The Circulating Selenium Concentration Is Positively Related to the Lipid Accumulation Product: A Population-Based Cross-Sectional Study

The lipid accumulation product (LAP) is a reliable marker of metabolic syndrome, which includes conditions like obesity. However, the correlation between the circulating selenium (CSe) concentration and the LAP is currently unclear. This study aimed to ascertain this correlation. Overall, 12,815 adults aged ≥20 years were enrolled in this study. After adjusting for all the confounding variables, CSe was positively correlated to the LAP (β = 0.41; 95% confidence interval [CI]: 0.28, 0.54; p < 0.001). Compared with the lowest quartile of CSe, the highest quartile of CSe was positively related to the LAP (β = 0.16; 95% CI: 0.12, 0.21; p < 0.001). Moreover, the correlation between CSe and the LAP revealed a positive non-linear trend. In the subgroup analysis, interaction effects were observed for age, sex, smoking, and stroke (p for interaction < 0.05). The effects were stronger for males (β = 0.64, 95% CI: 0.47, 0.80; p < 0.001) and individuals who smoke at the time of the trial (β = 0.64, 95% CI: 0.37, 0.91; p < 0.001). In conclusion, our results indicated that CSe was positively correlated with the LAP in a non-linear manner. Future research is warranted to explore their relationship and better understand the mechanisms underlying this association.


Introduction
Obesity is a global problem [1,2], contributing to the incidence and mortality of a range of diseases, including hypertension, coronary heart disease, type 2 diabetes mellitus (T2DM), dyslipidemia, cerebrovascular accidents, and cancer [3][4][5][6].It also results in a substantial increase in healthcare costs.However, to date, no country has implemented a successful public health model to reduce the prevalence of obesity, despite continued efforts to do so [7].In light of this, tackling obesity has become a global health priority.
Nutrients 2024, 16, 933 2 of 13 Thus, it is hypothesized that Se may influence obesity levels.Notably, a previous metaanalysis, including 65 articles, highlighted that the relationship between Se and being overweight or obese was controversial [26].Furthermore, to date, the association between the circulating selenium (CSe) concentration and LAP has not been explored.Hence, based on the National Health and Nutrition Examination Survey (NHANES), this research aimed to assess the correlation between CSe and the LAP.

Study Population
The NHANES was conducted in order to make a health evaluation for all Americans [27].The NHANES project was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of the National Center for Health Statistics (NCHS) (protocol code: 2011-17, date of approval 2011; protocol code: 2018-01, date of approval 2018).A written notice was submitted to all adult individuals.
The survey utilized open-access data from the NHANES.A secondary analysis in our study was based on the NHANES.The data analyzed in our study combined four survey cycles from the NHANES (2011-2012, 2013-2014, 2015-2016, and 2017-2018).After screening the data of 39,156 participants, 26,341 participants were eliminated due to the following reasons: Age < 20 years (n = 16,539), cancer (n = 2184), pregnancy (n = 245), missing CSe (n = 6423), missing WC (n = 684), and missing TG (n = 266).A total of 12,815 participants were eligible for inclusion and therefore incorporated into the analyses (Figure 1).
function dysfunction, and energy metabolism disorders are key in the pathogenesis of obesity [8,[23][24][25].Thus, it is hypothesized that Se may influence obesity levels.Notably, a previous meta-analysis, including 65 articles, highlighted that the relationship between Se and being overweight or obese was controversial [26].Furthermore, to date, the association between the circulating selenium (CSe) concentration and LAP has not been explored.Hence, based on the National Health and Nutrition Examination Survey (NHANES), this research aimed to assess the correlation between CSe and the LAP.

Study Population
The NHANES was conducted in order to make a health evaluation for all Americans [27]

Acquisition of Variables
Age, sex, race, education, family poverty ratio of income (FPRI), marital status, and pregnancy status were acquired from the demographic data.The BMI, WC, and blood pressure were retrieved from the examination data.A history of various self-reported diseases, medication use status, alcohol consumption, and smoking status were acquired from the questionnaire data.Serum lipid, fasting plasma glucose, hemoglobin A1c (HbA1c), and whole blood Se levels were retrieved from laboratory data.More detailed information is available from the NHANES.

Acquisition of Variables
Age, sex, race, education, family poverty ratio of income (FPRI), marital status, and pregnancy status were acquired from the demographic data.The BMI, WC, and blood pressure were retrieved from the examination data.A history of various self-reported diseases, medication use status, alcohol consumption, and smoking status were acquired from the questionnaire data.Serum lipid, fasting plasma glucose, hemoglobin A1c (HbA1c), and whole blood Se levels were retrieved from laboratory data.More detailed information is available from the NHANES.

Case Definition
Hypertension was defined using the following criteria: Self-reported hypertension; diagnosed by a physician; patients taking antihypertensive medication; an average systolic blood pressure (SBP) of ≥130 mmHg; and/or diastolic blood pressure (DBP) of ≥80 mmHg [28].T2DM was defined as follows: a diagnosis of diabetes mellitus; taking hypoglycemic medications or using insulin; HbA1c ≥ 6.5%; fasting plasma glucose ≥ 7.0 mmol/L; and/or a 2 h plasma glucose ≥ 11.1 mmol/L [29].Stroke or coronary heart disease was defined as a self-reported stroke, or a coronary heart disease as diagnosed by a physician.

Lipid Accumulation Product Index Calculation
The LAP index was calculated from the WC and TG using the following formula [11]:

Statistical Analysis
Regarding the NHANES guidelines, the statistical analysis adopted suitable sampling weights.Categorical variables are presented as numbers (%), whereas continuous variables are presented as medians (interquartile range) for skewed distributions.Differences were calculated via applying the chi-square test for categorical variables and the rank sum test for continuous variables.The potential confounders were explored using univariate linear regression analysis.After adjusting for multiple factors in different models, multivariate linear regression analyses were used to analyze the independent correlation between CSe and the LAP.Covariates with p < 0.05 in univariate analysis were included in the multivariate analysis as adjusting confounders.Three different models were adopted to verify independent correlations according to the guidelines of the STROBE statement.Model A was not adjusted for any variables.Model B was adjusted for age, sex, and race.Model C was adjusted for age, sex, race, education, marital status, alcohol consumption, BMI, hypertension, T2DM, stroke, coronary heart disease, TC, glucose, SBP, DBP, current injection of insulin, and current taking of hypotensive drugs.Dose-response analysis was adopted to test linear or non-linear relationships after adjusting for the same variables in model C. Subgroup analyses were applied to explore the variable interactions.The natural logarithmic (Ln) transformation of CSe and the LAP was carried out due to its non-normal distribution.

Baseline Characteristics
The participants were divided into three groups based on the LAP tertiles (Table 1).When compared with individuals in the low LAP, subjects in the high LAP were significantly more likely to be elderly, male, Mexican American, less educated, poor, married, and obese.They were also more likely to be currently injecting insulin or taking hypotensive drugs, or display symptoms of hypertension, T2DM, stroke, coronary heart disease, and higher levels of SBP, DBP, TC, glucose, and CSe (all p < 0.05).

Univariate Analysis
As shown in Table 2, the results of the univariate linear regression analysis demonstrated that age, alcohol consumption, hypertension, coronary heart disease, T2DM, stroke, glucose, TC, SBP, DBP, BMI, the taking of hypotensive drugs, and the injecting of insulin and Ln CSe were positively related to Ln LAP.In contrast, sex, race, education, and marital status were negatively correlated with Ln LAP (all p < 0.05).The FPRI and current smoking status were not associated with Ln LAP (p > 0.05).

Multivariate Analysis
A multivariate linear regression analysis was carried out to detect the correlation between CSe and the LAP.In model A, which had no adjustments for variables, the correlation between CSe and the LAP was positive (β = 0.76, 95% confidence interval [CI]: 0.56, 0.95; p < 0.001).In model B, which was adjusted for age, sex, and race, the correlation between CSe and the LAP was also positive (β = 0.66; 95% CI: 0.47, 0.84; p < 0.001).In model C, which was adjusted for all significant variables in the univariate analysis, the correlation between CSe and the LAP was also positive (β = 0.41; 95% CI: 0.28, 0.54; p < 0.001).For sensitivity analysis, we also processed CSe as a categorical variable (quartiles), and a similar trend was observed (p for trend < 0.001), as shown in Table 3.

Discussion
As an indispensable trace element in the human body, Se plays an important role in antioxidation, anti-inflammation, anti-aging, energy metabolism regulation, etc. [13][14][15][16][17][18][19][20][21][22].In nature, Se exists in two forms: inorganic Se and organic Se.Se is absorbed by the small intestine, and then distributed into various tissues of the body.After being absorbed by the small intestine, it is divided into various tissues of the body, which are then used to synthesize various selenoproteins that play important biological roles [14].There are twenty-five types of human selenoproteins, all of which are very small, including five glutathione peroxidases (GPx), three thioredoxin reductases, three iodothyronine deiodinases, and others [20].The synthesis of these selenoproteins requires the insertion of a Se-containing homolog of cysteine and 25 coding genes [13,14].GPx1 is the most abundant selenoprotein in mammals, and it is an enzyme that is universally expressed in various cell types.Along with the consumption of reduced glutathione, GPx1 consumes reduced glutathione in order to convert lipid peroxides to their respective alcohols, and to convert H 2 O 2 to water [17].This physiological process is beneficial, as it alleviates oxidative damage to biomolecules such as lipids, lipoproteins, and DNA [17], in addition to maintaining membrane integrity, and reducing the related risks of various diseases [17].Se can also intervene in energy metabolism by activating adipose tissue and regulating thyroid hormones [22].
To date, our analysis is the first to explore the correlation between CSe and the LAP.A positive non-linear correlation between CSe and the LAP was observed in our study.Moreover, the positive relationship between CSe and the LAP was more substantial in participants who were male and currently smoking.Previous literature has shown that the connection between Se levels and obesity is complex and contradictory.Previous observational research has demonstrated that the plasma Se content was negatively related to obesity among children [33].In contrast, a case-control study of 847 adults reported that a high serum Se concentration was related to a high BMI [34].A separate study in women revealed that hair Se levels increased in obese individuals [35].However, a study on French adults reported that the serum Se content was not correlated with the BMI, but rather with serum cholesterol levels [36].Furthermore, a previous NHANES study also reported that the Se dietary intake was unrelated to the BMI and WC [37].Nevertheless, another study revealed a positive correlation between Se dietary intake and obesity in adults [38].
Not only are the results of observational studies inconsistent, but those of interventional studies are also.In animal models, the BMI of obese mice was reduced following the dietary selenomethionine intake, which facilitated the browning reaction [39].However, a randomized prospective survey observed that the BMI was not changed, but there was a significant increase in lean muscle mass and a decrease in leptin levels after 3 months in participants taking oral 240 µg L of selenomethionine per day [40].
In the subgroup analysis, we observed that the connection between CSe and the LAP was influenced by age, sex, current smoking status, and stroke.Serum Se levels were higher in the older group [34].It is known that differences in adipose distribution and proportions between males and females directly affect the evaluation of the LAP.In addition, lifestyle, behavior, and sex hormones differ between males and females [41,42].The gene expression of selenoproteins differs between the sexes [43,44].A Japanese study found a strong connection between the LAP and diabetes mellitus in both sexes [45].The clinical features of diabetes differ between the sexes [46].An American study reported that the whole blood Se concentration was higher in male non-smokers [47].Our previous study found that CSe levels were negatively correlated with stroke [48], meaning that CSe levels were decreased in stroke patients.However, CSe was positively related to the LAP.This contrast amplified the relationship between CSe and the LAP, and made this relationship more significant.
This study has several strengths.First, this is the first analysis of the relationship between CSe and the LAP.Second, we found that CSe and the LAP were positively correlated in a non-linear manner.Third, the sample size of our study was relatively large.
Nevertheless, several additional limitations existed in this research.First, because our analysis was based on an observational survey, we can only draw a correlation, not a causal conclusion.Second, recall biases existed in our study due to some diseases identified based on self-reported diagnostic histories.This is despite the fact that selfreported diagnostic histories were consistent in medical records, particularly for stroke, hypertension, and diabetes mellitus [49].Nonetheless, the individuals enrolled into our analysis were American adults.Thus, there may be inherent population bias, and further investigation is required to generalize our results to other populations.

Conclusions
In conclusion, our results indicated that CSe was positively correlated with the LAP in a non-linear manner.Future investigations are warranted to explore their relationship and better understand the mechanisms underlying this association.Informed Consent Statement: A written notice was submitted to all adult individuals before enrollment.
. The NHANES project was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of the National Center for Health Statistics (NCHS) (protocol code: 2011-17, date of approval 2011; protocol code: 2018-01, date of approval 2018).A written notice was submitted to all adult individuals.The survey utilized open-access data from the NHANES.A secondary analysis in our study was based on the NHANES.The data analyzed in our study combined four survey cycles from the NHANES (2011-2012, 2013-2014, 2015-2016, and 2017-2018).After screening the data of 39,156 participants, 26,341 participants were eliminated due to the following reasons: Age < 20 years (n = 16,539), cancer (n = 2184), pregnancy (n = 245), missing CSe (n = 6423), missing WC (n = 684), and missing TG (n = 266).A total of 12,815 participants were eligible for inclusion and therefore incorporated into the analyses (Figure 1).

Figure 2 .
Figure 2. A positive non-linear manner between CSe and the LAP.A positive non-linear manner was identified after adjusting for multiple confounders in model C. The solid and dashed lines describe the β value and 95% CI, respectively.CI, confidence interval; Ln, natural logarithmic; LAP, lipid accumulation product; CSe, circulating selenium.

Table 4 .
The threshold effect analysis.Inflection Point of Ln CSe (μmol/L) p for log-likelihood ratio test 0.003 CI, confidence interval; Ln, natural logarithmic; CSe, circulating selenium.

Figure 2 .
Figure 2. A positive non-linear manner between CSe and the LAP.A positive non-linear manner was identified after adjusting for multiple confounders in model C. The solid and dashed lines describe the β value and 95% CI, respectively.CI, confidence interval; Ln, natural logarithmic; LAP, lipid accumulation product; CSe, circulating selenium.

Author Contributions:
Conceptualization, Y.Z.; methodology, K.Z. and W.S.; analysis, K.Z.; writingoriginal draft preparation, K.Z.; writing-review and editing, Y.Z. and W.S.; funding acquisition, Y.Z. and W.S. All authors have read and agreed to the published version of the manuscript.Funding: This work was supported by grants from the National Natural Science Foundation of China (82000411, 82030051), the National Key R & D Program of China (2021YFF0501403), and the Key R & D Program of Shandong Province (ZR2020QH023, 2021SFGC0503, 2021ZDSYS05, 2020ZLYS05).Institutional Review Board Statement: The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of the NCHS (protocol code: 2011-17, date of approval 2011; protocol code: 2018-01, date of approval 2018).

Table 1 .
Baseline characteristics of individuals based on the LAP tertiles.

Table 2 .
Results of univariate linear regression analysis of each variable.

Table 3 .
The detection of the independent relationship between Cse and the LAP using multivariate linear regression analysis.
Model A: no adjustment.Model B: adjusted for age, sex, and race.Model C: adjusted for age, sex, race, education, marital status, alcohol consumption, BMI, hypertension, T2DM, stroke, coronary heart disease, TC, glucose, SBP, DBP, current taking of hypotensive drugs, and current injection of insulin.CI, confidence interval; Ln, natural logarithmic; Ref, reference; LAP, lipid accumulation product; CSe, circulating selenium.

Table 5 .
Interaction effects in the subgroup analysis.

Table 4 .
The threshold effect analysis.

Table 5 .
Interaction effects in the subgroup analysis.