted that this excess risk of MetS among night shift workers was attributable to the short sleep duration; this conclusion, however, tends to be lacking of support given only one study provided sleep data, while the other eight studies exclusively reported rotating shifts covering the period that may damage circadian rhythm – another major aspect of night shift work. This study aims to summarize all published epidemiological studies in terms of the association between the risk of MetS and the night shift work that potentially involved circadian disruption using approaches of systematic review and meta-analysis, with special reference to the dose–response relationship with years of night shift work.
Methods Search strategy and eligible studies We systematically searched all observational epidemiological studies about night shift work and MetS risk in PubMed and Embase. All pertinent studies published from January 1971 to July 2013 were identified according to the search strategies as described in Supporting Information Supplement S1. Only English articles were included.
Inclusion criteria and data extraction The studies included in the meta-analysis had met all the following inclusion criteria: (i) evaluated the association between the risk of MetS and night shift work (night work or rotating work covering the period 24:00–05:00); (ii) used cross-sectional, case-control or cohort study design and (iii) provided sufficient data for calculation of odds ratio (OR) or RR with 95% confidence interval (CI). The following information was extracted from each study: (i) name of the first author; (ii) year of publication; (iii) country of origin; (iv) number of MetS cases; (v) gender and age of the study population; (vi) study design; (vii) night shift work category; (viii) MetS criteria and (ix) covariates used in the adjustment.
while only one item regarding Reporting had a score of 0 to 2 score and another item on Power scored 0 to 5. The maximum possible score of each article is 32 based on this checklist. All studies were reviewed by two reviewers (FW and LT) independently. There was no obvious difference between the judgments of these two reviewers concerning the quality of individual studies.
Meta-analysis The standard error of each RR or OR was calculated according to the 95% CI and they were used in the further meta-analysis. We recalculated the OR and the 95% CI for Copertaro et al.’s study according to the number of cases and controls by performing univariate logistic regression model, because no OR and 95% CI were provided in the original report (11). Two-stage meta-analysis was used to estimate the pooled effect of night shift work on the risk of MetS. We initially performed fixed-effect model to estimate the pooled effect of night shift work by combining all exposure categories if the individual studies reported night shift work of more than two levels of exposure. We then performed random effect model if the I2 index for heterogeneity test was larger than 50%; otherwise, fixed-effect model was performed. We evaluated potential publication bias according to the Begg’s regression asymmetry test (12) and examined the potential source of heterogeneity using sensitivity analysis. In addition, we performed subgroup analysis by the study design, origin countries, gender, the definition of MetS and intensity of night shifts. All statistical analyses were conducted by using Stata 11.0 (StataCorp, College Station, TX, USA).
Results Description of studies We identified 13 eligible articles from the database (Fig. 1), including three cohort studies (13–15), one nested
Quality assessment The quality of included articles was assessed according to the checklist developed by Downs and Black (10). There were totally 27 items in the checklist distributed among five subscales: Reporting (10 items), External validity (three items), Bias (seven items), Confounding (six items) and Power (one item). Most of the questions scored 0 or 1,
Figure 1 Selection strategy of articles in the meta-analysis.
© 2014 The Authors obesity reviews © 2014 World Obesity
obesity reviews
case-control study (16) and nine cross-sectional studies (11,17–24). The main characteristics are summarized in Table 1. One out of the 13 included studies focused on healthcare provider (14) and another concerned about the police officers (21); the occupations for the remaining 11 studies were mixed. The participants of five studies were male workers (13,19,22–24) and two studies were female workers (15,20), while the others included both male and female workers. Five studies were conducted in Asian countries (15,16,19,20,24), involving a total of 15,594 workers (including 953 MetS cases).
Summary of the definition of night shift work and metabolic syndrome Seven studies defined night shift work according to the frequency and time range of night work (11,14,15,18,19,22,23). All the included studies had the time ranges of night work covering the period from 0:00 to 06:00, with an exception of Esquirol et al.’s study (22) in which an earlier end at 05:00 was reported. Six of these seven studies using a similar definition of night shift work to the International Labor Organization (25) showed a positive association with a RR from 1.11 to 5.10; the negative association (RR = 0.89, P > 0.05) (it was recalculated after combining the ORs from the subgroups of workers with two-shift and three-shift work) was obtained from a cross-sectional study conducted by Kawada et al. in Japan, but this study did not adjust for any confounding factors (19). The remaining six studies used either frequency of night work (13,24), time range of night work (20,21) or binary category of shift work (ever vs. never) (16,17) to define the night shift work. All of these six studies showed a positive association between night shift work and the risk of MetS. All 13 included studies defined the presence of MetS according to these five components (any three positive), including high blood pressure, hyperglycaemia, hypertriglyceridemia, low high-density lipoprotein and obesity, with the exception of an early study of Nagaya et al. in which only four components were involved (not including obesity). Specifically, eight of them adopted The National Cholesterol Education Program Expert Panel on Education-Adult Treatment Panel III (NCEP-ATPIII) criteria to define the presence of MetS; three studies used the International Diabetes Federation (IDF) criteria (one with slight modified); one study employed the Japan Society for the Study of Obesity criteria, while another took the number of insulin-resistant components to define the MetS. The prevalence or incidence of MetS ranged from 6.2% to 20.3% for studies using the NCEP-ATPIII criteria, while it ranged from 3.9% to 28.5% for those adopting the IDF criteria. © 2014 The Authors obesity reviews © 2014 World Obesity
Night shift work and metabolic syndrome F. Wang et al. 3
Assessing quality of studies Twenty-seven questions included in the Downs and Black checklist were available for quality assessment; however, eight items were not suitable for our review as these eight items were designed for assessing the quality of randomized controlled trial, while our review involves observational studies. Therefore, only 19 items were applied to our eligible studies and 24 is the maximum score representing the best quality for an individual study (Supporting Information Supplement S2). The total scores ranged from 14 to 18 for studies included in this review with a median score of 16; only four articles have scores below the median. The lower scored articles were crosssectional studies, for which the adjustment of major confounders was considered not adequate. Articles with higher scores tended to take more consideration on the adjustment of confounding factors. In general, the cohort or nested case-control studies had a better quality (Supporting Information Supplement S3).
Meta-analysis results A two-stage meta-analysis was performed to estimate the association between night shift work and the MetS risk by combining results from all eligible studies, giving an adjusted RR of 1.57 (95% CI = 1.24–1.98, pheterogeneity = 0.001) (Fig. 2); however, a higher combined RR was indicated for cohort and nested case-control studies (RR = 2.03, 95% CI = 1.31–3.15, pheterogeneity = 0.07) than that of the cross-sectional studies (RR = 1.39, 95% CI = 1.08–1.80, pheterogeneity = 0.015). Table 2 showed an indication of a positive dose–response relationship between years of night shift and the MetS risk. Results from further meta-analysis (Fig. 3) revealed that a higher combined RR was associated with longer years of night shift work compared with the shorter one (RR = 1.77 95% CI = 1.32–2.36 vs. RR = 1.16 95% CI = 0.62–2.15). We performed subgroup analyses by the MetS criteria, gender and countries to address the sources of heterogeneity (Fig. 4). The pooled effects were 1.84 (95% CI = 1.45– 2.34, pheterogeneity = 0.37) for studies using the NCEP-ATPIII criteria and the effect was slightly higher than those using the IDF criteria (RR = 1.57, 95% CI = 1.29–1.92, pheterogeneity = 0.82). A slightly higher association was observed for studies of female workers (RR = 1.61, 95% CI = 1.10–2.34 vs. RR = 1.36, 95% CI = 1.03–1.81 for the males) and in countries other than Asia (RR = 1.65, 95% CI = 1.39–1.95 vs. RR = 1.35, 95% CI = 0.92–1.92 for Asian countries). Only six of all eligible studies reported specific effects from individual components of the MetS. Results from meta-analyses based on these available data showed that night shift work was significantly associated with an
18:00–06:00 night shift and 10 nights/month
Three rotating shift, and 21:00–05:00 for night shift
20:00–04:00 for night work
19:00–07:00 for night work
34.4 ± 8.4 years, male
35–69 years, male and female
39–60 years, male
39.5 ± 7.6 years, male and female
33.6 ± 7.1 years, female
34–64 years, male
Sookoian et al. Argentina 2007 (23)
Copertaro et al. Italy 2008 (11)
Esquirol et al. France 2009 (22)
Violanti et al. USA 2009 (21)
Chen et al. Taiwan 2010 (20)
Kawada et al. Japan 2010 (19)
Two or three rotating shift 15:00–06:30 for two shift 22:30–06:30 for three shift
22:00–6:00 night shift and ≥1.5 nights/month
Night work ≥1 night/week
30–59 years, male
Cross-sectional study Nagaya et al. Japan 2002 (24)
Definition of night shift
Age, gender
Studies, location
JASSO
NCEP-ATPIII
NCEP-ATPIII
NCEP-ATPIII
NCEP-ATPIII
NCEP-ATPIII
Self-defined, ≥3 insulin resistance markers
MS criteria
None
Age, smoking, drinking, education and duration of work
Gender, age, smoking, education, alcohol intake, education, marital status, police rank and physical activity
Age, work organization, total physical activity, job strain index, smoking status, alcohol consumption, carbohydrate intake, total energy intake and contributions to daily energy intake of meals from breakfast, morning, afternoon and night light meals
None
Age, physical activity
Body mass index, job, drinking, smoking and exercise
Adjusted confounders
Table 1 Main characteristics of the metabolic syndrome studies and night shift work included in the meta-analysis
386
N.A.
16
37
16
274
108
MS cases
Two-shift: 0.77 (0.61–0.98) Three-shift: 1.39 (0.92–2.09)
1.80 (0.70–4.30)
1.57 (0.41–5.95)
2.33 (1.04–5.23)
1.11 (0.39–3.14)*
1.51 (1.01–2.25)
For 30–39 years old: 2.77 (0.77–10.00); For 40–49 years old: 1.42 (0.67–2.87); For 50–59 years old: 0.75 (0.32–1.57)
RR or OR (95% CI)
14
14
17
16
14
15
17
Score
4 Night shift work and metabolic syndrome F. Wang et al.
obesity reviews
© 2014 The Authors obesity reviews © 2014 World Obesity
© 2014 The Authors obesity reviews © 2014 World Obesity
Healthy worker 24–65 years, male and female
32–62 years, male and female
Puttonen et al. Finland 2012 (18)
Tucker et al. France 2012 (17)
21:00–07:00 night shift and >4 nights/month
38.4 ± 8.4 years†, male and female
25–59 years, male and female
Pietroiusti et al. Italy 2010 (14)
Li et al. Japan 2011 (16)
NCEP-ATPIII
NCEP-ATPIII
NCEP-ATPIII
Modified IDF
IDF
IDF
MS criteria
Age, sex, job type, drinking habits, smoking habits, leisure-time physical activity, work intensity, sleep duration, vegetable intake and snack food habits
Age, gender, smoking, alcohol intake, familiar history, physical activity and work schedule
Age, baseline insulin resistance status, metabolic syndrome components, job and lifestyle factors.
Age, smoking, physical activity outside work, educational level, job strain, physical job demands, waist circumference, diastolic blood pressure, high-density lipoprotein cholesterol
Age, sex, socioeconomic status, smoking, alcohol, stress and sleep difficulty.
Age, education, physical activity, alcohol use, smoking and insomnia symptoms
Adjusted confounders
417
42
42
364
68
516
MS cases
1.87 (1.13–3.08)
5.10 (2.15–12.11)
Night shift: 3.70 (1.50–9.20) Former shift: 0.70 (0.20–2.40)
1.46 (1.04–2.07)
1.78 (1.03–3.08)
Current night shift: 1.37 (0.84–2.22) [male] 1.31 (0.54–3.17) [female] Former shift: 2.00 (1.26–3.19) [male] 1.51 (0.83–2.77) [female]
RR or OR (95% CI)
16
18
17
18
17
16
Score
*Recalculated with the original data. † Recalculated with means, standard deviation and sample size of day time and night shift workers. F, female; IDF, International Diabetes Federation; JASSO, The Japan Society for the Study of Obesity; M, male; NCEP-ATPIII, The National Cholesterol Education Program Expert Panel on Education-Adult Treatment Panel III; N.A., not available.
Ever shift work
Two rotating shift, 19:30–07:30 for night shift
32.8 ± 7.9 years, female
Two or three rotating shifts
Ever shift work
Three rotating shift work including night work >3 h, between 23:00 and 06:00 for night shift
Definition of night shift
Lin et al. Taiwan 2009 (15)
Cohort/nested case-control study De Bacquer et al. 35–59 years, male Belgium 2009 (13)
Age, gender
Studies, location
Table 1 Continued
obesity reviews Night shift work and metabolic syndrome F. Wang et al. 5
6
obesity reviews
Night shift work and metabolic syndrome F. Wang et al.
Studies (year)
Years of night shift work
De Bacquer et al. 2009 (13)
doi: 10.1111/obr.12194
Obesity Comorbidity/Etiology and Pathophysiology
Meta-analysis on night shift work and risk of metabolic syndrome F. Wang1,2, L. Zhang3, Y. Zhang3, B. Zhang1, Y. He1, S. Xie1, M. Li1, X. Miao4, E. Y. Y. Chan1, J. L. Tang1,2,5, M. C. S. Wong1, Z. Li3, I. T. S. Yu1 and L. A. Tse1,2,5
1
JC School of Public Health and Primary
Care, The Chinese University of Hong Kong, Hong Kong SAR, China; 2CUHK Centre for Public Health and Primary Care (Shenzhen), Shenzhen Research Institute of the Chinese University of Hong Kong, Shenzhen, China; 3
Shenzhen Prevention and Treatment Center
for Occupational Diseases, Shenzhen, China; 4
Department of Epidemiology and
Biostatistics, Tongji School of Public Health, Huazhong University of Science and Technology, Wuhan, China; 5Shenzhen Municipal Key Laboratory for Health Risk Analysis, Shenzhen Research Institute of the Chinese University of Hong Kong, Shenzhen, China
Received 12 March 2014; revised 30 April 2014; accepted 30 April 2014
Address for correspondence: LA Tse, JC
Summary This study aims to quantitatively summarize the association between night shift work and the risk of metabolic syndrome (MetS), with special reference to the dose–response relationship with years of night shift work. We systematically searched all observational studies published in English on PubMed and Embase from 1971 to 2013. We extracted effect measures (relative risk, RR; or odd ratio, OR) with 95% confidence interval (CI) from individual studies to generate pooled results using meta-analysis approach. Pooled RR was calculated using random- or fixed-effect model. Downs and Black scale was applied to assess the methodological quality of included studies. A total of 13 studies were included. The pooled RR for the association between ‘ever exposed to night shift work’ and MetS risk was 1.57 (95% CI = 1.24–1.98, pheterogeneity = 0.001), while a higher risk was indicated in workers with longer exposure to night shifts (RR = 1.77, 95% CI = 1.32–2.36, pheterogeneity = 0.936). Further stratification analysis demonstrated a higher pooled effect of 1.84 (95% CI = 1.45–2.34) for studies using the NCEP-ATPIII criteria, among female workers (RR = 1.61, 95% CI = 1.10–2.34) and the countries other than Asia (RR = 1.65, 95% CI = 1.39–1.95). Sensitivity analysis confirmed the robustness of the results. No evidence of publication bias was detected. The present meta-analysis suggested that night shift work is significantly associated with the risk of MetS, and a positive dose–response relationship with duration of exposure was indicated.
School of Public Health and Primary Care, The Chinese University of Hong Kong. 4/F
Keywords: Metabolic syndrome, night work, shift work.
School of Public Health and Primary Care, Prince of Wales Hospital, Sha Tin, N.T., Hong Kong SAR, China.
Abbreviations: CI, confidence interval; IDF, International Diabetes Federation; MetS, metabolic syndrome; OR, odd ratio; RR, relative risk.
E-mail: [email protected]
obesity reviews (2014)
Introduction Shift work that involves night work is a popular job nature in around 20% of the work force in industrialized areas, with high prevalence in service industries, particularly for the sectors of health care, retail and transportation (1). Decrease of night sleep duration and desynchronization of circadian rhythm are the two major aspects of night shift work (2) that were found to disturb metabolism (3–5) and
© 2014 The Authors obesity reviews © 2014 World Obesity
impair components of metabolic syndrome (MetS) (6,7), which may consequently result in the occurrence of MetS among night shift workers. Sufficient evidence has shown that the presence of MetS was linked to an increased risk of type 2 diabetes, cardiovascular diseases and all-cause mortality (8). Growing attentions have been raised by the public about the impact of night shift work on the risk of MetS, while the findings tended to be inconsistent or even 1
2
obesity reviews
Night shift work and metabolic syndrome F. Wang et al.
conflicted. Canuto et al. published the first systematic review (without meta-analysis) on night shift work and MetS risk by including 10 pertinent studies (9); eight of them showed a positive association with a relative risk (RR) ranging from 1.51 to 5.10. Canuto et al. interpreted that this excess risk of MetS among night shift workers was attributable to the short sleep duration; this conclusion, however, tends to be lacking of support given only one study provided sleep data, while the other eight studies exclusively reported rotating shifts covering the period that may damage circadian rhythm – another major aspect of night shift work. This study aims to summarize all published epidemiological studies in terms of the association between the risk of MetS and the night shift work that potentially involved circadian disruption using approaches of systematic review and meta-analysis, with special reference to the dose–response relationship with years of night shift work.
Methods Search strategy and eligible studies We systematically searched all observational epidemiological studies about night shift work and MetS risk in PubMed and Embase. All pertinent studies published from January 1971 to July 2013 were identified according to the search strategies as described in Supporting Information Supplement S1. Only English articles were included.
Inclusion criteria and data extraction The studies included in the meta-analysis had met all the following inclusion criteria: (i) evaluated the association between the risk of MetS and night shift work (night work or rotating work covering the period 24:00–05:00); (ii) used cross-sectional, case-control or cohort study design and (iii) provided sufficient data for calculation of odds ratio (OR) or RR with 95% confidence interval (CI). The following information was extracted from each study: (i) name of the first author; (ii) year of publication; (iii) country of origin; (iv) number of MetS cases; (v) gender and age of the study population; (vi) study design; (vii) night shift work category; (viii) MetS criteria and (ix) covariates used in the adjustment.
while only one item regarding Reporting had a score of 0 to 2 score and another item on Power scored 0 to 5. The maximum possible score of each article is 32 based on this checklist. All studies were reviewed by two reviewers (FW and LT) independently. There was no obvious difference between the judgments of these two reviewers concerning the quality of individual studies.
Meta-analysis The standard error of each RR or OR was calculated according to the 95% CI and they were used in the further meta-analysis. We recalculated the OR and the 95% CI for Copertaro et al.’s study according to the number of cases and controls by performing univariate logistic regression model, because no OR and 95% CI were provided in the original report (11). Two-stage meta-analysis was used to estimate the pooled effect of night shift work on the risk of MetS. We initially performed fixed-effect model to estimate the pooled effect of night shift work by combining all exposure categories if the individual studies reported night shift work of more than two levels of exposure. We then performed random effect model if the I2 index for heterogeneity test was larger than 50%; otherwise, fixed-effect model was performed. We evaluated potential publication bias according to the Begg’s regression asymmetry test (12) and examined the potential source of heterogeneity using sensitivity analysis. In addition, we performed subgroup analysis by the study design, origin countries, gender, the definition of MetS and intensity of night shifts. All statistical analyses were conducted by using Stata 11.0 (StataCorp, College Station, TX, USA).
Results Description of studies We identified 13 eligible articles from the database (Fig. 1), including three cohort studies (13–15), one nested
Quality assessment The quality of included articles was assessed according to the checklist developed by Downs and Black (10). There were totally 27 items in the checklist distributed among five subscales: Reporting (10 items), External validity (three items), Bias (seven items), Confounding (six items) and Power (one item). Most of the questions scored 0 or 1,
Figure 1 Selection strategy of articles in the meta-analysis.
© 2014 The Authors obesity reviews © 2014 World Obesity
obesity reviews
case-control study (16) and nine cross-sectional studies (11,17–24). The main characteristics are summarized in Table 1. One out of the 13 included studies focused on healthcare provider (14) and another concerned about the police officers (21); the occupations for the remaining 11 studies were mixed. The participants of five studies were male workers (13,19,22–24) and two studies were female workers (15,20), while the others included both male and female workers. Five studies were conducted in Asian countries (15,16,19,20,24), involving a total of 15,594 workers (including 953 MetS cases).
Summary of the definition of night shift work and metabolic syndrome Seven studies defined night shift work according to the frequency and time range of night work (11,14,15,18,19,22,23). All the included studies had the time ranges of night work covering the period from 0:00 to 06:00, with an exception of Esquirol et al.’s study (22) in which an earlier end at 05:00 was reported. Six of these seven studies using a similar definition of night shift work to the International Labor Organization (25) showed a positive association with a RR from 1.11 to 5.10; the negative association (RR = 0.89, P > 0.05) (it was recalculated after combining the ORs from the subgroups of workers with two-shift and three-shift work) was obtained from a cross-sectional study conducted by Kawada et al. in Japan, but this study did not adjust for any confounding factors (19). The remaining six studies used either frequency of night work (13,24), time range of night work (20,21) or binary category of shift work (ever vs. never) (16,17) to define the night shift work. All of these six studies showed a positive association between night shift work and the risk of MetS. All 13 included studies defined the presence of MetS according to these five components (any three positive), including high blood pressure, hyperglycaemia, hypertriglyceridemia, low high-density lipoprotein and obesity, with the exception of an early study of Nagaya et al. in which only four components were involved (not including obesity). Specifically, eight of them adopted The National Cholesterol Education Program Expert Panel on Education-Adult Treatment Panel III (NCEP-ATPIII) criteria to define the presence of MetS; three studies used the International Diabetes Federation (IDF) criteria (one with slight modified); one study employed the Japan Society for the Study of Obesity criteria, while another took the number of insulin-resistant components to define the MetS. The prevalence or incidence of MetS ranged from 6.2% to 20.3% for studies using the NCEP-ATPIII criteria, while it ranged from 3.9% to 28.5% for those adopting the IDF criteria. © 2014 The Authors obesity reviews © 2014 World Obesity
Night shift work and metabolic syndrome F. Wang et al. 3
Assessing quality of studies Twenty-seven questions included in the Downs and Black checklist were available for quality assessment; however, eight items were not suitable for our review as these eight items were designed for assessing the quality of randomized controlled trial, while our review involves observational studies. Therefore, only 19 items were applied to our eligible studies and 24 is the maximum score representing the best quality for an individual study (Supporting Information Supplement S2). The total scores ranged from 14 to 18 for studies included in this review with a median score of 16; only four articles have scores below the median. The lower scored articles were crosssectional studies, for which the adjustment of major confounders was considered not adequate. Articles with higher scores tended to take more consideration on the adjustment of confounding factors. In general, the cohort or nested case-control studies had a better quality (Supporting Information Supplement S3).
Meta-analysis results A two-stage meta-analysis was performed to estimate the association between night shift work and the MetS risk by combining results from all eligible studies, giving an adjusted RR of 1.57 (95% CI = 1.24–1.98, pheterogeneity = 0.001) (Fig. 2); however, a higher combined RR was indicated for cohort and nested case-control studies (RR = 2.03, 95% CI = 1.31–3.15, pheterogeneity = 0.07) than that of the cross-sectional studies (RR = 1.39, 95% CI = 1.08–1.80, pheterogeneity = 0.015). Table 2 showed an indication of a positive dose–response relationship between years of night shift and the MetS risk. Results from further meta-analysis (Fig. 3) revealed that a higher combined RR was associated with longer years of night shift work compared with the shorter one (RR = 1.77 95% CI = 1.32–2.36 vs. RR = 1.16 95% CI = 0.62–2.15). We performed subgroup analyses by the MetS criteria, gender and countries to address the sources of heterogeneity (Fig. 4). The pooled effects were 1.84 (95% CI = 1.45– 2.34, pheterogeneity = 0.37) for studies using the NCEP-ATPIII criteria and the effect was slightly higher than those using the IDF criteria (RR = 1.57, 95% CI = 1.29–1.92, pheterogeneity = 0.82). A slightly higher association was observed for studies of female workers (RR = 1.61, 95% CI = 1.10–2.34 vs. RR = 1.36, 95% CI = 1.03–1.81 for the males) and in countries other than Asia (RR = 1.65, 95% CI = 1.39–1.95 vs. RR = 1.35, 95% CI = 0.92–1.92 for Asian countries). Only six of all eligible studies reported specific effects from individual components of the MetS. Results from meta-analyses based on these available data showed that night shift work was significantly associated with an
18:00–06:00 night shift and 10 nights/month
Three rotating shift, and 21:00–05:00 for night shift
20:00–04:00 for night work
19:00–07:00 for night work
34.4 ± 8.4 years, male
35–69 years, male and female
39–60 years, male
39.5 ± 7.6 years, male and female
33.6 ± 7.1 years, female
34–64 years, male
Sookoian et al. Argentina 2007 (23)
Copertaro et al. Italy 2008 (11)
Esquirol et al. France 2009 (22)
Violanti et al. USA 2009 (21)
Chen et al. Taiwan 2010 (20)
Kawada et al. Japan 2010 (19)
Two or three rotating shift 15:00–06:30 for two shift 22:30–06:30 for three shift
22:00–6:00 night shift and ≥1.5 nights/month
Night work ≥1 night/week
30–59 years, male
Cross-sectional study Nagaya et al. Japan 2002 (24)
Definition of night shift
Age, gender
Studies, location
JASSO
NCEP-ATPIII
NCEP-ATPIII
NCEP-ATPIII
NCEP-ATPIII
NCEP-ATPIII
Self-defined, ≥3 insulin resistance markers
MS criteria
None
Age, smoking, drinking, education and duration of work
Gender, age, smoking, education, alcohol intake, education, marital status, police rank and physical activity
Age, work organization, total physical activity, job strain index, smoking status, alcohol consumption, carbohydrate intake, total energy intake and contributions to daily energy intake of meals from breakfast, morning, afternoon and night light meals
None
Age, physical activity
Body mass index, job, drinking, smoking and exercise
Adjusted confounders
Table 1 Main characteristics of the metabolic syndrome studies and night shift work included in the meta-analysis
386
N.A.
16
37
16
274
108
MS cases
Two-shift: 0.77 (0.61–0.98) Three-shift: 1.39 (0.92–2.09)
1.80 (0.70–4.30)
1.57 (0.41–5.95)
2.33 (1.04–5.23)
1.11 (0.39–3.14)*
1.51 (1.01–2.25)
For 30–39 years old: 2.77 (0.77–10.00); For 40–49 years old: 1.42 (0.67–2.87); For 50–59 years old: 0.75 (0.32–1.57)
RR or OR (95% CI)
14
14
17
16
14
15
17
Score
4 Night shift work and metabolic syndrome F. Wang et al.
obesity reviews
© 2014 The Authors obesity reviews © 2014 World Obesity
© 2014 The Authors obesity reviews © 2014 World Obesity
Healthy worker 24–65 years, male and female
32–62 years, male and female
Puttonen et al. Finland 2012 (18)
Tucker et al. France 2012 (17)
21:00–07:00 night shift and >4 nights/month
38.4 ± 8.4 years†, male and female
25–59 years, male and female
Pietroiusti et al. Italy 2010 (14)
Li et al. Japan 2011 (16)
NCEP-ATPIII
NCEP-ATPIII
NCEP-ATPIII
Modified IDF
IDF
IDF
MS criteria
Age, sex, job type, drinking habits, smoking habits, leisure-time physical activity, work intensity, sleep duration, vegetable intake and snack food habits
Age, gender, smoking, alcohol intake, familiar history, physical activity and work schedule
Age, baseline insulin resistance status, metabolic syndrome components, job and lifestyle factors.
Age, smoking, physical activity outside work, educational level, job strain, physical job demands, waist circumference, diastolic blood pressure, high-density lipoprotein cholesterol
Age, sex, socioeconomic status, smoking, alcohol, stress and sleep difficulty.
Age, education, physical activity, alcohol use, smoking and insomnia symptoms
Adjusted confounders
417
42
42
364
68
516
MS cases
1.87 (1.13–3.08)
5.10 (2.15–12.11)
Night shift: 3.70 (1.50–9.20) Former shift: 0.70 (0.20–2.40)
1.46 (1.04–2.07)
1.78 (1.03–3.08)
Current night shift: 1.37 (0.84–2.22) [male] 1.31 (0.54–3.17) [female] Former shift: 2.00 (1.26–3.19) [male] 1.51 (0.83–2.77) [female]
RR or OR (95% CI)
16
18
17
18
17
16
Score
*Recalculated with the original data. † Recalculated with means, standard deviation and sample size of day time and night shift workers. F, female; IDF, International Diabetes Federation; JASSO, The Japan Society for the Study of Obesity; M, male; NCEP-ATPIII, The National Cholesterol Education Program Expert Panel on Education-Adult Treatment Panel III; N.A., not available.
Ever shift work
Two rotating shift, 19:30–07:30 for night shift
32.8 ± 7.9 years, female
Two or three rotating shifts
Ever shift work
Three rotating shift work including night work >3 h, between 23:00 and 06:00 for night shift
Definition of night shift
Lin et al. Taiwan 2009 (15)
Cohort/nested case-control study De Bacquer et al. 35–59 years, male Belgium 2009 (13)
Age, gender
Studies, location
Table 1 Continued
obesity reviews Night shift work and metabolic syndrome F. Wang et al. 5
6
obesity reviews
Night shift work and metabolic syndrome F. Wang et al.
Studies (year)
Years of night shift work
De Bacquer et al. 2009 (13)
