Group-Based Fat Mass Index Trajectories in Boys and Girls from Elementary to High School

Abstract

Background:

There is a pressing need to understand developmental patterns of adiposity as well as factors that influence it. Group-based trajectories, usually based on body mass index (BMI), have been used in adults and children to investigate adiposity trajectories. The objectives of this study were to identify sex-specific groups with similar longitudinal trajectories for fat mass index (FMI) and to compare sex-specific longitudinal trajectory groups on demographic and behavioral characteristics.

Methods:

This study employed a longitudinal design to examine patterns of age-related change in FMI-assessed adiposity in children transitioning from elementary to high school. FMI, moderate-to-vigorous physical activity (MVPA), sedentary behavior, and diet quality were measured at 5^th, 6^th,7^th, 9^th, and 11^th grades when the children were 10–16 years of age. Sex-specific group-based trajectory analyses were conducted for FMI, and repeated ANOVAs were used to compare behaviors between trajectories.

Results:

Both boys’ (n = 320) and girls’ (n = 389) group-based trajectory analysis resulted in three groups, each with a low FMI and stable trajectory. In girls, two trajectories increased from 10 to 16 years of age, while in boys, one trajectory increased, and one decreased over time. In girls only, the groups differed by race and parent education. There were also group differences in MVPA for boys and in diet quality for girls.

Conclusions:

An accurate assessment of adiposity, FMI, was used to establish the trajectories in this study. The characteristics of the groups provide valuable information about when and for whom to target obesity prevention efforts.

Keywords

adiposity children diet longitudinal physical activity

Introduction

Overweight and obesity are associated with adverse health outcomes in adults^1,2 and youth.^3,4 Risk factors for overweight and obesity can appear during childhood,^4,5 and childhood obesity can track into adulthood.^6,7 The increased prevalence of overweight and obesity in all age groups in recent decades^8,9 has created a pressing need to understand developmental patterns of adiposity and the factors that influence it.^10,11

The prevalence of obesity in children varies by age, race/ethnicity, and family income.¹² Cross-sectional surveillance studies have shown that the prevalence of overweight and obesity increases with age during childhood and adolescence.^12,13 However, longitudinal studies have observed high levels of inter-child variability in age-related change in weight status and adiposity.^14,15 Group-based trajectory modeling has been used to identify population subgroups for adiposity in youth and adults.^16,17 However, few studies have compared groups of children following similar developmental patterns on demographic and behavioral characteristics. Interventions to prevent overweight and obesity would be enhanced by understanding which young people are most likely to gain excessive weight and adiposity and how they differ behaviorally from other children. Specifically, information about how diet, physical activity, and sedentary behaviors differ between groups over time could help plan more effective obesity prevention interventions.¹⁸

Longitudinal studies of weight status have typically examined body mass index (BMI).^16,17,19 While BMI is a practical and useful metric in cross-sectional surveillance studies, it is less useful in longitudinal research aimed at understanding factors that influence change in adiposity.²⁰ BMI is a measure of weight adjusted for height and is not a measure of adiposity.^21,22 Fat mass index (FMI, fat mass/height²) is a better measure of adiposity^21,22 particularly for application in longitudinal studies,²⁰ but it is seldom used to create trajectories, although some studies have examined FMI as an outcome.^23,24 Accordingly, the objectives of this study were to: 1) identify sex-specific groups with similar longitudinal trajectories for FMI in children transitioning from elementary to high school and 2) compare sex-specific longitudinal trajectory groups in terms of demographic and behavioral characteristics.

Methods

Study Design

This study employed a longitudinal design to examine patterns of age-related change in FMI-assessed adiposity in children transitioning from elementary to high school. Analyses were conducted separately for boys and girls since FMI trajectories differ by sex.^25,26

Participants and Setting

Participants were from the Transitions and Activity Changes in Kids (TRACK) study, which followed children from elementary school (2010) to high school (2017).^27–29 Fifth grade participants were enrolled in TRACK from 21 elementary schools in two school districts in South Carolina. Informed consent packets were sent home with the children for parents to read, complete, and return. Parents provided written consent in 5^th grade (for 5^th–7^th grades) and 9^th (for 9^th–11^th grades), and children provided verbal assent. Assessments were completed at 5^th, 6^th, 7^th, 9^th and 11^th grades, with children ranging from 10–16 years of age. The Institutional Review Board at the University of South Carolina approved all study protocols.

Students self-reported age and sex, race (American Indian or Alaskan Native, Asian, Black or African American, Native Hawaiian or Pacific Islander, White, or Other) and ethnicity (Hispanic or Latino). In 5^th grade, 1083 children (502 boys and 581 girls) were recruited into the TRACK study. Approximately 35% of the children were Black, 36% White, 11% Hispanic, and 17% other race/ethnicity. Mean age at 5^th grade was 10.6 ± 0.6 years and mean BMI was 21.2 ± 4.9 kg/m². FMI was available for 921 children in 5^th grade. Students were excluded from the analyses due to not having two or more measures from 6^th, 7^th, 9^th and 11^th grade (n = 140), missing parent education (n = 25), missing race/ethnicity (n = 1), or missing 5^th grade MVPA (n = 62). The analytic sample (n = 709) did not differ from those who were excluded (n = 211) by sex or parent education but had a higher percentage of Black and White participants than those excluded (Supplementary Table).

Measures

Fat mass index

Adiposity was operationally defined as FMI, which provides an accurate assessment of adiposity in children and adolescents.^21,22 Beginning at 5^th grade and for each subsequent measure, total body bioelectrical impedance (RJL Quantum II BIA systems; Clinton Township, MI) was measured according to the manufacturer’s standard protocol on the right side (hand to foot). Fat-free mass (FFM) was calculated using the following equation: FFM = [3.474 + (0.459 × Height²/Resistance) + (0.064 × Weight)]/(0.769 − (0.009 × Age) − 0.016 × Sex), where 1 = boys, 0 = girls.³⁰ Fat mass was determined by subtracting FFM from body weight, and FMI was calculated by dividing fat mass by height (m²). Height and weight were measured by trained staff after participants removed heavy clothing and shoes. Children’s standing and seated heights were measured to the nearest 0.1 cm using a portable stadiometer (Seca, Hamburg, Germany). Weight was measured to the nearest 0.1 kg using an electronic scale (Model 770; Seca, Hamburg, Germany). BMI was calculated using the standard equation (kg/m²).

Physical activity and sedentary behaviors

At each time point, children wore an accelerometer (ActiGraph GT1M and GT3X models, Pensacola, FL) on the right hip for seven consecutive days during waking hours. Data were collected and stored in 60-second epochs. Any period of 60 or more minutes of consecutive zeros was considered non-wear time and recoded to missing. Moderate-to-vigorous physical activity (MVPA) was calculated using an age-specific prediction equation³¹ generalized to the mean age of the TRACK cohort. The threshold for MVPA was 2200 counts/minute, corresponding to 4.0 metabolic equivalents (METs, 1 MET = 3.5 mL O₂ kg⁻¹ min⁻¹). Counts < 100/minute were considered sedentary. Data for Sundays were excluded due to low wear time. Eighty percent of the 5^th grade sample and 87% of the 11^th grade sample met accelerometer data inclusion criteria. Missing values were imputed using a sex-specific multiple imputation method via PROC MI in SAS (Version 9.4; SAS Institute, Inc., Cary, NC, USA) for each year using children with ≥8 hours on ≥2 days. MVPA and sedentary time per day were divided by minutes of wear time to produce minutes per hour to adjust for inter-individual differences in accelerator wear time.

Diet quality

At each time point, dietary intake was assessed by the Block Food Screener for Kids, previously validated with 24-hour recalls in 10–17-year-olds.³² The screener asks whether 41 specific food items were consumed in the last week. If a child responded yes to a food item, he/she was asked on how many days the item was consumed and the usual amount eaten (e.g., a little, some, a lot). For the present study, a measure of diet quality using information from the food screener was created based on a protocol from the Healthy Eating Index (HE-2010).³³ Briefly, specific components were chosen to reflect healthy dietary patterns, including total vegetables, total fruit, whole grains, dairy, and total protein. The overall score was calculated on a per calorie basis and adjusted for energy intake. Values >5000 kcal/day and <50 kcal/day were considered implausible, and those children were excluded from the analysis for that measurement period.³⁴ Higher scores indicated better diet quality, with each component having a maximum score of 10 points, for a total maximum score of 50 points.

Maturity offset

Maturity offset was calculated using sex-specific equations.^35,36 This noninvasive method predicts peak velocity using anthropometric variables. A negative maturity offset represents the number of years the child is from reaching peak height velocity, while a positive maturity offset represents the number of years a child is beyond peak height velocity. Maturity offset at 5^th grade was included as a covariate in analyses.

Statistical Analysis

Descriptive statistics were calculated by sex. Chi-square analyses and t-tests were used to assess differences between boys and girls in demographic variables, 5^th grade maturity offset, MVPA, sedentary behavior, and diet quality. T-tests were also used to determine if differences between boys and girls existed for FMI.

To identify patterns of trajectories of the five FMI measurements by child’s age, group-based trajectory analyses with PROC TRAJ³⁷ in SAS (version 9.4) were conducted by sex, using the CNORM distribution for continuous data. Both linear and quadratic trajectories for 1, 2, 3, and 4 groups were tested. The final number of groups was determined by Bayesian Information Criterion (BIC), the proportion of participants in the group, and the change in BIC between models (estimate of logged Bayes factor-2ΔBIC).³⁷ A 10-fold difference in Bayes factor is considered a meaningful difference. As confirmation for the number of groups chosen, posterior probabilities and odds of correct classification (OCC) were calculated.^38,39 After the number of groups was determined, models were rerun eliminating nonsignificant quadratic terms, and analyses examined differences in demographic variables among the groups. Using the same methods, trajectories were also examined by grades 5–11.

To examine differences between the FMI groups for MVPA, sedentary behaviors, and diet quality, two-way repeated measure ANOVA (FMI group = 3 by grade = 5) was conducted in Proc Mixed by sex. Least-square means for the behavioral outcomes MVPA, sedentary behavior, and diet quality were examined for FMI groups by grade after adjustment for race, parent education, and 5^th grade maturity offset.

Results

Of the 709 participants, 45% were boys (Table 1), ∼36% were Black, 39% White, 9% Hispanic, and 16% other race/ethnicities. Fifty-nine percent of the children had one or more parents with higher than a high school education. At baseline, boys had lower physical maturity, lower sedentary behavior, lower FMI, and higher MVPA compared with 5^th grade girls (p < 0.001). The trajectory of FMI declined from 10 to 16 years of age in boys but increased in girls. Consistent with findings from others,^40–42 we found that weight gain was greatest in the subgroup of children who were heaviest at baseline.

Table 1.

Characteristics of 709 Participants and p-Values (t-Tests or Chi-Square) Determine If There Were Differences Between Boys and Girls

	Boys	Girls
	320	389
Characteristic	Mean (SD) or %	Mean (SD) or %	p-Value
Age, years	10.6 (0.5)	10.5 (0.5)	0.64
Race	n, %	n, %
Black	129, 40.3%	123, 31.6%	0.11
White	119, 37.2%	160, 41.1%
Hispanic	24, 7.5%	38, 9.8%
Other	48, 15.0%	68, 17.5%
Parent education, >high school	194, 60.6%	226, 58.1%	0.50
Maturity offset at 5^th grade	−2.6 (0.6)	−0.8 (0.7)	<0.001
5^th grade MVPA (min/h)	3.7 (2.1)	2.2 (1.1)	<0.001
5^th grade sedentary (min/h)	31.0 (4.4)	32.7 (4.3)	<0.001
5^th grade diet quality	29.8 (5.5)	30.4 (5.5)	0.11

Fat mass index, by age and grade
	n
10 years	333	4.9 (3.2)	6.5 (3.6)	<0.001
11 years	675	5.2 (3.5)	6.9 (3.9)	<0.001
12 years	671	4.9 (3.7)	7.1 (4.2)	<0.001
13 years	316	4.5 (3.8)	6.9 (4.0)	<0.001
14 years	126	4.7 (4.0)	9.4 (5.5)	<0.001
15 years	128	4.4 (4.1)	7.6 (4.1)	<0.001
16 years	177	5.3 (4.8)	9.7 (5.8)	<0.001

	n, % follow-up
5^th grade	709	5.1 (3.4)	6.6 (3.7)	<0.001
6^th grade	701, 98.9%	5.0 (3.6)	6.7 (4.0)	<0.001
7^th grade	700, 98.7%	4.8 (3.8)	7.2 (4.2)	<0.001
9^th grade	274, 38.6%	4.5 (4.0)	8.5 (4.8)	<0.001
11^th grade	240, 33.9%	4.8 (4.7)	9.4 (5.5)	<0.001

Three FMI trajectories were identified separately for boys and girls (Table 2) using BIC, 2Δ BIC, and consideration of group size and were supported by the posterior probability and OCC. In boys (Fig. 1), FMI in Group 1 declined over time, in Group 2 remained constant, and in Group 3, increased from age 10 to age 16. In girls (Fig. 2), FMI in Group 1 remained constant, while FMI increased in Groups 2 and 3. Analyses using grade are similar to age trajectories (See Supplementary Figs.).

Figure 1.

FMI trajectory groups (solid lines with 95% confidence intervals) in boys ages 10–16 years. Group 2, Lower, n = 183, 57% Group 1, Mid, n = 95, 30% Group 3, Upper, n = 42, 13%.

Figure 2.

FMI trajectory groups (solid lines with 95% confidence intervals) in girls ages 10–16 years. Group 1, Lower, n = 230, 59% Group 2, Mid, n = 115, 30% Group 3, Upper, n = 44, 11%.

Table 2.

Determining Number of Groups for Boys and Girls and Assignment Accuracy (Posterior Probability and OCC)

Boys
1 group	2 groups		1 versus 2*	3 groups		2 versus 3*	4 groups		3 versus 4*
BIC	BIC	% per group	2Δ bIC	BIC	% per group	2Δ BIC	BIC	% per group	2Δ BIC
−2923.9	−2586.8	74.9%, 25.1%	674.2	−2445.8	29.9%, 57.3%,12.8%	282.0	−2375.1	15.6%, 53.8%,26.2%, 4.6%	141.4

	3-Group model
	Group 1Midn, %95, 30%	Group 2Lowern, %183, 57%	Group 3Uppern, %42, 13%
Posterior probability	0.94	0.98	0.94
OCC	36.7	36.5	106.7

Girls
1 group	2 groups		1 versus 2*	3 groups		2 versus 3*	4 groups		3 versus 4
BIC	BIC	% per group	2Δ BIC	BIC	% per group	2Δ	BIC	% per group	2Δ
−3896.5	−3416.2	76.2%, 23.8%		−3210.7	58.7%, 30.0%,11.3%		−3117.4	31.4%, 45.0%,16.6%, 7.0%	93.3

	3-Group model
	Group 1Lowern, %230, 59%	Group 2Midn, %115, 30%	Group 3Uppern, %44, 11%
Posterior probability	0.97	0.95	0.97
OCC	22.7	44.3	253.8

Posterior probability values > 0.70 indicate that the trajectory includes subjects with similar patterns of change, and an OCC of 5 or more is generally recommended for all groups.^38,39

* Comparison between groups, interpretation of 2Δ BIC = estimate if 2log_e (β₁₀).³⁷ Evidence against simple model; 0–2 = Not worth mentioning; 2–6 = Positive; 6–10 = Strong; >10 = Very strong.

BIC, Bayesian Information Criterion; smaller is better. OCC, odds of correct classification.

Characteristics of the 3 groups by sex are presented in Table 3. For boys, the lower trajectory (Group 2) was the largest (n = 183) and was the least physically mature. The mid trajectory (Group 1) was midway between Groups 2 and 3. The upper trajectory (Group 3) was the smallest group (n = 42) . In girls, the lower trajectory (Group 1) was the largest (n = 230), had the largest percentage of parents with greater than a high school education, and was the least physically mature at 5^th grade. The mid trajectory (Group 2), had the highest proportion of White children. The upper trajectory (Group 3) was the smallest n = 44), had the lowest percentage of parents with greater than a college education, and was predominantly Black.

Table 3.

Characteristics of the FMI Trajectory Groups by Sex

	Boys				Girls
	Group 1mid,declinedover time	Group 2lower, consistentover time	Group 3upper, increasedover time	p-Value	Group 1lower,consistent over time	Group 2mid,increased over time	Group 3upper,increased over time	p-Value
n, %=	95, 30%	183, 57%	42, 13%		230, 59%	115, 30%	44, 11%
Race				0.16				<0.001
Black	42.1%	39.3%	40.5%		26.1%	28.7%	68.2%
White	30.5%	39.9%	40.5%		45.2%	45.2%	9.1%
Hispanic	13.7%	5.5%	2.4%		11.7%	6.1%	9.1%
Other	13.7%	15.3%	16.7%		17.0%	20.0%	13.6%
Parent education, > high school	63.2%	58.5%	64.3%	0.65	65.7%	52.2%	34.1%	<0.001
Age at 5^th grade, mean (SD)	10.5 (0.5)	10.6 (0.5)	10.6 (0.5)	0.48	10.6 (0.5)	10.5 (0.5)	10.5 (0.5)	0.46
Maturity offset at 5^th grade, mean (SD)	−2.4 (0.5)	−2.8 (0.5)	−2.0 (0.5)	<0.001	−1.1 (0.6)	−0.6 (0.6)	−0.3 (0.6)	<0.001
Fat mass index, mean (SD)	6.9 (1.8)	2.8 (1.4)	11.2 (2.4)	<0.001	4.3 (1.7)	8.6 (1.9)	13.8 (2.9)	<0.001
% Over Weight at 5^th grade	89.5%	12.0%	100%	<0.001	17.0%	91.3%	100%	<0.001

p-Value from analysis of variance for continuous variables, and Chi-square for categorical variables.

The results of the two-way (FMI Group by Grade) analyses for diet quality, sedentary behavior, and MVPA are presented in Table 4 by sex. In boys, sedentary behavior increased and diet quality declined from 5^th to 11^th grade (p < 0.001). There were group differences for MVPA (p < 0.01), with the lower trajectory group (Group 2), having the highest MVPA compared with the other groups, but this difference disappeared by 9^th grade. In girls, MVPA declined and sedentary behavior increased from grade 5–11 (p < 0.001). There were group differences (p = 0.04) for diet quality with declines in Groups 1 and 3, but Group 2 (mid trajectory) remained stable over time.

Table 4.

Least Square Means from Two-Way ANOVAs (FMI Group by Grade) for MVPA, Sedentary Behavior, and Diet Quality Adjusted for Race, Parent Education, and 5^th Grade Maturity

		LS means (SE)
		Group 1	Group 2	Group 3
		Mid	Lower	Upper	Groupp-Value	Gradep-Value
	Grade	n = 95	n = 183	n = 42	Groupp-Value	Gradep-Value
MVPA, min/h	5	3.4 (0.2)	4.1 (0.2)	2.3 (0.4)	0.01	<0.001
	6	3.1 (0.2)	3.7 (0.2)	2.4 (0.4)
	7	3.0 (0.3)	3.6 (0.2)	2.2 (0.4)
	9	2.1 (0.4)	2.3 (0.3)	2.0 (0.7)
	11	2.7 (0.5)	3.8 (0.3)	4.9 (0.7)
Sedentary behavior,	5	31.1 (0.4)	30.5 (0.4)	33.8 (0.7)	0.30	<0.001
Min/h	6	34.4 (0.5)	34.1 (0.4)	35.7 (0.7)
	7	36.1 (0.5)	35.6 (0.4)	37.5 (0.7)
	9	39.6 (0.8)	40.2 (0.6)	39.0 (1.2)
	11	40.0 (0.9)	41.4 (0.6)	40.2 (1.2)
Diet quality	5	29.0 (0.6)	30.2 (0.5)	30.9 (0.9)	0.37	<0.001
	6	28.2 (0.6)	28.9 (0.5)	29.5 (1.0)
	7	28.2 (0.6)	29.1 (0.5)	29.4 (0.9)
	9	27.3 (1.0)	28.5 (0.8)	28.7 (1.7)
	11	27.0 (1.2)	26.7 (0.8)	26.5 (1.6)

		Lower	Mid	Upper
		n = 230	n = 115	n = 44
MVPA, min/h	5	2.3 (0.1)	2.1 (0.1)	1.7 (0.2)	0.09	<0.001
	6	1.9 (0.1)	1.8 (0.1)	1.4 (0.2)
	7	1.8 (0.1)	1.7 (0.1)	1.7 (0.1)
	9	0.9 (0.1)	1.0 (0.2)	0.8 (0.3)
	11	1.6 (0.1)	1.1 (0.2)	1.5 (0.2)
Sedentary behavior, min/h	5	32.5 (0.3)	32.8 (0.4)	34.0 (0.7)	0.76	<0.001
	6	36.9 (0.3)	36.6 (0.5)	38.4 (0.8)
	7	38.6 (0.3)	39.0 (0.5)	39.9 (0.8)
	9	42.9 (0.5)	42.1 (0.7)	41.3 (1.1)
	11	43.5 (0.5)	43.2 (0.7)	42.5 (0.9)
Diet quality	5	30.4 (0.4)	30.7 (0.5)	30.5 (0.9)	0.04	0.01
	6	29.5 (0.4)	30.0 (0.6)	30.8 (0.9)
	7	29.0 (0.4)	30.7 (0.6)	28.9 (0.9)
	9	29.3 (0.6)	29.5 (0.9)	27.6 (1.2)
	11	27.1 (0.7)	30.6 (1.0)	29.2 (1.3)

Discussion

This study identified three FMI trajectory groups in both boys and girls, although the shapes of the FMI trajectories differed by sex. In girls, the upper and mid trajectories increased over time, while in boys, the upper trajectory declined and the upper trajectory increased over time. In both girls and boys, one trajectory group maintained a low FMI. These findings identify sub-groups of youth that tend to gain FMI over time and are at risk of overweight or obesity, which could lead to targeting obesity prevention programs more effectively.

The FMI trajectory groups did not differ by race or parent education in boys. Kim et al.,⁴³ using BMI group trajectories, found race and parent education differences among boys < 11 years of age. Another study⁴⁴ that used growth curve analysis, reported that non-Black boys had higher FMI at age 8 than Black boys, but by age 14, there were no differences. The present study found differences by race and by parent education between the trajectory groups in girls. Most girls in the lower FMI trajectory group were White and had parents with greater than a high school education, whereas most girls in the high trajectory group were Black and had a low percentage of parents with greater than a high school education. Similarly, Kim et al.⁴³ reported that the upper trajectory of girls was composed of Black girls and the lower trajectory of White girls.

We observed differences across FMI trajectory groups in behaviors known to influence adiposity. Both boys and girls in the lower FMI trajectory group had the highest MVPA compared with the other groups in 5^th grade. Sedentary behaviors increased over time for both boys and girls. Others have also found negative relationships between adiposity and MVPA^45,46 and positive relationships between sedentary behavior and adiposity.^19,47,48 Very little has been reported using group trajectories and behavioral factors, such as accelerometer-assessed MVPA and sedentary behavior. Lin et al.⁴⁰ reported that boys in the persistently obese BMI trajectory had low levels of afterschool exercise, and girls in the persistently obese BMI trajectory frequently watched television or used computers. Kwon et al.,²³ using percent body fat group trajectories, reported that the consistently inactive group was more likely to be in the trajectory group that was becoming obese, compared with the other active group. In boys, diet quality declined over time, but with less decline in the FMI trajectory group that declined from 5^th to 11^th grade. Girls in the mid trajectory maintained diet quality from 5^th to 11^th grade, while diet quality declined for girls in the lower and upper trajectories. Other studies have found that maintaining a healthy body weight was related to a healthy diet.^48–50 Kwon et al.⁵¹ reported no difference for energy intake among group-based percent body fat trajectory groups. Similarly, growth curve analyses revealed no relationship between FMI and energy intake. The present study suggests that interventions to promote physical activity after 5^th grade may help to prevent further increases in FMI.

This study was unique in forming FMI-based trajectory groups, which is a more accurate assessment of adiposity in children and adolescents than BMI, and comparing these groups on MVPA, sedentary behavior, and diet quality, rather than on early life factors or health outcomes. Furthermore, trajectory groups were formed for boys and girls separately in a racially diverse cohort that was followed from 5^th to 11^th grade. Knowing how MVPA, sedentary behavior, and diet quality differ over time may help to determine when to promote healthy eating and physical activity in divergent weight groups in boys and girls.

Using BMI with group trajectory analysis usually manifests with one stable group,¹⁷ with other trajectory groups increasing over time in both boys and girls,^40,42 but in the current study, the shapes of the trajectories differed between boys and girls. Marshall et al.⁴² compared trajectories using BMI, FMI, and percent body fat in boys and girls using cluster analysis. They reported three trajectories that increased in both boys and girls but, as in the present study, boys and girls differed in trajectories based on percent body fat and FMI. Differences between FMI and BMI trajectories have also been found using growth curve analyses. Both FMI- and percent body fat-based trajectories decreased among boys over time, but increased among girls, while BMI increased in both boys and girls.²⁶ Fulton et al.²⁵ reported declines in FMI and increases in BMI using growth curve analysis when samples contained both boys and girls.

Strengths of this study include assessment at five time-points from 5th to 11^th grade, spanning the transition from elementary to middle school and high school. An accurate assessment of adiposity, FMI, was used to establish the trajectories, and students were diverse in race/ethnicity, had parents with different educational backgrounds, and included both boys and girls. Physical activity and sedentary behavior were measured using accelerometry, and diet quality was derived from an instrument with established measurement properties.³² Limitations include that the children were from only two school districts in one state, lower sample size in 9^th and 11^th grades, which limits generalizability, and that many of children were already overweight or obese (48%) when they entered the study. The diet screener, while validated, provides limited insight into nutrient intake and may limit interpretation. Also, other factors such as school environment or public health infrastructure may influence adiposity.

Conclusions

This study applied a trajectory analysis to longitudinal adiposity data in children who were transitioning from 5^th to 11^th grade. The primary purpose was to identify, separately for boys and girls, subgroups of the samples that demonstrated similar patterns for longitudinal change in adiposity as assessed by FMI. These findings support tailored interventions for groups of youth considering sex, race, socioeconomic status, and behavioral status well before the 5^th grade. Health care providers could provide physical activity and diet advice to parents of children in need of attention (i.e., those with higher weight status; high sedentary behavior, low physical activity, and/or poor diet; and lower socioeconomic status). To support the health care providers/family interventions, schools should improve opportunities for physical activity and diet quality. Community programs for youth should involve families to improve diet and physical activity and to limit sedentary behaviors. These measures should continue throughout the school years to prevent declines in physical activity and diet quality.

Impact Statement

Group-based trajectory modeling of fat mass index (FMI), which accurately assesses adiposity, established three subgroups of boys and three of girls ages 10–16. Groups of girls differed by race, parent education, and diet quality. Groups of boys differed by moderate-to-vigorous physical activity. These results suggest different prevention strategies in FMI subgroups.

Authors’ Contributions

M.D. and R.R.P. were responsible for designing the study. M.D. managed the data and performed analysis and original article drafting. R.P.S. and R.R.P. were involved in writing, review, and editing.

Footnotes

Acknowledgments

The authors wish to thank all the parents and children who participated, the data collection staff, and Gaye Groover Christmas, MPH, who provided editorial assistance in the preparation of the article.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was funded by the National Heart, Lung, and Blook Institute (R01HL091002-1 A1: PI: R.R.P.).

Supplemental Material

References

Field

, Coakley

, Must

, et al. Impact of overweight on the risk of developing common chronic diseases during a 10-year period. Arch Intern Med, 2001; 161(13):1581–1586.

Lingvay

, Cohen

, Le Roux

, et al. Obesity in adults. Lancet, 2024; 404(10456):972–987.

Güngör

. Overweight and obesity in children and adolescents. J Clin Res Pediatr Endocrinol, 2014; 6(3):129–143.

Dietz

. Health consequences of obesity in youth: Childhood predictors of adult disease. Pediatrics, 1998; 101(3 Pt 2):518–525.

Faienza

, Wang

, Frühbeck

, et al. The dangerous link between childhood and adulthood predictors of obesity and metabolic syndrome. Intern Emerg Med, 2016; 11(2):175–182.

Singh

, Mulder

, Twisk

JWR

, et al. Tracking of childhood overweight into adulthood: A systematic review of the literature. Obes Rev, 2008; 9(5):474–488.

Sun

, Xi

, Yang

, et al. Weight change from childhood to adulthood and cardiovascular risk factors and outcomes in adulthood: A systematic review of the literature. Obes Rev, 2021; 22(3):e13138.

Fryar

, Carroll

, Afful

. Prevalence of overweight, obesity, and severe obesity among children and adolescents aged 2-19 years: United States 1963–1965 through 2017–2018. NCHS Health E-Stats, 2020.

Wang

, Beydoun

, Min

, et al. Has the prevalence of overweight, obesity and central obesity levelled off in the United States? Trends, patterns, disparities, and future projections for the obesity epidemic. Int J Epidemiol, 2020; 49(3):810–823.

10.

O’Connor

, Williams

, Blair

, et al. Predictors of developmental patterns of obesity in young children. Front Pediatr, 2020; 8:109.

11.

Ross

, Flynn

, Pate

. What is really causing the obesity epidemic? A review of reviews in children and adults. J Sports Sci, 2016; 34(12):1148–1153.

12.

Stierman

, Ogden

, Yanovski

, et al. Changes in adiposity among children and adolescents in the United States, 1999–2006 to 2011–2018. Am J Clin Nutr, 2021; 114(4):1495–1504.

13.

Iyun

, Okobi

, Nwachukwu

, et al. Analyzing obesity trends in American children and adolescents: Comprehensive examination using the National Center for Health Statistics (NCHS) database. Cureus, 2024; 16(6):e61825.

14.

Williamson

, Han

, Johnson

, et al. Longitudinal study of body weight changes in children: Who is gaining and who is losing weight. Obesity (Silver Spring), 2011; 19(3):667–670.

15.

Bichteler

, Gershoff

. Identification of children’s BMI trajectories and prediction from weight gain in infancy. Obesity (Silver Spring), 2018; 26(6):1050–1056.

16.

De Rubeis

, Andreacchi

, Sharpe

, et al. Group‐based trajectory modeling of body mass index and body size over the life course: A scoping review. Obesity Science & Practice, 2021; 7(1):100–128.

17.

Mattsson

, Maher

, Boland

, et al. Group‐based trajectory modelling for BMI trajectories in childhood: A systematic review. Obes Rev, 2019; 20(7):998–1015.

18.

Chen

T-A

, Baranowski

, Moreno

, et al. Obesity status trajectory groups among elementary school children. BMC Public Health, 2016; 16:526.

19.

Must

, Tybor

. Physical activity and sedentary behavior: A review of longitudinal studies of weight and adiposity in youth. Int J Obes (Lond), 2005; 29 (,Suppl 2):S84–S96.

20.

Demerath

, Schubert

, Maynard

, et al. Do changes in body mass index percentile reflect changes in body composition in children? Data from the Fels Longitudinal Study. Pediatrics, 2006; 117(3):e487–e495.

21.

Freedman

, Zemel

, Weber

, et al. The cross-sectional relation of body mass index to adiposity among 8-to 19-year-olds within several races and ethnicities. J Pediatr, 2025; 277:114375.

22.

Freedman

, Wang

, Maynard

, et al. Relation of BMI to fat and fat-free mass among children and adolescents. Int J Obes (Lond), 2005; 29(1):1–8.

23.

Kwon

, Janz

, Letuchy

, et al. Active lifestyle in childhood and adolescence prevents obesity development in young adulthood. Obesity (Silver Spring), 2015; 23(12):2462–2469.

24.

Rzehak

, Oddy

, Mearin

, et al.; WP10 working group of the Early Nutrition Project. Infant feeding and growth trajectory patterns in childhood and body composition in young adulthood. Am J Clin Nutr, 2017; 106(2):568–580.

25.

Fulton

, Dai

, Steffen

, et al. Physical activity, energy intake, sedentary behavior, and adiposity in youth. Am J Prev Med, 2009; 37(1 Suppl):S40–S49.

26.

Dowda

, Taverno Ross

, McIver

, et al. Physical activity and changes in adiposity in the transition from elementary to middle school. Child Obes, 2017; 13(1):53–62.

27.

Pate

, Schenkelberg

, Dowda

, et al. Group-based physical activity trajectories in children transitioning from elementary to high school. BMC Public Health, 2019; 19(1):323.

28.

Pate

, Saunders

, Taverno Ross

, et al. Patterns of age-related change in physical activity during the transition from elementary to high school. Prev Med Rep, 2022; 26:101712.

29.

Pate

, Dowda

, Dishman

, et al. Determinants of change in physical activity in children during the transition from elementary to high school. Med Sci Sports Exerc, 2024; 56(7):1275–1284.

30.

Horlick

, Arpadi

, Bethel

, et al. Bioelectrical impedance analysis models for prediction of total body water and fat-free mass in healthy and HIV-infected children and adolescents. Am J Clin Nutr, 2002; 76(5):991–999.

31.

Freedson

, Pober

, Janz

. Calibration of accelerometer output for children. Med Sci Sports Exerc, 2005; 37(11 Suppl):S523–S530.

32.

Hunsberger

, O’Malley

, Block

, et al. Relative validation of Block Kids Food Screener for dietary assessment in children and adolescents. Matern Child Nutr, 2015; 11(2):260–270.

33.

Guenther

, Casavale

, Reedy

, et al. Update of the healthy eating index: HEI-2010. J Acad Nutr Diet, 2013; 113(4):569–580.

34.

Willett

. Nutritional Epidemilogy. 2nd ed. New York: Oxford University Press; 1998.

35.

Mirwald

, Baxter-Jones

, Bailey

, et al. An assessment of maturity from anthropometric measurements. Med Sci Sports Exerc, 2002; 34(4):689–694.

36.

Malina

, Kozieł

. Validation of maturity offset in a longitudinal sample of Polish boys. J Sports Sci, 2014; 32(5):424–437.

37.

Jones

, Nagin

, Roeder

. A SAS procedure based on mixture models for estimating developmental trajectories. Sociol Method Res, 2001; 29(3):374–393.

38.

Niyonkuru

, Wagner

, Ozawa

, et al. Group-based trajectory analysis applications for prognostic biomarker model development in severe TBI: A practical example. J Neurotrauma, 2013; 30(11):938–945.

39.

Nagin

. Group-based Modeling of Development. Cambridge, MA: Harvard University Press; 2005.

40.

Lin

L-J

, Chang

H-Y

, Luh

D-L

, et al. The trajectory and the related physical and social determinants of body mass index in elementary school children: Results from the child and adolescent behaviors in long‐term evolution study. J Obes, 2014; 2014(1):728762.

41.

Ziyab

, Karmaus

, Kurukulaaratchy

, et al. Developmental trajectories of Body Mass Index from infancy to 18 years of age: Prenatal determinants and health consequences. J Epidemiol Community Health, 2014; 68(10):934–941.

42.

Marshall

, Curtis

, Cavanaugh

, et al. Identification of and associations among low, middle, and high body composition trajectories from age 5-to 17-years. Children (Basel), 2020; 7(10):192.

43.

Kim

, Rajbhandari

, Krile

, et al. Body mass index trajectories among the healthy communities study children: Racial/ethnic and socioeconomic disparities in Childhood Obesity. J Racial Ethn Health Disparities, 2024; 11(1):203–215.

44.

Eissa

, Dai

, Mihalopoulos

, et al. Trajectories of fat mass index, fat free–mass index, and waist circumference in children: Project HeartBeat!. Am J Prev Med, 2009; 37(1 Suppl):S34–S39.

45.

Fisher

, Hill

, Webber

, et al. MVPA is associated with lower weight gain in 8-10 year old children: A prospective study with 1 year follow-up. PLoS One, 2011; 6(4):e18576.

46.

Kwon

, Janz

, Burns

, et al. Effects of adiposity on physical activity in childhood: Iowa bone development study. Med Sci Sports Exerc, 2010.

47.

Tanaka

, Reilly

, Huang

. Longitudinal changes in objectively measured sedentary behaviour and their relationship with adiposity in children and adolescents: Systematic review and evidence appraisal. Obes Rev, 2014; 15(10):791–803.

48.

Dowda

, Saunders

, Dishman

, et al. Association of physical activity, sedentary behavior, diet quality with adiposity: A longitudinal analysis in children categorized by baseline weight status. Int J Obes (Lond), 2024; 48(2):240–246.

49.

Watts

, Loth

, Peterson

, et al. Characteristics of a favorable weight status change from adolescence to young adulthood. J Adolesc Health, 2016; 58(4):403–409.

50.

Jennings

, Welch

, van Sluijs

, et al. Diet quality is independently associated with weight status in children aged 9-10 years. J Nutr, 2011; 141(3):453–459.

51.

Kwon

, Burns

, Levy

, et al. 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):1169–1174.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.08 MB

0.02 MB