Weight-Based Teasing in Youth: Associations with Metabolic and Inflammatory Markers

Abstract

Background:

Research among adults suggests that weight stigma is associated with worsened cardiometabolic health. However, these relationships have not been examined among youth.

Objective:

Assess associations between weight-based teasing (WBT) and metabolic and inflammatory markers among two samples of youth: (1) a non-treatment-seeking sample and (2) a weight loss treatment-seeking sample with obesity.

Method:

Weight, height, adiposity, waist circumference and blood pressure were measured. Fasting blood samples were collected for metabolic (triglycerides, glucose, high-density lipoprotein cholesterol) and inflammatory analytes (high-sensitivity C-reactive protein in Study 1 and erythrocyte sedimentation rate in both studies). Youths completed the Perception of Teasing Scale, a measure of WBT. Metabolic and inflammatory indices were compared between those with and without teasing, adjusting for demographics and body composition.

Results:

Study 1 enrolled 201 non-treatment-seeking youth (Mage = 13.1y; 54.2% female; 44.8% non-Hispanic White; 32.8% with overweight/obesity); 15.4% reported WBT. Study 2 enrolled 111 treatment-seeking adolescents with obesity (Mage = 14.0y; 66.7% female; 37.8% non-Hispanic White); 73.0% reported WBT. Adjusting for covariates, WBT was not associated with cardiometabolic risk factors in either study.

Conclusions:

WBT was not associated with worsened cardiometabolic health. Longitudinal research is needed to elucidate associations between WBT and health in youth.

Introduction

Many obesity-related comorbidities are only partially explained by excess adiposity, e.g.,¹ Research, therefore, has begun to elucidate psychosocial factors that, in conjunction with high body weight, may be associated with impaired metabolic function, hypertension, and susceptibility to metabolic syndrome, a cluster of risk factors for adverse cardiometabolic outcomes.² One such factor may be the presence of certain psychosocial stressors. Repeated exposure to psychosocial stressors elicits a cascade of physiologic responses that cumulatively confer additional risk for metabolic dysfunction,^{e.g., 3} inflammation, ⁴ and weight gain.⁵

A salient psychosocial stressor commonly reported by persons with high body weight is weight-based stigmatization, which refers to negative attitudes towards and derogation of persons perceived to have excess body weight. Weight stigma, which is experienced as both an acute and chronic stressor, appears to trigger activation of the hypothalamic pituitary adrenal (HPA) axis.^{e.g., 6,7} In support of this theory, research has found that observed and experienced weight stigma contribute to sustained salivary cortisol secretion in adults ^{e.g., 7,8} independent of the effect of body mass index (BMI, kg/m²). Further, among adults, weight-based discrimination is both cross-sectionally and longitudinally associated with greater concentration of scalp hair cortisol, a marker of chronic stress, even after adjusting for BMI.^9,10

Plausibly due to its associations with biochemical stress,¹¹ research among adults indicates that experiences of weight stigma may confer additional risk for the pathophysiology of obesity, above and beyond the contribution of excess adiposity. For instance, weight stigma has been found to amplify the link between central adiposity and nondiabetic glycemic control among adults with overweight.¹² In addition, a 10-year longitudinal study among adults found that, after adjusting for baseline BMI, perceived weight discrimination predicted twice the risk for high allostatic load, in addition to greater lipid and metabolic dysregulation.¹³

Given their appreciable role in obesity-related health conditions such as cardiovascular disease and type 2 diabetes,¹⁴ and associations with psychological factors, such as stress and depression,¹⁵ research has also sought to explore the links between weight stigma and inflammatory markers, such as C-reactive protein (CRP). Preliminary findings have indeed indicated that perceived weight-based discrimination is associated with higher circulating CRP among adults with overweight and obesity.^13,16 These findings indicate that weight stigma may be a salient biochemical stressor among adults that contributes to and exacerbates the pathophysiology of obesity.

To date, studies assessing the associations between weight stigma, markers of cardiometabolic health, and inflammation have focused on adults.¹⁷ However, youths with high body weight report frequent instances of weight-based teasing (WBT). Upwards of 60% of youths with overweight and obesity,^18,19 and up to 20% of youths without overweight¹⁸ report WBT; in fact, high body weight is the most frequently reported reason for teasing within schools.²⁰ Further, WBT may be experienced across contexts and relationships; youths experience WBT and negative weight-related comments from siblings, parents, teachers, coaches, and medical providers.^21,22

Of note, the experience of WBT among youth is associated with a number of adverse psychological and behavioral correlates, including body dissatisfaction, low self-esteem, depression, anxiety, and suicidal thoughts and behaviors.²³ Further, WBT is associated with behaviors that may contribute to and exacerbate excess weight, including unhealthy weight control behaviors, avoidance of physical activity and exercise, eating in secret, and binge-eating.²³ Perhaps due to these correlates, longitudinal research also demonstrates that WBT in childhood is associated with disordered eating and excess weight and fat gain through adolescence and into early adulthood, even after adjusting for baseline body composition.^18,24 Thus, WBT may place children and adolescents at increased risk for mood- and eating-related psychopathology and weight gain throughout development.

Despite the burgeoning literature documenting associations between weight-related stigma and metabolic dysregulation and markers of inflammation among adults, few studies have assessed these indices among youth. One prior study found that a distinct, but related construct, perceived pressure to be thin (e.g., being complimented for appearing to have lost weight), was associated with impaired insulin action and hyperinsulinemia among a sample of non-treatment-seeking youths, even after adjusting for body composition.²⁵ Given that insulin resistance naturally increases during puberty,²⁶ childhood and adolescence may be sensitive periods for worsening insulin resistance and cardiometabolic health among susceptible youths. Further, findings among adults suggest that identifying whether the associations between WBT and cardiometabolic risk factors emerge in youth has important implications for prevention and early intervention. To date, however, no research has assessed the associations between WBT, metabolic syndrome components, and markers of inflammation among children and adolescents.

Given the significant proportion of youths reporting WBT,¹⁹ in conjunction with data from adults identifying links between weight stigma and worsened metabolic function and inflammation,^7,11,13,16 research is needed to determine the associations between WBT, components of the metabolic syndrome, and inflammatory markers among youth. The present study, therefore, aimed to elucidate whether reported WBT is associated with components of the metabolic syndrome and markers of inflammation in two samples of youth: 1) a non-treatment-seeking sample of youth with and without overweight/obesity, and 2) a treatment-seeking sample of adolescents with obesity and obesity-related comorbidities presenting for a weight-loss medication trial. We hypothesized that, in both samples, the presence of WBT would be associated with elevated cardiometabolic risk factors and inflammatory markers above and beyond the contribution of demographic variables and body composition.

Methods

The current study is a secondary analysis of baseline data drawn from two studies conducted at the National Institutes of Health Hatfield Clinical Research Center. All participants and a parent/guardian completed two baseline visits, from which current data were drawn. At the first visit, written consent and assent were provided by parents and children, respectively. Participants completed interviews and questionnaires and underwent a physical examination. Anthropometric measurements were collected and blood was drawn by a phlebotomist or registered nurse following an overnight fast. A convenience sample restricted to participants who completed both the WBT questionnaire and a blood draw at the baseline visit were included in the current study. Study procedures were approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development Institutional Review Board.

Participants and Procedure

Study 1 is an ongoing observational longitudinal study initiated in 2015 (ClinicalTrials.gov ID# NCT02390765); other findings have been previously reported.^27-29 Participants were non-treatment-seeking boys and girls, 8-17 years old. Individuals were excluded from the study if they had underweight (i.e., BMI < 5^th percentile for age and sex)³⁰; a Full Scale Intelligent Quotient ≤ 70; a major psychiatric or medical condition, including type 2 diabetes; a history of significant or recent brain injury; regularly used medications or substances known to impact eating and/or weight; or reported a ≥ 5% reduction in weight during the past three months.

Study 2 was a randomized controlled trial designed to test the safety and efficacy of a weight-loss medication (orlistat); all participants completed the baseline visits prior to administration of the weight-loss medication. Study recruitment began in 1999 and concluded in 2008. Participants were treatment-seeking adolescents (12-17 years) with obesity (defined as BMI percentile ≥ 95^th for age and sex) and at least one obesity-related comorbidity (e.g., type 2 diabetes, hyperinsulinemia, significant hyperlipidemia, sleep apnea) but in otherwise good general health (ClinicalTrials.gov ID# NCT00001723). Inclusion was limited to non-Hispanic white and non-Hispanic black participants. Participants were excluded if they had other major medical conditions (e.g., Cushing syndrome), if they were using prescription medication not related to the obesity-related comorbidity or for weight loss, or if they or their parent/guardian had a psychiatric condition that could impede study completion. Results have been reported in prior studies.^31,32

Measures

Weight-based teasing.

The Perception of Teasing Scale (POTS)³³ was used to assess the presence of WBT due to high body weight. The POTS weight-based teasing score is derived from six items asking about the frequency of experiences such as, “People call me names like ‘Fatso’” and “People laugh at me for trying out for sports because I am heavy.” Participants respond to each item on a 5-point Likert scale of 1 (“Never”) to 5 (“Frequently”); total scores range from 6 to 30 with higher scores indicating more lifetime experiences of WBT. The POTS has demonstrated reliability in previous studies of youth.³⁴

Depressive symptoms.

Participants completed the Children’s Depression Inventory,³⁵ a 27-item self-report questionnaire of depressive symptoms over the previous two weeks, with greater scores indicating more pathology. The Children’s Depression Inventory is a valid and reliable measurement of depressive symptoms in youth.³⁶

Body composition.

Fasting weight was measured on a calibrated scale to the nearest 0.1 kg, and height was measured in triplicate on a stadiometer. Age- and sex-adjusted BMI-z was then calculated according to the CDC growth charts.³⁰ Fat mass (kg) was measured using dual-energy x-ray absorptiometry (Study 1: GE Lunar iDXA, GE Healthcare, Madison WI; software GE enCore 15, Study 2: Hologic QDR-4500A, Waltham, MA), or air displacement plethysmography (Study 2 only; Bod Pod, Life Measurement Instruments, Concord, CA). Both are validated measures of body composition.^37,38 In accordance with previous studies that have combined data from these two methods,^39,40 we added 2.29 kg to the measurement obtained by dual energy x-ray absorptiometry and multiplied fat mass obtained by air displacement plethysmography by 1.03.

Metabolic syndrome components.²

Waist circumference (cm) was measured at the iliac crest using a non-elastic tape measure. While participants were seated, systolic and diastolic blood pressure was measured using an automated blood pressure monitor (Dynamap, Philips). The NIH Clinical Research Center Department of Laboratory Medicine measured fasting triglycerides, plasma glucose, and high-density lipoprotein cholesterol (HDL-C) from blood samples using a Roche Cobas 6000 analyzer and reagents from Roche Diagnostics (Indianapolis, IN). Given the lack of consensus for specific cut-off values for a metabolic syndrome diagnosis in youth, ^{e.g., 41} all metabolic syndrome components were examined continuously.

Inflammation.

Inflammatory markers were measured at the NIH Clinical Research Center Department of Laboratory Medicine. Serum high-sensitivity C-reactive protein (hsCRP), a marker of chronic inflammation (Study 1), was measured using a Roche Cobas 6000 analyzer and reagents from Roche Diagnostics (Indianapolis, IN). Erythrocyte sedimentation rate (ESR) (Study 1 & Study 2), another marker of chronic inflammation, was measured using an Excyte 40 analyzer from ELITech Group (South Logan, UT).

Statistical Analysis

For both studies, all analyses were performed with IBM SPSS 25.0 (IBM Corp, Armonk, NY). Data were screened for normality. Extreme but plausible outliers across all variables, defined a priori as more than three standard deviations from the mean, were recoded to three standard deviations from the mean (<1.2% of data points) in order to retain these cases while minimizing their influence on analysis.^42,43 In accordance with prior research on weight stigma and indices of health in adults,^{e.g., 9,10,13,16} scores on the POTS were dichotomized to indicate presence versus absence of WBT. For both studies, participant characteristics by WBT status were compared using independent samples t-tests or chi-square tests, as appropriate.

For all primary analyses, the independent variable was presence vs. absence of WBT. All analyses adjusted for age, sex (female = 0, male = 1), and race (0 = non-Hispanic White, 1 = other). Analyses for metabolic syndrome components and inflammatory markers included total fat mass (kg) as an additional covariate.^44,45 Height was considered as a covariate in all analyses, but was removed from analyses as it was significantly correlated with age (r = .81) in Study 1, a violation of the assumption of non-multicollinearity among variables.⁴⁶ Given the potential overlap between WBT and depression, ^{e.g., 47,48} as well as the documented relationships between depression and metabolic health,^15,49 all models were repeated including depressive symptoms as an additional covariate. All tests were two-tailed, and differences were considered significant when p-values were < .05. Missing data for covariates and dependent variables in the primary analyses were minimal (< 7.5%) and were handled in each analysis by using listwise deletion. Study-specific analyses are detailed below:

Study 1.

HsCRP was log-transformed to improve normality. To examine the relationships between WBT and body composition, two analyses of covariance (ANCOVAs) were conducted, with BMI-z and fat mass as the dependent variables. The body composition variables could not be examined using a multivariate analysis of covariance (MANCOVA) due to violations of the assumption of homogeneity of covariance matrices. To examine the relationship between WBT and metabolic syndrome components,² a MANCOVA was conducted, with waist circumference, systolic blood pressure, diastolic blood pressure, triglycerides, glucose, and HDL-C as the dependent variables. Lastly, to examine the relationship between WBT and inflammatory markers, two ANCOVAs were conducted, with hsCRP and ESR as dependent variables. The inflammatory variables could not be examined using a MANCOVA due to violations of the assumption of homogeneity of covariance matrices. POTS was not assessed continuously in Study 1 due to the low prevalence of WBT, as well as the lack of variability in scores for youth who did report WBT.

Study 2.

Triglycerides were log-transformed to improve normality. To examine the relationships between WBT and body composition, a MANCOVA was conducted, with BMI-z and fat mass as the dependent variables. To examine the relationships between WBT and metabolic syndrome components,² a MANCOVA was conducted, with waist circumference, systolic blood pressure, diastolic blood pressure, triglycerides, glucose, and HDL-C as the dependent variables. Lastly, to examine the relationship between WBT and inflammatory markers, an ANCOVA was conducted, with ESR as the dependent variable. Given the variation in frequency of WBT reported by respondents in Study 2, exploratory secondary analyses were conducted using multiple linear regressions to determine whether there was an association between POTS score (continuous) and variables of interest.

Study 1

Results

A total of 201 children and adolescents (M_age= 13.1y ± 2.8, 54.2% female) completed the POTS. The racial/ethnic distribution was: 44.8% non-Hispanic White, 28.9% non-Hispanic Black, 9.0% Hispanic, and 17.4% other, multiple races, or unknown. Based on BMI-z (M_BMI-z: 0.6 ± 1.0, range: −1.6 to 2.8), 14.4% met criteria for overweight and 18.4% met criteria for obesity.³⁰ Fifteen percent (15.4%, n = 31) of the sample reported at least one experience of WBT; among those reporting WBT, the average score was 9.0 (SD = 4.1). Youths who reported at least one experience of WBT were older (p = .01), reported more depressive symptoms (p < .001), had higher BMI-z (p < .001), were more likely to have overweight or obesity (p < .001), and had greater fat mass (p < .001), higher waist circumference (p < .001), lower HDL-C (p = .003), and higher hsCRP (p = .003) than their peers who reported no WBT. There were no significant differences between those reporting teasing and those not reporting teasing in distribution of sex or race/ethnicity, triglycerides, systolic/diastolic blood pressure, plasma glucose, or ESR (ps > .05). Participant characteristics overall, as well as by WBT status, are shown in Table 1.

Table 1.

Study 1: Participant Characteristics Overall and by Weight-Based Teasing Status

	Total Sample (n = 201)	WBT Reported (n = 31)	WBT Not Reported (n = 170)	p
Age (years)	13.1 (2.8)	14.2 (2.9)	12.9 (2.7)	.01*
Sex, % (n)				.94
Female	54.2 (109)	54.8 (17)	54.1 (92)
Male	45.8 (92)	45.2 (14)	45.9 (78)
Race, % (n)				.13
Non-Hispanic White	44.8 (90)	32.3 (10)	47.1 (80)
Non-Hispanic Black	28.9 (58)	45.2 (14)	25.9 (44)
Hispanic	9.0 (18)	6.5 (2)	9.4 (16)
Other/Unknown	17.4 (35)	16.1 (5)	17.6 (30)
BMI-z	0.6 (1.0)	1.4 (0.8)	0.4 (1.0)	<.001*
Weight Status, % (n)				<.001*
Healthy Weight	67.2 (135)	32.3 (10)	73.5 (125)
Overweight	14.4 (29)	29.0 (9)	11.8 (20)
Obesity	18.4 (37)	38.7 (12)	14.7 (25)
Depressive Symptoms	6.6 (5.2)	9.9 (5.6)	6.0 (4.8)	<.001*
Fat Mass (kg)	16.2 (10.0)	25.6 (12.5)	14.5 (8.4)	<.001*
Waist Circumference (cm)	76.6 (13.7)	88.5 (14.8)	74.3 (12.3)	<.001*
Triglycerides (mg/dL)	73.0 (35.8)	74.1 (35.5)	72.8 (35.9)	.86
HDL-C (mg/dL)	57.0 (14.9)	50.8 (10.5)	58.0 (15.3)	.003*
Systolic Blood Pressure (mmHg)	110.8 (10.3)	112.8 (10.2)	110.4 (10.3)	.24
Diastolic Blood Pressure (mmHg)	61.4 (7.6)	62.7 (7.0)	61.2 (7.7)	.33
Glucose (mg/dL)	90.7 (6.5)	92.5 (5.8)	90.4 (6.6)	.14
ESR (mm/hr)	8.1 (6.0)	9.7 (7.0)	7.9 (5.7)	.17
HsCRP1 (mg/L)	1.2 (2.0)	2.6 (3.1)	0.9 (1.6)	.003*

Data presented as M (SD) unless otherwise noted.

¹Untransformed means and standard deviations shown.

^*Group differences are significant at p < .05 for independent samples t-tests or Chi-square analyses, as appropriate. Abbreviations: BMI-z, body mass index adjusted for age and sex; HDL-C, high-density lipoprotein cholesterol; ESR, erythrocyte sedimentation rate; hsCRP, high-sensitivity C-reactive protein.

Weight-Based Teasing and Metabolic and Inflammatory Markers

After adjusting for covariates, youths who reported WBT had significantly higher BMI-z (1.41 ± 0.18 versus 0.40 ± 0.08; p < .001) and fat mass (24.20 ± 1.5 kg versus 14.76 ± 0.65 kg; p < .001) than youths who did not report WBT; findings remained the same after adjusting for depressive symptoms (ps < .001). Using Pillai’s trace, there was no significant association of teasing status with metabolic syndrome components (waist circumference, systolic blood pressure, diastolic blood pressure, triglycerides, glucose, and HDL-C) [V = 0.04, F(6, 164) = 1.06, p = .39]. Results remained the same after adjusting for depressive symptoms (p = .36). There were also no significant differences in inflammatory markers between those with and without reported teasing (hsCRP: 0.65 ± 0.09 mg/L versus 0.55 ± 0.04 mg/L; p = .31; ESR: 7.52 ± 1.00 mm/hr versus 8.33 ± 0.40 mm/hr; p = .47); these results remained the same after adjusting for depressive symptoms (hsCRP: p = .21, ESR: p = .56).

Summary of Findings: Study 1

In this study of non-treatment-seeking children and adolescents with and without overweight/obesity, we found that, in unadjusted analyses, those who reported WBT were more likely to have overweight or obesity and had greater BMI-z and waist circumference, greater depressive symptoms, lower HDL-C, and greater inflammation (higher hsCRP) than peers who did not report WBT. After adjusting for demographics, significant differences in BMI-z and fat mass persisted. However, after adjusting for demographics and body composition, there were no significant associations between teasing status and metabolic syndrome components nor inflammatory markers. Including depressive symptoms as a covariate did not impact the pattern of findings.

Study 2

Results

A total of 111 adolescents (M_age= 14.0y ± 1.4; 66.7% female) completed the POTS and were included in the present study. The racial/ethnic distribution was: 62.2% non-Hispanic Black and 37.8% non-Hispanic White. All participants had obesity (M_BMI-z: 2.5 ± 0.3, range: 1.8 to 3.5). The majority (73.0%, n = 81) of youths reported at least one experience of WBT; among those reporting WBT, the average score was 13.0 ± 5.4. Compared to youths reporting no WBT, youths reporting at least one experience of WBT reported more depressive symptoms (p = .01) and had lower plasma glucose (p = .003). There were no significant differences between those with and without reported teasing on age, sex distribution, race distribution, BMI-z, fat mass, waist circumference, triglycerides, HDL-C, systolic/diastolic blood pressure, nor ESR (ps > .05). Participant characteristics overall as well as by WBT status are shown in Table 2; Correlations with WBT score are shown in Table 3.

Table 2.

Study 2: Participant Characteristics Overall and by Weight-Based Teasing Status

	Total Sample (n = 111)	WBT Reported (n = 81)	WBT Not Reported (n = 30)	p
Age (years)	14.0 (1.4)	14.0 (1.4)	14.0 (1.4)	.94
Sex, % (n)				.65
Female	66.7 (74)	65.4 (53)	70.0 (21)
Male	33.3 (37)	34.6 (28)	30.0 (9)
Race, % (n)				.47
Non-Hispanic White	37.8 (42)	35.8 (29)	43.3 (13)
Non-Hispanic Black	62.2 (69)	64.2 (52)	56.7 (17)
BMI-z	2.5 (0.3)	2.6 (0.3)	2.5 (0.3)	.43
Depressive Symptoms	7.5 (5.8)	8.4 (6.0)	5.1 (4.2)	.01*
Fat Mass (kg)	53.6 (17.4)	55.1 (18.5)	49.8 (13.6)	.16
Waist Circumference (cm)	108.3 (13.6)	109.3 (13.7)	105.6 (109.3)	.22
Triglycerides (mg/dL)1	110.7 (76.7)	107.8 (65.5)	118.5 (101.5)	.69
HDL-C (mg/dL)	43.7 (9.0)	43.5 (9.1)	44.2 (8.7)	.71
Systolic Blood Pressure (mmHg)	123.2 (12.7)	123.0 (13.1)	123.5 (11.7)	.85
Diastolic Blood Pressure (mmHg)	67.8 (8.5)	67.2 (7.9)	69.4 (9.9)	.21
Glucose (mg/dL)	88.5 (9.0)	87.0 (8.4)	92.6 (9.4)	.003*
ESR (mm/hr)	28.4 (19.3)	29.7 (20.5)	24.9 (15.5)	.25

Data presented as M (SD) unless otherwise noted.

¹Untransformed means and standard deviations shown.

^*Group differences are significant at p < .05 for independent samples t-tests or Chi-square analyses, as appropriate. Abbreviations: BMI-z, body mass index adjusted for age and sex; HDL-C, high-density lipoprotein cholesterol; ESR, erythrocyte sedimentation rate.

Table 3.

Study 2: Correlations between Weight-Based Teasing Score and Continuous Study Variables

	r	p
Age (years)	−0.24	.01*
Depressive Symptoms	0.27	.01*
BMI-z	0.26	.01*
Fat Mass (kg)	0.32	.001*
Waist Circumference (cm)	0.30	.001*
Triglycerides (mg/dL)1	−0.02	.82
HDL-C (mg/dL)	−0.06	.55
Systolic Blood Pressure (mmHg)	−0.12	.20
Diastolic Blood Pressure (mmHg)	−0.01	.90
Glucose (mg/dL)	−0.17	.09
ESR (mm/hr)	0.18	.06

¹Log-transformed to improve normality.

^*Significant at p < .05. Abbreviations: BMI-z, body mass index adjusted for age and sex; HDL-C, high-density lipoprotein cholesterol; ESR, erythrocyte sedimentation rate.

Weight-Based Teasing and Metabolic and Inflammatory Markers

After adjusting for covariates, there was no significant association of teasing status with body composition [BMI-z and fat mass; V = 0.02, F(2, 103) = 1.10, p = .34]. Using Pillai’s trace, there was no significant association of teasing with metabolic syndrome components [waist circumference, systolic blood pressure, diastolic blood pressure, triglycerides, glucose, and HDL-C; V = 0.09, F(6, 95) = 1.58, p = .16]. There was also no significant difference in ESR between youths who reported WBT and youths who did not (29.1 ± 2.1 mm/hr versus 26.6 ± 3.3; p = .52). The pattern of results remained the same after adjusting for depressive symptoms (ps = .07 – .53).

Exploratory Analyses Using Continuous Measure of Weight-Based Teasing

After adjusting for specified covariates, the frequency of WBT (continuous) was significantly associated with BMI-z (β = .20, p = .02) and fat mass (β = .32, p = .001). However, teasing frequency was not significantly associated with waist circumference (β = .05, p = .34), systolic blood pressure (β = −.19, p = .06), diastolic blood pressure (β = −.08, p = .47), triglycerides (β = .04, p = .66), glucose (β = −.18, p = .09), nor HDL-C (β = −.04, p = .72). Lastly, frequency of WBT was not significantly associated with ESR (β = .02, p = .82). All findings remained the same after adjusting for depressive symptoms.

Summary of Findings: Study 2

In this sample of treatment-seeking adolescents with obesity, we found that almost three-quarters of participants reported at least one experience of WBT. In unadjusted analyses, presence of WBT was associated with greater depressive symptoms and lower plasma glucose. After adjusting for relevant covariates, the presence (vs. absence) of WBT was not associated with body composition, whereas WBT severity was significantly and positively associated with both BMI-z and fat mass, corroborating prior findings indicating that WBT increases with BMI in weight loss treatment-seeking youths.¹⁹ After adjusting for covariates, neither the presence nor frequency of WBT was associated with metabolic syndrome components or markers of chronic inflammation. The patterns of findings all remained the same after adjusting for depressive symptoms.

General Discussion

Across two samples of youth, one comprising non-treatment-seeking youth of a broad weight distribution and the other comprising weight loss treatment-seeking youth with obesity, WBT was significantly associated with BMIz and fat mass before adjusting for covariates. However, after adjusting for relevant covariates including body composition, neither the presence nor frequency of WBT was significantly associated with a worsened metabolic state nor markers of inflammation.

Our finding that WBT was not associated with worsened cardiometabolic health in youth after adjusting for body composition contrasts with previous studies in adults observing links between weight stigma and discrimination and markers of inflammation and poor health.^{e.g.,9,10,12,13,16} It is possible that any associations between WBT and health in youth are primarily accounted for by body composition. It is also possible that these associations only begin to emerge in adulthood, perhaps due to greater chronicity of weight stigma and resultant chronic stress,^{e.g., 6,8} or after longer duration of obesity. Additional research should seek to clarify the nature of these associations across more heterogeneous samples. More specifically, research should be conducted among considerably larger, more heterogeneous samples enriched for overweight and obesity in order to elucidate whether associations between WBT and indices of health are primarily observed at certain BMI strata or among individuals with a more advanced obesity-related disease state. This is plausible given the moderating role of weight status observed among prior adult studies.¹⁶

As it is well-documented that youth with overweight and obesity report greater frequency of WBT than youth without,^{18,19,21,50,51} it will also be important to assess severity of WBT in future studies of youth with high body weight. Although studies of adults typically dichotomize weight-based discrimination, ^{e.g., 9,12,13} future studies should aim to determine whether there is a “dose-response” to teasing, such that youths with more frequent or persistent teasing present with greater cardiometabolic risk as compared to youths with less frequent or no reported teasing. Longitudinal research is also warranted to track how WBT impacts and interacts with metabolic health and markers of inflammation across pubertal development and into early adulthood.

Future research should also consider how different aspects of WBT relate to cardiometabolic health. Though a widely used and well-validated measure, the POTS assesses relatively high-threshold teasing experiences pertaining specifically to high body weight (e.g., “people snickered about your heaviness when you walked into a room alone”) and may not reflect common subtle forms of teasing experienced by youths such as weight-related comments made over social media or exclusion from peer groups. These high-threshold items may also have been less relevant for youth across the broader weight spectrum who still experience forms of weight-based teasing even in the absence of overweight.

It will also be important to consider timing of WBT assessment in future studies of youth. Assessments for both samples included in the current study were conducted year-round; however, it is possible that the frequency or severity of WBT (and its sequelae) may vary across the school year, for instance, with children recalling and thus reporting more teasing incidents during the school year. It is also possible that the physiologic effects of WBT may be strongest immediately after the teasing incident, as suggested by studies among adults observing increased cortisol reactivity following laboratory exposure to weight stigma.^{e.g., 8} Alternatively, chronic and persistent teasing may be more relevant for the markers of cardiometabolic health assessed in the current study, which may manifest over time. Notably, the POTS does not specify chronicity, nor source of teasing. It is likely that certain forms of teasing are more recurrent and aversive than others; for instance, it may be that repeated teasing by one’s parents or siblings is more stressful and potentially more robustly linked with adverse outcomes than infrequent negative remarks made by strangers or acquaintances. For example, a recent systematic review found that WBT in childhood was associated with elevated depressive symptoms, and that this association was strongest among children reporting teasing from both family members and peers (versus either alone), suggesting a dose-response relationship.⁵² Assessing the sources, frequency, and chronicity of WBT and elucidating their associations with physical health among youth will be a critical next step. Further, future research should prioritize the development and validation of updated measures of WBT in youth that address the limitations of the POTS; given the range of methods currently used to assess WBT, ^{e.g., 52} greater measurement consistency will allow for comparisons across studies and samples.

Strengths of the current study include the use of two diverse samples of youths of a broad age and BMI range. Whereas many prior studies of WBT in youth utilize either non-treatment-seeking or clinical samples, the current study replicated measures and analyses across two distinct samples: one non-treatment-seeking, and the other, a weight loss treatment-seeking sample with obesity and obesity-related comorbidities, thereby enabling greater generalizability and translation of findings. Laboratory assessment of metabolic function and inflammation was another strength of the study, as was the objective measurement of adiposity, which was included as a covariate in all analyses.

The lack of significant findings in the current study should be considered in light of several factors. Given that assessing WBT and its physiologic sequelae was not a primary aim of either original study, no a priori power estimates were made for the current study aims; thus, the current studies should be considered exploratory. In addition, the two studies were not conducted contemporaneously; thus it is possible that various cohort effects exist that may have influenced findings and limited the ability to draw conclusions from the combined results. Further, fasting measures of insulin resistance and sensitivity are considered optimal for the more subtle dysregulation in glucose homeostasis that may be present in samples without diabetes. In contrast, fasting glucose concentrations, assessed in the present studies, may remain in the normal range through alterations of insulin concentrations and thus may not reflect subtle underlying dysregulation.⁵³

The cross-sectional nature of the study is another limitation which precludes our understanding of the associations between WBT and markers of health over the course of development. Although the wide age range of the sample participants (8-17 years) is a unique aspect of our study and could be considered a strength, it also may have inhibited our ability to detect differences based on teasing status. For instance, it is possible that the adverse physiologic effects of WBT documented among adults only begins to manifest in late adolescence or early adulthood, or only among those for whom WBT is a chronic stressor. Although WBT and anti-fat attitudes begin in early childhood, it is possible that youths in elementary school may not be exposed to the same frequency, severity, or overtness of teasing as youths in middle school and high school.²¹ Therefore, it is possible that the inclusion of both children and adolescents in the current study may have blunted our ability to detect associations that manifest later in adolescence or early adulthood. However, Study 2 enrolled only adolescents, and no significant relationships were observed, though the inclusion criteria and subsequent restricted range of values may have influenced these findings. Finally, the relatively low prevalence of metabolic syndrome in youth,⁵⁴ and the fact that our sample comprised generally healthy youths with no significant medical conditions (other than at least one obesity-related comorbidity in Study 2) may have also limited our ability to detect associations between WBT and markers of physical health. Future research should examine these relationships in samples across the weight spectrum, including youth with severe obesity.

Despite these considerations, no previous research has assessed the associations between WBT and markers of cardiometabolic health among children and adolescents; therefore, the current study adds to our understanding of the physiologic sequelae of WBT among youth. Although WBT was associated with some components of metabolic syndrome and markers of inflammation, these associations were attenuated and no longer significant once relevant covariates, including adiposity, were added to the models. Thus, our findings did not corroborate studies among adults linking experiences of weight stigma with metabolic health and markers of inflammation.^12,13,16 Nonetheless, the present study is a first step in the effort to elucidate the physiologic sequelae of WBT among youth. Future research is needed among more heterogeneous samples of youth to confirm or challenge current findings.