A 4-Yr Mixed Longitudinal Study of Health Behaviors and Fat Mass Accrual during Adolescence and Early Adulthood : Medicine & Science in Sports & Exercise

Journal Logo

APPLIED SCIENCES

A 4-Yr Mixed Longitudinal Study of Health Behaviors and Fat Mass Accrual during Adolescence and Early Adulthood

MCCONNELL-NZUNGA, JENNIFER1; GABEL, LEIGH2; MACDONALD, HEATHER M.3,4; RHODES, RYAN E.5; HOFER, SCOTT M.6; NAYLOR, PATTI-JEAN5; MCKAY, HEATHER A.3,4

Author Information
Medicine & Science in Sports & Exercise: December 2022 - Volume 54 - Issue 12 - p 2178-2187
doi: 10.1249/MSS.0000000000003003
  • Free

Abstract

Overweight and obesity are defined as abnormal or excessive accumulation of body fat that poses a health risk (1). Fat mass accrual in childhood and adolescence contributes to overweight and obesity (2) and tracks into adulthood (3,4). Global prevalence of overweight and obesity in childhood and adolescence has increased significantly in the last four decades—from only 4% in 1975 to greater than 18% in 2016 (1). In Canada, nearly one in seven children have obesity, defined as a body mass index (BMI; body mass/height2) that is 2 standard deviations (SDs) above the age- and sex-specific mean (5,6). These global trends underscore the need to specify determinants of overweight and obesity and patterns of fat mass accrual more clearly in children and youth, so that we can use these data to design effective prevention approaches.

Across a host of genetic, environmental, biological, psychological, socioeconomic, and behavioral factors that influence fat mass in children and youth (7), physical activity (PA) (8), sedentary time (SED) (9), and dietary factors such as sugar-sweetened beverages (10) are widely studied correlates. Among these studies and others, BMI is the most common surrogate measure of body fatness. However, BMI cannot discriminate which proportion of fat versus fat-free mass contributes to total body mass (body mass in the BMI equation) (11). In pediatric studies, growth- and maturation-related changes in BMI often reflect increases in fat-free mass as compared with fat mass (12). Thus, accurate and reliable measurements of fat mass that discriminate fat mass from fat-free mass are critical to help clarify the influence of lifestyle factors on fat mass accrual during growth.

Dual-energy x-ray absorptiometry (DXA) is an objective and accurate tool to measure fat and fat-free mass in children and adolescents (13). To our knowledge, only eight longitudinal studies have examined fat mass accrual (by DXA) and PA (by accelerometry) during adolescence (14–21). Of these, only three controlled for maturational status or biological age (all were from the Iowa Bone Development Study) (16,18,20), and only one study considered the influence of energy intake (EI) on fat mass accrual (19). Assessing maturational status is critical as girls experience substantially greater fat mass accrual relative to boys during adolescence (22,23).

Therefore, we aimed to extend the body of knowledge that describes how fat mass is accrued during growth and maturation. To do so, we used objective measures of total body fat mass (DXA), PA, and SED (accelerometry) and considered the influence of EI. Importantly, we used prospective height data to identify age at peak height velocity (APHV) for each child; we aligned boys and girls on this common maturational landmark before comparing within or across sexes. Our primary objective was to evaluate sex- and maturity-related associations between PA, SED, EI, and fat mass accrual in a cohort of primarily Asian and White boys and girls. We hypothesized that 1) higher levels of total PA and moderate-to-vigorous PA (MVPA) would negatively predict fat mass accrual in boys and girls, and 2) increased SED and higher EI would positively predict fat mass accrual in boys and girls.

METHODS

Study design and participants

Participants were drawn from a cohort of healthy girls (n = 556) and boys (n = 515) age 8 to 12 yr at baseline who participated in the mixed-longitudinal University of British Columbia’s Healthy Bones Study (HBS) III. HBSIII was initially designed to determine the effects of weight-bearing PA on bone health during adolescence (24,25). The HBSIII cohort includes participants from three school-based studies conducted between 1999 and 2012 (24–26); participants were measured annually for up to 11 yr. We describe the cohort and our methods in detail elsewhere (27,28). In brief, children were enrolled in 1999, 2003, and 2009. Participants from the earlier cohorts were followed annually until 2011, whereas participants in the latter cohort were followed until 2012.

For the present analysis, we used height, weight, PA, EI, and fat mass data from 399 children measured annually between 2008 and 2012. Of the full cohort, n = 312 had PA data measured by accelerometry (all years) and fat mass data measured by DXA (all years, Fig. 1). Participants in our cohort were measured between one and four times, with 748 total observations. We refer to data acquired when PA was first measured (by accelerometry) as baseline. We obtained written informed consent from the parents or legal guardians, written assent from participants younger than 18 yr, and informed consent from participants 18 yr and older. The University of British Columbia’s Clinical Research Ethics Board approved the studies (H15-01194, H07-02013, H2-70537).

F1
FIGURE 1:
Participant inclusion diagram.

Health history and ethnicity

Parents completed a questionnaire at baseline from which we determined participants’ health history and ethnicity. Participants completed another health history questionnaire during annual visits, from which we noted any changes from previous years. No participants were identified as having a health condition that excluded them from our sample.

We determined each participant’s ethnicity based on their parents’ response to questions regarding parents’ and/or grandparents’ place of birth and their self-identified ethnicity (27). Participants were classified as Asian if both parents or three of four grandparents were born in Hong Kong, China, Japan, Taiwan, Philippines, Korea, or India; White if both parents or three of four grandparents were born in North America or Europe; and “other” if parents were of mixed or other ethnicities (e.g., African, First Nations).

Anthropometry

We measured each participant’s standing (in centimeters) and sitting height (in centimeters) using standard stretch stature methods, and recorded measurements to the nearest millimeter using a wall-mounted stadiometer (Seca Model 242, Hanover, MD). We measured body mass to the nearest 0.1 kg using a calibrated electronic scale (Seca Model 840) with participants dressed in light clothing. Trained research assistants took all measurements in duplicate (mean was used), unless differences were >0.4 cm or 0.2 kg, in which case they obtained a third measure (median was used). We calculated BMI (body mass (kg)/height (m2)) and report weight status for participants younger than 19 yr based on the World Health Organization BMI-for-age growth charts, where overweight is defined as between the 85th and 95th percentiles and obese above the 95th percentile. In participants 19 yr or older, a BMI between 25 and 30 kg·m−2 was classified as overweight and 30 kg·m−2 or greater as obese.

Total body fat mass

We determined total body fat mass (in kilograms) using DXA total body scans (QDR 4500W; Hologic, Inc., Waltham, MA). One trained technician acquired and analyzed all scans using standard manufacturer protocols (Hologic, Inc.). Spine and anthropomorphic phantoms were scanned daily to maintain quality assurance. The coefficient of variation in our laboratory for fat mass is 1.9% (UBC PA and Bone Health Research Group, unpublished data).

Maturity

As in previous analyses (27), we controlled for maturational differences between children of the same chronological age by estimating APHV in years as an indicator of biological maturity (29). To obtain APHV, we first fit an interpolating cubic spline to each participant’s height velocity data (using all height measurements obtained between 1999 and 2012). Magnitude of PHV was identified as growth in standing height per year (in centimeters per year) that occurred at APHV. Because of missing and mistimed height measurements surrounding APHV, the cubic spline method to determine APHV was only feasible in 157 participants. We used the Moore equation (29) and anthropometric data from the measurement occasion closest to normative APHV (approximately 11.6 yr in girls and 13.5 yr in boys [29]) to estimate APHV for remaining participants of White and other ethnicities, and ethnic-specific equations (unpublished) for Asian participants. Once we determined APHV for all participants, we subtracted APHV from chronological age at the time of measurement to generate maturity offset, a continuous measure of biological age (e.g., −1 yr equals 1 yr before attainment of APHV; +1 equals 1 yr after APHV). For descriptive purposes, we also assessed maturity at each visit using the method of Tanner; self-reported pubic hair stage for boys and self-reported breast stage for girls (30).

Energy intake

Trained researchers assisted participants to complete a 24-h dietary recall (31); data were entered into Food Processor SQL for Kids (Version 10.6; ESHA Research, Salem, OR). For these analyses, we report values for total daily EI (in kilocalories per day). Visual inspection of the data revealed 14 outlier observations from six girls (one with two outlier observations) and seven boys (seven measures below the 1st percentile and seven measures above the 99th percentile). Outliers under the 1st and over the 99th percentiles were imputed with the value of the 1st (747.1 kcal) and 99th percentiles (5703.5 kcal), respectively (32).

PA and SED

We measured PA and SED using accelerometers (ActiGraph GT1M, Pensacola, FL) with a 15-s epoch. We asked participants to position the accelerometer (attached to an elastic waist belt) at the iliac crest and to wear the device during all waking hours for 7 consecutive days, except during water-based activities (e.g., showering, swimming). We used KineSoft software (version 3.3.75; KineSoft, Loughborough, United Kingdom) to analyze all data. We included accelerometer data for those participants who recorded at least 10 h of data on 3 or more days (33), and defined nonwear time as 60 min of consecutive zero counts. We applied the Evenson cut points (34) to classify MVPA (>2296 counts per minute) and cut points >100 and <100 counts per minute to classify total PA and SED, respectively (35). We calculated total PA, MVPA, and SED as a percent of wear time.

Statistical analysis

Before modeling, we examined scatter plots for fat mass and its determinants against maturity offset for each participant. We used a biological age-as-time, sex-specific model to estimate annual change in fat mass in boys and girls, with biological age, or maturity offset. We considered polynomial multilevel models with effects up to cubic and estimated the models using Stata (version 16; StataCorp, College Station, TX). Time-varying predictors include both interindividual variation (i.e., between-person differences) and intraindividual variation (i.e., within-person differences over time) (36). We modeled total PA, MVPA, SED, and EI to represent the between-person and within-person effects on fat mass separately. In this analysis, TotalPABP, MVPABP, SEDBP, and EIBP contain the individual’s mean value across years and represent the level 2 between-person differences, whereas TotalPAWP, MVPAWP, SEDWP, and EIWP represent the level 1 within-person change from the person-mean–centered variable at each occasion (calculated by subtracting the covariate mean for that individual across years from the covariate observation on each measurement occasion [37]). We included random effects to allow each individual his/her own slope for the effect of maturity (37).

We began our analysis with an empty between-person model and built up to a final conditional model that included all time-invariant and time-varying predictors. We first fit an empty means random intercept model and used it to determine the amount of variance in fat mass at the between- and within-person levels. We used the intraclass correlation coefficient to determine the contribution of between-person differences and within-person fluctuation to the variation in fat mass. Next, with maturity offset as the time variable, we fit a fixed linear time random intercept model, a random linear time model (allowing each participant his or her own slope for the effect of maturity), and a random linear, fixed quadratic time model (allowing each participant’s slope to increase or decrease). We proceeded to investigate models with fixed cubic time parameters (as random cubic parameters cannot be estimated with only four measurement occasions [37]). We used Wald test P values and relative reduction in the deviance test (−2ΔLL) to determine significance of individual fixed effects and −2ΔLL and χ2 test of significance of random-effects variances and covariances between nested models.

We determined that a random linear, fixed cubic functional form was the best unconditional growth model to represent fat mass accrual across maturity for boys and a random linear, fixed quadratic functional form was the best unconditional growth model to represent fat mass accrual across maturity for girls. Next, we added baseline weight status (healthy weight or overweight/obese) then two ethnicity dummy variables (to represent the three ethnicity categories) to the models and tested the corresponding significance using the Wald test and relative reduction in the deviance test (−2ΔLL). Weight status predicted fat mass accrual in boys and girls. Ethnicity did not predict fat mass accrual in boys but did for girls, so we retained the fixed effect of ethnicity in girls’ models only. To adjust for body size, we added each participant’s within-person and between-person effects of height. Neither within- nor between-person height predicted fat mass accrual in boys or girls, so we did not retain height in the model.

Next, we developed a series of models to investigate the longitudinal associations between Total PA, MVPA, SED, and EI and fat mass accrual in boys and girls. For each of Total PA, MVPA, SED, and EI, we started with an empty means, random intercept model through to a random linear, fixed quadratic model in girls, and a random linear, fixed cubic model in boys to determine the amount of variance attributed to between- and within-person differences and how much individual change each covariate exhibited over maturity offset. We tested the significance of model improvement using the relative reduction in the deviance test (−2ΔLL) and χ2 test of significance.

Final models were fit separately for Total PA, MVPA, SED, and EI and included both within- and between-persons effects of Total PA, MVPA, or SED along with EI. We fit an additional model that included MVPA and SED, along with EI. We added interaction terms to examine potential moderation effects of Total PA, MVPA, SED, and EI by maturity and weight status. Interaction terms for Total PA, MVPA, SED, and EI did not significantly improve model fit (based on −2ΔLL and Akaike information criterion and Bayesian information criterion values). We visually inspected models using residual plots; diagnostics revealed adequate model fit.

RESULTS

The sample included 312 participants (174 girls; 56%) who were 9 to 21 yr of age (mean [SD], 14.6 [3.2] yr) at first accelerometry measurement, with 748 total observations (Table 1). Participants self-identified as Asian (47%), White (44%), or mixed or other ethnicity (9%). We provide participant characteristics for all demographics and covariates at first measurement in Table 2. Twelve percent of girls and 14% of boys were overweight, and 6% of girls and 9% of boys had obesity (by BMI [6]) at first assessment. Twenty-six percent of participants (14% girls, 41% boys) met the MVPA recommendation of 60 min·d−1 (38). Across measurement occasions, on average, participants accumulated 4 h·d−1 of total PA (mean [SD], 236.7 [69.2] min), 10 h·d−1 of SED, (mean [SD], 600.9 [95.5] min), and EI of 2441.7 (SD, 1054.7) kcal·d−1.

TABLE 1 - Number of participant observations by maturity offset (years from APHV) for boys and girls in the HBSIII cohort.
Maturity Offset Girls Boys
−4 1
−3 3 18
−2 18 26
−1 44 28
0 54 14
1 37 17
2 15 27
3 9 39
4 29 50
5 41 41
6 37 31
7 38 24
8 40 12
9 18 5
10 16 3
11 12
12 1
Total 412 336

TABLE 2 - Characteristics of boys and girls from the HBSIII cohort at first accelerometry measurement.
Girls (n = 174) Boys (n = 138)
Mean (SD) Min Max Mean (SD) Min Max
Age (yr) 14.4 (3.5) 9.5 21.3 14.9 (2.9) 9.6 21.6
No. Asian/White/Other 84/75/15 63/63/12
Maturity offset (yr) 2.9 (3.6) −2.7 10.6 1.7 (3.0) −4.1 9.3
APHV (yr) 11.5 (0.7) 9.4 14.1 13.1 (0.9) 10.9 15.9
Weight (kg) 49.5 (13.8) 22.2 87.5 58.0 (16.0) 27.8 108.6
Height (cm) 154.9 (11.4) 130.0 181.6 165.3 (14.5) 129.7 192.2
BMI (kg·m−2) 20.3 (3.8) 12.6 33.7 20.9 (3.5) 14.8 31.8
Body fat mass (kg) 13.8 (6.5) 2.9 37.5 10.9 (6.2) 3.7 38.8
Wear time (min·d−1) 834.6 (74.6) 655.9 1073.0 840.5 (69.7) 662.0 999.5
Total PA (min·d−1) 242.6 (62.9) 112.0 413.4 256.7 (78.6) 105.0 458.1
MVPA (min·d−1) 40.6 (17.5) 4.8 104.1 59.6 (26.7) 14.7 142.8
SED (min·d−1) 592.4 (101.7) 324.6 868.1 584.1 (105.3) 347.4 804.0
EI (kcal·d−1) 2130.9 (883.0) 747.1 5703.1 2891.5 (1196.5) 757.2 5703.5
No. Tanner stage 1/2/3/4/5 28/35/26/43/42 16/13/10/45/53
Maturity offset is years from APHV. Tanner breast stage for girls and pubic hair stage for boys.

We present our sex-specific models in Figure 2. Based on our mixed models for boys, which adjusted for maturity and weight status, average fat mass for a healthy weight boy was 8.9 kg (SE, 0.5 kg) at APHV. Boys’ fat mass trajectory was best described by a cubic relationship with maturity, such that fat mass increased by 0.4 kg·yr−1 (SE, 0.1 kg·yr−1) at APHV, with lower rates of accrual thereafter. Overweight/obese boys had an 11.2-kg (SE, 0.9 kg) greater fat mass than healthy weight boys across maturity.

F2
FIGURE 2:
Body fat mass unadjusted individual curves (thin, light gray lines) and adjusted group curves based on the mixed model (solid line) for boys (left) and girls (right). The vertical line indicates maturity offset (years from APHV) of 0. Mixed-model growth curve is adjusted for maturity and weight status in boys and represents the average curve for a healthy weight boy. For girls, mixed-model growth curve is adjusted for maturity, weight status, and ethnicity and represents the average curve for a healthy weight, White girl.

For girls, average fat mass was 10.2 kg (SE, 0.4 kg) for a healthy weight White girl at APHV. Fat mass accrual demonstrated a quadratic relationship with maturity, such that fat mass accrual was 1.4 kg·yr−1 (SE, 0.1 kg·yr−1) at APHV and slowed with increasing maturity (Fig. 2). Ethnicity significantly predicted fat mass, such that White girls had a 2.2-kg (SE, 0.6 kg) greater fat mass, on average, compared with Asian girls across adolescence. Overweight/obese girls had a 10.2-kg (SE 0.7 kg) greater fat mass than healthy weight girls across adolescence.

Influence of MVPA, SED, and EI on fat mass accrual

We provide results of the mixed models that examined the influence of Total PA, MVPA, SED, and EI (both within- and between-persons) on fat mass accrual for boys across maturity in Figure 3 and for boys and girls in Tables 3 and 4, respectively. As a percent of accelerometer wear time, average daily total PA was 28.1% for girls and 28.8% for boys, MVPA was 4.9% for girls and 6.9% for boys, whereas SED was 72.0% for girls and 71.2% for boys.

F3
FIGURE 3:
Boys’ fat mass unadjusted individual curves (thin, light gray lines) and adjusted mixed-model curves for participants in the upper (thick, solid black line) and lower (thick, dashed black line) quartiles of within-person change in Total PA (left; upper quartile ~+2.5%, lower quartile −2.5%, or change of ±6 min·d−1), MVPA (center; upper quartile ~+1%, lower quartile −1%, or change of ±6 min·d−1) and SED (right; upper quartile ~+2.5%, lower quartile −2.5%, or change of ±20 min·d−1). The vertical dashed line indicates maturity offset (years from APHV) of 0. Mixed-model growth curves reported in Table 3 are adjusted for maturity, EI, and either Total PA, MVPA, or SED, and represent the average curve for a healthy weight boy with between-person EI of 2500 kcal.
TABLE 3 - Boys’ model coefficients (95% confidence interval) body fat mass aligned on maturity offset (n = 138).
Model Including Total PA and EI Model Including MVPA and EI Model Including SED and EI Model Including MVPA, SED, and EI
Fixed effect
 Intercept 9.37 (8.39 to 10.35)* 9.38 (8.40 to 10.35)* 9.38 (8.40 to 10.35)* 9.37 (8.40 to 10.35)*
 Maturity offset 0.30 (−0.02 to 0.62) 0.43 (0.15 to 0.71)* 0.30 (−0.02 to 0.61) 0.41 (0.12 to 0.70)*
 Maturity offset2 −0.15 (−0.22 to −0.08)* −0.18 (−0.25 to −0.11)* −0.15 (−0.22 to −0.08)* −0.17 (−0.25 to −0.10)*
 Maturity offset3 0.02 (0.01 to 0.03)* 0.02 (0.01 to 0.03)* 0.02 (0.01 to 0.03)* 0.02 (0.01 to 0.03)*
 Weight status 11.02 (9.44 to 12.61)* 11.00 (9.41 to 12.52)* 11.03 (9.44 to 12.61)* 11.00 (9.41 to 12.51)*
 Total PA
 TotalPAWP −0.08 (−0.12 to −0.03)*
 TotalPABP −0.05 (−0.15 to 0.04)
 MVPA
 MVPAWP −0.23 (−0.35 to −0.12)* −0.21 (−0.36 to −0.05)*
 MVPABP −0.22 (−0.46 to 0.02) −0.23 (−0.47 to 0.01)
 SED
 SEDWP 0.08 (0.03 to 0.12)* 0.02 (−0.05 to 0.08)
 SEDBP 0.05 (−0.04 to 0.15)
 EI
 EIWP 0.00 (−0.03 to 0.03) −0.00 (−0.03 to 0.03) −0.00 (−0.03 to 0.03) −0.00 (−0.03 to 0.03)
 EIBP −0.15 (−0.21 to −0.08)* −0.14 (−0.21 to −0.08)* −0.15 (−0.21 to −0.08)* −0.14 (−0.21 to −0.08)*
Random effects
 Intercept 13.80 (9.18 to 20.73) 14.06 (9.31 to 21.21) 13.79 (9.18 to 20.73) 13.99 (9.29 to 21.09)
 Maturity offset 0.37 (0.16 to 0.86) 0.37 (0.16 to 0.85) 0.37 (0.16 to 0.86) 0.36 (0.16 to 0.84)
 Residual 2.84 (2.16 to 3.74) 2.70 (2.04 to 3.57) 2.84 (2.15 to 3.74) 2.71 (2.05 to 3.58)
Maturity is years from APHV. Weight status is classified as 0 for healthy weight and 1 for overweight or obese. EI represents change or difference per 100 kcal·d−1. WP variables represent level 1, within-person change and are person-mean centered. BP variables are level 2, between-person effects (an individual’s mean value) and are group-mean centered. Therefore, the model intercept represents the body fat mass value for a healthy weight boy at maturity offset of 0, MVPA of 7% of wear time, SED of 70% of wear time, and EI of 2500 kcal. SEDBP was removed from combined MVPA/SED model because of collinearity.
**P < 0.01.

TABLE 4 - Girls’ model coefficients (95% confidence interval) for body fat mass aligned on maturity offset (n = 174).
Model Including Total PA and EI Model Including MVPA and EI Model Including Sedentary Behavior and EI Model Including MVPA, SED, and EI
Fixed effect
 Intercept 9.91 (8.96 to 10.85)** 10.21 (9.35 to 11.08)** 9.91 (8.97 to 10.85)** 10.25 (9.38 to 11.12)**
 Maturity offset 1.45 (1.23 to 1.68)** 1.43 (1.24 to 1.62)** 1.45 (1.23 to 1.68)** 1.39 (1.19 to 1.59)**
 Maturity offset2 −0.06 (−0.08 to −0.04)** −0.06 (−0.08 to −0.04)** −0.06 (−0.08 to −0.04)** −0.05 (−0.07 to −0.03)**
 Weight status 9.97 (8.55 to 11.38)** 9.95 (8.56 to 11.34)** 9.97 (8.55 to 11.38)** 9.94 (8.56 to 11.33)**
 Ethnicity
  Asian −2.07 (−3.25 to −0.89)** −2.64 (−3.79 to −1.49)** −2.07 (−3.25 to −0.89)** −2.62 (−3.77 to −1.47)**
  Other −0.20 (−2.36 to 1.97) −0.06 (−2.75 to 1.47) −0.19 (−2.36 to 1.97) −0.59 (−2.70 to 1.53)
 Total PA
  TotalPAWP −0.03 (−0.07 to 0.02)
  TotalPABP 0.03 (−0.06 to 0.11)
 MVPA
  MVPAWP −0.06 (−0.18 to 0.07) −0.02 (−0.16 to 0.15)
  MVPABP −0.30 (−0.6 to −0.01)* −0.31 (−0.61 to −0.01)*
 SED
  SEDWP 0.03 (−0.02 to 0.07) 0.03 (−0.25 to 0.08)
  SEDBP −0.03 (−0.11 to 0.06)
 EI
  EIWP −0.00 (−0.03 to 0.02) −0.00 (−0.03 to 0.02) −0.00 (−0.03 to 0.02) −0.00 (−0.03 to 0.02)
  EIBP −0.07 (−0.14 to 0.01) −0.07 (−0.15 to −0.00)* −0.07 (−0.14 to 0.01) −0.07 (−0.15 to 0.00)
Random effects
 Intercept 8.14 (6.23 to 10.64) 7.76 (5.91 to 10.19) 8.14 (6.23 to 10.64) 7.73 (5.89 to 10.15)
 Maturity offset 0.06 (0.03 to 0.16) 0.07 (0.03 to 0.16) 0.06 (0.03 to 0.16) 0.07 (0.03 to 0.16)
 Residual 1.90 (1.58 to 2.30) 1.92 (1.59 to 2.31) 1.90 (1.58 to 2.30) 1.91 (1.59 to 2.31)
Maturity is years from APHV. Weight status is classified as 0 for healthy weight and 1 for overweight or obese. Ethnicity represents 0 for White, 1 for Asian, and 2 for other/mixed; EI represents change or difference per 100 kcal·d−1. WP variables represent level 1, within-person change and are person-mean centered. BP variables are level 2, between-person effects (an individual’s mean value) and are group-mean centered. Therefore, the model intercept represents the body fat percent value for a healthy weight White girl at maturity offset of 0, MVPA of 5% of wear time, SED of 70% of wear time, and EI of 2500 kcal. SEDBP was removed from combined MVPA/SED model because of collinearity.
*P < 0.05.
**P < 0.01.

For boys, within-person change in Total PA and MVPA negatively predicted fat mass accrual, indicating each percent increase in Total PA or MVPA above an individual’s average Total PA or MVPA (increase of 2 min·d−1 or Total PA or 6 min·d−1 of MVPA) was associated with a 0.08- or 0.23-kg lower fat mass, respectively (Table 3, Fig. 3). Between-person Total PA or MVPA (an individual’s average Total PA or MVPA across maturity) was not associated with fat mass accrual, although there was a trend for a negative relationship between person MVPA and fat mass accrual (P = 0.07). In the SED-only model, within-person change in SED positively predicted fat mass accrual; each 1% increase in SED above the individual’s average SED (increase of 6 min·d−1 of SED) predicted a 0.08-kg greater fat mass accrual (Table 3, Fig. 3). Between-person SED (an individual’s average SED across maturity) did not predict fat mass accrual. In combined models that included within-person change in MVPA and SED (between-person effects of SED excluded because of collinearity), change in MVPA persisted as a negative predictor of fat mass accrual, whereas SED was no longer associated with fat mass accrual (Table 3).

For all models in boys, within-person change in EI was not related to fat mass accrual. However, between-person EI demonstrated a negative relationship with fat mass accrual, such that boys with lower EI across adolescence had significantly greater fat mass accrual than boys with greater EI.

For girls, between-person average MVPA significantly negatively predicted fat mass accrual (Table 4), indicating that each percent (or 4 min·d−1) greater average MVPA across adolescence was associated with a 0.3% lower fat mass. Within-person change in Total PA and MVPA and between-person Total PA were not associated with fat mass accrual. Similarly, neither between- nor within-person change in SED predicted fat mass accrual (Table 4). Between-person MVPA persisted as a negative predictor of fat mass accrual in combined models that included MVPA and SED (between-person effects of SED excluded because of collinearity; Table 4). Within-person change in EI was not related to girls’ fat mass accrual (Table 4). However, between-person EI demonstrated a negative relationship with fat mass accrual in the MVPA only model, indicating that greater average EI was associated with lower fat mass accrual across adolescence.

DISCUSSION

In this mixed longitudinal study, MVPA negatively predicted fat mass accrual as adolescents advanced through young adulthood. Relationships between MVPA and fat mass were independent of SED, suggesting that increasing MVPA during childhood and adolescence may shift fat mass accrual trajectories downward. This study extends previous investigations that examined the influence of lifestyle factors on fat mass accrual during growth by aligning boys and girls on a common maturational landmark so as to enable comparisons between and within boys and girls of different chronological ages. We also incorporated objective measures of PA and SED and self-reported dietary EI. Our cohort of primarily Asian and White Canadian children and youth provide insight into how fat mass is accrued across maturity, and the influence of PA, including MVPA independent of SED. We note important sex differences in fat mass accrual across maturation and sex-specific influences of Total PA, MVPA, SED, and EI on change in fat mass.

Sex differences in fat mass accrual across maturation

Most boys and girls in our study had healthy amounts of body fat (39). The sex-specific pattern of fat mass accrual across maturation in the HBSIII cohort mirrored trajectories of fat mass accrual reported in previous studies (23,40) and is consistent with the known influence of endocrine factors on body composition during puberty (41). Girls experience a steady increase in fat mass during adolescence, whereas boys’ fat mass plateaus mid-puberty (concurrent with steep increases in fat-free mass). As we followed participants annually for up to 11 yr, our growth curves extend from adolescence through early adulthood (5 yr post-APHV) and into young adulthood. Fat mass accrual plateaued in boys during adolescence and increased slightly during young adulthood. Fat mass continued to increase in girls for 10 yr after APHV. These data are similar to Saskatchewan Pediatric Bone Mineral Accrual Study reports (42). However, our cohort included fewer observations for boys beyond 8 yr after APHV (early adulthood), which may have skewed our curves for boys. The inherently different trajectories for fat mass accrual in boys and girls through adolescence and into young adulthood underscore the need for sex-specific analyses in these kinds of longitudinal studies.

Physical activity

In line with our hypothesis, Total PA and MVPA negatively predicted fat mass accrual in boys in the HBSIII cohort, who were more physically active as compared with population-based Canadian data for adolescents age 15–19 yr (43). The relationship between Total PA and fat mass was largely driven by MVPA, not light PA, and we therefore focus on MVPA in this discussion. From our models (MVPA and EI model), an annual increase in MVPA of 12 min·d−1 (increase of ~2% of wear time) would generate an annual decrease in fat mass of 0.46 kg for boys, with EI held constant. Taken over time (5 yr of “adolescence”), this difference equates to accumulation of a 2.3-kg less body fat. Even a modest annual increase of 2 min·d−1 of MVPA over time (5 yr of “adolescence”) would equate to a 0.38-kg lower body fat. This finding and those of others (17,20,23) highlight MVPA as an important target for interventions aimed at preventing overweight and obesity during adolescence and into young adulthood. PA interventions in the upper quartile of effectiveness have increased MVPA by as much as 9.6 min·d−1 in children and adolescents (44). Similarly, PA interventions can reduce the risk of obesity in children and adolescents (45); however, most intervention studies focused on change in BMI or BMI z-scores as the primary outcome. Future trials that investigate how PA programs influence fat mass trajectories and distribution of fat mass in children and adolescents are warranted.

For girls, only average MVPA across adolescence was associated with fat mass accrual. Assuming all other variables in our model are held constant, a girl at APHV who participates in an average of 60 min·d−1 of MVPA across adolescence, as recommended by current guidelines (38) (which is 20 min·d−1 greater than the average for girls observed in our cohort), would have a 1.5-kg (or 15%) lower fat mass compared with her peers engaging in only 40 min·d−1 of MVPA. Both boys and girls in the Iowa Bone Development Study demonstrated an inverse relationship between fat mass accrual (by DXA) and MVPA (by accelerometry) (20). However, investigators did not report between- and within-person differences for the time-varying covariates, and average fat mass was higher among girls in the Iowa study as compared with girls in the HBSIII cohort, so it is not possible to directly compare their data with data from our study. Despite these differences, MVPA was lower in girls as compared with boys across maturity in both studies. In our cohort, girls engaged in 20 min·d−1 less MVPA compared with boys, on average, across maturity. Girls’ daily total PA declined by approximately 18 min·yr−1 until plateauing in late adolescence (~6 yr after APHV), whereas daily MVPA declined by approximately 2 min·yr−1 across adolescence. In contrast, boys’ daily total PA was greater than girls’ across maturity, and declines were less steep (approximately 16 min·yr−1 until plateauing in late adolescence; ~7 yr after APHV). Similarly, boys’ daily MVPA was significantly greater than girls’ across maturity, and declines were less severe (declines of approximately 1 min·yr−1 until adulthood). Although a myriad of factors influence fat mass accrual, this sex difference in PA is notable and confirms similar trends in population-based Canadian data (43). Importantly, fat mass as a proportion of total body mass (percent body fat) for girls in our study was considered “healthy” (39). However, adequate PA during growth and maturation generates health benefits (e.g., cardiovascular health, mental health, bone health) that extend far beyond fat mass accumulation.

Sedentary time

SED did not predict fat mass accrual in boys or girls independent of MVPA. Similarly, the Iowa study (18,20) found that SED (by accelerometry) was not associated with fat mass accrual in boys or girls (by DXA) across adolescent growth after accounting for MVPA. Janz and colleagues (20) also examined the unique influence of children’s TV viewing habits on fat mass. For boys and girls, TV time (by parental proxy) positively predicted absolute fat mass; mixed models suggested that a 1-h increase in TV time would increase fat mass by 2.5% in boys and 1.6% in girls (20). TV viewing may exemplify low energy expenditure and may also be a surrogate for unhealthy eating habits (46) and change in sleep patterns (47). Therefore, multitargeted interventions may be most appropriate to alter unhealthy fat mass trajectories.

Energy intake

Energy in—energy out hypotheses (48) suggest that increases in EI beyond what is needed to serve daily energy expenditures may result in storage of excess energy as body fat. However, the true “healthy weight” picture encompasses genetic predisposition, the endocrine environment, the microbiome, and many other systems and is much more complex. In our cohort, average daily EI for boys and girls fell within what is considered a healthy/normal range for physically active children and youth (49), but was higher than the national average (50). Relatively few (18% of girls and 23% of boys) had overweight or obesity, a proportion that is below the national average of 29% and 33% for Canadian girls and boys, respectively, age 6 to 17 yr (6). Dietary recall can be faulty, particularly among adolescents (51). An estimated 6 to 9 d of dietary recall provides a reliable, accurate estimate of usual dietary intake in youth (52)—as compared with the one measure we acquired for our study. Future studies might explore digital dietary assessment methods to assess EI in children and youth. We also did not examine the influence of specific macronutrient or micronutrients (e.g., carbohydrates, calcium) that may be more strongly associated with adiposity than total EI. These factors would all have contributed to the associations (or lack thereof) we observed between EI and fat mass accrual in boys and girls.

Strengths and limitations

We have noted the strengths of our study throughout the discussion. They include that we are among very few studies to use objective measures of fat mass (by DXA) and PA and SED (by accelerometry) to monitor fat mass accrual over time. Similarly, few studies have longitudinal growth data across enough years to calculate maturity offset as a common maturity landmark for boys and girls, as we did. This allowed us to compare children and youth across a range of sizes and shapes at the same maturational time point, regardless of their chronological age. Our multilevel modeling approach allowed us to evaluate the complex interplay between growth and maturation and lifestyle factors (PA, SED, and EI) on fat mass accrual.

Our study also has several limitations. First, we did not account for regional distribution of body fat, which is strongly associated with health outcomes in children and youth (53). Second, because participants in our study were early to peripubertal at study entry, it is unclear if relationships between fat mass and Total PA, MVPA, SED, and EI differ at earlier biological ages (i.e., >2 yr before APHV). Third, annual 24-h dietary recalls may not accurately represent typical patterns of food consumption (54). Finally, participants did not wear uniaxial accelerometers during water-based activities such as swimming, and accelerometers are also unable to accurately capture certain activities (i.e., bicycling or carrying loads), nor do they distinguish between sedentary postures such as sitting, standing, and lying down. Therefore, some activities were likely missed in our measures of PA, and standing time may have been included in estimates of SED.

CONCLUSIONS

Patterns of fat mass accrual may reflect healthy weight trajectories, as children grow, mature, and approach adulthood. Given the known relationship between unhealthy body weight and myriad health problems throughout life, insights into how fat mass is accrued and its relationship to lifestyle factors provide opportunities to intervene with PA and healthy eating strategies to support healthy body weights. We note a sex difference in how fat mass is accrued and how fat mass accrual may be influenced by PA. It seems important to better understand sex and gendered influences on PA and eating behaviors and how they, together, influence fat mass accrual and subsequently body weight (55). Multifaceted, sex/gender-specific interventions and analyses that consider the many systems, physiological and sex and gendered influences on fat mass trajectories during growth and maturation, may be the way of the future.

The authors gratefully acknowledge the Healthy Bones Study III participants, their families, and the support from principals and teachers at participating schools in the Richmond and Vancouver School Districts. The authors also acknowledge the dedication of the Healthy Bones Study III research teams, the efforts of the research coordinator, Douglas Race, and the supervision of imaging acquisition and processing from Dr. Danmei Liu (Centre for Hip Health and Mobility, Vancouver Coastal Research Institute).

This work was supported by funding from the Canadian Institutes of Health Research (MOP-84575). H. A. M. was supported by the British Columbia Health Research Foundation (2400-2 and 10898-2) and the Michael Smith Foundation for Health Research. L. G. was supported by a Canadian Institutes of Health Research Doctoral Research Award. The authors have no conflicts of interest. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The results of the present study do not constitute endorsement by the American College of Sports Medicine.

REFERENCES

1. World Health Organization. Obesity and Overweight [Internet]. Geneva (Switzerland): World Health Organization. Available from: https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight. Accessed January 15, 2022.
2. Helba M, Binkovitz LA. Pediatric body composition analysis with dual-energy x-ray absorptiometry. Pediatr Radiol. 2009;39(7):647–56.
3. Eriksson J, Forsen T, Osmond C, Barker D. Obesity from cradle to grave. Int J Obes Relat Metab Disord. 2003;27(6):722–7.
4. Barbour-Tuck E, Erlandson M, Muhajarine N, Foulds H, Baxter-Jones A. Influence of childhood and adolescent fat development on fat mass accrual during emerging adulthood: a 20-year longitudinal study. Obesity (Silver Spring). 2018;26(3):613–20.
5. WHO Multicentre Growth Reference Study Group. WHO child growth standards based on length/height, weight and age. Acta Paediatr Suppl. 2006;450:76–85.
6. Rao DP, Kropac E, Do MT, Roberts KC, Jayaraman GC. Childhood overweight and obesity trends in Canada. Health Promot Chronic Dis Prev Can. 2016;36(9):194–8.
7. Di Cesare M, Soric M, Bovet P, et al. The epidemiological burden of obesity in childhood: a worldwide epidemic requiring urgent action. BMC Med. 2019;17(1):212.
8. Reichert FF, Baptista Menezes AM, Wells JC, Carvalho Dumith S, Hallal PC. Physical activity as a predictor of adolescent body fatness: a systematic review. Sports Med. 2009;39(4):279–94.
9. Biddle SJ, García Bengoechea E, Wiesner G. Sedentary behaviour and adiposity in youth: a systematic review of reviews and analysis of causality. Int J Behav Nutr Phys Act. 2017;14(1):43.
10. Malik VS, Pan A, Willett WC, Hu FB. Sugar-sweetened beverages and weight gain in children and adults: a systematic review and meta-analysis. Am J Clin Nutr. 2013;98(4):1084–102.
11. Prentice AM, Jebb SA. Beyond body mass index. Obes Rev. 2001;2(3):141–7.
12. Horlick M. Body mass index in childhood—measuring a moving target. J Clin Endocrinol Metab. 2001;86(9):4059–60.
13. Zanini Rde V, Santos IS, Chrestani MA, Gigante DP. Body fat in children measured by DXA, air-displacement plethysmography, TBW and multicomponent models: a systematic review. Matern Child Health J. 2015;19(7):1567–73.
14. Janz KF, Burns TL, Levy SM; Iowa Bone Development Study. Tracking of activity and sedentary behaviors in childhood: the Iowa bone development study. Am J Prev Med. 2005;29(3):171–8.
15. Stevens J, Suchindran C, Ring K, et al. Physical activity as a predictor of body composition in American Indian children. Obes Res. 2004;12(12):1974–80.
16. Janz KF, Kwon S, Letuchy EM, et al. Sustained effect of early physical activity on body fat mass in older children. Am J Prev Med. 2009;37(1):35–40.
17. Riddoch CJ, Leary SD, Ness AR, et al. Prospective associations between objective measures of physical activity and fat mass in 12–14 year old children: the Avon Longitudinal Study of Parents and Children (ALSPAC). BMJ. 2009;339:b4544.
18. Kwon S, Burns TL, Levy SM, Janz KF. Which contributes more to childhood adiposity—high levels of sedentarism or low levels of moderate-through-vigorous physical activity? The Iowa Bone Development Study. J Pediatr. 2013;162(6):1169–74.
19. Hjorth MF, Chaput JP, Ritz C, et al. Fatness predicts decreased physical activity and increased sedentary time, but not vice versa: support from a longitudinal study in 8- to 11-year-old children. Int J Obes (Lond). 2014;38(7):959–65.
20. Janz KF, Boros P, Letuchy EM, Kwon S, Burns TL, Levy SM. Physical activity, not sedentary time, predicts dual-energy x-ray absorptiometry-measured adiposity age 5 to 19 years. Med Sci Sports Exerc. 2017;49(10):2071–7.
21. Aars NA, Beldo S, Jacobsen BK, et al. Association between objectively measured physical activity and longitudinal changes in body composition in adolescents: the Tromsø study Fit Futures cohort. BMJ Open. 2020;10(10):e036991.
22. O’Keeffe LM, Frysz M, Bell JA, Howe LD, Fraser A. Puberty timing and adiposity change across childhood and adolescence: disentangling cause and consequence. Hum Reprod. 2020;35(12):2784–92.
23. Mundt CA, Baxter-Jones AD, Whiting SJ, Bailey DA, Faulkner RA, Mirwald RL. Relationships of activity and sugar drink intake on fat mass development in youths. Med Sci Sports Exerc. 2006;38(7):1245–54.
24. MacKelvie KJ, McKay HA, Khan KM, Crocker PR. A school-based exercise intervention augments bone mineral accrual in early pubertal girls. J Pediatr. 2001;139(4):501–8.
25. Macdonald HM, Kontulainen SA, Khan KM, McKay HA. Is a school-based physical activity intervention effective for increasing tibial bone strength in boys and girls?J Bone Miner Res. 2007;22(3):434–46.
26. McKay HA, MacLean L, Petit M, et al. “Bounce at the Bell”: a novel program of short bouts of exercise improves proximal femur bone mass in early pubertal children. Brit J Sports Med. 2005;39(8):521–6.
27. Gabel L, Macdonald HM, Nettlefold L, McKay HA. Physical activity, sedentary time, and bone strength from childhood to early adulthood: a mixed longitudinal HR-pQCT study. J Bone Miner Res. 2017;32(7):1525–36.
28. McConnell-Nzunga J, Naylor PJ, Macdonald HM, Rhodes RE, Hofer SM, McKay HA. Body fat accrual trajectories for a sample of Asian–Canadian and Caucasian–Canadian children and youth: a longitudinal DXA-based study. Pediatr Obes. 2020;15(1):e12570.
29. Moore SA, McKay HA, Macdonald H, et al. Enhancing a somatic maturity prediction model. Med Sci Sports Exerc. 2015;47(8):1755–64.
30. Tanner JM. Foetus into Man. Cambridge (MA): Harvard University Press; 1978.
31. Lytle LA, Murray DM, Perry CL, Eldridge AL. Validating fourth-grade students’ self-report of dietary intake: results from the 5 A Day Power Plus program. J Am Diet Assoc. 1998;98(5):570–2.
32. Guttman I, Smith DE. Investigation of rules for dealing with outliers in small samples from the normal distribution: I: estimation of the mean. Dent Tech. 1969;11(3):527–50.
33. Mattocks C, Ness A, Leary S, et al. Use of accelerometers in a large field-based study of children: protocols, design issues, and effects on precision. J Phys Act Health. 2008;5(1 Suppl):S98–111.
34. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG. Calibration of two objective measures of physical activity for children. J Sports Sci. 2008;26(14):1557–65.
35. Trost SG, Loprinzi PD, Moore R, Pfeiffer KA. Comparison of accelerometer cut points for predicting activity intensity in youth. Med Sci Sports Exerc. 2011;43(7):1360–8.
36. Hoffman L. Longitudinal Analysis. New York, NY:Routledge; 2015. p. 654.
37. Curran PJ, Bauer DJ. The disaggregation of within-person and between-person effects in longitudinal models of change. Annu Rev Psychol. 2011;62:583–619.
38. Tremblay MS, Carson V, Chaput J-P. Introduction to the Canadian 24-hour movement guidelines for children and youth: an integration of physical activity, sedentary behaviour, and sleep. Appl Physiol Nutr Metab. 2016;41(6 Suppl 3):iii–v.
39. Williams DP, Going SB, Lohman TG, et al. Body fatness and risk for elevated blood-pressure, total cholesterol, and serum-lipoprotein ratios in children and adolescents. Am J Public Health. 1992;82(3):358–63.
40. van Lenthe FJ, van Mechelen W, Kemper HC, Post GB. Behavioral variables and development of a central pattern of body fat from adolescence into adulthood in normal-weight Whites: the Amsterdam Growth and Health Study. Am J Clin Nutr. 1998;67(5):846–52.
41. Loomba-Albrecht LA, Styne DM. Effect of puberty on body composition. Curr Opin Endocrinol Diabetes Obes. 2009;16(1):10–5.
42. Barbour-Tuck E, Erlandson M, Muhajarine N, Foulds H, Baxter-Jones A. Longitudinal patterns in BMI and percent total body fat from peak height velocity through emerging adulthood into young adulthood. Am J Hum Biol. 2018;30(1):23056.
43. Canadian Fitness and Lifestyle Research Institute. Bulletin 8: Achieving Sufficient Steps per Day among Canadian Children and Youth. Ottawa (ON): Canadian Fitness and Lifestyle Research Institute; 2018. Available from: Canadian Fitness and Lifestyle Research Institute Available from: https://cflri.ca/bulletin-8-achieving-sufficient-steps-day-among-canadian-children-and-youth. Accessed January 10, 2022.
44. Wright CE, Rhodes RE, Ruggiero EW, Sheeran P. Benchmarking the effectiveness of interventions to promote physical activity: a metasynthesis. Health Psychol. 2021;40(11):811–21.
45. Brown T, Moore TH, Hooper L, et al. Interventions for preventing obesity in children. Cochrane Database Syst Rev. 2019;7(7):CD001871.
46. Liang T, Kuhle S, Veugelers PJ. Nutrition and body weights of Canadian children watching television and eating while watching television. Public Health Nutr. 2009;12(12):2457–63.
47. Owens J, Maxim R, McGuinn M, Nobile C, Msall M, Alario A. Television-viewing habits and sleep disturbance in school children. Pediatrics. 1999;104(3):e27.
48. Hill JO, Wyatt HR, Peters JC. Energy balance and obesity. Circulation. 2012;126(1):126–32.
49. Gidding SS, Dennison BA, Birch LL, et al. Dietary recommendations for children and adolescents: a guide for practitioners: consensus statement from the American Heart Association. Circulation. 2005;112(13):2061–75.
50. Statistics Canada. Table 13-10-0772-01 Total Energy Intake from Foods, by Dietary Reference Intake Age-Sex Group, Household Population Age 1 and Over, Canadian Community Health Survey (CCHS)—Nutrition, Canada and Provinces. Ottawa (ON): Statistics Canada; 2017. Available from: Statistics Canada. Available from: https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1310077201. Accessed January 10, 2022.
51. Livingstone MB, Robson PJ, Wallace JM. Issues in dietary intake assessment of children and adolescents. Br J Nutr. 2004;92(2 Suppl):S213–22.
52. Ollberding NJ, Couch SC, Woo JG, Kalkwarf HJ. Within- and between-individual variation in nutrient intake in children and adolescents. J Acad Nutr Diet. 2014;114(11):1749–58.e5.
53. Daniels SR, Morrison JA, Sprecher DL, Khoury P, Kimball TR. Association of body fat distribution and cardiovascular risk factors in children and adolescents. Circulation. 1999;99(4):541–5.
54. Beaton GH, Milner J, Corey P, et al. Sources of variance in 24-hour dietary recall data: implications for nutrition study design and interpretation. Am J Clin Nutr. 1979;32(12):2546–59.
55. Shah B, Tombeau Cost K, Fuller A, Birken CS, Anderson LN. Sex and gender differences in childhood obesity: contributing to the research agenda. BMJ Nutr Prev Health. 2020;3(2):387–90.
Keywords:

ADIPOSITY; CHILDREN; YOUTH; PHYSICAL ACTIVITY; SEDENTARY TIME

Copyright © 2022 by the American College of Sports Medicine