Associations between different triglyceride glucose index-related obesity indices and eating disorders: results from NHANES 2005–2018

doi:10.21203/rs.3.rs-5317726/v1

Download PDF

Research Article

Associations between different triglyceride glucose index-related obesity indices and eating disorders: results from NHANES 2005–2018

https://doi.org/10.21203/rs.3.rs-5317726/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

This research aimed to determine the possible links between obesity measures related to the triglyceride glucose (TyG) index and the prevalence of eating disorders (ED) for the United States residents.

Methods

This observational investigation analyzed data from the National Health and Nutrition Examination Survey (NHANES) 2005–2018. It assessed the relationship of the TyG index, TyG combined with waist circumference (TyG-WC), or TyG combined with body mass index (TyG-BMI) with ED. The analysis employed a multivariable regression model, stratified analyses, and a ROC curve assessment.

Results

This research included a total of 10,324 adults. In the comprehensive analysis model, the TyG, TyG-BMI, along with TyG-WC all had a significant positive correlation with ED. The adjusted graphical representations revealed a rising trend in the association of TyG-BMI index with ED. Subgroup analyses indicated that individuals with hypertension exhibited even stronger positive associations between these indices and ED. The areas under the curve (AUC) values indicates the value for TyG-related indicators in predicting ED.

Conclusions

The research uncovered a significant and enduring connection between obesity measures related to the TyG-related index and ED, indicating a robust association of increased insulin resistance with the probability of ED among the U.S. population.

eating disorders

TyG-associated obesity index

TyG index

NHANES

insulin resistance

There is a pressing need for increased awareness of eating disorders (ED) worldwide, encompassing conditions of bulimia nervosa (BN), binge-eating disorder (BED), and anorexia nervosa (AN). Those are defined by the emergence of irregular eating patterns, disruptions in body weight regulation, and excessive preoccupation with weight [1]. ED disrupt an individual's social interactions and psychological health [2]. ED exert significant impacts on the entire spectrum of bodily systems [3] which are linked to the highest rates specific to their causes among all mental health disorders [4]. The development of ED stems from a complex interplay of numerous factors [5]. Athif I et al. [6] suggested AN is linked to heightened insulin sensitivity, whereas BN and BED are connected to reduced insulin sensitivity. A notably higher proportion of individuals in the binge-drinking cohort exhibited insulin resistance (IR). Akshay K et al. [7] concluded that IR had a positive correlation with BED while it showed negative correlation with AN. Hence, a profound comprehension of the relationship between IR and ED is essential for effectively managing the associated conditions.

Considering the intricacies of conventional IR index evaluations, the triglyceride glucose (TyG) index has been introduced, which gauges insulin sensitivity by utilizing levels of triglycerides (TG) and fasting blood glucose (FBG) [8, 9]. Concurrently, a variety of novel indices have been suggested, integrating anthropometric factors associated with obesity to enhance the precision of insulin IR assessment. For example, body mass index (BMI) is a standardized measure for assessing overall adiposity and health status, while waist circumference (WC) serves as a frequently employed metric for gauging central obesity and its metabolic risks [10, 11]. Considering the robust association between a high TyG index and an enhanced likelihood of ED, it has been theorized that a new composite index, which includes TyG along with obesity metrics, could be a valuable tool for predicting ED risk in population with IR. Specifically, the composite indicators TyG-WC and TyG-BMI demonstrated a pronounced relationship with metabolic syndrome like diabetes, and other conditions, providing a more precise reflection of disease risk than TyG alone [12–14].

Nevertheless, there is a scarcity of studies exploring the connection between these indices and ED specifically within the American demographic, underscoring the urgent requirement for further investigation to substantiate these associations. Consequently, the research delved into the National Health and Nutrition Examination Survey (NHANES) database to scrutinize a range of TyG-associated indicators, aiming to uncover the association of insulin resistance related to obesity with the propensity for ED.

Research methodology and participant selection

This research retrieved information on 70,190 participants from the NHANES spanning the years 2005–2018. NHANES applied a multi-stage stratified probability sampling technique to gather comprehensive information regarding health conditions, lifestyle, and nutritional status of population in America [15]. The complete dataset was rigorously filtered according to exclusionary parameters which involved: (1) aged < 20 years, aged > 60 years; (2) the presence of missing data on ED; (3) incomplete information of TyG, TyG-BMI and TyG-WC and (4) invalid questionnaires. Ultimately, the research involved a total of 10,324 participants. The overall process of selecting participant is exhibited in Fig. 1.

Definition and computation of obesity indices associated with TyG

The TyG index assesses IR by integrating FBG and TG levels. These indicators’ measurements were taken at the outset when collected their blood samples at the first time. Concurrently, body weight, height, and WC were recorded after physical examinations conducted at a mobile examination center. The TyG, TyG-WC as well as TyG-BMI indicators, were derived using the subsequent calculations: (1) TyG = ln [triglycerides (mg/dl) × glucose (mg/dl)/2]; (2) BMI = body mass (kg)/height²(m²); (3) TyG-WC = TyG×waist circumference; TyG-BMI = TyG×BMI.

Eating Status Questionnaires

The eating status was evaluated through NHANES questionnaires administered between 2005–2006 and 2017–2018, which inquired about the frequency of experiencing ED symptoms over the preceding half months. Their feedbacks were categorized as "not at all," "several days," "more than half the days," and "nearly every day," with scores allocated from 0 to 3.

Assessment of covariates

Age and poverty income ratio (PIR) was deemed as a continuous variable. According to gender, participants were grouped into males and females. Race/ethnicity involved in Mexican and non-Mexican Hispanics, non-Hispanic whites and blacks, along with others. They were divided into three groups by education level, including less than high school, high school graduate or equivalent as well as high school graduate or above. Marital status was divided into three groups, including married, widowed/divorced or separated and living with partner. Alcohol consumption was defined as how many alcoholic assumption per day in the past year. Weight, height and WC were assessed when attending a physical examination as continuous type variables. Besides, FBG, insulin, TG, high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) were tested at baseline when providing blood samples as continuous variables. Diabetes, hypertension, sleep trouble, congestive heart failure, tumor or malignancy and liver status were assessed (Doctor informed you had any condition yes or no). Furthermore, we incorporated consideration for weight and attempt to lose weight from weight history questionnaires

Statistical analyses

All data from the NHANES 2005–2006 to 2017–2018 was handled through Empower Stats (version 2.0), along with Microsoft Office Excel. Descriptive statistics were represented as mean ± standard deviation (SD) for continuous variables, and as the count (n) and percentage (%) for categorical variables. Taking into account the complex multistage probability sampling methodology used in the NHANES, the study paid careful attention to underlying impacts of sample weights, stratification, and clustering within the dataset. As a result, the research categorized the TyG index into quartiles and utilized a weighted multivariable logistic regression model to examine the relationships between TyG-associated obesity indicators and ED. To conduct a thorough analysis, we formulated three models into adjustment. Model 1 served as an unadjusted; model 2 was was adjusted for basic demographic variables concerning age, gender, and race; and model 3 subjected to adjustments for multiple possible confounders: age, gender, race, education level, marital status, PIR, alcohol consumption, diabetes, hypertension, sleep trouble, congestive heart failure, liver condition, cancer or malignancy, consideration for weight, attempt to lose weight, WC, BMI, HDL-C, LDL-C, insulin, fasting blood glucose, and fasting triglycerides.

Baseline information

The fundamental features of participants are detailed in Table 1. We classified never experiencing disordered eating and occasional ED (several days) as non-severe conditions. We classified frequent ED (more than half the days) and almost daily ED as severe conditions. The 10,324 participants exhibited an overall severe ED prevalence of 9.2% in the cohort. It comprised 48.534% males and 51.466% females in gender distribution. It included 18.965% Mexican Americans, 36.938% non-Hispanic whites, 19.366% non-Hispanic blacks, 11.428% with other Hispanic backgrounds, and 13.302% of other ethnicities. Compared to the non-severe ED group, the severe ED group presented with numerous distinct characteristics: higher BMI and WC, greater alcohol intake, higher rates of diabetes, sleep disorders, hypertension, congestive heart failure, and liver conditions, and higher levels of fasting triglycerides, fasting glucose, fasting insulin, TyG, TyG-BMI, as well as TyG-WC. Additionally, individuals in this group were more likely to perceive themselves as overweight or underweight and made more efforts to lose weight.

Table 1

Initial demographic and clinical information of the cohort
Variety	Non-severe (n = 9374)	Severe (n = 950)	P-value
Age (years)	40.294 ± 11.735	40.672 ± 11.682	0.37026
Gender (%)			0.84201
Male	48.175	48.534
Female	51.825	51.466
Race (%)			0.44684
Mexican American	17.360	18.965
Other Hispanic	10.118	11.428
Non-Hispanic White	38.874	36.938
Non-Hispanic Black	20.695	19.366
Other race	12.682	13.302
Educational level (%)			0.86185
Less than high school	22.468	23.406
High school graduate or equivalent	21.636	21.123
More than high school	55.716	55.470
Marital status (%)			0.34750
Married	49.859	52.376
Widowed/divorced or separated	38.967	37.432
Living with partner	11.175	10.192
PIR (%)	2.443 ± 1.570	2.434 ± 1.613	0.87936
Alcohol consumption (%)			0.00018
Light	23.626	19.131
Moderate	45.016	43.104
Heavy	31.358	37.765
Diabetes (%)			< 0.00001
Yes	5.881	9.236
No	92.535	87.622
Borderline	1.585	3.142
Hypertension (%)			< 0.00001
Yes	23.387	33.017
No	76.613	66.983
Sleep trouble (%)			< 0.00001
Yes	24.052	50.157
No	75.948	49.843
Congestive heart failure (%)			< 0.00001
Yes	0.820	2.581
No	99.180	97.419
Liver disease (%)			< 0.00001
Yes	3.362	6.881
No	96.638	93.119
Cancer or malignancy (%)			0.05865
Yes	4.946	6.440
No	95.054	93.560
Consideration for weight (%)			< 0.00001
Overweight	56.079	66.912
Underweight	4.186	9.033
About the right weight	39.735	24.055
Attempt to lose weight (%)			0.00028
Yes	35.285	41.563
No	64.715	58.437
HDL-C (mg/dL)	53.742 ± 15.861	51.302 ± 15.057	0.00002
LDL-C (mg/dL)	115.390 ± 33.963	114.010 ± 36.467	0.26210
Insulin (uU/mL)	12.339 ± 14.327	15.772 ± 16.600	< 0.00001
TG (mg/dL)	125.203 ± 112.786	139.177 ± 124.049	0.00065
FBG (mg/dL)	103.469 ± 29.137	107.445 ± 37.131	0.00021
BMI (kg/cm2) (%)			< 0.00001
< 25	34.853	25.683
25–30	33.683	29.709
> 30	31.464	44.609
WC (cm)	97.913 ± 16.746	102.437 ± 18.336	< 0.00001
TyG	8.551 ± 0.676	8.672 ± 0.721	< 0.00001
TyG-BMI	247.439 ± 67.788	271.075 ± 78.756	< 0.00001
TyG-WC	841.673 ± 181.261	893.545 ± 199.530	< 0.00001

Continuous variables were listed as Mean ± Standard deviation (SD), the P-value was derived using a weighted linear regression model

Categorical variables were listed as %, the P-value was derived using a weighted chi-square test

Multifactor regression analysis

To examine the regression equations for TyG index and tooth loss, we used a multifactor regression analysis and adjusted three models. Associations between TyG-BMI and ED were listed in Table 2. TyG [β (95% confidence intervals (CI)] = 0.031 (0.005, 0.056), TyG-BMI [β (95% CI] = 0.001 (0.000, 0.002) and TyG-WC [β (95% CI] = 0.000 (0.000, 0.001) indicated a positive correlation with ED in the adjusted model. Concurrently, we employed one-way analysis of variance for trend analysis. The findings indicated a consistent trend across all three models for the TyG-BMI quartiles, suggesting that an enhanced TyG-BMI is correlated with a heightened ED risk.

Table 2

Associations between TyG-BMI and ED
TyG-BMI (continuous)	0.002 (0.001, 0.002) < 0.00001	0.002 (0.001, 0.002) < 0.00001	0.001 (0.000, 0.002) 0.00471
TyG-BMI (quartile)
Q1	Reference	Reference	Reference
Q2	0.028 (-0.014, 0.070) 0.19313	0.028 (-0.014, 0.070) 0.19134	0.019 (-0.030, 0.067) 0.44972
Q3	0.089 (0.047, 0.131) 0.00003	0.089 (0.047, 0.131) 0.00003	0.034 (-0.026, 0.095) 0.26941
Q4	0.266 (0.224, 0.308) < 0.00001	0.265 (0.223, 0.307) < 0.00001	0.086 (0.000, 0.172) 0.04965
P for trend	< 0.001	< 0.001	< 0.001

Model 1: No covariates were adjusted; Model 2: Adjusted for gender, age and race; Model 3: Adjusted for all variables: gender, age, race, education level, PIR, marital status, waist, alcohol consumption status, hypertension status, diabetes status, liver condition, congestive heart failure, cancer or malignancy, sleep disorders, HDL, LDL, insulin, consideration for weight and attempt to lose weight.

Model 1 sample size: 10324; Model 2 sample size: 10324; Model 3 sample size: 10324

Identification of nonlinear associations

Given that prior multivariate analysis revealed a non-linear relationship between the baseline TyG-related index and ED, we utilized a smooth curve fitting technique (penalized spline method) to delve deeper into this correlation which were showed in Fig. 2. Saturation effect analysis of TyG, TyG-BMI and TyG-WC on ED were demonstrated in Table 3. In model 1, after adjusting for age, gender, race, BMI, alcohol consumption, education level, hypertension, and other covariates, the morbidity of ED rose by 0.002 (95% Cl = 0.001, 0.003; p < 0.001) with each unit increase in the TyG-BMI index respectively. In model 2, the inflection point was determined as 171.211 (P values for log-likelihood ratio = 0.039). the baseline TyG-BMI index was significantly and positively associated with the ED when TyG-BMI index exceeded 171.211 (β = 0.002; 95% CI = 0.001, 0.003), while TyG-BMI index was negatively associated with ED when TyG-BMI index was above 171.211 (β = -0.003; 95% CI = -0.006, -0.000).

Table 3

Threshold effect analysis of TyG-related index on ED
	OR (95% Cl) P-value
Model 1
Total	0.028 (0.002,0.054) 0.0332
Model 2
Inflection point	8.259
TyG index < 8.259	0.011 (-0.056, 0.078) 0.7493
TyG index > 8.259	0.034 (0.001, 0.067) 0.0422
P for Log-likelihood ratio	0.575
Model 1
Total	0.002 (0.001, 0.003) < 0.0001
Model 2
Inflection point	171.211
TyG-BMI index < 171.211	-0.003 (-0.006, -0.000) 0.0276
TyG-BMI index > 171.211	0.002 (0.001, 0.003) < 0.0001
P for Log-likelihood ratio	< 0.001
Model 1
Total	-0.000 (-0.000, 0.000) 0.1503
Model 2
Inflection point	589.442
TyG-WC index < 589.442	-0.002 (-0.004, -0.001) 0.0014
TyG-WC index > 589.442	-0.000 (-0.000, 0.000) 0.3703
P for Log-likelihood ratio	0.002

Subgroup analyses

The study conducted an in-depth examination of the intricate relationships between the TyG-related index and periodontitis through subgroup analyses accounting for gender, race, marital status, educational level, alcohol consumption, diabetes, sleep trouble, hypertension, congestive heart failure, cancer or malignancy, and liver condition. The results of subgroup analyses were listed in Table 4. A noteworthy correlation between the TyG and TyG-BMI index and ED was identified among participants suffering from hypertension.

Table 4

Subgroup analysis between TyG-related index and ED
Subgroup	β (95%CI) ^*	P for interaction ^*	β (95%CI) ^**	P for interaction **	β (95%CI) ^***	P for interaction ^***
Gender		0.7853		0.7853		0.0475
Male	0.025 (-0.012, 0.062)		0.025 (-0.012, 0.062)		-0.000 (-0.000, 0.000)
Female	0.032 (-0.004, 0.068)		0.032 (-0.004, 0.068)		-0.000 (-0.000, 0.000)
Race		0.4207		0.4207		0.1431
Mexican American	0.042 (-0.018, 0.102)		0.042 (-0.018, 0.102)		-0.000 (-0.000, 0.000)
Other Hispanic	-0.026 (-0.108, 0.056)		-0.026 (-0.108, 0.056)		-0.000 (-0.001, 0.000)
Non-Hispanic White	0.051 (-0.108, 0.056)		0.051 (-0.108, 0.056)		-0.000 (-0.000, 0.000)
Non-Hispanic Black	0.001 (-0.057, 0.058)		0.001 (-0.057, 0.058)		-0.000 (-0.000, 0.000)
Other race	0.019 (-0.056, 0.095)		0.019 (-0.056, 0.095)		-0.001 (-0.001, -0.000)
Marital status		0.4059		0.4059		0.2772
Married	0.014 (-0.023, 0.050)		0.014 (-0.023, 0.050)		-0.000 (-0.000, -0.000)
Widowed/divorced or separated	0.031 (-0.009, 0.072)		0.031 (-0.009, 0.072)		-0.000 (-0.000, 0.000)
Living with partner	0.070 (-0.007, 0.148)		0.070 (-0.007, 0.148)		0.000 (-0.000,0.001)
Education level		0.4422		0.4422		0.3923
Less than high school	0.055 (0.003, 0.108)		0.055 (0.003, 0.108)		0.000 (-0.000,0.000)
High school graduate or equivalent	0.022 (-0.033, 0.077)		0.022 (-0.033, 0.077)		-0.000 (-0.000, 0.000)
More than high school	0.015 (-0.021, 0.051)		0.015 (-0.021, 0.051)		-0.000 (-0.000, -0.000)
Alcohol consumption		0.6831		0.6831		0.7924
Light	0.046 (-0.015, 0.107)		0.046 (-0.015, 0.107)		-0.000 (-0.000, 0.000)
Moderate	0.021 (-0.017, 0.058)		0.021 (-0.017, 0.058)		-0.000 (-0.000, 0.000)
Heavy	0.013 (-0.033, 0.058)		0.013 (-0.033, 0.058)		-0.000 (-0.000, 0.000)
Diabetes		0.1339		0.1339		0.1996
Yes	0.093 (0.024, 0.162)		0.093 (0.024, 0.162)		0.000 (-0.000, 0.001)
No	0.018 (-0.011, 0.047)		0.018(-0.011, 0.047)		-0.000 (-0.000, -0.000)
Borderline	-0.004 (-0.184, 0.177)		-0.004 (-0.184, 0.177)		0.000 (-0.001, 0.001)
Hypertension		0.0442		0.0442		0.0530
Yes	0.068 (0.021, 0.115)		0.068 (0.021, 0.115)		0.000 (-0.000, 0.000)
No	0.013 (-0.017, 0.043)		0.013 (-0.017, 0.043)		-0.000 (-0.000, -0.000)
Sleep trouble		0.8296		0.8296		0.1677
Yes	0.033 (-0.017, 0.084)		0.033 (-0.017, 0.084)		0.000 (-0.000, 0.000)
No	0.027 (-0.002, 0.056)		0.027 (-0.002, 0.056)		-0.000 (-0.000, 0.000)
Congestive heart failure		0.4022		0.4022		0.3006
Yes	0.117 (-0.095, 0.329)		0.117 (-0.095, 0.329)		-0.001 (-0.002, 0.000)
No	0.026 (-0.000, 0.052)		0.026 (-0.000, 0.052)		-0.000 (-0.000, 0.000)
Liver condition		0.0604		0.0604		0.0866
Yes	0.152 (0.021, 0.284)		0.152 (0.021, 0.284)		0.001 (-0.000, 0.001)
No	0.024 (-0.003, 0.051)		0.024 (-0.003, 0.051)		-0.000 (-0.000, -0.000)
Cancer or malignancy		0.3325		0.3325		0.6929
Yes	0.094 (-0.042, 0.230)		0.094 (-0.042, 0.230)		-0.000 (-0.001, 0.001)
No	0.026 (-0.000, 0.053)		0.026 (-0.000, 0.053)		-0.000 (-0.000, 0.000)
^* TyG; ^ TyG-BMI; * TyG-WC

Sensitivity and specificity analysis

ROC analysis was conducted to evaluate the prognostic value of TyG-related index for ED. The results of the ROC curves were shown in Table 5 and Fig. 3. The areas under the curve (AUC) of TyG index, TyG-BMI index and TyG-WC index in predicting ED were 0.550 (95%CI: 0.524, 0.577), 0.587 (95%CI: 0.559, 0.616) and 0.582 (95%CI: 0.555, 0.610), respectively. Since the AUC of TyG-BMI was the largest, its value of predicting ED seemed greater than TyG and TyG-WC index. (P > 0.05).

Table 5

Model predictions of association between TyG-related index and ED
Objects	Cutoff (Sensitivity, Specificity)	AUC (95%CI)
TyG	0.050 (0.287, 0.798)	0.550 (0.524, 0.577)
TyG-BMI	0.046 (0.496, 0.661)	0.587 (0.559, 0.616)
TyG-WC	0.044 (0.571, 0.569)	0.582 (0.555, 0.610)

The predominant discoveries of this study revealed that TyG-related index was positively correlated with the occurrence of ED in the full adjustment model. Furthermore, an increased TyG-BMI index is associated with a heightened risk of ED. Subgroup analyses showed that strong relationship between TyG and TyG-BMI index and ED was more likely to be observed among participants with hypertension. Threshold effect analysis indicated that higher TyG-BMI was significantly associated with a higher risk of ED. Results of ROC analysis indicated that the TyG-related index had valid predictive value for ED.

Previous studies have indicated that metabolic syndrome associated with insulin resistance (IR) is notably influential in in increasing the risk of ED [16]. The homeostasis model assessment of insulin resistance (HOMA-IR), is a traditional indicator of IR [17]. Nevertheless, considering the expense of insulin testing and the constraints of HOMA-IR, there is a need for a more effective, and user-friendly metric. The TyG index offers equivalent or superior results for assessing IR, owing to its ease of computation, sensitivity, and specificity [18–20].

Previous studies have validated the effectiveness of WC and BMI in predicting obesity [21]. A nationwide cohort study indicated that the combination of obesity indicators with the TyG index could more accurately predict the risk of metabolic syndrome than the use of these metrics alone [22]. Furthermore, TyG-WC, and TyG-BMI demonstrated greater strength in predicting IR than TyG alone [23]. Consequently, the integration of obesity metrics with TyG might more precisely capture the relationship between IR and ED. Although the precise physiological processes remain unclear, several elements could account for this link.

Flint et al. [24] deduced that the insulin response following a meal could serve as a significant signal for satiety, and that central nervous system IR in overweight individuals might be responsible for the diminished impact on appetite control. The homeostatic approach to appetite regulation involves feedback which is attributed to the inhibitory action of insulin. Gastrointestinal hormones, such as cholecystokinin, and glucagon-like peptide-1 initiate a series of neural and hormonal signals that operate both peripherally and centrally to facilitate satiation [25]. These hormones collaborate with the appetite-stimulating hormone ghrelin to manage food intake in a cyclical manner. Insulin provides a steady influence on appetite by fine-tuning the intensity of the cyclical signals [25]. Peripheral neuroendocrine signals converge in the hypothalamus. Upon transmission to the arcuate nucleus through the nucleus tractus solitarius, these signals activate the orexigenic neurons that express agouti-related peptide (AgRP) and neuropeptide Y (NPY), as well as the anorexigenic pro-opiomelanocortin (POMC) neurons that release alpha-melanocyte-stimulating hormone [26]. The equilibrium in the release of these functionally opposing neuropeptides confers the satiation [27]. The ghrelin hormone activates the orexigenic NPY/AgRP neurons in the arcuate nucleus, while glucose and insulin have an inhibitory effect on them. Conversely, the POMC/CART neurons are activated by insulin and glucose, but their activity is suppressed when NPY/AgRP neurons are stimulated [27].

While glucose and insulin regulate appetite together, there may be specific conditions where glucose homeostasis has a greater impact on eating behavior. The glucostatic theory of appetite, proposed by Mayer [28], proposed that alterations in the utilization of glucose were crucial for encoding feelings of hunger, with "metabolic hypoglycemia" serving as a trigger for starting a meal. In support of this, some studies assessed the link among glycemic load, glycemic responses and appetite [29] and suggest that lower glycemic loads and glycemic responses leads to reduced post-meal appetite. Hypoglycemia, common in those with diabetes [30], is met with strong neuroendocrine reactions aimed at restoring euglycemia [31]. Hypoglycemia boosts the activity of hypothalamic NPY/AgRP neurons and diminishes that of POMC/CART neurons, leading to overeating [32].

Beyond its peripheral effects, insulin also centrally modulates neuronal activity in brain areas associated with eating behavior, sensory processing, and reward [33]. It is hypothesized that brain IR coincides with peripheral resistance in areas for appetite [34]. Vagal fibers that innervate the gastrointestinal tract are pivotal to the gut-brain axis since fibers facilitates communication between the hypothalamus and peripheral organs, such as the liver and adipose tissue [35]. However, the expression of vagal gastrointestinal hormone receptors and sensitivity to peripheral stimuli might be diminished in individuals with obesity and diabetes [36], which implies that an excessive fat, commonly observed in type 2 diabetes (T2D), could interfere with the vagal perception of peripheral appetite signals and their recognition in brain areas critical to ED. The insulin-reward pathways could be compromised in T2D, given that peripheral IR is linked to IR in brain areas that govern the rewarding aspects of food [34]. Additionally, hippocampal damage has been demonstrated to lead to ED, such as a diminished ability to sense hunger and fullness and a reduced mental representation of food [37].

This research contributes significant findings that highlight the positive associations between TyG-related index and ED. The study benefits from a substantial sample size and nationwide coverage. Additionally, the analysis controlled for various confounding factors associated with ED, which strengthens the reliability of the outcomes. However, there are limitations to acknowledge. First, the study limits the ability to determine causality between TyG-related index and ED. Second, it is difficult to completely eliminate the potential impact of unmeasured confounders. Lastly, the study's focus on the US population may restrict the applicability of the results to a global setting. Future studies should aim to include larger and more diverse samples to identify the most predictive TyG-related indices for ED and to establish universally applicable threshold values.

The research confirmed a significant association between the TyG-related index and ED in the U.S. population. This finding highlights the potential of TyG-related obesity indices for assessing the risk of ED and accentuates the importance of IR in the development of ED. Regular monitoring of the TyG-related index could enable clinicians to pinpoint individuals at a higher risk for ED, facilitating the application of targeted preventative measures. However, the precise molecular mechanisms that connect the TyG-related index with ED remain to be fully understood. Future studies should aim to explore the possible biological pathways that associate the TyG-related index with ED and to determine whether interventions that lower the TyG-related index could lead to a reduction in the prevalence of ED.

Ethics approval and consent to participate

Not applicable.

Competing interests

The authors declare no competing interests.

Funding

This work was supported by the National Natural Science Foundation of China (82370601).

Author Contribution

H.X. was responsible for the conception and writing. Y.D. contributed to making figures and tables. K.S. performed the investigation. Y.R. reviewed and edited this manuscript. All the authors reviewed the manuscript and agreed to publish it.

Acknowledgments

Not applicable.

Availability of data and material

All data for this study can be found on the NHANES website.

Himmerich H, Keeler JL, Davies HL, Tessema SA, Treasure J. The evolving profile of eating disorders and their treatment in a changing and globalised world. Lancet. 2024;403:2671–5. 10.1016/s0140-6736(24)00874-2.
Santomauro DF, Melen S, Mitchison D, Vos T, Whiteford H, Ferrari AJ. The hidden burden of eating disorders: an extension of estimates from the Global Burden of Disease Study 2019. Lancet Psychiatry. 2021;8:320–8. 10.1016/s2215-0366(21)00040-7.
Wu J, Liu J, Li S, Ma H, Wang Y. Trends in the prevalence and disability-adjusted life years of eating disorders from 1990 to 2017: results from the Global Burden of Disease Study 2017. Epidemiol Psychiatr Sci. 2020;29:e191. 10.1017/s2045796020001055.
Global regional. and national burden of 12 mental disorders in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet Psychiatry 2022, 9:137–150.10.1016/s2215-0366(21)00395-3
Culbert KM, Racine SE, Klump KL. Research Review: What we have learned about the causes of eating disorders - a synthesis of sociocultural, psychological, and biological research. J Child Psychol Psychiatry. 2015;56:1141–64. 10.1111/jcpp.12441.
Ilyas A, Hübel C, Stahl D, Stadler M, Ismail K, Breen G, Treasure J, Kan C. The metabolic underpinning of eating disorders: A systematic review and meta-analysis of insulin sensitivity. Mol Cell Endocrinol. 2019;497:110307. 10.1016/j.mce.2018.10.005.
Kumar A, Alam S, Bano S, Prakash R, Jain V. Association of eating disorders with glycaemic control and insulin resistance in patients of type 2 diabetes mellitus. Int J Biochem Mol Biol. 2023;14:40–50.
Lee J, Kim B, Kim W, Ahn C, Choi HY, Kim JG, Kim J, Shin H, Kang JG, Moon S. Lipid indices as simple and clinically useful surrogate markers for insulin resistance in the U.S. population. Sci Rep. 2021;11:2366. 10.1038/s41598-021-82053-2.
Son DH, Lee HS, Lee YJ, Lee JH, Han JH. Comparison of triglyceride-glucose index and HOMA-IR for predicting prevalence and incidence of metabolic syndrome. Nutr Metab Cardiovasc Dis. 2022;32:596–604. 10.1016/j.numecd.2021.11.017.
Moosaie F, Abhari SMF, Deravi N, Behnagh AK, Esteghamati S, Firouzabadi FD, Rabizadeh S, Nakhjavani M, Esteghamati A. Waist-To-Height Ratio Is a More Accurate Tool for Predicting Hypertension Than Waist-To-Hip Circumference and BMI in Patients With Type 2 Diabetes: A Prospective Study. Front Public Health. 2021;9. 10.DOI:10.3389/fpubh.2021.726288.
Katzmarzyk PT, Hu G, Cefalu WT, Mire E, Bouchard C. The Importance of Waist Circumference and BMI for Mortality Risk in Diabetic Adults. Diabetes Care. 2013;36:3128–30. 10.2337/dc13-0219.
Li X, Sun M, Yang Y, Yao N, Yan S, Wang L, Hu W, Guo R, Wang Y, Li B. Predictive Effect of Triglyceride Glucose-Related Parameters, Obesity Indices, and Lipid Ratios for Diabetes in a Chinese Population: A Prospective Cohort Study. Front Endocrinol (Lausanne). 2022;13:862919. 10.3389/fendo.2022.862919.
Raimi TH, Dele-Ojo BF, Dada SA, Fadare JO, Ajayi DD, Ajayi EA, Ajayi OA. Triglyceride-Glucose Index and Related Parameters Predicted Metabolic Syndrome in Nigerians. Metab Syndr Relat Disord. 2021;19:76–82. 10.1089/met.2020.0092.
Yan S, Wang D, Jia Y. Comparison of insulin resistance-associated parameters in US adults: a cross-sectional study. Horm (Athens). 2023;22:331–41. 10.1007/s42000-023-00448-4.
Ghimire S, Baral BK, Feng D, Sy FS, Rodriguez R. Is selenium intake associated with the presence of depressive symptoms among US adults? Findings from National Health and Nutrition Examination Survey (NHANES) 2011–2014. Nutrition. 2019;62:169–76. 10.1016/j.nut.2018.12.007.
Berthoud HR. Mind versus metabolism in the control of food intake and energy balance. Physiol Behav. 2004;81:781–93. 10.1016/j.physbeh.2004.04.034.
Tahapary DL, Pratisthita LB, Fitri NA, Marcella C, Wafa S, Kurniawan F, Rizka A, Tarigan TJE, Harbuwono DS, Purnamasari D, Soewondo P. Challenges in the diagnosis of insulin resistance: Focusing on the role of HOMA-IR and Tryglyceride/glucose index. Diabetes Metab Syndr. 2022;16(102581). 10.1016/j.dsx.2022.102581.
Guerrero-Romero F, Villalobos-Molina R, Jiménez-Flores JR, Simental-Mendia LE, Méndez-Cruz R, Murguía-Romero M, Rodríguez-Morán M. Fasting Triglycerides and Glucose Index as a Diagnostic Test for Insulin Resistance in Young Adults. Arch Med Res. 2016;47:382–7. 10.1016/j.arcmed.2016.08.012.
Simental-Mendía LE, Rodríguez-Morán M, Guerrero-Romero F. The product of fasting glucose and triglycerides as surrogate for identifying insulin resistance in apparently healthy subjects. Metab Syndr Relat Disord. 2008;6:299–304. 10.1089/met.2008.0034.
Wang S, Shi J, Peng Y, Fang Q, Mu Q, Gu W, Hong J, Zhang Y, Wang W. Stronger association of triglyceride glucose index than the HOMA-IR with arterial stiffness in patients with type 2 diabetes: a real-world single-centre study. Cardiovasc Diabetol. 2021;20(82). 10.1186/s12933-021-01274-x.
Park Y, Kim NH, Kwon TY, Kim SG. A novel adiposity index as an integrated predictor of cardiometabolic disease morbidity and mortality. Sci Rep. 2018;8:16753. 10.1038/s41598-018-35073-4.
Gui J, Li Y, Liu H, Guo LL, Li J, Lei Y, Li X, Sun L, Yang L, Yuan T, et al. Obesity- and lipid-related indices as a predictor of obesity metabolic syndrome in a national cohort study. Front Public Health. 2023;11:1073824. 10.3389/fpubh.2023.1073824.
Lim J, Kim J, Koo SH, Kwon GC. Comparison of triglyceride glucose index, and related parameters to predict insulin resistance in Korean adults: An analysis of the 2007–2010 Korean National Health and Nutrition Examination Survey. PLoS ONE. 2019;14:e0212963. 10.1371/journal.pone.0212963.
Flint A, Gregersen NT, Gluud LL, Møller BK, Raben A, Tetens I, Verdich C, Astrup A. Associations between postprandial insulin and blood glucose responses, appetite sensations and energy intake in normal weight and overweight individuals: a meta-analysis of test meal studies. Br J Nutr. 2007;98:17–25. 10.1017/s000711450768297x.
Schwartz MW, Seeley RJ, Zeltser LM, Drewnowski A, Ravussin E, Redman LM, Leibel RL. Obesity Pathogenesis: An Endocrine Society Scientific Statement. Endocr Rev. 2017;38:267–96. 10.1210/er.2017-00111.
Morton GJ, Cummings DE, Baskin DG, Barsh GS, Schwartz MW. Central nervous system control of food intake and body weight. Nature. 2006;443:289–95. 10.1038/nature05026.
López-Gambero AJ, Martínez F, Salazar K, Cifuentes M, Nualart F. Brain Glucose-Sensing Mechanism and Energy Homeostasis. Mol Neurobiol. 2019;56:769–96. 10.1007/s12035-018-1099-4.
Mayer J. Glucostatic mechanism of regulation of food intake. N Engl J Med. 1953;249:13–6. 10.1056/nejm195307022490104.
Gonzalez-Anton C, Lopez-Millan B, Rico MC, Sanchez-Rodriguez E, Ruiz-Lopez MD, Gil A, Mesa MD. An enriched, cereal-based bread affects appetite ratings and glycemic, insulinemic, and gastrointestinal hormone responses in healthy adults in a randomized, controlled trial. J Nutr. 2015;145:231–8. 10.3945/jn.114.200386.
Leese GP, Wang J, Broomhall J, Kelly P, Marsden A, Morrison W, Frier BM, Morris AD. Frequency of severe hypoglycemia requiring emergency treatment in type 1 and type 2 diabetes: a population-based study of health service resource use. Diabetes Care. 2003;26:1176–80. 10.2337/diacare.26.4.1176.
Watts AG, Donovan CM. Sweet talk in the brain: glucosensing, neural networks, and hypoglycemic counterregulation. Front Neuroendocrinol. 2010;31:32–43. 10.1016/j.yfrne.2009.10.006.
Sindelar DK, Ste Marie L, Miura GI, Palmiter RD, McMinn JE, Morton GJ, Schwartz MW. Neuropeptide Y is required for hyperphagic feeding in response to neuroglucopenia. Endocrinology. 2004;145:3363–8. 10.1210/en.2003-1727.
Arnold SE, Arvanitakis Z, Macauley-Rambach SL, Koenig AM, Wang HY, Ahima RS, Craft S, Gandy S, Buettner C, Stoeckel LE, et al. Brain insulin resistance in type 2 diabetes and Alzheimer disease: concepts and conundrums. Nat Rev Neurol. 2018;14:168–81. 10.1038/nrneurol.2017.185.
Anthony K, Reed LJ, Dunn JT, Bingham E, Hopkins D, Marsden PK, Amiel SA. Attenuation of insulin-evoked responses in brain networks controlling appetite and reward in insulin resistance: the cerebral basis for impaired control of food intake in metabolic syndrome? Diabetes 2006, 55:2986–92.10.2337/db06-0376
Cammisotto P, Bendayan M. A review on gastric leptin: the exocrine secretion of a gastric hormone. Anat Cell Biol. 2012;45:1–16. 10.5115/acb.2012.45.1.1.
de Lartigue G. Role of the vagus nerve in the development and treatment of diet-induced obesity. J Physiol. 2016;594:5791–815. 10.1113/jp271538.
Stevenson RJ, Francis HM. The hippocampus and the regulation of human food intake. Psychol Bull. 2017;143:1011–32. 10.1037/bul0000109.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Associations between different triglyceride glucose index-related obesity indices and eating disorders: results from NHANES 2005–2018

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Introduction

Methods

Research methodology and participant selection

Definition and computation of obesity indices associated with TyG

Eating Status Questionnaires

Assessment of covariates

Statistical analyses

Results

Baseline information

Multifactor regression analysis

Identification of nonlinear associations

Subgroup analyses

Sensitivity and specificity analysis

Discussion

Conclusion

Declarations

Competing interests

Funding

Author Contribution

Acknowledgments

Availability of data and material

References

Additional Declarations

Status:

Version 1