Investigating the Risk of Obesity in European Night Shift Workers: A Study Protocol for Cross-Sectional and Mechanistic Studies in the SHIFT2HEALTH Project

Abstract

Introduction

Approximately one in five workers in Europe is engaged in shift work. Studies reveal that night shift work leads to an increased risk of overweight, obesity, and related diseases. Yet, the biological and behavioural mechanisms underlying these associations are not fully understood. The cross-sectional and mechanistic studies within the European SHIFT2HEALTH project aim to investigate biological, behavioural, and psychosocial key risk factors responsible for the association between night work and obesity across five European countries.

Methods

A multi-centric cross-sectional study is designed to unravel obesogenic risk factors, eating habits and sleep patterns in night shift workers and day workers from the health- and various industrial sectors. Recruitment takes place in Austria, Denmark, Germany, Poland, and the Netherlands, aiming at 500 night shift workers and 500 day workers. Anthropometric measurements, sensory perception and food preference tests are performed, alongside extensive questionnaires. In addition, biological samples (blood, hair, urine, faeces) are collected for biomarker measurements of inflammation, oxidative stress, glycaemic and lipaemic parameters, for microbiome and metabolomics analyses and chronotype assessment. In a nested mechanistic study, night shift workers (N = 200) recruited in Austria and in the Netherlands, additionally collect urine samples from all voids over 24 h during a day shift and a night shift, as well as dried blood spots and tongue swabs at four time points and undergo continuous sleep, activity, and light exposure monitoring through actigraphy. The association between night shift work and its metrics with levels of pre-obesity biomarkers will be evaluated in crude and multivariable-adjusted regression models, adjusting for potential confounders. Stratified analyses by age, gender, sector, and chronotype will be conducted.

Conclusion

In the cross-sectional and mechanistic studies of the SHIFT2HEALTH project, biological, behavioural, and psychosocial factors of night shift workers will be compared with those of day workers across sectors. The outcomes of these studies will serve as a basis for future intervention studies and, together, will contribute to the development of strategies to prevent and reduce overweight and obesity to improve the health and well-being of night shift workers.

Introduction

Shift work is an integral part of the 24/7 society. Nearly 18% of workers across Europe work in shifts, which add to 29 million people [1], and the numbers are expected to remain high or even rise in certain sectors in the future [2]. Several studies in the last decades have focused on better understanding the complex biological and behavioural determinants linked to night shift work exposure leading to negative health effects. Also, due to its relation to breast, prostate and colorectal cancer, International Agency for Research on Cancer (IARC) classified night shift work in 2019 as “probably carcinogenic to humans-Group 2A” [3, 4].

Among other factors, regular night shift work has also been linked to an increased risk of developing overweight or obesity [5], in particular abdominal obesity [6] and normal weight obesity [7] leading to type 2 diabetes, elevated blood pressure and metabolic syndrome [8–11]. However, the sex-dependent vulnerability to the obesogenic effects of shift work still needs further investigation [6]. Sleep deprivation and chronic circadian disruption account as important causes of the increased risk of obesity, besides other diseases [12, 13]. Indeed, insufficient sleep can lead to an imbalance of the appetite hormones, food intake, energy expenditure which would altogether contribute to the development of obesity [14–16]. Night shift workers were found to present higher body mass index (BMI), waist circumference, and hip circumference compared to non-shift workers [16], along with elevated levels of 28 serum metabolites highly related to lipid metabolism [17]. Additionally, individual metabolic responses with notable distinct variability were seen by metabolomics analyses in serum samples [18], further suggesting that chronobiological differences, such as chronotype, may differentially influence metabolic health in shift workers, and that a mismatch of chronotype-schedule could increase obesity risk [19].

A growing body of research also highlights the gut microbiome as a crucial factor in regulating both circadian rhythm and host metabolism [20, 21], since arrhythmicity in the gut microbiome could also be associated with risk of type 2 diabetes. Furthermore, intestinal microbiota has been shown to regulate circadian lipid metabolism and lipid absorption [22].

Night shift work affects physical health by increasing the risk of chronic conditions such as cardiovascular disease, metabolic syndrome or type 2 diabetes, with stronger effects for individuals with over 5 years of exposure to shift work [23]. These conditions could be triggered by the increased levels of various inflammatory markers [24–26]. In fact, a number of studies demonstrated the increase in circulating inflammatory markers in night shift workers, especially of circulating hs-CRP [27–29] in both male and female shift workers [24, 30, 31], besides other classical inflammatory markers such as increased IL-6, IL-1 β, or TNF-α both in plasma [32] and in saliva (IL-1 β and TNF-α only) [33]. Sleep deprivation is accounted as one important factor leading to this inflammatory state [34]. Indeed, the altered cytokine levels could be due to the effects of disrupted circadian rhythm on the pineal gland [33]. Additionally, irregular mealtimes associated with shift work have been shown to affect levels of reactive oxygen species, leading to increased oxidative damage [35–37] which in turn affects inflammation.

Given the work-related restrictions, night shift workers eat at unconventional times, including eating during the biological night, and their dietary habits may differ from day workers [38]. In particular, increased snacking with higher levels of saturated fats, was observed in afternoon and evening shifts [39]. These behaviours can contribute to worsening their metabolic health, but could also affect their gut microbiome [40]. Night eating syndrome, known to develop from unfavourable meal timing, was shown to affect mental health, particularly through depressive symptoms [41]. Therefore, novel preventive and therapeutic strategies for weight loss and management of overweight should consider the timing of meals [42] as well as internal timing. In one study, night shift workers with an evening chronotype tended to report worse eating habits and sleep quality compared to non-shift workers [43]. Also, daytime eating during simulated night work mitigated changes in cardiovascular risk factors and circadian misalignment [44, 45].

In addition, night shifts disrupt the physiological sleep rhythm, often resulting in insufficient sleep, poor sleep quality, and sleep disorders [4]. Besides physical issues due to sleep problems, night shift work has also been associated with mental health issues, including an increased risk of depression or anxiety [46, 47], which adds on the already above-mentioned negative health effects on the immune system, cardiovascular system, and metabolism.

Knowledge Gaps

EU-wide studies, across sectors with different shift schedules that use comprehensive biomarker or omics approaches and assess key factors for metabolic health, such as diet and sleep, are currently missing. Thus, an integrative approach in multiple sectors needs to be considered [48]. Indeed most studies are limited either by small study sizes, narrow focus on a single occupation or shift work schedule, or evaluation of single biomarkers. A large UK-population based study with participants from the UK Biobank confirmed the association between night shift work and obesity across multiple sectors [47], although detailed information on night shift work schedule and its exposure was not reported.

SHIFT2HEALTH Description and Aims

The cross-sectional study of the EU-funded SHIFT2HEALTH project (https://shift2health.eu/), intends to fill this gap through the analysis of key pre-obesity biomarkers, as well as behavioural and psychosocial determinants characteristic for night shift workers from 5 European countries, making also use of comprehensive analytical tools, such as metabolomics and microbiome analyses.

In addition, to supplement the missing information on acute exposure to night shift, night shift workers recruited in Austria and the Netherlands participate in a nested mechanistic study, where some of these effects are assessed. These are rarely investigated in large cohort studies.

Both the cross-sectional and the mechanistic studies include night shift workers from the female-dominated healthcare, and the more male-dominated industry sector. Comparisons across the sectors, across countries and with the control group contribute to a better understanding of the challenges different types of workers face.

Methods

This protocol outlines the design and the procedures of the cross-sectional and mechanistic studies, which have started in May 2024.

Database and Data Management

During the study, data from questionnaires, anthropometric measurements, sensory analyses, breakfast preference, and biological samples are registered in REDCap (Research Electronic Data Capture; https://project-redcap.org/), in a pseudo-anonymised form and according to EU data protection regulations. Data that cannot be directly registered on REDCap, such as data from the Leeds Food Preference Questionnaire (LFPQ) and from the BIA is saved on local servers and, as backup, on the u:cloud server of the University of Vienna, password-protected, pseudo-anonymised and in adherence to the EU regulations.

Inclusion and Exclusion Criteria

Participants have to fulfil the criteria indicated in Table 1, in order to participate in the cross-sectional and mechanistic studies. For this protocol, night shift is defined as exposure to at least 3 h of working time between 00:00 and 05:00. Participants with at least 3 years of night shift history, whether intermittent or continuous, and currently working in night shifts are included in the target group.

Table 1.

Inclusion and exclusion criteria to the cross-sectional and mechanistic studies

Night shift workers (target group)	Day shift workers (control group)
Inclusion criteria
General inclusion criteria
✓ Working in the healthcare or industrial sector
✓ Being employed or self-employed
✓ Being 21 years of age or older
✓ Working at least 24 h a week
Specific inclusion criteria
✓ At least 3 years of work in shift	✓ Not have worked in night shift (permanent or rotating) in the past 5 years
✓ Still working in night shifts	✓ No history of night shift work (permanent or rotating) for more than 5 years
✓ Working at least 4 or more night shifts in a month, at least 2 of which consecutive	​
Exclusion criteria
✓ Pregnancy
✓ Lactation period
✓ Body mass index (BMI) of 40 kg/m2 or above
✓ Current treatment of a disease e.g., cancer radio- or chemotherapy
✓ Chronic diseases if in an ongoing therapy but not after a remission (renal failure, active hepatitis, cirrhosis, myocardial infarction, chronic obstructive pulmonary disease and cancer)
✓ Immunodeficiency syndrome, any acute episode of auto-immune or auto-inflammatory diseases (e.g., type-1 diabetes, multiple sclerosis, lupus, rheumatoid arthritis) and acute episodes of atopic diseases (atopic dermatitis, asthma, type 1 allergies such as hay fever)
✓ Bariatric surgery
✓ Antibiotics in the month previous the study visit

Above, the general and specific inclusion criteria applied to the night shift workers group and the day shift workers group are listed. Below, the list of the exclusion criteria, valid for both groups.

Cross-Sectional Study

The study aim is to recruit 1,000 participants from five countries: Austria, Denmark, Germany, the Netherlands and Poland, with 200 participants each. Altogether 500 participants from the target group (night shift workers) and 500 from the control group (day workers) are included, evenly divided between healthcare sector and industry. The control group is frequency matched with the target group according to age and gender. Both subjective methods, based on validated questionnaires, and objective methods, relying on assessments, anthropometric measurements and biological analyses, are applied.

Design

Figure 1 indicates the design of the cross-sectional study. It consists of three parts: enrolment, study visit, and collection of biological samples. Some parts, such as the installation and guideline of the dietary assessment Traqq^® app and the instructions and materials for the self-collection of biological samples, can happen during the enrolment or the study visit, without affecting the outcome. These procedures are adapted locally, mainly due to logistical reasons. Likewise, the biological samples can be collected during the study visit or on a separate occasion, on a day shift or a day off, within the same week from the study visit. The most important criteria are that participants are not exposed to night shift in the 3 days before the visit or the collection of the samples, as the main outcomes are influenced by this exposure. The “Recent exposure questionnaire” needs to be filled in the 24 h before the collection of biological samples, ensuring uniformity across centres in the recording of the data. This questionnaire consists of 25 items and is designed to capture the shift schedule, the types of activities performed, the exposure to light, the diet and health around the collection of the biological samples.

Fig. 1.

Study design of the cross-sectional study of the SHIFT2HEALTH project. The study consists of 3 parts: the enrolment, the study visit and the collection of biological samples. Each part is further divided into sections, as seen in the white boxes. The grey boxes indicate parts that can be done at the enrolment or study visit without affecting the outcome.

Recruitment

In-person hospital meetings with unit heads are organised. Additionally, face-to-face meetings are arranged with companies that may be interested in joining the study, and recruitment may start after approval by the management team. Informational sessions are held at both hospitals and companies to engage with potential participants directly. Flyers containing key information about the study are also distributed during peak working hours to maximise visibility and engagement. Finally, social media campaigns, media articles, radio broadcasts are implemented to reach a broader group of both the target and the control groups.

Feasibility of this study is predicted at high rate, given the one-time participation requirement. In addition, to keep compliance high for the entire data collection period, participants receive incentives for their participation and are informed about their results at the end of the trial.

Study Visit

The study visit occurs either at least 3 days after a night shift (for night shift workers) or any day (for day workers) and consists of nine sections:

• Information regarding diseases and medication and dietary supplements

• Instructions on the self-collection of the biological samples and schedule of the collection

• Collection of hair follicles

• Anthropometric measurements and BIA

• Sensory perception tests

• LFPQ

• Breakfast preference choice

• Download and explanation of the dietary intake monitoring app (Traqq^®)

• Filling of a questionnaire on health and well-being status, socio-economic status, diet and eating habits, and sleep hygiene and psychological determinants.

Outcome Measurements

Aim

The aim of the study is to analyse biological and behavioural obesogenic changes occurring to night shift workers in comparison to the control group, based on both subjective and objective methods, in different countries and sectors.

Primary Outcome

Hs-CRP is selected a priori as the primary outcome for its central pathway role linking night shift work-related circadian misalignment to metabolic dysfunction. In this context, hs-CRP is not merely a generic inflammatory marker, but also a well-validated downstream indicator of obesity-related inflammatory load and metabolic stress.

Key Secondary Outcomes

While hs-CRP is designated as the principal primary endpoint for sample size determination and confirmatory inference, the study also assesses biologically relevant domains, including systemic inflammation, glucose-insulin homeostasis, lipid metabolism, and microbiome dysbiosis. Together, these markers form a panel of pre-obesity biomarkers, due to their link with the development of obesity. Anthropometric measurements are also considered as key secondary outcomes, linked to all other outcomes.

Further Secondary Outcomes

Further important data are generated from sensory perception tests, food preference, nutritional assessment, and the subjective data from the questionnaires.

Data Collection Methods

Collection of Biological Samples for Biomarkers Analyses

A panel of pre-obesity biomarkers, intended as key elements of inflammation, glucose and insulin homeostasis and lipid metabolism, is analysed in various biological samples. A scheme is provided in Table 2. On the morning of the collection, fastened participants provide 45 mL of blood (63 mL in the Netherlands), of which 27 mL (36 mL in the Netherlands) in tubes for the collection of serum (Vacuette^R, Greiner BIO) and 18 mL (27 mL in the Netherlands) in EDTA ones (Vacuette^R, K3E K3EDTA, Greiner BIO) through venepuncture by authorised professional staff. In addition, participants provide a faecal sample collected within 48 h from the study visit, and an urine sample from the first morning void.

Table 2.

Biomarker analysis for the cross-sectional study

Biomarker in Cross-sectional study	Specimen	Analysing centre
Markers of inflammation
hs-C-reactive protein (Hs-CRP)	Serum	Erasmus University Medical Center-Rotterdam
Interleukin-6 (IL-6)
Tumour necrosis factor α (TNF-α)
Interferon γ (IFNγ)
Interleukin-10 (IL-10)
Markers of gut permeability
Zonulin	Serum	University of Applied Sciences FH Joanneum-Graz
Intestinal FABP (fatty acid binding protein)
Lipopolysaccharide-binding protein (LBP)
Gut on a chip
Short chain fatty acids (SCFA)
Short chain fatty acids (SCFA)	Faeces
Markers of glycaemia
Insulin	Serum	Erasmus University Medical Center-Rotterdam
Glucose
Markers of lipid metabolism
Cholesterol	Serum	Erasmus University Medical Center-Rotterdam
Low-density lipoprotein cholesterol (LDL-C)
High-density lipoprotein cholesterol (HDL-C)
Triglycerides (TG)
Hormones
Cortisol	Serum	Erasmus University Medical Center-Rotterdam
Leptin
6-sulfatoxymelatonin (aMT6s)	Urine	University of Vienna
Markers of oxidative stress
8-oxo-Guo//8-OHdG (8-Oxoguanosine)	Urine	University of Vienna
DNA damage	Whole blood
Malondialdehyde (MDA)	Plasma
Protein carbonyls (PC)
Metabolomics
Unconjugated bilirubin (UCB)	Serum
Microbiome dysbiosis
Microbiota (16S amplicon sequencing)	Faeces	University of Vienna
Circadian markers
Chronotype (clock genes)	Hair follicles	Charité – Universitätsmedizin Berlin
Biobank
Biobank for backup and future analyses	Whole blood	University of Vienna
	Serum
	Plasma
	Urine

A panel of biomarkers as biological outcome of the cross-sectional study is listed. As indicated, the biomarkers can be deriving from various biological specimens and are analysed in 4 SHIFT2HEALTH centres.

Upon receiving the samples, study staff processes them immediately. Faecal and urine samples are frozen at −80°C; the latter ones after being aliquoted. Totally, 3.4 mL of whole blood is aliquoted and frozen at −80°C and the remaining 41.6 mL are centrifuged at 3,500 rpm for 10 min at 4°C to separate the plasma and serum which are then aliquoted and stored at −80°C. All aliquots are stored locally before being shipped centrally to the analysis laboratories (University of Applied Sciences FH Joanneum-Graz, Erasmus University Medical Center-Rotterdam and University of Vienna) or to the biobank (University of Vienna).

Furthermore, 15–20 hair follicles are collected from the head to determine chronotype [49]. Sampling of hair follicles is performed at least 3 days after a night shift, and within 3 h from wake up, as recent night exposure could influence the assessment of chronotype. The hair follicle samples are collected in tubes containing 1.5 mL of RNAlater and are shipped at room temperature to Charité Universitätsmedizin Berlin. Isolation of RNA and determination of gene expression from time telling genes are performed at the Charité as described in Maier et al. [49].

Anthropometry

Waist and hip circumferences are measured with a measuring tape in double determination. Before this measurement, participants are asked to deeply breathe in and exhale in a relaxed manner. The height of participants is measured through analogue stadiometers while participants stand without wearing shoes, upright. All measurements are reported in cm. Similarly, for the weight determination, participants remove their shoes and clothes except for the underwear and stand in an upright position. The weight is measured on a calibrated mechanical or an electronic approval class III scale and reported in kg.

Body composition is measured with bioelectric impedance analysis (BIA). For this, participants are required to refrain from visiting a sauna or perform intense physical activity 24 h prior measurement. All recruiting centres use the same equipment for BIA measurements (BIA Nutribox, Data Input, DE). Participants with pacemakers are excluded from this assessment, for safety reasons. Electrodes are applied on the dominant side of the body, on the hand and foot respectively. Body fat, body water, extracellular mass, muscle and body cell mass (BCM), cell quota, basal metabolic rate, and phase angle (quality of lean body mass) are calculated.

Sensory Perception Tests

Participants are required to abstain from eating, drinking anything else than water or smoking 1 h prior to the visit and are asked to not intensively wear deodorant and perfume. Sensory analyses are performed by using “Taste Strips” (Burghart, DE) and “The ODOFIN Sniffin’ Sticks Identification Test” (Burghart, DE), validated in previous studies [50–54].

The taste strips contain six medium concentrated paper strips, emulating basic tastes, impregnated with basic taste qualities, namely bitter (quinine hydrochloride dihydrate), sweet (sucrose), sour (citric acid), salty (sodium chloride), umami (monosodium glutamate), and another bitter strip (PROP: 6-n-propylthiouracil). Participants are given the strips, one after the other, in a semi-randomised order and apply those on the tongue for 5–10 seconds before discarding them. The first 3 strips are given randomly to the participants, while umami and bitter recognition follow, with the bitter PROP strip being given the last. Participants are asked to identify each taste by a choice of answers (6 possibilities = all five basic tastes and PROP) and to indicate its intensity on a scale from 0 to 10 (10 being most intense).

For the odour perception, 16 felt-tip pens filled with everyday odours are given to the participants in a randomised order and one after the other, with 30 s intervals. In order to identify the odour, a forced choice procedure is performed. After smelling the pens for 3–4 s, participants are given a card with four plausible options to choose from, one of them being the correct answer. Participants are then asked to identify each odour and to indicate their liking on a scale from 1 to 9 (9 being most liking).

Food Preference

Food preference is assessed through the LFPQ and the breakfast preference. The order of these tests is used as an outcome itself.

Leeds Food Preference Questionnaire (LFPQ)

Participants are tested for the liking and wanting of food preference and food reward through the behavioural computerised leeds food preference questionnaire (LFPQ; University of Leeds) [55]. During this test, food pictures from validated databases of the countries studied are displayed. They contain high or low fat and are similar in familiarity, protein content, sweet or non-sweet/savoury taste, and palatability. Participants are asked to choose which food they prefer at that moment of the test, how much they like its taste and how much they would like to eat it at the moment of the test. The food preferences are automatically scored for their composition in high or low fat, sweet, and savoury types.

Breakfast Preference

Participants are also asked about their breakfast preference from a menu card, indicating a savoury (ham, cheese, spreads) or a sweet (jam, honey, chocolate spread) option with or without butter (fatty preference). The main outcome for this test is the conscious choice of the participants and not the breakfast consumption itself.

Traqq^® App

Participants record their dietary intake through the ecological momentary dietary assessment app “Traqq^®” [56, 57]. Traqq^® allows detailed assessment of consumed foods, energy, and nutrients on different types of days, including day shifts, night shifts (for the night shift workers), and days off. Night shift workers are invited to record their food intake throughout two night shifts, two day shifts, and two days off work whereas their day working controls record their food intake during two day shifts and one day off. For night shift workers, the reporting windows are tailored to their shift schedules to ensure capturing the different types of days. Each country uses a tailored version of Traqq^®, in the local language and linked to the national food composition database.

Questionnaires

A combined questionnaire, referred to as “the baseline questionnaire,” has been designed to capture the sociodemographic, lifestyle, occupation and shift schedules, sleep behaviour, diet, health and well-being using validated questionnaires. When needed, validated questionnaires were complimented by additional items, for a total of 150 questions. To create standardised questionnaires in multiple languages, a validated version; forward and backward translations and pre-testing were performed, in agreement with the WHO guidelines [58]. The baseline questionnaire can be completed before the day of the study visit +/−3 days and comprehends eight main sections:

• Sociodemographic background and environment [59–62] (EuropeanSocialSurvey, I. Family study, European Union Statistics on Income and Living Conditions)

• Shift schedules, including start and end-times, different type of shifts and work activities for the analysis of frequency of shifts and shift duration (not validated questions)

• Sleeping behaviours [63–66] (µMCTQ, Insomnia Severity Index, Shift Work Disorder Questionnaire, The Epworth Sleepiness Scale)

• Diet [67–70] (HEAftWE: Healthy eating assessment for the Workplace Environment, Control of Eating Questionnaire, Food choice motives Control of Eating Questionnaire, Chrononutrition Questionnaire Profile Questionnaire)

• Mental health and quality of life [71–74] (Work-SoC, PSS-4, social support, SF-12)

• Smoking, alcohol use [75], and physical activity [76] (EHIS)

• Diseases

• Literacy and interests in ways to change lifestyle (Nutritional Health Literacy Scale NHLS-EU-Q12, two not validated questions)

Nested Mechanistic Study

The aim of the mechanistic study is to deep-phenotype a subset of night workers (N = 200) recruited in 2 countries (AT, NL), and characterise acute changes occurring during the night shifts by collecting additional biological samples on a night shift (in addition to a day shift). The study integrates repeated urine and blood samples around the clock to characterise circadian disruption and dynamics of metabolic dysregulation as well as objective measures of sleep and light to continuously monitor sleep-activity patterns and light exposure with the use of actigraphy over at least 7 days including the sampling days. Figure 2 summarises the design of the mechanistic study, while Table 3 indicates more in detail the analytes measured for each sample type.

Fig. 2.

Study design of the mechanistic study. 200 night shift workers recruited in Austria and in the Netherlands provide additional samples at 4 time points: before and after a day shift and before and after a night shift. Also, they wear the MotionWatch for a period including 2 day shifts, 2 night shifts and 2 days off.

Table 3.

Biomarker analysis for the nested mechanistic study

Biomarker in mechanistic study	Specimen	Analysing centre
6-sulfatoxymelatonin	Urine	University of Vienna
Creatinine
Oral microbiome	Tongue swab
Metabolomics	Dried blood spot

24-h Urine Samples

The main outcome of the mechanistic study is rhythmicity of 6-sulfatoxymelatonin (aMT6s) production (total levels, peak time), main melatonin metabolite, measured in urine samples from all natural voids across 24 h on a night shift compared to a day shift. Participants are asked to collect time-stamped samples from all natural urine voids during a span of 24 h covering a day shift and another 24 h period covering a night shift. Concentration of aMT6s in urine samples is determined using a radioimmunoassay and will be corrected for creatinine to account for dilution variability of the samples.

Dried Blood Samples and Tongue Swabs

Secondary outcomes include metabolomics in dried blood spots and oral microbiome analyses in tongue swabs. Participants collect finger blood drops as Dried Blood Spots (DBS) on Whatman^® protein saver cards (Sigma Aldrich) and tongue swabs (COPAN eSWAB 480CE, IT) repeatedly at four time points (before and after a day shift and a night shift) for metabolomics and microbiome analyses respectively. Participants are given safety lancets (Sarsted) to prick their fingers to provide at least one drop on each of the cards. Once collected, the cards are air dried for maximum 4 h and then kept at −80°C with desiccants (Sigma Aldrich) to protect from O₂ scavenging. Tongue swabs are stored at −80°C until analysed.

Actigraphy

Sleep-wake patterns, light exposure and physical activity are tracked with the use of MotionWatch8 (CamNtech). Depending on their shift schedules, participants wear the MotionWatch8 on their non-dominant hand for continuous recording over a period of at least 7 days including 2 day shifts, 2 night shifts, and 2 days off.

Biobank

Part of the biological samples is collected for a centralised biobank. The central biobank is organised at University of Vienna (Coordinator SHIFT2HEALTH). Samples collected at University of Copenhagen are stored at a local CUBE biobank due to ethical constraints, while the sample leftovers are destroyed after the protocol approved analyses are performed. In the Netherlands, additional samples of serum, plasma, and urine are stored in the local WUR biobank for possible future analyses.

The biobanked samples are stored for a period 10 years in the central biobank, 15 years at the local CUBE in Denmark and 15 years at the WUR biobank in the Netherlands. All samples are kept at −80°C and backed up by emergency power supply.

Statistical Analyses

Sample Size Calculation

To calculate the power of the cross-sectional study, hs-CRP was chosen as example outcome. Merged means of hs-CRP (1.36 mg/L) from shift workers and subjects with overweight were calculated and for the control group a merged mean (hs-CRP) of 0.83 mg/L in non-shift worker and persons with normal weight was used as previously described [30, 31]. A mean expected population SD of 1.61 mg/L, an alpha of 0.0005 (0.005/100 variables) and a power (1-β) of 90% was assumed resulting in two equal groups of 419 (=838) subjects. Accounting for an anticipated dropout rate of about 15%, 1000 participants for the cross-section trial are planned to be recruited in total in all participating countries (200 study participants per country).

Cross-Sectional Study

Primary and key secondary outcome data will be checked for normality, and log-transformation will be applied to skewed distributions. Outliers will be investigated and documented. Values above and below the limit of detection will be imputed (e.g., by the limit of detection/2) and submitted to sensitivity analyses. Batch correction methods will be applied for robustness. Biomarkers levels will be compared visually using box-plots categorised by shift type, country, sex, time of blood sampling, BMI, and age. Linear mixed regression models will be used to assess the different biomarkers as continuous outcomes and shift type (night vs. day) and key shift work metrics such as night shift frequency (nights/month) and duration (years in night shift work) as the exposure. In analyses targeting a specific subset of biomarkers, either Tukey’s test (for pairwise comparisons) or Bonferroni adjustment will be used.

Exploratory analysis with unconstrained PCA of microbiota data will be organised in ordination plot, depending on shift type and time point of collection. Similarly, exploratory untargeted metabolomics data will be classified. FDR will be used for all omics data. Adjusted regression per microbe/metabolite and FDR control will be considered.

Secondary analyses will explore modelling using categorical biomarker levels (e.g., tertiles, quintiles). Basic models adjusting for age, sex, country, and adjusted models for different confounders (to be defined according to the outcome) will be run. Confounder selection will be based on a DAG for each specific set of exposure-outcome combination. The final decisions on confounder selection would require a DAG for each specific set of exposure-outcome combination. Effect modification analyses will be conducted according to age, sex, and chronotype. Mediation analyses will be conducted to quantify the different mechanistic paths (e.g., dietary patterns, sleep disruption, circadian disruption) underlying the obesogenicity of night work.

Behavioural dietary outcomes (e.g., LFPQ liking and wanting domain scores and breakfast choices) will be summarised analogously and visualised across key strata (shift type, country, sex, BMI, and age). Between-group differences will be estimated using generalised linear mixed models with a random intercept for country. Shift type (day vs. night) will be the primary exposure; night-shift frequency and duration (years of night work) will be modelled continuously to assess dose-response. Minimally adjusted (age, sex, country) and extended models including prespecified sociodemographic and lifestyle covariates (e.g., education/SES, smoking, physical activity; BMI/WHR as appropriate) will be reported.

Descriptive statistics (mean, SD, median, IQR, and frequencies) will be conducted to assess the sociodemographic, lifestyle, dietary, sleep, anthropometric, and other characteristics of the study population and stratified by night vs. day workers. Logistical and contextual differences in data completeness across centres, countries and work sectors will be adjusted for country, or use random intercept for country, and conduct secondary grouped analyses by country.

Healthy worker effect will be addressed by adjusting for age, sex, and length of occupation. Detailed lifetime occupational histories will be obtained to characterise cumulative shift work exposure also in day workers, prior schedule changes, thereby reducing exposure misclassification and reverse causation. Day workers will be stratified into never-exposed and formerly exposed groups to minimise dilution of exposure contrasts. Duration of shift work will be modelled taking into account age to evaluate potential depletion of susceptible subjects among long-term workers.

Differences between participants with complete and incomplete data will be evaluated to assess the plausibility of missing completely at random, missing at random (MAR), or missing not at random mechanisms. For primary analyses, we will use a missing category for missing values in confounder variables or multiple imputation by chained equations under a MAR assumption. Imputation models will include all covariates in the analytic model, as well as auxiliary variables associated with missingness to strengthen the plausibility of the MAR assumption. Biomarkers with skewed distributions will be appropriately transformed prior to imputation. Sensitivity analyses will include complete-case analysis.

Nested Mechanistic Study

Cosinor analysis will be used to evaluate the aMT6s 24-h production for each subject and 24 h period (night shift and day shift). Geometric means and 95% confidence intervals for each derived aMT6s cosinor parameter (mesor, amplitude, and acrophase) will be generated for each shift work period. Linear mixed models will be used to examine the intraindividual associations between night work and log-transformed 6-sulfatoxymelatonin cosinor parameters, adjusting for confounders.

Biomarker levels (metabolomics and oral microbiome) and sleep parameters measured repeatedly in the nested mechanistic study will be presented descriptively across time points (e.g., before and after a day and before and after a night shift) using box-plots and summary statistics e.g., mean (SD) or median (IQR). Differences between time points (e.g., same clock time, or before vs. after a shift) will be tested using paired t test and repeated-measures ANOVA. Subsequent mixed models will investigate within-person differences in the levels between shifts, adjusting for potential time-varying confounders.

Microbiome data will be filtered for abundance and then transformed. Subsequently, multivariate analyses will be done, as unconstrained PCA and redundancy analysis.

Discussion

Shift work has a significant impact on health. Despite the growing number of studies on the biological and psychosocial connection between night shift work and risk of developing obesity, a clearer understanding and a comprehensive view of the mechanisms leading to this is needed. This includes a more precise biological classification of inflammatory markers, hormones, blood parameters, and metabolites involved in the association between night shift and obesity, as well as deeper insights into lifestyle factors such as sleep hygiene, dietary patterns, physical activity, and overall well-being. Furthermore, new results are intended to highlight the link between optimal sleep and circadian health when looking into metabolic health and body weight regulation [14], but also timing of food intake [77]. The innovative aspect of the SHIFT2HEALTH cross-sectional and mechanistic studies relies on the comprehensive analysis of the effect of night shift in workers from both the healthcare and the industry sectors, based in first instance on the analysis of a broad panel of inflammatory and oxidative stress markers, hormones, blood parameters implemented with valuable- and rarely seen in big cohorts- data from metabolomics and microbiome analyses as objective measurements, but also integrating those data with subjective outcomes, based on questionnaires. Furthermore, mechanistic insights analysing acute responses to night shift exposure remain limited in human studies, which often tend to capture only the more chronic effects, such as the increased risk of obesity, observed in cross-sectional studies. This information is valuable, as detailed information about shift schedule is often missing in larger cohorts and aims to advance mechanistic research on circadian disruption and chronic disease risk.

This is the largest Europe-wide study to investigate the obesogenic risk and other health effects of shift work across multiple dimensions. The various results predicted from the cross-sectional and mechanistic studies of the SHIFT2HEALTH project serve as a valuable basis for designing intervention studies, including the planned specific product and lifestyle intervention studies within SHIFT2HEALTH.

Given the ambitious scope of these studies and the scale of recruitment required, several challenges have to be acknowledged. The nature of night shift work poses a particular obstacle, as it affects availability of workers to receive information about the project and to attend study visits, potentially disrupting their leisure time or sleep routine. Similarly, the strict inclusion criteria for day shift workers, specifically regarding their history of night shift work, further challenge the recruitment. However, these stringent criteria ensure that the control group remains unexposed to recent chronic night shift work, thereby enabling a higher-quality comparison between the exposed target group and the unexposed control group. These inclusion criteria also guarantee the participation of the target group with a health status similar to the control group, preventing selection effects.

Although the studies are designed with a comprehensive approach, some limitations have to be taken into consideration. The recruitment occurring in different countries in Europe during different seasonal periods might affect results, as participants could be exposed differently to day light. Also, involving different recruiting centres would inevitably lead to some small differences in the recruiting process and on the collection of samples. To reduce this potential impact, stringent SOPs were developed to ensure consistency in all procedures across all centres.

Conclusion

The findings will inform evidence-based recommendations and targeted interventions aimed at reducing the risk of obesity and overweight among night shift workers in Europe, thereby improving their quality of life, health, and overall well-being. This project is part of OBEClust, the European Cluster of Obesity Research Projects, which aims to reduce obesity risk in Europe. Ultimately, SHIFT2HEALTH, aims to influence policy and workplace practices across Europe to create healthier and more sustainable working conditions for shift workers. This is expected to lead to improved evidence-based guidelines, lifestyle advice, and policies for shift workers and recommendations for employers in the future.

Acknowledgments

The authors would like to thank Kilian Gandolf for designing and programming the questionnaires on REDCap and all participants who volunteer in these studies. Further acknowledgments go to all study nurses at the Medical University of Vienna for supporting the recruitment of participants in Vienna. University of Copenhagen acknowledges Sofie Skov Frost and the kitchen staff from the Nutrition and Health section. University of Bremen acknowledges the support of BIPS BioBank at the Leibniz-Institute for Prevention Research and Epidemiology – BIPS. Wageningen University acknowledges Hanne de Jong, Michelle van Alst, Verena Hasenegger, and Anneli Rost for their assistance in adapting Traqq^® for use in the different countries in SHIFT2HEALTH.

Statement of Ethics

This study enrols participants under ethical principles in accordance with the Declaration of Helsinki [78]. This study protocol was reviewed and approved by ethics committees at each of the participating sites. This full list of participating site and ethics committees can be found at online supplementary Table S1 (for all online suppl. material, see https://doi.org/10.1159/000552231). Furthermore, the study has been registered on clinicaltrials.gov with the registration ID: NCT06288568 (https://clinicaltrials.gov/study/NCT06288568). At each recruiting centre, potential participants, who show interest in the study, are asked to complete a screening questionnaire to assess eligibility. All eligible participants sign the informed consent before being officially enrolled in the study. Subsequently, participants are given a unique anonymised ID number. For the safety of participants, any abnormal results observed during the study, are promptly reported to the participant to seek professional medical support, except for those participants, that opted out from receiving this information when filling out the informed consent (Denmark, Germany). In case of ambiguities, the study staff is available to provide clarification. Participation can be withdrawn at any time, without giving any explanation.

Conflict of Interest Statement

The authors have no conflicts of interest to declare.

Funding Sources

The SHIFT2HEALTH Project is funded by the EU Commission under the HORIZON-HLTH-2022-STAYHLTH-01-05-two-stage – Prevention of obesity throughout the life course program, grant agreement ID: 101080788.

Author Contributions

K.H.W., V.S., K.S., C.C., M.U., M.V., S.H., H.J., D.A.L., E.J.M.F., M.H.V., E.W., M.W., M.Wal., K.B.O., J.R., M.A.W., B.M., M.Ri., M.R., H.L.H., A.K., I.M., K.P., H.B.: conceptualisation. K.H.W., V.S., I.V., K.S., C.C, S.H., M.H.V., K.B.O., J.R., M.A.W., H.L..H., K.P.: writing-original draft. K.H.W., V.S., I.V., K.S., C.C, D.A.L., M.H.V., S.H., E.W., M.W., M.Wal., K.B.O., J.R., M.A.W., H.L.H., K.P.: writing-review and editing. K.H.W., V.S., I.V., H.F., H.J., D.A.L., E.Z., M.U., R.C., G.J., T.H., M. Wo, M.Wal., I.M., K.S., C.C., M.U., K.B.O., J.R. M.A.W.,P.P., M.Ri., M.R., B.F.N., C.H., A.L.A., C.D., B.M. H.L.H., K.P., and A.O.: investigation. K.H.W., K.S., and M.Wal.: project administration. All authors approved the manuscript for publication.

Funding Statement

The SHIFT2HEALTH Project is funded by the EU Commission under the HORIZON-HLTH-2022-STAYHLTH-01-05-two-stage – Prevention of obesity throughout the life course program, grant agreement ID: 101080788.

Data Availability Statement

Metadata from the described studies will be openly available on Zenodo (https://zenodo.org/). The data supporting these findings will not be publicly available due to ethical considerations of participants confidentiality and will be made available upon reasonable request to the corresponding author provided the data transfer requested is allowed under the GDPR.