Comprehensive assessment of chrononutrition behaviors among nationally representative adults: Insights from National Health and Nutrition Examination Survey (NHANES) data

Abstract

Background: Aligning the time of food intake, i.e., chrononutrition, with the body's circadian clock can have a significant impact on overall health, particularly cardiometabolic health. However, there is a lack of population-based information on various chrononutrition behaviors in the United States, where the prevalence of obesity is high.

Objective: Our primary aim was to characterize chrononutrition behaviors and their 15-year trends among US adults. We also explored the temporal associations between trends in chrononutrition behaviors and trends in obesity.

Design: We utilized data from 8 cycles (2003-2018) of the National Health and Nutrition Examination Survey (NHANES) on 34,470 adults (age >19 years). The clock time of food/beverage consumption was extracted from two 24-h food recalls. The following chrononutrition behaviors were defined: 1) The clock time of the first, last, and midpoint (when 50% of total daily energy was consumed) of food/beverage intake; 2) Eating window (the time elapsed between the first and last intake); 3) Late-night eating (food intake between 21:00-23:59); and 4) Eating frequency. Survey-weighted % or mean ± standard error (SE) was used to demonstrate chrononutrition behaviors and survey-weighted regression models were utilized to evaluate trends in chrononutrition behaviors, BMI, and obesity over a 15-year period.

Results: Thirty five percent of US adults had long eating windows lasting 13 h or more, with 59% of individuals consuming calories after 9 PM. The patterns of food intake among American adults were skewed, with the highest proportion of their daily energy intake (36%) being consumed during dinner meals. Notable differences in chrononutrition behaviors observed among different population subgroups. Young adults and men had longer eating windows with a higher prevalence of late-night eating compared to their age- and sex-counterparts. Black individuals had shorter eating periods due to delayed breakfast, the highest proportion (68%) of late-night eating, and obtained a greater amount of energy intake from snacks compared to other racial/ethnic groups. Over the 15-year span, there were only minor changes in a few aspects of chrononutrition behaviors, including 2% reduction in the time of eating window, while most other meal timing behaviors remained unchanged. Trends in chrononutrition behaviors were disproportionately smaller than the trends in obesity rates.

Conclusions: US adults persistently consume higher amounts of daily energy intake later in the day. Despite calls for Americans to shift intake to earlier parts of the day, this study shows that there is little change in the overall population over the 15-year period reviewed.

Keywords: Body's internal clock; Chrono diet; Circadian nutrition; Mealtimes; Weight.

Figure 1

Mean clock time of food/beverage intake (hours/day) within a day among US adults (N = 34,470).

Figure 2

Proportion of energy intake according to self-identified eating occasions among US adults in NHANES (N = 34,470). Values are survey-weighted mean ± standard error to account for the NHANES complex sampling design.

Figure 3

Percentage of daily energy intake during self-identified eating occasions (i.e., breakfast, lunch, dinner, and snack) across NHANES cycles (from 2003 to 2018; N = 34,470).

Figure 4

Trends in chrononutrition behaviors (i.e., eating window, eating frequency, last intake time, and late-night eating) alongside trends in obesity among US adults by NHANES cycles (2003-2018). P _trend for eating window < 0.01; P _trend for eating frequency = 0.19; P _trend for last intake time food/beverage < 0.01; P _trend for late-night eating < 0.01; P _trend for obesity < 0.01. Late Night Eaters were individuals in the "yes" category of the dichotomous late night eating variable (yes/No), indicating that they consumed food or beverages containing more than 0 kcal energy between 21:00 and 23:59 within a day. P values for trends are from survey-weighted general linear regression for continuous variables and from survey-weighted logistic regression for categorical variables.