The National Health and Nutrition Examination Survey (NHANES) is an extensive and continuous program. It is conducted by the National Center for Health Statistics (NCHS) under the Centers for Disease Control and Prevention (CDC). For our analysis, we collected and analyzed data from the 2017–2018 NHANES cycle. Among the 24,814 participants from the NHANES study conducted between 2017 and 2020, 14,657 were non-pregnant adults aged 20 and older. After excluding 2,346 individuals with viral hepatitis and those who had excessive drinking, we retained 12,311 participants. Viral hepatitis was diagnosed through a positive hepatitis B surface antigen (HBsAg) test. It can also be identified by the presence of hepatitis C virus (HCV) RNA. High alcohol consumption is characterized by a daily intake of at least three drinks for women. For men, it is defined as a daily intake of four drinks. Additionally, consuming five or more drinks on a single occasion within a month also indicates high alcohol consumption (9). Participants with incomplete NHHR data were excluded, resulting in 10,494 individuals eligible for further selection. The primary analysis included 9,864 individuals after excluding 630 participants lacking VCTE data. For the supplementary analysis investigating the relationship between NHHR and MASLD diagnosis using the fatty liver index (FLI), 4,860 participants were enrolled. The evaluation of the association between NHHR and advanced liver fibrosis, based on fibrosis-4 index (FIB-4) and BARD criteria, included 10,398 and 10,494 participants, respectively (Figure 1). The NCHS Ethics Review Board approved NHANES 2017–2020, and all participants provided informed consent. Data collection and analysis were undertaken in accordance with NHANES guidelines.

Liver stiffness and CAP, important indicators of liver steatosis, were assessed using VCTE with the FibroScan 502 V2 Touch device (Echosens, North America). The device at the NHANES Mobile Examination Center (MEC) was fitted with either a medium (M) or extra-large (XL) probe. FLI was ranged from 0 to 100 using the followed formula: (e^{0.953*loge (triglycerides) +0.139*BMI+0.718*loge (ggt)+0.053*waist circumference–15.745})/(1+e^{0.953*loge (triglycerides)+0.139*BMI+0.718*loge (ggt)+0.053*waist circumference–15.745}) ∗ 100 (10). FIB-4 was calculated by followed formula: age (years) × AST [U/L] / (platelets [10⁹/L] × (ALT [U/L])1/2) (10). In contrast, the BARD score, which ranges from 0 to 4, is based on BMI, the AST/ALT ratio, and a history of diabetes. A BMI above 28 kg/m² earns 1 point, an AST/ALT ratio greater than 0.8 gives 2 points, and a history of diabetes adds another point. MASLD was defined as the coexistence of hepatic steatosis and at least one of the five cardiometabolic criteria (11): (1) BMI ≥ 25 kg/m² or waist circumference ≥ 94 cm for males and ≥ 80 cm for females, (2) fasting glucose ≥ 100 mg/dl or 2-h post-load glucose levels ≥ 140 mg/dl or hemoglobin A1c ≥ 5.7% or diabetes mellitus or treatment for diabetes mellitus, (3) blood pressure ≥ 130/85 mmHg or antihypertensive drug treatment, (4) fasting plasma triglycerides ≥ 150 mg/dl or lipid-lowering treatment, (5) plasma HDL-cholesterol < 40 mg/dl for men and < 50 mg/dl for women or lipid-lowering treatment. Hepatic steatosis diagnosis requires meeting these criteria: (1) liver steatosis (CAP ≥ 248 DB/m) (12); (2) FLI ≥ 60 (13). For the identification of advanced liver fibrosis, three conditions are necessary: (1) VCTE ≥ 8.8 E/kPa (indicating liver stiffness F3 and F4) (14), (2) FIB-4 > 3.25 (15), and (3) BARD ≥ 2 (10, 16).

The NHHR functions as an independent variable for assessing exposure. To compute NHHR, we utilized the Non-HDL-C/HDL-C ratio method established in previous studies (17). Non-HDL-C is measured by subtracting HDL-C from total cholesterol (TC) using the lipid profiles of fasting subjects. The levels of TC and HDL-C were detected by automatic biochemical analyzer. For determining TC concentrations, both the Roche Cobas 6000 and Roche Modular P chemical analyzers were employed during the analytical process.

We analyzed sociodemographic factors including gender, age, education level (categorized as below high school, high school or equivalent, and above high school), and race/ethnicity (classified as Mexican American, non-Hispanic Black, non-Hispanic White, other Hispanic, and other races). We examined the poverty-income ratio (PIR), assessing family income against the poverty threshold. For analytical purposes, we divided PIR into three categories: less than 1.30, 1.30 to 3.49, and 3.50 or higher. This framework allowed us to examine the potential effects of sociodemographic factors on our results. We also evaluated health-related covariates such as body mass index (BMI in kg/m²), waist circumference (WC in cm), smoking status, alcohol use, and physical activity levels. BMI is calculated by dividing a person’s weight in kilograms by their height in meters squared. Smoking status was classified as follows: individuals who had never smoked 100 cigarettes were considered never smokers. Individuals were classified as former smokers if they answered “No” to the current smoking question, and as current smokers if they responded “Yes” (18). Alcohol consumption was categorized into three groups: “never” drinkers (fewer than 12 drinks per year), “former” drinkers (previously consumed 12 or more drinks annually but have stopped), and “current” drinkers (consume 12 or more drinks annually) (18). Moderate drinking is characterized by a daily intake of at least two drinks for women and three for men, or engaging in binge drinking twice or more monthly (9).

Participants were categorized based on their weekly metabolic equivalent (MET) minutes of vigorous activity into four groups: No moderate to vigorous physical activity (NMVPA) for 0 MET-minutes/week, Low (LMVPA) for 1–599 MET-minutes/week, Moderate (MMVPA) for 600–1,199 MET-minutes/week, and High (HMMVPA) for 1,200 or more MET-minutes/week.

Venous blood samples were collected to assess various biomarkers, including levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST). Based on fasting blood glucose and insulin levels, we evaluated insulin resistance using the homeostasis model assessment of insulin resistance (HOMA-IR). The formula for HOMA-IR is:

We also assessed the presence of various coexisting conditions, particularly Diabetes Mellitus (DM) and hypertension. Diabetes diagnosis was determined by a glycohemoglobin level of at least 6.5%, the use of diabetes medications or insulin, or a self-reported diabetes diagnosis (19). Hypertension was defined by criteria such as a mean systolic blood pressure ≥ 140 mmHg, a mean diastolic blood pressure ≥ 90 mmHg, or a self-reported hypertension diagnosis (20).

We utilized the “mice” package to create five imputed datasets through chained equations for addressing missing covariate data. This method was designed to minimize the effect of missing data on our analyses. Post-imputation, we evaluated the robustness of our findings through sensitivity analyses across five data sets (refer to Supplementary Tables 1, 2).

Throughout the study, we utilized a variety of strategies to reduce duplicate information. Baseline characteristics by NHHR quartiles were presented using weighted means and standard errors for continuous variables, and weighted proportions for categorical variables. By incorporating sampling weights, the population is represented more accurately.

We employed multivariable weighted linear regression models to examine the association between NHHR and both MASLD and advanced liver fibrosis diagnosed through VCTE, allowing for an analysis of NHHR’s influence on these conditions while adjusting for pertinent factors. Restricted cubic splines were employed to assess non-linear relationships by modeling complex patterns deviating from linearity. In addition, threshold effect analyses were performed to identify significant cut-off points. Data were stratified by various factors for subgroup analyses to evaluate how the association between NHHR and body fat distribution varies among different groups. To assess NHHR’s effectiveness, we examined ROC curves to depict sensitivity against 1-specificity and determine cut-off points from the results. Our approach seeks to provide a comprehensive analysis, minimizing redundancies and highlighting critical nuances in the relationship between NHHR, MASLD, and advanced liver fibrosis. A p-value below 0.05 was deemed statistically significant. All analyses were conducted using R software (version 4.4.1).

The main analysis of this study involved 9,864 participants, divided into four groups according to quartiles of NHHR. Specifically, 2,469 participants were assigned to Group Q1 (48.62–109.24), 2,461 to Group Q2 (109.24–135.57), 2,468 to Group Q3 (135.57–168.38), and 2,466 to Group Q4 (168.38–1,075.92). Table 1 displays the baseline characteristics of participants categorized by NHHR strata. We analyzed baseline characteristics across the NHHR quartiles, finding significant differences in age, gender, race, education, smoking status, alcohol consumption, physical activity, BMI, waist circumference, ALT levels, HOMA-IR, and the prevalence of Diabetes Mellitus and hypertension. We conducted a stratified analysis based on MASLD presence and advanced liver fibrosis diagnosed through VCTE to explore potential NHHR differences among the groups. We compared NHHR values concerning the presence of MASLD and advanced liver fibrosis, diagnosed via vibration-controlled transient elastography. Figure 2 illustrates significant differences in NHHR between the MASLD and non-MASLD groups, as well as between individuals with and without advanced liver fibrosis.

In the main analysis, multiple logistic regression was employed to evaluate the independent association between NHHR and MASLD. Model 1 did not involve any adjustments. Higher NHHR levels in the Q2, Q3, and Q4 groups were significantly associated with increased odds of MASLD, with odds ratios of 1.70 (95% CI 1.48–1.96), 3.14 (95% CI 2.70–3.66), and 5.26 (95% CI 4.50–6.16), respectively, compared to the Q1 group (P for trend < 0.0001). Model 2 was adjusted for age, race, gender, PIR, education level, physical activity, BMI, waist circumference, smoking status, and alcohol consumption. After adjustments, elevated NHHR levels in Q2, Q3, and Q4 groups remained significantly associated with MASLD, with odds ratios of 1.11 (95% CI: 0.83–1.49), 1.65 (95% CI: 1.27–2.12), and 1.99 (95% CI: 1.56–2.54), respectively (P for trend < 0.0001). To further account for additional covariates, Model 3 included adjustments for age, sex, education, race, PIR, physical activity, BMI, waist circumference, smoking status, alcohol consumption, ALT, AST, HOMA-IR, Diabetes Mellitus, and hypertension. After adjustments, elevated NHHR levels in Q3 and Q4 groups were significantly linked to MASLD, with odds ratios of 1.59 (95% CI: 1.20–2.11) and 1.83 (95% CI: 1.40–2.39), respectively, and a trend P-value of < 0.0001 (Table 2).

In the secondary analysis, we redefined MASLD based on serum levels of FLI to further explore its correlation with NHHR (Supplementary Table 3). Compared to diagnoses made with vibration-controlled transient elastography, higher NHHR levels in the Q3 and Q4 groups maintained a significant positive association with MASLD across all models: Model 1 (OR = 3.48, 95% CI: 2.63–4.59; OR = 6.49, 95% CI: 4.68–9.00; P for trend < 0.0001), Model 2 (OR = 4.65, 95% CI: 2.96–7.32; OR = 13.32, 95% CI: 7.53–23.54; P for trend < 0.0001), and Model 3 (OR = 3.90, 95% CI: 2.46–6.17; OR = 11.66, 95% CI: 6.56–20.72; P for trend < 0.0001).

In the main analysis, multiple logistic regression was utilized to assess the independent association between NHHR and advanced liver fibrosis. In Model 2, adjustments were made for age, gender, education level, race, physical activity, PIR, BMI, smoking status, waist circumference, and alcohol consumption. Higher NHHR levels in the Q2 and Q3 groups were significantly associated with advanced liver fibrosis compared to the reference group, with odds ratios of 0.60 (95% CI 0.41–0.89, P = 0.01) and 0.60 (95% CI 0.43–0.83, P = 0.004), respectively. To further account for additional covariates, Model 3 included adjustments for age, sex, education, race, PIR, physical activity, BMI, waist circumference, smoking status, alcohol consumption, ALT, AST, HOMA-IR, Diabetes Mellitus, and hypertension. In this model, elevated NHHR levels in the Q2 and Q3 groups were significantly linked to a lower risk of advanced liver fibrosis, with odds ratios of 0.61 (95% CI 0.40–0.94, P = 0.03) and 0.64 (95% CI 0.47–0.89, P = 0.01), respectively (Table 3).

In the secondary analysis, we redefined advanced liver fibrosis utilizing serum levels of FIB-4 and BARD to further examine its correlation with NHHR (Supplementary Tables 4, 5). For the FIB-4 metric, elevated NHHR levels in the Q3 group were significantly positively associated with MASLD in all models: Model 1 (OR = 0.20, 95% CI 0.11–0.36, P < 0.0001), Model 2 (OR = 0.20, 95% CI 0.08–0.51, P = 0.002), and Model 3 (OR = 0.32, 95% CI 0.13–0.80, P = 0.02). In the BARD analysis, when NHHR was considered a continuous variable in Model 3, it continued to serve as an independent protective factor against advanced liver fibrosis (OR = 0.85, 95% CI: 0.77–0.94, P = 0.004), compared to diagnoses based on vibration-controlled transient elastography.

Utilizing restricted cubic splines and adjusting for age, education level, sex, race, PIR, physical activity, BMI, waist circumference, smoking status, alcohol consumption, ALT, AST, HOMA-IR, Diabetes Mellitus, and hypertension, we identified a U-shaped nonlinear relationship between NHHR and both MASLD (P for all = 0.000; P for nonlinear = 0.029) (see Figure 3) and advanced liver fibrosis (P for all = 0.0001; P for nonlinear = 0.000) (see Figure 3), indicating that both low and high NHHR values are associated with an increased risk of advanced fibrosis, while moderate NHHR values may be associated with a lower risk.

Additionally, subgroup analyses were conducted based on age, sex, Diabetes Mellitus, hypertension, and BMI to explore the independent association between NHHR, MASLD, and advanced liver fibrosis diagnosed via VCTE. The findings revealed that NHHR was a significant independent risk factor for MASLD in specific groups, including older adults (over 60), females, individuals without diabetes, those with hypertension, and overweight or obese participants (see Figure 4A). In contrast, NHHR served as a significant independent protective factor against advanced liver fibrosis primarily among younger and normal-weight individuals. Conversely, in middle-aged participants, as well as those with diabetes, hypertension, and obesity, NHHR emerged as an independent risk factor for advanced liver fibrosis (see Figure 4B).

Figure 5 illustrates the areas under the curve (AUC) for MASLD diagnosed via VCTE (roc1), advanced liver fibrosis via VCTE (roc2), MASLD using FLI (roc3), advanced liver fibrosis assessed by FIB-4 (roc4), and advanced liver fibrosis using BARD (roc5), which were 0.643 (95% CI: 0.632–0.654), 0.554 (95% CI: 0.538–0.569), 0.680 (95% CI: 0.664–0.695), 0.677 (95% CI: 0.637–0.716), and 0.510 (95% CI: 0.499–0.521), respectively. The diagnostic capability of NHHR for predicting MASLD using FLI surpasses that of MASLD via VCTE, advanced liver fibrosis via VCTE, advanced liver fibrosis assessed by FIB-4, and advanced liver fibrosis using BARD. The optimal cutoff point for NHHR in predicting MASLD using FLI was found to be 2.476, with sensitivity and specificity values of 0.589 and 0.698, respectively.

This article has several advantages: (1) First, it is the first study to confirm the association between NHHR and both MASLD and advanced liver fibrosis in a large population. (2) The research includes both primary and supplementary analyses, making the conclusions more robust and reliable. (3) This study adjusted for potential covariates to reduce the influence of confounding factors, thereby clarifying the independent relationship between NHHR and MASLD as well as advanced liver fibrosis. (4) We not only employed conventional regression analysis to demonstrate the relationship between NHHR and the diseases but also utilized restricted cubic splines (RCS) and receiver operating characteristic (ROC) analysis to enrich our study. This multifaceted approach provides robust evidence of the relationship between NHHR and liver diseases from various angles.

Also, the current study has several limitations: (1) The diagnosis of MASLD has not yet been widely implemented in clinical practice, which may restrict the applicability of the results. (2) It must be acknowledged that some covariates in the data used in this study were not recorded or measured, thus it is not possible to completely exclude the influence of unmeasured or unknown confounding factors on the results. (3) This study only analyzed the predictive value of baseline NHHR for MASLD and liver fibrosis recorded during a single examination, and it does not confirm whether the continuous dynamic changes of NHHR during follow-up would outperform the baseline NHHR alone. However, our findings indicate that the NHHR recorded at a single time point provides good predictive value for the occurrence of MASLD and liver fibrosis, and its assessment is more economical and demonstrates better patient compliance compared to other invasive tests. (4) The current study population primarily consists of individuals from the United States, highlighting the need for future research involving diverse racial groups. (5) This study is a cross-sectional study, and it cannot establish a causal relationship between NHHR and MASLD as well as advanced Liver Fibrosis. (6) While our chosen methods provided useful insights, they may not capture all the complexities of the data. Some alternative methods, such as random forest, might align better with the experimental hypotheses. (7) Using secondary databases may inevitably encounter issues such as sampling bias, variable constraints, self-reported data and missing data in the data collection process.