Abstract
Objectives
This study examined the associations between actigraphy-derived RAR and meal timing. It also investigated whether these associations are modified by race/ethnicity.
Methods
We used data from the National Health and Nutrition Examination Survey 2013−-2014 (n=2,944). Twenty-four-hour rest-activity metrics (mean, amplitude, acrophase, robustness) were derived from actigraphy data via cosinor analysis. Continuous meal-timing exposures were mean-centered (weighted means) so that coefficients represented associations within the reference race/ethnicity group (Non-Hispanic White) at the sample mean. Models included race × meal-timing interaction terms to test whether associations between meal timing and acrophase differed by race/ethnicity. Significant interactions were further examined using race-stratified simple slopes, with β estimates, standard errors (SE), and 95% confidence intervals (CIs) reported.
Results
Significant interactions were observed for acrophase and last mealtime (p-interaction = 0.02). Later last meals were associated with later acrophase among Mexican American (β = 0.77 ± 0.28, p-interaction = 0.02) and Non-Hispanic Black (β = 0.40 ± 0.12, p-interaction = 0.02) adults. No other exposure–outcome pairs were significant (p > 0.05).
Conclusions
Significant race/ethnicity interactions indicate that Mexican American and Black adults with later last meals exhibited correspondingly later acrophase compared to White adults. Our findings suggest possible associations between meal timing and acrophase varying by race. These findings highlight potential racial/ethnic differences in circadian responses to meal timing, underscoring the need for culturally tailored measures in circadian nutrition research.
Keywords: circadian rhythms, rest-activity rhythm, meal timing, time-restricted eating, cosinor analysis, accelerometer
Introduction
Circadian rhythms (circa= around, dies= one day) are the expression of 24-hour internal clocks present in almost all tissues and organs of the human body under modulation by the external light–dark cycle [1]. The main zeitgeber (time cue) of the circadian system is light. However, nonphotic stimuli such as meal timing can also provide input to the circadian system. Monitoring of rest-activity rhythms (RARs) is a way of assessing circadian rhythmicity comprising activities such as sleep timing, sleep regularity, and physical activity occurring during a 24-hour period. The use of wearable devices such as accelerometers has made it possible to assess 24-hour RAR as a reliable marker of the circadian system [2]. Research in diverse population samples in the US has shown an association between disrupted RAR, as assessed via actigraphy, and markers of impaired metabolic health [3, 4].
Circadian rhythms play important roles in various functions of the human body, including the regulation of the sleep–wake cycle, feeding–fasting cycle, and temperature. When internal clocks become uncoupled from light–dark rhythms in the external environment, poor health can ensue [1]. Shift work and other lifestyle factors can alter the timing of sleep, meals, and exercise timing. When these behaviors do not align with internal central and peripheral clocks, misalignment can occur [5, 6].
Meal timing is influenced by the human circadian system, and irregular meal timing has been associated with adverse metabolic outcomes. However, given the cross-sectional nature of most of the existing literature, the direction of the existing association cannot be determined. Therefore, mealtimes that align with our internal clock are thought to be associated with optimal health [7]. Chrononutrition, an emerging science, describes the optimization of food timing according to the circadian phase to improve metabolic health [8–10]. Earlier meal timing has been linked to better metabolic outcomes, consistent with evidence that insulin sensitivity and glucose tolerance are lower later in the day, yielding larger postprandial glucose responses to evening carbohydrate intake [11–13]. Time-restricted eating (TRE) is a dietary pattern whereby eating is restricted to a certain time of day (~8–10 hours) without restricting caloric intake. The influence of meal timing on circadian rhythms is a rapidly expanding area of chrononutrition research, with growing evidence linking mistimed eating to adverse health outcomes. The benefits of TRE on body weight, blood pressure, and lipid and glucose levels have been documented in both human and animal studies [14–17].
Simultaneously, racial and ethnic disparities in health behaviors and outcomes are well-documented. For instance, previous research, including a study [18] utilizing National Health and Nutrition Examination Survey (NHANES) data revealed significant racial/ethnic differences in meal timing. However, a critical gap remains. It is unclear whether these observed variations in meal timing behavior directly contribute to circadian disruption across groups. This gap is particularly important due to the disproportionate burden of cardiometabolic diseases that are often experienced by racial and ethnic minority groups. Thus, using data from NHANES 2013--2014, the main aim of this study was to examine the association between actigraphy-derived RAR and meal timing. We also sought to investigate if these associations are modified by race/ethnicity. We hypothesized that 1) participants with later acrophase (i.e., later timing of peak activity in actigraphic RAR) would have later first and last mealtimes and longer daily eating window, and 2) racial and ethnic minority adults would have later acrophase and lower amplitude and robustness compared to White adults.
Methods
Study design and population
The data used for this study were obtained from NHANES 2013−-2014, a cross-sectional study of civilian noninstitutionalized US population. The NHANES study is conducted every two years and uses a complex, multistage probability design in which data are collected through interviews, physical examinations, and laboratory assessments. For the 2013–2014 cycle, we downloaded the Demographics (DEMO_H), 24-hour dietary recall Day 1 and Day 2 files (DR1IFF_H, DR2IFF_H), Body Measures (BMX_H), and Urine Pregnancy Test (UCPREG_H) data. Wrist accelerometer data were available only in the 2011–2012 and the 2013–2014 cycles; we selected the 2013–2014 because it was the most recent cycle with complete accelerometer and dietary recall data, allowing for the most current assessment of meal timing and rest–activity rhythms. Both accelerometer and dietary recall data were collected during the same cycle. A detailed description of the NHANES study has been published previously [19].
We included adults ≥ 20 years with complete actigraphy and mealtime data (calories between 600 and 8,000) and non-pregnant women. The actigraphy time series contained 10,080 time points each, which corresponds to 7 consecutive days with 1-minute resolution. Some time series were longer than 10,080 time points, and they were truncated to preserve equal sampling for all subjects. Participants without this data, as well as those with a sleep duration of less than 120 minutes or greater than 1,200 minutes, and those with unreliable diet recalls (DR1DRSTZ = 2 i.e., study participants’ 24-hour dietary recall was not reliable or complete) were excluded. A total of 2,348 participants were included in the final analysis. Figure 1 shows the flowchart of study participants. The study protocol for NHANES was approved by the Ethics Review Board of the National Center for Health Statistics. All study participants signed informed consent.
Figure 1: Flowchart of participants selection in NHANES 2013–2014.
A flow diagram illustrating the selection of the final analytic sample from the total participants of the NHANES 2013–2014 study. The diagram shows the initial number of participants and details the reasons and corresponding numbers for exclusion, including age, pregnancy status, unreliable dietary data, missing covariates, and actigraphy data. The diagram concludes with the final analytic sample size of eligible participants for the study.
Outcome and exposure assessment
The main outcome measures for this study were four RAR metrics (robustness, amplitude, acrophase, and mesor). The robustness assesses the overall strength of rhythmicity; amplitude measures (half) the difference between the peak and nadir of the fitted curve; and the acrophase is the timing of the peak activity. The mesor (mean of the rhythm) is the midline estimation statistic of the rhythm. The actigraphy data were analyzed by cosinor rhythmometry [20] using the cosinor.exe program [21] for the 7-day time series of each individual. The cosinor procedure uses a sinusoidal function to fit the waveform of the time series without assuming that the waveform of the rhythm is sinusoidal. A sinusoidal function is used to infer the phase of the oscillatory process with the understanding that, as demonstrated by Fourier 200 years ago, any oscillatory process can be described by a sum of sine and cosine waves [22]. The fundamental frequency (which is what the cosinor procedure provides) is the best estimate of the main component of the true rhythm. Thus, the computed acrophase is the best estimate of the phase of the true rhythm. Related, custom software was used to parse and pre-process the time series data after spurious time series were removed. A time series was considered spurious if it had null mean activity (mesor = 0) or a non-significant rhythmic pattern (robustness = 0). Actigraphy data was excluded if it had null RAR or no data for actigraphy (n=2421).
In accordance with the NHANES protocol, all study participants were required to wear the physical activity monitor (PAM) ActiGraph GT3X+ accelerometer (Pensacola, FL) on their nondominant hand continuously for seven consecutive days. The PAM data was used to collect objective information on study participants’ movement during wake and sleep times. A full description of the PAM procedures used during the NHANES 2013−-2014 cycle can be found on the website (https://wwwn.cdc.gov/Nchs/Nhanes/2013-2014/PAXMIN_H.htm). Physical activity was quantified using the monitor-independent movement summary (MIMS) unit, expressed as MIMS per minute of wear time. MIMS is a device-independent, open-source accelerometer-derived metric that summarizes raw acceleration signals into a standardized measure of movement intensity. For each participant, MIMS values were averaged across valid days and used as the input variable in cosinor models to characterize rest–activity rhythms [23]. Days with 0 values for mesor or amplitude were considered non-wear and excluded. A “day” was defined as midnight to midnight, consistent with NHANES accelerometer protocols. Acrophase values were modeled on a 24-hour circular scale; thus, values near 0 (just after midnight) and near 24 (just before midnight) are interpreted as proximate in time. Shift work status was not available in the dataset and was therefore not included as a covariate. We acknowledge this as a limitation, as night shift schedules may influence acrophase timing.
The main exposures of interest were first mealtime, last mealtime and the daily eating window (i.e., the average last mealtime minus the average first mealtime). The mealtimes were derived from two nonconsecutive 24-hour dietary recalls. Data for day one dietary recall was collected in-person by a trained interviewer at the Mobile Examination Center while that for day 2 is collected over the telephone within 3–10 days. The interviews utilize the USDA Automated Multiple-Pass Method (AMPM) which ensures that participants are probed on intake to ensure that misreporting is minimized. These recalls capture meal timing across multiple days, and the use of two recalls has demonstrated acceptable reliability for estimating habitual eating patterns at the population level [24]. The USDA’s Food and Nutrient Database for Dietary Studies (FNDDS) is used to determine the nutrient values for dietary intake reported by NHANES participants. In line with 24-hour recall protocols, the first and last meals were considered those consumed after midnight and before midnight, respectively. Only 6% and 3.7% of our final analytic sample had a first eating episode and acrophase between 00:00–05:00 hours respectively. Additionally, the last mealtime was a combination of both snacks and meals. All the timing variables were converted to hours. For example, meal intake at 7:30 am was recorded at 7.5 hours.
Covariates
The covariates that were considered in this study included self-reported age (years), and gender (male and female). Educational status was self-reported as less than high school, high school/GED, some college, and college graduate and above. Weight and height were measured, yielding an individual’s body mass index (BMI; weight (kg)/height (m2)). Participants were categorized into four groups based on their body mass index (BMI): underweight (< 18.5 kg/m2), normal weight (18.5 to 24.9 kg/m2), overweight (25.0 to 29.9 kg/m2), and obese (≥ 30 kg/m2). Socioeconomic status was assessed with the poverty-to-income ratio (PIR), which is the ratio of one’s household income to the federal poverty level by household size. The PIR of study participants was categorized into three groups: low income (PIR < 1), middle income (PIR 1–4), and high income (PIR > 4). Finally, the models were adjusted for total energy intake, which was assessed via 24-hour dietary recall. These covariates were selected based on prior literature demonstrating their potential influence on both dietary behaviors and rest–activity rhythms [6, 25] Age, gender, and race/ethnicity are key demographic factors associated with circadian patterns. BMI and total energy intake capture aspects of body composition and dietary load that may confound associations. Education and socioeconomic status reflect social determinants of health that can affect both meal timing and activity patterns.
Statistical analysis
The analysis conducted in this study considered the complex survey design of NHANES by applying sample weights, stratification, and clustering of the design in accordance with NHANES analytic guidelines. In accordance with NCHS guidance, the reliability of estimates was evaluated using criteria outlined in the NHANES Reliability of Estimates module, and all estimates met recommended standards [26]. Study characteristics are reported as weighted means (standard errors) for continuous variables and percentages for categorical variables. Continuous exposures were mean-centered (weighted means) so that the coefficient for the exposure corresponded to the slope within the reference race/ethnicity group (Non-Hispanic White) at the sample mean of the exposure. Models included exposure and race/ethnicity interaction terms to test moderation. We first performed a joint Wald test of the interaction (CONTRAST in SURVEYREG) for each outcome and exposure pair. For a statistically significant interaction, we estimated race-specific simple slopes (β per 1-unit change in the centered exposure) using ESTIMATE statements and report β, SE, and 95% CIs for each race/ethnicity stratum. For transparency, we also provide a matrix of Wald p-values for all outcome–exposure interactions and, for significant interactions, forest plots of the race-specific slopes with 95% CIs. When interaction terms between meal-timing exposures and race/ethnicity were included, the coefficients for meal-timing exposures represent associations within the reference group (Non-Hispanic White), while interaction terms reflect differences in associations for other racial/ethnic groups. Sensitivity analyses were conducted by (1) excluding participants with eating episodes between 00:00 and 05:00 and (2) reclassifying eating episodes between 00:00 and 02:00 as the last meal of the previous day, and models were re-estimated using the same analytic approach. All models were estimated as multivariable models that included all exposures of interest simultaneously and adjusted for age, gender, BMI, education, marital status, socioeconomic status, and total energy intake. Each meal timing variable was modeled separately while accounting for covariates. Statistical analyses were performed via SAS version 9.4 (Cary, North Carolina, USA). A p-value < 0.05 was considered statistically significant for each multiple regression analysis.
Results
The study demographics, average RAR metrics, and average meal timing for the study participants are presented in Table 1. This study included a total of 2910 participants with an average age of 47.39 (0.49) years. Approximately half (52.7%) of the participants were female. The mean first and last mealtimes were 7.55 (7:33 am) and 20.60 (8:36 pm), respectively, with a daily eating window of approximately 13 hours. Most of the participants (81%) had a daily eating window of 10 hours or more. The average time of peak activity (acrophase) for all study participants was 2:08 pm. A histogram of the time of day for the first and last meal distributions is shown in Figure 2. Figure 2 shows that most participants consumed their first meals between 6 am and 9 am Last mealtimes also occurred in the evening with the highest concentration between 6:00 and 9:00 pm. Late-night eating (00:00–05:00) was uncommon, representing approximately 2.8% of the weighted sample, with distribution across racial/ethnic groups broadly reflecting the overall sample composition (Supplementary Table 2). Supplementary Table 3 presents associations between meal timing and actigraphy-derived outcomes across the primary and sensitivity analyses. Results were highly consistent across all scenarios, including exclusion of participants with late-night eating (00:00–05:00) and reclassification of early-morning eating episodes (00:00–02:00) as the last meal of the previous day. Key associations for average start time and eating duration with actigraphy-derived mean activity and acrophase remained similar in direction, magnitude, and statistical significance, suggesting that potential misclassification of meal timing did not materially affect the findings.
Table 1.
Sample Characteristics of US Adults in NHANES 2013—2014
| Mexican American | Non-Hispanic Asian | Non-Hispanic Black | Non-Hispanic White | Other Hispanic | Other Race | Total (n=2910) | |
|---|---|---|---|---|---|---|---|
| Variables | Frequency (Weighted %) | Frequency (Weighted %) | Frequency (Weighted %) | Frequency (Weighted %) | Frequency (Weighted %) | Frequency (Weighted %) | Frequency (Weighted %) |
| Gender | |||||||
| Female | 170 (3.9) | 169 (2.6) | 316 (6.2) | 724 (36.1) | 129 (2.6) | 48 (1.3) | 1556 (52.7) |
| Education | |||||||
| College and above | 27 (0.6) | 183 (2.8) | 109 (2.1) | 396 (23.2) | 38 (0.9) | 27 (0.7) | 780 (30.4) |
| Some college | 112 (2.8) | 72 (1.1) | 210 (4) | 493 (25.1) | 82 (2.0) | 39 (1.3) | 1008 (36.2) |
| High school and GED | 88 (2.1) | 43 (0.6) | 183 (3.2) | 305 (15) | 60 (1.1) | 17 (0.5) | 696 (22.5) |
| 9–11th Grade | 98 (2.4) | 18 (0.2) | 108 (2) | 148 (5.3) | 48 (0.9) | 6 (0.1) | 426 (11) |
| Marital Status | |||||||
| Married | 178 (4.2) | 216 (3.2) | 222 (3.6) | 740 (41.2) | 120 (2.6) | 31 (1) | 1507 (55.8) |
| Widowed | 10 (0.1) | 11 (0.2) | 33 (0.5) | 116 (4.1) | 10 (0.1) | 5 (0.2) | 185 (5.2) |
| Divorced | 27 (0.5) | 25 (0.3) | 91 (1.6) | 171 (7.7) | 25 (0.4) | 13 (0.5) | 352 (11.1) |
| Separated | 11 (0.2) | 3 (0.04) | 26 (0.5) | 23 (1) | 12 (0.3) | 2 (0.03) | 77 (2.1) |
| Never married | 61 (1.8) | 52 (0.9) | 195 (4) | 210 (10.9) | 39 (1) | 27 (0.7) | 584 (19.2) |
| Living with partner | 37 (1.1) | 9 (0.2) | 43 (0.9) | 82 (3.7) | 21 (0.5) | 11 (0.3) | 203 (6.6) |
| Body mass index | |||||||
| Normal weight | 52 (1.4) | 166 (2.5) | 125 (2.3) | 352 (18.1) | 64 (1.4) | 27 (0.8) | 786 (26.6) |
| Overweight | 119 (2.8) | 90 (1.3) | 164 (3) | 421 (22.5) | 69 (1.4) | 23 (0.7) | 886 (31.8) |
| Obese | 149 (3.7) | 44 (0.7) | 303 (5.6) | 545 (27.5) | 92 (2) | 37 (1) | 1170 (40.5) |
| Underweight | 2 (0.04) | 15 (0.3) | 12 (0.2) | 14 (0.5) | 2 (0.1) | 1 (0.04) | 46 (1.1) |
| SES | |||||||
| High income | 41 (1.1) | 133 (2) | 117 (2.1) | 414 (27.7) | 37 (0.8) | 19 (0.8) | 761 (34.4) |
| Middle income | 168 (4.1) | 133 (2) | 279 (5.1) | 651 (31.7) | 115 (2.5) | 43 (1.3) | 1389 (46.6) |
| Low income | 84 (2) | 29 (0.5) | 163 (3.1) | 220 (6.6) | 54 (1.2) | 21 (0.4) | 571 (13.7) |
| Mean (SE) | Mean (SE) | Mean (SE) | Mean (SE) | Mean (SE) | Mean (SE) | Total | |
| Age (years) | 38.4 (0.5) | 43.90 (1.37) | 45.27 (0.75) | 49.73 (0.60) | 41.10 (0.95) | 40.50 (2.30) | 47.39 (0.49) |
| Age (range) | 20–80 | 20–80 | 20–80 | 20–80 | 20–80 | 20–78 | 20–80 |
| RAR Metrics N= 2910 | |||||||
| Amplitude (MIMS/min) | 7.38 (0.18) | 6.24 (0.33) | 5.66 (0.13) | 6.26 (0.08) | 7.06 (0.27) | 6.24 (0.33) | 6.31 (0.06) |
| Acrophase (hh:mm) | 13.84 (0.26) | 14.71 (0.46) | 14.31 (0.12) | 14.12 (0.09) | 14.55 (0.17) | 14.71 (0.47) | 14.18 (0.06) |
| Mesor (MIMS/min) | 9.84 (0.14) | 8.81 (0.31) | 8.42 (0.13) | 8.43 (0.11) | 9.48 (0.27) | 8.81 (0.31) | 8.60 (0.08) |
| Robust | 19.58 (0.58) | 17.05 (1.27) | 15.34 (0.45) | 17.87 (0.29) | 18.64 (0.87) | 17.05 (1.27) | 17.63 (0.23) |
| Meal Timing | |||||||
| Mean first meal (hrs.) | 7.78 (0.09) | 7.72 (0.13) | 8.18 (0.08) | 7.40 (0.08) | 7.77 (0.14) | 7.73 (0.38) | 7.56 (0.07) |
| Mean last meal (hrs.) | 20.54 (0.14) | 20.90 (0.08) | 20.63 (0.05) | 20.57 (0.06) | 20.77 (0.14) | 20.61 (0.19) | 20.60 (0.05) |
| Mean daily eating window (hrs.) | 12.77 (0.13) | 13.18 (0.17) | 12.45 (0.07) | 13.17 (0.11) | 13 (0.20) | 12.88 (0.47) | 13.05 (0.09) |
| Mean total energy (kcal) | 2224.18 (82.69) | 1957.08 (39.37) | 2064.19 (31.27) | 2091.18 (26.28) | 2097.28 (55.94) | 2183 (86.36) | 2095.2 (18.14) |
MIMS/min = Monitor-Independent Movement Summary units per minute, a standardized accelerometer-derived measure of physical activity.
SES: Socioeconomic status
Kcal: kilocalories
Hrs.: hours
Figure 2. Distribution of the First and Last Mealtimes of Study Participants.
A histogram showing the distribution of the “Begin first meal” (maroon bars) and “End last meal” (green bars) times for the study participants. The x-axis represents the time of day in hours, and the y-axis indicates the number of people.
Associations between RAR metrics and meal timing variables
Continuous meal-timing exposures were mean-centered so that the main effects represent associations within the reference race/ethnicity group (Non-Hispanic White). Models included race × meal-timing interaction terms; p-interaction values (Wald tests) are reported to assess whether associations differed by race/ethnicity. Table 2 shows race-stratified associations between acrophase and meal-timing variables for exposure pairs with statistically significant interactions (p < 0.05). It was observed for last mealtime that there was evidence of interaction (p-interaction = 0.02) by race/ethnicity. Later last mealtimes were associated with a later acrophase among Mexican American (β = 0.77 ± 0.29, p-interaction = 0.02), Non-Hispanic Black (β = 0.40 ± 0.12, p-interaction = 0.002) participants, whereas associations were not statistically significant among Non-Hispanic White, Other Hispanic, Other Race, and Non-Hispanic Asian participants. No statistically significant associations were observed for other meal-timing variables.
Table 2.
Race-stratified slopes for Acrophase × Daily eating window and Last meal
| Outcome | Exposure | Race | Beta | SE | 95% CI (Lower) | 95% CI (Upper) | p-value (Slope) | p-value (Interaction) |
|---|---|---|---|---|---|---|---|---|
| Ethnicity | ||||||||
| Acrophase | Average last mealtime | Non-Hispanic White (ref) | 0.06 | 0.07 | −0.10 | 0.22 | 0.43 | 0.02 |
| Mexican American | 0.77 | 0.28 | 0.19a | 1.36 | 0.02 | 0.02 | ||
| Other Hispanic | 0.14 | 0.22 | −0.33 | 0.61 | 0.53 | 0.02 | ||
| Non-Hispanic Black | 0.40 | 0.12 | 0.14 a | 0.66 | 0.005 | 0.02 | ||
| Non-Hispanic Asian | 0.08 | 0.16 | −0.26 | 0.42 | 0.62 | 0.02 | ||
| Other Race | 0.08 | 0.18 | −0.31 | 0.46 | 0.68 | 0.02 | ||
| Average first mealtime | Non-Hispanic White (ref) | 0.30 | 0.05 | 0.19 | 0.41 | <.001 | 0.34 | |
| Mexican American | 0.32 | 0.11 | 0.08 | 0.56 | 0.010 | 0.34 | ||
| Other Hispanic | 0.26 | 0.13 | −0.02 | 0.55 | 0.07 | 0.34 | ||
| Non-Hispanic Black | 0.14 | 0.08 | −0.03 | 0.30 | 0.09 | 0.34 | ||
| Non-Hispanic Asian | 0.15 | 0.21 | −0.30 | 0.61 | 0.49 | 0.34 | ||
| Other Race | 0.21 | 0.10 | −0.002 | 0.43 | 0.05 | 0.34 | ||
| Average daily eating window | Non-Hispanic White (ref) | −0.18 | 0.04 | −0.27 | −0.10 | 0.001 | 0.09 | |
| Mexican American | 0.03 | 0.11 | −0.20 | 0.26 | 0.78 | 0.09 | ||
| Other Hispanic | −0.11 | 0.14 | −0.40 | 0.19 | 0.45 | 0.09 | ||
| Non-Hispanic Black | 0.01 | 0.07 | −0.14 | 0.15 | 0.94 | 0.09 | ||
| Non-Hispanic Asian | −0.08 | 0.16 | −0.43 | 0.26 | 0.62 | 0.09 | ||
| Other Race | −0.09 | 0.09 | −0.29 | 0.10 | 0.31 | 0.09 | ||
| Amplit ude | Average last mealtime | Non-Hispanic White (ref) | −0.01 | 0.05 | −0.12 | 0.10 | 0.84 | 0.26 |
| Mexican American | 0.07 | 0.12 | −0.19 | 0.32 | 0.60 | 0.26 | ||
| Other Hispanic | −0.08 | 0.16 | −0.43 | 0.27 | 0.63 | 0.26 | ||
| Non-Hispanic Black | −0.05 | 0.11 | −0.28 | 0.18 | 0.65 | 0.26 | ||
| Non-Hispanic Asian | −0.22 | 0.11 | −0.45 | 0.01 | 0.06 | 0.26 | ||
| Other Race | −0.33 | 0.22 | −0.80 | 0.14 | 0.16 | 0.26 | ||
| Average first mealtime | Non-Hispanic White (ref) | −0.02 | 0.04 | −0.10 | 0.07 | 0.71 | 0.30 | |
| Mexican American | −0.06 | 0.11 | −0.29 | 0.18 | 0.62 | 0.30 | ||
| Other Hispanic | 0.15 | 0.13 | −0.14 | 0.43 | 0.28 | 0.30 | ||
| Non-Hispanic Black | 0.07 | 0.06 | −0.05 | 0.19 | 0.25 | 0.30 | ||
| Non-Hispanic Asian | 0.03 | 0.06 | −0.10 | 0.16 | 0.62 | 0.30 | ||
| Other Race | 0.05 | 0.16 | −0.29 | 0.39 | 0.75 | 0.30 | ||
| Average daily eating window | Non-Hispanic White (ref) | 0.01 | 0.04 | −0.08 | 0.09 | 0.88 | 0.10 | |
| Mexican American | 0.06 | 0.09 | −0.13 | 0.26 | 0.49 | 0.10 | ||
| Other Hispanic | −0.12 | 0.12 | −0.37 | 0.14 | 0.35 | 0.10 | ||
| Non-Hispanic Black | −0.06 | 0.04 | −0.15 | 0.02 | 0.14 | 0.10 | ||
| Non-Hispanic Asian | −0.08 | 0.05 | −0.19 | 0.04 | 0.16 | 0.10 | ||
| Other Race | −0.17 | 0.11 | −0.40 | 0.06 | 0.14 | 0.10 | ||
| Mean | Average last mealtime | Non-Hispanic White (ref) | 0.02 | 0.05 | −0.08 | 0.12 | 0.71 | 0.64 |
| Mexican American | 0.13 | 0.13 | −0.15 | 0.41 | 0.34 | 0.64 | ||
| Other Hispanic | −0.05 | 0.21 | −0.49 | 0.40 | 0.83 | 0.64 | ||
| Non-Hispanic Black | 0.03 | 0.11 | −0.20 | 0.26 | 0.79 | 0.64 | ||
| Non-Hispanic Asian | −0.17 | 0.13 | −0.45 | 0.11 | 0.21 | 0.64 | ||
| Other Race | −0.04 | 0.14 | −0.34 | 0.25 | 0.75 | 0.64 | ||
| Average first mealtime | Non-Hispanic White (ref) | −0.15 | 0.03 | −0.22 | −0.08 | <.001 | 0.25 | |
| Mexican American | −0.17 | 0.10 | −0.37 | 0.04 | 0.10 | 0.25 | ||
| Other Hispanic | 0.01 | 0.12 | −0.25 | 0.28 | 0.91 | 0.25 | ||
| Non-Hispanic Black | −0.05 | 0.06 | −0.18 | 0.08 | 0.45 | 0.25 | ||
| Non-Hispanic Asian | −0.11 | 0.12 | −0.36 | 0.14 | 0.37 | 0.25 | ||
| Other Race | −0.14 | 0.17 | −0.50 | 0.23 | 0.45 | 0.25 | ||
| Average daily eating window | Non-Hispanic White (ref) | 0.11 | 0.03 | 0.04 | 0.18 | 0.004 | 0.21 | |
| Mexican American | 0.17 | 0.06 | 0.04 | 0.31 | 0.010 | 0.21 | ||
| Other Hispanic | −0.02 | 0.13 | −0.30 | 0.26 | 0.87 | 0.21 | ||
| Non-Hispanic Black | 0.05 | 0.05 | −0.0 | 0.15 | 0.38 | 0.21 | ||
| Non-Hispanic Asian | 0.03 | 0.10 | −0.17 | 0.23 | 0.76 | 0.21 | ||
| Other Race | 0.06 | 0.12 | −0.18 | 0.31 | 0.60 | 0.21 | ||
| Robustness | Average last mealtime | Non-Hispanic White (ref) | 0.05 | 0.20 | −0.38 | 0.47 | 0.81 | 0.30 |
| Mexican American | 0.13 | 0.36 | −0.64 | 0.90 | 0.73 | 0.30 | ||
| Other Hispanic | −0.13 | 0.53 | −1.26 | 1.01 | 0.82 | 0.30 | ||
| Non-Hispanic Black | −0.54 | 0.32 | −1.23 | 0.15 | 0.12 | 0.30 | ||
| Non-Hispanic Asian | −0.49 | 0.42 | −1.38 | 0.39 | 0.25 | 0.30 | ||
| Other Race | −1.22 | 1.24 | −3.86 | 1.42 | 0.34 | 0.30 | ||
| Average first mealtime | Non-Hispanic White (ref) | 0.18 | 0.20 | −0.24 | 0.60 | 0.38 | 0.64 | |
| Mexican American | 0.10 | 0.34 | −0.63 | 0.84 | 0.77 | 0.64 | ||
| Other Hispanic | 0.46 | 0.49 | −0.57 | 1.49 | 0.36 | 0.64 | ||
| Non-Hispanic Black | 0.35 | 0.19 | −0.05 | 0.75 | 0.08 | 0.64 | ||
| Non-Hispanic Asian | 0.20 | 0.22 | −0.27 | 0.67 | 0.38 | 0.64 | ||
| Other Race | 0.37 | 0.50 | −0.70 | 1.44 | 0.47 | 0.64 | ||
| Average daily eating window | Non-Hispanic White (ref) | −0.11 | 0.16 | −0.45 | 0.24 | 0.53 | 0.29 | |
| Mexican American | −0.04 | 0.32 | −0.72 | 0.65 | 0.91 | 0.29 | ||
| Other Hispanic | −0.32 | 0.38 | −1.14 | 0.49 | 0.41 | 0.29 | ||
| Non-Hispanic Black | −0.41 | 0.12 | −0.66 | −0.17 | 0.003 | 0.29 | ||
| Non-Hispanic Asian | −0.27 | 0.21 | −0.72 | 0.19 | 0.23 | 0.29 | ||
| Other Race | −0.74 | 0.44 | −1.68 | 0.21 | 0.12 | 0.29 |
P-value (interaction) a indicates slope p<0.05 within stratum. Models adjusted for all covariates: age, gender, BMI, education, marital status, socioeconomic status, and total energy intake (n=2348); ref: reference.
Discussion
Using a nationally representative U.S. sample, we aimed to examine the association between actigraphy-derived RAR and meal timing. We also sought to investigate the associations between race/ethnicity and RAR metrics.
In this study, we found that later last mealtime was associated with a delayed acrophase among Mexican Americans and Non-Hispanic Black adults, with no significant associations among other racial/ethnic groups. The interaction between average end time and race/ethnicity was statistically significant (p-interaction = 0.02), and race-stratified simple-slope estimates confirmed that later last meals were associated with later acrophase among Mexican American and Non-Hispanic Black adults. These findings, supported by visual inspection of race-stratified slopes (Figure 3), suggest that the relationship between meal timing and circadian activity phase differs across racial/ethnic groups rather than reflecting a uniform shift across the population. These factors may contribute to the variability in the degree of alignment between eating behaviors and circadian rhythmicity [27, 28].
Figure 3.
Race-stratified simple slopes (β ± 95 % CI) for the association between average eating end time and activity acrophase. Models were survey-weighted, adjusted for age, gender, BMI, education, marital status, socioeconomic status, and total energy intake, and used mean-centered exposures. Wald p-interaction = 0.02 for the overall race × end-time interaction.
Meal timing may regulate the circadian system
The central clock of the human circadian system is in the suprachiasmatic nucleus (SCN) of the hypothalamus of the brain. There are also peripheral clocks throughout all tissues and organs of humans. The primary zeitgeber of the central clock is light; however, peripheral clocks can be synchronized by nonphotic cues such as the fasting-feeding cycle or rest–activity cycles independently of the central clock [29]. Thus, consuming meals during the biological night when the endogenous circadian system does not expect the intake of food can result in circadian misalignment and uncouple peripheral clocks from the central pacemaker of the brain, possibly leading to metabolic dysfunction [30]. The association between the timing of meals and acrophase could have implications for health. Research has revealed positive associations between early (vs. late) mealtimes and improved health outcomes [31–33]. In a study by Dashti and colleagues [31], compared with early eaters, late eaters were found to have a higher BMI, lower insulin sensitivity, and higher triglyceride levels (p < 0.01). Similarly, other studies have shown associations between late mealtimes and metabolic health, describing greater risks for dyslipidemia and insulin resistance when meals are consumed later in the day [34–36]. The consumption of meals later in the day has been associated with a delay in the acrophase of the circadian system (i.e., timing of physiological function), thus resulting in circadian disruption [37]. However, due to this study being cross-sectional, we cannot determine whether later meal timing contributes to a delayed acrophase if individuals who have a later circadian timing tend to consume meals later in the day. The race-specific effects observed in our analyses suggest that sociocultural, behavioral, or biological factors may influence how late-night eating affects circadian timing. For example, the associations among Mexican American and Non-Hispanic Black adults may reflect group differences in evening activity patterns, light exposure, or food timing consistency.
Other factors, such as having a regular eating schedule, consuming nutrient-dense meals, and reducing caloric intake, may provide health benefits. In the present study, later last meals were associated with later acrophase specifically in Mexican American and Non-Hispanic Black adults, indicating potential racial/ethnic differences in the relationship between later meal timing and circadian activity patterns.
It is important to note that the association between eating and activity rhythms is likely complex and multifactorial. First, the relationship may be bidirectional, whereby activity patterns influence eating timing, and eating timing, in turn, shapes activity rhythms [38]. Second, both behaviors are regulated by the circadian clock, and their association may reflect underlying inter-individual differences in circadian phase or chronotype rather than causal effects. Third, social schedules and environmental constraints, such as nightshift work or family routines, strongly influence both the timing of meals and physical activity. Thus, our findings may reflect the interplay of biological, behavioral, and social determinants of daily rhythms [6]. From an epidemiological perspective, these results underscore the importance of considering meal timing when examining rest–activity rhythms, either as a potential confounder, effect modifier, or covariate in future studies. Such an approach may help clarify the extent to which eating behaviors contribute to circadian organization and health outcome.
Race and ethnicity differences in circadian rhythms
In the present study, we were interested in studying the relationship between RAR and meal timing in the context of different racial and ethnic groups. The results we described in this study could be explained by factors such as genetics, culture, shift work, neighborhood quality, and economics, among others, that may leave some individuals more vulnerable to experiencing circadian disruption than others [39]. For example, African Americans have been shown to have a slightly shorter free-running circadian period (tau) than European Americans do (0.25 hours shorter). Therefore, they need a corresponding larger phase advance (i.e., earlier start of the biological rhythm than usual) each day to be able to synchronize to the 24-hour day [40]. Additionally, African Americans have been found to be at greater risk of circadian rhythm disruption—reasons include socioeconomic status and environmental factors [41]. Additionally, African Americans and Hispanics are at greater risk of developing certain chronic diseases than White Americans are [42], all of which contribute to a greater prevalence of circadian disruption among these populations. The results from a study (n=12,526) using NHANES (2011−-2014) data revealed that the differences in RAR metrics by race and ethnicity begin during childhood and persist throughout adulthood. Understanding how these disparities develop may allow for more tailored interventions in different populations. [43]
The strengths of our study include the use of a large representative sample, which improves the generalizability of the findings. Second, we evaluated meal timing and rest–activity rhythms, a previously understudied connection, despite the importance of the feeding–fasting cycle in understanding circadian misalignment. Finally, we utilized rigorous analytical methods from accelerometry data to estimate RAR measures.
The limitations of the study must be considered when these results are interpreted. First, these data are cross-sectional and thus cannot comprehensively capture changes across the lifespan within individuals. Second, although we used two 24-hour dietary recalls, meal timing may not represent participants’ habitual patterns, and the data are subject to recall error. Third, participants did not report sleep periods, limiting our ability to determine whether the first reported eating occasion reflects breakfast or late-night intake. In addition, meal timing variables may include both meals and snacks, which may differ in their metabolic and circadian effects and contribute to heterogeneity in the observed associations, particularly for last meal timing. Although a small proportion of participants reported eating between midnight and 5 am, this may introduce minor exposure misclassification; however, sensitivity analyses were conducted to address this. Fourth, the dietary recall days did not overlap with the 7-day actigraphy period, which may have limited the ability to capture concurrent meal timing and rest–activity rhythms. Although, we adjusted for several demographic and lifestyle covariates, no model-building process was conducted to evaluate their impact on associations, and residual confounding may remain. Additionally, although excluding cases participants with zero RAR values improved data quality, participants with partial non-wear were included and that could limit standardization across observations. Lastly, the reliability of estimates was evaluated using NCHS-recommended methods, however, some racial/ethnic subgroup sample sizes were relatively small, which may limit the precision and generalizability of those estimates. Future research should examine the combined effects of time-restricted eating on circadian rhythms and metabolic health.
Conclusion
In summary, this study examined associations between actigraphy-derived rest–activity rhythms (RARs) and meal timing and whether these associations are modified by race/ethnicity. We found evidence of possible associations between later acrophase and later timing of meals, suggesting that behavioral rhythms and dietary timing may be linked. Additionally, we observed that racial and ethnic differences exist in RARs, with Mexican Americans and Black adults showing later last meals appearing to delay acrophase compared to White adults. These findings underscore the importance of considering both meal timing and racial/ethnic context when characterizing circadian rhythms. Future research should continue to investigate these patterns using multiple days of dietary intake and concurrent measures of circadian behavior.
Supplementary Material
What was known:
Disruptions in circadian rhythms and irregular meal timing have been associated with adverse health outcomes. Previous studies have suggested that eating later in the day may be associated with circadian misalignment resulting in adverse health effects. However, the relationship between meal timing and rest–activity rhythms (RAR) in a nationally representative U.S. population remains poorly characterized, and potential differences by race/ethnicity are not well understood.
What study adds:
This study provides evidence associating meal timing with actigraphy-derived RAR metrics and highlights how later mealtimes may relate to daily activity rhythms. We demonstrate the differing associations by race/ethnicity to emphasize the importance of considering social and behavioral factors when examining circadian and sleep-related patterns.
Funding statement:
VYAB is supported by the Promotion of Academic Workforce Diversity in Translational Behavioral & Cardiometabolic Research (PINNACLE; Grant #5T32HL166609–02).
Footnotes
Declaration of conflicts of interest: Given their role as Sleep Health editorial board members, Drs. Grandner and Jean-Louis were not involved in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to another journal editor.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Data availability statement:
The datasets supporting the conclusions of this article are publicly available at [https://wwwn.cdc.gov/nchs/nhanes/continuousnhanes/default.aspx?BeginYear=2013].
REFERENCES
- 1.Gutman R, Pendergast JS, Nakamura W, Kojima S. Circadian desynchrony: Consequences, mechanisms, and Open Issues. Frontiers Media SA; 2023. p. 1177643. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Mitchell JA, Quante M, Godbole S, James P, Hipp JA, Marinac CR, et al. Variation in actigraphy-estimated rest-activity patterns by demographic factors. Chronobiol Int. 2017;34(8):1042–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Xiao Q, Matthews CE, Playdon M, Bauer C. The association between rest-activity rhythms and glycemic markers: the US National Health and Nutrition Examination Survey, 2011–2014. Sleep. 2022;45(2):zsab291. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Makarem N, German CA, Zhang Z, Diaz KM, Palta P, Duncan DT, et al. Rest-Activity Rhythms Are Associated With Prevalent Cardiovascular Disease, Hypertension, Obesity, and Central Adiposity in a Nationally Representative Sample of US Adults. J Am Heart Assoc. 2024;13(1):e032073. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Baron KG, Reid KJ. Circadian misalignment and health. Int Rev Psychiatry. 2014;26(2):139–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ansu Baidoo V, Knutson KL. Associations between circadian disruption and cardiometabolic disease risk: a review. Obesity. 2023;31(3):615–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wehrens SM, Christou S, Isherwood C, Middleton B, Gibbs MA, Archer SN, et al. Meal timing regulates the human circadian system. Current Biology. 2017;27(12):1768–75. e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Hawley JA, Sassone-Corsi P, Zierath JR. Chrono-nutrition for the prevention and treatment of obesity and type 2 diabetes: from mice to men. Diabetologia. 2020;63(11):2253–9. [DOI] [PubMed] [Google Scholar]
- 9.Flanagan A, Bechtold DA, Pot GK, Johnston JD. Chrono-nutrition: From molecular and neuronal mechanisms to human epidemiology and timed feeding patterns. J Neurochem. 2021;157(1):53–72. [DOI] [PubMed] [Google Scholar]
- 10.Raji OE, Kyeremah EB, Sears DD, St-Onge MP, Makarem N. Chrononutrition and Cardiometabolic Health: An Overview of Epidemiological Evidence and Key Future Research Directions. Nutrients. 2024;16(14). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hatanaka M, Hatamoto Y, Tajiri E, Matsumoto N, Tanaka S, Yoshimura E. An Earlier First Meal Timing Associates with Weight Loss Effectiveness in A 12-Week Weight Loss Support Program. Nutrients. 2022;14(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lopez-Minguez J, Gómez-Abellán P, Garaulet M. Timing of breakfast, lunch, and dinner. Effects on obesity and metabolic risk. Nutrients. 2019;11(11):2624. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Leung GKW, Huggins CE, Bonham MP. Effect of meal timing on postprandial glucose responses to a low glycemic index meal: A crossover trial in healthy volunteers. Clin Nutr. 2019;38(1):465–71. [DOI] [PubMed] [Google Scholar]
- 14.Hatori M, Vollmers C, Zarrinpar A, DiTacchio L, Bushong EA, Gill S, et al. Time-restricted feeding without reducing caloric intake prevents metabolic diseases in mice fed a high-fat diet. Cell metabolism. 2012;15(6):848–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gabel K, Hoddy KK, Haggerty N, Song J, Kroeger CM, Trepanowski JF, et al. Effects of 8-hour time restricted feeding on body weight and metabolic disease risk factors in obese adults: A pilot study. Nutr Healthy Aging. 2018;4(4):345–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hutchison AT, Regmi P, Manoogian EN, Fleischer JG, Wittert GA, Panda S, et al. Time-restricted feeding improves glucose tolerance in men at risk for type 2 diabetes: a randomized crossover trial. Obesity. 2019;27(5):724–32. [DOI] [PubMed] [Google Scholar]
- 17.Wilkinson MJ, Manoogian ENC, Zadourian A, Lo H, Fakhouri S, Shoghi A, et al. Ten-Hour Time-Restricted Eating Reduces Weight, Blood Pressure, and Atherogenic Lipids in Patients with Metabolic Syndrome. Cell Metab. 2020;31(1):92–104 e5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Ansu Baidoo VY, Zee PC, Knutson KL. Racial and Ethnic Differences in Eating Duration and Meal Timing: Findings from NHANES 2011–2018. Nutrients. 2022;14(12). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Johnson CL, Dohrmann SM, Burt VL, Mohadjer LK. National health and nutrition examination survey: sample design, 2011–2014: US Department of Health and Human Services, Centers for Disease Control and; …; 2014. [Google Scholar]
- 20.Refinetti R, Cornelissen G, Halberg F. Procedures for numerical analysis of circadian rhythms. Biol Rhythm Res. 2007;38(4):275–325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Refinetti R Circadian physiology: CRC press; 2016. [Google Scholar]
- 22.Chatfield SC, Volpicelli FM, Adler NM, Kim KL, Jones SA, Francois F, et al. Bending the cost curve: time series analysis of a value transformation programme at an academic medical centre. BMJ Qual Saf. 2019;28(6):449–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.John D, Tang Q, Albinali F, Intille S. An Open-Source Monitor-Independent Movement Summary for Accelerometer Data Processing. J Meas Phys Behav. 2019;2(4):268–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Moshfegh AJ, Rhodes DG, Baer DJ, Murayi T, Clemens JC, Rumpler WV, et al. The US Department of Agriculture Automated Multiple-Pass Method reduces bias in the collection of energy intakes. Am J Clin Nutr. 2008;88(2):324–32. [DOI] [PubMed] [Google Scholar]
- 25.Boege HL, Bhatti MZ, St-Onge MP. Circadian rhythms and meal timing: impact on energy balance and body weight. Curr Opin Biotechnol. 2021;70:1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Johnson CL, Paulose-Ram R, Ogden CL, Carroll MD, Kruszan-Moran D, Dohrmann SM, et al. National health and nutrition examination survey. Analytic guidelines, 1999–2010. 2013. [PubMed] [Google Scholar]
- 27.Eastman CI, Tomaka VA, Crowley SJ. Circadian rhythms of European and African-Americans after a large delay of sleep as in jet lag and night work. Sci Rep. 2016;6:36716. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Hughes AJ, Gunn H, Siengsukon C, Stearns MA, James E, Donley T, et al. Eliminating sleep health disparities and achieving health equity: Seven areas for action in the behavioral sleep medicine community. Behavioral sleep medicine. 2023;21(5):633–45. [DOI] [PubMed] [Google Scholar]
- 29.Wehrens SMT, Christou S, Isherwood C, Middleton B, Gibbs MA, Archer SN, et al. Meal Timing Regulates the Human Circadian System. Curr Biol. 2017;27(12):1768–75 e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Damiola F, Le Minh N, Preitner N, Kornmann B, Fleury-Olela F, Schibler U. Restricted feeding uncouples circadian oscillators in peripheral tissues from the central pacemaker in the suprachiasmatic nucleus. Genes Dev. 2000;14(23):2950–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Dashti HS, Gómez-Abellán P, Qian J, Esteban A, Morales E, Scheer FA, et al. Late eating is associated with cardiometabolic risk traits, obesogenic behaviors, and impaired weight loss. The American journal of clinical nutrition. 2021;113(1):154–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Allison KC, Hopkins CM, Ruggieri M, Spaeth AM, Ahima RS, Zhang Z, et al. Prolonged, controlled daytime versus delayed eating impacts weight and metabolism. Current Biology. 2021;31(3):650–7. e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ruiz-Lozano T, Vidal J, De Hollanda A, Scheer F, Garaulet M, Izquierdo-Pulido M. Timing of food intake is associated with weight loss evolution in severe obese patients after bariatric surgery. Clinical nutrition. 2016;35(6):1308–14. [DOI] [PubMed] [Google Scholar]
- 34.Yoshida J, Eguchi E, Nagaoka K, Ito T, Ogino K. Association of night eating habits with metabolic syndrome and its components: a longitudinal study. BMC public health. 2018;18:1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Chen H-J, Chuang S, Chang H, Pan W. Energy intake at different times of the day: Its association with elevated total and LDL cholesterol levels. Nutrition, Metabolism and Cardiovascular Diseases. 2019;29(4):390–7. [DOI] [PubMed] [Google Scholar]
- 36.Bandin C, Scheer FA, Luque AJ, Avila-Gandia V, Zamora S, Madrid JA, et al. Meal timing affects glucose tolerance, substrate oxidation and circadian-related variables: A randomized, crossover trial. Int J Obes (Lond). 2015;39(5):828–33. [DOI] [PubMed] [Google Scholar]
- 37.Sundaram S, Johnson LK, Yan L. High-Fat Diet Alters Circadian Rhythms in Mammary Glands of Pubertal Mice. Front Endocrinol (Lausanne). 2020;11:349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Zuraikat FM, Wood RA, Barragan R, St-Onge MP. Sleep and Diet: Mounting Evidence of a Cyclical Relationship. Annu Rev Nutr. 2021;41:309–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Bower KM, Thorpe RJ Jr, Rohde C, Gaskin DJ. The intersection of neighborhood racial segregation, poverty, and urbanicity and its impact on food store availability in the United States. Preventive medicine. 2014;58:33–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Zhang C, Tait C, Minacapelli CD, Bhurwal A, Gupta K, Amin R, et al. The role of race, sex, and age in circadian disruption and metabolic disorders. Gastro Hep Advances. 2022;1(3):471–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Bower KM, Thorpe RJ, Jr., Rohde C, Gaskin DJ. The intersection of neighborhood racial segregation, poverty, and urbanicity and its impact on food store availability in the United States. Prev Med. 2014;58:33–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Odlum M, Moise N, Kronish IM, Broadwell P, Alcántara C, Davis NJ, et al. Trends in poor health indicators among Black and Hispanic middle-aged and older adults in the United States, 1999–2018. JAMA Network Open. 2020;3(11):e2025134–e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Wallace DA, Johnson DA, Redline S, Sofer T, Kossowsky J. Rest-activity rhythms across the lifespan: cross-sectional findings from the US representative National Health and Nutrition Examination Survey. Sleep. 2023;46(11):zsad220. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets supporting the conclusions of this article are publicly available at [https://wwwn.cdc.gov/nchs/nhanes/continuousnhanes/default.aspx?BeginYear=2013].