Diet and Physical Activity by Prediabetes Status Among U.S. Adolescents: National Health and Nutrition Examination Survey,2007

Abstract

Background:

Prediabetes among adolescents is on the rise, yet it is unclear if modifiable risk factors vary by prediabetes status.

Methods:

This study examined associations between diet (primary objective) and physical activity (secondary objective) by prediabetes status among U.S. adolescents (12–19 years) who participated in the National Health and Nutrition Examination Survey from 2007–2018. Differences in Healthy Eating Index (HEI)-2015–2020 scores (total and 13 component scores), nutrients of public health concern, and physical activity were examined by prediabetes status (no prediabetes vs. prediabetes).

Results:

Adolescents (n = 2,487) with prediabetes had significantly lower whole grains component scores and intakes of vitamin D, phosphorus, and potassium (all p < .05), than adolescents without prediabetes. Physical activity levels were not optimal for either group, there were no differences by prediabetes status (n = 2,188).

Conclusion:

Diabetes prevention interventions for adolescents are needed and should promote a healthy diet target and encourage physical activity.

Introduction

The incidence of adolescents with prediabetes is growing at an alarming rate, impacting about 18% of U.S. adolescents (12–19 years).¹ According to the American Diabetes Association, a healthy diet and physical activity are modifiable risk factors shown to reduce prediabetes risk.² In general, U.S. adolescents (regardless of prediabetes status) are not meeting the Dietary Guidelines for Americans (DGAs)³ or Physical Activity Guidelines for Americans⁴ recommendations. What remains unclear is if adolescents with prediabetes have a less healthy diet and are less physically active than adolescents without prediabetes. This study examines differences for adolescents in overall diet quality and nutrients of public health concern (primary objective),³ and physical activity (secondary objective) by prediabetes status (no prediabetes vs. prediabetes) among a nationally representative sample of U.S. adolescents.

Materials and Methods

The current study analyzed data from adolescents (12–19 years) who participated in the 2007–2018 National Health and Nutrition Examination Survey (NHANES), a public access, cross-sectional survey that examines the health and nutrition of noninstitutionalized U.S. citizens.⁵ Data were collected in 2-year cycles and six cycles were included in analyses (2007/2008, 2009/2010, 2011/2012, 2013/2014, 2015/2016, 2017/2018). Data were publicly available and thus Institutional Review Board approval was not required.

The following individual and household characteristics were analyzed: age (12–15 or 16–19 years), sex (male or female), ethnicity/race (Mexican America/Other Hispanic, Non-Hispanic White, Non-Hispanic Black, or Other Race/Multiracial), BMI status, and family poverty to income ratio (PIR). For BMI status, weight (kilograms)/height (meters²) was calculated, and sex-specific BMI-for-age percentiles were reported: underweight/normal weight (<85th percentile), overweight (≥85th–<95th percentile), or obesity (≥95th percentile). The PIR was calculated (total family income/poverty threshold for a specific family size) and reported as low income (PIR <1.3), middle income (PIR ≥1.3 and <3.5), or high income (PIR ≥3.5).

Adolescents completed two 24-hour diet recalls (one in person and one telephone). Diet recalls were averaged to calculate Healthy Eating Index (HEI)-2015–2020 scores (total and 13 components),⁶ measure of adherence to the DGA recommendations (0–100 points for the total score, with higher scores indicating greater adherence to the DGA).³ The 13 components include items recommended for adequate consumption (total fruits, whole fruits, total vegetables, greens and beans, whole grains, dairy, total protein foods, seafood and plant proteins, fatty acids) and items recommended in moderation (refined grains, sodium, added sugars, saturated fats).³ Also, intake of nutrients under consumed by adolescents were analyzed (i.e., Vitamin D vitamin B6, Vitamin B12, choline, total folate, calcium, iron, magnesium, phosphorous, potassium, fiber).⁵

Physical activity level (PAL) was examined using two questions, as previously done in a similar study⁷: (1) “In a typical week do you do any vigorous-intensity sports, fitness, or recreational activities that cause large increases in breathing or heart rate like running or basketball for at least 10 minutes continuously?,” and (2) “In a typical week, do you do any moderate-intensity sports, fitness, or recreational activities that cause a small increase in breathing or heart rate such as brisk walking, bicycling, swimming, or volleyball for at least 10 minutes continuously?.” Findings were reported as low PAL (i.e., answered no to both questions), moderate PAL (i.e., answered yes to one question and no to the other question), or high PAL (i.e., answered yes to both questions).⁷

Prediabetes was defined using three laboratory measures: fasting plasma glucose (100 to <126 mg/dL), glycosylated hemoglobin (HbA1c; 5.7% to <6.5%), or 2-hour plasma glucose of the oral glucose tolerance test (140–199 mg/dL).⁸ If an adolescent met at least one of the aforementioned criteria for prediabetes, they were included in the prediabetes sample.

The primary objective was to examine associations between diet and prediabetes status (i.e., prediabetes or no prediabetes), and adolescents with two reliable diet recalls and at least one prediabetes laboratory measure were included in the analytic sample (n = 2487). The secondary objective was to examine associations between PAL and prediabetes status, and adolescents with responses to the two physical activity questions and at least one prediabetes laboratory measure were included in the analytic sample (n = 2188). Adolescents with diabetes were excluded from the analysis.

Statistical Analyses

Data analyses were conducted using SAS^® Survey Procedures (version 9.4; SAS Institute, Inc., Cary, NC) with fasting sampling weights⁵ and appropriate survey weights to account for the complex survey design. Differences in prediabetes status (no prediabetes vs. prediabetes) by demographic variables were determined using Rao-Scott chi-square test. Multivariate regression models were conducted to determine differences in HEI scores, nutrient intakes (PROC SURVEYREG), and PAL (PROC SURVEYLOGISTIC) by prediabetes status. Regression models of total HEI score and components were adjusted for age (categorical), sex (categorical), race and ethnicity (categorical), BMI (continuous), and PIR (categorical).

Regression models of nutrient intakes and PAL were adjusted for the same covariates and total energy intake (kcal; continuous). Square root transformation of nutrient variables was used in models to approximate normality, with nontransformed values presented in tables. Research suggests that adjusting for multiple comparisons in large observational studies increases Type II error,⁹ thus adjustments were not made. All reported p-values were two-tailed and p < 0.05 was considered statistically significant.

Results

Regarding the primary objective, the analytic sample of adolescents (n = 2487), 72.8% had no prediabetes and 27.2% had prediabetes (Table 1). Adolescents who were male (vs. female) with obesity (vs. normal weight) were more likely to have prediabetes (p < 0.05). No differences in the remaining individual and household characteristics were observed by prediabetes status (Table 1). Adolescents with prediabetes had minimally lower mean HEI total scores [45.0 (standard error [SE] 1.1)] than those without prediabetes [47.0 (SE 0.5); p = 0.09] (Table 2). Adolescents with prediabetes had lower whole grains HEI component scores [1.6 (SE 0.3)] than those without prediabetes [2.2 (SE 0.1); p = 0.04]. No other differences in HEI component scores were observed by prediabetes status.

Table 1.

Individual and Household Characteristics by No Prediabetes and Prediabetes Among U.S. Adolescents (12–19 Years), National Health and Nutrition Examination Survey 2007–2018 ^a

Primary objective characteristic (n = 2487)	No prediabetes (n = 1778)	Prediabetes ^b (n = 709)	p ^e
Primary objective characteristic (n = 2487)	n ^c (% ^d )	n ^c (% ^d )	p ^e
Age (years)			0.20
12–15	872 (49.3)	387 (53.1)
16–19	906 (50.7)	322 (46.9)
Sex			<0.001
Male	809 (45.2)	444 (63.9)
Female	969 (54.8)	265 (36.1)
Ethnicity and race			0.29
Mexican American/other Hispanic	590 (20.4)	258 (23.1)
Non-Hispanic White	538 (56.9)	188 (52.1)
Non-Hispanic Black	429 (14.3)	165 (15.2)
Other Race/Multi-racial	221 (8.4)	98 (9.6)
BMI status^f,g			<0.001
Underweight/normal	1141 (66.4)	373 (55.9)
Overweight	294 (16.3)	102 (14.5)
Obesity^h	310 (17.3)	228 (29.6)
Family PIR			0.12
Low income (<1.3)	821 (32.8)	360 (39.6)
Middle income (≥1.3 and <3.5)	623 (37.6)	238 (35.6)
High income (≥3.5)	334 (29.6)	111 (24.8)

Secondary objective characteristic (n = 2188)	No prediabetes (n = 1571)	Prediabetes ^b (n = 617)
Physical activity level
Low	383 (21.9)	146 (22.4)	0.61
Moderate	677 (40.9)	256 (43.3)
High	511 (37.1)	215 (34.3)

To examine associations between diet and prediabetes status, adolescents with two reliable diet recalls and at least one prediabetes laboratory measure were included in the analytic sample (n = 2487). To examine associations between PAL and prediabetes status, adolescents with responses to the two physical activity questions and at least one prediabetes laboratory measure were included in the analytic sample (n = 2188).

Prediabetes defined as fasting plasma glucose (100 to <126 mg/dL), glycosylated hemoglobin (HbA1c; 5.7% to <6.5%), or 2-hour plasma glucose of the oral glucose tolerance test (140–199 mg/dL).

Unweighted frequencies.

Weighted column percentages for categorical variables by prediabetes status.

For complex survey data such as NHANES, using the Rao-Scott F adjusted chi-square statistic is recommended to yield a more conservative interpretation than the Wald chi-square. P < 0.05.

n = 39 missing BMI status.

BMI status is based on BMI percentile adjusted for age and sex; underweight/normal includes ≤84th percentile; overweight includes 85 to ≤94th percentile; obesity includes ≥95th percentile.

Significant pairwise comparison via odds ratios (reference = underweight/normal).

NHANES, National Health and Nutrition Examination Survey; PAL, physical activity level; PIR, poverty-to-income ratio.

Table 2.

Healthy Eating Index-2015 Scores and Nutrients by No Prediabetes and Prediabetes Among U.S. Adolescents (12–19 Years), National Health and Nutrition Examination Survey 2007–2018 (n = 2487)

HEI-2015 component and total scores^a	Score range	No prediabetes (n = 1778)	Prediabetes^b (n = 709)	p ^c
HEI-2015 component and total scores^a	Score range	LS mean (SE)	LS mean (SE)	p ^c
Total fruits	0–5	2.27 (0.11)	2.00 (0.14)	0.09
Whole fruits	0–5	2.13 (0.12)	1.86 (0.16)	0.19
Total vegetables	0–5	2.46 (0.07)	2.39 (0.14)	0.66
Greens and beans	0–5	1.06 (0.09)	1.28 (0.16)	0.19
Whole grains	0–10	2.16 (0.11)	1.59 (0.26)	0.04
Dairy	0–10	6.76 (0.14)	6.28 (0.21)	0.10
Total protein foods	0–5	4.18 (0.07)	4.04 (0.11)	0.27
Seafood and plant proteins	0–5	1.91 (0.12)	1.77 (0.12)	0.44
Fatty acids	0–10	4.17 (0.15)	4.11 (0.17)	0.72
Refined grains	0–10	4.62 (0.20)	4.17 (0.31)	0.26
Sodium	0–10	4.13 (0.18)	4.45 (0.22)	0.33
Added sugars	0–10	5.20 (0.18)	5.13 (0.26)	0.83
Saturated fats	0–10	5.91 (0.18)	5.91 (0.25)	1.00
Total HEI Score^d	0–100	46.96 (0.48)	44.97 (1.13)	0.09
Energy and nutrients of public health concern for adolescents^d,e
Energy (kcal)		2093 (47)	2061 (71)	0.44
Vitamin D (D2+D3) (mcg)		5.2 (0.2)	4.5 (0.3)	0.03
Vitamin B6 (mg)		2.0 (0.04)	1.85 (0.08)	0.07
Vitamin B12 (mcg)		5.4 (0.1)	5.1 (0.3)	0.25
Choline (mg)		273.5 (5.5)	261.8 (9.8)	0.09
Total folate (mcg)		233.2 (8.8)	219.0 (9.7)	0.61
Calcium (mg)		1019 (15)	973.9 (24)	0.16
Iron (mg)		15.8 (0.3)	15.5 (0.5)	0.84
Magnesium (mg)		248.5 (2.9)	239.6 (9.7)	0.16
Phosphorus (mg)		1334 (11)	1284 (25)	0.01
Potassium (mg)		2281 (26)	2151 (77)	0.02
Fiber (gm)		14.4 (0.3)	13.9 (0.6)	0.33

All HEI models adjusted for age, sex, race, and ethnicity, BMI (continuous variable), and PIR.

Prediabetes defined as fasting plasma glucose (100 to <126 mg/dL), glycosylated hemoglobin (HbA1c; 5.7% to <6.5%), or 2-hour plasma glucose of the oral glucose tolerance test (140–199 mg/dL).

p-values reflect models of log-transformed dietary variables (P < 0.05); LS means presented are nontransformed (i.e., actual 2 day mean of nutrient intake).

Total HEI Scores may range from 0 to 100, with 100 indicating greatest adherence to the Dietary Guidelines for Americans.

All nutrient models (except energy) adjusted for age, sex, race and ethnicity, BMI (continuous variable), PIR, and total energy (kcal).

HEI, Healthy Eating Index; LS, least squares; SE, standard error.

On average, adolescents with prediabetes consumed significantly less vitamin D, phosphorous, and potassium than adolescents without prediabetes (all p < 0.05). Regarding our secondary objective (n = 2188), PALs (low vs. moderate vs. high) did not vary by prediabetes status (data not shown).

Discussion

The incidence of adolescents with prediabetes is concerning, especially since prediabetes is shown to track from adolescence to adulthood.¹ In the current study, males (vs. females) and those with obesity (vs. normal weight) were more likely to have prediabetes. Previous research suggests similar differences in type 2 diabetes by sex and BMI among U.S. adolescents.¹⁰

We found that U.S. adolescents with prediabetes consumed fewer whole grains, potassium, Vitamin D, and phosphorus than adolescents without prediabetes. An inverse relationship between whole grain intake and various glycemic measures is well documented.¹¹ Observational studies of adults with type 2 diabetes reveal that potassium intake improves glucose metabolism,¹² supporting our results. Vitamin D intake is shown to decrease insulin resistance and increase phosphorus absorption,¹³ further underscoring our findings. Similar to Al-Ibrahim and Jackson who examined associations between total HEI and prediabetes among U.S. adults in NHANES (≥20 years),¹⁴ we did not observe associations between total HEI and prediabetes status among adolescents.

In addition, added sugar intake did not vary by prediabetes status, possibly because added sugar consumption is universally high among adolescents due to consumption of ready to eat cereals and sugar sweetened beverages.¹⁵ Sneed et al. also did not report differences in added sugar intake by prediabetes status among adults in NHANES.¹⁶

The current study did not find differences in PAL by prediabetes status, despite evidence that better metabolic control is associated with increased physical activity among youth.¹⁷ It is important to note that ∼60% of adolescents with prediabetes and ∼60% without prediabetes reported that they do not even engage in moderate or vigorous physical activity for at least 10 minutes in a typical week. Thus, suggesting that national efforts are needed to increase physical activity for all adolescents.

This study has several limitations. For one, NHANES is a cross-sectional study and casual inferences cannot be made. Second, 24-hour diet recall data are subject to recall and social desirability bias, which may lead to mis- and/or underreporting, although NHANES diet recall data are reliable at the population level.¹⁸ Third, the physical activity questions in NHANES vary across cycles and by age, and do not ask if ≥60 minutes of daily physical activity is achieved per the American Academy of Pediatrics.¹⁹ The only questions examining overall physical activity across cycle years for 12–18-year-olds were the two questions we reported, which have been used to approximate PAL in other publications.^7,20 Fourth, NHANES does not collect prediabetes measures at multiple time points, which may lead to a misassessment of prediabetes status given that puberty occurs during adolescence.^1,8

Finally, NHANES does not collect pregnancy data on children 8–19 years (<1% of the sample) of age, and these individuals could not be excluded.

Conclusion

Given the prevalence of U.S. adolescents with prediabetes, interventions are needed to improve diet and physical activity. Future studies are needed to explore the mechanisms for which diet and physical activity effect prediabetes status among adolescents.

Impact Statement

The incidence of prediabetes among adolescents is growing at an alarming rate, yet, research about this population is limited. This study addresses a major knowledge gap regarding if dietary intake and physical activity varies by prediabetes status. Findings can inform the development of diabetes prevention interventions for adolescents.

Footnotes

Authors' Contributions

T.M.L.: conceptualization (lead); writing—original draft (lead); methodology (lead); and writing—review and editing (equal). F.O.: conceptualization (equal); writing—original draft (equal); methodology (equal); writing—review and editing (equal); software (lead); and formal analysis (lead). M.R.: conceptualization (equal); writing—original draft (equal); methodology (equal); and writing—review and editing (equal). C.W.L.: conceptualization (equal); writing—original draft (equal); methodology (equal); writing—review and editing (equal); software (lead); and formal analysis (lead). M.C.: conceptualization (equal); writing—original draft (equal); and writing—review and editing (equal). D.B.A.: writing—original draft (supporting) and review and editing (supporting).

Funding Information

No funding was received for this article.

Author Disclosure Statement

No competing financial interests exist.

References

Andes

, Cheng

, Rolka

, et al. Prevalence of prediabetes among adolescents and young adults in the United States, 2005–2016. JAMA Pediatr, 2020; 174(2):e194498; doi: 10.1001/jamapediatrics.2019.4498

Committee

ADAPP.

14. Children and adolescents: Standards of medical care in diabetes—2022. Diabetes Care, 2021; 45(Supplement_1):S208–S231; doi: 10.2337/dc22-S014

U.S. Department of Agriculture. Dietary Guidelines for Americans Make Every Bite Count With the Dietary Guidelines. 2020. Available from: https://www.dietaryguidelines.gov/sites/default/files/2020-12/Dietary_Guidelines_for_Americans_2020-2025.pdf

U.S. Department of Health and Human Services. Physical Activity Guidelines for Americans, 2nd edition. Washington, DC; 2018. Available from: https://health.gov/sites/default/files/2019-09/Physical_Activity_Guidelines_2nd_edition.pdf

Centers for Disease Control and Prevention. NHANES Questionnaires, Datasets, and Related Documentation. 2019. Available from: https://wwwn.cdc.gov/nchs/nhanes/

National Cancer Institute Division of Cancer Control & Population Sciences. Healthy Eating Index (HEI) | USDA-FNS. Available from: https://www.fns.usda.gov/resource/healthy-eating-index-hei

Murakami

, Livingstone

MBE

. Meal and snack frequency in relation to diet quality in US children and adolescents: The National Health and Nutrition Examination Survey 2003–2012. Public Health Nutr, 2016; 19(9):1635–1644; doi: 10.1017/S1368980016000069

American Diabetes Association. 2. Classification and diagnosis of diabetes: Standards of medical care in diabetes-2021. Diabetes Care, 2021; 44:S15–S33; doi: 10.2337/dc21-S002

Rothman

KJ.

No adjustments are needed for multiple comparisons. Epidemiology, 1990; 1(1):43–46.

10.

Demmer

, Zuk

, Rosenbaum

, et al. Prevalence of diagnosed and undiagnosed type 2 diabetes mellitus among US adolescents: Results from the continuous NHANES, 1999–2010. Am J Epidemiol, 2013; 178(7):1106–1113; doi: 10.1093/aje/kwt088

11.

, Fu

, Pan

, et al. Role of whole grain consumption in glycaemic control of diabetic patients: A systematic review and meta-analysis of randomized controlled trials. Nutrients, 2021; 14(1):109; doi: 10.3390/nu14010109

12.

Chatterjee

, Yeh

, Edelman

, et al. Potassium and risk of type 2 diabetes. Expert Rev Endocrinol Metab, 2011; 6(5):665–672; doi: 10.1586/eem.11.60

13.

Cândido

, Bressan

. Vitamin D: Link between osteoporosis, obesity, and diabetes?. Int J Mol Sci, 2014; 15(4):6569–6591; doi: 10.3390/ijms15046569

14.

Al-Ibrahim

, Jackson

. Healthy eating index versus alternate healthy index in relation to diabetes status and health markers in

U.S

. adults: NHANES 2007–2010. Nutr J, 2019; 18(1); doi: 10.1186/s12937-019-0450-6

15.

Bailey

, Fulgoni VL

III

, Cowan

, et al. Sources of added sugars in young children, adolescents, and adults with low and high intakes of added sugars. Nutrients, 2018; 10(1):102; doi: 10.3390/nu10010102

16.

Sneed

, Azuero

, Moss

, et al. Total added sugar consumption is not significantly associated with risk for prediabetes among U.S. adults: National Health and Nutrition Examination Survey, 2013–2018. PLoS One, 2023; 18(6):e0286759; doi: 10.1371/journal.pone.0286759

17.

McGavock

, Dart

, Wicklow

. Lifestyle therapy for the treatment of youth with type 2 diabetes. Curr Diab Rep, 2015; 15(1):568; doi: 10.1007/s11892-014-0568-z

18.

Blanton

, Moshfegh

, Baer

, et al. The USDA automated multiple-pass method accurately estimates group total energy and nutrient intake. J Nutr, 2006; 136(10):2594–2599.

19.

Lobelo

, Muth

, Hanson

, et al. Physical activity assessment and counseling in pediatric clinical settings. Pediatrics, 2020; 145(3); doi: 10.1542/peds.2019-3992

20.

Setiono

, Guerra

, Leung

, et al. Sociodemographic characteristics are associated with prevalence of high-risk waist circumference and high-risk waist-to-height ratio in U.S. adolescents. BMC Pediatr, 2021; 21(1):215; doi: 10.1186/s12887-021-02685-1

Diet and Physical Activity by Prediabetes Status Among U.S. Adolescents: National Health and Nutrition Examination Survey,2007–2018