Chrononutrition and risk of type 2 diabetes: a prospective cohort study

Chrononutrition and risk of type 2 diabetes: a prospective cohort study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Chrononutrition and risk of type 2 diabetes: a prospective cohort study Camille Lassale, Elisa Gallo, Catalina Ramírez-Contreras, Joana Llauradó-Pont, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9107412/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract The timing of food intake interacts with circadian biology and may influence metabolic disease risk, yet correlations among chrononutritional behaviours complicate their joint evaluation. In a Spanish prospective cohort of 6,858 adults aged 40-65y, we examined associations between chrono-nutritional behaviors and incident type 2 diabetes (T2D) using regression models and mixture-based methods to assess both individual and joint effects. During a 6-year follow-up, 152 participants developed T2D. In single-exposure analyses, a later first meal (OR = 0.82 per hour; 95%CI 0.71–0.95) and longer nighttime fasting (OR = 0.78 per hour; 0.68–0.90) were associated with lower T2D risk, whereas a later last meal was associated with higher risk (OR = 1.34 per hour; 1.07–1.69). When modeled jointly, a one interquartile range increase across chrono-nutritional variables was associated with a significant lower T2D risk, with nighttime fasting contributing most strongly to the overall association. These findings highlight nighttime fasting as a potentially modifiable chrono-nutritional target for T2D prevention. Health sciences/Endocrinology/Endocrine system and metabolic diseases/Diabetes/Type 2 diabetes Health sciences/Medical research/Epidemiology Health sciences/Health care/Disease prevention/Lifestyle modification Chrono-nutrition meal timing circadian rhythm diabetes obesity cardiovascular disease epidemiology Figures Figure 1 1. Introduction Type 2 diabetes (T2D) is one of the most pressing non-communicable diseases worldwide: according to the International Diabetes Federation's (IDF) 2025 Atlas, an estimated 589 million adults (11.1%) worldwide are living with diabetes, a figure that is expected to rise to 853 million by 2050, equivalent to almost one in eight adults 1 . Identifying modifiable and actionable risk factors is key to prevention and reduce T2D burden. An emerging but under-researched area is chrononutrition, the study of how the timing of food intake interacts with circadian biology and influences metabolic health. Chrononutrition includes behaviors such as timing of meals, number of eating occasions, and nighttime fasting, each of which is regulated by endogenous circadian rhythms and exogenous lifestyle and social influences 2,3 . There are multiple potential mechanisms linking chrononutrition to metabolic dysfunction. Central and peripheral circadian clocks regulate glucose homeostasis, insulin secretion, lipid metabolism and energy expenditure 2 . Disruption of these clocks, whether through late-night eating, irregular mealtimes or shortened fasting intervals, can lead to internal misalignment, resulting in impaired glucose regulation, decreased insulin sensitivity in adipose tissue and dysregulation of appetite-related hormones such as ghrelin and leptin 3,4 . Late-night food intake may also blunt the diurnal amplitude of cortisol and melatonin, further impairing metabolic control 5 . These pathways suggest that timing of eating behaviors may act as an active driver of metabolic dysfunction when it is not in sync with endogenous biological rhythms. While individual components of chrononutritional behaviors have been associated with metabolic outcomes, a key limitation is that most existing studies isolate individual timing variables, neglecting the possibility that synergistic patterns of chrononutritional behaviors may exert a more substantial influence on disease risk 6,7 . This represents a critical gap in the current evidence base. Recent cohort studies suggest that late eating, skipping breakfast and irregular meal frequency are associated with unfavorable glucose metabolism and an increased risk of T2D 8–10 . For instance, when following 100,000 French adults from the NutriNet-Santé cohort over 7 years, Palomar-Cros et al. 11 found that participants who ate their first meal after 9 am had a 59% increased risk of incident T2D compared to those who ate before 8 am, whereas extended nighttime fasting (> 13 hours) provided a metabolic benefit only when combined with early breakfast intake. These results highlight the interaction between multiple eating behaviors, not just their individual effects. Different chrononutrition variables, such as the timing of the first and last meal, nighttime fasting duration, or the interval between the last meal and bedtime, are often highly interrelated, leading to multicollinearity when analyzed jointly, or complicating their interpretation when analyzed separately 12 . To address this complexity, cluster analysis has been used as a dimension-reduction technique to group individuals with similar chrononutritional behaviors and to capture overall eating-timing patterns. Our group 13 previously applied this approach in the GCAT study, identifying distinct sex-specific chrononutritional patterns in adults from Catalonia and related them to body mass index (BMI), while Santonja et al. 12 reported different clusters in Austrian surveys associated with obesity, depression, and poor self-rated health. However, neither study focused on T2D as an outcome, and cluster analysis primarily served a descriptive purpose, limiting its ability to quantify the associations of chrononutrition variables with the outcome risk. Recognizing these gaps, the present study explores alternative analytical approaches to better capture the multidimensional structure of chrononutritional behaviors and their potential links to T2D risk. 2. Methods 2.1 Participants and study design The Genomes for Life (GCAT) study 31 is a population-based, prospective cohort study including over 19,000 adult volunteers in Catalonia region (Spain), designed to investigate the role of environmental and genetic factors in the development of chronic diseases. The full description of the population is reported in the Extended Data. Briefly, inclusion criteria for participation in the study included being between 40 and 65 years of age, understanding at least one of the two official languages of the region, possession of an individual Health System Identification Card, and current residence in Catalonia. Exclusion criteria included mental or health impairment disorders to give written informed consent or efficient communication or planning to leave Catalonia in the next five years. The baseline visit was in 2014–2017, when information on demographic, socioeconomic data, biological and clinical data were collected. A follow-up was conducted in 2018 and included a questionnaire on lifestyle, the Mental Health Inventory score (MHI5) 32 , night shift work history, circadian habits and the food-frequency questionnaire (FFQ). The GCAT cohort study was approved by a local Ethics committee (PI13-020) and written informed consent was obtained from participants. The data was extracted in an anonymized format so that participants could not be identified. The GCAT cohort collaborates with the Catalan Health Department to link electronic health records (EHRs) through the “Programa d'analítica de dades per a la recerca i la innovació en salut” or PADRIS system. Sample selection and flow chart From then original cohort, we included in this study only those who completed the Follow-up 1 questionnaire in 2018 and a semiquantitative FFQ (n = 8079). The FFQ 33 , specifically developed and validated for the Spanish adult population, includes over 128 items covering the most commonly eaten food groups in Spain and refers to habitual intake during the previous 12 months. Subjects with missing or implausible chrononutrition values (see 2.3 below), such as first meal happening before the wake-up time or last meal happening after bedtime were excluded (n = 921). Participants who engaged in night shift work were also excluded, as such schedules can disrupt circadian rhythms (including eating patterns, physical activity, and sleep) and lead to chrono-disruption (n = 55) 34 . In addition, subjects with a diagnosis of diabetes prior to 2018 were excluded in order to model incidence of T2D. Finally, 24 participants with missing data in covariates were removed from the analysis. This resulted in a final sample of 6858 subjects (Extended Data Figure 1). Since the chrononutrition information was collected for the first time at Follow-up 1 in 2018, this constitutes the baseline of the present analysis. 2.2 Outcome variables Type 2 Diabetes Mellitus The diagnosis of T2D was extracted from the electronic health records (EHR) and defined according to the harmonized ICD-10 diagnosis code E11.XX. Subjects with a diagnosis of T2D prior to 2018 (prevalent cases) were excluded. Incident cases and diagnosis date were identified between 01/01/2018 and 31/12/2024. 2.3 Exposure variables Chrononutrition variables Concomitant to the FFQ (2018 follow-up questionnaire), additional questions were asked about the usual times of main eating occasions, spanning 5 eating occasions (breakfast, lunch, afternoon snack, dinner, post-dinner snack), separately for the week/working days and the weekend/free days (see Extended Data for exact questions). A weighted average was calculated for each chrononutrition variable listed below (X) as (Xweekday*5 + Xweekend*2)/7, except for eating jet-lag. The following variables were derived: i. Time of first meal in hours and minutes ii. Time of last meal in hours and minutes iii. Nighttime fasting duration: time in hours between the last meal of a day and the first meal of the following day. iv. Wakeup-to-firstmeal: difference in hours between the time of awakening and the time of the first meal. v. Lastmeal-to-bedtime: difference in hours between the time of the last meal and bedtime. vi. Number of eating occasions: number of meals reported by each participant, ranging from 1 to 5 possible eating occasions. vii. Eating jet lag: difference in hours between the eating midpoint on weekend and the eating midpoint on weekday, where eating midpoint is defined as the midpoint between the first and the last meal of the day. Analyses conducted as continuous variable were using the absolute value of the eating jet lag. 2.4 Covariates The baseline questionnaire included information on age (years), sex (men/women), education (primary/secondary/higher), and was updated with information from the 2018 follow-up on smoking (smoker, former smoker and never smoker), physical activity (METs/week) and BMI (kg/m 2 ). From the FFQ data, food groups and nutrient intake was derived, and we calculated a score of adherence to the Mediterranean diet (rMED) 35 . The total rMED score ranges from 0 to 18 with higher scores reflecting greater adherence to the Mediterranean diet. We also derived the share of the diet that come from ultra-processed food (UPF). All food items were classified into four groups according to the Nova classification system: 1) unprocessed or minimally processed foods, 2) processed culinary ingredients, 3) processed foods, and 4) UPF 36 . We calculated the proportional (%) of UPF consumed by each participant relative to the total weight of food and beverage (grams/day). The sleep timing questions referred to bedtime and wake-up time on weekdays and weekends. The weighted average was used to calculate sleeping midpoint as the midpoint between bedtime and wake-up time. We identified prevalent cardiovascular diseases or risk factor (CVD), diagnosed previous to T2D onset, from the EHR using ICD-10 codes, including I20" (angina), "I21" (myocardial infarction), "I10" (hypertension), "E780" (hypercholesterolemia), "I22", "I23", "I24", "I25", (ischemic heart disease), "I60", "I61", "I62", "I63", "I64", "I65", "I66", "I67", "I68", "I69", (cerebrovascular), "I50" (heart failure). A full characterization of the covariates is provided in the Extended Data. 2.5 Statistical Analyses Descriptive characteristics of the study population are presented as median and IQR for continuous variables and proportions for categorical variables. The description was performed overall, according to T2D, and separately for men and women, and we report the p-value for chi-square or Kruskal-Wallis Rank Sum test. The explanatory variables (time of first meal, time of last meal, number of eating occasions, nighttime fasting, wakeup-to-firstmeal, lastmeal-to-bedtime, and eating jetlag) were modelled as continuous variables. We explored their correlation by using Spearman’s rank correlation and plotted a heatmap. Logistic regression and non-linearity We fitted logistic regression models to assess the association between chrononutrition variables and T2D, and reported odds ratios (ORs) with 95% confidence intervals (CIs) for one meal increase in the number of eating occasions and for one hour increase in all the other chrononutrition variables. We first examined potential non-linear associations between chrononutrition variables and T2D risk. For each chrononutrition variable, we fitted a base logistic regression model adjusted for age and sex. To evaluate whether associations deviated from linearity, we compared two models for each chrononutrition variable: one in which the exposure was modeled linearly and one in which it was modeled using natural cubic splines with three degrees of freedom. Model fit was compared using likelihood ratio tests (ANOVA) to assess whether the spline-based model provided a significantly better fit than the linear model. Since no clear deviation from linearity was observed, we modelled chrononutrition variables as linear terms. A series of logistic models with increasing levels of adjustment was fitted to assess the robustness of the associations: Model 0: age, sex Model 1: Model 0 + education, smoking habits, physical activity, rMED, energy intake and UPF intake; Model 2: Model 1 + sleeping midpoint; Model 3: Model 2 + BMI; Model 4: Model 3 + presence of CVD. Since BMI and the presence of CVD might act as mediators, we separately included them in the last two models. Moreover, we applied two complementary modeling approaches. First, we conducted a single-exposure model, where each chrononutrition variable was analyzed separately, second, we conducted mutually adjusted models, where multiple chrononutrition variables entered simultaneously in the same model. In both cases, covariates were added stepwise following Models 0–4 above. This strategy allows single-exposure results for each chrononutrition variable, and mutually adjusted results showing if associations vary when multiple meal metrics are jointly modeled and when covariates are added stepwise. To assess potential multicollinearity among predictors, we computed the generalized variance inflation factor (GVIF), which extends the standard VIF to handle categorical variables with multiple degrees of freedom. All models showed GVIF values < 5, indicating no concerning collinearity. In mutually adjusted analyses, time of first meal, time of last meal, and number of eating occasions were modelled jointly; Nighttime fasting was adjusted for number of eating occasions and eating midpoint to account for the timing of the fasting period; For the remaining chrononutrition exposures, each model was further adjusted for number of eating occasions. To explore potential effect modifications, we tested interactions between nighttime fasting duration and both energy and UPF intake using likelihood ratio tests. We also conducted stratified analyses to further explore heterogeneity in associations. Specifically, we estimated associations separately for participants who broke their nighttime fast at or before 8:30 AM and those who broke it after 8:30 AM. In addition, we performed sex-stratified analyses to assess potential differences in the association between chrononutrition variables and T2D. In all stratified analyses, given the limited number of incidents T2D events, we only computed base models to avoid overfitting. Quantile g-computation Chrononutrition variables often show high collinearity due to their interdependent definitions, which limits the analysis of their combined effects. Indeed, multicollinearity poses challenges for traditional Generalized Linear Models (GLM), leading to inflated standard errors and unstable coefficient estimates 37 . We thus applied quantile g-computation, a method originally developed in environmental epidemiology to assess the joint effect of highly correlated exposure, such as pollutants or chemicals 30 . In nutrition research, this method has recently been applied to examine mixtures of dietary nutrients 38 but, to the best of our knowledge, it has never been used in chrononutrition. Quantile g-computation involves transforming each exposure variable into quantiles and fitting a GLM that includes all quantized exposures as predictors. The method then assigns weights to each variable, reflecting its relative contribution to the overall effect. Finally, the mixture effect is calculated as the weighted sum of the exposure coefficients and interpreted as the change in the outcome associated with a quantile increase in all the mixture variables simultaneously, while adjusting for covariates 30 . Quantile g-computation allows exposure weights to act in both positive and negative directions, reflecting potentially heterogeneous associations between mixture components and the outcome. As weights represent the relative contribution of each exposure to the mixture effect in a particular direction, they should therefore be interpreted separately for positive and negative associations. We generated a priori the group of chrononutrition variables to include in the mixture analysis to ensure biological interpretability and to avoid combining variables that are directly derived from one another. Specifically, we excluded time of first and last meal, as they are used to calculate nighttime fasting, and instead included: nighttime fasting, wakeup-to-firstmeal, lastmeal-to-bedtime, eating jetlag, and number of eating occasions. The same gradual five levels of adjustment as described before were used. All statistical analyses were performed using R studio version 4.5.0, statistical tests were considered significant when p < 0.05. 3. Results Overall, 6858 participants (58.6% women) with a median age of 52 (IQR 47–58) years at baseline were included, of whom 152 developed T2D between 2018 and 2024. Baseline characteristics are shown in Table 1 for the overall population and stratified by T2D case status, and in Extended Data Table 1 stratified by sex. The prevalence of pre-existing CVD was of 32% and much higher among participants who developed T2D during follow-up compared with those who did not (53% vs. 31%), as was BMI (median 30.0 [IQR: 28–34 vs 26 [24, 29] kg/m 2 ). Regarding chrononutrition, the median time of the first meal of the day was at 8:34 (IQR 7:47 − 9:30) while the last meal of the day was typically eaten late in the evening, with a median time of 21:08 (IQR 20:49 − 21:38), resulting in a median nighttime fasting duration of 11:26 (10:34 − 12:22) hours. Figure 1 shows the correlations between chrononutrition variables where the most correlated variables were the time of first meal and nighttime fasting duration, and the time of first meal with wakeup-to-firstmeal (ρ = 0.82 and 0.74 respectively). The distribution of all the chrononutrition variables recorded in the FFQ are reported in Extended Data (Extended Data Table 2 ). Table 1 Baseline characteristics of study population overall and stratified by T2D level Overall Incident T2D test No Yes n 6858 6706 152 Sex (%) Women 4017 (58.6) 3950 (58.9) 67 (44.1) < 0.001 Age (median [IQR]) 52 [47, 58] 52 [47, 58] 57 [51, 61] < 0.001 Education (%) Primary education 1322 (19.3) 1287 (19.2) 35 (23.0) 0.012 Secondary education 2102 (30.7) 2043 (30.5) 59 (38.8) Higher education 3434 (50.1) 3376 (50.3) 58 (38.2) CVD risk factor (%) Yes 2170 (31.6) 2089 (31.2) 81 (53.3) < 0.001 Smoking habit (%) Smoker 1013 (14.8) 988 (14.7) 25 (16.4) 0.013 Former smoker 3023 (44.1) 2941 (43.9) 82 (53.9) Never smoker 2822 (41.1) 2777 (41.4) 45 (29.6) Physical activity (METs/week) (median [IQR]) 62 [40, 97] 63 [40, 98] 57 [35, 90] 0.107 BMI (median [IQR]) 26.13 [23.73, 29.08] 26.06 [23.69, 28.96] 30.09 [27.93, 34.19] < 0.001 Energy intake (median [IQR]) 1942 [1581, 2363] 1939 [1580, 2363] 2016 [1657, 2427] 0.064 rMED score categories (%) Low adherence 1603 (23.4) 1566 (23.4) 37 (24.3) 0.264 Medium adherence 3276 (47.8) 3196 (47.7) 80 (52.6) High adherence 1979 (28.9) 1944 (29.0) 35 (23.0) % UPF (weight) (median [IQR]) 8.15 [4.94, 13.03] 8.12 [4.94, 13.00] 9.46 [5.78, 14.49] 0.025 Sleeping midpoint (median [IQR]) 3:30 [3:04, 3:56] 3:30 [3:04, 3:56] 3:40 [3:10, 4:03] 0.027 N of eating occasions (median [IQR]) 3.0 [3.0, 3.7] 3.00 [3.00, 3.71] 3.00 [3.00, 3.71] 0.672 Time of first meal (median [IQR]) 8:34 [7:47, 9:30] 8:34 [7:47, 9:30] 8:34 [7:58, 9:26] 0.903 Time of first meal category (%) Before 8:30 3344 (48.8) 3272 (48.8) 72 (47.4) 0.791 Time of last meal (median [IQR]) 21:08 [20:49, 21:38] 21:08 [20:47, 21:38] 21:17 [21:00, 21:58] 0.003 Nighttime fasting (median [IQR]) 11:26 [10:34, 12:22] 11:26 [10:34, 12:22] 11:16 [10:26, 12:00] 0.037 Eating jetlag (median [IQR]) (minutes) 40 [15, 67] 40 [15, 67] 40 [15, 60] 0.334 Time wakeup to first meal (median [IQR]) (min) 52 [22, 128] 52 [22, 128] 52 [20, 108] 0.22 Time last meal to bedtime (median [IQR]) 2:27 [2:00, 3:00] 2:28 [2:00, 3:00] 2:53 [1:49, 3:00] 0.361 Eating midpoint (median [IQR] 14.89 [14.39, 15.43] 14.89 [14.39, 15.43] 15.04 [14.50, 15.46] 0.107 Table 2 Associations between chrononutrition variables and T2D estimated using logistic regression. Exposure Model 0 OR (95% CI) Single exposure Model 4 OR (95% CI) Single exposure Model 0 OR (95% CI) Mutually adjusted Model 4 OR (95% CI) Mutually adjusted Time of first meal 0.94 [0.83, 1.05] 0.84 [0.72, 0.96] 0.90 [0.79, 1.02] 0.82 [0.71, 0.95] Meal time model Time of last meal 1.43 [1.16, 1.76] 1.29 [1.05, 1.61] 1.48 [1.19, 1.85] 1.34 [1.07, 1.69] N of eating occasions 1.19 [0.88, 1.62] 1.21 [0.89, 1.64] 0.98 [0.71, 1.36] 0.96 [0.68, 1.35] Nighttime fasting 0.84 [0.75, 0.95] 0.80 [0.71, 0.91] 0.78 [0.68, 0.90] 0.78 [0.68, 0.90] Nighttime fasting model Eating midpoint 1.07 [0.89, 1.27] 0.92 [0.75, 1.12] 1.33 [1.06, 1.67] 1.10 [0.86, 1.42] N of eating occasions - - 0.98 [0.71, 1.36] 0.96 [0.68, 1.35] Wakeup-to-first meal 0.90 [0.78, 1.01] 0.85 [0.73, 0.97] 0.91 [0.79, 1.03] 0.86 [0.74, 0.98] Wake up to first meal model N of eating occasions - - 1.12 [0.81, 1.55] 1.10 [0.79, 1.52] Lastmeal-to-bedtime 0.88 [0.74, 1.05 0.81 [0.66, 0.99] 0.90 [0.75, 1.07] 0.82 [0.67, 1.01] Last meal to bedtime model N of eating occasions - - 1.14 [0.83, 1.56] 1.13 [0.82, 1.54] Eating jetlag 0.96 [0.76, 1.20] 0.98 [0.77, 1.23] 0.99 [0.77, 1.23] 1.01 [0.78, 1.27] Eating jetlag model N of eating occasions - - 1.19 [0.87, 1.62] 1.21 [0.88, 1.65] Model 0 is adjusted for age and sex; Model 4 is adjusted for age and sex, education, smoking habits, physical activity, Mediterranean diet, energy intake and ultra-processed food intake, sleeping midpoint, BMI and presence of CVD. The results for the number of eating occasions in the single-exposure model are reported only once, within the mealtime model. Logistic regression and non-linearity Spline models did not reveal evidence of non-linear associations with T2D for any chrononutrition variable (Extended Data Fig. 2), confirmed by the non-significant likelihood ratio tests (Extended Data Fig. 2). Table 2 shows the results of the single exposure and mutually adjusted models, with Model 0 used to assess the base associations, and Model 4 representing the best fitting model according to the AIC. Results from intermediate adjustment models are provided in Extended Data Table 3 . Table 3 Association between the mixture of chrononutrition variables and T2D risk. Quantile g-computation OR and 95% CI of the overall mixture and individual weights (the relative contribution of each exposure to the mixture effect in a particular direction) for each chrononutrition variable. Variable Model 0 Model 1 Model 2 Model 3 Model 4 OR overall (95% CI) 0.73 [0.50, 1.07] 0.70 [0.48, 1.03] 0.66 [0.45, 0.98] 0.61 [0.41, 0.91] 0.61 [0.41, 0.90] Chrononutrition variables individual weights Nighttime fasting -0.54 -0.55 -0.49 -0.43 -0.47 Wake up to first meal + 1 + 1 + 0.95 + 1 + 1 Last meal to bedtime -0.17 -0.14 -0.29 -0.25 -0.23 N of eating occasions -0.26 -0.28 -0.22 -0.28 -0.27 Eating jetlag -0.03 -0.04 + 0.05 -0.04 -0.04 In mutually adjusted models, later first meal was associated with a lower risk of developing T2D (model 4 OR for 1 hour increase 0.82 [95% CI: 0.71, 0.95]), whereas a later last meal was associated with an increased risk (model 4 OR 1.34 [95% CI: 1.07, 1.69]). These associations remained significant across models with different sets of covariates and in single-exposure analyses. A longer nighttime fasting duration was significantly associated with a reduced risk of T2D across all models. In particular, model 4, further adjusted for eating midpoint and number of eating occasions, showed an OR of 0.78 [95% CI: 0.68, 0.90] for 1 hour increase in nighttime fasting. A longer wakeup-to-firstmeal duration was associated with a reduced risk of T2D in models 3 and 4 (OR 0.86 [95% CI: 0.74, 0.98] and OR 0.85 [95% CI: 0.73, 0.97] respectively), as well as in the single-exposure model. Likewise, longer lastmeal-to-bedtime was associated with a reduced risk of T2D, albeit with a smaller effect size and greater uncertainty (single exposure model 4 OR 0.81 [95% CI: 0.66, 0.99], mutually adjusted model 4 OR 0.82 [95% CI: 0.67, 1.01]). Eating jetlag and the number of eating occasions were not associated with T2D risk in any model in the overall population. Stratified results show little differences across sex (Extended Data Table 4), and a longer nighttime fasting showed a slightly stronger negative association with T2D among participants eating their first meal after 8:30 (Extended Data Table 5). There was no significant interaction between nighttime fasting and energy intake nor UPF intake. Quantile g-computation The mixture model simultaneously considered all chrononutrition variables, estimating on the one hand their joint effect on T2D risk, and on the other hand the relative contribution of each variable to the overall association (weights) (Table 3 ). A one-IQR increase across all variables was associated with a decrease in the T2D risk (model 4 OR 0.61 [95% CI: 0.41, 0.90]). In particular, the association was protective in all the four models adjusted for an increased number of confounders, and resulted stronger in models from 2 to 4. Moreover, in all the four models, nighttime fasting had the biggest role in contributing to the decrease of risk. A higher number of eating occasions and a longer lastmeal-to-bedtime also carried substantial protective weight. In contrast, a longer wakeup-to-firstmeal showed a positive weight on risk, though this did not offset the overall protective association of the mixture. Eating jetlag had only minimal weight, with inconsistent direction across models. 4. Discussion This study provides new insights into the role of chrononutritional behaviors in the development of T2D. To the best of our knowledge, this is the first study to combine both the widely used single-exposure logistic regression and a novel mixture method to analyze the joint effects of chrononutrition variables. Our findings consistently indicated that longer nighttime fasting, achieved through both a later first meal and an earlier last meal, were associated with a lower risk of incident T2D. This suggests that fasting duration may be a key characteristic of chrononutritional patterns relevant to metabolic health. Less consistent association across models were observed for time elapsed between wakeup-to-firstmeal, as well as between the lastmeal-to-bedtime, where longer intervals were associated with lower T2D risk. Taken together, these findings support the hypothesis that the duration of the fasting window, rather than the meal timing itself, may be the key element driving these associations. Our results are consistent with Carew et al. 14 who observed a protective association for later first meals (after 9AM) in a prospective study of a US population. In contrast, Palomar-Cros et al. 11 reported that later first meals were associated with a higher risk of T2D in a French cohort, and a protective effect of nighttime fasting was only evident when the first meal happened before 8AM. Although their cohort study is the only one exploring nighttime fasting in relation to T2D, it differs from ours in participant age and chrononutrition behaviors, particularly the time of last meal and the number of eating occasions. The higher risk of T2D associated with a later last meal observed in our study aligns with the hypothesis that late dinners may shorten the overnight fasting window and contribute to circadian misalignment. Controlled trials showed that late-day carbohydrate intake impairs glycemic control 3 , and Verde et al. 6 and López-Prieto 4 et al. highlighted the role of late-night eating in circadian misalignment and metabolic disruption. Moreover, late eating has been associated to obesity and poor glucose tolerance in prediabetic subjects 15,16 . Our findings thus reinforce these results and support growing concerns about evening eating patterns in modern societies. A longer fasting window has been suggested to be metabolically advantageous, since extended fasting intervals promote ketogenesis, enhance lipolysis, and reduce hepatic glucose output, improving insulin sensitivity and lowering inflammation 17 . They may also restore the diurnal rhythm of hormones such as cortisol, leptin, and melatonin, which are often blunted by irregular or late-night eating 18 ; this interaction has not been fully explored in previous studies. Longer fasting (also known as time restricted eating) is increasingly being investigated as a dietary strategy for the prevention and therapy of glucose and lipid metabolic disturbances. It has generally been shown to lower fasting and post-meal glucose, improve 24-hour glucose control, reduce fasting insulin, and enhance insulin sensitivity 19 . However, some trials show no benefit or even negative effects 19 . These mixed results likely stem from differences in study design, such as eating window length, timing, calorie changes, and intervention duration, as well as variations in participant characteristics like metabolic health, age, sex, and chronotype 20 . Studies in animal models have shown that longer fasting can prevent and even reverse T2D through multiple molecular mechanisms. In mice and rats fed high-fat diets, restricting feeding time (e.g., 8 hours feeding, 16 hours fasting) improves insulin sensitivity, reduces systemic inflammation, and prevents hyperglycemia without reducing total caloric 21 . These effects are linked to increased efficiency of insulin receptor signaling, AMPK activation, reduction of the mTOR pathway, increased mitochondrial biogenesis, and stimulation of autophagy. Furthermore, intermittent fasting increases the expression of BDNF and CREB, factors involved in regulating energy metabolism and cell protection, promoting the preservation of pancreatic β cells and glycemic homeostasis 22 . We observed no significant associations for eating jetlag, differing from other studies that suggest metabolic dysregulation with irregular eating schedules 23,24 . This discrepancy may be due to limited variability or measurement sensitivity in our cohort. We also found no association between the number of eating occasions and the risk of T2D. However, evidence from literature showed that a higher number of meals was associated with a reduced risk of T2D 11 and this may be explained by a reduced serum insulin and lipid concentrations between meals 25–27 . Conversely, an observational study found that additional snacks beyond the three main meals were associated with an increased risk of T2D 28 . The lack of findings in our population may reflect the lack of variability in the number of meals compared to other studies. The mixture analysis largely confirmed results from the logistic models, showing an overall protective association of the combined set of chrononutritional behaviors with T2D risk. In particular, nighttime fasting had the highest negative weight followed by lastmeal-to-bedtime. Wakeup-to-firstmeal had a positive weight in the mixture, although its influence on the overall estimate appeared limited, as the mixture OR remained consistently protective across models adjustments. Interestingly, the number of eating occasions received a negative weight in the mixture, despite showing no clear association in logistic models. These observations suggest the presence of potential synergistic relationships among correlated chrononutritional behaviors that may not be fully captured when exposures are assessed in single-exposure models. Cluster analyses are often used to identify and describe groups of individuals with similar chrononutritional patterns, but the derived groups depend on data-driven decisions, may vary across populations, and do not allow estimation of the risk associated with modifying the variables 29 . In contrast, quantile g-computation is a quantitative method that provides an estimate of the overall effect of the mixture, identifies the relative contribution of each exposure, accounts for multicollinearity and allows insights into potential synergistic effects 30 . Overall, few epidemiological studies have examined these synergistic patterns, and our results illustrate how mixture methods, in particular quantile g-computation, can offer a valuable methodological option for further investigating complex behavioral exposures. We recognize in our study the correlation between chrononutrition variables were not extremely high, but by introducing this analytic strategy to the field, our work offers a practical guideline for researchers seeking methods that more appropriately reflect the multidimensional nature of temporal eating patterns. The findings of this study underscore the importance of incorporating chrononutrition into public health strategies for T2D prevention. While traditional nutritional guidance has focused primarily on dietary quantity and quality, our results highlight that when people eat may be relevant for metabolic health. In particular, the consistent protective association observed for longer nighttime fasting suggests that aligning food intake with endogenous circadian rhythms could offer a practical, low-cost behavioral target for diabetes prevention. These insights have potential implications for dietary guidelines, workplace health programs, and clinical counseling, particularly in Mediterranean populations where late-night eating is culturally embedded. Public health interventions that promote earlier cessation of food intake, reduce eating close to bedtime, and encourage more regular temporal patterns of intake may help reduce T2D incidence at the population level. As chrononutrition behaviors are modifiable, integrating temporal eating recommendations into existing nutritional policies could represent a scalable and equitable approach to mitigate the growing global burden of T2D. Of note, some limitations of our study should be acknowledged. First, given the observational design, causality cannot be inferred. Meal timing data were self-reported and collected at a single time point, and eating occasions were limited to 5 possible meals, which may have introduced misclassification bias, recall bias and rounding errors. Moreover, the questionnaire did not allow us to assess the energy intake distribution throughout the day, factor that may be relevant for glucose metabolism and T2D risk. The relatively small number of incident cases of T2D (152) has also limited the power of stratified analyses. Finally, generalizability may be limited to similar adult populations in Mediterranean countries. Conclusions Our results indicate that a longer nighttime fasting duration is associated with a reduced risk of developing T2D, and that chrononutrition variables may act synergistically in ways that are not captured by single-exposure models. The implementation of quantile g-computation can contribute to a better understanding of these complex and interrelated behaviors. Integrating temporal eating recommendations into public health dietary guidance might represent a simple and widely applicable strategy to help reduce the burden of T2D at the population level. Declarations Acknowledgements ChatGPT-5, a language model developed by OpenAI, was used to assist with improving the grammar and flow of certain paragraphs of the manuscript. The content was originally written by the authors, and ChatGPT was used solely for enhancement purposes, not for generating new content or ideas. Data availability: The data that support the findings of this study are not publicly available due to patient privacy protections and ethical restriction but are available from the corresponding author upon reasonable request and with appropriate approvals. Code availability: The code and a sample database are available at https://github.com/joanall Conflict of Interests The authors declare no conflict of interest. Funding: E.G. gratefully acknowledge the support of grant JDC2024-055407-I, funded by MICIU/AEI/10.13039/501100011033 and the ESF+. Camille Lassale is supported by a Ramon y Cajal Fellowship RYC2020-029599 funded by MICIU/AEI/10.13039/501100011033 and the European Social Fund “Invest in your future”. This project was funded by the Fundació La MaratóTV3 – Cardiovascular disease (grant agreement 202323-30-31, CUPID project). This study makes use of data generated by the GCAT-Genomes for Life. Cohort study of the Genomes of Catalonia, Fundacio IGTP. IGTP is part of the CERCA Program / Generalitat de Catalunya. GCAT was funded by Acción de Dinamización del ISCIII-MINECO and the Ministry of Health of the Generalitat of Catalunya (ADE 10/00026); and have additional suport by the Agència de Gestió d’Ajuts Universitaris i de Recerca (AGAUR) (SGR 01537), Spanish National Grant PI18/01512, TED2021-130626B-I00, La MaratoTV3 167/C/2021, VEIS project (001-P-001647) (co-funded by European Regional Development Fund (ERDF), “A way to build Europe”), the European Union under grant agreement no. 101046314 (END-VOC). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. Authors’ contributions All authors contributed to the study conception and design. Data acquisition was performed by RdC, MK. Data curation was performed by EG, JL-P, CR, MO, CL, AP-C. Funding acquisition was carried by RdC, MK. The formal analysis was performed by EG, JL-P, CR. The supervision of the present study was done by CL. The first draft of the manuscript was written by EG and CR and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. The work reported in the paper has been performed by the authors, unless clearly specified in the text. References International Diabetes Federation (IDF), I. D. F. (IDF). I. D. A., 11th Edition. Brussels, Belgium: IDF; 2025. https://diabetesatlas.org. IDF Diabetes Atlas . (Brussels, Belgium, 2025). Reytor-González, C. et al. Chrononutrition and Energy Balance: How Meal Timing and Circadian Rhythms Shape Weight Regulation and Metabolic Health. Nutrients 17 , 2135 (2025). Henry, C. J., Kaur, B. & Quek, R. Y. C. Chrononutrition in the management of diabetes. Nutr. Diabetes 10 , 6 (2020). López-Prieto, R. S. et al. Time Matters: An Insight into the Relationship Between Chrononutrition and Diabetes. Rev. Investig. Clínica 76 , 80–90 (2024). Kim, Y.-I., Kim, E., Lee, Y. & Park, J. Role of late-night eating in circadian disruption and depression: a review of emotional health impacts. Phys. Act. Nutr. 29 , 18–24 (2025). Verde, L. et al. Chrononutrition in type 2 diabetes mellitus and obesity: A narrative review. Diabetes Metab. Res. Rev. 40 , e3778 (2024). O’Connor, S. G. et al. Conceptualization and Assessment of 24-H Timing of Eating and Energy Intake: A Methodological Systematic Review of the Chronic Disease Literature. Adv. Nutr. 15 , 100178 (2024). Knutson, K. L. et al. Role of Circadian Health in Cardiometabolic Health and Disease Risk: A Scientific Statement From the American Heart Association. Circulation 152 , e408–e419 (2025). Buscemi, C. et al. The impact of breakfast skipping on plasma glucose levels in non-diabetic individuals: gender-based differences and implications. Int. J. Food Sci. Nutr. 76 , 203–208 (2025). Minari, T. P. & Pisani, L. P. Skipping breakfast and its wide-ranging health consequences: A systematic review from multiple metabolic disruptions to socioeconomic factors. Nutr. Res. 141 , 34–45 (2025). Palomar-Cros, A. et al. Associations of meal timing, number of eating occasions and night-time fasting duration with incidence of type 2 diabetes in the NutriNet-Santé cohort. Int. J. Epidemiol. 52 , 1486–1497 (2023). Santonja, I. et al. Meal-timing patterns and chronic disease prevalence in two representative Austrian studies. Eur. J. Nutr. 62 , 1879–1890 (2023). Pons-Muzzo, L. et al. Sex-specific chrono-nutritional patterns and association with body weight in a general population in Spain (GCAT study). Int. J. Behav. Nutr. Phys. Act. 21 , 102 (2024). Carew, A. S. et al. Prospective study of breakfast frequency and timing and the risk of incident type 2 diabetes in community-dwelling older adults: the Cardiovascular Health Study. Am. J. Clin. Nutr. 116 , 325–334 (2022). McHill, A. W. et al. Later circadian timing of food intake is associated with increased body fat†. Am. J. Clin. Nutr. 106 , 1213–1219 (2017). Díaz-Rizzolo, D. A. et al. Late eating is associated with poor glucose tolerance, independent of body weight, fat mass, energy intake and diet composition in prediabetes or early onset type 2 diabetes. Nutr. Diabetes 14 , 90 (2024). Ruppert, P. M. M. & Kersten, S. Mechanisms of hepatic fatty acid oxidation and ketogenesis during fasting. Trends Endocrinol. Metab. 35 , 107–124 (2024). Kim, Y.-I., Kim, E., Lee, Y. & Park, J. Role of late-night eating in circadian disruption and depression: a review of emotional health impacts. Phys. Act. Nutr. 29 , 018–024 (2025). Chen, Y.-E., Tsai, H.-L., Tu, Y.-K. & Chen, L.-W. Effects of timing and eating duration of time restricted eating on metabolic outcomes: systematic review and network meta-analysis. BMJ Med. 5 , (2026). Schuppelius, B., Peters, B., Ottawa, A. & Pivovarova-Ramich, O. Time Restricted Eating: A Dietary Strategy to Prevent and Treat Metabolic Disturbances. Front. Endocrinol. 12 , (2021). Belkacemi, L. et al. Intermittent Fasting Modulation of the Diabetic Syndrome in Streptozotocin-Injected Rats. Int. J. Endocrinol. 2012 , 1–12 (2012). Mattson, M. P., Longo, V. D. & Harvie, M. Impact of intermittent fasting on health and disease processes. Ageing Res. Rev. 39 , 46–58 (2017). Makarem, N. et al. Variability in Daily Eating Patterns and Eating Jetlag Are Associated With Worsened Cardiometabolic Risk Profiles in the American Heart Association Go Red for Women Strategically Focused Research Network. J. Am. Heart Assoc. 10 , e022024 (2021). Chen, Y.-E. et al. Associations of >1-h compared with 1-h meal timing variability (eating jetlag) with plasma glycemic parameters and continuous glucose monitoring measures among pregnant females: a prospective cohort study. Am. J. Clin. Nutr. 122 , 244–254 (2025). Jenkins, D. J. A. et al. Nibbling versus Gorging: Metabolic Advantages of Increased Meal Frequency. The New England Journal of Medicine https://www.nejm.org/doi/abs/10.1056/NEJM198910053211403 (1989) doi:10.1056/NEJM198910053211403. Bertelsen, J. et al. Effect of Meal Frequency on Blood Glucose, Insulin, and Free Fatty Acids in NIDDM Subjects. Diabetes Care 16 , 4–7 (1993). Titan, S. M. O. et al. Frequency of eating and concentrations of serum cholesterol in the Norfolk population of the European prospective investigation into cancer (EPIC-Norfolk): cross sectional study. BMJ 323 , 1286 (2001). Mekary, R. A., Giovannucci, E., Willett, W. C., van Dam, R. M. & Hu, F. B. Eating patterns and type 2 diabetes risk in men: breakfast omission, eating frequency, and snacking1234. Am. J. Clin. Nutr. 95 , 1182–1189 (2012). Zhao, J. et al. A review of statistical methods for dietary pattern analysis. Nutr. J. 20 , 37 (2021). Keil, A. P. et al. A Quantile-Based g-Computation Approach to Addressing the Effects of Exposure Mixtures. Environ. Health Perspect. 128 , 47004 (2020). Obón-Santacana, M. et al. GCAT|Genomes for life: a prospective cohort study of the genomes of Catalonia. BMJ Open 8 , e018324 (2018). Hoeymans, N., Garssen, A. A., Westert, G. P. & Verhaak, P. F. Measuring mental health of the Dutch population: a comparison of the GHQ-12 and the MHI-5. Health Qual. Life Outcomes 2 , 23 (2004). Fernández-Ballart, J. D. et al. Relative validity of a semi-quantitative food-frequency questionnaire in an elderly Mediterranean population of Spain. Br. J. Nutr. 103 , 1808–1816 (2010). Saulle, R., Bernardi, M., Chiarini, M., Backhaus, I. & La Torre, G. Shift work, overweight and obesity in health professionals: a systematic review and meta-analysis. Clin. Ter. 169 , e189–e197 (2018). Buckland, G. et al. Adherence to a Mediterranean diet and risk of gastric adenocarcinoma within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort study. Am. J. Clin. Nutr. 91 , 381–390 (2010). Martinez-Steele, E. et al. Best practices for applying the Nova food classification system. Nat. Food 4 , 445–448 (2023). Yu, L. et al. A review of practical statistical methods used in epidemiological studies to estimate the health effects of multi-pollutant mixture. Environ. Pollut. 306 , 119356 (2022). Liu, Y. et al. Synergistic Effects of Nutrients on Musculoskeletal Health in Gerontology: Understanding the Combined Impact of Macronutrients and Micronutrients. Nutrients 16 , 1640 (2024). Additional Declarations There is NO Competing Interest. Supplementary Files ExtendedData.docx Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9107412","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":625164946,"identity":"083444bc-256c-45ab-b8e1-b524796134f1","order_by":0,"name":"Camille Lassale","email":"data:image/png;base64,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","orcid":"","institution":"ISGlobal","correspondingAuthor":true,"prefix":"","firstName":"Camille","middleName":"","lastName":"Lassale","suffix":""},{"id":625164947,"identity":"ce548ca7-539d-48a7-bb4d-ebbbca64acf6","order_by":1,"name":"Elisa Gallo","email":"","orcid":"https://orcid.org/0000-0002-0827-4474","institution":"Barcelona Institute for Global Health","correspondingAuthor":false,"prefix":"","firstName":"Elisa","middleName":"","lastName":"Gallo","suffix":""},{"id":625164948,"identity":"26ed1476-c690-47ba-abb5-fe1bae44a8c6","order_by":2,"name":"Catalina Ramírez-Contreras","email":"","orcid":"","institution":"Departamento de Nutrición y Salud Pública, Facultad de Ciencias de la Salud y de los Alimentos, Universidad del Bío-Bío","correspondingAuthor":false,"prefix":"","firstName":"Catalina","middleName":"","lastName":"Ramírez-Contreras","suffix":""},{"id":625164949,"identity":"f127bb5e-0897-4b7f-9a4a-4e252fa6eb54","order_by":3,"name":"Joana Llauradó-Pont","email":"","orcid":"","institution":"ISGlobal","correspondingAuthor":false,"prefix":"","firstName":"Joana","middleName":"","lastName":"Llauradó-Pont","suffix":""},{"id":625164950,"identity":"b81911c1-9155-4fd5-9593-8da47a1015b0","order_by":4,"name":"Kyriaki Papantoniou","email":"","orcid":"","institution":"Department of Epidemiology, Center for Public Health, Medical University of Vienna","correspondingAuthor":false,"prefix":"","firstName":"Kyriaki","middleName":"","lastName":"Papantoniou","suffix":""},{"id":625164951,"identity":"6ea957cc-bf85-47fe-8ccf-3bc29fc33e9b","order_by":5,"name":"Manolis Kogevinas","email":"","orcid":"","institution":"ISGlobal, Barcelona Institute for Global Health","correspondingAuthor":false,"prefix":"","firstName":"Manolis","middleName":"","lastName":"Kogevinas","suffix":""},{"id":625164952,"identity":"97675014-3433-4375-83b0-749974dda026","order_by":6,"name":"Anna Palomar-Cros","email":"","orcid":"","institution":"Fundació Institut Universitari per a la recerca a l’Atenció Primària de Salut Jordi Gol i Gurina (IDIAPJGol)","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Palomar-Cros","suffix":""},{"id":625164953,"identity":"e4912ee6-8e4c-40ae-9248-0827080b8611","order_by":7,"name":"Andrea Montanari","email":"","orcid":"","institution":"Instituto de Salud Global Barcelona","correspondingAuthor":false,"prefix":"","firstName":"Andrea","middleName":"","lastName":"Montanari","suffix":""},{"id":625164954,"identity":"1a6c62d0-ecc1-4281-bd92-3fff2bf1a5ae","order_by":8,"name":"Rafael de Cid","email":"","orcid":"","institution":"IGTP","correspondingAuthor":false,"prefix":"","firstName":"Rafael","middleName":"","lastName":"de Cid","suffix":""},{"id":625164955,"identity":"8acbcea0-6fc2-43d1-8b4b-4ec4de3c981d","order_by":9,"name":"Mireia Obón-Santacana","email":"","orcid":"","institution":"Oncology Data Analytics Program, Catalan Institute of Oncology","correspondingAuthor":false,"prefix":"","firstName":"Mireia","middleName":"","lastName":"Obón-Santacana","suffix":""}],"badges":[],"createdAt":"2026-03-12 17:39:48","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9107412/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9107412/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107916754,"identity":"52f4e881-e372-4b33-aa66-35804955a981","added_by":"auto","created_at":"2026-04-27 14:19:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":92128,"visible":true,"origin":"","legend":"\u003cp\u003eSpearman’s correlation between chrononutrition variables.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-9107412/v1/a34c874d84396a7b96a81e92.png"},{"id":108007305,"identity":"c7e716ae-51a3-4cc7-a4c9-8627b231485e","added_by":"auto","created_at":"2026-04-28 12:59:26","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":561279,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9107412/v1/9f3c4d86-59b0-4c4c-8ca8-105d252fd6f4.pdf"},{"id":107916755,"identity":"f190ff0d-a3da-41d9-8768-b04bd8116e4d","added_by":"auto","created_at":"2026-04-27 14:19:05","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":289456,"visible":true,"origin":"","legend":"","description":"","filename":"ExtendedData.docx","url":"https://assets-eu.researchsquare.com/files/rs-9107412/v1/d4b691d2d5ee92ee15c5e352.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Chrononutrition and risk of type 2 diabetes: a prospective cohort study","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eType 2 diabetes (T2D) is one of the most pressing non-communicable diseases worldwide: according to the International Diabetes Federation's (IDF) 2025 Atlas, an estimated 589\u0026nbsp;million adults (11.1%) worldwide are living with diabetes, a figure that is expected to rise to 853\u0026nbsp;million by 2050, equivalent to almost one in eight adults\u003csup\u003e1\u003c/sup\u003e. Identifying modifiable and actionable risk factors is key to prevention and reduce T2D burden.\u003c/p\u003e \u003cp\u003eAn emerging but under-researched area is chrononutrition, the study of how the timing of food intake interacts with circadian biology and influences metabolic health. Chrononutrition includes behaviors such as timing of meals, number of eating occasions, and nighttime fasting, each of which is regulated by endogenous circadian rhythms and exogenous lifestyle and social influences\u003csup\u003e2,3\u003c/sup\u003e. There are multiple potential mechanisms linking chrononutrition to metabolic dysfunction. Central and peripheral circadian clocks regulate glucose homeostasis, insulin secretion, lipid metabolism and energy expenditure\u003csup\u003e2\u003c/sup\u003e. Disruption of these clocks, whether through late-night eating, irregular mealtimes or shortened fasting intervals, can lead to internal misalignment, resulting in impaired glucose regulation, decreased insulin sensitivity in adipose tissue and dysregulation of appetite-related hormones such as ghrelin and leptin\u003csup\u003e3,4\u003c/sup\u003e. Late-night food intake may also blunt the diurnal amplitude of cortisol and melatonin, further impairing metabolic control\u003csup\u003e5\u003c/sup\u003e. These pathways suggest that timing of eating behaviors may act as an active driver of metabolic dysfunction when it is not in sync with endogenous biological rhythms.\u003c/p\u003e \u003cp\u003eWhile individual components of chrononutritional behaviors have been associated with metabolic outcomes, a key limitation is that most existing studies isolate individual timing variables, neglecting the possibility that synergistic patterns of chrononutritional behaviors may exert a more substantial influence on disease risk\u003csup\u003e6,7\u003c/sup\u003e. This represents a critical gap in the current evidence base. Recent cohort studies suggest that late eating, skipping breakfast and irregular meal frequency are associated with unfavorable glucose metabolism and an increased risk of T2D\u003csup\u003e8\u0026ndash;10\u003c/sup\u003e. For instance, when following 100,000 French adults from the NutriNet-Sant\u0026eacute; cohort over 7 years, Palomar-Cros et al.\u003csup\u003e11\u003c/sup\u003e found that participants who ate their first meal after 9 am had a 59% increased risk of incident T2D compared to those who ate before 8 am, whereas extended nighttime fasting (\u0026gt;\u0026thinsp;13 hours) provided a metabolic benefit only when combined with early breakfast intake. These results highlight the interaction between multiple eating behaviors, not just their individual effects.\u003c/p\u003e \u003cp\u003eDifferent chrononutrition variables, such as the timing of the first and last meal, nighttime fasting duration, or the interval between the last meal and bedtime, are often highly interrelated, leading to multicollinearity when analyzed jointly, or complicating their interpretation when analyzed separately\u003csup\u003e12\u003c/sup\u003e. To address this complexity, cluster analysis has been used as a dimension-reduction technique to group individuals with similar chrononutritional behaviors and to capture overall eating-timing patterns. Our group\u003csup\u003e13\u003c/sup\u003e previously applied this approach in the GCAT study, identifying distinct sex-specific chrononutritional patterns in adults from Catalonia and related them to body mass index (BMI), while Santonja et al.\u003csup\u003e12\u003c/sup\u003e reported different clusters in Austrian surveys associated with obesity, depression, and poor self-rated health. However, neither study focused on T2D as an outcome, and cluster analysis primarily served a descriptive purpose, limiting its ability to quantify the associations of chrononutrition variables with the outcome risk. Recognizing these gaps, the present study explores alternative analytical approaches to better capture the multidimensional structure of chrononutritional behaviors and their potential links to T2D risk.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cp\u003e\u003cstrong\u003e2.1 Participants and study design\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Genomes for Life (GCAT) study\u003csup\u003e31\u003c/sup\u003e is a population-based, prospective cohort study including over 19,000 adult volunteers in Catalonia region (Spain), designed to investigate the role of environmental and genetic factors in the development of chronic diseases. The full description of the population is reported in the Extended Data. Briefly, inclusion criteria for participation in the study included being between 40 and 65 years of age, understanding at least one of the two official languages of the region, possession of an individual Health System Identification Card, and current residence in Catalonia. Exclusion criteria included mental or health impairment disorders to give written informed consent or efficient communication or planning to leave Catalonia in the next five years. The baseline visit was in 2014–2017, when information on demographic, socioeconomic data, biological and clinical data were collected. A follow-up was conducted in 2018 and included a questionnaire on lifestyle, the Mental Health Inventory score (MHI5)\u003csup\u003e32\u003c/sup\u003e, night shift work history, circadian habits and the food-frequency questionnaire (FFQ). \u0026nbsp;The GCAT cohort study was approved by a local Ethics committee (PI13-020) and written informed consent was obtained from participants. The data was extracted in an anonymized format so that participants could not be identified.\u003c/p\u003e\n\u003cp\u003eThe GCAT cohort collaborates with the Catalan Health Department to link electronic health records (EHRs) through the “Programa d'analítica de dades per a la recerca i la innovació en salut” or PADRIS system.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eSample selection and flow chart\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eFrom then original cohort, we included in this study only those who completed the Follow-up 1 questionnaire in 2018 and a semiquantitative FFQ (n = 8079). The FFQ\u003csup\u003e33\u003c/sup\u003e, specifically developed and validated for the Spanish adult population, includes over 128 items covering the most commonly eaten food groups in Spain and refers to habitual intake during the previous 12 months.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSubjects with missing or implausible chrononutrition values (see 2.3 below), such as first meal happening before the wake-up time or last meal happening after bedtime were excluded (n = 921). Participants who engaged in night shift work were also excluded, as such schedules can disrupt circadian rhythms (including eating patterns, physical activity, and sleep) and lead to chrono-disruption (n = 55)\u003csup\u003e34\u003c/sup\u003e. In addition, subjects with a diagnosis of diabetes prior to 2018 were excluded in order to model incidence of T2D. Finally, 24 participants with missing data in covariates were removed from the analysis. This resulted in a final sample of 6858 subjects (Extended Data Figure 1).\u003c/p\u003e\n\u003cp\u003eSince the chrononutrition information was collected for the first time at Follow-up 1 in 2018, this constitutes the baseline of the present analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2 Outcome variables\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eType 2 Diabetes Mellitus\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe diagnosis of T2D was extracted from the electronic health records (EHR) and defined according to the harmonized ICD-10 diagnosis code E11.XX. Subjects with a diagnosis of T2D prior to 2018 (prevalent cases) were excluded. Incident cases and diagnosis date were identified between 01/01/2018 and 31/12/2024.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3 Exposure variables\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eChrononutrition variables\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eConcomitant to the FFQ (2018 follow-up questionnaire), additional questions were asked about the usual times of main eating occasions, spanning 5 eating occasions (breakfast, lunch, afternoon snack, dinner, post-dinner snack), separately for the week/working days and the weekend/free days (see Extended Data for exact questions). A weighted average was calculated for each chrononutrition variable listed below (X) as (Xweekday*5 + Xweekend*2)/7, except for eating jet-lag. The following variables were derived:\u003c/p\u003e\n\u003cp\u003ei.\u0026nbsp; \u0026nbsp;\u0026nbsp;Time of first meal in hours and minutes\u003c/p\u003e\n\u003cp\u003eii.\u0026nbsp; \u0026nbsp;Time of last meal in hours and minutes\u003c/p\u003e\n\u003cp\u003eiii. Nighttime fasting duration: time in hours between the last meal of a day and the first meal of the following day.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eiv.\u0026nbsp;\u0026nbsp;Wakeup-to-firstmeal: difference in hours between the time of awakening and the time of the first meal.\u003c/p\u003e\n\u003cp\u003ev.\u0026nbsp; \u0026nbsp;Lastmeal-to-bedtime: difference in hours between the time of the last meal and bedtime.\u003c/p\u003e\n\u003cp\u003evi.\u0026nbsp;\u0026nbsp;Number of eating occasions: number of meals reported by each participant, ranging from 1 to 5 possible eating occasions.\u003c/p\u003e\n\u003cp\u003evii. Eating jet lag: difference in hours between the\u0026nbsp;eating midpoint\u0026nbsp;on weekend and the\u0026nbsp;eating midpoint on weekday, where eating midpoint is defined as the midpoint between the first and the last meal of the day. Analyses conducted as continuous variable were using the absolute value of the eating jet lag.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4 Covariates\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe baseline questionnaire included information on age (years), sex (men/women), education (primary/secondary/higher),\u0026nbsp;\u0026nbsp;and was updated with information from the 2018 follow-up on smoking (smoker, former smoker and never smoker), physical activity (METs/week) and BMI (kg/m\u003csup\u003e2\u003c/sup\u003e). From the FFQ data, food groups and nutrient intake was derived, and we calculated a score of adherence to the Mediterranean diet (rMED)\u003csup\u003e35\u003c/sup\u003e. The total rMED score ranges from 0 to 18 with higher scores reflecting greater adherence to the Mediterranean diet. We also derived the share of the diet that come from ultra-processed food (UPF). All food items were classified into four groups according to the Nova classification system: 1) unprocessed or minimally processed foods, 2) processed culinary ingredients, 3) processed foods, and 4) UPF\u003csup\u003e36\u003c/sup\u003e. We calculated the proportional (%) of UPF consumed by each participant relative to the total weight of food and beverage (grams/day).\u003c/p\u003e\n\u003cp\u003eThe sleep timing questions referred to bedtime and wake-up time on weekdays and weekends. The weighted average was used to calculate sleeping midpoint as the midpoint between bedtime and wake-up time.\u003c/p\u003e\n\u003cp\u003eWe identified prevalent cardiovascular diseases or risk factor (CVD), diagnosed previous to T2D onset, from the EHR using ICD-10 codes, including I20\" (angina), \"I21\" (myocardial infarction), \"I10\" (hypertension), \"E780\" (hypercholesterolemia), \"I22\", \"I23\", \"I24\", \"I25\", (ischemic heart disease), \"I60\", \"I61\", \"I62\", \"I63\", \"I64\", \"I65\", \"I66\", \"I67\", \"I68\", \"I69\", (cerebrovascular), \"I50\" (heart failure). A full characterization of the covariates is provided in the Extended Data.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.5 Statistical Analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDescriptive characteristics of the study population are presented as\u0026nbsp;median and IQR for continuous variables and proportions for categorical variables. The description was performed overall, according to T2D, and separately for men and women, and we report the p-value for chi-square\u0026nbsp;or Kruskal-Wallis Rank Sum test.\u003c/p\u003e\n\u003cp\u003eThe explanatory variables (time of first meal, time of last meal, number of eating occasions, nighttime fasting, wakeup-to-firstmeal, lastmeal-to-bedtime, and eating jetlag) were modelled as continuous variables. We explored their correlation by using Spearman’s rank correlation and plotted a heatmap.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLogistic regression and non-linearity\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eWe fitted logistic regression models to assess the association between chrononutrition variables and T2D, and reported odds ratios (ORs) with 95% confidence intervals (CIs) for one meal increase in the number of eating occasions and for one hour increase in all the other chrononutrition variables.\u003c/p\u003e\n\u003cp\u003eWe first examined potential non-linear associations between chrononutrition variables and T2D risk. For each chrononutrition variable, we fitted a base logistic regression model adjusted for age and sex. To evaluate whether associations deviated from linearity, we compared two models for each chrononutrition variable: one in which the exposure was modeled linearly and one in which it was modeled using natural cubic splines with three degrees of freedom. Model fit was compared using likelihood ratio tests (ANOVA) to assess whether the spline-based model provided a significantly better fit than the linear model. Since no clear deviation from linearity was observed, we modelled chrononutrition variables as linear terms.\u003c/p\u003e\n\u003cp\u003eA series of logistic models with increasing levels of adjustment was fitted to assess the robustness of the associations:\u003c/p\u003e\n\u003cp\u003eModel 0: age, sex\u003c/p\u003e\n\u003cp\u003eModel 1: Model 0 + education, smoking habits, physical activity, rMED, energy intake and UPF intake;\u003c/p\u003e\n\u003cp\u003eModel 2: Model 1 + sleeping midpoint;\u003c/p\u003e\n\u003cp\u003eModel 3: Model 2 + BMI;\u003c/p\u003e\n\u003cp\u003eModel 4: Model 3 + presence of CVD.\u003c/p\u003e\n\u003cp\u003eSince BMI and the presence of CVD might act as mediators, we separately included them in the last two models.\u003c/p\u003e\n\u003cp\u003eMoreover, we applied two complementary modeling approaches. First, we conducted a single-exposure model, where each chrononutrition variable was analyzed separately, second, we conducted mutually adjusted models, where multiple chrononutrition variables entered simultaneously in the same model. In both cases, covariates were added stepwise following Models 0–4 above. This strategy allows single-exposure results for each chrononutrition variable, and mutually adjusted results showing if associations vary when multiple meal metrics are jointly modeled and when covariates are added stepwise. To assess potential multicollinearity among predictors, we computed the generalized variance inflation factor (GVIF), which extends the standard VIF to handle categorical variables with multiple degrees of freedom. All models showed GVIF values \u0026lt; 5, indicating no concerning collinearity.\u003c/p\u003e\n\u003cp\u003eIn mutually adjusted analyses, time of first meal, time of last meal, and number of eating occasions were modelled jointly; Nighttime fasting was adjusted for number of eating occasions and eating midpoint to account for the timing of the fasting period; For the remaining chrononutrition exposures, each model was further adjusted for number of eating occasions. To explore potential effect modifications, we tested interactions between nighttime fasting duration and both energy and UPF intake using likelihood ratio tests. We also conducted stratified analyses to further explore heterogeneity in associations. Specifically, we estimated associations separately for participants who broke their nighttime fast at or before 8:30 AM and those who broke it after 8:30 AM. In addition, we performed sex-stratified analyses to assess potential differences in the association between chrononutrition variables and T2D. In all stratified analyses, given the limited number of incidents T2D events, we only computed base models to avoid overfitting.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eQuantile g-computation\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eChrononutrition variables often show high collinearity due to their interdependent definitions, which limits the analysis of their combined effects. Indeed, multicollinearity poses challenges for traditional Generalized Linear Models (GLM), leading to inflated standard errors and unstable coefficient estimates\u003csup\u003e37\u003c/sup\u003e. We thus applied quantile g-computation, a method originally developed in environmental epidemiology to assess the joint effect of highly correlated exposure, such as pollutants or chemicals\u003csup\u003e30\u003c/sup\u003e. In nutrition research, this method has recently been applied to examine mixtures of dietary nutrients\u003csup\u003e38\u003c/sup\u003e but, to the best of our knowledge, it has never been used in chrononutrition.\u003c/p\u003e\n\u003cp\u003eQuantile g-computation involves transforming each exposure variable into quantiles and fitting a GLM that includes all quantized exposures as predictors. The method then assigns weights to each variable, reflecting its relative contribution to the overall effect. Finally, the mixture effect is calculated as the weighted sum of the exposure coefficients and interpreted as the change in the outcome associated with a quantile increase in all the mixture variables simultaneously, while adjusting for covariates\u003csup\u003e30\u003c/sup\u003e . Quantile g-computation allows exposure weights to act in both positive and negative directions, reflecting potentially heterogeneous associations between mixture components and the outcome. As weights represent the relative contribution of each exposure to the mixture effect in a particular direction, they should therefore be interpreted separately for positive and negative associations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe generated \u003cem\u003ea priori\u003c/em\u003e the group of chrononutrition variables to include in the mixture analysis to ensure biological interpretability and to avoid combining variables that are directly derived from one another. Specifically, we excluded time of first and last meal, as they are used to calculate nighttime fasting, and instead included: nighttime fasting, \u0026nbsp; \u0026nbsp; \u0026nbsp;wakeup-to-firstmeal, lastmeal-to-bedtime, eating jetlag, and number of eating occasions. The same gradual five levels of adjustment as described before were used.\u003c/p\u003e\n\u003cp\u003eAll statistical analyses were performed using R studio version 4.5.0, statistical tests were considered significant when p \u0026lt; 0.05.\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003eOverall, 6858 participants (58.6% women) with a median age of 52 (IQR 47\u0026ndash;58) years at baseline were included, of whom 152 developed T2D between 2018 and 2024. Baseline characteristics are shown in Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e for the overall population and stratified by T2D case status, and in Extended Data Table \u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e stratified by sex. The prevalence of pre-existing CVD was of 32% and much higher among participants who developed T2D during follow-up compared with those who did not (53% vs. 31%), as was BMI (median 30.0 [IQR: 28\u0026ndash;34 vs 26 [24, 29] kg/m\u003csup\u003e2\u003c/sup\u003e). Regarding chrononutrition, the median time of the first meal of the day was at 8:34 (IQR 7:47\u0026thinsp;\u0026minus;\u0026thinsp;9:30) while the last meal of the day was typically eaten late in the evening, with a median time of 21:08 (IQR 20:49\u0026thinsp;\u0026minus;\u0026thinsp;21:38), resulting in a median nighttime fasting duration of 11:26 (10:34\u0026thinsp;\u0026minus;\u0026thinsp;12:22) hours. Figure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e shows the correlations between chrononutrition variables where the most correlated variables were the time of first meal and nighttime fasting duration, and the time of first meal with wakeup-to-firstmeal (\u0026rho;\u0026thinsp;=\u0026thinsp;0.82 and 0.74 respectively). The distribution of all the chrononutrition variables recorded in the FFQ are reported in Extended Data (Extended Data Table \u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eBaseline characteristics of study population overall and stratified by T2D\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003elevel\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003eOverall\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\n \u003cp\u003eIncident T2D\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003etest\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003cth align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003en\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e6858\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e6706\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e152\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSex (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eWomen\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e4017 (58.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e3950 (58.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e67 (44.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eAge (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e52 [47, 58]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e52 [47, 58]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e57 [51, 61]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eEducation (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003ePrimary education\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1322 (19.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e1287 (19.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e35 (23.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eSecondary education\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e2102 (30.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e2043 (30.5)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e59 (38.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eHigher education\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e3434 (50.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e3376 (50.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e58 (38.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eCVD risk factor (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e2170 (31.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e2089 (31.2)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e81 (53.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSmoking habit (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eSmoker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1013 (14.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e988 (14.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e25 (16.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eFormer smoker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e3023 (44.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e2941 (43.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e82 (53.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eNever smoker\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e2822 (41.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e2777 (41.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e45 (29.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003ePhysical activity (METs/week) (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e62 [40, 97]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e63 [40, 98]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e57 [35, 90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.107\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eBMI (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e26.13 [23.73, 29.08]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e26.06 [23.69, 28.96]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e30.09 [27.93, 34.19]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eEnergy intake (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1942 [1581, 2363]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e1939 [1580, 2363]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e2016 [1657, 2427]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.064\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003erMED score categories (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eLow adherence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1603 (23.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e1566 (23.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e37 (24.3)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.264\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eMedium adherence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e3276 (47.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e3196 (47.7)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e80 (52.6)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eHigh adherence\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1979 (28.9)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e1944 (29.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e35 (23.0)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e% UPF (weight) (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e8.15 [4.94, 13.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e8.12 [4.94, 13.00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e9.46 [5.78, 14.49]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.025\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eSleeping midpoint (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e3:30 [3:04, 3:56]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e3:30 [3:04, 3:56]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e3:40 [3:10, 4:03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.027\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eN of eating occasions (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e3.0 [3.0, 3.7]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e3.00 [3.00, 3.71]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e3.00 [3.00, 3.71]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.672\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTime of first meal (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e8:34 [7:47, 9:30]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e8:34 [7:47, 9:30]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e8:34 [7:58, 9:26]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.903\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTime of first meal category (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eBefore 8:30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e3344 (48.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e3272 (48.8)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e72 (47.4)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.791\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTime of last meal (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e21:08 [20:49, 21:38]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e21:08 [20:47, 21:38]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e21:17 [21:00, 21:58]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eNighttime fasting (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e11:26 [10:34, 12:22]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e11:26 [10:34, 12:22]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e11:16 [10:26, 12:00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.037\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eEating jetlag (median [IQR]) (minutes)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e40 [15, 67]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e40 [15, 67]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e40 [15, 60]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.334\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTime wakeup to first meal (median [IQR]) (min)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e52 [22, 128]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e52 [22, 128]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e52 [20, 108]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eTime last meal to bedtime (median [IQR])\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e2:27 [2:00, 3:00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e2:28 [2:00, 3:00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e2:53 [1:49, 3:00]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.361\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eEating midpoint (median [IQR]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e14.89 [14.39, 15.43]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e14.89 [14.39, 15.43]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e15.04 [14.50, 15.46]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\n \u003cp\u003e0.107\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\n \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003cbr\u003e\u003c/div\u003e\n \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003cbr\u003e\u003c/div\u003e\u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eAssociations between chrononutrition variables and T2D estimated using logistic regression.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eExposure\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eModel 0\u003c/p\u003e\n \u003cp\u003eOR (95% CI)\u003c/p\u003e\n \u003cp\u003eSingle exposure\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003eModel 4\u003c/p\u003e\n \u003cp\u003eOR (95% CI)\u003c/p\u003e\n \u003cp\u003eSingle exposure\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003eModel 0\u003c/p\u003e\n \u003cp\u003eOR (95% CI)\u003c/p\u003e\n \u003cp\u003eMutually adjusted\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003eModel 4\u003c/p\u003e\n \u003cp\u003eOR (95% CI)\u003c/p\u003e\n \u003cp\u003eMutually adjusted\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eTime of first meal\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.94 [0.83, 1.05]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.84 [0.72, 0.96]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e0.90 [0.79, 1.02]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e0.82 [0.71, 0.95]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eMeal time model\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eTime of last meal\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.43 [1.16, 1.76]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1.29 [1.05, 1.61]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e1.48 [1.19, 1.85]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e1.34 [1.07, 1.69]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eN of eating occasions\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.19 [0.88, 1.62]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e1.21 [0.89, 1.64]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e0.98 [0.71, 1.36]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e0.96 [0.68, 1.35]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eNighttime fasting\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.84 [0.75, 0.95]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.80 [0.71, 0.91]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e0.78 [0.68, 0.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e0.78 [0.68, 0.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\" morerows=\"2\" rowspan=\"3\"\u003e\n \u003cp\u003e\u003cstrong\u003eNighttime fasting model\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eEating midpoint\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e1.07 [0.89, 1.27]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.92 [0.75, 1.12]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e1.33 [1.06, 1.67]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e1.10 [0.86, 1.42]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eN of eating occasions\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e0.98 [0.71, 1.36]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e0.96 [0.68, 1.35]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eWakeup-to-first meal\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.90 [0.78, 1.01]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.85 [0.73, 0.97]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e0.91 [0.79, 1.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e0.86 [0.74, 0.98]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eWake up to first meal model\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eN of eating occasions\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e1.12 [0.81, 1.55]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e1.10 [0.79, 1.52]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eLastmeal-to-bedtime\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.88 [0.74, 1.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.81 [0.66, 0.99]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e0.90 [0.75, 1.07]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e0.82 [0.67, 1.01]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eLast meal to bedtime model\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eN of eating occasions\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e1.14 [0.83, 1.56]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e1.13 [0.82, 1.54]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eEating jetlag\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.96 [0.76, 1.20]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.98 [0.77, 1.23]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e0.99 [0.77, 1.23]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e1.01 [0.78, 1.27]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\n \u003cp\u003e\u003cstrong\u003eEating jetlag model\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eN of eating occasions\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\n \u003cp\u003e1.19 [0.87, 1.62]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\n \u003cp\u003e1.21 [0.88, 1.65]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"6\"\u003eModel 0 is adjusted for age and sex; Model 4 is adjusted for age and sex, education, smoking habits, physical activity, Mediterranean diet, energy intake and ultra-processed food intake, sleeping midpoint, BMI and presence of CVD. The results for the number of eating occasions in the single-exposure model are reported only once, within the mealtime model.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cem\u003eLogistic regression and non-linearity\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eSpline models did not reveal evidence of non-linear associations with T2D for any chrononutrition variable (Extended Data Fig.\u0026nbsp;2), confirmed by the non-significant likelihood ratio tests (Extended Data Fig.\u0026nbsp;2).\u003c/p\u003e\n\u003cp\u003eTable \u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e shows the results of the single exposure and mutually adjusted models, with Model 0 used to assess the base associations, and Model 4 representing the best fitting model according to the AIC. Results from intermediate adjustment models are provided in Extended Data Table \u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eAssociation between the mixture of chrononutrition variables and T2D risk. Quantile g-computation OR and 95% CI of the overall mixture and individual weights (the relative contribution of each exposure to the mixture effect in a particular direction) for each chrononutrition variable.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003eModel 0\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003eModel 1\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003eModel 2\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003eModel 3\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003eModel 4\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eOR overall (95% CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e0.73 [0.50, 1.07]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e0.70 [0.48, 1.03]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e0.66 [0.45, 0.98]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e0.61 [0.41, 0.91]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e0.61 [0.41, 0.90]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eChrononutrition variables individual weights\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eNighttime fasting\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-0.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-0.55\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e-0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e-0.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e-0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eWake up to first meal\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e+\u0026thinsp;1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e+\u0026thinsp;1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e+\u0026thinsp;0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e+\u0026thinsp;1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e+\u0026thinsp;1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eLast meal to bedtime\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-0.17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-0.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e-0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e-0.25\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e-0.23\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eN of eating occasions\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e-0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e-0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e-0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\" colname=\"c1\"\u003e\n \u003cp\u003e\u003cstrong\u003eEating jetlag\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c2\"\u003e\n \u003cp\u003e-0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c3\"\u003e\n \u003cp\u003e-0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c4\"\u003e\n \u003cp\u003e+\u0026thinsp;0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c5\"\u003e\n \u003cp\u003e-0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\" colname=\"c6\"\u003e\n \u003cp\u003e-0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eIn mutually adjusted models, later first meal was associated with a lower risk of developing T2D (model 4 OR for 1 hour increase 0.82 [95% CI: 0.71, 0.95]), whereas a later last meal was associated with an increased risk (model 4 OR 1.34 [95% CI: 1.07, 1.69]). These associations remained significant across models with different sets of covariates and in single-exposure analyses. A longer nighttime fasting duration was significantly associated with a reduced risk of T2D across all models. In particular, model 4, further adjusted for eating midpoint and number of eating occasions, showed an OR of 0.78 [95% CI: 0.68, 0.90] for 1 hour increase in nighttime fasting.\u003c/p\u003e\n\u003cp\u003eA longer wakeup-to-firstmeal duration was associated with a reduced risk of T2D in models 3 and 4 (OR 0.86 [95% CI: 0.74, 0.98] and OR 0.85 [95% CI: 0.73, 0.97] respectively), as well as in the single-exposure model. Likewise, longer lastmeal-to-bedtime was associated with a reduced risk of T2D, albeit with a smaller effect size and greater uncertainty (single exposure model 4 OR 0.81 [95% CI: 0.66, 0.99], mutually adjusted model 4 OR 0.82 [95% CI: 0.67, 1.01]).\u003c/p\u003e\n\u003cp\u003eEating jetlag and the number of eating occasions were not associated with T2D risk in any model in the overall population.\u003c/p\u003e\n\u003cp\u003eStratified results show little differences across sex (Extended Data Table\u0026nbsp;4), and a longer nighttime fasting showed a slightly stronger negative association with T2D among participants eating their first meal after 8:30 (Extended Data Table\u0026nbsp;5). There was no significant interaction between nighttime fasting and energy intake nor UPF intake.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eQuantile g-computation\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eThe mixture model simultaneously considered all chrononutrition variables, estimating on the one hand their joint effect on T2D risk, and on the other hand the relative contribution of each variable to the overall association (weights) (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). A one-IQR increase across all variables was associated with a decrease in the T2D risk (model 4 OR 0.61 [95% CI: 0.41, 0.90]). In particular, the association was protective in all the four models adjusted for an increased number of confounders, and resulted stronger in models from 2 to 4.\u003c/p\u003e\n\u003cp\u003eMoreover, in all the four models, nighttime fasting had the biggest role in contributing to the decrease of risk. A higher number of eating occasions and a longer lastmeal-to-bedtime also carried substantial protective weight. In contrast, a longer wakeup-to-firstmeal showed a positive weight on risk, though this did not offset the overall protective association of the mixture. Eating jetlag had only minimal weight, with inconsistent direction across models.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study provides new insights into the role of chrononutritional behaviors in the development of T2D. To the best of our knowledge, this is the first study to combine both the widely used single-exposure logistic regression and a novel mixture method to analyze the joint effects of chrononutrition variables. Our findings consistently indicated that longer nighttime fasting, achieved through both a later first meal and an earlier last meal, were associated with a lower risk of incident T2D. This suggests that fasting duration may be a key characteristic of chrononutritional patterns relevant to metabolic health. Less consistent association across models were observed for time elapsed between wakeup-to-firstmeal, as well as between the lastmeal-to-bedtime, where longer intervals were associated with lower T2D risk. Taken together, these findings support the hypothesis that the duration of the fasting window, rather than the meal timing itself, may be the key element driving these associations.\u003c/p\u003e \u003cp\u003eOur results are consistent with Carew et al.\u003csup\u003e14\u003c/sup\u003e who observed a protective association for later first meals (after 9AM) in a prospective study of a US population. In contrast, Palomar-Cros et al.\u003csup\u003e11\u003c/sup\u003e reported that later first meals were associated with a higher risk of T2D in a French cohort, and a protective effect of nighttime fasting was only evident when the first meal happened before 8AM. Although their cohort study is the only one exploring nighttime fasting in relation to T2D, it differs from ours in participant age and chrononutrition behaviors, particularly the time of last meal and the number of eating occasions. The higher risk of T2D associated with a later last meal observed in our study aligns with the hypothesis that late dinners may shorten the overnight fasting window and contribute to circadian misalignment. Controlled trials showed that late-day carbohydrate intake impairs glycemic control\u003csup\u003e3\u003c/sup\u003e, and Verde et al.\u003csup\u003e6\u003c/sup\u003e and L\u0026oacute;pez-Prieto\u003csup\u003e4\u003c/sup\u003e et al. highlighted the role of late-night eating in circadian misalignment and metabolic disruption. Moreover, late eating has been associated to obesity and poor glucose tolerance in prediabetic subjects\u003csup\u003e15,16\u003c/sup\u003e. Our findings thus reinforce these results and support growing concerns about evening eating patterns in modern societies.\u003c/p\u003e \u003cp\u003eA longer fasting window has been suggested to be metabolically advantageous, since extended fasting intervals promote ketogenesis, enhance lipolysis, and reduce hepatic glucose output, improving insulin sensitivity and lowering inflammation\u003csup\u003e17\u003c/sup\u003e. They may also restore the diurnal rhythm of hormones such as cortisol, leptin, and melatonin, which are often blunted by irregular or late-night eating\u003csup\u003e18\u003c/sup\u003e; this interaction has not been fully explored in previous studies. Longer fasting (also known as time restricted eating) is increasingly being investigated as a dietary strategy for the prevention and therapy of glucose and lipid metabolic disturbances. It has generally been shown to lower fasting and post-meal glucose, improve 24-hour glucose control, reduce fasting insulin, and enhance insulin sensitivity\u003csup\u003e19\u003c/sup\u003e. However, some trials show no benefit or even negative effects\u003csup\u003e19\u003c/sup\u003e. These mixed results likely stem from differences in study design, such as eating window length, timing, calorie changes, and intervention duration, as well as variations in participant characteristics like metabolic health, age, sex, and chronotype\u003csup\u003e20\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eStudies in animal models have shown that longer fasting can prevent and even reverse T2D through multiple molecular mechanisms. In mice and rats fed high-fat diets, restricting feeding time (e.g., 8 hours feeding, 16 hours fasting) improves insulin sensitivity, reduces systemic inflammation, and prevents hyperglycemia without reducing total caloric \u003csup\u003e21\u003c/sup\u003e. These effects are linked to increased efficiency of insulin receptor signaling, AMPK activation, reduction of the mTOR pathway, increased mitochondrial biogenesis, and stimulation of autophagy. Furthermore, intermittent fasting increases the expression of BDNF and CREB, factors involved in regulating energy metabolism and cell protection, promoting the preservation of pancreatic β cells and glycemic homeostasis\u003csup\u003e22\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe observed no significant associations for eating jetlag, differing from other studies that suggest metabolic dysregulation with irregular eating schedules\u003csup\u003e23,24\u003c/sup\u003e. This discrepancy may be due to limited variability or measurement sensitivity in our cohort.\u003c/p\u003e \u003cp\u003eWe also found no association between the number of eating occasions and the risk of T2D. However, evidence from literature showed that a higher number of meals was associated with a reduced risk of T2D\u003csup\u003e11\u003c/sup\u003e and this may be explained by a reduced serum insulin and lipid concentrations between meals\u003csup\u003e25\u0026ndash;27\u003c/sup\u003e. Conversely, an observational study found that additional snacks beyond the three main meals were associated with an increased risk of T2D\u003csup\u003e28\u003c/sup\u003e. The lack of findings in our population may reflect the lack of variability in the number of meals compared to other studies.\u003c/p\u003e \u003cp\u003eThe mixture analysis largely confirmed results from the logistic models, showing an overall protective association of the combined set of chrononutritional behaviors with T2D risk. In particular, nighttime fasting had the highest negative weight followed by lastmeal-to-bedtime. Wakeup-to-firstmeal had a positive weight in the mixture, although its influence on the overall estimate appeared limited, as the mixture OR remained consistently protective across models adjustments. Interestingly, the number of eating occasions received a negative weight in the mixture, despite showing no clear association in logistic models. These observations suggest the presence of potential synergistic relationships among correlated chrononutritional behaviors that may not be fully captured when exposures are assessed in single-exposure models. Cluster analyses are often used to identify and describe groups of individuals with similar chrononutritional patterns, but the derived groups depend on data-driven decisions, may vary across populations, and do not allow estimation of the risk associated with modifying the variables\u003csup\u003e29\u003c/sup\u003e. In contrast, quantile g-computation is a quantitative method that provides an estimate of the overall effect of the mixture, identifies the relative contribution of each exposure, accounts for multicollinearity and allows insights into potential synergistic effects\u003csup\u003e30\u003c/sup\u003e. Overall, few epidemiological studies have examined these synergistic patterns, and our results illustrate how mixture methods, in particular quantile g-computation, can offer a valuable methodological option for further investigating complex behavioral exposures. We recognize in our study the correlation between chrononutrition variables were not extremely high, but by introducing this analytic strategy to the field, our work offers a practical guideline for researchers seeking methods that more appropriately reflect the multidimensional nature of temporal eating patterns.\u003c/p\u003e \u003cp\u003eThe findings of this study underscore the importance of incorporating chrononutrition into public health strategies for T2D prevention. While traditional nutritional guidance has focused primarily on dietary quantity and quality, our results highlight that \u003cem\u003ewhen\u003c/em\u003e people eat may be relevant for metabolic health. In particular, the consistent protective association observed for longer nighttime fasting suggests that aligning food intake with endogenous circadian rhythms could offer a practical, low-cost behavioral target for diabetes prevention. These insights have potential implications for dietary guidelines, workplace health programs, and clinical counseling, particularly in Mediterranean populations where late-night eating is culturally embedded. Public health interventions that promote earlier cessation of food intake, reduce eating close to bedtime, and encourage more regular temporal patterns of intake may help reduce T2D incidence at the population level. As chrononutrition behaviors are modifiable, integrating temporal eating recommendations into existing nutritional policies could represent a scalable and equitable approach to mitigate the growing global burden of T2D.\u003c/p\u003e \u003cp\u003eOf note, some limitations of our study should be acknowledged. First, given the observational design, causality cannot be inferred. Meal timing data were self-reported and collected at a single time point, and eating occasions were limited to 5 possible meals, which may have introduced misclassification bias, recall bias and rounding errors. Moreover, the questionnaire did not allow us to assess the energy intake distribution throughout the day, factor that may be relevant for glucose metabolism and T2D risk. The relatively small number of incident cases of T2D (152) has also limited the power of stratified analyses. Finally, generalizability may be limited to similar adult populations in Mediterranean countries.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eOur results indicate that a longer nighttime fasting duration is associated with a reduced risk of developing T2D, and that chrononutrition variables may act synergistically in ways that are not captured by single-exposure models. The implementation of quantile g-computation can contribute to a better understanding of these complex and interrelated behaviors. Integrating temporal eating recommendations into public health dietary guidance might represent a simple and widely applicable strategy to help reduce the burden of T2D at the population level.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eChatGPT-5, a language model developed by OpenAI, was used to assist with improving the grammar and flow of certain paragraphs of the manuscript. The content was originally written by the authors, and ChatGPT was used solely for enhancement purposes, not for generating new content or ideas.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data that support the findings of this study are not publicly available due to patient privacy protections and ethical restriction but are available from the corresponding author upon reasonable request and with appropriate approvals.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode availability:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe code and a sample database are available at https://github.com/joanall\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e E.G. gratefully acknowledge the support of grant JDC2024-055407-I, funded by MICIU/AEI/10.13039/501100011033 and the ESF+. Camille Lassale is supported by a Ramon y Cajal Fellowship RYC2020-029599 funded by MICIU/AEI/10.13039/501100011033 and the European Social Fund \u0026ldquo;Invest in your future\u0026rdquo;. This project was funded by the Fundaci\u0026oacute; La Marat\u0026oacute;TV3 \u0026ndash; Cardiovascular disease (grant agreement 202323-30-31, CUPID project). This study makes use of data generated by the GCAT-Genomes for Life. Cohort study of the Genomes of Catalonia, Fundacio IGTP. IGTP is part of the CERCA Program / Generalitat de Catalunya. GCAT was funded by Acci\u0026oacute;n de Dinamizaci\u0026oacute;n del ISCIII-MINECO and the Ministry of Health of the Generalitat of Catalunya (ADE 10/00026); and have additional suport by the Ag\u0026egrave;ncia de Gesti\u0026oacute; d\u0026rsquo;Ajuts Universitaris i de Recerca (AGAUR) (SGR 01537), Spanish National Grant PI18/01512, TED2021-130626B-I00, La MaratoTV3 167/C/2021, VEIS project (001-P-001647) (co-funded by European Regional Development Fund (ERDF), \u0026ldquo;A way to build Europe\u0026rdquo;), the European Union under grant agreement no. 101046314 (END-VOC). The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026rsquo; contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll authors contributed to the study conception and design. Data acquisition was performed by RdC, MK. Data curation was performed by EG, JL-P, CR, MO, CL, AP-C. Funding acquisition was carried by RdC, MK. The formal analysis was performed by EG, JL-P, CR. The supervision of the present study was done by CL. The first draft of the manuscript was written by EG and CR and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. The work reported in the paper has been performed by the authors, unless clearly specified in the text.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eInternational Diabetes Federation (IDF), I. D. F. (IDF). I. D. A., 11th Edition. Brussels, Belgium: IDF; 2025. https://diabetesatlas.org. \u003cem\u003eIDF Diabetes Atlas\u003c/em\u003e. (Brussels, Belgium, 2025).\u003c/li\u003e\n \u003cli\u003eReytor-Gonz\u0026aacute;lez, C. \u003cem\u003eet al.\u003c/em\u003e Chrononutrition and Energy Balance: How Meal Timing and Circadian Rhythms Shape Weight Regulation and Metabolic Health. \u003cem\u003eNutrients\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 2135 (2025).\u003c/li\u003e\n \u003cli\u003eHenry, C. J., Kaur, B. \u0026amp; Quek, R. Y. C. Chrononutrition in the management of diabetes. \u003cem\u003eNutr. Diabetes\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 6 (2020).\u003c/li\u003e\n \u003cli\u003eL\u0026oacute;pez-Prieto, R. S. \u003cem\u003eet al.\u003c/em\u003e Time Matters: An Insight into the Relationship Between Chrononutrition and Diabetes. \u003cem\u003eRev. Investig. Cl\u0026iacute;nica\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e, 80\u0026ndash;90 (2024).\u003c/li\u003e\n \u003cli\u003eKim, Y.-I., Kim, E., Lee, Y. \u0026amp; Park, J. Role of late-night eating in circadian disruption and depression: a review of emotional health impacts. \u003cem\u003ePhys. Act. Nutr.\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 18\u0026ndash;24 (2025).\u003c/li\u003e\n \u003cli\u003eVerde, L. \u003cem\u003eet al.\u003c/em\u003e Chrononutrition in type 2 diabetes mellitus and obesity: A narrative review. \u003cem\u003eDiabetes Metab. Res. Rev.\u003c/em\u003e \u003cstrong\u003e40\u003c/strong\u003e, e3778 (2024).\u003c/li\u003e\n \u003cli\u003eO\u0026rsquo;Connor, S. G. \u003cem\u003eet al.\u003c/em\u003e Conceptualization and Assessment of 24-H Timing of Eating and Energy Intake: A Methodological Systematic Review of the Chronic Disease Literature. \u003cem\u003eAdv. Nutr.\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 100178 (2024).\u003c/li\u003e\n \u003cli\u003eKnutson, K. L. \u003cem\u003eet al.\u003c/em\u003e Role of Circadian Health in Cardiometabolic Health and Disease Risk: A Scientific Statement From the American Heart Association. \u003cem\u003eCirculation\u003c/em\u003e \u003cstrong\u003e152\u003c/strong\u003e, e408\u0026ndash;e419 (2025).\u003c/li\u003e\n \u003cli\u003eBuscemi, C. \u003cem\u003eet al.\u003c/em\u003e The impact of breakfast skipping on plasma glucose levels in non-diabetic individuals: gender-based differences and implications. \u003cem\u003eInt. J. Food Sci. Nutr.\u003c/em\u003e \u003cstrong\u003e76\u003c/strong\u003e, 203\u0026ndash;208 (2025).\u003c/li\u003e\n \u003cli\u003eMinari, T. P. \u0026amp; Pisani, L. P. Skipping breakfast and its wide-ranging health consequences: A systematic review from multiple metabolic disruptions to socioeconomic factors. \u003cem\u003eNutr. Res.\u003c/em\u003e \u003cstrong\u003e141\u003c/strong\u003e, 34\u0026ndash;45 (2025).\u003c/li\u003e\n \u003cli\u003ePalomar-Cros, A. \u003cem\u003eet al.\u003c/em\u003e Associations of meal timing, number of eating occasions and night-time fasting duration with incidence of type 2 diabetes in the NutriNet-Sant\u0026eacute; cohort. \u003cem\u003eInt. J. Epidemiol.\u003c/em\u003e \u003cstrong\u003e52\u003c/strong\u003e, 1486\u0026ndash;1497 (2023).\u003c/li\u003e\n \u003cli\u003eSantonja, I. \u003cem\u003eet al.\u003c/em\u003e Meal-timing patterns and chronic disease prevalence in two representative Austrian studies. \u003cem\u003eEur. J. Nutr.\u003c/em\u003e \u003cstrong\u003e62\u003c/strong\u003e, 1879\u0026ndash;1890 (2023).\u003c/li\u003e\n \u003cli\u003ePons-Muzzo, L. \u003cem\u003eet al.\u003c/em\u003e Sex-specific chrono-nutritional patterns and association with body weight in a general population in Spain (GCAT study). \u003cem\u003eInt. J. Behav. Nutr. Phys. Act.\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 102 (2024).\u003c/li\u003e\n \u003cli\u003eCarew, A. S. \u003cem\u003eet al.\u003c/em\u003e Prospective study of breakfast frequency and timing and the risk of incident type 2 diabetes in community-dwelling older adults: the Cardiovascular Health Study. \u003cem\u003eAm. J. Clin. Nutr.\u003c/em\u003e \u003cstrong\u003e116\u003c/strong\u003e, 325\u0026ndash;334 (2022).\u003c/li\u003e\n \u003cli\u003eMcHill, A. W. \u003cem\u003eet al.\u003c/em\u003e Later circadian timing of food intake is associated with increased body fat\u0026dagger;. \u003cem\u003eAm. J. Clin. Nutr.\u003c/em\u003e \u003cstrong\u003e106\u003c/strong\u003e, 1213\u0026ndash;1219 (2017).\u003c/li\u003e\n \u003cli\u003eD\u0026iacute;az-Rizzolo, D. A. \u003cem\u003eet al.\u003c/em\u003e Late eating is associated with poor glucose tolerance, independent of body weight, fat mass, energy intake and diet composition in prediabetes or early onset type 2 diabetes. \u003cem\u003eNutr. Diabetes\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 90 (2024).\u003c/li\u003e\n \u003cli\u003eRuppert, P. M. M. \u0026amp; Kersten, S. Mechanisms of hepatic fatty acid oxidation and ketogenesis during fasting. \u003cem\u003eTrends Endocrinol. Metab.\u003c/em\u003e \u003cstrong\u003e35\u003c/strong\u003e, 107\u0026ndash;124 (2024).\u003c/li\u003e\n \u003cli\u003eKim, Y.-I., Kim, E., Lee, Y. \u0026amp; Park, J. Role of late-night eating in circadian disruption and depression: a review of emotional health impacts. \u003cem\u003ePhys. Act. Nutr.\u003c/em\u003e \u003cstrong\u003e29\u003c/strong\u003e, 018\u0026ndash;024 (2025).\u003c/li\u003e\n \u003cli\u003eChen, Y.-E., Tsai, H.-L., Tu, Y.-K. \u0026amp; Chen, L.-W. Effects of timing and eating duration of time restricted eating on metabolic outcomes: systematic review and network meta-analysis. \u003cem\u003eBMJ Med.\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e, (2026).\u003c/li\u003e\n \u003cli\u003eSchuppelius, B., Peters, B., Ottawa, A. \u0026amp; Pivovarova-Ramich, O. Time Restricted Eating: A Dietary Strategy to Prevent and Treat Metabolic Disturbances. \u003cem\u003eFront. Endocrinol.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, (2021).\u003c/li\u003e\n \u003cli\u003eBelkacemi, L. \u003cem\u003eet al.\u003c/em\u003e Intermittent Fasting Modulation of the Diabetic Syndrome in Streptozotocin-Injected Rats. \u003cem\u003eInt. J. Endocrinol.\u003c/em\u003e \u003cstrong\u003e2012\u003c/strong\u003e, 1\u0026ndash;12 (2012).\u003c/li\u003e\n \u003cli\u003eMattson, M. P., Longo, V. D. \u0026amp; Harvie, M. Impact of intermittent fasting on health and disease processes. \u003cem\u003eAgeing Res. Rev.\u003c/em\u003e \u003cstrong\u003e39\u003c/strong\u003e, 46\u0026ndash;58 (2017).\u003c/li\u003e\n \u003cli\u003eMakarem, N. \u003cem\u003eet al.\u003c/em\u003e Variability in Daily Eating Patterns and Eating Jetlag Are Associated With Worsened Cardiometabolic Risk Profiles in the American Heart Association Go Red for Women Strategically Focused Research Network. \u003cem\u003eJ. Am. Heart Assoc.\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, e022024 (2021).\u003c/li\u003e\n \u003cli\u003eChen, Y.-E. \u003cem\u003eet al.\u003c/em\u003e Associations of \u0026gt;1-h compared with 1-h meal timing variability (eating jetlag) with plasma glycemic parameters and continuous glucose monitoring measures among pregnant females: a prospective cohort study. \u003cem\u003eAm. J. Clin. Nutr.\u003c/em\u003e \u003cstrong\u003e122\u003c/strong\u003e, 244\u0026ndash;254 (2025).\u003c/li\u003e\n \u003cli\u003eJenkins, D. J. A. \u003cem\u003eet al.\u003c/em\u003e Nibbling versus Gorging: Metabolic Advantages of Increased Meal Frequency. \u003cem\u003eThe New England Journal of Medicine\u003c/em\u003e https://www.nejm.org/doi/abs/10.1056/NEJM198910053211403 (1989) doi:10.1056/NEJM198910053211403.\u003c/li\u003e\n \u003cli\u003eBertelsen, J. \u003cem\u003eet al.\u003c/em\u003e Effect of Meal Frequency on Blood Glucose, Insulin, and Free Fatty Acids in NIDDM Subjects. \u003cem\u003eDiabetes Care\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 4\u0026ndash;7 (1993).\u003c/li\u003e\n \u003cli\u003eTitan, S. M. O. \u003cem\u003eet al.\u003c/em\u003e Frequency of eating and concentrations of serum cholesterol in the Norfolk population of the European prospective investigation into cancer (EPIC-Norfolk): cross sectional study. \u003cem\u003eBMJ\u003c/em\u003e \u003cstrong\u003e323\u003c/strong\u003e, 1286 (2001).\u003c/li\u003e\n \u003cli\u003eMekary, R. A., Giovannucci, E., Willett, W. C., van Dam, R. M. \u0026amp; Hu, F. B. Eating patterns and type 2 diabetes risk in men: breakfast omission, eating frequency, and snacking1234. \u003cem\u003eAm. J. Clin. Nutr.\u003c/em\u003e \u003cstrong\u003e95\u003c/strong\u003e, 1182\u0026ndash;1189 (2012).\u003c/li\u003e\n \u003cli\u003eZhao, J. \u003cem\u003eet al.\u003c/em\u003e A review of statistical methods for dietary pattern analysis. \u003cem\u003eNutr. J.\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 37 (2021).\u003c/li\u003e\n \u003cli\u003eKeil, A. P. \u003cem\u003eet al.\u003c/em\u003e A Quantile-Based g-Computation Approach to Addressing the Effects of Exposure Mixtures. \u003cem\u003eEnviron. Health Perspect.\u003c/em\u003e \u003cstrong\u003e128\u003c/strong\u003e, 47004 (2020).\u003c/li\u003e\n \u003cli\u003eOb\u0026oacute;n-Santacana, M. \u003cem\u003eet al.\u003c/em\u003e GCAT|Genomes for life: a prospective cohort study of the genomes of Catalonia. \u003cem\u003eBMJ Open\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, e018324 (2018).\u003c/li\u003e\n \u003cli\u003eHoeymans, N., Garssen, A. A., Westert, G. P. \u0026amp; Verhaak, P. F. Measuring mental health of the Dutch population: a comparison of the GHQ-12 and the MHI-5. \u003cem\u003eHealth Qual. Life Outcomes\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 23 (2004).\u003c/li\u003e\n \u003cli\u003eFern\u0026aacute;ndez-Ballart, J. D. \u003cem\u003eet al.\u003c/em\u003e Relative validity of a semi-quantitative food-frequency questionnaire in an elderly Mediterranean population of Spain. \u003cem\u003eBr. J. Nutr.\u003c/em\u003e \u003cstrong\u003e103\u003c/strong\u003e, 1808\u0026ndash;1816 (2010).\u003c/li\u003e\n \u003cli\u003eSaulle, R., Bernardi, M., Chiarini, M., Backhaus, I. \u0026amp; La Torre, G. Shift work, overweight and obesity in health professionals: a systematic review and meta-analysis. \u003cem\u003eClin. Ter.\u003c/em\u003e \u003cstrong\u003e169\u003c/strong\u003e, e189\u0026ndash;e197 (2018).\u003c/li\u003e\n \u003cli\u003eBuckland, G. \u003cem\u003eet al.\u003c/em\u003e Adherence to a Mediterranean diet and risk of gastric adenocarcinoma within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort study. \u003cem\u003eAm. J. Clin. Nutr.\u003c/em\u003e \u003cstrong\u003e91\u003c/strong\u003e, 381\u0026ndash;390 (2010).\u003c/li\u003e\n \u003cli\u003eMartinez-Steele, E. \u003cem\u003eet al.\u003c/em\u003e Best practices for applying the Nova food classification system. \u003cem\u003eNat. Food\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 445\u0026ndash;448 (2023).\u003c/li\u003e\n \u003cli\u003eYu, L. \u003cem\u003eet al.\u003c/em\u003e A review of practical statistical methods used in epidemiological studies to estimate the health effects of multi-pollutant mixture. \u003cem\u003eEnviron. Pollut.\u003c/em\u003e \u003cstrong\u003e306\u003c/strong\u003e, 119356 (2022).\u003c/li\u003e\n \u003cli\u003eLiu, Y. \u003cem\u003eet al.\u003c/em\u003e Synergistic Effects of Nutrients on Musculoskeletal Health in Gerontology: Understanding the Combined Impact of Macronutrients and Micronutrients. \u003cem\u003eNutrients\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 1640 (2024).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Chrono-nutrition, meal timing, circadian rhythm, diabetes, obesity, cardiovascular disease, epidemiology","lastPublishedDoi":"10.21203/rs.3.rs-9107412/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9107412/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe timing of food intake interacts with circadian biology and may influence metabolic disease risk, yet correlations among chrononutritional behaviours complicate their joint evaluation. In a Spanish prospective cohort of 6,858 adults aged 40-65y, we examined associations between chrono-nutritional behaviors and incident type 2 diabetes (T2D) using regression models and mixture-based methods to assess both individual and joint effects. During a 6-year follow-up, 152 participants developed T2D. In single-exposure analyses, a later first meal (OR\u0026thinsp;=\u0026thinsp;0.82 per hour; 95%CI 0.71\u0026ndash;0.95) and longer nighttime fasting (OR\u0026thinsp;=\u0026thinsp;0.78 per hour; 0.68\u0026ndash;0.90) were associated with lower T2D risk, whereas a later last meal was associated with higher risk (OR\u0026thinsp;=\u0026thinsp;1.34 per hour; 1.07\u0026ndash;1.69). When modeled jointly, a one interquartile range increase across chrono-nutritional variables was associated with a significant lower T2D risk, with nighttime fasting contributing most strongly to the overall association. These findings highlight nighttime fasting as a potentially modifiable chrono-nutritional target for T2D prevention.\u003c/p\u003e","manuscriptTitle":"Chrononutrition and risk of type 2 diabetes: a prospective cohort study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-27 14:19:01","doi":"10.21203/rs.3.rs-9107412/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"communications-medicine","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"commsmed","sideBox":"Learn more about [Communications Medicine](http://www.nature.com/commsmed)","snPcode":"43856","submissionUrl":"https://mts-commsmed.nature.com/cgi-bin/main.plex","title":"Communications Medicine","twitterHandle":"@commsmedicine","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Communications Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4ffe8644-1bfe-4ad0-9c3e-c5684ffa3725","owner":[],"postedDate":"April 27th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":66543735,"name":"Health sciences/Endocrinology/Endocrine system and metabolic diseases/Diabetes/Type 2 diabetes"},{"id":66543736,"name":"Health sciences/Medical research/Epidemiology"},{"id":66543737,"name":"Health sciences/Health care/Disease prevention/Lifestyle modification"}],"tags":[],"updatedAt":"2026-04-27T14:19:01+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-27 14:19:01","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9107412","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9107412","identity":"rs-9107412","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}