Zhang WH, Shi RDQ, Chen K, Li XM, Yu GH, Ye XY, Xu Q, Ding Y, Luo SH. Association between the high-sensitivity C-reactive protein-triglyceride-glucose index and metabolic dysfunction-associated steatotic liver disease across different glycemic states. World J Diabetes 2026; 17(7): 120647 [DOI: 10.4239/wjd.120647]
Corresponding Author of This Article
Si-Hui Luo, MD, Department of Endocrinology and Metabolism, The First Affiliated Hospital of University of Science and Technology of China, University of Science and Technology of China, No. 17 Lujiang Road, Hefei 230001, Anhui Province, China. luosihui@ustc.edu.cn
Research Domain of This Article
Endocrinology & Metabolism
Article-Type of This Article
research-article
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Baishideng Publishing Group Inc, 7041 Koll Center Parkway, Suite 160, Pleasanton, CA 94566, USA
Share the Article
Zhang WH, Shi RDQ, Chen K, Li XM, Yu GH, Ye XY, Xu Q, Ding Y, Luo SH. Association between the high-sensitivity C-reactive protein-triglyceride-glucose index and metabolic dysfunction-associated steatotic liver disease across different glycemic states. World J Diabetes 2026; 17(7): 120647 [DOI: 10.4239/wjd.120647]
World J Diabetes. Jul 15, 2026; 17(7): 120647 Published online Jul 15, 2026. doi: 10.4239/wjd.120647
Association between the high-sensitivity C-reactive protein-triglyceride-glucose index and metabolic dysfunction-associated steatotic liver disease across different glycemic states
Wen-Hao Zhang, Rong-Dong-Qing Shi, Xian-Ming Li, Gui-Hua Yu, Xin-Yuan Ye, Yu Ding, Si-Hui Luo, Department of Endocrinology and Metabolism, The First Affiliated Hospital of University of Science and Technology of China, University of Science and Technology of China, Hefei 230001, Anhui Province, China
Wen-Hao Zhang, Rong-Dong-Qing Shi, Xian-Ming Li, Gui-Hua Yu, Xin-Yuan Ye, Yu Ding, Si-Hui Luo, Anhui Provincial Key Laboratory of Metabolic Health and Panvascular Diseases, Hefei 230001, Anhui Province, China
Kui Chen, Health Management Medicine Center, The Third Xiangya Hospital, Central South University, Changsha 410013, Hunan Province, China
Kui Chen, State Key Laboratory of Ultrasound in Medicine and Engineering, College of Biomedical Engineering, Chongqing Medical University, Chongqing, 400016, China
Qing Xu, Graduate School, Bengbu Medical University, Bengbu 233030, Anhui Province, China
Co-first authors: Wen-Hao Zhang and Rong-Dong-Qing Shi.
Co-corresponding authors: Yu Ding and Si-Hui Luo.
Author contributions: Zhang WH and Shi RDQ contributed equally to the manuscript as co-first authors; Ding Y and Luo SH contributed equally to the manuscript as co-corresponding authors; Zhang WH contributed to study conceptualization, methodology, data analysis, data interpretation and visualization, wrote, and revised the manuscript; Shi RDQ reviewed the manuscript, and contributed to methodology, data analysis and data interpretation; Chen K contributed to data collection, reviewed and revised the manuscript; Li XM, Yu GH, Ye XY, and Xu Q provided insightful comments on the interpretation of the data, and the structure and revision of the manuscript; Ding Y contributed to study conceptualization, administrative support, data interpretation, reviewed and revised the manuscript; Luo SH contributed to study conceptualization, administrative support, fund acquisition, data interpretation, reviewed and revised the manuscript; and the corresponding authors attest that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.
AI contribution statement: No part of the main text of the manuscript or the response to reviewers was Al-generated. The study rationale, methodology, data analysis, results, interpretation, and conclusions were independently developed, written, reviewed, and approved by the authors. No AI tool was used for language polishing, translation, data analysis, or writing assistance. The authors take full responsibility for the accuracy, integrity, originality, and scientific validity of all content in the manuscript and the response to reviewers.
Supported by National Natural Science Foundation of China, No. 82300904 and No. 82100857.
Institutional review board statement: The studies were reviewed and approved by the Institutional Review Board of the First Affiliated Hospital of University of Science and Technology of China (Approval No: 2021 KY-034) and conformed to the guidelines of the Helsinki declaration and its later amendments or comparable ethical standards.
Informed consent statement: Written informed consent was obtained from all participants.
Conflict-of-interest statement: All authors declare that they have no conflicts of interest related to this work.
STROBE statement: The authors have read the STROBE Statement—checklist of items, and the manuscript was prepared and revised according to the STROBE Statement—checklist of items.
Data sharing statement: The data supporting this study are available from the corresponding author upon reasonable request, subject to privacy and ethical restrictions.
Corresponding author: Si-Hui Luo, MD, Department of Endocrinology and Metabolism, The First Affiliated Hospital of University of Science and Technology of China, University of Science and Technology of China, No. 17 Lujiang Road, Hefei 230001, Anhui Province, China. luosihui@ustc.edu.cn
Received: March 5, 2026 Revised: April 9, 2026 Accepted: June 9, 2026 Published online: July 15, 2026 Processing time: 127 Days and 2.1 Hours
Abstract
BACKGROUND
Metabolic dysfunction-associated steatotic liver disease (MASLD) is the most common chronic liver disease worldwide, but non-invasive biomarkers for its detection remain limited. The high-sensitivity C-reactive protein-triglyceride-glucose index (CTI), which integrates inflammatory and metabolic parameters, has emerged as a potential marker. However, its association with MASLD and diagnostic performance across different glycemic states remain unclear.
AIM
To investigate the association between CTI and MASLD and to evaluate its diagnostic performance across different glycemic states.
METHODS
This study included 7413 adults who underwent routine health examinations between January 2020 and March 2024. Participants were categorized into three groups according to glycemic status: Normal glucose regulation (NGR; n = 2933), prediabetes mellitus (pre-DM; n = 3549), and diabetes mellitus (DM; n = 931). CTI was calculated as 0.412 × ln(high-sensitivity C-reactive protein) + ln(triglycerides × fasting plasma glucose/2). The Boruta algorithm was applied to identify variables relevant to MASLD. Multivariable logistic regression assessed CTI-MASLD associations, restricted cubic spline analyses were performed to examine dose-response relationships, and diagnostic performance was evaluated using receiver operating characteristic curves.
RESULTS
After multivariable adjustment, each unit increase in CTI was significantly associated with MASLD in the NGR (adjusted OR = 3.07; 95%CI: 2.19-4.31) and pre-DM groups (adjusted OR = 3.26; 95%CI: 1.70-6.23), but not in the DM group (adjusted OR = 1.58; 95%CI: 0.65-3.84). CTI achieved an overall area under the curve (AUC) of 0.808 (95%CI: 0.798-0.817), with AUCs of 0.792, 0.783, and 0.749 in the NGR, pre-DM, and DM groups, respectively. Optimal CTI thresholds increased progressively with worsening glycemic status (NGR: 8.596; pre-DM: 8.908; DM: 9.456).
CONCLUSION
CTI was significantly associated with MASLD in normoglycemic and prediabetic populations but not in diabetic individuals, with optimal thresholds varying by glycemic status.
Core Tip: This large cross-sectional study (n = 7413) evaluated the high-sensitivity C-reactive protein-triglyceride-glucose index (CTI) as a non-invasive biomarker for metabolic dysfunction-associated steatotic liver disease (MASLD) across glycemic states. The CTI showed robust diagnostic performance (area under the curve: 0.808) with strongest predictive value in normal glucose regulation (adjusted OR = 3.07) and prediabetes (OR = 3.26), but weaker association in diabetes (OR = 1.58). Optimal CTI thresholds increased progressively from normoglycemia to diabetes (8.596, 8.908, 9.456, respectively). These findings suggest CTI is a practical biomarker for early MASLD risk identification, particularly in individuals without overt diabetes.
Citation: Zhang WH, Shi RDQ, Chen K, Li XM, Yu GH, Ye XY, Xu Q, Ding Y, Luo SH. Association between the high-sensitivity C-reactive protein-triglyceride-glucose index and metabolic dysfunction-associated steatotic liver disease across different glycemic states. World J Diabetes 2026; 17(7): 120647
Metabolic dysfunction-associated steatotic liver disease (MASLD), formerly known as non-alcoholic fatty liver disease (NAFLD), has emerged as the most prevalent chronic liver disease worldwide, affecting approximately 30% of the global population[1-4]. This metabolic liver disorder, characterized by hepatic fat accumulation without excessive alcohol consumption, represents a major public health challenge with substantial economic burden[5-7]. The recent name change from NAFLD to MASLD reflects advances in understanding disease mechanisms, highlighting central role to metabolic dysfunction rather than merely the absence of alcohol[6].
MASLD development involves complex interactions between metabolic disorders and chronic inflammation[8]. Insulin resistance (IR), a hallmark of metabolic dysfunction, promotes hepatic lipogenesis and reduces fatty acid oxidation, leading to fat accumulation in the liver[9,10]. Concurrently, adipose tissue dysfunction triggers the release of pro-inflammatory cytokines and adipokines, creating an environment for that promotes hepatic inflammation and fibrosis progression[11]. This inflammatory-metabolic interaction, together with gut dysbiosis[12], endoplasmic reticulum stress, and mitochondrial dysfunction, drives disease progression from simple steatosis to steatohepatitis and cirrhosis[13,14].
Current diagnostic approaches for MASLD rely primarily on imaging modalities or invasive liver biopsy, with limited non-invasive biomarkers available for general population-based screening[1]. The triglyceride-glucose index (TyG) effectively captures metabolic dysfunction[15] and predicts cardiovascular outcomes in MASLD patients[16], but lacks the inflammatory component important for disease progression. High-sensitivity C-reactive protein (hs-CRP) correlates with hepatic fibrosis in MASLD[17], but neither marker alone captures the complex inflammatory metabolic interactions in disease progression. The fibrosis-4 index (FIB-4) and NAFLD fibrosis score (NFS) were developed primarily to detect advanced fibrosis and therefore have limited ability to identify early-stage MASLD[18]. This diagnostic gap is particularly important because early intervention offers the greatest potential for disease reversal[19].
Given the limitations of single-dimension biomarkers, the hs-CRP-TyG (CTI) has recently been developed as a composite biomarker that combines both inflammatory (hs-CRP) and metabolic (TyG) components[20]. Previous studies have shown CTI’s value in predicting cardiovascular disease, stroke, and cancer outcomes[20-23]. Recent analysis revealed that CTI is associated with NAFLD and liver fibrosis, with diagnostic performance better than either TyG or hs-CRP alone[24]. Despite these promising findings, critical knowledge gaps remain. The diagnostic accuracy of CTI for MASLD across different glycemic states, normal glucose regulation (NGR), prediabetes mellitus (pre-DM), and diabetes mellitus (DM), has not been systematically evaluated. This is particularly important given the well-established association between MASLD and DM, and the likelihood that the relative contributions of metabolic vs inflammatory factors vary with glycemic status[25]. Understanding biomarker performance across the glycemic spectrum is essential for developing risk-stratified screening strategies.
In this study, we analyzed data from Chinese adults undergoing routine health examinations to investigate the association of CTI with MASLD and to assess its diagnostic performance across different glycemic states. Using the Boruta algorithm for feature selection, we examined the relevance of CTI among candidate biomarkers for MASLD. We further explored dose-response relationships across glycemic strata and evaluated whether CTI adds diagnostic information beyond existing non-invasive indices. These analyses may help clarify the role of integrated inflammatory-metabolic biomarkers in MASLD detection across varying metabolic conditions.
MATERIALS AND METHODS
Study design and population
This retrospective cross-sectional study was conducted using data from participants who underwent routine health examinations at the Health Management Center of the Third Xiangya Hospital, Central South University (Changsha, China) between January 2020 and March 2024. The participant screening process is illustrated in Figure 1. Initially, 24350 individuals who underwent routine health examinations were identified. Exclusion criteria were: (1) Missing baseline hs-CRP data (n = 15480); (2) Incomplete metabolic parameters including fasting plasma glucose (FPG), triglycerides (TG), or glycated haemoglobin (HbA1c; n = 13); (3) Age under 18 years or missing age data (n = 3); (4) History of viral hepatitis (n = 179); and (5) Excessive alcohol consumption (> 20 g/day for women and > 30 g/day for men; n = 1262). After applying these criteria, 7413 participants were included in the final analysis.
Figure 1 Study flowchart of participant selection.
FPG: Fasting plasma glucose; hs-CRP: High-sensitivity C-reactive protein; TG: Triglycerides; HbA1c: Glycated haemoglobin.
Data collection and measurements
Demographic characteristics, lifestyle factors, and medical history were collected using a standardized web-based questionnaire as previously described[26]. Data included sociodemographic characteristics (age, gender, education), lifestyle factors (smoking, alcohol consumption, physical activity), medical history (diabetes, hypertension, dyslipidemia), and current medications. Height and weight were measured using an automated anthropometric instrument with participants in light clothing without shoes. Body mass index (BMI) was calculated as weight/height2 (kg/m2). Waist circumference (WC) was measured at the midpoint between the lower rib and iliac crest. Hip circumference (HC) was measured at the maximum circumference of the buttocks. Blood pressure was measured using an Omron HEM-9020 automatic sphygmomanometer after 5 minutes of rest, with the average of three readings recorded[27].
Fasting blood samples were obtained in the morning after a 12-hour overnight fast. A range of biochemical parameters including FPG, fasting insulin (FINS), total cholesterol (TCHO), TG, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), alanine aminotransferase (ALT), aspartate aminotransferase (AST), serum creatinine (Cr), blood urea nitrogen (BUN), uric acid, total protein (TP), total bilirubin (TBIL), direct bilirubin (DBIL), albumin (ALB), HbA1c, and hs-CRP was measured using an automatic biochemical analyzer (Hitachi 7600; Hitachi, Tokyo, Japan)[27]. Complete blood count including platelet count (PLT) and white blood cell count (WBC) was performed using an automated hematology analyzer. The sample analysis was performed in accordance with the manufacturer's specifications. The estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease formula for Chinese subjects: EGFR = 175 × Cr (mg/dL)-1.234 × age (year)-0.179 × 0.79 (if female)[28].
Calculation of indices
The CTI was calculated as[20]: CTI = 0.412 × ln[hs-CRP (mg/L)] + ln{[TG (mg/dL) × FPG (mg/dL)]/2}. The TyG[29] was calculated as ln{[TG (mg/dL) × FPG (mg/dL)]/2}. Additional indices included waist to hip ratio (WHR)[30] = WC (cm)/HC (cm); body roundness index (BRI)[31] = 364.2 − 365.5 × √{1 − [WC (cm)/(2π)]²/[0.5 × height (cm)]²} and homeostatic model assessment of IR (HOMA-IR)[32] = [FINS (μU/mL) × FPG (mmol/L)]/22.5. Liver fibrosis indices were calculated as follows: FIB-4[1] = [age (year) × AST (U/L)]/[PLT (109/L) × √ALT (U/L); AST to platelet ratio index (APRI)[1] = [AST (U/L)/top normal AST (U/L)] × [100/PLT (109/L)]; NFS[1] = -1.675 + 0.037 × age (year) + 0.094 × BMI (kg/m2) + 1.13 × DM (no = 0, yes = 1) + 0.99 × AST/ALT - 0.013 × PLT (109/L) - 0.66 × ALB (g/dL).
Definitions
MASLD was diagnosed by hepatic steatosis on imaging plus at least one cardiometabolic risk factor[1]: (1) Overweight or obesity: BMI ≥ 23 kg/m2 (Chinese) , or WC ≥ 90 cm in men and ≥ 80 cm in women (Chinese); (2) Dysglycemia or type 2 diabetes: Pre-DM defined as HbA1c 5.7%-6.4% or FPG 5.6-6.9 mmol/L or 2-hour plasma glucose during oral glucose tolerance test (OGTT) 7.8-11 mmol/L, or type 2 diabetes defined as HbA1c ≥ 6.5% or FPG ≥ 7.0 mmol/L or 2-hour plasma glucose during OGTT ≥ 11.1 mmol/L or treatment for type 2 diabetes; (3) TG ≥ 1.7 mmol/L or lipid-lowering treatment; (4) HDL-C ≤ 1.0 mmol/L in men and ≤ 1.3 mmol/L in women or lipid-lowering treatment; and (5) Blood pressure ≥ 130/85 mmHg or treatment for hypertension.
Hypertension was defined as blood pressure ≥ 140/90 mmHg, a reported history of hypertension, or antihypertensive medication use[33]. DM was defined as FPG ≥ 7.0 mmol/L, HbA1c ≥ 6.5%, antidiabetic medication use, or a reported diabetes history. Pre-DM was FPG 5.6-6.9 mmol/L or HbA1c 5.7%-6.4%. NGR was FPG < 5.6 mmol/L and HbA1c < 5.7%[34]. Hypertriglyceridemia was defined as TG ≥ 1.7 mmol/L or use of lipid-lowering medications. Low HDL-C cholesterol was defined as serum levels of < 1.0 mmol/L. Obesity was defined as BMI ≥ 28 kg/m2 for general obesity or WC ≥ 90 cm (men) or ≥ 85 cm (women) for central obesity according to the Chinese criteria[35]. Current smoking was defined as ≥ 1 cigarette/day for > 6 months[27]. Current drinking was alcohol consumption ≥ 2 days/week[27]. Regular exercise was physical activity > 3 times/week for ≥ 30 minutes[27].
Statistical analysis
Continuous variables were expressed as mean ± SD (normal distribution) or median (IQR; non-normal distribution) and compared using ANOVA or Kruskal-Wallis test, respectively. Categorical variables were presented as n (%) and compared using the χ2 test. Participants were stratified into CTI quartiles: Q1 (< 8.30), Q2 (8.30-8.88), Q3 (8.88-9.46), and Q4 (> 9.46). To identify variables relevant of MASLD, we employed the Boruta algorithm (500 iterations) for feature selection. This analysis was conducted for the overall population and stratified by glycemic status (NGR, pre-DM, DM) to examine whether feature importance varied across metabolic states[36]. Variables were classified as confirmed, rejected, or tentative based on Z-score comparisons with shadow features[36].
To determine appropriate confounders and minimize overadjustment bias, a directed acyclic graph (DAG) was constructed based on prior knowledge and published evidence[37]. Informed by the DAG, we examined the CTI-MASLD association using multivariable logistic regression with three models: Model 1 (unadjusted); Model 2 (adjusted for age, sex, current smoking, current drinking, regular exercise); Model 3 (Model 2 plus systolic blood pressure, diastolic blood pressure, HC, AST, TP, TBIL, DBIL, TCHO, HDL-C, LDL-C, PLT, ALB, HbA1c, eGFR, BUN, history of hypertension, history of hypertriglyceridemia, history of low HDL-C). Both per-unit increase and quartile analyses were performed, with trend P values calculated. Dose-response relationships were further characterized using restricted cubic splines (RCS) with three knots. Diagnostic performance was evaluated using receiver operating characteristic curves, with area under the curve (AUC) and 95%CI calculated for CTI, hs-CRP, TyG, liver fibrosis indices (FIB-4, NFS, and APRI), liver enzymes (ALT and AST), anthropometric indices (BMI, WC, HC WHR, and BRI), and conventional risk factors. The incremental value of adding CTI was assessed, with optimal cut-offs determined using the Youden index and AUCs compared using DeLong’s test. Effect modification was assessed through subgroup analyses stratified by age (< 60 years vs ≥ 60 years), sex, lifestyle factors (current smoking, current drinking, regular exercise), BMI categories (< 24 kg/m2, 24-27.9 kg/m2, ≥ 28 kg/m2), and medical history (history of hypertension, history of hyperlipidemia), with interaction P-values calculated using likelihood ratio tests. All primary analyses were conducted in the overall population and stratified by glycemic status.
To test the robustness of our findings, we performed five sensitivity analyses: (1) Excluding current smokers; (2) Excluding current drinkers; (3) Excluding individuals with obesity (BMI ≥ 28 kg/m2); (4) Excluding participants with acute inflammation (WBC > 10 × 109/L); and (5) Excluding those using lipid-lowering medications. Finally, we calculated the E-value to quantify the minimum strength of association that an unmeasured confounder would require with both CTI and MASLD risk to explain away the observed associations. The calculation method of E-value was as follows: E = risk ratio (RR) + sqrt [RR × (RR - 1)][38]. All analyses were performed using R version 4.3.0 with two-sided P < 0.05 considered significant.
RESULTS
Baseline characteristics of the study population
Among 7413 participants, the mean age was 50.0 ± 10.9 years and 56.1% were male. Table 1 shows baseline characteristics by CTI quartiles (Q1: < 8.30, Q2: 8.30-8.88, Q3: 8.88-9.46, Q4: > 9.46). Higher CTI values were associated with older age, men predominance (34.1% in Q1 vs 77.0% in Q4, P < 0.001), and unfavorable lifestyle factors including higher smoking (8.57% vs 28.6%) and drinking rates (11.2% vs 24.1%), with lower physical activity (50.8% vs 46.4%, all P < 0.05). Cardiometabolic risk factors increased across CTI quartiles. BMI increased from 22.4 ± 2.83 kg/m2 to 26.6 ± 3.16 kg/m2, WC from 76.5 ± 8.69 cm to 90.7 ± 8.64 cm, and SBP from 118 ± 15.5 mmHg to 129 ± 16.5 mmHg (all P < 0.001). TG levels increased 3.8-fold (median: 0.81 mmol/L vs 3.11 mmol/L), while HDL-C decreased by 25% (1.50 ± 0.31 mmol/L vs 1.12 ± 0.23 mmol/L). FPG increased from 5.04 ± 0.57 mmol/L to 6.50 ± 2.30 mmol/L, and hs-CRP showed 3.9-fold elevation (median: 0.59 mg/L vs 2.32 mg/L, all P < 0.001). HOMA-IR and FINS approximately doubled from Q1 to Q4. DM prevalence was nearly 10-fold higher in Q4 compared to Q1 (29.5% vs 3.18%), while hypertriglyceridemia increased from 2.10% to 94.5% (P < 0.001). MASLD prevalence showed stepwise increases: 19.1%, 44.1%, 67.9%, and 87.4% across quartiles (P for trend < 0.001).
Table 1 Patient demographics and baseline characteristics stratified by C-reactive protein-triglyceride-glucose index quartiles.
Feature selection by the Boruta algorithm across glycemic states
The Boruta algorithm (500 iterations) identified 34 features significantly associated with MASLD, with all features achieving stable classification before convergence (Figure 2; Supplementary Figures 1-4). Anthropometric indices dominated with the highest importance: BMI, BRI, and WC (Z-scores > 30), followed by CTI (fourth) and WHR (fifth). Metabolic markers including TyG, TG, HOMA-IR, and FINS showed Z-scores of 20-30, while liver enzymes (ALT and AST) had Z-scores of 10-25.
Figure 2 Feature importance analysis using the Boruta algorithm across glycemic states.
A-D: Boruta algorithm results showing feature importance (Z-scores) for metabolic dysfunction-associated steatotic liver disease identification across different glycemic states: Total population (A), normal glucose regulation (B), prediabetes mellitus (C), and diabetes mellitus (D). NGR: Normal glucose regulation; Pre-DM: Prediabetes mellitus; DM: Diabetes mellitus; BMI: Body mass index; BRI: Body roundness index; WC: Waist circumference; CTI: C-reactive protein-triglyceride-glucose index; WHR: Waist to hip ratio; TyG: Triglyceride-glucose index; TG: Triglycerides; HOMA-IR: Homeostatic model assessment of insulin resistance; FINS: Fasting insulin; ALT: Alanine aminotransferase; HC: Hip circumference; HDL: High-density lipoprotein; UA: Uric acid; HbA1c: Glycated haemoglobin; HTG: Hypertriglyceridemia; FPG: Fasting plasma glucose; hs-CRP: High-sensitivity C-reactive protein; DBP: Diastolic blood pressure; SBP: Systolic blood pressure; AST: Aspartate aminotransferase; FIB-4: Fibrosis-4 index; Cr: Serum creatinine; APRI: Aspartate aminotransferase to platelet ratio index; NFS: Non-alcoholic fatty liver disease fibrosis score; ALB: Albumin; TP: Total protein; TCHO: Total cholesterol; PLT: Platelet count; eGFR: Estimated glomerular filtration rate; DBIL: Direct bilirubin; LDL: Low-density lipoprotein; TBIL: Total bilirubin; BUN: Blood urea nitrogen; HTN: Hypertension.
In NGR (n = 2933), 27 features were confirmed with anthropometric measures (BMI, BRI, WC, and WHR) dominating (Z-scores > 20). CTI ranked fifth, showing higher importance than standalone hs-CRP. In pre-DM (n = 3549), 29 features were confirmed with CTI rising to fourth position, coinciding with increased importance of metabolic indices (TyG, TG, HOMA-IR, FINS; Z-scores > 15). In DM (n = 931), 22 features were confirmed. TyG and TG surpassed CTI, which dropped to seventh position (Z-score > 10). Notably, hs-CRP was rejected while CTI remained confirmed.
Association between CTI and MASLD
Logistic regression revealed dose-dependent CTI-MASLD associations (Table 2). Each unit increase in CTI was associated with 5.26-fold increased MASLD odds (95%CI: 4.83-5.73, P < 0.001) in unadjusted analysis, with quartile ORs of 3.34, 8.98, and 29.5 for Q2-Q4 vs Q1. After full adjustment (Model 3) based on the DAG (Supplementary Figure 5), per-unit OR = 2.74 (95%CI: 1.79-4.20) with quartile OR = 2.59, 3.34, and 5.09, respectively (P for trend < 0.001).
Table 2 Logistic regression analyses the association between C-reactive protein-triglyceride-glucose index and metabolic dysfunction-associated steatotic liver disease.
Stratified analyses examined CTI-MASLD associations across glycemic states (Table 3). In NGR (n = 2933), adjusted per-unit OR = 3.07 (95%CI: 2.19-4.31, P < 0.001) with quartile OR = 2.62, 3.40, and 5.15, respectively for Q2-Q4 (P for trend < 0.001). Pre-DM (n = 3549) showed similar per-unit associations (OR = 3.26, 95%CI: 1.70-6.23) with the highest quartile OR = 5.57 (95%CI: 1.58-19.68). In DM (n = 931), associations were attenuated, adjustment for metabolic factors eliminated significance (Model 3 per-unit OR = 1.58, 95%CI: 0.65-3.84).
Table 3 Logistic regression analyses the association between C-reactive protein-triglyceride-glucose index and metabolic dysfunction-associated steatotic liver disease in different glycemic status.
RCS analyses characterized CTI-MASLD relationships across glycemic states (Figure 3). The overall population showed a linear relationship without threshold effects (P for non-linearity = 0.308; reference point = 9.48). This linearity persisted in NGR (P for non-linearity = 0.498), pre-DM (P for non-linearity = 0.625), and DM (P for non-linearity = 0.207), with reference points of 10.74, 9.32, and 11.61, respectively.
Figure 3 Dose-response relationships between C-reactive protein-triglyceride-glucose index and metabolic dysfunction–associated steatotic liver disease using restricted cubic splines.
A-D: Restricted cubic splines curves depicting the dose-response relationship between C-reactive protein-triglyceride-glucose index and metabolic dysfunction-associated steatotic liver disease odds ratios across glycemic states: Total population (A), normal glucose regulation (B), prediabetes mellitus (C), and diabetes mellitus (D). Adjusted for age, sex, current smoking, current drinking, regular exercise, systolic blood pressure, diastolic blood pressure, hip circumference, aspartate aminotransferase, total protein, total bilirubin, direct bilirubin, total cholesterol, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol, platelet count, albumin, glycated haemoglobin, estimated glomerular filtration rate, blood urea nitrogen, history of hypertension, history of hypertriglyceridemia, history of low HDL-C. MASLD: Metabolic dysfunction-associated steatotic liver disease; NGR: Normal glucose regulation; Pre-DM: Prediabetes mellitus; DM: Diabetes mellitus.
Diagnostic performance of CTI for MASLD
CTI demonstrated good diagnostic performance with an overall AUC of 0.808 (95%CI: 0.798-0.817) at an optimal cut-off 8.908 (Table 4). Performance remained stable across glycemic states (Table 5): AUCs of 0.792 (NGR), 0.783 (pre-DM), and 0.749 (DM), with progressively higher cut-offs (8.596, 8.908, 9.456).
Table 4 Receiver operating characteristic analysis of biomarkers and risk scores for metabolic dysfunction-associated steatotic liver disease prediction.
Table 5 Receiver operating characteristic curve analysis of biomarkers and risk scores for metabolic dysfunction-associated steatotic liver disease prediction in different glycemic status.
CTI outperformed individual components and traditional markers. Compared to hs-CRP (AUC = 0.662), CTI showed superior discrimination (P < 0.001), with the gap widening in diabetes (0.615 vs 0.749). Traditional fibrosis indices (FIB-4, NFS, APRI) achieved AUCs below 0.65. While TyG showed comparable discrimination, CTI provided more balanced sensitivity specificity profiles.
Anthropometric indices achieved the highest discrimination but showed declining performance with metabolic progression (AUC from 0.887 in NGR to 0.790 in DM). Adding CTI yielded consistent improvements, with sensitivity increasing from 70.1% to 79.9% in diabetes. In sex-stratified analyses (Supplementary Tables 1 and 2), CTI showed acceptable diagnostic performance in both men and women across glycemic states.
Subgroup analyses
Subgroup analyses examined CTI-MASLD associations across populations (Figure 4). Younger participants (< 60 years) showed stronger associations than older adults (OR = 3.00, 95%CI: 2.57-3.50 vs OR = 1.63, 95%CI: 1.22-2.19; P for interaction < 0.001). BMI showed significant interaction (P < 0.001): Robust effects in normal weight (OR = 2.82, 95%CI: 2.28-3.50) and overweight (OR = 2.40, 95%CI: 1.98-2.92), but marked attenuation in obesity (OR = 1.28, 95%CI: 0.7-2.33). Sex, lifestyle factors, and metabolic comorbidities showed no effect modification (all P for interaction > 0.05).
Five sensitivity analyses supported CTI-MASLD association robustness. Excluding smokers (n = 2962) or drinkers (n = 2965) yielded adjusted ORs of 3.07 and 3.19, with maintained significance in NGR and pre-DM but attenuation in DM (Supplementary Tables 3-6). Removing individuals with obesity (BMI ≥ 28 kg/m2, n = 6320) preserved associations (OR = 3.16, 95%CI: 2.56-3.91; Supplementary Tables 7 and 8). Excluding acute inflammation (WBC > 10 × 109/L, n = 7304) minimally affected results (OR = 3.32, 95%CI: 2.71-4.06; Supplementary Tables 9 and 10). Among participants without lipid-lowering medication (n = 772), associations were stronger in NGR (OR = 5.79, 95%CI: 1.92-17.51; Supplementary Tables 11 and 12). The E-value of 4.92 indicates substantial unmeasured confounding would be required to explain observed relationships.
DISCUSSION
This study evaluated the association between CTI and MASLD in 7413 Chinese adults. We found that CTI, which combines inflammatory and metabolic parameters, showed good diagnostic performance across different glycemic states, with stronger associations in individuals with NGR and pre-DM. This differential performance across glycemic states may reflect the varying contributions of inflammation and metabolic dysfunction during MASLD development.
Our analysis revealed several important findings. CTI showed significant associations with MASLD across all glycemic states, with stronger discriminative ability in individuals with NGR and pre-DM. The diagnostic performance of CTI (AUC = 0.808) was comparable to established metabolic indices (TyG, AUC = 0.802) and better than inflammatory markers (hs-CRP, AUC = 0.662) alone. The optimal CTI cut-off values increased with worsening glycemic status (8.596 for NGR, 8.908 for pre-DM, 9.456 for diabetes), reflecting the metabolic continuum in MASLD development. The progressive increase in MASLD prevalence across CTI quartiles (19.1%-87.4%) demonstrates the index’s ability to stratify disease likelihood. This relationship is consistent with recent evidence[17,24] showing that composite indices incorporating both metabolic and inflammatory parameters may better capture MASLD status than single markers. The persistence of significant associations after adjustment for multiple confounders (OR = 2.74, 95%CI: 1.79-4.20) suggests that CTI captures disease information beyond conventional covariates.
The better performance of CTI compared to its individual components provides insights into MASLD development. While IR promotes liver lipid accumulation through increased lipogenesis and reduced fatty acid oxidation, inflammation contributes to liver injury through various pathways[9,10]. The combination of these processes in CTI likely explains its improved diagnostic capability. This is consistent with evidence that both the TyG and hs-CRP are correlates of metabolic diseases, supporting the value of combined assessment. Our finding that CTI outperformed hs-CRP alone across all glycemic states (AUC = 0.808 vs 0.662, P < 0.001) indicates that inflammation assessment alone provides insufficient discrimination for MASLD[17]. This observation aligns with the current understanding that metabolic dysfunction remains the primary driver of hepatic steatosis, while inflammation mainly influences disease progression.
The differential performance of CTI across glycemic states provides insights into MASLD heterogeneity. In NGR individuals, where traditional metabolic markers may be less sensitive, CTI maintained good discriminative ability (adjusted OR = 3.07, 95%CI: 2.19-4.31). This suggests that subtle inflammatory-metabolic changes may already be present before overt metabolic dysfunction becomes apparent, supporting CTI’s utility in identifying early-stage disease. The Boruta analysis showed that CTI was the fifth most important feature in NGR, while pure metabolic indices ranked lower, supporting this interpretation. The weaker associations in diabetes (adjusted OR = 1.58, 95%CI: 0.65-3.84) likely reflect the dominant contribution of established metabolic dysfunction in this population, compounded by the smaller sample size in the diabetes subgroup (n = 931 vs 2933 and 3549). This finding is similar to observations from recent studies showing that CTI associations with cardiovascular outcomes were also attenuated in diabetes[23]. The flattened dose-response curves in diabetes suggest a ceiling effect where additional inflammatory-metabolic information provides limited incremental value beyond the already elevated metabolic burden.
The recent change from NAFLD to MASLD emphasizes metabolic dysfunction as central to disease definition[6]. Our findings support this change, as CTI captures metabolic parameters while adding inflammatory assessment. The diagnostic performance of CTI (AUC = 0.808, 95%CI: 0.798-0.817) is comparable to other non-invasive indices reported in the literature[39]. CTI addresses a gap in current non-invasive assessment tools. While FIB-4, NFS and APRI were designed mainly for advanced fibrosis detection[39] and showed poor performance for MASLD identification in our study (AUC < 0.65), CTI specifically targets the metabolic-inflammatory phenotype characteristic of MASLD. However, adding CTI to anthropometric indices (BMI, BRI) yielded only marginal AUC improvements (0.002-0.004), as these measures already achieved AUCs of 0.874-0.887. The value of CTI therefore lies not in surpassing anthropometric assessment, but in its derivation from routine laboratory parameters, making it applicable in large-scale screening settings where standardized anthropometric measurements are unavailable or impractical. This distinction is important as non-invasive techniques have improved early detection and monitoring of MASLD, and CTI could complement these methods in comprehensive assessment.
Our findings extend previous work on CTI in several ways. While earlier studies demonstrated CTI’s value in cardiovascular and cancer risk assessment[20,21,23,40,41], this is among the first comprehensive evaluations in MASLD across the glycemic spectrum. A recent meta-analysis found that each unit increase in the TyG index was associated with higher NAFLD prevalence (OR = 2.36, 95%CI: 1.88-2.97)[42]. Our observed association for CTI (unadjusted OR = 5.26, 95%CI: 4.83-5.73) suggests that incorporating inflammation improves disease discrimination. The diagnostic accuracy in our study is broadly comparable to that reported in a recent meta-analysis on TyG-related indices (AUC = 0.75, 95%CI: 0.71–0.79), although cross-study comparisons should be interpreted with caution given differences in populations and diagnostic criteria[43].
This study has several strengths, including the comprehensive metabolic phenotyping, and robust analytical approach using machine learning based feature selection. A DAG-informed covariate selection strategy was employed to minimize overadjustment, and the consistency of results across adjustment levels supports the robustness of the observed associations. The stratified analysis by glycemic status provides new insights into CTI’s differential utility across the metabolic spectrum. The extensive sensitivity analyses support the robustness of our findings. However, several limitations should be considered. First, the cross-sectional design limits causal inference. Longitudinal studies are needed to evaluate whether CTI changes are associated with incident MASLD or progression to advanced fibrosis. Second, MASLD diagnosis relied on abdominal ultrasonography, which has limited sensitivity for mild steatosis[44]. This may have led to misclassification, particularly in the NGR group where hepatic fat content might be lower. Third, our study population was Chinese, and validation in other ethnic groups is necessary given known ethnic variations in MASLD susceptibility and metabolic profiles. The lack of histological confirmation represents another limitation, as we could not assess CTI’s performance for detecting metabolic dysfunction-associated steatohepatitis or staging fibrosis. Future studies incorporating liver biopsy or advanced imaging, prospective follow up, and diverse populations could further clarify CTI’s utility for these endpoints.
CONCLUSION
In conclusion, CTI showed significant cross-sectional associations with MASLD across glycemic states, combining inflammatory and metabolic parameters. Its good diagnostic performance in NGR and pre-DM populations supports potential clinical utility. While CTI cannot replace comprehensive clinical evaluation or advanced imaging, its derivation from routine laboratory parameters makes it a potential complementary tool for MASLD identification. Future prospective studies evaluating whether CTI is associated with disease progression and treatment response will be important for establishing its role in clinical practice.
ACKNOWLEDGEMENTS
We thank all the participants for donating the data.
European Association for the Study of the Liver (EASL); European Association for the Study of Diabetes (EASD); European Association for the Study of Obesity (EASO). EASL-EASD-EASO Clinical Practice Guidelines on the management of metabolic dysfunction-associated steatotic liver disease (MASLD).J Hepatol. 2024;81:492-542.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 1461][Cited by in RCA: 1320][Article Influence: 660.0][Reference Citation Analysis (6)]
Quek J, Chan KE, Wong ZY, Tan C, Tan B, Lim WH, Tan DJH, Tang ASP, Tay P, Xiao J, Yong JN, Zeng RW, Chew NWS, Nah B, Kulkarni A, Siddiqui MS, Dan YY, Wong VW, Sanyal AJ, Noureddin M, Muthiah M, Ng CH. Global prevalence of non-alcoholic fatty liver disease and non-alcoholic steatohepatitis in the overweight and obese population: a systematic review and meta-analysis.Lancet Gastroenterol Hepatol. 2023;8:20-30.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 563][Cited by in RCA: 506][Article Influence: 168.7][Reference Citation Analysis (2)]
Karlsen TH, Sheron N, Zelber-Sagi S, Carrieri P, Dusheiko G, Bugianesi E, Pryke R, Hutchinson SJ, Sangro B, Martin NK, Cecchini M, Dirac MA, Belloni A, Serra-Burriel M, Ponsioen CY, Sheena B, Lerouge A, Devaux M, Scott N, Hellard M, Verkade HJ, Sturm E, Marchesini G, Yki-Järvinen H, Byrne CD, Targher G, Tur-Sinai A, Barrett D, Ninburg M, Reic T, Taylor A, Rhodes T, Treloar C, Petersen C, Schramm C, Flisiak R, Simonova MY, Pares A, Johnson P, Cucchetti A, Graupera I, Lionis C, Pose E, Fabrellas N, Ma AT, Mendive JM, Mazzaferro V, Rutter H, Cortez-Pinto H, Kelly D, Burton R, Lazarus JV, Ginès P, Buti M, Newsome PN, Burra P, Manns MP. The EASL-Lancet Liver Commission: protecting the next generation of Europeans against liver disease complications and premature mortality.Lancet. 2022;399:61-116.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 542][Cited by in RCA: 489][Article Influence: 122.3][Reference Citation Analysis (5)]
Rinella ME, Lazarus JV, Ratziu V, Francque SM, Sanyal AJ, Kanwal F, Romero D, Abdelmalek MF, Anstee QM, Arab JP, Arrese M, Bataller R, Beuers U, Boursier J, Bugianesi E, Byrne CD, Castro Narro GE, Chowdhury A, Cortez-Pinto H, Cryer DR, Cusi K, El-Kassas M, Klein S, Eskridge W, Fan J, Gawrieh S, Guy CD, Harrison SA, Kim SU, Koot BG, Korenjak M, Kowdley KV, Lacaille F, Loomba R, Mitchell-Thain R, Morgan TR, Powell EE, Roden M, Romero-Gómez M, Silva M, Singh SP, Sookoian SC, Spearman CW, Tiniakos D, Valenti L, Vos MB, Wong VW, Xanthakos S, Yilmaz Y, Younossi Z, Hobbs A, Villota-Rivas M, Newsome PN; NAFLD Nomenclature consensus group. A multisociety Delphi consensus statement on new fatty liver disease nomenclature.J Hepatol. 2023;79:1542-1556.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 2209][Cited by in RCA: 1961][Article Influence: 653.7][Reference Citation Analysis (0)]
Mózes FE, Lee JA, Selvaraj EA, Jayaswal ANA, Trauner M, Boursier J, Fournier C, Staufer K, Stauber RE, Bugianesi E, Younes R, Gaia S, Lupșor-Platon M, Petta S, Shima T, Okanoue T, Mahadeva S, Chan WK, Eddowes PJ, Hirschfield GM, Newsome PN, Wong VW, de Ledinghen V, Fan J, Shen F, Cobbold JF, Sumida Y, Okajima A, Schattenberg JM, Labenz C, Kim W, Lee MS, Wiegand J, Karlas T, Yılmaz Y, Aithal GP, Palaniyappan N, Cassinotto C, Aggarwal S, Garg H, Ooi GJ, Nakajima A, Yoneda M, Ziol M, Barget N, Geier A, Tuthill T, Brosnan MJ, Anstee QM, Neubauer S, Harrison SA, Bossuyt PM, Pavlides M; LITMUS Investigators. Diagnostic accuracy of non-invasive tests for advanced fibrosis in patients with NAFLD: an individual patient data meta-analysis.Gut. 2022;71:1006-1019.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in Crossref: 389][Cited by in RCA: 375][Article Influence: 93.8][Reference Citation Analysis (6)]
Williams B, Mancia G, Spiering W, Agabiti Rosei E, Azizi M, Burnier M, Clement DL, Coca A, de Simone G, Dominiczak A, Kahan T, Mahfoud F, Redon J, Ruilope L, Zanchetti A, Kerins M, Kjeldsen SE, Kreutz R, Laurent S, Lip GYH, McManus R, Narkiewicz K, Ruschitzka F, Schmieder RE, Shlyakhto E, Tsioufis C, Aboyans V, Desormais I; ESC Scientific Document Group. 2018 ESC/ESH Guidelines for the management of arterial hypertension.Eur Heart J. 2018;39:3021-3104.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 8066][Cited by in RCA: 6720][Article Influence: 840.0][Reference Citation Analysis (3)]
ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, Collins BS, Hilliard ME, Isaacs D, Johnson EL, Kahan S, Khunti K, Leon J, Lyons SK, Perry ML, Prahalad P, Pratley RE, Seley JJ, Stanton RC, Gabbay RA; on behalf of the American Diabetes Association. 2. Classification and Diagnosis of Diabetes: Standards of Care in Diabetes-2023.Diabetes Care. 2023;46:S19-S40.
[RCA] [PubMed] [DOI] [Full Text][Cited by in Crossref: 2057][Cited by in RCA: 1731][Article Influence: 577.0][Reference Citation Analysis (6)]
Jin XF, Su ZJ, Zhong JF, Wang J, Shi BJ, Liu ZY, Li PS, Xiong X, Xu SQ, Qiu C, Xiao Y, Li SB, Tong WH, Ge JP, Li HX, Liu XY, Li ZZ, Kong XR, Chen JL, Wen XJ, Sun P, Chen PP, Li YY, Chen Y, Mao LT, Peng ZH, Li H, Zhang T. Association of the C-reactive protein-triglyceride-glucose index with metabolic dysfunction-associated steatotic liver disease and long-term all-cause and cardiovascular mortality: evidence from two nationwide prospective cohort studies.Cardiovasc Diabetol. 2026;25:75.
[RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)][Cited by in RCA: 3][Reference Citation Analysis (0)]
Creativity or innovation: Grade B, Grade B, Grade B, Grade B
Scientific significance: Grade B, Grade B, Grade B, Grade B
P-Reviewer: Dabla PK, MD, Chief Physician, Professor, India; He J, MD, PhD, Associate Research Scientist, China; Priego Parra BA, MD, PhD, Assistant Professor, Mexico S-Editor: Lin C L-Editor:P-Editor: Yang YQ
