Obesogenic Behaviors and Depressive Symptoms’ Influence on Cardiometabolic Risk Factors in American Indian Children

Abstract

BACKGROUND:

American Indian (AI) populations suffer disproportionately from cardiovascular disease and depression as compared to other racial/ethnic groups. Behaviors that contribute to obesity are considered obesogenic and include poor diet, low physical activity, and high screen time. This study examined the relationship between depressive symptoms and obesogenic behaviors on cardiometabolic risk factors in AI youth.

METHODS:

Participants (n=121) were evaluated for depressive symptoms, obesogenic behaviors, weight, blood pressure, lipids, and glucose levels.

RESULTS:

All participants failed to meet guidelines for intake of sugar-sweetened beverages and fruits/vegetables, 74% did not meet physical activity guidelines, and 85% did not meet screen time guidelines. Lower physical activity was associated with higher body fat percentage (β=−4.20 ± 1.82, p=0.022). Elevated depressive symptoms and presence of at-risk cardiometabolic risk factors were found. Higher depressive symptoms were associated with higher blood glucose (random, fasting, and hemoglobin A1c)

CONCLUSIONS:

Low physical activity, high screen time, and the presence of depressive symptomology heighten cardiometabolic risk factors in AI children. Associations between depressive symptoms and blood glucose underscore the impact of emotional health on cardiometabolic disease and emphasize need for proper depression assessment in chronic disease prevention efforts.

DISEASES that are apparently unrelated may share common etiological pathways. For example, depressive symptoms instigate biological responses that lead to cardiometabolic disease.⁽¹⁾ Obesity and cardiometabolic disease can lead to depressive symptoms,^(2,3) creating a cyclical metabolic pathology. This constellation of diseases is particularly troubling for children who have a high prevalence of these disorders, such as American Indians (AI).⁽⁴⁾ In a recent study, 63% of 7- to 13-year-old AI youth were overweight/obese,⁽⁵⁾ compared to 34% in the general population.⁽⁶⁾ Furthermore, AI children have the highest prevalence of type 2 diabetes as compared to US Blacks, Hispanics, and Whites.⁽⁷⁾ Twelve⁽⁸⁾ to 14%⁽⁴⁾ of AI children were hypertensive, compared to 2% in the general population of 8- to 12-year-olds.⁽⁹⁾

A review of 21 studies found that depressive symptoms are related to obesogenic behaviors in children and youth.⁽¹⁰⁾ Children who engage in obesogenic behaviors are at higher risk for obesity,^(11,12) cardiovascular disease,⁽¹³⁾ pediatric metabolic syndrome,^(14,15) and depressive symptoms.^(16–19) However, no previous studies concomitantly examined the emotional (specifically depression) and cardiometabolic risk factors or specifically included AI children. AI youth participate in more obesogenic behaviors than non-AI peers, which exacerbate the early development of pathological progression.⁽⁵⁾ For example, AI youth 7- to 13-years-old had higher daily consumption of sugar-sweetened beverages than the general population (309 vs 178 kcal for 2–111 year olds] and 286 kcal for 12–19 year olds),^(5,20) and the percentage of AI children meeting 1 hour of physical activity (PA)/day was lower than the general population (32% vs 70%).^(5,21) Understanding the relationship between behavioral risk factors and physical and mental health is important for effective interventions within high-risk populations.⁽²²⁾ Therefore, the purpose of this project was to examine the relationship between depressive symptoms and obesogenic behaviors on cardiometabolic risk factors in 7- to 13-year-old AI youth.

Methods

Study Design

This cross-sectional study included 7-to 13-year-old children (n=121) who attended a wellness summer camp. The wellness summer camp, Native Youth Preventing Diabetes, is a 5-day, 4-night diabetes prevention intervention. The intervention was offered at low or no-cost to participants from eligible coalition partners. Eligibility included freedom from diagnosed metabolic-disease. Self-reported behaviors (diet, physical activity, and screen time), self-reported depressive symptoms, and biological measures were ascertained by trained technicians. Study personnel assisted with reading comprehension as needed. Parents voluntarily consented and children voluntarily assented. University ethics committee and tribal coalition reviewed and approved this study and procedures.

Dietary Intake Behavior

An abbreviated Beverage Questionnaire-15 (Bevq-15)⁽²³⁾ was used to assess sugar-sweetened beverage consumption. Questions included frequency and volume of sweetened juice/beverages, regular soft drinks, diet soft drinks, sweetened teas, fruit juice, and energy/sports drinks. Summing the caloric value of the daily consumption of each beverage estimates the total daily sugar-sweetened beverage consumption in kilocalories.⁽²⁴⁾ The Bevq-15 is validated in children 9+ years.⁽²³⁾ While this tool has not been validated for younger children, no difference was observed in responses between the 7- and 8-year-olds and the older children.⁽⁵⁾ The selected psychometric properties for these questions are r²=0.52–0.95 (p<0.001), indicating a moderate-to-strong test-retest reliability.⁽²³⁾

Frequency of breakfast, lunch, and dinner consumption was assessed in three questions using the Project Eating Among Teens (EAT)-2010 Survey.^(25,26) Median values were used to calculate total daily meal frequency.⁽²⁷⁾

Select questions from the Youth Behavior Risk Surveillance Survey (YBRSS) examined fruit and vegetable intake frequency.⁽²⁸⁾ These questions assessed the weekly frequency of fruit, green salad, potato, carrot, and “other” vegetable consumption. The median value for each answer option range was used to calculate daily fruit and vegetable intake.⁽²⁹⁾ Psychometric properties for this survey instrument are unavailable.

Physical Activity and Screen Time Behavior

Time spent in physical activity (PA) and screen time (ST) was reported using the Project EAT 2010 Survey.⁽²⁶⁾ Three questions assessed participation in mild, moderate, and vigorous PA.⁽³⁰⁾ Eight questions assessed time in hours spent watching TV, using a computer, playing sedentary electronic games, and playing non-sedentary (i.e., exergaming) electronic games for week and weekend days.⁽³¹⁾ Median values were used to calculate daily time spent in PA and ST.⁽²⁷⁾ Daily moderate and vigorous PA were summed to create moderate + vigorous PA (MVPA).⁽²⁷⁾ Psychometric properties for this tool show a moderate test-retest reliability in 12- to 18-year-old adolescents: mild PA r=0.54, moderate PA r=0.53, and vigorous PA r=0.72, watching TV/DVD; r=0.81 using computer; r=0.84 sedentary video games; and r=0.73 non-sedentary video games.⁽³¹⁾ This tool has not been validated in children <12 years; however, no difference in responses between children of younger and older ages were previously observed.⁽⁵⁾

Depressive Symptoms

The 27-item Child Depression Inventory (CDI) was used to evaluate depressive symptoms.^(32,33) Each item of the CDI score was assigned a value from 0–2, with final score representing an average of all items. Participants with an average score ≤0.86 were considered to be negative for depressive symptomology, while participants with an average score ≥0.90 were considered to be positive for depressive symptomology. For this study, participants who scored between 0.87–0.89 or who showed an inclination for suicide were individually reassessed by an on-site licensed clinical psychologist for depressive symptomology.⁽³²⁾ The continuous average score ranging from 0 to 2 was an independent variable in this study. It has been validated in children ages 6–17 years (Cronbach’s alpha=0.86).

Cardiometabolic Risk Factors

Height in inches, without shoes, was measured using a portable Seca stadiometer (Seca Corp., Chino, CA). Participants were measured for weight (lbs) and body fat percentage using a Tanita TBF-310 Body Composition Analyzer (Tanita Corporation, Arlington Heights, IL). Participants’ body mass index (BMI) percentiles for age and sex were calculated⁽³⁴⁾ and classified as under-weight, healthy weight, overweight, or obese (i.e., at-risk ≥85th percentile).⁽³⁵⁾

Blood pressure was measured in triplicate using an automatic Omron blood pressure machine (Kyoto, Japan) and averaged. Blood pressure percentile was determined for age and gender.⁽³⁶⁾ At-risk blood pressure was ≥90th percentile as defined by the National High Blood Pressure Education Program.⁽³⁷⁾

Waist circumference was measured at midline using an abdominal circumference tension-tape by a registered nurse or registered dietitian and recorded in centimeters. Participants with a waist circumference measurement ≥90th percentile for age and sex were classified as at-risk.⁽³⁸⁾ Acanthosis nigricans, a thickening and darkening of skin associated with insulin resistance, was evaluated at the neck, armpit, elbow, and knee joint by a physician or physician assistant.⁽³⁹⁾ Those participants with acanthosis nigricans were classified as at-risk.

Fasting lipids and glucose, total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, VLDL cholesterol, and blood sugar were measured by a point-of-care lipid analyzer, the Alere Cholestech Analyzer (Alere Corporation, Waltham, MA), via finger-stick blood test after a controlled 10-hr fast. Participants’ outcomes were reported on a continuous scale and additionally classified as normal or at-risk. At-risk lipids were defined as LDL cholesterol >110 mg/dL, triglycerides ≥100 mg/dL, and HDL cholesterol <50 mg/dL.⁽⁴⁰⁾ At-risk fasting glucose was defined as ≥110 mg/dL.⁽¹⁵⁾ With the exception of the LDL cholesterol analysis, which is not utilized in metabolic syndrome classification, the Alere Cholestech Analyzer meets the National Cholesterol Education Program guidelines for acceptable laboratory accuracy.⁽⁴¹⁾

Hemoglobin A1c (HgbA1c) was measured by a point-of-care HgbA1c analyzer, the Siemens DCA Analyzer (Siemens Corp., Malvern, PA). Participants were assessed after a controlled 10-hr fast via finger-stick blood test. Participants’ outcomes were reported on a continuous scale and classified as normal or at-risk.⁽⁴²⁾ At-risk HgbA1c values were >5.7%.⁽⁴³⁾ The Siemens DCA analyzer shows good validity and reliability in a field setting,⁽⁴⁴⁾ and additionally, recent research suggests that in obese youth without diabetes, HgbA1c results combined with 2-hr post-meal blood glucose show a high rate of sensitivity to impaired glucose.⁽⁴⁵⁾ Participants were measured for post-prandial glucose tolerance via finger-stick blood test using a Freestyle Freedom Lite monitor (Abbott Laboratories, Abbott Park, IL) 2 hrs after a 75-g glucose challenge meal. Participant outcomes were reported on a continuous scale and classified as normal or at-risk.^(42,46) At-risk post-prandial glucose values were classified as ≥140 mg/dL.

Individual cardiometabolic outcomes were clustered using two mechanisms: a composite cardiometabolic risk score and presence of metabolic syndrome.⁽¹⁵⁾ The composite cardiometabolic risk score was the sum of presence of at-risk scores for each of the following: waist circumference percentile >75th, blood pressure >90th, LDL cholesterol >110 mg/dL, triglycerides >100 mg/dL, HDL cholesterol <50 mg/dL, fasting glucose >110 mg/dL, and acanthosis nigricans positive. For example, if a participant were at risk for waist circumference, triglycerides, HDL cholesterol, and acanthosis nigricans, their composite cardiometabolic risk score was 4. Metabolic syndrome is determined by the presence or absence of three or more of the following five risk factors: waist circumference percentile (>75th), blood pressure percentile (>90th), HDL cholesterol (<50 mg/dL), triglycerides (>100 mg/dL), and fasting glucose (>110 mg/dL).⁽¹⁵⁾ With the exception of adding LDL cholesterol and acanthosis nigricans to the continuous cardiometabolic risk score, the same constituent risk factors were used for both definitions; the former highlights the risk for chronic disease, the latter highlights presence of disease.

Data Analysis

Descriptive statistics included means ± standard deviation (SD) and frequency. Differences between sexes for obesogenic behaviors, depressive symptoms, and continuous scale cardiometabolic risk factors were analyzed using independent t-test. Differences between sexes for acanthosis nigricans were analyzed with Pearson’s chi-square. Association of obesogenic behaviors and depressive symptomology with individual and clustered cardiometabolic risk factors was assessed with a series of linear regression models. Logistic regression models were used to determine the association between obesogenic behaviors, depressive symptomology, acanthosis nigricans, and metabolic syndrome. Covariates included age and sex. Cohen’s d and r were calculated to determine effect size for independent samples t-tests. Data analyses used SPSS 10^® (IBM Corporation, Armonk, NY). p-Values <0.05 were significant.

Results

Obesogenic behaviors, depressive symptoms, and cardiometabolic risk factors for the total sample and by sex are shown in Table 1. The study participants’ mean age was 10.5 ± 1.6 yrs and 60% (n=73) were female. All participants exceeded the federal guidelines of zero sugar-sweetened beverages and consumed insufficient fruit/vegetable servings per day. Additionally, 74% had insufficient PA, while 85% exceeded ≥2 hrs of ST/day. The prevalence of at-risk for variables is as follows: BMI% (64%), waist circumference (60%), HDL (57%), triglycerides (17%), blood pressure (13%), total cholesterol (13%), depressive symptomology (12%), and LDL cholesterol (8%). Females had a significantly lower sugar-sweetened beverage intake, higher body fat percentage, lower HDL cholesterol level, and higher prevalence of acanthosis nigricans.

Table 1:

Obesogenic Behaviors, CDI score, Cardio-metabolic risk factor, metabolic syndrome for AI children ages 7–13 years by total sample and sex (n=121)

	Total Sample		Females (n=73)		Males (n=48)
	Mean ± SD	At-Risk %	Mean ± SD	At-risk %	Mean ± SD	At-risk %	p-value
Obesogenic Behaviors (daily)
All Sugar-sweetened beverages (kcal intake)	381.9 ± 248.1	100	323.7 ± 207.7	100	470.7 ± 279.0	100	0.002
All Fruit and Vegetable intake frequency (servings)	1.9 ± 0.9	100	1.9 ± 0.9	100	2.1 ± 1.0	100	0.338
All Meal frequency intake	2.4 ± 0.73		2.5 ± 0.7		2.3± 0.8		0.486
Moderate +vigorous physical activity (hours)	0.7 ± 0.5	74.4	0.7 ± 05	74	0.7 ± 0.6	75	0.503
Total screen time per average weekday (hours)	7.2 ± 5.7	83.5	6.5 ± 4.7	79.5	8.5 ± 6.9	89.6	0.079
Total screen time per average weekend day (hours)	7.3 ± 6.1	84.3	6.8 ± 5.4	82.2	8.1 ± 7.0	87.5	0.295
CDI score	0.4 ± 0.3	11.6	0.4 ± 0.3	11	0.4 ± 0.3	12.5	0.812
Body mass index percentile	80.7 ± 24.2	64	80.5 ± 25.7	66	80.9 ± 22.0	60	0.929
Waist circumference percentile	67.5 ± 26.3	60	66.9 ± 26.9	60	68.4 ± 25.6	60	0.744
Body fat percentage	26.8 ± 10.9	-	28.7 ±10.6	-	23.9 ± 11.0	-	0.019
Blood pressure percentile	55.6 ± 14.6	13	54.0 ± 12.5	10	58.2 ± 17.3	19	0.156
Total cholesterol (mg/dl)	144.3 ± 27.9	13	140.4 ± 25.8	8	150.2 ± 30.2	21	0.067
HDL cholesterol (mg/dl)	48.8 ±14.5	57	46.3 ± 13.3	66	52.6 ± 15.4	44	0.023
LDL cholesterol (mg/dl)	86.8 ± 27.1	8	84.5 ± 26.1	8	93.6 ± 29.8	8	0.303
Triglycerides(mg/dl)	88.7 ± 44.5	17	90.4 ± 47.8	23	83.4 ± 33.5	8	0.536
Post-prandial glucose(mg/dl)	95.3 ± 24.9	1	94.9 ± 31.3	1	95.8 ± 9.0	0	0.806
Fasting blood glucose(mg/dl)	87.0 ± 15.5	7	86.3 ± 17.2	4	88.0± 12.4	10	0.539
HgbA1c~ (%)	5.3 ± 0.9	5	5.4 ± 1.0	5	5.3 ± 0.7	7	0.759
Acanthosis nigricans(+ or −)	n/a	30	n/a	37	n/a	19	0.032
Cardio-metabolic risk factor score (range 0–7)*	1.3 ± 1.3 (0–5)	n/a	2.2 ± 1.5 (0–4)	n/a	2.0 ± 1.8 (0–5)	n/a	0.229
Metabolic syndrome risk score (range 0–5)**	1.6 ± 1.1 (0–5)	21	1.6 ± 1.0	23	1.4 ± 1.2	19	0.310

^~Does not include incomplete values, to include low or abnormal hemoglobin results (n=112)

^*Cardio Metabolic Risk is a continuous sum of risk scores for: blood pressure percentile, waist circumference percentile, HDL cholesterol, LDL cholesterol, triglycerides, fasting glucose, acanthosis nigricans

^**Metabolic syndrome is categorically defined by having at least three at-risk scores from the following: blood pressure percentile, waist circumference percentile, HDL cholesterol, triglycerides, fasting glucose. Pearson’s Chi-square used for categorical analysis such as acanthosis nigricans and metabolic syndrome.

Independent means t-tests used for continuous variables to include: BMI, Waist Circumference, Blood pressure, body fat, lipid and glucose variables.

Body mass index percentile was calculated using height, weight, gender and age in months

Blood pressure percentile was determine via the average of three separate blood pressures and plotted by gender, age and height percentile

Body fat percent was calculated by the Tanita 300 scale

Waist circumference percentile was measured in centimeters at midline and calculated using gender and age

Total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, fasting glucose, HgbA1c was obtained (via fingerstick) after a fast (>8 hours)

Post-prandial glucose was obtained 2 hours after a high carbohydrate, low fat, low fiber meal

Acanthosis nigricans was assessed by a trained medical professional (physician, nurse practitioner, physician assistant) and recorded as negative or positive

Depressive Symptoms—Child Depression Inventory (CDI) score is an average of a 27-item continuous score

Associations between obesogenic behaviors and depressive symptomology and individual and composite cardiometabolic risk factors are shown in Table 2. In models including both obesogenic behaviors and depressive symptomology, clustered cardiometabolic risk score and metabolic syndrome were not found to be associated; however, several individual scores were (Table 2). Lower MVPA was associated with higher body fat percentage. Higher ST was associated with higher HDL cholesterol, independent of depressive symptoms. Independent of obesogenic behaviors, depressive symptoms were associated with higher post-prandial glucose, fasting glucose, and HgbA1C. For example, for each point increase in CDI score, random glucose was 24.22 percentiles higher in this analysis.

Table 2.

Association between depressive symptoms (CDI score) and individual obesogenic behaviors on individual and clustered metabolic risk, adjusted for age and gender in 7 – 13 year old American Indian children without diabetes.

	Child’s Depression Inventory and Obesogenic Behaviors
	CDI	Daily Sugar Sweet Beverage Intake (kcal)	CDI	Daily Fruit and Vegetable Intake (servings)	CDI	Daily Meal Intake Frequency	CDI	Daily Moderate and Vigorous Physical Activity (time in hours)	CDI	Daily Weekday Screen time (time in hours)	CDI	Daily Weekend Screen time (time in hours)
Outcomes	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE	Beta ± SE
Body mass index percentile	7.13 ± 7.69	0.01 ± 0.01	6.44 ± 7.79	−2.98 ± 2.46	5.67 ± 8.24	−13.59 ± 17.32	6.26 ± 7.70	−1.04 ± 0.60	5.72 ± 7.92	0.53 ± 0.64	6.21 ± 8.14	0.30 ± 0.45
Waist circ. percentile	6.07 ± 8.34	0.02 ± 0.01	6.06 ± 8.50	−2.24 ± 2.69	4.83 ± 8.97	−14.12 ± 18.87	4.89 ± 8.32	−1.42 ± 0.64	4.25 ± 8.60	0.70 ± 0.49	3.48 ± 8.82	0.65 ± 0.49
Body fat percent	4.37 ± 3.36	0.01 ± 0.00	4.37 ± 3.41	−0.64 ± 1.09	3.94 ± 3.61	−4.48 ± 7.61	3.84 ± 3.33	−3.95 ± 1.83	4.31 ± 3.41	0.09 ± 0.21	4.15 ± 3.68	0.06 ± 0.20
Blood pressure percentile	−0.26 ± 4.71	0.00 ± 0.01	0.24 ± 4.77	0.42 ± 1.51	0.70 ± 5.03	43.03 ± 10.58	−0.21 ± 4.75	−0.145 ± 0.36	−0.81 ± 4.84	0.20 ± 0.28	−0.18 ± 4.97	0.04 ± 0.28
Total cholesterol (mg/dl)	−8.42 ± 8.68	0.00 ± 0.01	−8.48 ± 8.77	−0.25 ± 2.80	−8.16 ± 9.27	1.08 ± 19.57	−8.55 ± 8.71	−0.13 ± 0.68	−9.88 ± 8.95	0.34 ± 0.53	−11.06 ± 9.167	0.45 ± 0.52
HDL cholesterol (mg/dl)	−6.22 ± 4.55	−0.00 ± 0.01	−6.30 ± 4.60	−0.02 ± 1.47	−7.27 ± 4.85	−5.76 ± 10.25	−6.05 ± 4.57	0.16 ± 0.36	−8.88 ± 4.61	0.58 ± 0.27	−9.83 ± 4.73	0.59 ± 0.27
LDL cholesterol (mg/dl)	−11.89 ± 12.918	−0.00 ± 0.02	−13.27 ± 13.45	−1.27 ± 4.20	−16.15 ± 12.77	48.01 ± 29.4	−10.16 ± 12.45	2.13 ± 1.12	−10.09 ± 13.04	−0.61 ± 0.85	−8.08 ± 13.43	−0.76 ± 0.84
Triglycerides (mg/dl)	−16.86 ± 19.79	0.00 ± 0.03	−23.09 ± 19.79	−9.06 ± 6.65	−13.03 ± 20.05	34.48 ± 47.89	−16.33 ± 19.50	−0.80 ± 1.83	−20.15 ± 20.32	0.83 ± 1.37	−27.58 ± 21.08	1.68 ± 1.32
Post-prandial glucose (mg/dl)	22.05 ± 7.74	0.01 ± 0.01	24.22 ± 7.83	2.76 ± 2.49	22.13 ± 8.31	−3.87 ± 17.55	24.33 ± 7.73	1.01 ± 0.61	19.70 ± 7.97	0.69 ± 0.47	16.74 ± 8.08	1.01 ± 0.46
Fasting glucose (mg/dl)	9.81 ± 4.85	0.01 ± 0.01	10.45 ± 4.92	0.59 ± 1.57	10.93 ± 5.19	4.61 ± 10.96	11.19 ± 4.82	0.68 ± 0.38	7.98 ± 4.97	0.49 ± 0.29	6.78 ± 5.08	0.56 ± 0.29
HgbA1c (%)	0.73 ± 0.27	0.00 ± 0.00	0.84 ± 0.27	.16 ± 0.09	0.75 ± 0.29	−0.09 ± 0.62	0.81 ± 0.27	0.03 ± 0.02	0.70 ± 0.28	0.02 ± 0.02	0.61 ± 0.29	0.03 ± 0.02
Cardio-metabolic Risk*	0.29 ± 0.43	0.00 ± 0.00	0.27 ± 0.43	−0.04 ± 0.14	0.33 ± 0.45	0.22 ± 0.96	0.28 ± 0.43	−0.07 ± 0.24	0.29 ± 0.44	0.00 ± 0.03	0.34 ± 0.45	−0.01 ± 0.03
Odds Ratio	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)	(95% CI)
Acanthosis nigricans (+ or −)	1.77 (0.45, 6.92)	1.00 (1.00, 1.00)	1.80 (0.45, 7.13)	1.03 (0.65, 1.63)	1.73 (0.402, 7.409)	0.888 (0.04, 21.33)	1.81 (0.46, 7.10)	1.16 (0.53, 2.50	2.11 (0.51, 8. 8)	0.96 (0.09, 1.05)	2.01 (0.46, 8.73)	0.98 (0.90, 1.07)
Metabolic Syndrome**	3.059 (0.707, 13.231)	1.001 (0.999, 1.003)	3.303 (0.759, 14.374)	1.091 (0.662, 1.797)	4.241 (0.884, 20.340)	5.905 (0.153, 228.465)	3.003 (0.694, 12.986)	0.734 (0.304, 1.774)	3.471 (0.750, 16.063)	0.98 (0.89, 1.08)	2.72 (0.57, 12.99)	1.02 (0.94, 1.12)

^*Cardio Metabolic Risk is a continuous sum of risk scores for: blood pressure percentile, waist circumference percentile, HDL cholesterol, LDL cholesterol, triglycerides, fasting glucose, acanthosis nigricans

^**Metabolic syndrome is categorically defined by having at least three at-risk scores from the following: blood pressure percentile, waist circumference percentile, HDL cholesterol, triglycerides, fasting glucose

^-Bolded values indicate statistical significance (p<0.05)

Predictor variables

Child Depression Inventory (CDI) score is an average of a 27-item continuous score

Total sugar sweetened beverage (SSB) intake is a daily kilocalorie average of all SSB variables

Total fruit and vegetable intake is a daily frequency sum of all fruit and vegetable variables

Total meal frequency is a daily frequency sum of all meal variables

Moderate and vigorous physical activity is a daily time average(hours) of moderate and vigorous physical activity variables

Weekday screen time is a daily time average (hours) of all weekday sedentary screen time variables

Weekend screen time is a daily time average(hours) of all weekend sedentary screen time variables

Outcome variables

Body mass index percentile was calculated using height, weight, gender and age in months

Blood pressure percentile was determine via the average of three separate blood pressures and plotted by gender, age and height percentile

Body fat percent was calculated by the Tanita 300 scale for

Waist circumference percentile was measured in centimeters at midline and calculated using gender and age

Total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, fasting glucose, HgbA1c was obtained (via fingerstick) after a fast (>8 hours)

Random glucose was obtained 2 hours after a high carbohydrate, low fat, low fiber meal

Acanthosis nigricans was assessed by a trained medical professional (physician, nurse practitioner, physician assistant) and recorded as negative or positive

Discussion

This study contributes to the knowledge gap regarding obesogenic behaviors and cardiometabolic risk and depressive symptoms in AI children. While composite scores that included multiple risk factors were not associated with obesogenic behaviors or depressive symptomology, individual cardiometabolic risk factors were, indicating a link between early behaviors and/or depressive symptomology before cardiometabolic risk is manifest. This study also highlights the prevalence of obesogenic behaviors, depressive symptoms, and cardiometabolic risk factors in AI children, data rarely available.

AI children have higher rates of cardiometabolic risk factors than children in the general population. Sixty-four percent of AI children in our study were overweight/obese. These findings were similar to a 63% overweight/obesity finding in a previous study of AI children⁽⁵⁾ and higher than the national prevalence of 34% for overweight/obesity.^(6,47) Approximately 13% were considered hypertensive as compared to 2% nationally.⁽⁹⁾ Nearly 60% of participants had depressed HDL cholesterol compared to 23% nationally.⁽⁹⁾ Prevalence of metabolic syndrome was higher in this population than the general U.S. pediatric population (21% vs 3.5%)⁽⁴⁸⁾ and similar to national adult prevalence of metabolic syndrome in the (34%).⁽⁴⁹⁾ This finding emphasizes that AI children have a higher prevalence of cardiometabolic disease than the U.S. general pediatric population and forecasts chronic health problems with subsequent need for more advanced health care and higher medical-related expenses.

Approximately 74% of participants did not meet the federal PA guidelines of 60 min/day (compared to 30% nationally)⁽²¹⁾ and 85% do not meet ST guidelines of <2 hrs/day (compared to 46% nationally).⁽²¹⁾ All study participants failed to meet federal recommendations of zero servings of sugar-sweetened beverages (compared to 20% nationally).⁽⁵⁰⁾ The significant difference between sugar-sweetened beverage intakes by sex compounded with a medium-sized effect strengthens the argument that sex influences dietary intake. No study participants met the fruit and vegetable intake recommendation of five servings per day, which is similar to the national intake.⁽⁵¹⁾ With the exception of fruit and vegetable intake where the AI and general population both had a prevalence of near zero, AI children had higher prevalence of obesogenic behaviors. It is evident that nutrition and movement behavior change and improvement are needed within this population.

Higher depressive symptoms were associated with higher blood glucose variables (post-prandial glucose, fasting glucose, and HgbA1c), even as obesogenic behaviors that could attenuate the relationship were included. These findings support previous studies and highlight the biologic effects of depressive symptoms on insulin resistance⁽⁵²⁾ and demonstrate that depressive symptoms have a profound and possibly singular influence on blood glucose, since both hypertension and dyslipidemia are included in the insulin resistance progression⁽⁵³⁾ and are not associated with depressive symptoms in this population. This emphasizes the underlying biological impact of severe depressive symptoms on insulin resistance⁽⁵²⁾ and embraces the effect of chronic stress through depression on essential brain functions via the hypothalamic-pituitary-adrenal (HPA) axis and subsequent glucose control.⁽⁵⁴⁾ Additionally, in populations with a high prevalence of overweight/obesity, as is observed in AI, the effects of depressive symptomology on metabolic disorder may be more severe.⁽⁵⁵⁾

Contrary to the study hypothesis and other studies,^(56–58) dietary behaviors were not associated with cardiometabolic risk factors. This may be a result of the instrument focus on healthy rather than unhealthy food intake. For example, sugar-sweetened beverages and meal skipping are associated with cardiometabolic risk but fruit and vegetable intake was not.⁽⁵⁹⁾ In support of previous literature,^(48,60–62) higher MVPA was associated with lower body fat percentages. It should be noted that higher ST is associated with higher post-prandial glucose and fasting glucose and that this association is stable when controlling for depressive symptoms, possibly indicating that ST may play a role in this relationship. This is consistent with another study in a similarly-aged population that showed higher ST was found to be associated with a decreased insulin response and resistance.⁽⁶³⁾

In contrast to a recent review inversely linking sedentary time with HDL cholesterol,⁽⁶⁴⁾ ST was positively associated with HDL cholesterol independent of depressive symptoms. However, the review did not include depressive symptoms,⁽⁶⁴⁾ which may explain the difference. Our outcomes do support reports that depressive symptomology is associated with elevated HDL cholesterol.^(65,66)

Strengths and Limitations

The investigators included depressive symptoms and cardiometabolic risk factors in AI youth, which should be considered a strength of the current study. Due to the invasive nature of conducting these tests, this type of information has historically been unavailable in controlled study. While this was a field-based study, the equipment used for blood analyses had been shown to remain accurate in previous field-based studies.^(41,44)

Study limitations are centered on study type, sample size, and tools used. This study was cross-sectional, and therefore outcomes inference was limited to correlational analysis. Study recruitment was restricted to camp participation, and therefore, sample size and variability were solely controlled by camp participation, and opportunities to increase or change the make-up of the sample did not exist. Camp participation was open to all eligible AI children; however, this did not control for parent perception of camp eligibility and therefore may have been more likely to send children who were at higher risk for diabetes than children who were not. Data collection opportunities were restricted to the duration of the camp week. Although a search for age-appropriate tools was conducted, such tools were scarce and further limited by time availability of participants and the environment in which the study took place. Therefore, some of the study tools utilized had not yet been validated for the younger ages of this sample. The ability to generalize outcomes is limited due to the nature of this project; however, AI children in Oklahoma do stand to benefit from the outcomes. These outcomes may also serve as a comparison group to other minority groups such as children of African-Americans and Hispanic-Americans.

Conclusions

Participants in this study are at higher-risk for obesogenic behaviors, elevated body weight, severe depressive symptomology, hypertension, dyslipidemia, and metabolic syndrome than their non-AI counterparts. The positive association of depressive symptomology and elevated blood sugar variables remained significant when obesogenic behaviors were considered in the analysis model. Depressive symptomology and obesogenic behaviors did not show an increased risk with clustered risk scores such as cardiometabolic risk and metabolic syndrome. As expected, physical activity was associated with weight indicators, and weekday and weekend screen time was associated with blood glucose variables. Unexpectedly, higher weekday and weekend screen times were associated with higher HDL cholesterol when depressive symptomology was also considered.

The effects of both obesogenic behaviors and depressive symptomology on individual cardiometabolic risk factors and the comparatively high prevalence of depressive symptomology in this population are justification for inclusion of mental health intervention in disease prevention models. While interventions highlighting diet, physical activity, and reduced screen time remain necessary, the inclusion of prevention and treatment of depressive symptomology would be beneficial. Future research is called for to include further evaluation of the effect of behavioral health intervention in the chronic disease model and the differences in macronutrient percent intake between males and females and how this may affect chronic disease development.