Understanding Accelerated Summer Body Mass Index Gain by Tracking Changes in Children's Height, Weight, and Body Mass Index Throughout the Year

Abstract

Background:

Drivers of summer body mass index (BMI) gain in children remain unclear. The Circadian and Circannual Rhythm Model (CCRM) posits summer BMI gain is biologically driven, while the Structured Days Hypothesis (SDH) proposes it is driven by reduced structure.

Objectives:

Identify the mechanisms driving children's seasonal BMI gain through the CCRM and SDH.

Methods:

Children's (N = 147, mean age = 8.2 years) height and weight were measured monthly during the school year, and once in summer (July–August). BMI z-score (zBMI) was calculated using CDC growth charts. Behaviors were measured once per season. Mixed methods regression estimated monthly percent change in children's height (%HΔ), weight (%WΔ), and monthly zBMI for school year vs. summer vacation, seasonally, and during school months with no breaks vs. school months with a break ≥1 week.

Results:

School year vs. summer vacation analyses showed accelerations in children's %WΔ (Δ = 0.9, Standard Error (SE) = 0.1 vs. Δ = 1.4, SE = 0.1) and zBMI (Δ = −0.01, SE = 0.01 vs. Δ = 0.04, SE = 0.3) during summer vacation, but %HΔ remained relatively constant during summer vacation compared with school (Δ = 0.3, SE = 0.0 vs. Δ = 0.4, SE = 0.1). Seasonal analyses showed summer had the greatest %WΔ (Δ = 1.8, SE = 0.4) and zBMI change (Δ = 0.05, SE = 0.03) while %HΔ was relatively constant across seasons. Compared with school months without a break, months with a break showed higher %WΔ (Δ = 0.7, SE = 0.1 vs. Δ = 1.6, SE = 0.2) and zBMI change (Δ = −0.03, SE = 0.01 vs. Δ = 0.04, SE = 0.01), but %HΔ was constant (Δ = 0.4, SE = 0.0 vs. Δ = 0.3, SE = 0.1). Fluctuations in sleep timing and screen time may explain these changes.

Conclusions:

Evidence for both the CCRM and SDH was identified but the SDH may more fully explain BMI gain. Interventions targeting consistent sleep and reduced screen time during breaks from school may be warranted no matter the season.

Introduction

Summer (i.e., May–August) is a period of excess body mass index (BMI) gain with a growing body of evidence indicating that school-aged children experience accelerations in BMI gain.^1–6 As a consequence of accelerated summer BMI gain, the percent of children with obesity increases during summer. Currently, there are two complementary conceptual models to explain accelerated summer BMI gain in children—the Circadian and Circannual Rhythm Model (CCRM)⁷ and the Structured Days Hypothesis (SDH).⁸ The CCRM posits that the observed seasonality in children's height and weight gain is regulated by the circannual clock that is entrained or synchronized by exposure to the earth's seasonal light/dark cycle.⁷ It is hypothesized that children are biologically primed to experience accelerated BMI gains during the summer due to seasonal rhythmicity in their chronobiology that likely evolved to maximize growth in the summer when food is plentiful. Previous studies suggest that children's height gain accelerates in the late spring and early summer and decelerates in fall and winter, whereas children's weight increases more in the late summer and early fall and decelerates in the winter and spring.^9,10

According to the CCRM,⁷ these fluctuations are believed to be driven by longer days leading to more exposure to sunlight in the summer and shorter days leading to less exposure to sunlight in the winter. Because BMI is a ratio of weight to height [i.e., weight (kg)/(height[m])²], this should lead to biologically driven increases in BMI during the summer and a flatlining of BMI gain during the winter. Compounding on these biologically driven increases, the CCRM also indicates that during the summer when children are away from school they are exposed to fewer social demands. These reduced social demands may lead to later and more variable bedtimes, which may lead to circadian misalignment during the summer, a risk factor for accelerated BMI gain.

The SDH⁸ posits that structure, defined as preplanned, segmented, and adult-supervised compulsory environments, protects children against behaviors that may lead to obesity and prevents excessive BMI gain. The SDH draws upon concepts in the “filled-time perspective” literature, which claims that time filled with favorable activities cannot be filled with unfavorable activities.¹¹ This perspective leads to the hypothesis that children engage in a greater number of behaviors that lead to increased weight gain during times that are less structured (e.g., nonschool days) than during times that are more structured (e.g., school days). Unhealthy behaviors include (1) increased time spent sedentary,^12,13 (2) reduced engagement in physical activity,^13–15 (3) displaced and unstable sleep patterns,^16–18 and (4) a less nutrient-rich diet.^19–21 The traditional school calendar in the United States may lead to accelerated BMI gain during summer, as children typically attend school from August to May and have a concentrated summer vacation with prolonged unstructured time during June and July.

It is unclear whether the degree to which accelerated summer BMI gain is a function of naturally occurring biological fluctuations in height and weight gain (as per the CCRM) or socially imposed shifts in structure (as per the SDH). This is a critical limitation of our understanding of the etiology of childhood obesity. If accelerated summer BMI gain is a naturally occurring, biologically driven phenomenon, then developing behavioral interventions to mitigate accelerated summer BMI gain would be of limited utility and waste valuable resources. If, however, accelerated summer BMI gain is a product of limited structure, then behavioral interventions that impose health-promoting structure during the summer are critical for preventing accelerated summer BMI gain.

The key to understanding the mechanisms driving accelerated summer BMI gain lies in the predictions that the CCRM and SDH make related to height, weight, and corresponding BMI change during the 9 months of the year outside of summer. Figure 1 summarizes the changes in height, weight, and BMI predicted by the CCRM and the SDH. While both the CCRM and SDH predict accelerated summer BMI gain, they predict dramatically different height, weight, and corresponding BMI changes during the 9 months of school. The CCRM predicts BMI should decrease during the winter, regardless of breaks from school, while the SDH predicts BMI gain should increase during breaks from school, like winter break. Furthermore, the CCRM predicts height gain will accelerate in the spring and early summer while the SDH predicts height gain will remain relatively constant throughout the year. Because of the different hypothesized mechanisms for change, the two theories would also predict that health behaviors would change differently throughout the year. The CCRM would hypothesize that primarily changes in children's sleep timing and duration throughout the year would lead to changes in BMI.

Figure 1.

Predicted height, weight, and BMI change predicted by the CCRM and the structured days hypothesis. BMI, body mass index; CCRM, Circadian and Circannual Rhythm Model.

Specifically, children would sleep the most and sleep timing would be the earliest in winter, followed by fall and spring, and finally duration would be the shortest while timing would be the latest in summer. Alternatively, the SDH would hypothesize that all of children's behaviors would remain relatively constant during days with more structure (i.e., school days), while during days with less structure (i.e., breaks from school) these behaviors would be less healthy.

Up to this point, we have been unable to answer the question about when children's height and weight increase because studies exploring accelerated summer BMI gain only measure children in the spring (before summer) and the fall (following summer). This only allows for BMI gain comparisons between the 3 months of summer and the other 9 months of the year. Thus, these studies cannot unpack the accelerations in height or weight (which comprise BMI gain) predicted by the CCRM and SDH that happen within the 9 months children are in school. The primary aim of this study was to explore children's seasonal BMI gain through the lens of the CCRM and SDH. As a secondary aim, this study also explored seasonal changes in children's health behaviors.

Methods

Setting and Participants

Table 1 presents the characteristics of the participating schools and children. A single school in the Midlands Region of South Carolina, United States, was invited to participate in the study. The school served 935 children from K-5th grade (i.e., ∼5–12 years) and 60.9% of the children came from families living in poverty based upon the Department of Health and Human Services' poverty guidelines.²² The school was purposefully selected because it had served a diverse population of students (i.e., 45.8% Black, 30.8% White, 5.5% Asian, 9.4% Hispanic, 8.4% Other). The school consisted of 44 classrooms with a random subsample of 8 classes selected to participate in this study. Fifth-grade classrooms (n = 7) were excluded because they would be transitioning to a different school following their fifth-grade year (i.e., transitioning to middle schools), whereas kindergarten classrooms (n = 8) were excluded to ensure as some of the behavioral measures were developed in samples of children that did not include kindergarten-aged children.²³ The remaining 29 classrooms were stratified by grade level with 2 classrooms randomly selected to participate in the study in each grade.

Table 1.

Number of Participants With Valid Measure by Wave

	School year vs. summer vacationa	Monthb	Seasonc	Break vs. no breakd	Number of participants
Height and weight	School year	September	Fall	No break ≥1 week	140
	School year	October	Fall	No break ≥1 week	143
	School year	November	Fall	No break ≥1 week	144
	School year	December	Winter	Break ≥1 week	140
	School year	January	Winter	Break ≥1 week	140
	School year	February	Winter	No break ≥1 week	136
	School year	March	Spring	Break ≥1 week	137
	School year	April	Spring	Break ≥1 week	136
	School year	May	Spring	No break ≥1 week	135
	Summer vacation	June/July	Summer	N/A	38
	Summer vacation	August	Summer	N/A	117
Behavioral	Fall				85
	Winter				79
	Spring				67
	Summer				52

^a Independent variable for changes predicted by both the SDH and CCRM during the school year vs. summer vacation.

^b Month of measure.

^c Independent variable for changes predicted by the CCRM during each season.

^d Independent variable for change predicted by the SDH during months during the school year that had at least 1 week break from school vs. months during the school year that did not have at least a 1 week break from school.

CCRM, Circadian and Circannual Rhythm Model; SDH, Structured Days Hypothesis.

All children (N = 172) in these classrooms were eligible to participate. Since height and weight are commonly collected in school, a passive, opt-out protocol was adopted for the height and weight measures. Parents received a letter informing them of the study protocols and aims and informed them to contact their child's classroom teacher if they did not want their child to take part. As part of that letter, parents were also given the option to opt-in to the behavioral measures portion of the study. Parents who consented to their child's participation in the behavioral measures portion of the study were asked to sign and return the letter to the school where it was retrieved by research staff. Child assent was requested and received before all anthropometric and behavioral measures.

Study Design and Procedures

All protocols and procedures were approved by the lead author's IRB before enrollment of the first participant. This observational cohort study followed children over one academic year and summer vacation taking place over a 12-month period. Eleven height and weight measures that typically occurred in the third week of the month were completed. Additionally, behavioral measures were collected four times throughout the year and were timed to coincide as closely as possible with the fall equinox (September 22), winter solstice (December 21), spring equinox (March 20), and summer solstice (June 21).

Measures

Height and weight (primary)

Children were measured monthly during the 2021/22 school year (September 2021–May 2022) and following the summer of 2022 (September 2022). The subset of children that consented to the behavioral measures also had their height and weight measured in July of 2022. All measures were during the same 1-week period at each measurement occasion, which occurred in the third week of the month. All measures were obtained during regularly scheduled physical education (PE) class by trained research assistants. During the summer, children were contacted to make an appointment with the research team and measures were completed at their school. At each measurement occasion (i.e., school and summer), children's height (nearest 0.1 cm) was collected through a portable stadiometer (Model S100; Ayrton Corp., Prior Lake, MN) and weight (nearest 0.01 kg) through a digital scale (Healthometer model 500KL, Health o meter, McCook, IL).

Height and weight were collected without shoes, in light clothing (e.g., no coats, heavy sweaters), and in duplicate. The mean of the two measures was calculated for use in analytic modeling. BMI was calculated (BMI = kg/m²) and transformed into age- and sex-specific z-scores (zBMI).²⁴

Physical activity and sleep (secondary)

Physical activity and sleep data were measured using the ActiGraph GT3X+. The accelerometers were programmed in ActiLife (software version 6.13.4) to record in 1 second epochs at 30 hertz with idle sleep mode enabled. Time spent in physical activity intensity categories was determined using intensity thresholds described by Hildebrand et al.^25,26 Sleep data were estimated using the approach described by van Hees et al.²⁷ A sleep log was collected from parents nightly through the daily diary (see below) to guide sleep detection. Accelerometer data were processed using the GGIR package (version 2.6.4) in R (R Foundation for Statistical Computing; Vienna, Austria). A valid day of data included a minimum of 16 hours of wear time. Strings of 30 minutes or more of consecutive zeros were counted as nonwear time and were removed. The participants were asked to wear the accelerometer on their nondominant wrist at all times for 2 weeks in the fall of 2021 (October 27th to November 7th), winter of 2022 (January 12th to January 26th), spring of 2022 (March 23rd to April 6th), and summer of 2022 (July 11th to July 25th).

For fall, winter, and spring measures, participants received their accelerometers at school from research staff and were requested to return them to the school at the end of the 2-week period. For the summer measure, participants scheduled a time to meet research staff at school to have their height and weight measured and receive their accelerometer. The participant then mailed their accelerometer back to research staff in a self-addressed postage-paid envelope.

Healthy and unhealthy food/drink consumption, and screen time (secondary)

Children's food and drink consumption and screen time were assessed through parent report. Parents received a text message with a link to a daily diary at 7 pm, which asked them to report their child's screen time and foods consumed daily during the four 14-day periods children wore the accelerometer. To enhance the accuracy of reports, parents were encouraged to complete the diaries along with their child. Parents/children estimated the total amount of time (hours and minutes) spent in front of a screen that day (e.g., TV, computer, video game, smartphone, tablet).²³ Similar to past studies,²⁸ foods and drinks were assessed using the Beverage and Snack Questionnaire.²⁹ For this study, individual food items were grouped in accordance with the Healthy Meal Index.³⁰ Food categories included: fruits, vegetables, dairy (nonsugar sweetened), convenience foods (fast food), sweets and desserts, and sugar-sweetened beverages (including dairy).

Two groups were created for analysis of foods consumed: healthy foods/drinks (fruits, vegetables, and unsweetened dairy, 100% fruit juice), and unhealthy foods/drinks (convenience foods, and sweets/desserts, sugar-sweetened beverages). Consumption was dichotomized (i.e., “did” vs. “did not” consume) and reported as mean healthy or unhealthy foods/drinks consumed per day.²⁹

Statistical Analyses

All analyses were completed in Stata (v16.1; College Station, TX) during August of 2022. Before completing the primary analyses, descriptive means and standard deviations of school and child characteristics were examined. For the primary analysis, children with at least one measure of height and weight at any time point were included.³¹ For secondary analyses, children with at least one valid day of physical activity, sleep, diet, or screen time at any time point were included.

For the primary analyses, multilevel mixed effects regression models estimated monthly percent change in height and weight and change in zBMI. To account for differing time between measurement periods, change was divided by the number of days between each child's measure to derive daily change. This value was then multiplied by 30.4, the mean number of days between measures, to derive monthly change. The formula for monthly change is presented in Supplementary File S1. Separate models estimated change¹ predicted by both the CCRM and SDH during the school year (September-May) vs. summer vacation from school (May–September),² predicted by the CCRM during each season (i.e., fall, winter, spring, and summer), and³ predicted by the SDH during months during the school year that had at least 1-week break from school (i.e., December, January, March, April) vs. months during the school year that did not have at least a 1-week break from school (i.e., August, September, October, November, February, May). See Table 1 for a representation of these measures. Meteorological seasons were defined as: fall (September–November), winter (December–February), spring (March–May), and summer (June–August).

For the secondary analyses examining differences in health behaviors by season, multilevel mixed effects linear regressions, with days nested within children, were estimated. Separate models were estimated for each of the measured health behaviors of (1) sedentary time, (2) light physical activity, (3) MVPA, (4) screen time, (5) foods and drinks consumed, (6) foods consumed, (7) drinks consumed, (8) total sleep time, (9) sleep onset, and (10) sleep offset. A categorical season variable was considered the independent variable.

All statistical models included sex, race/ethnicity, age, and baseline value of the dependent variable as covariates. Physical activity models included wear time as an additional covariate. All primary and secondary analyses used full information maximum likelihood estimators to account for missing data, which have been shown to produce unbiased results.^32,33 Furthermore, because only a subset of children completed a measure at the midpoint of summer (i.e., June/July), sensitivity analyses were conducted that was stratified by children who had a summer measure and those that did not.

Results

Figure 2 shows the flow of participants through the height and weight and behavioral measures. Supplementary Tables S1 and S2 show the pattern of measures for participant's height and weight, and behaviors, respectively. Briefly, 70% of participants were missing one or fewer height and weight measures and over 90% missing 2 or fewer measures. For behavioral measures, 69% of participants were missing one or fewer measures and over 90% were missing 2 or fewer measures. Table 1 displays the total number of participants with valid height, weight, and behavioral measures at each measurement wave. Table 2 presents the characteristics of the participating school and children. Figure 3 presents the percent change in weight and height and change in zBMI by month.

Figure 2.

Flow of participants through the study.

Table 2.

Demographics of Participating School and Children

Characteristics of participants with height and weight measures
Number of participating teachers	8
Number of participating students (at least one height/weight measure)	147
Female (%)	48.9
Age (SD), years	8.2 (1.1)
Grades	1st–4th
Race/ethnicity
Black (%)	52.4
White (%)	37.4
Asian (%)	7.5
Hispanic (%)	0.7
Other (%)	2.0
Characteristics of participants with behavioral measures
Number of participating students (at least one valid day of accelerometer wear)	90
Female (%)	49.4
Age (SD), years	8.2 (1.1)
Race/ethnicity
Black (%)	48.3
White (%)	34.8
Asian (%)	9.0
Hispanic (%)	0.6
Other (%)	7.3

SD, standard deviation.

Figure 3.

Monthly percent change in weight, height, and change in BMI z-score. *Indicates a statistically significant difference at p ≤ 0.05 from baseline change.

Findings related to the primary analyses can be found in Figure 4.

Figure 4.

Percent change in children's weight, height, and change in BMI z-score. Percent weight change (a) summer vs. school, (b) seasonally, (c). During school months with no breaks ≥1 week vs. school months with breaks ≥1 week. Percent height change (d) summer vs. school, (e) seasonally, (f). During school months with no breaks ≥1 week vs. school months with breaks ≥1 week. BMI z-score change (g) summer vs. school, (h) seasonally, (i). during school months with no breaks ≥1 week vs. school months with breaks ≥1 week. *Indicates a statistically significant difference at p ≤ 0.05 from referent change (i.e., school, fall, no break months).

School Year Vs. Summer

Children's monthly percent weight change (Δ = 0.9, Standard Error (SE) = 0.1 vs. Δ = 1.4, SE = 0.1) and zBMI change (Δ = −0.01, SE = 0.01 vs. Δ = 0.04, SE = 0.03) were greater during summer vacation compared with the school year. The same increase was not evident during summer vacation for monthly percent height change (Δ = 0.3, SE = 0.0 vs. Δ = 0.4, SE = 0.1). Sensitivity analyses were not different from the primary analyses.

Seasonal Analyses (Summer Vs. Spring, Winter, and Fall)

For seasonal analyses, summer showed the highest monthly percent change in weight (Δ = 1.8, SE = 0.4) followed by spring (Δ = 1.4, SE = 0.3), fall (Δ = 1.0, SE = 0.2), and winter (Δ = 0.8, SE = 0.3). The same pattern of change in zBMI was evident with summer the highest (Δ = 0.05, SE = 0.03), followed by spring (Δ = 0.04, SE = 0.02), fall (Δ = −0.02, SE = 0.01), and winter (Δ = −0.04, SE = 0.02). Monthly percent change in height demonstrated a different pattern of change than what was seen in weight and zBMI with winter showing the largest change (Δ = 0.6, SE = 0.1), followed by fall (Δ = 0.5, SE = 0.1), summer (Δ = 0.3, SE = 0.1), and spring (Δ = 0.2, SE = 0.1). Sensitivity analyses were not different from the primary analyses.

Months During the School Year Without a Break Vs. Months With a Break ≥1 Week

Children's monthly percent weight change (Δ = 1.6, SE = 0.2 vs. Δ = 0.7, SE = 0.1) and zBMI change (Δ = −0.03, SE = 0.01 vs. Δ = 0.04, SE = 0.01) was greater during school year months with ≥1-week break from school compared with months during school without a 1-week break. The same pattern was not evident during school year months with ≥1-week break from school compared with months during school without a 1-week break for monthly percent height change (Δ = 0.4, SE = 0.0 vs. Δ = 0.3, SE = 0.1). Sensitivity analyses were not different from the primary analyses.

Health Behaviors

Changes in children's health behaviors are presented in Supplementary Figures S1–S9. Total sleep time was relatively constant across seasons at 411.9 (SE = 10.0) minutes in the fall, 408.6 (SE = 10.1) minutes in the winter, 415.1 (SE = 10.3) minutes in the spring, and 418.1 (SE = 10.9) minutes in the summer. Children's sleep onset time was also relatively constant at 21.5 (SE = 0.1) in the fall, 21.9 (SE = 0.1) in the winter, and 21.9 (SE = 0.1) in the spring. However, a large 1.7-hour shift in sleep onset time occurred in the summer with onset occurring at 23.2 (SE = 0.1). The same pattern was apparent for sleep offset time across the fall (6.5, SE = 0.1), winter (6.9, SE = 0.1), spring (6.8, SE = 0.1), and summer (8.2, SE = 0.1).

Children spent 546.7 (SE = 7.6) minutes sedentary in the fall, with sedentary increasing to 558.3 (SE = 7.7) minutes in the winter, then decreasing to 545.9 (SE = 7.9) minutes in the spring, and 531.5 (SE = 8.6) minutes in the summer. Light physical activity was the inverse of this pattern with children spending 307.5 (SE = 5.6), 296.6 (SE = 5.6), 312.0 (SE = 5.8), and 313.4 (SE = 6.2) minutes in light physical activity in the fall, winter, spring, and summer, respectively. MVPA also followed this pattern with children engaging in 47.9 (SE = 1.8), 44.0 (SE = 1.8), 49.2 (SE = 1.9), and 53.2 (SE = 2.0) minutes in MVPA in the fall, winter, spring, and summer, respectively.

Screen time displayed a sawtooth pattern with total daily screen time fluctuating from 129.2 (SE = 7.7) minutes in the fall, to 176.7 (SE = 7.9) minutes in the winter, to 127.3 (SE = 8.0) minutes in the spring, and 175.7 (SE = 8.4) minutes in the summer.

The mean number of healthy and unhealthy foods consumed per day was constant across seasons. Sensitivity analyses were not different from the secondary analyses.

Discussion

The primary aim of this study was to explore changes in children's body composition metrics through the lens of the CCRM and SDH. A secondary aim was to explore seasonal changes in health behaviors. This study has the potential to shed light on potential biological and social/behavioral mechanisms that may be driving accelerated summer BMI gain. This is critical given that accelerated summer BMI gain is poorly understood. These findings are critical understanding accelerated summer BMI gain and can inform when and how to intervene to prevent accelerated summer BMI gain. Furthermore, these findings can inform the field of childhood obesity about critical windows for intervention outside of the summer.

There are several key findings of the current study. First, children's percent weight change and zBMI change were the greatest during the summer. This is predicted by both the CCRM and the SDH. However, percent change in weight and change in zBMI were the highest in months when children had a break from school of ≥1 week. This is consistent with the SDH that claims weight gain, and subsequent zBMI,^8,34 will accelerate when children are exposed to relatively less structure (e.g., breaks from school like summer vacation). The SDH asserts that when children attend school, they are exposed to more structure, which will help regulate health behaviors and lead to less weight gain. This finding is consistent with past literature that has shown children's BMI gain accelerates during summer vacation, when children are generally not at school for extended periods of time (i.e., 10–12 weeks).^35–37 Some of the most compelling evidence for this pattern has been produced over the last 3 years when schools were closed due to the COVID-19 pandemic.

A number of studies have shown that BMI gain for children accelerated during this time period.^38–43 This study extends and corroborates findings from previous studies by demonstrating the same trends when children have breaks from school. This is important because it confirms that periods of time with relatively less structure may be an important mechanism that could be targeted for intervention to prevent childhood obesity. Furthermore, it identifies important windows for intervention outside of summer like winter and spring break from school.

This study showed a pattern of height and weight change that more closely reflects changes predicted by the SDH. Given BMI is a ratio of weight-to-height [i.e., weight (kg)/(height[m])²], changes in height, weight, and/or the combination of the two drive changes in BMI metrics. The SDH posits height gain throughout the year is constant while weight gain fluctuates based on when children are exposed to more (e.g., school days) or less (e.g., breaks from school) structure. According to the CCRM, seasonal rhythmicity in chronobiology due to fluctuations in daylight (longer days during summer and shorter during winter) leads to seasonal fluctuations in height and weight gain.⁷ Specifically, children's height gain accelerates in the late spring and early summer and decelerates in fall and winter, whereas children's weight gain accelerates in the late summer and early fall and decelerates in the winter and spring. In the current sample, accelerations and decelerations in height gain predicted by the CCRM were largely absent. Furthermore, percent weight change in the current sample peaked in March, following spring break, and June/July, during summer break, and the third highest change was during January, immediately following winter break from school.

Contrary to the CCRM, children's height gain was the greatest during the fall and winter, and the lowest in the spring and summer. These results are consistent with previous findings suggesting that children increase their height at a faster rate during the school year compared with the summer.⁴⁴ Weight gain was also the lowest in the winter while peaking in the spring and summer. The data collected herein suggest that during summer the mechanisms posited by the CCRM are operating in concert with and compounding those proposed by the SDH (i.e., biologically driven gains in weight are exacerbated in the summer by the reduction in exposure to structure). However, these data demonstrate that the biological mechanisms slowing weight gain in the winter may be overridden due to the extended break from school for winter break. These data suggest that regardless of season, breaks from school are key times to intervene on children's BMI.

This study found that sleep timing and increases in screen time may be behavioral mechanisms that are driving accelerated summer BMI gain. Sleep timing was relatively constant across fall, winter, and spring but shifted ∼90 minutes later during the summer. Screen time was greatest in the summer and winter compared with other seasons. A similar shift in sleep timing has been observed during summer among 5–8-year-olds.⁴⁵ A similar but less pronounced shift in later sleep timing was also observed during the winter. These findings are similar to past research that has found sleep duration is longer by about 10–20 minutes when children are on break from school (summer and spring break) compared with times when children are attending school.^28,46–48 These findings may be explained by social jetlag, which is characterized by shifting sleep later on days that are free from social obligations like school.⁴⁹ Social jetlag is negatively related to a variety of markers of adiposity.⁴⁹ These findings are complicated to explain in the context of summer BMI gain; on the one hand longer sleep duration is protective against childhood obesity,⁵⁰ while on the other hand later bed and wake times have been associated with increased risk for obesity in youth.^51–53

The reason for this relationship between sleep timing and accelerated BMI gain during summer may be explained by the CCRM, which posits that circadian misalignment due to shifting social demands, is a risk factor for BMI gain.⁷ At least one study has found later bed times were associated with increased daily screen use,⁵⁴ which is consistent with our findings. This suggests that the benefits of longer sleep duration may be overridden by the risk for obesity incurred by later bed and wake times, which could be related to increased screen use. There is a suggestive precedence for this in the literature as past studies have shown that, even while holding sleep duration constant, children who go to bed and wake later are at increased risk of obesity.⁵⁵ Thus the benefits of longer sleep may be outweighed by the risks of later sleep timing.

Physical activity in addition to health and unhealthy foods and drinks consumed were relatively constant throughout the year with slight reductions in sedentary time and increases in activity at all intensities during the summer. This finding is contradictory with previous research that has shown increases in sedentary time and decreases in physical activity during breaks from school.^8,28,34,56 It is unclear why the findings herein are contradictory, especially given that, consistent with past research, increases in weight and zBMI were observed in this sample. This may suggest that changes in sleep timing are more critical for maintaining weight than physical activity.

The current study found that changes in food and drink consumption were minimal across seasons. Studies that have examined changes in diet during summer vacation in the past have produced mixed findings.^28,56 One study found that children ate more fruit during the summer vacation²⁸ while the other found the opposite.⁵⁶ The findings here are possibly due to the use of parent proxy report and food frequency questionnaires. Food frequency questionnaires may not be sensitive enough to capture changes in children's diet between seasons. Furthermore, it is likely that parents are less aware of the foods and drinks children consume on days that they are in the care of others, like school days. Finally, data presented herein only represent the number of foods and drinks consumed. Thus, there is no information on serving sizes. Thus, more sensitive measures that are not dependent on parent report may be necessary in future studies. Nevertheless, a healthy diet is critical to children's health, but it is unclear the role diet plays in accelerated summer BMI gain.

While sleep timing and increased screen time may explain accelerated summer BMI gain, these patterns are less clear during the 9 months of the academic year. For instance, screen time was lower in spring than winter but weight gain was higher in spring and lowest in winter. Furthermore, sleep timing was similar in spring and winter, yet BMI gain was greater in the spring. It is important to note that, in the current study, all behavioral measures in the fall, winter, and spring were completed while school was in session. Thus, it is impossible to ascertain what behaviors children engaged in during breaks from school when the SDH would hypothesize that health behaviors would deteriorate. At least from the SDH perspective, this may explain why behaviors during the school year do not align with the BMI and weight trajectories. It is also important to note that BMI gain is attributable to a complex web of behavioral, social, environmental, and genetic factors.^57,58 Thus, findings in this study may, at least partially, be driven by social, environmental, and genetic mechanisms that were not measured in the current study. Future studies examining accelerated BMI gain should consider these factors and incorporate measures that would tap into these mechanisms.

This study has several strengths. First, the study collected height and weight data monthly and behavioral data seasonally over a calendar year. To our knowledge, this is the first study to collect anthropometric data monthly, which allowed us to test the CCRM and SDH. Correspondingly, the use of the CCRM and SDH to guide conclusion is a strength as using theory to guide hypotheses is best practice in research.⁵⁹ Monthly data collection allowed for the comparison of these two theories' ability to explain fluctuations in weight, height, and zBMI, and provides a clearer picture of the mechanisms driving children's accelerated summer BMI gain. The number of validated behavioral measures collected on children's behaviors also is a strength of the study.

This study also has limitations. First, only one school in a southeastern metropolitan U.S. city was included in the study and students were primarily from traditionally minoritized race/ethnicities. Thus, follow-up studies that include larger and more diverse samples are needed. There was also significant participant dropout over the course of the 1-year study for the behavioral measures, and some dropouts, but to a lesser degree, for the height and weight measures. Missing data for the behavioral measures during the summer was especially concerning. Thus, it is possible that the dropouts were systematically different from those participants who remained in the study. This study is also limited because behavioral measures during fall, summer, and spring all occurred while school was in session. This limited the current study's ability to compare and contrast behaviors between the CCRM SDH. Future studies should seek to measure children's behaviors during each season while children are and are not on breaks from school. This would help unravel the behavioral mechanisms that are driving BMI gain during seasons outside of summer.

Finally, diet measures were not comprehensive and relied on proxy parent report. Furthermore, diet measures did not assess quantity or frequency of the foods consumed. Thus, it may not have captured important changes in diet across seasons.

This study sought to examine changes in children's body composition and related behaviors through the lens of the CCRM and SDH. While there were elements of both the CCRM and SDH evident in the data, it appears as if the SDH better explains children's BMI gain. Interventions to mitigate accelerated BMI gain during breaks from school may be warranted no matter the season. Consistent sleep timing and reduced screen time may be important intervention targets for preventing summer BMI gain.

Impact Statement

This study shed light on potential biological and social/behavioral mechanisms that may be driving accelerated summer body mass index gain through the Circadian and Circannual Rhythm Model (CCRM) and Structured Days Hypothesis (SDH). The study identified important windows for intervention outside of summer, like winter and spring break from school, and important levers for intervention like sleep and screen time.

Authors' Contributions

R.G.W.: Conceptualization Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Visualization, Writing—original draft, and Writing—review and editing; JW: Project administration, Writing—review and editing; OF: Project administration, and Writing—review and editing; BA: Funding acquisition, and Investigation Writing—review and editing; EA: Writing—review and editing; SB: Writing—review and editing; RD: Project administration and Writing—review and editing; HP: Project administration and Writing—review and editing; LVK: Project administration and Writing—review and editing; MB: Writing—review and editing; AB: Project administration and Writing—review and editing; LRO Project administration and Writing—review and editing; CDP Project administration and Writing—review and editing; JM Writing—review and editing

Disclaimer

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Funding Information*

Research reported in this publication was supported in part by the National Institute of General Medical Sciences Award Number P20GM130420. Olivia Finnegan was supported by National Institute of General Medical Sciences Award Number T32GM081740.

*Correction added on July 14, 2023 after first online publication of April 19, 2023: Funding Information was inadvertently incomplete. Funding Information has now been corrected.

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

Supplementary File S1

Supplementary Table S1

Supplementary Table S2

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Figure S7

Supplementary Figure S8

Supplementary Figure S9