Adolescents’ diet quality, cardiometabolic risk, and adiposity: A prospective cohort research brief

Abstract

Objectives.

Examine the prospective association among diet with adolescents’ cardiometabolic risk (CMR) and anthropometrics.

Methods.

Secondary analysis of an observational study of adolescents ages 10–16. Twenty-four-hour food recall was used to calculate Healthy Eating Index-2015 (HEI-2015) scores. Anthropometrics were assessed using magnetic resonance imaging, dual energy x-ray absorptiometry, and height/weight measurements. CMR included mean arterial pressure, homeostatic model assessment for insulin resistance (HOMA-IR), high-density lipoprotein cholesterol (HDL-c), and triglycerides. Associations between baseline-HEI-2015 score with follow-up adiposity and CMR were examined using regression models.

Results.

A total of 192 adolescents were included. Baseline HEI-2015 scores were inversely associated with follow-up total CMR-z (p = 0.01), HOMA-IR (p<0.01), waist circumference-z (p=0.02), body mass index percentile (p=0.01), fat mass (p=0.04), lean mass (p=0.02), and visceral adipose tissue mass (p=0.01).

Conclusions and Implications.

Adolescents with lower adherence to dietary guidelines and who had greater CMR and anthropometry measurements at baseline continued this trajectory across the observation.

INTRODUCTION

Obesity in adolescence is associated with cardiometabolic risk (CMR) during adolescence and later hypertension, hyperglycemia, and dyslipidemia into adulthood.¹ A key association with obesity and CMR is dietary intake. Diets that are high in energy, fat, and refined carbohydrates and that are low in fruits, vegetables, and fiber are cross-sectionally associated with higher CMR factors and higher adiposity in adolescence.² Unfortunately, these poor dietary patterns are relatively reflective of the current U.S. patterns among youth.³ The diet quality of adolescents is among the worst across all age groups in the U.S.,⁴ preventing risk of chronic disease through improving diet patterns is an area of great need. ^5,6

The Healthy Eating Index-2015 (HEI-2015) is a diet quality index used to measure alignment of diet with the 2015–2020 Dietary Guidelines for Americans.^7–9 These guidelines are the national recommendations co-authored by the U.S. Department of Health and Human Services and the U.S. Department of Agriculture intended for use and dissemination by health professionals and policy makers for all individuals ages 2 years and older. Among adults in the U.S., research supports a link between diet quality as measured by HEI and CMR.^10–12 Several studies have examined the associations between diet quality and the risk of obesity, adiposity, and CMR in adolescents, but results thus far have been inconsistent and largely cross-sectional. Inverse associations between HEI-2010 and change in BMI in girls have been supported, but not among boys.¹³ Total HEI-2015 and Alternative Healthy Eating Index-2010 (AHEI-2010) scores were cross-sectionally inversely associated with the metabolic syndrome risk among a large sample of adolescents ¹⁴ and African American boys,¹⁵ respectively. Overall, there is some research to support a connection between diet quality and adolescent weight, adiposity, and CMR factors. Current findings are inconsistent and rely on various measures of diet quality other than the current 2015–2020 Dietary Guidelines for Americans. Part of this inconsistency may be a reliance on measures of BMI instead of the more precise imaging estimates of body fat distribution and cardiometabolic risk factors, which are more precise ways to monitor adolescents’ health still with clinical utility.

There is a dearth of prospective data among U.S. adolescents that examines the relationship among compliance with national dietary guidelines and CMR factors and body composition. A prospective examination of diet quality may help identify targets for future interventions in families, homes, and communities. The objective of the present study was to examine the association between diet quality at baseline with CMR factors and body composition 2-years later among a sample of adolescents.

METHODS

Participants

The current study is a secondary analysis of the Translational Investigation of Growth and Everyday Routine in Kids (TIGER Kids) cohort, a prospective observational study of adolescents, 10–16 years of age recruited between 2016 and 2018 to examine combined associations between meeting physical activity, sleep, and overall dietary guidelines with CMR factors and adiposity (NCT02784509).¹⁶ Recruitment took place using convenience sampling in a metropolitan of Louisiana, a medically underserved area characterized by high levels of poverty, food insecurity, obesity, and related diseases.¹⁷ Recruitment efforts for parents included email listserv, community events, social media, and health fairs. Participants provided follow-up measures 2 years later (18 months - 30 months), between 2018 and 2020 and were offered a total compensation of $100 for the completion of this study. Inclusion criteria for the study were having a body weight < 500 lbs and having the ability to understand instructions and complete all study procedures. Exclusion criteria were adolescent pregnancy, restrictive diet due to illness, or significant physical or mental disability that would impede walking, wearing an accelerometer or global positioning system (GPS) monitoring, or responding to ecological momentary assessment. The study protocol and all procedures were approved through a full board review by the Pennington Biomedical Research Center Institutional Review Board (IRB). At baseline, parents provided written informed consent, and adolescents provided written informed assent.

Of the 342 eligible and enrolled adolescents, the final sample included 192 adolescents with complete baseline and follow-up data. There were no significant differences between the included and excluded adolescents in the present study for age, sex, race, household income, puberty, in-school status,, mean values of adiposity indicators, and CMR factor z-scores except for Homeostatic Model Assessment of Insulin Resistance (HOMA-IR; 0.15 vs. −0.11, p = 0.021).

Procedures

Parents and adolescents attended an in-person orientation with study staff to orient to the study procedures and ask questions and to learn how to accurately wear accelerometers and complete assessments. At the baseline clinic visit, parents completed a demographic survey to provide reports of adolescent age, sex (male, female), race (American Indian/Alaska Native, Asian, Black/African American, Native Hawaiian or Other Pacific Islander, White) and ethnicity (Hispanic or Latino/a, non-Hispanic or Latino/a), household income (4 options to select income range), and if the adolescent was in school term or on holiday. At both baseline and follow-up, adolescents were asked to wear an accelerometer for at least 7 days and to complete 2 24-hour dietary recalls for their food and drink intake prior to the appointment plus a 24-hour dietary recall during the appointment (for a total of at least 2–3 dietary recalls at both baseline and 2-year follow-up). Anthropometrics, body composition, blood pressure, and clinical chemistry measurements were collected on the same day, as described below.

Measures

Diet quality.

Twenty-four hour food recall was assessed using the web-based Automated Self-Administered 24-hour Dietary Assessment Tool (ASA-24).¹⁸ The ASA-24 has been validated in a sample of adolescents¹⁹ and uses multiple prompts to elicit recall of food and beverages consumed the prior day.^20,21 Prior the baseline visit, adolescents received instructions on how to complete the dietary recalls. Adolescents were sent at preselected intervals that were unannounced 1 weekday and 1 weekend 24-hour recall before their baseline visit, which then included a third dietary recall on a weekday. If 1 or no recalls were available after the baseline visit, adolescents were contacted within 30 days to complete additional recalls to capture average intake of at least 2 recalls, as done similarly by the National Health and Nutrition Examination Survey⁴ due to participant burden and the fact that the current study is examining the population and not any individual participant. These surveys were sent to the parent’s email and the parent was asked to have the adolescent respond and could provide the adolescent with assistance as needed. Diet quality was calculated with the Healthy Eating Index-2015 (HEI-2015) score, which was derived from the ASA24 results and a SAS macro. The HEI-2015 was chosen for its alignment with the U.S. dietary guidelines. The simple HEI scoring algorithm method was utilized to create this score, all available recalls were averaged for each participant.²² The total score of the HEI-2015 ranges from 0–100, with scores closer to 100 representing a higher quality of diet in alignment with the 2015–2020 Dietary Guidelines for Americans.^8,9 The HEI-2015 total score is derived from scoring 13 dietary components, and maximum scores of 5 or 10 are possible for each component. HEI-2015 component scores of adequacy include total fruits, whole fruits, total vegetables, greens and beans, whole grains, dairy, total protein foods, seafood and plant proteins, and fatty acids. HEI-2015 component scores of moderation are reverse scored and include refined grains, sodium, added sugars, and saturated fats.

Weight was measured to the nearest 0.1 kg by trained research staff using a Michelli GSE 460 scale (G.T. Michelli Co., Baton Rouge, LA). Research staff measured participants wearing a gown and no shoes, then subtracted gown weight to calculate final weight. Two measurements were averaged, or the closest 2 of 3 when measurements differed by more than 0.5 kg.

Standing height was measured to the nearest 0.1 cm using a Harpenden stadiometer (Holtain Limited, Crymych, UK) by trained research staff. Two measurements were averaged, or the closest 2 of 3 when measurements differed by more than 0.5 cm.

BMI %ile and BMIz were calculated using participant’s age, sex, height, and weight using the CDC 2000 growth chart.²³ Participants were grouped into the following BMI categories: healthy weight or underweight (< 85th %ile), overweight (> 85th %ile to < 95th %ile), and obesity (> 95th %ile).

Waist circumference (WC) was measured at the natural waist, between the inferior border of the rib cage and superior aspect of the iliac crest, with clothing moved out of the way to the nearest 0.1 cm.²⁴

Total fat mass and lean mass were measured by whole-body dual energy x-ray absorptiometry (DXA) using a GE Lunar iDXA scanner (GE Medical Systems, Milwaukee, WI) and standard imaging and positioning protocol.²⁴ Body fat percentage was calculated as total fat mass divided by total mass (fat mass and lean mass).

Visceral adipose tissue from the highest point of the liver to the bottom of the right kidney was measured by water-fat shifting MRI using the General Electric Discovery 750w 3.0 Tesla (GE Medical Systems, Milwaukee, WI). IDEAL-IQ imaging was used to capture images during a single acquisition with a 20-second breath hold. A trained technician drew the visceral depot at each fifth slice, starting at 2 slices under the L4/L5 to the diaphragm. A validated algorithm was then used to calculate visceral adipose tissue and estimates of total and subcutaneous fat at the abdomen. These procedures are described in greater detail elsewhere.²⁵

To measure Cardiometabolic Risk, a fasting sample of blood was collected by a trained phlebotomist. High density lipoprotein cholesterol (HDL-c) and triglycerides were obtained from a Trinity DXC600 manufactured by Beckman Coulter. Insulin and glucose were assayed on the Siemens Immulite 2000 and used to calculate Homeostatic Model Assessment of Insulin Resistance (HOMA-IR) using the formula HOMA-IR = (insulin * glucose) / 22.5.²⁶ High HOMA-IR increases insulin resistance, and vice versa for low HOMA-IR. Resting blood pressure was assessed using standard clinical procedures on a sphygmomanometer. Mean arterial pressure (MAP) was calculated as MAP = (SBP + 2 [DBP]) / 3.²⁷

Physical activity was measured for 7 continuous days (24 h/day) using Actigraph GT3X+ accelerometers (Pensacola, FL). Moderate-to-vigorous physical activity (MVPA) was defined as meeting more than 574 counts out of 15 second epochs of accelerometry data.²⁸ This was used as a covariate.

Adolescents responded to questions about pubertal development based on a series of standardized, validated drawings.²⁹ Adolescents self-reported using a scale of 1 (no development) to 5 (complete development) for female breasts or male genitalia and for pubic hair. Pubertal stage was included as a covariate.

Statistical Analyses

A standardized cardiometabolic risk z-score was calculated based on the sample and in accordance with accepted standards for pediatric practices based on their reflection of adult metabolic syndrome criteria:²⁷ BMI, MAP, fasting blood glucose, and triglycerides were regressed based on age, sex, and race, and then the standardized residuals were summed (HDL-C). HDL-C is inversely related to cardiometabolic risk thus, it is multiplied by −1. Paired t-tests were used to compare values from baseline to follow-up. Linear regression was used to assess associations of diet quality using baseline total HEI-2015 with each CMR and body composition outcome separately. For the general linear models, all analyses were first conducted after adjusting for baseline age and sex, race, household income, in-school status, puberty (also during follow-up), and moderate-to-vigorous physical activity, and then further baseline corresponding CMR and body composition. For the first series of models, differences in CMR factors and body composition based on different levels of total HEI-2015 (tertile 1, n = 64, range 23.8–41.7; tertile 2, n = 64, range 41.9–52.3; and tertile 3, n = 64, range 52.8–86.4 calculated from the sample) were tested using a general linear model. Additionally, associations between changes in total HEI-2015 from baseline to follow-up and CMR and body composition was assessed using the general linear model. Sensitivity analyses were used to examine associations for specific HEI-2015 component scores and associations stratified by pubertal development groups at baseline and follow-up. All statistical analyses were performed using IBM SPSS Statistics for Windows version 24.0 (IBM Corp., Armonk, NY, USA).³⁰

RESULTS

A total of 192 participants with complete data collection from baseline to follow-up were included in the analyses from an initial sample of 342 participants at baseline. Those excluded included those missing data at baseline (7 missing DXA or MRI, 2 missing anthropometry, 21 did not complete dietary recalls, 3 did not complete blood draw) or those who did not return to follow-up (n=84) or did not complete follow-up measures (8 missing DXA/MRI, 1 missing anthropometry, 18 missing dietary recalls, 6 did not complete blood draw; Figure 1). The mean time between baseline and follow-up measurements was 1.96 ± 0.22 years.

Figure 1.

CONSORT Flow Diagram of Participants Included in Final Analyses

Table 1 describes the baseline and follow-up characteristics of the sample. At baseline, adolescents were 12.9 ± 1.9 years of age, 47.9% were male, 57.8% were White, and 33.3% were African American. Across the sample, 50.6% of adolescents were categorized as having healthy weight or underweight (n=4) with a BMI %ile <85th. There were no significant differences in the total HEI-2015 or any CMR factor between the baseline and the follow-up values. Body composition measurements, including BMI %ile, fat mass, lean mass, and VAT mass, increased from baseline to follow-up, whereas body fat percentage decreased from baseline to follow-up (Table 1).

Table 1.

Demographic Characteristics, Total 2015 Healthy Eating Index (HEI-2015) Score, Body Composition and Cardiometabolic Risk Factors Among 192 Adolescents

Characteristic	Baseline	Follow-up	P value a
	(n = 192)	(n = 192)
Demographic	Mean (SD)	Mean (SD)
Age (year)	12.9 (1.88)	14.9 (1.91)
	n (%)	n (%)
Male	92 (47.9)	-
Race
White	111 (57.8)	-
African American	64 (33.3)	-
Other b	17 (8.9)	-
Annual household income
< $29,999	19 (9.9)	-
$30,000–69,999	45 (23.4)	-
$70,000–139,999	67 (34.9)	-
≥ $140,000	50 (26.0)	-
Missing/refused	11 (5.7)	-
In school (vs. on school holiday)	139 (72.4)	116 (60.4)
Puberty status
Pre-puberty	23 (12.0)	4 (2.1)
In puberty	102 (53.1)	76 (39.6)
Completed puberty	67 (34.9)	112 (58.3)
	Means (SDs)	Means (SDs)
Moderate-to-vigorous physical activity (minutes/day)	36.3 (21.0)	25.5 (16.3)	<0.001
Total HEL-2015 score c	47.6 (11.8)	46.3 (11.5)	0.21
Body composition
Body mass index percentile	71.3 (30.2)	73.4 (28.4)	0.028
Body fat (%)	34.6 (10.3)	33.8 (11.2)	0.035
Fat mass (kg)	21.9 (14.4)	25.3 (17.0)	<0.001
Lean mass (kg)	36.8 (10.4)	43.8 (10.6)	<0.001
Visceral adipose tissue mass (kg)	0.55 (0.47)	0.63 (0.55)	<0.001
Cardiometabolic risk factors d
High-density lipoprotein cholesterol Z-score	0.02 (0.98)	0.05 (0.97)	0.39
Homeostatic Model Assessment for Insulin	−0.11 (0.83)	−0.07 (0.85)	0.55
Resistance Z-score
Mean arterial pressure Z-score	−0.08 (1.02)	0.002 (1.02)	0.23
Triglycerides Z-score	−0.04 (1.05)	−0.004 (1.06)	0.68
Waist circumference Z-score	−0.03 (1.01)	−0.04 (0.93)	0.82
Total cardiometabolic risk factors Z score	−0.26 (3.20)	−0.07 (3.31)	<0.001

^aPaired t-tests were used to compare values from baseline to follow-up.

^bParticipants marked “Other” when asked to report race given the following options: American Indian/Alaska Native, Asian, Native Hawaiian or Other Pacific Islander.

^cThe total HEL-2015 score is derived from scoring 13 dietary components, and maximum scores of 5 and 10 are possible for each component.

^dA standardized cardiometabolic risk z-score was calculated with five individual cardiometabolic risk components (high-density lipoprotein cholesterol, Homeostatic Model Assessment for Insulin Resistance, mean arterial pressure, triglycerides, and waist circumference) in accordance with accepted standards for pediatric practices.³⁰

Total HEI-2015 scores of the current sample indicated adolescents met around 50% of the recommendations at baseline (47.6 +/− 11.8) and follow up (46.3+/− 11.5). HEI-2015 mean component scores are shown in Figure 2.

Figure 2.

HEI-2015 Component Mean Scores and Standard Deviations^a Scored out of 5(A) and 10(B)

^a Panel A components have a maximum score of 5, and Panel B components have a maximum score of 10. Higher scores indicate greater conformance with 2015–2020 Dietary Guidelines for Americans. HEI-2015 component scores of moderation are reverse scored and include refined grains, sodium, added sugars, and saturated fats.

In multivariable-adjusted models (Table 2), baseline total HEI-2015 score was inversely associated with follow-up CMR z-score, HOMA-IR z-score, waist circumference z-score, BMI percentile, fat mass, lean mass, and VAT mass. When controlling baseline values, associations were still significant for HOMA-IR z-score, waist circumference z-score, BMI percentile, body fat, fat mass, and VAT mass.

Table 2.

Linear Regression Between Baseline Total Healthy Eating Index-2015 and Follow-up Body Composition and Cardiometabolic Risk Factors ^a

	Baseline Total Healthy Eating Index-2015 score
	Multivariable adjusted β	P value	Multivariable adjusted β	P value
Body composition during follow-up
Body mass index percentile	−0.428 c	0.010	−0.164 d	0.036
Body fat (%)	−0.110 b	0.066	−0.063 d	0.040
Fat mass (kg)	−0.191 b	0.046	−0.090 d	0.016
Lean mass (kg)	−0.136 b	0.014	−0.035 d	0.162
Visceral adipose tissue mass (kg)	−0.008 b	0.015	−0.004 d	0.009
Cardiometabolic risk factors during follow-up
High-density lipoprotein cholesterol Z-score	−0.004 c	0.499	−0.002 e	0.685
Homeostatic Model Assessment for Insulin Resistance Z-score	−0.016 c	0.002	−0.009 e	0.038
Mean arterial pressure Z-score	−0.012 c	0.059	−0.008 e	0.135
Triglycerides Z-score	−0.007 c	0.296	−0.003 e	0.594
Waist circumference Z-score	−0.014 c	0.012	−0.005 e	0.015
Total cardiometabolic risk factors Z-score	−0.052 c	0.008	−0.022 e	0.120

^aAssociations between baseline total Healthy Eating Index-2015 and follow-up body composition and cardiometabolic risk factors were determined using Linear Regression.

^bMultivariable adjusted for baseline age, sex, household income, in-school status, puberty (during follow-up also), and moderate-to-vigorous physical activity.

^cMultivariable adjusted for baseline household income, in-school status, puberty (during follow-up also), and moderate-to-vigorous physical activity.

^dMultivariable adjusted for baseline age, sex, household income, in-school status, puberty (during follow-up also), moderate-to-vigorous physical activity, and baseline corresponding cardiometabolic risk factors and body composition.

^eMultivariable adjusted for baseline household income, in-school status, puberty (during follow-up also), moderate-to-vigorous physical activity, and baseline corresponding cardiometabolic risk factors and body composition.

The multivariable-adjusted follow-up means of CMR factors and adiposity across tertiles of baseline HEI-2015 scores are presented in Table 3, indicating an inverse association between baseline total HEI-2015 score and follow-up BMI %ile, VAT mass, HOMA-IR z-score, waist circumference z-score, and total CMR z-score. After additional adjustment for baseline CMR factors and body composition, however, these associations became no longer significant except for BMI %ile and VAT mass.

Table 3.

Multivariable Adjusted Means of Follow-up Body Composition and Cardiometabolic Risk Factors According to Different Levels of Baseline Total 2015 Healthy Eating Index Score

	Baseline total 2015 Healthy Eating Index score			P for trend
	Tertile 1	Tertile 2	Tertile 3
No. of participants	64	64	64
Baseline total HEL-2015 score (range)	23.8–41.7	41.9–52.3	52.8–86.4
Multivariable adjusted means (SEs)
Body composition during follow-up
Body mass index percentile c	79.7 (3.32)	73.9 (3.31)	66.7 (3.32)	0.023
Body fat (%) b	35.3 (1.20)	33.9 (1.20)	32.3 (1.21)	0.204
Fat mass (kg) b	28.2 (1.92)	25.8 (1.92)	21.9 (1.92)	0.073
Lean mass (kg) b	45.8 (1.11)	44.2 (1.11)	41.4 (1.11)	0.022
Visceral adipose tissue mass (kg) b	0.78 (0.06)	0.61 (0.06)	0.50 (0.06)	0.011
Cardiometabolic risk factors during follow-up c
High-density lipoprotein cholesterol Z-score	0.17 (0.12)	−0.07 (0.12)	0.03 (0.12)	0.373
Homeostatic Model Assessment for Insulin Resistance Z-score	0.13 (0.10)	−0.08 (0.10)	−0.27 (0.10)	0.032
Mean arterial pressure Z-score	0.13 (0.13)	0.12 (0.12)	−0.24 (0.12)	0.059
Triglycerides Z-score	0.15 (0.13)	−0.15 (0.13)	−0.01 (0.13)	0.300
Waist circumference Z-score	0.14 (0.11)	0.02 (0.11)	−0.28 (0.11)	0.022
Total cardiometabolic risk factors Z-score	0.71 (0.40)	−0.15 (0.39)	−0.77 (0.39)	0.031
Multivariable adjusted means (SEs)
Body composition during follow-up
Body mass index percentile e	74.8 (1.57)	75.4 (1.55)	70.1 (1.56)	0.036
Body fat (%) d	34.6 (0.61)	34.2 (0.61)	32.7 (0.61)	0.068
Fat mass (kg) d	26.6 (0.75)	25.2 (0.74)	24.0 (0.75)	0.054
Lean mass (kg) d	44.4 (0.50)	43.5 (0.50)	43.6 (0.51)	0.407
Visceral adipose tissue mass (kg) d	0.69 (0.03)	0.63 (0.03)	0.57 (0.03)	0.008
Cardiometabolic risk factors during follow-up
High-density lipoprotein cholesterol Z-score e	0.16 (0.08)	−0.09 (0.08)	0.07 (0.08)	0.109
Homeostatic Model Assessment for Insulin Resistance Z-score e	0.05 (0.09)	−0.11 (0.09)	−0.16 (0.09)	0.219
Mean arterial pressure Z-score e	0.09 (0.11)	0.07 (0.11)	−0.16 (0.11)	0.204
Triglycerides Z-score e	0.11 (0.12)	−0.15 (0.12)	0.03 (0.12)	0.313
Waist circumference Z-score e	0.04 (0.04)	−0.06 (0.04)	−0.11 (0.04)	0.063
Total cardiometabolic risk factors Z-score e	0.40 (0.28)	−0.35 (0.28)	−0.26 (0.28)	0.128

^aAssociations between baseline tertiles of total 2015 Healthy Eating Index and follow-up body composition and cardiometabolic risk factors were determined using General Linear Model.

^bMultivariable adjusted for baseline age, sex, household income, in-school status, puberty (also during follow-up), and moderate-to-vigorous physical activity.

^cMultivariable adjusted for baseline household income, in-school status, puberty (also during follow-up), and moderate-to-vigorous physical activity.

^dMultivariable adjusted for baseline age, sex, household income, in-school status, puberty (also during follow-up), moderate-to-vigorous physical activity, and baseline corresponding cardiometabolic risk factors and body composition.

^eMultivariable adjusted for baseline household income, in-school status, puberty (during follow-up also), moderate-to-vigorous physical activity, and baseline corresponding cardiometabolic risk factors and body composition.

Sensitivity analysis.

In multivariable-adjusted models, the HEI-2015 component scores for greens and beans were inversely associated with total CMR factors (β =−0.273, p=0.031), HOMA-IR z-score (β=−.088, p=0.008), and BMI percentile (β=−2.637, p=0.015). Seafood and plant proteins were also inversely associated with total CMR factors (β=−0.237, p=0.033), HOMA-IR z-score (β=−0.076, p=0.009), triglycerides z-score (β=−0.085, p=0.022), BMI percentile (β=−1.939, p=0.042), and lean mass (β=−0.626, p=0.046). HEI-2015 component score for whole fruit was inversely associated with waist circumference z-score (β=−0.086, p=0.008), BMI percentile (β=−2.766, p=0.006), fat mass (β=−1.409, p=0.014), and lean mass (β=−0.914, p=0.006).

When stratified by puberty development, the inverse association was still significant among participants who remained in the completed puberty stage at baseline and follow-up for fat mass (β=−0.444, p=0.050), HOMA-IR z-score (β=−0.023, p=0.026), triglycerides z-score (β=−0.022, p=0.006), and total cardiometabolic risk z-score (β=−0.09, p=0.013) and was borderline significant for HOMA-IR z-score (p=0.073), waist circumference z-score (p=0.077), and total CMR z-score (p=0.072); and among participants who went from the pre-puberty stage at baseline to the in-puberty stage at follow-up for body fat (β=−0.35, p=0.017) and VAT mass (β=−0.013, p=0.034). The inverse association was also borderline significant among participants who remained in the pre-puberty stage at baseline and follow-up for BMI percentile (β=−0.53, p=0.082) and fat mass (β=−0.241, p=0.081). Analyses among participants who remained in the pre-puberty stage at baseline and follow-up yielded no significant associations.

DISCUSSION

This study examined the relationships among adolescents’ adherence to the 2015–2020 Dietary Guidelines for Americans and measures of CMR factors and adiposity. These findings indicated baseline HEI-2015 scores were inversely associated with CMR z-score, HOMA-IR z-score, waist circumference z-score, BMI percentile, fat mass, lean mass, and VAT mass among adolescents at a 2-year follow-up. However, when the baseline value of the dependent variable was included as a covariate, associations were attenuated to non-significance except for BMI percentile and VAT mass. In other words, the adolescent’s baseline adiposity or CMR z-score was a more powerful predictor of their cardiometabolic profile than their dietary intake, suggesting the adverse effects of a poor diet had already established a trajectory of adiposity and health risk in these adolescents.

The overall mean HEI-2015 score in this present study was 47.6 at baseline and 46.3 during follow-up, which is similar but slightly worse than the average HEI-2015 score for U.S. adolescents aged 9–13 years old (M=53) and aged 14–18 (M=49) based on the nationally representative adolescents surveyed in the National Health and Nutrition Examination Survey (NHANES) 2017–2018.⁴ Diet quality is a major factor associated with obesity;² thus, the low overall score in this sample may contribute to the higher prevalence of obesity in Louisiana as compared to the larger U.S. population.³¹

Considering overall eating patterns, these findings showed that adolescents with poor adherence to the 2015–2020 Dietary Guidelines for Americans and associated CMR factors continued this same trajectory over the 2-year observation. Trends in U.S. youth eating patterns have shown modest improvements between 1999 and 2016, with more youth moving from poor-quality diets to intermediate quality diets,³² which tracks with slowing, but not yet receding, youth obesity prevalence in the U.S.³³ Liu and colleagues (2020) suggest these improvements may be attributed to a shift towards focusing on healthy diet patterns, initiatives for increasing physical activity, the strengthening of child nutrition programs, and more rigorous standards for school meals throughout this time period (1999–2016).³² In fact, adolescents are more likely to receive meals in greater alignment with the 2015–2020 Dietary Guidelines for Americans while at school or when purchasing food from stores, compared to full-service and quick-service options.³⁴

The results found HEI-2015 component scores of greens and beans, seafood and plant proteins, whole fruit, and whole grains to be inversely related to CMR factors and body composition, and refined grain component scores to be positively related. It appears that participants within this sample are not consuming enough of the adequacy components, as compared to the moderation components. Current USDA-FNS requirements under the National School Lunch Program (NSLP) focus on the inclusion of fruits, vegetables, grains (80% whole grain-rich), meats/meat alternates, and fluid milk.³⁵ Based on present findings, to protect against CMR and increased body composition, guidelines and policies may consider increasing seafood and plant proteins. Additional research examining the associations between dietary patterns using alternative scoring procedures and guidelines and prospective anthropometric and cardiometabolic risk in adolescents may provide insight to additional options of health-promoting intake.

HEI-2015 added sugar component scores were not significantly associated with any of the anthropometric or cardiometabolic outcomes. The HEI-2015 version, compared with the HEI-2010 version, divided the previous scale of “empty calories” into added sugars and saturated fat scales, recognizing differences in the way carbohydrates and lipids are digested and metabolized.⁷ This finding is interesting as added sugar intake is often reported in the media as a target to improve childhood rates of obesity,³⁶ although research fails to support the contribution of sugar intake above and beyond matched macronutrient distributed groups.³⁷ The results of the current study suggest that patterns of intake, versus sugar intake specifically, contribute to associations with anthropometric and cardiometabolic outcomes.

Puberty is also a considerable risk factor associated with CMR factors and body composition. ^{38–40 41}. Our study indicated that the inverse association between baseline total HEI-2015 and select CMR factors and body composition was present for participants who went from the pre-puberty to in-puberty stages, in-puberty to completed puberty stages, and remained in the completed puberty stages from baseline to follow-up. Adolescence is a critical time to provide nutritional intervention to promote positive prospective anthropometric and cardiometabolic associated outcomes.

The present study has several strengths. First, the longitudinal study design allowed an assessment of baseline and 2-year follow-up change in diet quality in association with CMR factors and adiposity. Second, body composition components were measured repeatedly with MRI and DXA. These methods are more precise measures of adiposity than traditional BMI measurements and are the gold standard of research.^25,42 Third, the use of the HEI-2015 score to examine diet quality can be put into context of following the national dietary guidelines and compared to other samples or populations. Limitations to the current study must also be recognized. First, the sample was recruited from a metropolitan area in Louisiana where prevalence of adolescent obesity is higher than the national average.³¹ Thus, these findings may have a ceiling-effect and be relevant to this particular population and/or those populations at greater risk. Moreover, adolescents were recruited using convenience sampling, which may introduce sampling or selection bias. Second, the accuracy of diet recalls is subjected to bias from social, economic, and cultural factors. Third, though the data analyses adjusted for some confounding factors, significant factors, such as maternal pre-pregnancy BMI and genetic factors, were not measured and therefore could not be evaluated. These factors may be relevant for future research examining similar relationships and outcomes.^43,44

IMPLICATIONS FOR RESEARCH AND PRACTICE

Adolescents’ diet patterns fall far from meeting national dietary guidelines. This study demonstrated static patterns of insufficient diet quality among adolescents over a 2-year time span. Those who started at baseline with poorer diet quality had greater overall CMR and anthropometric measurements at follow-up among a sample of adolescents from Louisiana. These findings provide a novel understanding of the prospective relationships among adolescent diet quality, CMR factors, and body composition. It is important to examine these clinical markers of health risk in addition to anthropometry. The current study found specific patterns of diet quality that were associated with adolescent CMR factors and highlighted these findings for suggested guidance on dietary pattern recommendations and requirements. Promotion of nutrition knowledge is necessary, but knowledge does not equal adherence.⁴⁵ Nutrition knowledge is not consistently linked with food consumption behaviors thus,⁴⁶identifying barriers to consuming a healthful diet, as well as investigating effective strategies to overcome these barriers in adolescence may curtail future overall CMR and adiposity. Effective and timely intervention with a focus on adherence to dietary guidelines is necessary in improving diet quality and reducing overall CMR and adiposity in this age range.