Association of High-Density Lipoprotein Cholesterol-Based Inflammatory Markers With MASLD and Significant Liver Fibrosis in US Adults: Insights From NHANES 2017–2020

Abstract

INTRODUCTION:

Systemic inflammation and lipid metabolism disturbances are important hallmarks of the onset and progression of metabolic dysfunction-associated steatotic liver disease (MASLD). We aimed to explore the association of lymphocyte-high-density lipoprotein-cholesterol ratio (LHR), monocyte-HDL-C ratio (MHR), neutrophil-HDL-C ratio (NHR), and platelet-HDL-C ratio (PHR) with MASLD and significant liver fibrosis using NHANES 2017–2020 data.

METHODS:

LHR, MHR, NHR, and PHR were calculated based on complete blood count parameters and serum HDL-C. MASLD and liver fibrosis were diagnosed based on transient elastography. Multivariate logistic regression analyses were used to explore these associations, and receiver operating characteristic was used to compare the predictive power of these markers.

RESULTS:

A total of 8,341 participants were included, and the prevalence of MASLD and significant liver fibrosis was 45.1% and 11.57%, respectively. In fully adjusted models, log-transformed LHR, MHR, NHR, and PHR were positively associated with the odds of MASLD (odds ratio 1.853, 1.685, 1.470, and 1.879, respectively) and significant liver fibrosis (odds ratio 1.570, 1.425, 1.396, and 1.384, respectively) (all P < 0.05). Most of these associations were nonlinear, and significant positive correlations existed only after their respective inflection points. The association of LHR with significant liver fibrosis was more pronounced in men. Receiver operating characteristic analysis showed that LHR/NHR was superior in predicting MASLD, whereas MHR/NHR distinguished significant liver fibrosis better than other markers.

DISCUSSION:

LHR, MHR, NHR, and PHR were independently associated with MASLD and liver fibrosis in US adults and may serve as emerging predictors. Future cohort studies are needed to confirm these findings and explore clinical predictive value.

INTRODUCTION

Metabolic dysfunction-associated steatotic liver disease (MASLD) is a novel nomenclature recently proposed by the Delphi consensus of three large pan-national liver associations as an alternative to what was previously known as nonalcoholic fatty liver disease (NAFLD) (1). MASLD is an umbrella term for a spectrum of steatotic liver disease (SLD) closely related to metabolic disorders that range from simple steatosis to metabolic dysfunction-associated steatohepatitis characterized by lobular inflammation and can continue to progress to cirrhosis and hepatocellular carcinoma (2,3). As the most common chronic liver disease worldwide, it affects more than one-third of the world's adult population (4). A large body of high-quality epidemiologic evidence suggests that MASLD is associated with an increased odds of major liver-related events such as liver fibrosis/cirrhosis, hepatocellular carcinoma, liver transplantation, and liver-related mortality (5). In addition, as a multisystemic disease, MASLD is implicated in an increased incidence of a variety of nonliver-related events, such as cardiovascular disease, multiple cancers, type 2 diabetes (T2D), and chronic kidney disease (6). The burden of MASLD has dramatically increased over the past few decades and will remain on the rise (4,7). Identifying modifiable factors for developing MASLD and implementing public education and intervention is critical to public health.

Chronic systemic inflammation and disorders of lipid metabolism are important pathogenic mechanisms and promotional hallmarks of disease progression in MASLD. In recent years, crosstalk between the inflammatory immune response and lipid metabolism has been suggested to play a major role in the development of MASLD (8). As a pathophysiological initiator of MASLD, lipotoxicity due to hepatic and systemic lipid overload induces important events such as mitochondrial dysfunction, endoplasmic reticulum stress, and autophagy dysfunction through complex molecular biological mechanisms and leads to oxidative stress, hepatic inflammation, and systemic immune-inflammatory activation (9–11). A variety of immune cells are involved in the inflammatory cascade that underlies the onset and progression of MASLD, including neutrophils, lymphocytes, monocytes, and platelets (12,13). These inflammatory immune cells promote inflammatory infiltration and hepatic fibrosis through the secretion of proinflammatory mediators such as cytokines and can promote lipid metabolism disorders through a vicious cycle (14). In addition, reduced blood high-density lipoprotein-cholesterol (HDL-C) levels have been observed in patients with MASLD and are potentially associated with disease progression and prognosis, suggesting that HDL-C may be a key marker of lipid metabolism disorders in MASLD (15,16). In recent years, multiple markers have been shown to be promising markers of disease development and progression by combining immune cells and HDL-C, including neutrophil-HDL-C ratio (NHR), lymphocyte-HDL-C ratio (LHR), monocyte-HDL-C ratio (MHR), and platelet-HDL-C ratio (PHR). A large body of observational evidence suggests that these novel inflammatory metabolic markers are associated with the development of multiple cardiometabolic disorders, such as metabolic syndrome (MetS), T2D, and cardiovascular disease (17–19).

However, the association of these markers with MASLD and liver fibrosis remains poorly investigated. To the best of our knowledge, only one recent cross-sectional study has shown a positive correlation between NHR and MASLD but not with liver fibrosis (20). The association of other markers such as LHR, MHR, and PHR with MASLD and clinically significant liver fibrosis remains unknown. Therefore, in this study, we aimed to use population-based data from the National Health and Nutrition Examination Survey (NHANES) to explore the associations of LHR, MHR, NHR, and PHR with MASLD and significant hepatic fibrosis and to explore the comparative predictive power of these markers in MASLD and/or significant hepatic fibrosis.

METHODS

Study design and population

NHANES is a research program conducted by the Centers for Disease Control and Prevention to assess the health and nutritional status of noninstitutionalized adults and children in the United States Overall, NHANES is a nationwide, population-based, multiethnic cross-sectional survey with a complex multistage probability sampling design. All survey protocols were approved by the National Center for Health Statistics Ethics Review Board, and all participants provided written informed consent. NHANES is a publicly accessible, free database for the public and researchers, and all participants were deidentified (21). Therefore, as a secondary analysis of the NHANES database, ethical approval from the local institution was waived.

The study population selection flowchart was presented in Figure 1. We first included 15,549 adult participants from NHANES 2017 through March 2020. Participants with vibration-controlled transient elastography (VCTE) not performed (n = 1,101) or ineligible (n = 1,101) were excluded. Next, we excluded participants with chronic viral hepatitis (n = 347), autoimmune hepatitis (n = 17), excessive alcohol consumption (n = 1,256), missing LHR/MHR/NHR/PHR data (n = 736), and missing covariates (n = 2,739). A total of 8,341 participants were finally included.

Figure 1.

Flowchart of study population selection, NHANES 2017–2020.

Evaluation of LHR, MHR, NHR, and PHR

These markers were assessed according to the respective inflammatory cell counts (1,000 cells/μL) and serum HDL-C levels (mmol/L). The specific formulas for the 4 markers were as follows: LHR = lymphocyte count/HDL-C, MHR = monocyte count/HDL-C, NHR = neutrophil count/HDL-C, and PHR = platelet count/HDL-C. Neutrophil, lymphocyte, monocyte, and platelet counts were derived from complete blood counts. The complete blood count parameters were derived from the Beckman Coulter counting and sizing method, combined with an automated dilution and mixing device for sample processing. Serum HDL-C levels were measured using a magnesium/dextran sulfate solution. In addition, because these marker levels were distributed skewed among participants when treated as continuous variables, we performed a logarithmic transformation for normalization (22).

Diagnosis of MASLD

MASLD is defined as the presence of at least one cardiometabolic risk factor underlying SLD and the exclusion of other causes of steatosis (1). In this study, we assessed SLD using the VCTE method (FibroScan). The Controlled Attenuation Parameter (CAP, decibels per meter [dB/m]) was used to assess liver fat content as measured by VCTE, a metric that has been shown to have excellent accuracy in a large body of clinical studies (23). We used CAP ≥274 dB/m for assessing the presence of SLD. This cutoff value was shown to have optimal sensitivity (for assessment of ≥5% steatosis) in a prospective cohort study using liver biopsy as a reference and has been used in numerous clinical studies for the diagnosis of MASLD (24–26). Cardiometabolic risk factor consisted of the following 5 items: overweight/obesity/central obesity (body mass index [BMI] ≥25 kg/m² or waist circumference>94 cm [male]/80 cm [female]), hyperglycemia (fasting glucose ≥5.6 mmol/L, 2 h postprandial glucose ≥7.8 mmol/L, HbA1c ≥ 5.6%, presence of T2D, or presence of T2D treatment), hypertension (blood pressure ≥ 130/85 mm Hg or on antihypertensive medication), elevated plasma triglyceride levels (≥1.7 mmol/L or on lipid-lowering therapy), and lowered serum HDL-C (≤1 mmol/L for men or ≤1.3 mmol/L for women or on lipid-lowering therapy) (1). Other factors contributing to hepatic steatosis such as chronic viral hepatitis, autoimmune hepatitis, and significant alcohol consumption (>30 g/day [men]/20 g/day [women]) were excluded (25).

Assessment of significant liver fibrosis

The liver stiffness measurement index of the VCTE test was used to assess the degree of hepatic fibrosis, which was judged to be significant (≥F2) based on median liver stiffness measurement ≥8.0 kPa (25).

Covariates

To adjust for the effects of potential confounders, we included a range of covariates into the multivariate logistic regression models. These covariates included age, sex, race/ethnicity, household income-poverty ratio (PIR), education level, marital status, smoking, physical activity, dietary energy intake, diabetes, hypertension, and BMI. Smoking history was categorized as never smokers (<100 lifetime cigarettes), former smokers (≥100 lifetime cigarettes but not smoking at the time of the interview date), and current smokers (≥100 lifetime cigarettes and still actively smoking), based on the participant's self-reported assessment on the smoking-related questionnaire (27). Participants were asked about weekly physical activity levels by a trained interviewer at home. Physical activity was categorized as no, moderate, or vigorous physical activity participation based on self-report on the Physical Activity Questionnaire (28). Total daily energy intake was obtained from the NHANES dietary interview section, which was assessed through its association with the US Department of Agriculture's Food and Nutrient Database for Dietary Studies (29). Diabetes and hypertension were assessed based on previous studies, including self-reported history of disease, met blood glucose (HbA1c)/blood pressure test values, or were taking medications (28).

Statistical analysis

To account for the complex NHANES study design (e.g., oversampling and nonresponse), all statistical analyses were performed using appropriate weighting criteria as specified in the NHANES Analytic Guidelines (30). To explore baseline differences in the included population, we performed baseline analyses according to MASLD or significant liver fibrosis status. Continuous variables were expressed as mean ± standard error and detected by weighted t-tests, and categorical variables were expressed as number (percentage) and tested by weighted χ² analysis. Multivariate logistic regression analyses were used to explore the association of LHR, MHR, NHR, and PHR with MASLD and significant liver fibrosis and calculated the odds ratio [OR] and 95% confidence interval (CI). Three multivariate logistic regression models were constructed, with varying degrees of adjustment. The crude model was not adjusted for any covariates. Model 1 was partially adjusted for age, sex, race/ethnicity, education level, PIR, and marital status, whereas Model 2 additionally adjusted for smoking, physical activity, dietary energy intake, diabetes, hypertension, and BMI in addition to Model 1. Restricted cubic spline (RCS) modeling was used to ascertain whether these associations exhibited nonlinear characteristics. To ascertain whether these associations persisted when analyzed according to sex subgroups, stratified analyses were conducted. Furthermore, interaction tests were used to identify whether sex was an effect modifier. To validate the stability of the results, we performed several sensitivity analyses. First, given the higher prevalence of cancer among patients with MASLD and the resulting increased burden of disease (for example, bladder cancer is one of the most serious cancers in the elderly, and insulin resistance worsens when chronic liver disease is more aggressive) (31), we additionally adjusted for cancer history in fully adjusted models in multivariable regression analyses. We then chose another commonly used CAP cutoff value for assessing hepatic steatosis for sensitivity analysis. The cutoff value CAP ≥285 dB/m was used to identify the presence of SLD, which was demonstrated to possess the greatest Youden index in a prospective liver biopsy cohort study (32). Finally, we compared the ability of LHR, MHR, NHR, and PHR in predicting MASLD and significant liver fibrosis by the area under the curve (AUC) of the receiver operating characteristic (ROC) curves (using the DeLong test). We also compared the AUC of these markers with common noninvasive markers of steatosis (including Framingham Steatosis Index [FSI], hepatic steatosis index, fatty liver index, alanine transaminase, aspartate aminotransferase)/fibrosis (including fibrosis-4 index [FIB-4], AST-to-platelet ratio index, steatosis-associated fibrosis estimator, and NAFLD fibrosis score [NFS]) and explored whether conventional markers of steatosis/fibrosis in combination with LHR/NHR/PHR/MHR had improved predictive power compared with their use alone. Data were processed and analyzed using R (version 4.2.3), with P values of less than 0.05 considered to be statistically significant.

RESULTS

Baseline characteristics

The total number of participants included in the study was 8,341 (mean age 48.289 years), and the prevalence of MASLD among them was 45.10%. When compared with participants without MASLD, those who had MASLD were found to be older, had a higher energy intake and BMI, and were more likely to be Mexican American, nonsingle, ≤high school education, former smokers, and have diabetes and hypertension. The MASLD population had significantly higher LHR, MHR, NHR, and PHR (all P < 0.0001) (Table 1). The baseline analysis based on significant liver fibrosis status was presented in Supplementary Table S1 (http://links.lww.com/CTG/B323). Significant liver fibrosis was present in 965 participants (11.57%). Notably, LHR, MHR, NHR, and PHR were significantly higher in patients with hepatic fibrosis compared with participants without significant hepatic fibrosis (all P < 0.0001).

Table 1.

Baseline analysis based on MASLD status, NHANES 2017–2020

Variable	Total (n = 8,341)	No (n = 4,579)	Yes (n = 3,762)	P-value
Age, years	48.289 ± 0.598	45.871 ± 0.685	51.356 ± 0.598	<0.0001
PIR	3.087 ± 0.042	3.098 ± 0.052	3.074 ± 0.053	0.71
Energy intake, kcal/d	2,107.596 ± 14.510	2043.043 ± 17.326	2,189.496 ± 18.570	<0.0001
BMI	30.070 ± 0.195	26.826 ± 0.190	34.186 ± 0.211	<0.0001
LHR	1.797 ± 0.026	1.614 ± 0.041	2.030 ± 0.032	<0.0001
MHR	0.467 ± 0.006	0.409 ± 0.005	0.541 ± 0.008	<0.0001
NHR	3.447 ± 0.047	2.991 ± 0.041	4.025 ± 0.063	<0.0001
PHR	195.475 ± 1.910	178.971 ± 1.904	216.415 ± 2.199	<0.0001
Sex				<0.0001
Male	3,893 (46.767)	1946 (42.290)	1947 (52.447)
Female	4,448 (53.233)	2,633 (57.710)	1815 (47.553)
Race				<0.0001
Mexican American	1,053 (8.468)	414 (6.028)	639 (11.564)
Non-Hispanic Black	1967 (10.471)	1,248 (12.172)	719 (8.314)
Non-Hispanic White	3,060 (64.594)	1,626 (64.878)	1,434 (64.233)
Other Hispanic	810 (6.642)	441 (6.900)	369 (6.316)
Other race	1,451 (9.825)	850 (10.023)	601 (9.573)
Marital status				<0.0001
No-single	4,949 (62.736)	2,572 (58.277)	2,377 (68.393)
Single	3,392 (37.264)	2007 (41.723)	1,385 (31.607)
Education				0.012
<high school	577 (3.115)	280 (2.807)	297 (3.505)
High school	2,884 (34.854)	1,534 (32.996)	1,350 (37.210)
>high school	4,880 (62.032)	2,765 (64.197)	2,115 (59.285)
Smoking				<0.001
Never	4,974 (59.562)	2,815 (60.953)	2,159 (57.797)
Former	2029 (25.319)	944 (22.209)	1,085 (29.266)
Now	1,338 (15.118)	820 (16.838)	518 (12.937)
Physical activity				0.116
No	4,228 (45.968)	2,324 (47.153)	1904 (44.464)
Moderate	2008 (26.872)	1,073 (25.505)	935 (28.607)
Vigorous	2,105 (27.160)	1,182 (27.342)	923 (26.929)
Hypertension				<0.0001
No	4,579 (61.330)	2,914 (72.854)	1,665 (46.709)
Yes	3,762 (38.670)	1,665 (27.146)	2097 (53.291)
Diabetes				<0.0001
No	6,516 (83.422)	4,012 (92.572)	2,504 (71.812)
Yes	1825 (16.578)	567 (7.428)	1,258 (28.188)

Continuous variables were expressed as mean ± standard error and detected by weighted t-tests, and categorical variables were expressed as number (percentage) and tested by weighted χ² analysis.

LHR, lipoprotein-cholesterol ratio; MASLD, metabolic dysfunction-associated steatotic liver disease; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; PHR, platelet-HDL-C ratio.

Association of LHR, MHR, NHR, and PHR with MASLD and significant liver fibrosis

In fully adjusted model 2, we found that logarithmically (Log2) transformed LHR, MHR, NHR, and PHR were all positively associated with odds of MASLD (OR and 95% CI 1.853 [1.633–2.103], 1.685 [1.494–1.901], 1.470 [1.266–1.707], and 1.879 [1.592–2.218], respectively; all P < 0.05). A comparison between the first and fourth quartiles for LHR, MHR, NHR, and PHR revealed a significantly higher prevalence of MASLD in the latter group (all P for trend < 0.0001). In comparison with the reference population (Q1), the ORs for developing MASLD in Q4 were 3.015, 2.397, 2.060, and 2.602 for LHR, MHR, NHR, and PHR, respectively (Table 2). Similarly, all these markers were significantly and positively correlated with the odds of significant liver fibrosis (Log2(LHR): OR = 1.570, P < 0.0001; Log2(MHR): OR = 1.425, P = 0.0249; Log2(NHR): OR = 1.396, P = 0.0283; Log2(PHR): OR = 1.384, P = 0.0276). Similarly, being in higher quartiles was associated with a significantly increased likelihood of significant liver fibrosis (all P for trend <0.05) (Table 3).

Table 2.

Association of LHR, MHR, NHR, and LHR with the prevalence of MASLD in US adults

	Crude model OR (95% CI) P value	Model 1 OR (95% CI) P value	Model 2 OR (95% CI) P value
Log2(LHR)	2.504 (2.266–2.766) <0.0001	2.953 (2.639–3.305) <0.0001	1.853 (1.633–2.103) <0.0001
LHR quartile
Q1	Ref.	Ref.	Ref.
Q2	1.791 (1.452–2.209) <0.0001	2.039 (1.639–2.536) <0.0001	1.429 (1.147–1.780) 0.0049
Q3	3.332 (2.766–4.014) <0.0001	4.132 (3.328–5.129) <0.0001	2.238 (1.761–2.843) <0.0001
Q4	4.879 (4.043–5.886) <0.0001	6.648 (5.424–8.148) <0.0001	3.015 (2.380–3.819) <0.0001
P for trend	<0.0001	<0.0001	<0.0001
Log2(MHR)	2.728 (2.519–2.954) <0.0001	2.828 (2.572–3.108) <0.0001	1.685 (1.494–1.901) <0.0001
MHR quartile
Q1	Ref.	Ref.	Ref.
Q2	2.201 (1.904–2.544) <0.0001	2.311 (2.002–2.668) <0.0001	1.479 (1.233–1.774) 0.0005
Q3	3.402 (2.839–4.076) <0.0001	3.506 (2.892–4.250) <0.0001	1.757 (1.393–2.216) 0.0001
Q4	5.623 (4.716–6.704) <0.0001	5.876 (4.866–7.095) <0.0001	2.397 (1.930–2.978) <0.0001
P for trend	<0.0001	<0.0001	<0.0001
Log2(NHR)	2.551 (2.313–2.813) <0.0001	2.618 (2.321–2.953) <0.0001	1.470 (1.266–1.707) 0.0001
NHR quartile
Q1	Ref.	Ref.	Ref.
Q2	1.673 (1.335–2.095) 0.0001	1.681 (1.312–2.155) 0.0003	1.121 (0.872–1.439) 0.3841
Q3	3.074 (2.624–3.600) <0.0001	3.181 (2.689–3.762) <0.0001	1.483 (1.260–1.747) 0.0001
Q4	5.800 (4.651–7.233) <0.0001	6.009 (4.601–7.846) <0.0001	2.060 (1.484–2.859) 0.0004
P for trend	<0.0001	<0.0001	<0.0001
Log2(PHR)	3.073 (2.795–3.378) <0.0001	3.678 (3.264–4.144) <0.0001	1.879 (1.592–2.218) <0.0001
PHR quartile
Q1	Ref.	Ref.	Ref.
Q2	1.672 (1.337–2.092) 0.0001	1.840 (1.481–2.287) <0.0001	1.322 (1.012–1.728) 0.0545
Q3	3.156 (2.564–3.886) <0.0001	3.694 (2.965–4.603) <0.0001	2.028 (1.600–2.569) <0.0001
Q4	5.087 (4.403–5.877) <0.0001	6.503 (5.530–7.647) <0.0001	2.602 (2.075–3.263) <0.0001
P for trend	<0.0001	<0.0001	<0.0001

The crude model was not adjusted for any covariates. Model 1 was partially adjusted for age, sex, race/ethnicity, education level, poverty income ratio (PIR), and marital status, whereas Model 2 additionally adjusted for smoking, physical activity, dietary energy intake, diabetes, hypertension, and BMI in addition to Model 1.

CI, confidence interval; LHR, lipoprotein-cholesterol ratio; MASLD, metabolic dysfunction-associated steatotic liver disease; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; OR, odds ratio; PHR, platelet-HDL-C ratio.

Table 3.

Association of LHR, MHR, NHR, and LHR with the prevalence of significant liver fibrosis in US adults

	Crude model OR (95% CI) P value	Model 1 OR (95% CI) P value	Model 2 OR (95% CI) P value
Log2(LHR)	1.840 (1.616–2.096) <0.0001	1.926 (1.691–2.193) <0.0001	1.570 (1.366–1.804) <0.0001
LHR quartile
Q1	Ref.	Ref.	Ref.
Q2	1.045 (0.643–1.700) 0.8593	1.115 (0.691–1.799) 0.6589	0.955 (0.577–1.581) 0.8601
Q3	2.050 (1.570–2.677) <0.0001	2.194 (1.732–2.780) <0.0001	1.738 (1.384–2.181) 0.0001
Q4	2.595 (1.915–3.516) <0.0001	2.839 (2.143–3.760) <0.0001	2.057 (1.508–2.805) 0.0002
P for trend	<0.0001	<0.0001	<0.0001
Log2(MHR)	1.883 (1.506–2.356) <0.0001	1.803 (1.383–2.349) 0.0001	1.425 (1.068–1.901) 0.0249
MHR quartile
Q1	Ref.	Ref.	Ref.
Q2	1.766 (1.250–2.496) 0.0027	1.719 (1.210–2.441) 0.0054	1.389 (0.957–2.016) 0.0992
Q3	2.212 (1.455–3.361) 0.0007	2.078 (1.312–3.291) 0.0043	1.612 (0.990–2.624) 0.0694
Q4	3.235 (2.184–4.792) <0.0001	2.956 (1.873–4.665) 0.0001	1.923 (1.167–3.166) 0.0183
P for trend	<0.0001	0.0001	0.0244
Log2(NHR)	1.855 (1.499–2.297) <0.0001	1.818 (1.422–2.323) <0.0001	1.396 (1.057–1.844) 0.0283
NHR quartile
Q1	Ref.	Ref.	Ref.
Q2	0.773 (0.524–1.140) 0.2015	0.760 (0.504–1.147) 0.2024	0.650 (0.425–0.994) 0.0609
Q3	1.627 (1.208–2.191) 0.0028	1.582 (1.142–2.189) 0.0102	1.174 (0.825–1.669) 0.3832
Q4	2.780 (1.983–3.897) <0.0001	2.644 (1.812–3.856) <0.0001	1.642 (1.072–2.516) 0.0337
P for trend	<0.0001	<0.0001	0.0017
Log2(PHR)	1.746 (1.397–2.182) <0.0001	1.823 (1.418–2.344) 0.0001	1.384 (1.057–1.813) 0.0276
PHR quartile
Q1	Ref.	Ref.	Ref.
Q2	0.969 (0.705–1.331) 0.8451	0.969 (0.708–1.327) 0.8469	0.761 (0.545–1.062) 0.1243
Q3	1.329 (0.955–1.850) 0.1002	1.359 (0.983–1.880) 0.0741	0.995 (0.686–1.445) 0.9799
Q4	2.273 (1.619–3.192) <0.0001	2.377 (1.635–3.455) 0.0001	1.527 (1.026–2.272) 0.0500
P for trend	<0.0001	<0.0001	0.0121

CI, confidence interval; LHR, lipoprotein-cholesterol ratio; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; OR, odds ratio; PHR, platelet-HDL-C ratio.

The crude model was not adjusted for any covariates. Model 1 was partially adjusted for age, sex, race/ethnicity, education level, poverty income ratio (PIR), and marital status, whereas Model 2 additionally adjusted for smoking, physical activity, dietary energy intake, diabetes, hypertension, and BMI in addition to Model 1.

Nonlinear association exploration

RCS analysis showed that log-transformed LHR, NHR, and PHR were nonlinearly associated with MASLD odds (P for nonlinearity all <0.0001), whereas a linear association existed for MHR (P for nonlinearity = 0.2762) (Figure 2a–d). All markers were nonlinearly associated with the odds of significant liver fibrosis (P for nonlinearity all <0.05) (Figure 3a–d). Threshold effect analysis indicated that LHR, NHR, and PHR were significantly and positively correlated with the prevalence of MASLD after the inflection point (OR 2.091, 1.703, and 2.042, respectively) (Table 4). Similarly, LHR, MHR, NHR, and PHR were all positively associated with significant liver fibrosis after the inflection point (OR 1.785, 1.587, 1.979, and 1.854, respectively). Except for LHR, all markers were negatively associated with the odds of liver fibrosis before the inflection point (OR 1.171, 0.311, 0.417, and 0.453, respectively) (Table 4).

Figure 2.

RCS analysis of the association of log LHR, MHR, NHR, and PHR with MASLD. (a) LHR; (b) MHR; (c) NHR; (d) PHR. LHR, lipoprotein-cholesterol ratio; MASLD, metabolic dysfunction-associated steatotic liver disease; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; PHR, platelet-HDL-C ratio.

Figure 3.

RCS analysis of the association of log LHR, MHR, NHR, and PHR with significant liver fibrosis. (a): LHR; (b) MHR; (c) NHR; (d) PHR. LHR, lipoprotein-cholesterol ratio; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; PHR, platelet-HDL-C ratio.

Table 4.

Threshold effect analyses of the association of LHR, MHR, NHR, and PHR with MASLD and significant liver fibrosis

	OR (95% CI) P value
MASLD
Log LHR ≤ –0.337	0.527 (0.187–1.484) 0.2401
Log LHR > –0.337	2.091 (1.783–2.451) <0.0001
Log NHR ≤ 0.871	0.835 (0.502–1.391) 0.4973
Log NHR > 0.871	1.703 (1.409–2.059) <0.0001
Log PHR ≤ 6.75	0.243 (0.116–0.508) 0.0013
Log PHR > 6.75	2.042 (1.770–2.356) <0.0001
Significant liver fibrosis
Log LHR ≤ –0.086	1.171 (0.561–2.443) 0.6792
Log LHR > –0.086	1.785 (1.494–2.134) <0.0001
Log MHR ≤ –1.72	0.311 (0.164–0.590) 0.0019
Log MHR > –1.72	1.587 (1.236–2.038) 0.0017
Log NHR ≤ 1.139	0.417 (0.198–0.876) 0.0317
Log NHR > 1.139	1.979 (1.604–2.441) <0.0001
Log PHR ≤ 7.183	0.453 (0.232–0.884) 0.0308
Log PHR > 7.183	1.854 (1.367–2.514) 0.0008

Adjusted for age, sex, race/ethnicity, education level, poverty income ratio (PIR), marital status, smoking, physical activity, dietary energy intake, diabetes, hypertension, and BMI.

CI, confidence interval; LHR, lipoprotein-cholesterol ratio; MASLD, metabolic dysfunction-associated steatotic liver disease; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; OR, odds ratio; PHR, platelet-HDL-C ratio.

Stratified analysis according to sex

Interaction tests demonstrated that sex did not influence the association of LHR, MHR, NHR, and PHR with the prevalence of MASLD (all P for interaction > 0.05) (Figure 4a). However, sex affected the association of log-transformed LHR with significant liver fibrosis (P for interaction = 0.019). This association was more significant in men (OR 1.719 compared with 1.426) (Figure 4b).

Figure 4.

Stratified analysis of the association of LHR, MHR, NHR, and PHR with MASLD and significant liver fibrosis according to sex. (a) MASLD; (b) significant liver fibrosis. LHR, lipoprotein-cholesterol ratio; MASLD, metabolic dysfunction-associated steatotic liver disease; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; PHR, platelet-HDL-C ratio.

Sensitivity analysis

Additional adjustment for cancer in the fully adjusted model did not significantly change the results. We found that LHR/MHR/NHR/PHR all remained positively associated with MASLD (OR 1.869, 1.687, 1.471, and 1.883, respectively) and significant fibrosis (OR 1.571, 1.425, 1.396, and 1.384, respectively) (Supplementary Tables S2 and S3, http://links.lww.com/CTG/B323). Using CAP ≥285 dB/m as a cutoff value for diagnosis of hepatic steatosis did not significantly alter the results. In Model 2, log-transformed LHR, MHR, NHR, and PHR remained positively associated with MASLD prevalence (OR 1.801, 1.618, 1.534, and 1.822, respectively; all P < 0.0001) (Supplementary Table S4, http://links.lww.com/CTG/B323).

Comparing the predictive performance of LHR, MHR, NHR, and PHR

The ROC curves indicated that the AUCs of LHR, MHR, NHR, and PHR for predicting MASLD were 0.669, 0.666, 0.658, and 0.654, respectively. Furthermore, the AUCs of these markers predicting significant liver fibrosis were 0.632, 0.626, 0.599, and 0.576, respectively (Figure 5). ROC-specific data were presented in Table 5. Compared with LHR, MHR and PHR had significantly lower MASLD predictive ability (P of 0.0359 and 0.0044, respectively), whereas NHR was not significantly different (P = 0.6250). In comparison with MHR, LHR, and PHR exhibited markedly diminished predictive capacity for significant liver fibrosis (P of 0.0001 and < 0.0001, respectively). Conversely, NHR displayed no statistically discernible discrepancy (P = 0.3687). However, we found that FSI had the best AUC for predicting MASLD (0.8264), which was significantly higher than LHR/MHR/NHR/PHR (all P < 0.0001) (Supplementary Table S5, http://links.lww.com/CTG/B323). NFS had the highest AUC for predicting significant fibrosis (0.7269), which was significantly higher than all HDL-related inflammatory markers (Supplementary Table S6, http://links.lww.com/CTG/B323). We further compared whether FSI/NFS in combination with these HDL-related inflammatory markers had improved predictive value for MASLD/significant fibrosis compared with them alone. FSI combined with any LHR/MHR/NHR/PHR did not improve the predictive value for MASLD compared with FSI alone (Supplementary Table S7, http://links.lww.com/CTG/B323). However, compared with NFS alone, NFS combined with either MHR, NHR, or PHR had significantly improved predictive value (all P < 0.0001). NFS combined with PHR had the highest AUC (0.7681) (Supplementary Table S8, http://links.lww.com/CTG/B323).

Figure 5.

ROC curves of LHR, MHR, NHR, and PHR for the prediction of MASLD and significant liver fibrosis. (a) MASLD; (b) significant liver fibrosis. MASLD, metabolic dysfunction-associated steatotic liver disease; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; PHR, platelet-HDL-C ratio.

Table 5.

ROC data for LHR, MHR, NHR, and PHR predicting MASLD and significant liver fibrosis

Test	AUC	95% CI low	95% CI upper	Best threshold	Specificity	Sensitivity	P value for AUC comparison
MASLD
LHR	0.6687	0.6571	0.6802	1.6030	0.6062	0.6621	Ref.
MHR	0.6577	0.6461	0.6693	0.3741	0.5119	0.7254	0.0359
NHR	0.6659	0.6543	0.6775	2.9182	0.5785	0.6712	0.6250
PHR	0.6540	0.6423	0.6657	178.7640	0.5739	0.6648	0.0044
Significant liver fibrosis
MHR	0.6319	0.6134	0.6504	0.4581	0.5963	0.6041	Ref.
NHR	0.6256	0.6068	0.6444	2.9346	0.4935	0.7047	0.3687
LHR	0.5992	0.5795	0.6189	1.7928	0.6047	0.5699	0.0001
PHR	0.5762	0.5567	0.5957	178.9232	0.4835	0.6560	<0.0001

AUC, area under the curve; LHR, lipoprotein-cholesterol ratio; MASLD, metabolic dysfunction-associated steatotic liver disease; MHR, monocyte-HDL-C ratio; NHR, neutrophil-HDL-C ratio; PHR, platelet-HDL-C ratio; ROC, receiver operating characteristic.

DISCUSSION

In a cross-sectional NHANES-based analysis, novel markers based on CBC-derived immune cells and HDL-C, including LHR, MHR, NHR, and PHR, were positively associated with the prevalence of both MASLD and significant hepatic fibrosis in US adults. RCS analysis showed that most of these associations were nonlinear, i.e., positively associated with MASLD and liver fibrosis only after their respective inflection points (above a certain threshold). Sex influenced the association of LHR with significant liver fibrosis, and a more significant trend was observed in male participants. ROC analysis showed that LHR/NHR and MHR/NHR had superior predictive ability for MASLD and significant liver fibrosis, respectively. Although these HDL-related inflammatory markers did not provide higher predictive value compared with traditional markers, our results suggest that NFS in combination with MHR, NHR, or PHR had improved predictive values for significant fibrosis compared with NFS alone.

To the best of our knowledge, this is the first comprehensive exploration of the associations of LHR, MHR, NHR, and PHR with the prevalence of MASLD and significant hepatic fibrosis in the general adult participants. Only one recently published study has explored the association of NHR with MASLD and liver fibrosis. Using NHANES 2017–2020, Lu et al showed that NHR was positively associated with the prevalence of MASLD in the general population (OR 1.20, 95% CI 1.09–1.31) (20). However, NHR was not significantly associated with significant liver fibrosis (OR 1.01, 95% CI 0.94–1.09) (20). A cross-sectional analysis including 155 Turkish adult participants demonstrated that NHR was associated with steatosis severity in ultrasound-diagnosed NAFLD (NHR was significantly increased in fatty liver grades 2 and 3 compared with controls; both P < 0.05) (33). This trend of increasing with steatosis grade was consistent with the trend of increasing noninvasive liver fibrosis score (FIB-4) scores (33). A recent large cross-sectional analysis showed that NHR was positively associated with the odds of NAFLD in the Chinese health checkup population (OR 1.248, P < 0.05) (34). A large cohort study that included 9,803 Iranian participants demonstrated that LHR and MHR at Q4 (compared to quartile Q1) was only associated with increased prevalence of NAFLD in female participants (OR = 1.31 and 1.30, respectively) (19). Another cross-sectional analysis from Iran suggests that MHR was associated with metabolic prevalence factors in the NAFLD population (35). Of note, the study by Kohsari et al used immune cell percentages rather than counts to calculate LHR and MHR, whereas NAFLD was diagnosed by disease history and drug use, which may have influenced the findings (19). Wang et al demonstrated that MHR was positively associated with the odds of NAFLD in the general population (OR 2.87, 95% CI 1.95–4.22, P < 0.0001) (36). Similarly, MHR was found to be positively associated with the odds of significant liver fibrosis (OR = 1.60, 95% CI 1.08–2.37, P = 0.0182) (36). Another cross-sectional analysis from a Chinese population of 14,189 participants in medical checkups suggested that MHR was significantly and positively associated with the odds of NAFLD, even after adjusting for confounders (OR = 1.026, 95% CI: 1.002–1.052; P = 0.037) (37). In a cross-sectional analysis from NHANES 2017–2020, Lu et al demonstrated that, in comparison with Q1, individuals in Q2, Q3, and Q4 PHR exhibited significantly elevated odds of NAFLD (OR 1.44, 1.95, and 2.36, respectively) (38). However, no significant association was observed between PHR categories and the odds of liver fibrosis (38). Currently, the association of LHR, MHR, and PHR with MASLD and liver fibrosis remains unknown. Our study fills this research gap by showing that LHR, MHR, NHR, and PHR were all positively associated with the prevalence of MASLD and significant liver fibrosis, independent of confounders. In addition, our findings suggest that these immunoinflammatory markers are positively associated with the odds of significant liver fibrosis, which is inconsistent with some previous findings in related studies. We speculate that this may be due to differences in population selection and covariate adjustment.

In addition, several observational studies have also shown significant associations between these markers and the prevalence of MetS (MASLD is proposed to be the hepatic manifestation of MetS). A large cross-sectional analysis from China showed that LHR and NHR were independent predictors of MetS in female participants (OR 3.671 and 1.728, respectively; P < 0.001), but not in male participants (39). Similarly, Chen et al demonstrated a significant positive correlation between LHR and MetS prevalence in a cross-sectional analysis from China (OR 4.117; P < 0.001) (40). Findings from the Northeast China Rural Cardiovascular Health Study suggested that higher LHR independently predicted newly diagnosed MetS (OR 1.57 for LHR in Q4 compared with Q1) (41). A recent cross-sectional investigation from Italy indicated that NHR, LHR, MHR, and PHR were all significantly correlated with and predictive of MetS prevalence, severity, and insulin resistance in adults with severe obesity, independent of other factors (42). Collectively, these novel markers may serve as independent predictors of MetS. Our findings suggest that these markers are similarly independently associated with MASLD and have potential predictive value. There is still a lack of information about the predictive value of these markers in MASLD. Only one study has explored the predictive ability of some of these markers for NAFLD and the AUC of the single marker were significantly lower than those in our study (34). Our study demonstrated for the first time that LHR/NHR and MHR/NHR were superior to other markers in predicting MASLD and significant hepatic fibrosis, respectively, suggesting their potential advantage as relevant predictors.

An interesting finding was that the positive association of LHR with significant liver fibrosis was more pronounced in men. The results from a previous Kurdish cohort study from Iran showed that the association between LHR and NAFLD was more significant in women (19). However, the association between LHR and hepatic fibrosis is still unexplored. Thus, the biological explanation for this variability remains unclear. A previous in vivo study showed that female mice demonstrated milder high-fat choline-deficient diet-induced steatohepatitis, exhibiting less liver inflammation and fibrosis (43). CD1d-deficient (lacking natural killer T cells [NKT]) male mice demonstrated more pronounced fibrosis and inflammation, whereas the absence of NKT cells had a negligible effect on steatohepatitis in female mice, suggesting a sex-specific role of NKT cells on the progression of metabolic dysfunction-associated steatohepatitis (43). Further experimental studies are needed to explore the underlying molecular biological mechanisms of these associations.

Infiltration and activation of immune cells may activate the innate and adaptive immune systems by secreting inflammatory cytokines, which in turn drive the development and progression of MASLD (9). Recruitment and activation of innate immune cells can directly drive hepatic steatosis, whereas reprogramming of the immunological landscape of conventional immune cells, including monocytes, B cells, T cells, and neutrophils, leads to an inflammatory microenvironment and causes hepatic fibrosis and hepatocellular injury (44–46). The pathophysiological role of platelets in MASLD is also being increasingly recognized in relation with the microthrombosis and procoagulant state induced by the inflammatory response (47). In addition, platelet activation may contribute to the immune-inflammatory response by releasing corresponding mediators that crosstalk with other immune cells (48). Hepatic and systemic lipid metabolism disorders are important pathogenesis in MASLD and induce subsequent mitochondrial dysfunction, dysregulated autophagy, oxidative stress, inflammation, and liver fibrosis. The role of HDL-C in MASLD is increasingly being investigated and may be a promising target (49).

Our study was based on the nationally representative NHANES database, which is characterized by a large sample and multiethnicity, thus enabling generalization of our study to other US adult populations. Hepatic steatosis was assessed by VCTE with good accuracy. In addition, we adequately corrected for confounders and reduced study bias. However, some limitations require attention and need to interpret our findings with caution. As this was a cross-sectional study, it was not possible to draw causal associations and unadjusted confounders remained. Furthermore, the absence of a prospective cohort study design precludes the possibility of ruling out a reverse causal association. This study only explored the association of these markers with MASLD and significant hepatic fibrosis. Owing to the lack of evaluation data, we were unable to explore their association with MASLD severity and other liver-related events. The assessment of hepatic steatosis uses imaging rather than liver biopsy. However, the invasiveness and potential operation-related risks associated with biopsy have limited its use in large epidemiological investigations of steatosis. By contrast, VCTE has demonstrated good accuracy.

CONCLUSIONS

LHR, MHR, NHR, and PHR were all positively associated with the prevalence of MASLD and significant liver fibrosis in US adults, mostly after the inflection point. The association of LHR with significant hepatic fibrosis was more pronounced in men. LHR/NHR had superior predictive ability for MASLD, whereas MHR/NHR distinguished significant liver fibrosis better than other markers. Our findings suggest that these novel markers may be independently associated with MASLD and significant hepatic fibrosis and used as predictors for clinical screening. However, future cohort studies are needed to validate these findings and elucidate the mechanisms involved through mechanistic studies.

CONFLICTS OF INTEREST

Specific author contributions: J.W. conducted and designed this work. Statistical analysis and manuscript preparation was performed by S.J. and X.Y. S.J. and Z.W. contributed to the accuracy of data analysis. Data collection and table preparation was conducted by Y.K. and X.Y. All authors reviewed the manuscript.

Financial support: This work is supported by the Beijing Hospitals Authority's Ascent Plan (DFL20221003) and the Sanming Project of Medicine in Shenzhen (No. SZSM202311023)

Potential competing interests: The authors declare no conflicts of interest, whether financial or personal, that could have influenced the work presented in this paper.

Ethics statement: All protocols were approved by the National Center for Health Statistics Ethics Review Board, and participants have provided written informed consent.

Data availability: This study analyzed publicly available datasets and can be found at https://wwwn.cdc.gov/nchs/nhanes/continuousnhanes/default.aspx?Cycle=2017-2020.

Supplementary Material

ct9-16-e00873-s001.docx ^{(52.7KB, docx)}