An examination of sex and racial/ethnic differences in the metabolic syndrome among adults: a confirmatory factor analysis and a resulting continuous severity score

Abstract

Objective: The metabolic syndrome (MetS) is typically diagnosed based on abnormalities in specific clustered clinical measures that are associated with increased risk for coronary heart disease (CHD) and Type 2 diabetes mellitus (T2DM). However, current MetS criteria result in racial/ethnic discrepancies. Our goals were to use confirmatory factor analysis (CFA) to delineate differential contributions to MetS by sub-group, and if contributions were discovered, develop sex and racial/ethnic-specific equations to calculate MetS severity.

Research design and methods: Using data on adults from the National Health and Nutrition Examination Survey 1999-2010, we performed a CFA of a single MetS factor that allowed differential loadings across groups, resulting in a sex and race/ethnicity-specific continuous MetS severity score.

Results: Loadings to the single MetS factor differed by sub-group for each MetS component (p<0.001), with lower factor loadings among non-Hispanic-blacks for triglycerides and among Hispanics for waist circumference. Systolic blood pressure exhibited low factor loadings among all groups. MetS severity scores were correlated with biomarkers of future disease (high-sensitivity C-reactive-protein, uric acid, insulin resistance). Non-Hispanic-black-males with diabetics had a low prevalence of MetS but high MetS severity scores that were not significantly different from other racial/ethnic groups.

Conclusions: This analysis among adults uniquely demonstrated differences between sexes and racial/ethnic groups regarding contributions of traditional MetS components to an assumed single factor. The resulting equations provide a clinically-accessible and interpretable continuous measure of MetS for potential use in identifying adults at higher risk for MetS-related diseases and following changes within individuals over time. These equations hold potential to be a powerful new outcome for use in MetS-focused research and interventions.

Keywords: AIC; ATP-III; AUC; Adult Treatment Panel III; Akaike’s Information Criteria; Area under the curve; Bentler–Bonett Normed Fit Index; CDC; CFA; CHD; CHF; CVD; Cardiovascular disease; Centers for Disease Control; Clinical studies; Confirmatory Factor Analysis; Congestive heart failure; Coronary heart disease; Epidemiology; GFI; Goodness of Fit Index; HOMA-IR; High-sensitivity C-reactive protein; Hisp; Hispanic; Homeostasis model of insulin resistance; MI; MetS; Metabolic syndrome; Myocardial infarction; NFI; NHANES; NHB; NHW; National Health and Nutrition Examination Survey; Non-Hispanic Black; Non-Hispanic White; Obesity; RMSEA; ROC; Racial/ethnic differences; Receiver operating characteristic; Root Mean Square Error of Approximation; SBP; SRMR; Standardized Root Mean Square Residual; Systolic blood pressure; T2DM; Type 2 diabetes mellitus; WC; Waist circumference; hsCRP.

Figure 1

MetS Severity Score Validation Studies: Age and Associated Serum Risk Markers by Racial/Ethnic Group. Each scatter plot represents Model 2 MetS severity score for each racial/ethnic group for each of the following analyses: A: correlation with cross-sectional age; B: correlation with ln(HOMA); C: correlation with ln(hsCRP); D: correlation with uric acid. See Table 3 for R² values by sex and racial/ethnic group. Abbreviations: high-sensitivity C-reactive protein: hsCRP; HOMA: homeostasis model of insulin resistance.

Figure 2

ATP-III MetS Prevalence and MetS Severity Score Levels Among Participants with Diabetes. Data shown are by sex- and racial/ethnic subgroup among NHANES participants with diabetes: (A) Prevalence of ATP-III MetS; (B) Model 1 score (i.e., without racial/ethnic specificity); (C) Model 2 Mets severity score (accounting for sex and racial/ethnic differences). Mean model scores presented are estimated from linear models that adjust for highest household education and household income (values presented for those with a highest household education = high school and an income-to-poverty ratio value of 2).