Joint Associations of Sleep Quality, Mediterranean Diet, and Physical Activity with Central and Visceral Adiposity in 88,343 Spanish Workers

Abstract

Background: Obesity remains a major public health challenge, with central and visceral fat distribution conferring particularly high cardiometabolic risk. Lifestyle factors, including diet, physical activity, and sleep quality, have been implicated in adiposity, yet their combined and interactive effects in working populations remain insufficiently characterized. Methods: We conducted a cross-sectional study in 88,343 Spanish employees (53,122 men, 35,221 women) attending occupational health examinations between 2021 and 2024. Obesity was assessed using four complementary indices: body mass index (BMI), waist-to-height ratio (WHtR), Clínica Universidad de Navarra–Body Adiposity Estimator (CUN-BAE), and Metabolic Score for Visceral Fat (METS-VF). Lifestyle factors included sleep quality (Pittsburgh Sleep Quality Index), Mediterranean diet adherence (MEDAS), and physical activity (IPAQ). Multivariable logistic regression models were adjusted for sociodemographic and lifestyle variables, with interaction, stratified, joint exposure, and dose–response analyses. Results: Obesity prevalence varied widely by index, ranging from 18.9% (BMI) to 55.6% (CUN-BAE). Poor sleep quality was independently associated with higher odds of obesity across all indices, particularly central obesity (WHtR OR 1.58, 95% CI 1.48–1.69), with stronger associations observed in women. Physical inactivity and non-adherence to the Mediterranean diet were robust predictors, with inactivity showing the largest effect sizes (METS-VF OR 9.92, 95% CI 8.70–11.15). Interaction analyses indicated that both Mediterranean diet adherence and regular physical activity attenuated the adverse association between poor sleep and obesity outcomes. Restricted cubic spline models revealed a progressive dose–response relationship between increasing PSQI score and central obesity. Joint exposure analyses showed nearly five-fold higher odds of central obesity among workers with concurrent poor sleep, physical inactivity, and low Mediterranean diet adherence. A graded inverse association was observed between a composite healthy lifestyle score (0–3) and obesity, with a score of 3 associated with 72–75% lower odds of BMI-obesity and WHtR-high. Conclusions: In this large occupational cohort, poor sleep quality, physical inactivity, and low Mediterranean diet adherence emerged as independent and combined determinants of general, central, and visceral obesity. Integrated workplace strategies promoting sleep hygiene, physical activity, and dietary quality—particularly among women and lower socioeconomic groups—may represent an effective approach to reducing obesity risk in working populations.

1. Introduction

Obesity and excess adiposity remain major public health challenges worldwide, increasingly affecting working-age populations and contributing substantially to cardiometabolic morbidity, reduced productivity, and rising healthcare costs. Importantly, cardiometabolic risk is not determined solely by total body mass, but also by fat distribution, with central and visceral adiposity conferring disproportionately higher metabolic risk compared with generalized obesity. Consequently, contemporary obesity research increasingly emphasizes phenotyping beyond body mass index (BMI), integrating indices that better capture body fat percentage and intra-abdominal adiposity, such as the Clínica Universidad de Navarra–Body Adiposity Estimator (CUN-BAE) and the Metabolic Score for Visceral Fat (METS-VF) [1,2,3].

Sleep has emerged as a critical yet underrecognized determinant of adiposity and metabolic health. Experimental and epidemiological evidence consistently demonstrates that insufficient or poor-quality sleep promotes positive energy balance through increased energy intake, hormonal dysregulation, circadian misalignment, and preferential abdominal fat accumulation [4,5,6]. In parallel, METS-VF has been validated as a practical surrogate of visceral adiposity and cardiometabolic risk, showing strong associations with metabolic outcomes and mortality across populations [7].

Dietary quality and physical activity constitute central pillars of obesity prevention. Adherence to the Mediterranean diet has been repeatedly associated with lower obesity prevalence and improved cardiometabolic profiles [8]. An expanding body of literature also supports bidirectional links between dietary patterns and sleep, with Mediterranean diet adherence associated with better sleep quality and duration, potentially mediated by anti-inflammatory, antioxidant, and chrono-nutritional mechanisms [9,10]. Likewise, regular physical activity is particularly effective in reducing central and visceral fat depots and may mitigate the metabolic consequences of poor sleep [11,12].

Despite this growing evidence base, important gaps remain. Most population studies rely primarily on BMI, limiting insights into central and visceral adiposity phenotypes. Moreover, relatively few investigations have simultaneously evaluated sleep quality, dietary patterns, and physical activity within the same analytical framework using validated instruments, particularly in large occupational cohorts. Evidence regarding sex-specific vulnerability and socioeconomic gradients across multiple adiposity indices also remains limited, despite their relevance for workplace health strategies.

Working populations represent a particularly important context in which to explore these relationships, as occupational factors can influence sleep, physical activity, and dietary behaviors, thereby shaping obesity risk in real-world settings. Furthermore, the combined and potentially synergistic effects of healthy lifestyle behaviors on adiposity outcomes have not been adequately quantified using composite approaches.

Therefore, leveraging a large occupational cohort of Spanish workers, the present study aimed to: (i) examine the independent associations of sleep quality, Mediterranean diet adherence, and physical activity with four complementary adiposity indices (BMI, waist-to-height ratio, CUN-BAE, and METS-VF); (ii) evaluate effect modification by sex and lifestyle behaviors; and (iii) assess dose–response relationships using a composite healthy lifestyle score. We hypothesized that poor sleep quality would be independently associated with higher odds of general, central, and visceral obesity, and that adherence to the Mediterranean diet and regular physical activity would attenuate these associations, with differential susceptibility by sex and occupational social class.

The hypothesized pathways linking sleep quality, Mediterranean diet adherence, physical activity, and visceral adiposity are summarized in Figure 1.

Poor sleep quality may promote positive energy balance through increased appetite, hormonal dysregulation, and circadian disruption, thereby favoring visceral fat accumulation. In contrast, adherence to the Mediterranean diet and regular physical activity may counteract inflammation, oxidative stress, and impaired fat oxidation, mitigating central and visceral adiposity. These lifestyle behaviors are hypothesized to interact synergistically, ultimately influencing cardiometabolic outcomes, including insulin resistance, dyslipidemia, and hypertension.

2. Materials and Methods

2.1. Study Design and Population

We performed a cross-sectional analysis within the Ibermutua occupational health cohort, a large database of actively employed adults from multiple economic sectors across Spain. For the present study, we included 88,343 workers who underwent standardized occupational health examinations between January 2021 and December 2024. The study protocol was approved by the institutional ethics committee, and all participants provided written informed consent prior to participation in accordance with the Declaration of Helsinki.

2.2. Inclusion and Exclusion Criteria

Eligible participants were men and women aged 18–69 years who attended a routine occupational health assessment during the study period. The age cut-off was chosen to reflect the active working population and to avoid distortions related to retirement age. Exclusion criteria were (Figure 2):

• Missing information on anthropometric variables (height, weight, or waist circumference).
• Absence of validated responses to lifestyle questionnaires (International Physical Activity Questionnaire [IPAQ], Mediterranean Diet Adherence Screener [MEDAS], or Pittsburgh Sleep Quality Index.
• Extreme anthropometric values suggestive of measurement error (BMI < 15 kg/m² or >60 kg/m²; waist circumference < 50 cm or >180 cm).
• Incomplete sociodemographic data (age, sex, or occupational social class).

No exclusions were applied based on pre-existing chronic diseases, as the primary aim was to reflect the obesity burden in a real-world working population. Sensitivity analyses were planned to test the robustness of results by excluding participants with extreme metabolic values.

2.3. Sociodemographic Variables

Age was recorded in years and classified into five categories: 18–29, 30–39, 40–49, 50–59, and 60–69 years. Sex was self-reported. Occupational social class was derived from the Spanish National Classification of Occupations (CNAE-11) and categorized into three groups: Class I (managers and professionals), Class II (clerical and intermediate occupations), and Class III (manual and service workers) [13].

2.4. Lifestyle Factors

Lifestyle behaviors were systematically assessed:

• Smoking status: Participants were classified as current smokers, former smokers, or never smokers. For regression models, smoking was analyzed as a dichotomous variable (current vs. non-smoker).
• Physical activity: Evaluated with the validated short version of the IPAQ. Participants reporting ≥150 min/week of moderate activity, ≥75 min/week of vigorous activity, or an equivalent combination were classified as physically active; others were considered inactive [14].
• Dietary habits: Adherence to the Mediterranean diet was assessed using the 14-item MEDAS questionnaire, validated in Spanish populations. A score ≥ 9 indicated adherence, whereas <9 indicated non-adherence [15].
• Lifestyle score: A composite score ranging from 0 to 3 was created by summing physical activity (yes = 1), Mediterranean diet adherence (yes = 1), and good sleep quality (yes = 1).

2.5. Sleep Quality

Sleep was measured using the Spanish-validated version of the Pittsburgh Sleep Quality Index (PSQI). A global score ≤ 5 defined good sleep quality, while >5 indicated poor sleep. This dichotomization has been widely applied in epidemiological research of occupational populations [16].

2.6. Anthropometric and Obesity Indices

Height, weight, and waist circumference were measured by trained staff following standardized procedures. BMI was calculated as weight (kg)/height² (m²), with obesity defined as BMI ≥ 30 kg/m². The waist-to-height ratio (WHtR) was computed as waist circumference/height, with central obesity defined as WHtR > 0.50. Body fat percentage was estimated with the Clínica Universidad de Navarra-Body Adiposity Estimator (CUN-BAE), validated in Spanish populations, using ≥25% in men and ≥35% in women as cut-offs [17]. Visceral adiposity was estimated using the Metabolic Score for Visceral Fat (METS-VF), applying previously validated sex-specific cut-off values (≥7.18 in men and ≥5.92 in women) [18].

2.7. Clinical and Biochemical Measurements

Blood samples were obtained by trained healthcare personnel during routine occupational health examinations following standardized procedures. Participants were instructed to attend after an overnight fast of at least 8–12 h. Venous blood samples were collected using standard vacuum tubes and processed in certified clinical laboratories.

Biochemical parameters, including total cholesterol, HDL cholesterol, LDL cholesterol, and glucose, were determined using automated enzymatic methods with validated analyzers according to standard laboratory protocols. Internal and external quality control procedures were applied to ensure accuracy and reliability of the measurements.

2.8. Statistical Analysis

Continuous variables are presented as mean ± standard deviation, and categorical variables as frequencies and percentages. Group comparisons were performed using Student’s t test for continuous variables and χ² test for categorical variables.

Associations between sociodemographic, lifestyle, and sleep quality variables and each obesity outcome (BMI-obesity, WHtR-high, CUN-BAE obesity, METS-VF high) were examined using multivariable logistic regression. Odds ratios (ORs) with 95% confidence intervals (CIs) were reported. Models were adjusted sequentially for (1) sociodemographic factors (age, sex, social class); (2) lifestyle factors (smoking, physical activity, Mediterranean diet); and (3) sleep quality. We tested multiplicative interaction terms (sex × sleep, physical activity × sleep, and Mediterranean diet × sleep) in multivariable logistic regression models. p-values for interaction were obtained from these product terms. Stratified analyses were conducted by sex, focusing on BMI-obesity and WHtR-high. A dose–response relationship was assessed for the lifestyle score (0–3 points) using logistic regression.

To explore potential non-linear dose–response associations, PSQI score was additionally modeled as a continuous variable using restricted cubic spline functions for central obesity (WtHR-high), with knots placed at the 10th, 50th, and 90th percentiles of its distribution. Models were adjusted for age, sex, occupational social class, smoking status, physical activity, and Mediterranean diet adherence. Additional spline models were fitted for METS-VF (continuous) and BMI-defined obesity, with results presented in Supplementary Figures S2–S4. All spline analyses were performed using logistic regression models with binary outcomes. Results are presented as predicted probabilities to ensure consistency and comparability across adiposity indices.

To evaluate socioeconomic gradients, social class was modeled as an ordinal variable (I–III), and p-trend values were obtained. Multicollinearity was checked using variance inflation factors (VIF), with values > 5 considered problematic. Sensitivity analyses were conducted, excluding extreme anthropometric values and using alternative PSQI cut-offs. Missing data were handled using complete-case analysis; the proportion of excluded cases was reported.

To improve consistency and interpretability, analyses were structured according to the type of exposure and outcome: (i) primary analyses used categorical outcomes (obesity indices defined by standard cut-offs) and categorical exposures; (ii) secondary analyses explored dose–response relationships using continuous exposure (PSQI score) modeled with restricted cubic splines; and (iii) joint exposure analyses evaluated combined lifestyle profiles. This structured approach allows comparison across analytical frameworks while maintaining clinical interpretability.

All analyses were systematically performed for the four adiposity indices (BMI, WHtR, CUN-BAE, and METS-VF), although only selected outcomes are presented in the main tables for clarity.

All statistical analyses were conducted with SPSS version 29.0 (IBM Corp., Armonk, NY, USA). Two-tailed p-values < 0.05 were considered statistically significant.

3. Results

Among 88,343 workers (60.1% men), obesity prevalence varied substantially according to the index used, ranging from 18.9% for BMI-defined obesity to 55.6% for CUN-BAE obesity. Central obesity (WHtR > 0.50) and high METS-VF increased markedly with age and lower social class. Poor sleep quality, physical inactivity, and low Mediterranean diet adherence were consistently associated with higher odds of obesity across all indices, with physical inactivity emerging as the strongest predictor. Significant interactions indicated that both physical activity and Mediterranean diet adherence attenuated the adverse association between poor sleep and obesity outcomes. Results are presented following a structured analytical framework, including (i) categorical exposure–categorical outcome models, (ii) joint exposure analyses, and (iii) continuous exposure models assessing dose–response relationships. For clarity, the main tables focus on BMI and WHtR as representative adiposity outcomes, while the full set of analyses for all indices is provided in the Supplementary Materials.

Table 1. Baseline sociodemographic, lifestyle, and clinical characteristics of the study population by sex.

	Men n = 53,122	Women n = 35,221
Variables	Mean (SD)	Mean (SD)	p-Value
Age (years)	39.8 (10.3)	39.2 (10.2)	<0.001
Height (cm)	174.0 (7.1)	161.2 (6.6)	<0.001
Weight (kg)	81.1 (13.9)	65.4 (13.2)	<0.001
Waist (cm)	87.7 (9.2)	73.9 (7.9)	<0.001
Hip (cm)	100.1 (8.4)	97.2 (9.0)	<0.001
Systolic BP (mm Hg)	124.4 (15.1)	114.3 (14.7)	<0.001
Diastolic BP (mm Hg)	75.4 (10.6)	69.6 (10.2)	<0.001
Cholesterol (mg/dL)	195.8 (38.8)	193.9 (36.7)	<0.001
HDL-c (mg/dL)	51.0 (7.0)	53.7 (7.7)	<0.001
LDL-c (mg/dL)	120.3 (37.7)	122.6 (37.3)	<0.001
Triglycerides (mg/dL)	123.8 (88.8)	88.2 (46.2)	<0.001
Glucose (mg/dL)	88.1 (12.9)	84.1 (11.6)	<0.001
Variables	n (%)	n (%)	p-Value
18–29 years	9555 (18.0)	6851 (19.4)	<0.001
30–39 years	17,423 (32.8)	11,817 (33.5)
40–49 years	15,759 (29.6)	10,270 (29.2)
50–59 years	8752 (16.5)	5382 (15.3)
60–69 years	1633 (3.1)	901 (2.6)
Social class I	2809 (5.3)	2454 (7.0)	<0.001
Social class II	9385 (17.7)	11,721 (33.3)
Social class III	40,928 (77.0)	21,046 (59.7)
Smokers	19,722 (37.1)	11,525 (32.7)	<0.001
Mediterranean diet adherence	21,876 (41.2)	17,999 (51.1)	<0.001
Physically active	24,318 (45.8)	18,294 (51.9)	<0.001
Good sleep quality (PSQI ≤ 5)	37,104 (69.8)	31,415 (89.2)	<0.001
Poor sleep quality (PSQI > 5)	16,018 (30.2)	3806 (10.8)

BP Blood Pressure. TC Total Cholesterol. HDL-c High-density lipoprotein-cholesterol. LDL-c Low-density lipoprotein-cholesterol. TG Triglycerides. SD Standard deviation.

summarizes baseline characteristics by sex. Men showed higher body weight, waist circumference, and blood pressure, whereas women had higher HDL-c levels and reported healthier lifestyle profiles, including greater Mediterranean diet adherence, higher physical activity, and better sleep quality (all p < 0.001).

Table 2. Prevalence (%) and sample size (n) of obesity according to four adiposity indices across sociodemographic and lifestyle categories.

		BMI Obesity	WHtR High	CUN BAE Obesity	METS-VF High
Men	n	%	%	%	%
18–29 years	9555	10.4	30.8	22.5	2.8
30–39 years	17,423	16.8	43.4	43.7	3.4
40–49 years	15,759	22.6	53.8	63.6	11.3
50–59 years	8752	27.2	60.0	78.3	20.6
60–69 years	1633	27.9	67.2	87.8	29.0
Social class I	2809	17.2	44.2	52.3	6.8
Social class II	9385	17.9	44.5	52.9	8.2
Social class III	40,928	19.9	48.6	53.4	12.5
Non smokers	33,400	16.8	44.3	46.4	7.8
Smokers	19,722	21.0	49.7	56.6	12.7
Mediterranean diet adherence	21,876	8.5	25.9	19.6	4.5
Non-adherence to Mediterranean diet	31,246	27.1	60.3	76.1	22.8
Physically active	24,318	7.9	24.6	19.9	4.0
Physically inactive	28,804	28.9	67.1	80.6	20.0
Good sleep quality (PSQI ≤ 5)	37,104	9.6	34.1	35.5	5.9
Poor sleep quality (PSQI > 5)	16,018	20.8	79.2	89.9	14.3
Women	n	%	%	%	%
18–29 years	6851	10.0	11.0	23.9	0.3
30–39 years	11,817	12.8	13.9	35.8	0.9
40–49 years	10,270	16.8	18.3	55.5	1.5
50–59 years	5382	21.4	22.5	77.5	2.5
60–69 years	901	26.3	26.9	90.1	3.9
Social class I	2454	9.5	10.8	32.9	1.5
Social class II	11,721	9.7	11.4	36.3	1.9
Social class III	21,046	18.8	19.7	54.6	3.5
Non smokers	23,696	11.9	14.8	40.4	1.9
Smokers	11,525	16.6	17.0	50.2	2.9
Mediterranean diet adherence	17,999	6.9	4.5	21.1	0.9
Non-adherence to Mediterranean diet	17,222	22.5	22.1	74.1	3.9
Physically active	18,294	7.3	3.9	17.9	0.7
Physically inactive	16,927	20.5	25.6	78.4	4.2
Good sleep quality (PSQI ≤ 5)	31,415	8.8	10.8	49.9	1.7
Poor sleep quality (PSQI > 5)	3806	31.5	39.9	88.6	4.0

BMI Body mass index. WHtR Waist to height ratio. CUN BAE Clinica Universitaria de Navarra Body adiposity estimator. METS-VF Metabolic score for visceral fat.

shows obesity prevalence across sociodemographic and lifestyle categories. A marked age gradient was observed for all indices, particularly for central obesity. Clear socioeconomic differences were evident, with higher prevalence in lower occupational classes. Unhealthy lifestyle behaviors were consistently associated with substantially higher obesity prevalence: non-adherence to the Mediterranean diet, physical inactivity, and poor sleep quality showed the largest contrasts, especially for WHtR. Across both sexes, physically inactive participants (men: n = 28,804; women: n = 16,927) and individuals with poor sleep quality (men: n = 16,018; women: n = 3806) exhibited the highest prevalence of central obesity.

Table 3. Multivariable logistic regression models for obesity by four indices in relation to sociodemographic and lifestyle factors.

Category	n	BMI Obesity	WHtR High	CUN BAE Obesity	METS-VF High
Men	53,122	1	1	1	1
Women	35,221	2.05 (1.95–2.16)	3.15 (2.70–3.61)	1.21 (1.16–1.27)	6.39 (5.40–7.39)
18–29 years	16,406	1	1	1	1
30–39 years	29,24	1.33 (1.27–1.39)	1.12 (1.09–1.16)	1.63 (1.48–1.79)	1.33 (1.25–1.42)
40–49 years	26,029	1.73 (1.60–1.86)	1.64 (1.50–1.79)	2.29 (1.96–2.62)	1.64 (1.45–1.84)
50–59 years	14,134	2.23 (2.02–2.44)	2.10 (1.65–2.55)	3.81 (3.25–4.38)	1.91 (1.65–2.17)
60–69 years	2534	2.88 (2.59–3.18)	2.98 (2.41–3.55)	7.44 (6.33–8.55)	3.92 (3.23–4.62)
Social class I	5263	1	1	1	1
Social class II	21,106	1.15 (1.12–1.19)	1.19 (1.15–1.24)	1.23 (1.18–1.29)	1.27 (1.21–1.33)
Social class III	61,974	1.31 (1.24–1.38)	1.38 (1.30–1.47)	1.49 (1.30–1.69)	1.59 (1.40–1.79)
Non smokers	57,096	1	1	1	1
Smokers	31,247	1.21 (1.16–1.27)	1.23 (1.18–1.29)	1.31 (1.23–1.38)	1.29 (1.20–1.39)
Physically active	42,612	1	1	1	1
Physically inactive	45,731	7.23 (6.30–8.16)	5.92 (5.62–6.23)	6.99 (5.88–8.10)	9.92 (8.70–11.15)
Good sleep quality (PSQI ≤ 5)	68,519	1	1	1	1
Poor sleep quality (PSQI > 5)	19,824	2.78 (2.50–3.07)	1.58 (1.48–1.69)	1.89 (1.70–2.08)	2.51 (2.11–2.92)

BMI Body mass index. WHtR Waist to height ratio. CUN BAE Clinica Universitaria de Navarra Body adiposity estimator. METS-VF Metabolic score for visceral fat. OR Odds ratio. CI Confidence Interval.

summarizes the multivariable-adjusted associations between sociodemographic and lifestyle factors and obesity outcomes. Female sex, older age, and lower occupational social class were consistently associated with higher odds of obesity across indices. Among lifestyle factors, non-adherence to the Mediterranean diet and especially physical inactivity showed the strongest associations, particularly for central and visceral adiposity. Poor sleep quality remained independently associated with obesity after full adjustment, with the largest effects observed for WtHR-defined central obesity.

Stratified analyses are presented in

Table 4. Stratified associations between sleep quality (PSQI) and obesity outcomes according to sex, physical activity, and Mediterranean diet adherence.

Sex × Sleep
Outcome	Group	n	Good Sleep (ref.)	Poor Sleep OR (95% CI)	p-Interaction
BMI-obesity	Men	53,122	1.00 (ref.)	1.38 (1.26–1.51)	<0.01
	Women	35,221	1.00 (ref.)	2.05 (1.85–2.27)
WHtR high	Men	53,122	1.00 (ref.)	1.44 (1.31–1.58)	<0.01
	Women	35,221	1.00 (ref.)	2.42 (2.10–2.78)
* Physical activity × Sleep
BMI-obesity	Active	42,612	1.00 (ref.)	1.22 (1.08–1.38)	<0.05
	Inactive	45,731	1.00 (ref.)	1.89 (1.71–2.09)
WHtR high	Active	42,612	1.00 (ref.)	1.35 (1.19–1.53)	<0.05
	Inactive	45,731	1.00 (ref.)	2.10 (1.89–2.34)
* Mediterranean diet × Sleep
BMI-obesity	Adherent	39,875	1.00 (ref.)	1.25 (1.10–1.42)	<0.05
	Non-adherent	48,468	1.00 (ref.)	2.10 (1.90–2.32)
WHtR high	Adherent	39,875	1.00 (ref.)	1.32 (1.15–1.51)	<0.05
	Non-adherent	48,468	1.00 (ref.)	2.25 (2.01–2.53)

* Models adjusted for age, social class, smoking, and the non-stratified lifestyle variables. Full results for CUN-BAE and METS-VF are provided in Supplementary Tables.

. The association between poor sleep quality and obesity outcomes varied across levels of sex, physical activity, and Mediterranean diet adherence. Formal tests of interaction were conducted using multiplicative interaction terms, with corresponding p-interaction values reported in the table. The adverse association between poor sleep quality and obesity was stronger in women than in men. Moreover, both regular physical activity and adherence to the Mediterranean diet attenuated the magnitude of the association between poor sleep and obesity outcomes. In contrast, the association was substantially amplified among physically inactive workers and those not adhering to a Mediterranean diet, particularly for central obesity (WHtR-high).

Joint exposure analyses combining sleep quality, physical activity, and Mediterranean diet adherence are shown in

Table 5. Joint associations of sleep quality, physical activity, and Mediterranean diet adherence with central obesity (WHtR-high).

Sleep	PA	Diet	n	OR (95% CI)
Good	Active	Mediterranean diet adherence	30,412	1.00
Good	Active	Non-adherence	25,981	1.42 (1.30–1.55)
Good	Inactive	Mediterranean diet adherence	22,115	1.88 (1.72–2.05)
Good	Inactive	Non-adherence	20,991	2.56 (2.33–2.82)
Poor	Active	Mediterranean diet adherence	9463	1.63 (1.48–1.79)
Poor	Active	Non-adherence	7168	2.21 (2.01–2.43)
Poor	Inactive	Mediterranean diet adherence	776	3.12 (2.82–3.46)
Poor	Inactive	Non-adherence	6636	4.67 (4.20–5.19)

Reference category: good sleep quality, physically active, and Mediterranean diet adherence. Models adjusted for age, sex, social class, and smoking status.

.

Joint exposure analyses across BMI, CUN-BAE, and METS-VF (Supplementary Tables S2–S4) showed consistent graded associations, with the strongest effects observed for visceral adiposity (METS-VF), supporting the robustness of the combined lifestyle model across adiposity phenotypes.

Marked combined effects were observed. Compared with participants reporting good sleep, regular physical activity, and Mediterranean diet adherence, those with concurrent poor sleep, physical inactivity, and low Mediterranean diet adherence exhibited nearly five-fold higher odds of central obesity (OR 4.67, 95% CI 4.20–5.19). Intermediate profiles showed graded increases in risk, indicating cumulative adverse effects of unhealthy lifestyle behaviors.

To explore the dose–response association between sleep quality and central obesity beyond the dichotomous cut-off, PSQI score was modeled as a continuous variable using restricted cubic splines (Figure 3).

A progressive dose–response association was observed between increasing PSQI score and the probability of central obesity (WHtR-high). The relationship was approximately linear across most of the distribution, with progressively higher odds at increasing PSQI values.

Consistent dose–response patterns were observed across all adiposity indices when PSQI was modeled as a continuous variable, and results were expressed as predicted probabilities of the corresponding obesity outcomes, including BMI-defined obesity, CUN-BAE-defined obesity, and high METS-VF (Supplementary Figures S2–S4).

Sex-stratified analyses are presented in

Table 6. Sex-stratified multivariable logistic regression models for BMI-obesity and WHtR-high.

Predictor	BMI-Obesity OR (95%CI)—Men	BMI-Obesity OR (95%CI)—Women	WHtR-High OR (95%CI)—Men	WHtR-High OR (95%CI)—Women
Age (per 10 y)	1.42 (1.35–1.49)	1.28 (1.22–1.35)	1.36 (1.29–1.44)	1.24 (1.18–1.31)
Social class (III vs. I)	1.25 (1.14–1.37)	1.42 (1.28–1.57)	1.31 (1.20–1.43)	1.48 (1.34–1.64)
Smoker (yes)	1.18 (1.09–1.28)	1.24 (1.13–1.36)	1.21 (1.12–1.31)	1.29 (1.18–1.42)
Physically inactive	2.80 (2.58–3.03)	3.12 (2.85–3.41)	2.95 (2.73–3.20)	3.28 (3.00–3.58)
Non-adherence to Mediterranean diet	3.45 (3.12–3.82)	3.76 (3.38–4.18)	3.62 (3.27–4.01)	3.89 (3.50–4.33)
Poor sleep quality (PSQI > 5)	1.29 (1.19–1.40)	1.72 (1.55–1.91)	1.33 (1.22–1.45)	1.85 (1.65–2.07)

Models adjusted sequentially: Block 1 (sociodemographic variables: age, social class); Block 2 (lifestyle variables: smoking, physical activity, Mediterranean diet); Block 3 (sleep quality: PSQI). Values shown are from fully adjusted models. BMI Body mass index. WHtR Waist to height ratio. Full results for CUN-BAE and METS-VF are provided in Supplementary Tables.

. Age, occupational social class, smoking, physical inactivity, low Mediterranean diet adherence, and poor sleep quality were consistently associated with both BMI-obesity and WHtR-high in men and women. Across all models, effect sizes were systematically larger in women, particularly for poor sleep and physical inactivity, indicating greater susceptibility. Sex-stratified analyses were additionally conducted for CUN-BAE and METS-VF using the same modeling approach. To maintain clarity in the main manuscript, these results are presented in the Supplementary Materials (Supplementary Tables S5 and S6), ensuring full comparability across adiposity indices.

Table 7. Multivariable-adjusted odds ratios (95%CI) for BMI-obesity and WHtR-high by lifestyle score (0–3).

Lifestyle Score	n	BMI-Obesity OR (95% CI)	WHtR-High OR (95% CI)
0 (ref.)	9842	1.00	1.00
1	21,376	0.62 (0.55–0.70)	0.58 (0.51–0.66)
2	34,915	0.41 (0.36–0.47)	0.38 (0.33–0.44)
3	22.21	0.28 (0.24–0.33)	0.25 (0.21–0.29)

Models adjusted for age, sex, social class, and smoking. p-trend < 0.001 for both outcomes. Full results for CUN-BAE and METS-VF are provided in Supplementary Tables.

shows a clear inverse dose–response relationship between the lifestyle score and obesity outcomes. Each additional healthy lifestyle component was associated with progressively lower odds of both BMI-obesity and WHtR-high, with participants scoring 3 exhibiting approximately 70–75% lower odds compared with those scoring 0 (p-trend < 0.001).

4. Discussion

Guided by the conceptual framework presented in Figure 1, this large occupational study demonstrates that poor sleep quality, physical inactivity, and low adherence to the Mediterranean diet are not only independent but also interacting determinants of obesity, particularly central and visceral adiposity. Among the most notable findings, poor sleep quality was associated with substantially higher odds of abdominal obesity, with stronger effects observed in women. Physical inactivity emerged as the dominant risk factor across all adiposity indices, while Mediterranean diet adherence and regular physical activity significantly attenuated the detrimental impact of poor sleep. Beyond previously reported individual associations, the main contribution of this study lies in the joint and integrative evaluation of sleep quality, physical activity, and Mediterranean diet adherence within a large occupational cohort.

These findings provide empirical support for the proposed combined effects model, indicating that lifestyle behaviors operate jointly rather than in isolation. The joint exposure analyses further showed that the coexistence of poor sleep, physical inactivity, and low Mediterranean diet adherence was associated with markedly higher odds of central obesity, suggesting cumulative adverse effects. Conversely, combining healthy behaviors conferred progressive protection against adverse adiposity profiles in working populations.

4.1. Comparison with Previous Literature

Our findings align with and extend prior evidence linking sleep disturbances to obesity and central adiposity. Large cohort studies have consistently demonstrated that poor sleep quality and reduced duration are associated with higher BMI, waist circumference, and visceral fat accumulation [19,20]. A recent meta-analysis reported that sleep restriction promotes positive energy balance and weight gain through increased appetite and altered hormonal regulation [21]. Importantly, our observation that women are more susceptible to the adverse effects of poor sleep parallels reports from both clinical and population-based studies, which suggest that hormonal, behavioral, and psychosocial factors may contribute to sex-specific vulnerabilities [22].

Physical inactivity showed the strongest associations with obesity indices in our cohort, confirming its central role in adiposity development. Evidence from prospective studies highlights that higher levels of physical activity reduce visceral adipose tissue independent of diet [23], and workplace-based interventions have shown beneficial effects on weight and metabolic risk [24]. The interaction between sleep quality and physical activity observed in our study is supported by recent reports indicating that regular exercise may buffer the metabolic consequences of inadequate sleep [25].

Dietary quality, assessed through adherence to the Mediterranean diet, was another strong determinant. Previous research has shown that the Mediterranean diet not only reduces the risk of obesity but also improves sleep quality [26]. Our results indicate that diet may attenuate the impact of poor sleep on adiposity, supporting mechanistic studies highlighting anti-inflammatory and chrono-nutritional effects of Mediterranean dietary patterns [27].

The consistency of results across four different obesity indices—including BMI, WtHR, CUN-BAE, and METS-VF—underscores the robustness of our findings. Prior studies have indicated that BMI underestimates adiposity and fails to capture central fat distribution, while indices such as WHtR and METS-VF better predict cardiometabolic risk [28,29]. Our work adds to this literature by demonstrating that lifestyle factors exert similar, yet more pronounced effects when obesity is assessed using refined measures of central and visceral adiposity. Importantly, these findings reflect combined and modifying effects rather than formally demonstrated biological synergy. To our knowledge, this is one of the largest occupational studies to simultaneously examine sleep quality, diet, and physical activity in relation to multiple adiposity phenotypes, including central and visceral obesity. Unlike most previous studies relying solely on BMI, our study incorporates multiple complementary indices (WHtR, CUN-BAE, and METS-VF), allowing a more precise characterization of central and visceral adiposity and strengthening the consistency of the findings.

In addition, tobacco consumption deserves consideration in the context of obesity. Although smoking has traditionally been associated with lower body weight due to appetite-suppressing and metabolic effects, accumulating evidence indicates a complex and paradoxical relationship. Current smokers in our cohort showed higher odds of obesity across all indices, consistent with reports linking smoking to greater central fat accumulation and adverse metabolic profiles despite lower average BMI in some populations. Recent occupational studies have highlighted that smoking may exacerbate visceral adiposity and cardiometabolic risk, possibly through chronic inflammation, oxidative stress, and unfavorable fat distribution rather than total body weight reduction [30]. These findings suggest that the apparent protective effect of smoking on body weight is misleading and reinforce the need for comprehensive workplace health interventions that address smoking cessation alongside diet, physical activity, and sleep hygiene.

4.2. Interaction Effects and Sex Differences

The interaction analyses underscore that the impact of poor sleep on obesity is not uniform across the population. Women appeared particularly susceptible, which may reflect sex-specific pathways involving hormonal regulation, circadian rhythms, and psychosocial stressors [31]. Furthermore, physical activity attenuated the association between poor sleep and both general and central obesity. These findings suggest that workplace health promotion strategies should adopt an integrated approach that simultaneously targets sleep hygiene and regular physical activity, especially among women and inactive employees. However, these interactions should be interpreted as effect modification rather than evidence of biological synergy, as additive interaction was not formally assessed. This approach goes beyond independent effects and provides evidence on the interaction and synergy between lifestyle behaviors, an aspect that remains insufficiently explored in large-scale epidemiological studies.

Socioeconomic position also emerged as a relevant determinant of obesity. We observed a clear gradient, with workers in lower occupational classes exhibiting higher odds of obesity across all indices. These findings are consistent with previous European and Spanish studies showing that lower socioeconomic status is associated with greater central adiposity and higher cardiometabolic burden, largely explained by differences in diet quality, physical activity opportunities, and psychosocial stressors [32]. Importantly, the persistence of this gradient after adjustment for lifestyle behaviors suggests that structural and contextual factors—such as work conditions, economic insecurity, and access to health-promoting resources—play a substantial role. Our results, therefore, highlight the need for equity-oriented public health strategies specifically targeting obesity prevention in lower occupational classes, where lifestyle interventions alone may not fully counterbalance social determinants.

An additional interaction between adherence to the Mediterranean diet and sleep quality revealed that diet may buffer the adverse effect of poor sleep. Among workers not adhering to a Mediterranean dietary pattern, poor sleep was strongly associated with both BMI-defined obesity and abdominal obesity, whereas among adherents the magnitude of these associations was substantially weaker. These findings reinforce previous evidence that the Mediterranean diet confers combined metabolic benefits and may help mitigate behavioral or circadian disruptions [33,34].

In this context, our study contributes to the literature by providing robust evidence from a large occupational cohort in which these factors have rarely been evaluated simultaneously.

4.3. Strengths and Limitations

The main strengths of this study include its exceptionally large sample size, the standardized collection of anthropometric and lifestyle data, and the use of multiple validated indices to capture different dimensions of obesity, including central obesity (WHtR) and visceral adiposity estimated by METS-VF. The comprehensive assessment of sleep quality, Mediterranean diet adherence, and physical activity within the same population enabled evaluation of both independent and interacting lifestyle effects. Moreover, modeling sleep quality as both a categorical and continuous exposure using restricted cubic splines provided a nuanced assessment of dose–response relationships. However, dose–response analyses using restricted cubic splines were not presented uniformly across all adiposity indices, which may limit the comparability of non-linear associations. The inclusion of workers across diverse occupational sectors further enhances the relevance of findings for the active working population.

Nevertheless, certain limitations should be acknowledged. First, the cross-sectional design precludes causal inference, and reverse causality cannot be ruled out, as obesity itself may impair sleep quality or reduce physical activity. Second, lifestyle factors were self-reported, which may introduce recall or social desirability bias, although validated instruments (PSQI, IPAQ, MEDAS) were used to mitigate this limitation. Third, residual confounding from unmeasured factors such as psychological stress or chronotype remains possible. Additionally, specific occupational factors such as shift work or occupational physical activity were not available in this cohort and could not be included in the analyses, which may contribute to residual confounding. Missing data were handled using complete-case analysis, which may introduce selection bias if excluded participants differ systematically from those included. Although sensitivity analyses using alternative PSQI cut-offs and exclusion of extreme anthropometric values yielded consistent results, the potential impact of this approach on the representativeness of the sample cannot be fully ruled out. However, given the large sample size and the consistency of associations across multiple adiposity indices, the overall impact of potential selection bias is likely limited, although it should be considered when interpreting the findings. Additionally, although interaction analyses were performed, the study did not assess additive interaction using formal measures (e.g., RERI or attributable proportion), which limits the ability to infer true biological synergy between lifestyle factors. In addition, METS-VF provides an indirect estimate of visceral adiposity and was not validated against imaging in this cohort. Finally, the study population consisted exclusively of actively employed adults, which may underestimate the true burden of obesity compared with unemployed or retired populations, but increases applicability to workplace health strategies.

4.4. Implications for Public Health and Occupational Health

These findings have important implications for both public health and occupational health policies. The workplace represents a strategic setting for implementing lifestyle interventions targeting diet, physical activity, and sleep hygiene. Given the observed combined effects, these findings support the need for integrated strategies addressing multiple lifestyle behaviors simultaneously, rather than isolated interventions. In particular, promoting sleep health should be considered a priority in occupational health strategies, as sleep remains an underrecognized determinant of obesity. Given the observed synergistic effects, these findings support the need for integrated strategies addressing multiple lifestyle behaviors simultaneously, rather than isolated interventions.

From a broader perspective, the observed socioeconomic gradients highlight the need for equity-oriented policies. Workers in lower occupational classes were disproportionately affected, suggesting that structural interventions—such as flexible work schedules, access to healthy food options, and workplace exercise programs—are essential to reduce inequalities in obesity risk.

In summary, this large occupational study shows that poor sleep quality, physical inactivity, and low adherence to the Mediterranean diet act jointly to shape central and visceral adiposity. Addressing these modifiable factors through integrated workplace health programs may represent an effective and scalable strategy to curb the growing burden of obesity and its cardiometabolic consequences, particularly among socioeconomically vulnerable workers.

5. Conclusions

This large occupational study demonstrates that poor sleep quality, physical inactivity, and low adherence to the Mediterranean diet act as independent and combined determinants of obesity, with particularly strong effects on central and visceral adiposity. Women appeared more susceptible to the adverse impact of poor sleep, while both physical activity and dietary adherence significantly attenuated these associations. The graded inverse relationship observed with the composite lifestyle score highlights the cumulative benefits of adopting multiple healthy behaviors simultaneously.

Our findings also emphasize the importance of moving beyond traditional anthropometric measures such as BMI and incorporating indices that better reflect adiposity distribution, including WHtR, CUN-BAE, and METS-VF. By capturing different dimensions of obesity, these tools provide a more comprehensive understanding of how lifestyle factors influence adverse fat accumulation.

From a public health and occupational health perspective, integrated interventions promoting sleep hygiene, regular physical activity, and Mediterranean diet adherence may substantially reduce obesity risk among workers. Tailored strategies addressing sex differences and socioeconomic disparities are especially warranted. Overall, workplace-based, multifaceted health promotion programs represent a promising and scalable approach to mitigate obesity and its cardiometabolic consequences.