Social jetlag and obesity: A systematic review and meta-analysis

Summary

Social jetlag, the weekly variation in sleep timing, is proposed to contribute to increased obesity risk, potentially because of the misalignment of behavioral cycles relative to the endogenous circadian timing system. This systematic review and meta-analysis aim to determine the association between social jetlag and adiposity-related measures using observational studies. We reviewed 477 references, of which 43 studies met inclusion criteria with a total sample size of 231,648. There was a positive association between social jetlag and body mass index (correlation coefficient [r]: 0.12; 95%CI, 0.07, 0.17; P < 0.001; I² = 94.99%), fat mass (r: 0.10; 95%CI, 0.05, 0.15; P < 0.001; I² = 0.00%), fat mass index (fat mass divided by height in meter squared, β: 0.14 kg/m²; 95%CI, 0.05, 0.23; P < 0.001; I² = 56.50%), percent of body fat (r: 0.37; 95%CI, 0.33, 0.41; P < 0.001; I² = 96.17%), waist circumference (r: 0.15; 95%CI, 0.06, 0.24; P = 0.001; I² = 90.83%), and the risk of having overweight/obesity (odds ratio: 1.20; 95%CI, 1.02, 1.140; P = 0.039; I² = 98.25%). Social jetlag is positively and consistently associated with multiple obesity-related anthropometric measures. Further studies are needed to test causality, underlying mechanisms, and whether obesity interventions based on increasing regularity of the sleep/wake cycle can aid in the battle against the obesity pandemic.

1 |. INTRODUCTION

More than 40% of U.S. adults have obesity^1,2 with rates still rising such that by 2030 half of the U.S. population is estimated to have obesity and a quarter severe obesity.³ Recent studies have provided convincing evidence that circadian misalignment, that is, mistiming between our behavioral cycles relative to our internal circadian timing system (our ‘biological clock’) has adverse cardiometabolic consequences^4–6. Social jetlag, the variation of sleep timing between work/school days and free days is thought to reflect the misalignment between our circadian system and our behavioral and environmental schedules caused by social demands such as work hours and school start times.⁷ While milder than the circadian misalignment experienced by, for example, night workers, it is highly prevalent in the general population.⁸ It affects millions of individuals in industrialized nations, being absent during the toddler age and increasing from 6 to 18 years, continuing through the employed years, and generally decreasing after retirement.⁹ Social jetlag has been quantified as the difference in the midpoint of the sleep episodes between work/school days and free days,⁷ and it has been linked with an increased risk of overweight/obesity.⁹ However, not all studies are in agreement regarding the link of social jetlag with obesity and related measures,^9–20 and the exact underlying mechanisms of such an association are not yet fully understood.²¹

While social jetlag is not a pathology per se, it has been associated with several pathophysiological alterations related to obesity and metabolic syndrome.^9,22,23 Furthermore, it has been related to a positive energy balance and alterations in several hormonal profiles.²⁴ Social jetlag has also been connected to obesity-related metabolic consequences, involving an elevation of leptin levels, a rise of ghrelin levels, and increased serum concentration of nocturnal cortisol.^24–26

Although a growing number of studies have proposed that social jetlag is related to adiposity measures, their findings are still controversial. For example, some provided evidence in agreement with a positive association between social jetlag and adiposity measures.^{13,14,19,22,27,28} Conversely, others suggested an inverse association between social jetlag and anthropometric indices^29,30 together with some other citations without significant results.^17,31,32

Furthermore, to our knowledge, no systematic review and meta-analysis has been carried out on the topic until now. Hence, the current systematic review and meta-analysis aimed to synthesize the available data in terms of the association between social jetlag and adiposity-related measures using observational studies in an attempt to achieve a better understanding of the potential association between social jetlag and obesity.

2 |. METHODS

2.1 |. Protocol and registration

The protocol of the current study was registered in the International Prospective Register of Systematic Reviews (PROSPERO; Registration No. CRD42022381790; available from https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42022381790) and was reported in agreement with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) Statements.³³

2.2 |. Eligibility criteria

The Population, Intervention (Exposure), Comparison, Outcomes, and Study design [PI(E)COS] framework was implemented to formulate a research question (that in this case was whether social jetlag was associated with obesity) and to facilitate the search and study selection process.³⁴

2.2.1 |. Population

The focus of the current review was on the general population and not on specific clinical populations. Therefore, although some of the analyzed studies included individuals with obesity, metabolic syndrome, or prediabetes, they mainly comprised healthy individuals. Studies that comprised pregnant or lactating women, patients in a hospital setting or acute care, and patients diagnosed with type 2 diabetes mellitus, were excluded from this review. There were no exclusion criteria in terms of the age range of participants.

2.2.2 |. Intervention (exposure)

The intervention (exposure) was social jetlag referred to as the midpoint of sleep on free days (MSF) – mid-point of sleep on workdays (MSW).⁷ Studies were eligible if they used subjective (e.g., self-reported) or objective (e.g., actigraphy) measures to define social jetlag.

2.2.3 |. Comparison

Comparisons were performed between individuals with social jetlag and those without social jetlag (or very low values) as determined in the original studies. However, there were no inclusion criteria in terms of having a control group or comparator.

2.2.4 |. Outcomes

The current study aimed to examine the association between social jetlag and several obesity-related anthropometric measures with a total of 11 outcomes including percent of body fat (PBF), fat mass index (FMI; FM divided by height in m²), fat mass (FM), neck circumference (NC), waist-to-height ratio (WHtR), waist-to-hip ratio (WHR), waist circumference (WC), weight, body mass index (BMI), and the risk of overweight/obesity and abdominal obesity were examined.

2.2.5 |. Study design

All types of observational studies, including case-control, longitudinal, and cross-sectional studies, were included. Only full-text peer-reviewed in-press or published studies were eligible for inclusion. Review articles, poster abstracts, editorials, commentary, and case reports were excluded. There was no restriction in terms of the sample size.

2.3 |. Search strategy

A systematic search of the selected electronic databases was carried out to identify eligible studies from the earliest available date up until January 2023 via predefined keywords by two independent reviewers (E.K. and A.A.). The following keywords were used: “social jet lag” OR “social jet-lag” OR “social jetlag” AND “obes*” OR “overweight” OR “BMI” OR “body mass index” OR “weight” OR “adiposity” OR “waist circumference” OR “waist-to-hip ratio” OR “waist-to-height ratio” OR “body mass” OR “body fat” OR “fat mass” OR “neck circumference”. Details of the tailored search strategy in each database are shown in Table S1. No filtering was employed in terms of publication time, language, and study design upon database searching. Even though Google Scholar does not enable full complex searches, this database was also manually searched using the above keywords to increase the power of the search strategy and minimize the chance of missing any eligible articles. Finally, the citation list of the eligible studies was also checked.

2.4 |. Study selection

Search results of the selected databases were imported into the EndNote 20 software (Thomson Corporation, Stamford, USA) to facilitate the study selection process and to remove duplicate references. The study selection process was performed by two independent investigators (A.A. and E.K.), and any discrepancy was resolved through discussion with a third investigator (F.A.J.L.S.), if required. In the first step of screening, titles, and abstracts of potentially eligible studies were screened. In the second step of screening, full-text of studies that fulfilled the initial screening were assessed by two independent investigators.

2.5 |. Data extraction

The process of data extraction was performed by one investigator (A.A.) and checked by another investigator (E.K.) using predefined Microsoft Excel sheets. Investigators were not blinded to the journals or authors while extracting data. When different models were reported, the fully adjusted models’ results were extracted for meta-analysis. Important study features (i.e., population, exposure, comparator, outcome, first author, publication year, study design, country, latitude zone, age, sample size, calculation of social jetlag, calculation of midpoint of sleep, the instrument used to calculate social jetlag, and covariates) were extracted.

2.6 |. Risk of bias and study quality assessment

The quality of the enrolled studies was assessed using the Newcastle–Ottawa Scale (NOS) by two independent investigators (A.A. and E.K.). The NOS evaluates the quality of articles in terms of three domains including selection, comparability, and outcome with a total score of nine for case-control and longitudinal studies and 10 for cross-sectional ones. Study quality was not eligible for inclusion, so all available evidence was included.³⁵

Using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework, we assessed the quality or certainty of evidence. The GRADE tool categorized the certainty of evidence into four categories (very low, low, moderate, and high), and the rating started at low for nonrandomized or observational studies and high for randomized studies. The quality of evidence can be downgraded if there are serious limitations across studies (e.g., imprecision, indirectness, inconsistency of relative treatment effects, serious risk of bias, or other factors).³⁶

2.7 |. Statistical analysis

Quantitative analysis of the association between social jetlag and adiposity measures was carried out using STATA software version 17.0 (Stata Corporation, College Station, TX, USA). In the current meta-analysis, we estimated three types of effect sizes: (1) Fisher’s Z of the correlation between social jetlag and BMI, FM, PBF, and WC; (2) Beta (β) of the association between social jetlag and FMI; and (3) Ln odds ratio (OR) of the relationship between social jetlag and the risk of overweight/obesity and abdominal obesity. Fisher’s Z and corresponding standard error (SE) were calculated using sample size (N) and correlation coefficient via the relevant formula.³⁷ OR in each study was converted to the effect size of interest by using natural logarithms of OR, and SE was also estimated using the corresponding 95% confidence interval of OR. Following proper conversion, quantitative analysis was carried out using a fixed (inverse-variance³⁸) or random (DerSimonian–Laird³⁹) weighted model. Inconsistency and heterogeneity between enrolled studies were investigated using the I² statistic and Chi-squared test, respectively. A significance level of the Chi-squared test was set at P < 0.05 while I² was interpreted as trivial (0–25%), low (25.1–50%), moderate (50.1–75%), or high (75.1–100%).⁴⁰ Publication bias was examined via qualitative (i.e., visual inspection of the funnel plot) and quantitative (Begg’s and Egger’s test) approaches. If evidence of publication bias was detected, the trim-and-fill analysis was implemented to impute possible missed studies and attenuate asymmetry in the funnel plot.⁴¹ Leave-one-out meta-analysis was implemented to assess the integrity of findings by keeping out individual studies to see if they affected overall findings. Subgroup analysis was performed, whenever feasible, to detect possible sources of heterogeneity. If a publication provided the results for the relationship between social jetlag and anthropometric parameters stratified by certain variables, we assumed that these groups were independent and therefore divided into two different datasets. A two-sided P value < 0.05 was considered statistically significant.

3 |. RESULTS

3.1 |. Studies characteristics and findings of the systematic search

PRISMA flow diagram of the study selection process is illustrated in Figure 1. Upon systematic search of selected databases, a total of 477 articles were found, which was reduced to 328 after deduplication. Then, the title/abstract of the remaining articles were screened, and 188 records were excluded because they did not meet the inclusion criteria. Finally, 140 articles went forward for eligibility assessment on the basis of PICOS components, and 43 studies met the inclusion criteria.

FIGURE 1.

Flow diagram for the selection of studies.

A total of 43 articles were enrolled in the current systematic review and meta-analysis with a sample size ranging from 69 to 34,651 (Table 1). Enrolled studies were published between 2012 and 2023 with a mean BMI and age ranging from 17.7 to 32.0 kg/m² and 9.6 to 63.6 years, respectively. The age group of the target population was children/adolescents in 17 studies^{12,13,15,16,18,27,28,42–51} and adults in 24 studies.^{10,11,14,17,19,20,22,23,29–32,53–64} However, two studies investigated the link between social jetlag and obesity in both children/adolescents and adults.^9,64 The included studies were conducted in the United States,^{10,12,15,19,27,43,44,48,50,52,56,62} United Kingdom,⁵² Egypt,³¹ Thailand,⁵³ Turkey,^11,46 Brazil,^17,29,54 the Netherlands,^13,23,62 Czech Republic,³⁰ Korea,^14,56,59 Japan,⁵⁷ Germany,^9,45 Norway,⁵⁸ China,^16,20,64 Taiwan,⁶⁰ Hong Kong,⁴⁸ New Zealand,^22,28,44,50 Iceland,¹⁸ Finland,³² and Spain.⁶³ Thirty-six citations were cross-sectional in design,^{9–12,14–20,23,27–32,42–44,48–56,58–60,62–64} six were longitudinal,^{13,22,45,47,57,61} and one was case-control.⁴⁶ Thirty-seven studies were conducted in the general population,^{9–20,22,23,27,28,30,32,42–52,54,55,58,59,61–64} two in patients with metabolic syndrome,^31,57 two in healthy workers,^29,60 one in patients with prediabetes,⁵³ and one in shift-working nurses.⁵⁶ Anthropometric indices were measured in the clinical setting in 30 studies^{10–13,15,17–19,22,23,27–29,31,32,42,44–48,52–54,57–62} and were self-reported in 11 studies,^{9,14,16,30,43,49–51,55,56,63} whereas the rest did not report the approach.^20,64 To measure social jetlag, 17 studies utilized open questions,^{11,15–17,28,31,32,48–53,55,57,59,63} 15 utilized Munich ChronoType Questionnaire (MCTQ),^{9,13,14,20,22,23,29,30,43,45,54,56,58,60,62} 10 utilized actigraphy,^{10,12,18,19,27,42,44,47,61,64} and one utilized Childhood Chronotype Questionnaire (CCQ) ⁴⁶ (Table 1).

TABLE 1.

Characteristics of included studies.

Author, year	Geographic location	Latitude zone	Sample size (F/M)	Age (mean)	BMI (kg/m2)	Study design
Abbott et al, 2019	United States	Midlatitude	1401/755	47	NM	Cross-sectional
Al Khatib et al, 2022	United Kingdom	Subarctic and midlatitude	2959/2056	42.6	27.5	Cross-sectional
Ali et al., 2020	Egypt	Subtropical	60/90	46	31.96	Cross-sectional
Anothaisintawee et al., 2018	Thailand	Equatorial and tropical	1401/732	63.6	25.8	Cross-sectional
Berry et al., 2021	United States	Midlatitude	142/119	15.2	NM	Cross-sectional
Bodur et al., 2021	Turkey	Midlatitude	513/197	21.58	22.12	Cross-sectional
Brum et al., 2020	Brazil	Subtropical	139/61	43.2	27.55	Cross-sectional
Cespedes Feliciano et al., 2019	United States	Midlatitude	418/386	13.2	20.9	Cross-sectional
Cetiner et al., 2021	United States	Midlatitude and subtropical	782/768	14.5	NM	Cross-sectional
Constantino et al., 2022	Brazil	Subtropical	221/144	47	26.5	Cross-sectional
Dashti et al., 2021	United States	Midlatitude	20,609/14,042	54.5	27.4	Cross-sectional
De Zwart et al., 2018	The Netherland	Midlatitude	46/37	16.6	21.6	Cohort
Farkova et al., 2020	Czech Republic	Midlatitude	1964/739	32.75	24.55	Cross-sectional
Higgins et al., 2021	New Zealand	Midlatitude	189/192	10.23	NM	Cross-sectional
Hwang et al., 2023	Korea	Midlatitude	149/34	27	21.5	Cross-sectional
Islam et al., 2018	Japan	Midlatitude	159/1005	44.6	NM	Cohort
Jang et al., 2021	Korea	Midlatitude	304 F	20.56	20.74	Cross-sectional
Jankovic et al., 2021	Germany	Midlatitude	95/118	12.9	19	Cohort
Johnsen et al., 2013	Norway	Subarctic	3,356/3,053	30–65	26.7	Cross-sectional
Johnson et al., 2020	United States	Midlatitude and subtropical	593/615	12.3	22.4	Cross-sectional
Karadag et al., 2021	Turkey	Midlatitude	75/86	12.25	NM	Case-control
Kim et al., 2020	Korea	Midlatitude	4,878/3,417	41.17	23.91	Cohort
Koopman et al., 2017	The Netherland	Midlatitude	840/745	59.33	26.66	Cross-sectional
LeMay-Russell et al., 2021	United States	Midlatitude	74/63	12.5	NM	Cohort
Liang et al., 2022	China	Subtropical	1,691/1,842	14.67	19.74	Cross-sectional
Li et al., 2022	Taiwan	Subtropical	3,011/880	33.69	-	Cross-sectional
Lo et al., 2019	Hong Kong	Equatorial and tropical	542/536	15.35	NM	Cross-sectional
Malone et al., 2016	United States	Midlatitude	51/18	15.50	NM	Cross-sectional
Mathew et al., 2020	United States	Midlatitude and subtropical	1473/1587	15.59	24.02	Cross-sectional
McMahon et al., 2019	United States	Subtropical	198/192	27.6	20–35	Cohort
Mota et al., 2017	Brazil	Equatorial and tropical	581/211	55.9	30	Cross-sectional
Parsons et al., 2015	New Zealand	Midlatitude	498/539	38	NM	Cohort
Randler et al., 2013	Germany	Midlatitude	507/406	13.7	20	Cross-sectional
Roenneberg et al., 2012	Germany	Midlatitude	64,110	NM	NM	Cross-sectional
Rognvaldsdottir et al., 2017	Iceland	Subarctic	179/122	15.9	21.95	Cross-sectional
Rutters et al., 2014	The Netherland		78/67	27.75	24.7	Cross-sectional
Schneider et al., 2020	United States	Midlatitude and subtropical	632/622	14.56	NM	Cross-sectional
Stoner et al., 2018	New Zealand	Midlatitude	169/172	9.6	17.7	Cross-sectional
Suikki et al., 2021	Finland	Subarctic	3757/3022	49.62	27.01	Cross-sectional
Wong et al., 2015	United States	Midlatitude	447	42.7	26.8	Cross-sectional
Zeron-Rugerio et al., 2019	Spain	Midlatitude	397/137	21.5	21.7	Cross-sectional
Zhang et al., 2018	China	Subtropical	564/413	20.06	NM	Cross-sectional
Zhang et al., 2019	China	Subtropical	71,176	NM	NM	Cross-sectional
Author, year	Population health status	The approach of adiposity measurement	Instrument used to define SJL	Calculation of SJL	Calculation of midpoint of sleep	Covariates
Abbott et al, 2019	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, Hispanic/Latino background, study site, income, acculturation, education level, sleep duration, shift work status
Al Khatib et al, 2022	General population	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, ethnicity, SES, smoking status, alcohol intake, number of children below 4 years of, age and long-standing illness
Ali et al., 2020	Metabolic syndrome	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	-
Anothaisintawee et al., 2018	Prediabetes	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	-
Berry et al., 2021	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	-
Bodur et al., 2021	General population	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	-
Brum et al., 2020	Healthy worker	Clinical setting	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Gender, age
Cespedes Feliciano et al., 2019	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	Age, race/ethnicity, pubertal status, season of measurement, parental SES, actigraphy-measured sleep duration, parent report of adolescent physical activity, adolescent report of television viewing, diet quality
Cetiner et al., 2021	General population	Self-report	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, ethnicity, family income, physical activity, dietary intake, sleep duration
Constantino et al., 2022	General population	Clinical setting	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender
Dashti et al., 2021	General population	Self-report	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, race and ethnicity, and employment status
De Zwart et al., 2018	General population	Clinical setting	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Gender, average sleep duration, physical activity, hours of television use, chronotype
Farkova et al., 2020	General population	Self-report	MCTQ	Weekend sleep duration-weekday sleep duration	SJL calculation did not use midpoint of sleep	-
Higgins et al., 2021	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, ethnicity, school decile, presence of a screen in the bedroom, moderate-to-vigorous intensity physical activity, and dietary intake
Hwang et al., 2023	Shift-working nurses	Self-report	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Chronotype, exercise time, and eating habits
Islam et al., 2018	Metabolic syndrome	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, site, job, marital status, overtime work, physical activity, smoking, alcohol, skipping breakfast, habitual snacking at night, energy intake, and sleep quality
Jang et al., 2021	General population	Self-report	MCTQ	MSFsc-MSW	Midpoint between sleep onset and sleep end	Age, living alone, chronotype, light exposure, sleep duration, sleep quality, premenstrual symptoms, depression
Jankovic et al., 2021	General population	Clinical setting	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Age at baseline, sex, time between last and first measurement, age at takeoff, persons in the household, maternal BMI
Johnsen et al., 2013	General population	Clinical setting	MCTQ	MSFsc-MSW	Midpoint between sleep onset and sleep end	Age
Johnson et al., 2020	General population	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, site, ethnicity, sleep duration, education, season, diet, physical activity,
Karadag et al., 2021	General population	Clinical setting	CCQ	MSFsc-MSW	Midpoint between sleep onset and sleep end	Number of siblings, parents’ age
Kim et al., 2020	General population	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, SES, lifestyle behaviors (smoking status, alcohol consumption, regular exercise, total energy intake, and fat intake), sleep duration
Koopman et al., 2017	General population	Clinical setting	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Gender, employment status, educational level, sleep duration
LeMay-Russell et al., 2021	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	Baseline fat mass and height
Liang et al., 2022	General population	Self-report	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, monthly household income, physical activity, intake of all beneficial foods, intake of all detrimental foods, and sleep duration
Li et al., 2022	Healthy worker	Clinical setting	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	age, gender, sleep quality, long working hours, alcohol consumption, and smoking
Lo et al., 2019	General population	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, time spent watching television, time spent on homework, time spent on computers, and eating 1 hour before going to bed
Malone et al., 2016	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	Gender, fruit/vegetable intake, screen time, school night sleep duration, daytime naps
Mathew et al., 2020	General population	Self-report	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, living arrangements, household income
McMahon et al., 2019	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	Time, sex, physical activity
Mota et al., 2017	General population	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, sex, employment status, use of sleeping pills, use of antidepressants
Parsons et al., 2015	General population	Clinical setting	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Gender, chronotype, sleep duration, smoking, SES
Randler et al., 2013	General population	Self-report	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Screen time, fast-food consumption, nutrition-related self-efficacy, parental control
Roenneberg et al., 2012	General population	Self-report	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender
Rognvaldsdottir et al., 2017	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	Gender
Rutters et al., 2014	General population	Clinical setting	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Gender, age
Schneider et al., 2020	General population	Self-report	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Gender, age, race, school type, parental marital status, home ownership, parental education, moderate-vigorous physical activity, teen sedentary time, teen frequency of beneficial foods eaten, teen frequency of detrimental foods eaten, and time spent watching television
Stoner et al., 2018	General population	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	School decile, ethnicity, gender, age
Suikki et al., 2021	General population	Clinical setting	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, education, smoking, leisure-time physical activity, and energy intake
Wong et al., 2015	General population	Clinical setting	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, race, educational attainment, family income, sleep characteristics, and work status
Zeron-Rugerio et al., 2019	General population	Self-report	Open questions	MSF-MSW	Midpoint between sleep onset and sleep end	Age, gender, physical activity, and sleep duration
Zhang et al., 2018	General population	NM	MCTQ	MSF-MSW	Midpoint between sleep onset and sleep end	Age; gender; race; grade; the type of university; mother’s education; father’s education; outdoor activities on work days or free days
Zhang et al., 2019	General population	NM	Actigraphy	MSF-MSW	Midpoint between sleep onset and sleep end	-

Abbreviations: BMI: body mass index; F: female; M: male; SJL: social jetlag; MSF: midpoint of sleep on free days; MSW: midpoint of sleep on work days; CCQ: childhood chronotype questionnaire; MCTQ: Munich chronotype questionnaire; NM: not mentioned; SES: socioeconomic status.

The results of the methodological quality assessment are based on the NOS (Tables S2–S4). It shows that from the total 43 studies included, the methodological quality of 33 studies^{9,10,12,14–20,22,23,27–29,32,42–45,48–52,54–56,58–60,62,63} ranked as good (low risk of bias), and 10 studies^{11,13,30,31,46,47,53,57,61,64} as fair (moderate risk of bias).

3.2 |. Findings from qualitative and quantitative analyses

3.2.1 |. Social jetlag and body weight

Three studies^17,29,59 assessed the relationship between social jetlag and body weight with an overall sample size of 9,287 participants. No quantitative analysis was performed because of the small number of available comparable studies (differences in study design and reported effect sizes). In a cross-sectional study, Brum et al.²⁹ indicated that each unit of increase in social jetlag was associated with 14% smaller odds of excess body weight among health-care workers. Another cross-sectional study failed to provide a link between social jetlag and body weight.¹⁷ In a prospective study, qualitative synthesis revealed a significant relationship between higher values of social jetlag with increased body weight gain.⁵⁹

3.2.2 |. Social jetlag and BMI

Twenty-two datasets from 20 publications with a total of 75,557 participants examined the association between social jetlag and BMI.^{12–15,17–19,22,27,30,31,44,46,47,49,50,53,63,64} There was a positive correlation between social jetlag and BMI (Correlation coefficient: 0.12; 95% CI, 0.07, 0.17; P < 0.001) with evidence of substantial heterogeneity (I² = 94.99%, P < 0.001) (Figure 2). Following subgroup analysis, social jetlag and BMI were linked only among good-quality studies (correlation coefficient: 0.14; 95% CI, 0.07, 0.21; I² = 92.6%) and those carried out in the midlatitude zone (correlation coefficient: 0.18; 95% CI, 0.09, 0.25; I² = 93.7%) (Table 2). The overall finding was not influenced by keeping out individual studies. No evidence of significant publication bias (Egger’s test: P = 0.722, Begg’s test: P = 0.955) was detected (Figure S1).

FIGURE 2.

Forest plot for the association between social jetlag and body mass index. The square in the middle of each straight line represents the mean of the correlation coefficient, whereas the left and right extremes of the horizontal lines represent the corresponding 95% CI. The middle of the diamond represents the overall mean for the correlation coefficient, whereas the left and right extremes of the diamond represent the corresponding 95% CI.

TABLE 2.

Subgroup analysis of the association between social jetlag and adiposity measures.

Sub-grouped by	No. of studies	Effect sizea	95% CI	I2 (%)	P for heterogeneity	P for between subgroup heterogeneity
BMI
Latitude zone						< 0.001
Midlatitude	14	0.18	0.09, 0.25a	93.7	< 0.001
Longitudinal/case-control	8	0.03	-0.01, 0.07	82.3	< 0.001
Continent						< 0.001
America	9	0.13	0.05, 0.22a	93.1	< 0.001
Europe and Africa	8	0.08	-0.03, 0.20	92.3	< 0.001
Asia and Oceania	5	0.12	0.03, 0.20a	94.6	< 0.001
Age						< 0.001
< 18 years	11	0.11	0.05, 0.16a	75.8	< 0.001
≥ 18 years	11	0.11	0.03, 0.19a	97.2	< 0.001
Study design						< 0.001
Cross-sectional	18	0.10	0.04, 0.16a	95.6	< 0.001
Others	4	0.17	0.09, 0.26a	42.2	0.158
Instrument used to assess SJL						< 0.001
Objective	9	0.12	0.01, 0.23a	96.7	< 0.001
Subjective	13	0.11	0.05, 0.18a	91.9	< 0.001
Study quality						< 0.001
Good	14	0.14	0.07, 0.21a	92.6	< 0.001
Fair	8	0.04	-0.01, 0.09	79.2	< 0.001
WC
Latitude zone						0.004
Midlatitude	6	0.17	0.10, 0.24a	75.7	0.001
Others	4	0.05	-0.19, 0.29	95.7	< 0.001
Continent						< 0.001
America	3	0.16	-0.04, 0.36	96.7	< 0.001
Europe and Africa	6	0.16	0.05, 0.26a	77.0	0.001
Asia and Oceania	1	0.07	0.009, 0.13a	-	-
Age						< 0.001
< 18 years	3	0.33	0.28, 0.37a	0.0	0.462
≥ 18 years	7	0.08	0.02, 0.14a	65.6	0.008
Study design						0.108
Cross-sectional	7	0.11	-0.005, 0.22	92.1	< 0.001
Longitudinal/case-control	3	0.24	0.02, 0.44a	89.6	< 0.001
Instrument used to assess SJL						0.671
Objective	1	0.15	0.06, 0.23a	-	-
Subjective	9	0.15	0.04, 0.25a	91.8	< 0.001
Study quality						0.178
Good	6	0.14	0.03, 0.24a	93.6	< 0.001
Fair	4	0.14	-0.11, 0.38	83.0	< 0.001
Overweight/obesity
Latitude zone						< 0.001
Midlatitude	8	1.39	1.04, 1.86a	89.7	< 0.001
Others	21	1.13	0.93, 1.37	98.26	< 0.001
Continent
America	16	1.14	0.92, 1.41	98.62	< 0.001
Europe and Africa	4	1.35	0.64, 2.84	94.11	< 0.001
Asia and Oceania	9	1.14	1.03, 1.25a	53.85	0.027
Age						< 0.001
< 18 years	10	1.22	0.93, 1.61	97.35	< 0.001
≥ 18 years	19	1.14	1.02, 1.27a	86.03	< 0.001
Study design						< 0.001
Cross-sectional	27	1.20	1.02, 1.42a	98.33	< 0.001
Longitudinal/case-control	2	1.06	0.82, 1.36	62.05	0.105
Instrument used to assess SJL						< 0.001
Objective	5	1.05	0.86, 1.28	44.73	0.124
Subjective	24	1.22	1.02, 1.45*	98.50	< 0.001
Study quality						< 0.001
Good	28	1.20	1.02, 1.41*	98.28	< 0.001
Fair	1	0.93	0.73, 1.17	-	-

Abbreviation: BMI: body mass index; WC: waist circumference; SJL: social jetlag.

^aCalculated by random-effects model.

^*P < 0.05 was considered statistically significant.

3.2.3 |. Social jetlag and FM

Three studies addressed the link between social jetlag and FM among 1,555 individuals ^22,44,47. Social jetlag was significantly associated with increased FM (correlation coefficient: 0.10; 95% CI, 0.05, 0.15; P < 0.001) with no evidence of heterogeneity (I² = 0.0%, P = 0.873) (Figure 3). The overall findings did not change after keeping out individual studies. No evidence of significant publication bias was also detected (Egger’s test: P = 0.741, Begg’s test: P > 0.99) (Figure S2).

FIGURE 3.

Forest plot for the association between social jetlag and fat mass. The square in the middle of each straight line represents the mean of the correlation coefficient, whereas the left and right extremes of the horizontal lines represent the corresponding 95% CI. The middle of the diamond represents the overall mean for the correlation coefficient, whereas the left and right extremes of the diamond represent the corresponding 95% CI.

3.2.4 |. Social jetlag and FMI

Quantitative analysis of four datasets reported in three publications^12,28,44 in a total of 1,410 individuals on the relationship between social jetlag and FMI showed a significant positive association between social jetlag values and FMI (β: 0.14 kg/m²; 95% CI, 0.05, 0.23; P < 0.001) among those with social jetlag compared to those without social jetlag with evidence of moderate heterogeneity (I² = 56.50%, P = 0.08) (Figure 4). Overall meta-analysis result was not influenced by keeping out individual studies. There was evidence of marginally significant publication bias (Egger’s test: P = 0.059, Begg’s test: P = 0.089) (Figure S3). Trim-and-fill analysis added one dataset to our previous one, and the new analysis revealed that there was no significant link between social jetlag and FMI (β: 0.19 kg/m²; 95% CI, 0.06, 0.44; P = 0.135).

FIGURE 4.

Forest plot for the association between social jetlag and fat mass index. The square in the middle of each straight line represents the mean of beta, whereas the left and right extremes of the horizontal lines represent the corresponding 95% CI. The middle of the diamond represents the overall mean for beta, whereas the left and right extremes of the diamond represent the corresponding 95% CI.

3.2.5 |. Social jetlag and PBF

Three datasets comprising 1,672 participants assessed the link between social jetlag and PBF.^13,15,44 Social jetlag was positively and significantly related to PBF (correlation coefficient: 0.37; 95% CI, 0.33, 0.41; P < 0.001) with evidence of significant heterogeneity (I² = 96.17%, P < 0.001) (Figure 5). The overall finding was not influenced by keeping out individual studies. Significant publication bias was not observed (Egger’s test: P = 0.690, Begg’s test: P > 0.99) (Figure S4).

FIGURE 5.

Forest plot for the association between social jetlag and percent of body fat. The square in the middle of each straight line represents the mean of the correlation coefficient, whereas the left and right extremes of the horizontal lines represent the corresponding 95% CI. The middle of the diamond represents the overall mean for the correlation coefficient, whereas the left and right extremes of the diamond represent the corresponding 95% CI.

3.2.6 |. Social jetlag and WC

Ten studies (eight publications) with 5,413 participants examined the association between social jetlag and WC.^{13,15,17,19,22,23,31,46} Overall, the meta-analysis showed that social jetlag was significantly and positively linked with WC (correlation coefficient: 0.15; 95% CI, 0.06, 0.24; P = 0.001) and heterogeneity was significant (I² = 90.83%, P < 0.001) (Figure 6). Subgroup analysis is shown in Table 2. Data show a significant result in the midlatitude zone (correlation coefficient: 0.17; 95% CI, 0.10, 0.24; I² = 75.7%) and studies undertaken in the population of Europe and Africa (correlation coefficient: 0.16; 95% CI, 0.05, 0.27; I² = 77.0%). Likewise, a significant correlation between social jetlag and WC was only observed in longitudinal/case-control (correlation coefficient: 0.24; 95% CI, 0.02, 0.44; I² = 89.6%) and good quality (correlation coefficient: 0.14; 95% CI, 0.03, 0.24; I² = 93.6%) studies (Table 2). The overall finding was not sensitive to the omission of an individual dataset. There was no significant publication bias (Egger’s test: P = 0.308, Begg’s test: P = 0.858) (Figure S5).

FIGURE 6.

Forest plot for the association between social jetlag and waist circumference. The square in the middle of each straight line represents the mean of the correlation coefficient, whereas the left and right extremes of the horizontal lines represent the corresponding 95% CI. The middle of the diamond represents the overall mean for the correlation coefficient, whereas the left and right extremes of the diamond represent the corresponding 95% CI.

3.2.7 |. Social jetlag and NC

Two citations^11,17 looked into the association between social jetlag and NC with an overall 1,502 individuals. No meta-analysis was carried out for NC because of the small number of available studies. Qualitative synthesis indicated a significant positive link between social jetlag and NC in one study¹¹ and a nonsignificant negative association in the other study.¹⁷

3.2.8 |. Social jetlag and WHR

Four studies^28,31,58,61 evaluated the association between social jetlag and WHR among 7,290 participants; however, no meta-analysis could be conducted because of the small number of comparable studies (difference in study design and reported effect sizes). Qualitative synthesis revealed that there was a positive association between social jetlag and WHR in two studies,^28,58 whereas two others found no significant correlation.^31,61

3.2.9 |. Social jetlag and WHtR

Three datasets^27,44,61 examined the link between social jetlag and WHtR comprising 840 subjects. We were unable to do a meta-analysis on social jetlag and WHtR, owing to the small number of comparable studies (difference in study design and reported effect sizes). Qualitative analysis indicated a positive relationship between social jetlag and WHtR in one study;²⁷ however, two others did not confirm this finding.^44,61

3.2.10 |. Social jetlag and risk of overweight/obesity

Twenty-nine citations (17 publications) with a total of 123,509 subjects investigated the link between social jetlag and the risk of having overweight/obesity.^{9,12,15–17,20,22,29,32,43,48,51,54–56,60,61} Social jetlag was significantly associated with higher odds of having overweight/obesity (OR: 1.20; 95% CI, 1.02, 1.40; P = 0.039). There was substantial heterogeneity among the enrolled studies (I² = 98.25%, P < 0.001) (Figure 7). Subgroup analysis revealed that the relationship between social jetlag and the risk of overweight/obesity was only significant between Asia and Oceania populations (OR: 1.14; 95% CI, 1.03, 1.25; I² = 53.85%), midlatitude zone (OR: 1.39; 95% CI, 1.04, 1.86; I² = 89.70%), cross-sectional studies (OR: 1.20; 95% CI, 1.02, 1.42; I² = 98.33%), studies measuring social jetlag via subjective approaches (OR: 1.22; 95% CI, 1.02, 1.45; I² = 98.50%), and among good-quality studies (OR: 1.20; 95% CI, 1.02, 1.41; I² = 98.28%) (Table 2). Meta-analysis results for the link between social jetlag and the risk of overweight/obesity were sensitive to the exclusion of Roenneberg et al. (OR: 1.15; 95% CI, 0.98, 1.35).⁹ Owing to the evidence of significant publication bias (Egger’s test: P = 0.067, Begg’s test: P < 0.001) (Figure S6), trim-and-fill analysis was utilized to identify possible un-detected citations. One citation was imputed; however, the overall finding was still significant regarding the link between social jetlag and the risk of overweight/obesity (OR: 1.20; 95% CI, 1.02, 1.39).

FIGURE 7.

Forest plot for the association between social jetlag and overweight/obesity. The square in the middle of each straight line represents the mean of the odds ratio, whereas the left and right extremes of the horizontal lines represent the corresponding 95% CI. The middle of the diamond represents the overall mean for odds ratio, whereas the left and right extremes of the diamond represent the corresponding 95% CI.

3.2.11 |. Social jetlag and risk of abdominal obesity

Three studies with 1,575 participants investigated the relationship between social jetlag and the risk of abdominal obesity.^29,48,57 Overall, the meta-analysis revealed that there was no significant association between social jetlag and the risk of abdominal obesity (OR: 1.07; 95% CI, 0.96, 1.20; P = 0.210) with evidence of significant heterogeneity (I² = 90.64%, P < 0.001) (Figure 8). Keeping out the study of Brum et al.²⁹ modified the overall effect (OR: 1.93; 95% CI, 1.47, 2.55). There was evidence of publication bias (Egger’s test: P = 0.142, Begg’s test: P > 0.99) (Figure S7).

FIGURE 8.

Forest plot for the association between social jetlag and abdominal obesity. The square in the middle of each straight line represents the mean of the odds ratio, whereas the left and right extremes of the horizontal lines represent the corresponding 95% CI. The middle of the diamond represents the overall mean for odds ratio, whereas the left and right extremes of the diamond represent the corresponding 95% CI.

3.3 |. Certainty of evidence

Certainty of our meta-analysis findings for the relationship between social jetlag and adiposity measures was investigated via the GRADE tool. Owing to the observational nature of the enrolled documents and significant heterogeneity, the level of evidence was very low for BMI, PBF, FMI, WC, risk of overweight/obesity, and abdominal obesity, and low for FM (Table 3).

TABLE 3.

GRADE summary of findings.

Outcome	Study design (number of participants)	Risk of bias	Inconsistency	Indirectness	Imprecision (ES [95%CI])	Publication bias	Other	Quality of evidence
BMI	Observational (N = 75,557)	Most information is from good-quality studies	I2 = 94.99%, P < 0.001	Not at all	0.12 (0.07 to 0.17)	Begg’s test: P = 0.955, Egger’s test: P = 0.722	No	Very low
FM	Observational (N = 1555)	Most information is from good-quality studies	I2 = 0.0%, P = 0.873	Not at all	0.10 (0.05 to 0.15)	Begg’s test: P > 0.99, Egger’s test: P = 0.741	No	Low
FMI	Observational (N = 1410)	Most information is from good-quality studies	I2 = 56.50%, P = 0.075	Not at all	0.14 (0.05 to 0.23)	Begg’s test: P = 0.089, Egger’s test: P = 0.059	No	Very low
PBF	Observational (N = 1672)	Most information is from good-quality studies	I2 = 96.17%, P < 0.001	Not at all	0.37 (0.33 to 0.41)	Begg’s test: P > 0.99, Egger’s test: P = 0.690	No	Very low
WC	Observational (N = 5413)	Most information is from good-quality studies	I2 = 90.83%, P < 0.001	Not at all	0.15 (0.06 to 0.24)	Begg’s test: P = 0.858, Egger’s test: P = 0.308	No	Very low
Overweight/obesity	Observational (N = 123,509)	Most information is from good-quality studies	I2 = 98.25%, P < 0.001	Not at all	1.20 (1.02 to 1.40)	Begg’s test: P < 0.001, Egger’s test: P = 0.067	No	Very low
Abdominal obesity	Observational (N = 1575)	Most information is from good-quality studies	I2 = 90.64%, P < 0.001	Not at all	1.07 (0.96 to 1.20)	Begg’s test: P > 0.99, Egger’s test: P = 0.142	No	Very low

Note: BMI: body mass index; FM: fat mass; FMI: fat mass index; PBF: percent of body fat; WC: waist circumference; ES: effect size; CI: confidence interval.

4 |. DISCUSSION

The current systematic review and meta-analysis aimed to provide a summarized report on the association between social jetlag and obesity-related anthropometric measures. Data from 43 studies comprising 231,648 individuals demonstrated a positive association between social jetlag and adiposity measures including BMI, FM, FMI, PBF, and WC. Moreover, we found that subjects with social jetlag had a significantly higher risk for overweight/obesity compared to those without social jetlag. We further qualitatively summarize the evidence for the link between social jetlag and other anthropometric indices including body weight, NC, WHR, and WHtR. Our findings were limited because of the presence of significant heterogeneity for all the outcomes except for FM. This together with the observational nature of the included studies suggests that the certainty of the evidence is very low, as demonstrated by the GRADE tool, and hence, our findings should be interpreted with caution. The findings of the current study contribute to the existing literature regarding the possible link between social jetlag and obesity, identify gaps in the evidence, and describe recommendations for future studies to address these knowledge gaps. A summary of recommendations for future studies are presented in Table 4.

TABLE 4.

Summary of recommendations for future research.

Overall design	Longitudinal studies with adequate sample size and follow-up duration should be carried out.
Exposure	Measure social jetlag using both objective (e.g., actigraphy) and subjective (e.g., sleep diary) approaches.
	Social jetlag should be calculated as the absolute difference between the mid-point of sleep on work/school days (MSW) and free days (MSF) without correction of the midpoint of sleep on free days for sleep debt. Social jetlag should be analyzed both as a categorical and continuous variable.
Outcome	Various anthropometric measures should be reported in relation to social jetlag including body weight, body mass index, fat mass, fat-free mass, waist circumference, waist-to-hip ratio, waist-to-height ratio, and neck circumference to find the most sensitive indicator related to obesity.
Confounders	A comprehensive set of confounders should be controlled for, including sex, age, sociodemographic characteristics of the population (traditions, behaviors, cultural background, education level, and work/school status and schedule), latitude zone, the season of sampling, and health status of participants.
Other issues	Investigate the underlying mechanisms regarding the link between social jetlag and obesity.

Social jetlag is becoming a primary public health concern because of its high prevalence; more than 50% of the U.S. population is suffering from this issue.⁸ Our meta-analysis showed that social jetlag was linked with an increased risk for overweight/obesity (20% increase). These data may be taken into consideration regarding obesity prevention. Health-care providers should be more aware of the possible role of circadian misalignment in the epidemiology of obesity. Moreover, efforts are needed to educate the general population about the importance of regular and optimal sleep timing.

Several underlying mechanisms may explain the link between social jetlag and obesity. Social jetlag represents a mismatch between socially driven behavioral schedules and the circadian system leading to circadian misalignment. By sleeping at irregular times, social jetlag disturbs the optimal alignment of behaviors such as sleep, eating, and physical activity relative to the circadian system and the related processes it controls.^65,66 For instance, such circadian misalignment caused by social jetlag may disrupt the hypothalamic–pituitary–adrenal (HPA) axis and glycemic and metabolic changes.^65,67 In addition, circadian misalignment may also impair the hypothalamic–pituitary–thyroid (HPT) axis leading to disturbed thyroid hormone secretion.⁶⁸ It was shown that the suprachiasmatic nucleus (SCN), the central pacemaker of the circadian timing system, has a dual control mechanism for thyroid activity by affecting the autonomic input to the thyroid gland on the one hand and neuroendocrine control of thyroid-stimulating hormone (TSH) release on the other hand.⁶⁹ It has been also reported that individuals with social jetlag consume more junk food⁴⁹ and sugar-sweetened beverages (SSBs),⁴³ as well as fewer fruits and vegetables.⁴⁹ In addition, total energy intake has been reported to be higher in individuals with social jetlag than in those without social jetlag.^70,71 Circadian misalignment can contribute to reduced 24-h energy expenditure and peptide YY, as well as an elevation in ghrelin and leptin levels which may explain the link between social jetlag and obesity.⁶ Furthermore, circadian misalignment alters sleep architecture and may disturb food reward, metabolic pathways, substrate oxidation, and glucose-insulin metabolism, and support a positive energy balance.^4,72–76

(Flight) jetlag, in contrast to social jetlag which is a conflict between our internal biological clocks and our daily lives, is a temporary circadian misalignment issue that occurs when your body’s internal clock (or circadian rhythm) is still timed according to your original time zone rather than that where you have traveled to. It is commonly experienced after flying across multiple time zones. Jetlag typically resolves over several days or weeks (roughly taking 1 day per hour difference in time zone) as the circadian system synchronizes to the new time zone by exposure to Zeitgebers (time cues), primarily light. Symptoms of jetlag include gastrointestinal disturbances, alterations in mood, reduced cognitive performance, fatigue, and sleep disruption; however, the pattern of symptoms varies widely across individuals.⁷⁷

Some points warrant consideration while interpreting the findings. There was substantial heterogeneity in the analysis of BMI, PBF, WC, risk of overweight/obesity, and abdominal obesity which may be explained in the context of clinical and methodological heterogeneity among studies. First, the exposure measurement approach and calculations were not identical. For instance, social jetlag was measured via objective techniques (i.e., actigraphy) in some studies versus questionnaire-based approaches in others. Second, anthropometric measures were assessed in clinical settings in some studies and using self-reported values in others. Third, the age and gender of the study population may also play a role as a source of heterogeneity because social jetlag varies according to age and also differs between females and males.^23,78 Other demographic characteristics such as cultural background, traditions, behaviors, education level, and work status were also dissimilar among the enrolled studies.⁷⁸ Fourth, most cross-sectional studies recruited participants in different seasons of the year. As the circadian system is influenced by photoperiod and seasonality,^79–81 the prevalence of social jetlag may vary throughout the year. Likewise, there were differences in study location and latitude zone which may contribute to heterogeneity. Higher latitude zones (e.g., subarctic and midlatitude) show evidence of more photoperiod changes during the year.⁸² Similarly, study design and study quality may also play a role as a source of heterogeneity. Our subgroup analyses aimed to rule out the effects of these factors on overall results and magnitude of heterogeneity; however, these analyses were effective only for BMI among non-cross-sectional studies (I² = 42.2%, P = 0.158), and the risk of overweight/obesity in non-cross-sectional studies (I² = 62.0%, P = 0.105) and those assessing social jetlag using objective approaches (I² = 44.7%, P = 0.124). However, previous evidence discussed the issue of I² interpretation and proposed that I² may be increased in response to the elevation in the number of enrolled studies.⁸³ Another important issue is significant publication bias on the basis of the results of Egger’s and Begg’s tests for the association between social jetlag and FMI and between social jetlag and the risk of abdominal obesity. However, the trim-and-fill analysis could not impute any additional study to attenuate asymmetry in the funnel plot for the risk of abdominal obesity. Therefore, our findings seem reliable, and observed publication bias does not seem to influence overall results. In contrast, the trim-and-fill analysis revealed that one study was needed for a symmetrical funnel plot for FMI, and including this dataset changed the final finding regarding the link between social jetlag and FMI. Hence, it seems that this issue diminishes the reliability of our meta-analysis result for the relationship between social jetlag and FMI. Observed publication bias may be explained in the context of our eligibility criteria because only peer-reviewed articles in English were included. In addition, it should be noted that the study population of Brum et al. study includes shift workers, in whom ‘social jetlag’ may have a very different meaning that for non-shift workers.²⁹

The current systematic review and meta-analysis are among the first qualitative and quantitative summaries of the available literature regarding the association between social jetlag and adiposity measures. Moreover, we comprehensively report all of the available anthropometric measures including body weight, BMI, FM, FMI, PBF, WC, NC, WHR, WHtR, and also the risk of overweight/obesity and abdominal obesity. However, some limitations warrant consideration. There was evidence of significant heterogeneity between included studies which may be explained through clinical and methodological heterogeneity. In addition, observed publication bias may also diminish the reliability of our findings. Most of the enrolled studies were cross-sectional, which precludes us from providing evidence for a cause-and-effect relationship between social jetlag and obesity. These issues led to a low level of evidence based on GRADE. Besides, confounders selection was not the same between studies which may underestimate or overestimate the effect size in each study.

5 |. CONCLUSION

In conclusion, we found that social jetlag is positively associated with adiposity measures including BMI, FM, FMI, PBF, and WC, and also the risk of overweight/obesity. Moreover, our qualitative synthesis summarizes evidence of a possible link between social jetlag and other anthropometric measures. However, owing to the low certainty of the evidence, our findings should be interpreted with caution, and further well-designed longitudinal studies are needed to confirm our findings and also investigate the causality.

Supplementary Material

SUPPORTING INFORMATION

Additional supporting information can be found online in the Supporting Information section at the end of this article.

Abbreviations:

BMI

Body mass index

CCQ

Childhood chronotype questionnaire

FM

Fat mass

FMI

Fat mass index

GRADE

Grading of recommendations assessment, development, and evaluation

HPA

Hypothalamic pituitary adrenal

MCTQ

Munich ChronoType Questionnaire

MSF

Mid-point of sleep on free days

MSW

Mid-point of sleep on workdays

NC

Neck circumference

NOS

Newcastle–Ottawa scale

OR

Odds ratio

PBF

Percent body fat

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analysis

SE

Standard error

SSB

Sugar-sweetened beverage

WC

Waist circumference

WHR

Waist-to-hip ratio

WHtR

Waist-to-height ratio

DATA AVAILABILITY STATEMENT

The data supporting this study’s findings are available from the corresponding author upon reasonable request.