{
  "abstract": "Introduction Type 2 diabetes mellitus (T2DM) and metabolic dysfunction-associated steatotic liver disease (MASLD) frequently coexist in the same individual. Dyslipidemia plays a central role in the pathogenesis of MASLD. Remnant cholesterol (RC) is the cholesterol content carried in triglyceride-rich lipoproteins. We aimed to examine the association of RC, its discordance with low-density lipoprotein cholesterol (LDL-C) in terms of MASLD risk in patients with T2DM.Research design and methods A total of 26 302 T2DM patients who were hospitalized were cross-sectionally assessed. Logistic regression models were employed to assess the associations. We used three approaches to assess the effects of the discordance between RC and LDL-C on MASLD risk: clinical cut-off points, differences of >10 percentile points and residuals.Results The median age was 62.0 years, and 48.3% had MASLD. RC was more closely associated with MASLD than LDL-C. Furthermore, discordantly high RC was associated with higher MASLD risk than discordantly high LDL-C, regardless of which method of discordance analysis is used. For example, for the residual approach, in a regression model including LDL-C and RC residual, the latter, representing the portion of RC not explained by LDL-C, was more associated with MASLD risk than LDL-C (OR=1.26, 95% CI 1.22 to 1.30 for RC residual and OR=1.04, 95% CI 1.01 to 1.07 for LDL-C); in a regression model including RC and LDL-C residual, the latter representing the portion of LDL-C not explained by RC, RC was more associated with MASLD risk than LDL-C residual (OR=1.24, 95% CI 1.20 to 1.29 for RC and OR=1.09, 95% CI 1.06 to 1.12 for LDL-C residual).Conclusions Discordantly high RC was associated with higher MASLD risk than discordantly high LDL-C. Our findings suggest that RC may serve as a potential target for prevention and intervention for MASLD.",
  "authors": [
    {
      "affiliations": [
        "Department of Endocrinology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, People's Republic of China",
        "Hubei Provincial Clinical Medical Research Center for Endocrinology and Metabolic Diseases, Wuhan, People's Republic of China"
      ],
      "name": "Ying Zhao"
    },
    {
      "affiliations": [
        "Department of Endocrinology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, People's Republic of China",
        "Hubei Provincial Clinical Medical Research Center for Endocrinology and Metabolic Diseases, Wuhan, People's Republic of China"
      ],
      "name": "Zhaoyue Li"
    },
    {
      "affiliations": [
        "Department of Endocrinology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, People's Republic of China",
        "Hubei Provincial Clinical Medical Research Center for Endocrinology and Metabolic Diseases, Wuhan, People's Republic of China"
      ],
      "name": "Tian Yu"
    },
    {
      "affiliations": [
        "Department of Endocrinology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, People's Republic of China",
        "Hubei Provincial Clinical Medical Research Center for Endocrinology and Metabolic Diseases, Wuhan, People's Republic of China"
      ],
      "name": "Yan Yang"
    },
    {
      "affiliations": [
        "Department of Endocrinology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, People's Republic of China",
        "Hubei Provincial Clinical Medical Research Center for Endocrinology and Metabolic Diseases, Wuhan, People's Republic of China"
      ],
      "name": "Tingting Du"
    }
  ],
  "full_text": "WHAT IS ALREADY KNOWN ON THIS TOPIC RC is more closely associated with metabolic dysfunction-associated steatotic liver disease (MASLD) risk than low-density lipoprotein cholesterol (LDL-C), and dyslipidemia in type 2 diabetes mellitus (T2DM) plays a central role in the pathogenesis of MASLD. However, data on the association of remnant cholesterol (RC), its discordance/concordance with LDL-C and MASLD risk in T2DM are limited.WHAT THIS STUDY ADDS RC was positively associated with MASLD risk independent of traditional cardiovascular risk factors. Discordantly high RC was associated with a higher risk of MASLD than discordantly high LDL-C, even in T2DM patients with optimal control of LDL-C and triglyceride levels.HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY RC may serve as a potential target for prevention and intervention for MASLD.Introduction Type 2 diabetes mellitus (T2DM) is a major global public health concern, affecting 10.5% of the global adult population. 1 It often coexists with metabolic dysfunction-associated steatotic liver disease (MASLD) in the same individual in a bi-directional relationship.2 3 A global prevalence of 65.04% of patients with T2DM suffered from MASLD.4 The two diseases can aggravate each other’s disease progression. For example, T2DM is a major driver of the progression of MASLD to MASH, advanced fibrosis and cirrhosis.5 An updated meta-analysis reported that among those with T2DM and MASLD, 35.54% had clinically significant fibrosis (F2–F4) and 14.95% had advanced fibrosis (F3–F4)4; vice versa, T2DM patients with MASLD have worse glycemic control and develop diabetic-related complications more rapidly than those without MASLD.6 7Dyslipidemia in T2DM plays a central role in the pathogenesis of MASLD.8 Among the spectrum of dyslipidemia, there is no doubt that increased low-density lipoprotein cholesterol (LDL-C) is the most important one.9–11 However, T2DM patients with a substantial reduction in LDL-C levels or even with optimal LDL-C levels as recommended by guidelines still encounter considerable MASLD burden.12 13 Recent research conducted in the general population indicates that the risk of MASLD is increased in individuals with elevated remnant cholesterol (RC).14 15 Insulin resistance and hyperglycemia can markedly increase the RC level in patients with T2DM.8 11 Furthermore, studies of the general population showed that RC correlated more closely with MASLD risk than LDL-C.16 Therefore, it is highly likely that elevated RC can contribute to the excess risk of MASLD in T2DM patients. However, whether RC can affect MASLD risk in T2DM is less well-defined. In addition, the effects of the discordance/concordance between RC and LDL-C on MASLD risk in T2DM remain unknown. Furthermore, RC can trigger an inflammatory response;17–19 the effect of the independent and combined effects of RC and inflammatory markers such as high-sensitive C reactive protein (hs-CRP) on risks of MASLD remains unknown. Hence, we aimed to evaluate the association of RC, its discordance/concordance with LDL-C and the independent and combined effects of RC and hs-CRP in terms of MASLD risk in patients with T2DM.Research design and methods Study population This cross-sectional study included 27 343 T2DM patients hospitalized in Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (Wuhan, China) with measurements of abdominal ultrasonography between 2014 and 2024. We excluded patients who were younger than 18 years (n=12); with excessive alcohol consumption (>30 g/day for men and >20 g/day for women) (n=358); with positive hepatitis B surface antigen or hepatitis C antibody, or autoimmune hepatitis (n=647); and without necessary information such as RC (n=76). The remaining available 26 302 T2DM patients were included in the present analyses. According to the Private Information Protection Law, information that might identify subjects was safeguarded by the Computer Center. This study was approved by the institutional review board of Tongji Hospital. Because we only retrospectively accessed a deidentified database for purposes of analysis, informed consent requirement was exempted by the institutional review board.Clinical measurements Data on patients’ age, sex, smoking, drinking, current and previous medical histories as well as treatments were obtained from the hospital’s electronic medical records. Data on patients’ age, sex, smoking status, alcohol consumption, current and previous medical history, and treatments were obtained from the hospital electronic medical records. 20 Smoking status was defined according to the Centers for Disease Control and Prevention criteria.21 Never smokers were defined as individuals who had smoked less than 100 cigarettes in their lifetime. Former smokers were those who had smoked at least 100 cigarettes in their lifetime and had quit for at least 30 days. Current smokers were participants who had smoked at least 100 cigarettes in their lifetime and reported smoking within the last 30 days. Drinking behavior was classified based on the frequency and quantity of alcohol consumption. Specifically, individuals were considered heavy drinkers if they consumed more than 30 g of alcohol per day for men, and 20 g/day for women, in line with the guidelines of the WHO. Height, weight and blood pressure (BP) were measured according to the standardized protocol of the WHO. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. After a 5 min rest, sitting BP was measured on the patient’s right arm with a sphygmomanometer two times every 5 min. Data were analyzed using the mean of the two readings.Laboratory measurements After an overnight fast of at least 8 hours, venous blood samples were obtained and tested immediately after collection. LDL-C, total cholesterol (TC), triglyceride (TG), high-density lipoprotein cholesterol (HDL-C), non-HDL-C, fasting plasma glucose (FPG), aspartate transferase, alanine aminotransferase, gamma-glutamyl transferase, alkaline phosphatase, creatinine, urinary albumin-to-creatinine ratio were determined by using an autoanalyzer (Cobas C8000, Roche, Mannheim, Germany). RC was calculated as TC-LDL-C-HDL-C. The estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation. Glycosylated hemoglobin A1c (HbA1c) was measured using high-performance liquid chromatography (D-10TM; Bio-Rad Laboratories, Hercules, California). The hs-CRP was detected with a ROCHE COBAS 8000 analyzer.Ultrasonography Ultrasound tests were performed by certified sonographers using a high-resolution, real-time scanner (model SSD-2000; Aloka, Tokyo, Japan). Certified radiologists used standard criteria in evaluating the presence or absence of hepatic fat. 22Definitions Obesity was defined as BMI ≥28.0 kg/m² according to the Working Group on Obesity in China (WGOC). 23 T2DM was diagnosed according to the 2025 American Diabetes Association (ADA) criteria.24 According to the ADA criteria, poor glycemic control was defined as HbA1c level ≥7.0%; poor BP control as BP ≥130/80 mm Hg; poor cholesterol control as LDL-C level ≥100 mg/dL; poor TG control as TG level ≥150 mg/dL; poor HDL-C control as HDL-C level ≤40/50 mg/dL for men/women. Liver steatosis was defined as the presence of stronger echoes in the hepatic parenchyma compared with echoes in the kidney or spleen parenchyma.22Statistical analyses All statistical analyses were conducted using R, V.4.2.2 (The R Foundation for Statistical Computing, Vienna, Austria). Continuous variables were presented as means (SDs) or medians (IQRs) depending on their distribution. Categorical variables were presented as numbers (percentages). Differences in continuous variables between groups were tested with Analysis of Variance or Kruskal-Wallis test. Differences in categorical variables were tested with χ 2 test. Logistic regression models were used to assess the associations (ORs, with 95% CIs) of RC with the risk of MASLD. Model 1 was a crude model. Model 2 was adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, antihypertensive medication use and HbA1c. Model 3 was further adjusted for hs-CRP. Among these, BMI, HbA1c and hs-CRP were adjusted as continuous variables, while smoking and drinking status were adjusted as categorical variables.We used three approaches using the same models mentioned above to assess the effects of the discordance/concordance between RC and LDL-C on MASLD risk: (1) clinical cut-off points; (2) differences of >10 percentile points and (3) residuals. First, for clinical cut-off points, we chose 2.60 mmol/L (100 mg/dL), 3.37 mmol/L (130 mg/dL) and 0.62 mmol/L (24 mg/dL) for LDL-C, non-HDL-C and RC, respectively, based on worldwide guideline recommendations.11 25–27 Therefore, four mutually exclusive concordance/discordance groups were defined based on the cut-off points of LDL-C and RC: low/low (less than the cut-off points of both LDL-C and RC), low/high (less than the cut-off points of LDL-C and greater than or equal to the cut-off points of RC), high/low (greater than or equal to the cut-off points of LDL-C and less than the cut-off points of RC) and high/high (greater than or equal to the cut-off points of both LDL-C and RC). We assessed the association between RC and LDL-C concordant/discordant groups and MASLD risk with group 1 as the reference. The similar process was done for RC and non-HDL-C concordant/discordant groups and LDL and non-HDL-C concordant/discordant groups. Second, for percentile differences, percentile rankings for RC and LDL-C were tabulated using the cumulative distribution function. Then we calculated the concordance/discordance by the RC percentile minus the LDL-C percentile. We defined discordance using >10 difference in percentile units (RC percentile minus LDL-C percentile).28 Therefore, the population was divided into three groups: discordant high (RC>10 percentile units higher than LDL-C), concordant (RC within 10 percentile units of LDL-C) and discordant low (RC 10 percentile lower than LDL-C). We assessed the association between RC and LDL-C concordant/discordant groups and MASLD risk with the concordant group as the reference. At last, for residuals, we proceeded as follows. Due to the correlation between RC and LDL-C, to accurately quantify the effect of RC versus LDL-C, we regressed RC on LDL-C and used the resulting residual to represent the amount of RC that is not explained by LDL-C residual regression analyses. Similarly, we regressed LDL-C or non-HDL-C on RC, and the resulting residual was used to represent the portion of LDL-C or non-HDL-C not explained by RC.29–31Logistic regression analysis was also utilized to examine the independent and combined effects of RC and hs-CRP in terms of MASLD risk. To assess if the associations of RC with MASLD differed by hs-CRP levels, likelihood ratio test was utilized to examine potential interaction effects between RC and hs-CRP levels.32 We further evaluated the additive interaction using the relative excess risk due to interaction (RERI) and attributable proportion indices.Further examination of the association of RC with MASLD risk in age, gender, BMI categories, BP, types of antidiabetic medications and HbA1c was also evaluated. There are two types of antidiabetic medications: those receiving novel agents (sodium-glucose cotransporter-2 inhibitors (SGLT2i) or glucagon-like peptide-1 receptor agonists (GLP-1RA)) and those receiving traditional agents (sulfonylureas, thiazolidinediones or insulin). To assess if the associations differed by BMI status, logistic regression analysis was utilized to examine potential interaction effects between RC and BMI categories. The process was repeated separately for subgroups stratified by age, gender, BP types of antidiabetic medications and HbA1c. Significance was accepted at a two-tailed p<0.05.To ensure the robustness of our findings, sensitivity analyses were conducted after excluding individuals taking lipid-lowering medications and excluding the history of receiving novel antidiabetic medications (SGLT2i or GLP-1RA). Sensitivity analyses were also conducted in participants with TG<1.70 mmol/L and LDL<2.60 mmol/L.Results The characteristics of the participants Of the included 26 302 T2DM patients, the median age was 62.0 years (54.0–70.0 years), 61.2% (16 098) were male and 48.3% (12,702) had MASLD. As shown in table 1, T2DM patients with MASLD had significantly higher levels of TC, TG, LDL-C, non-HDL-C and RC and poorer glycemic control compared with T2DM patients without MASLD (all p value <0.05).Table 1The characteristics of the study participants with diabetes according to MASLD statusTotaln=26 302Without MASLDn=13 600MASLDn=12 702P valueAge (years)62.0 (54.0–70.0)65.0 (57.0–72.0)59.0 (49.0–67.0)＜0.001Male sex, n (%)16 098 (61.2)8594 (63.2)7504 (59.1)＜0.001Smoking, n (%)7355 (28.0)3806 (28.0)3549 (27.9)0.946BMI (kg/m2)24.7 (22.4–27.2)23.4 (21.3–25.6)26.0 (23.9–28.4)＜0.001SBP (mm Hg)128.0 (117.0–141.0)128.0 (116.0–143.0)128.0 (117.0–140.0)0.012DBP (mm Hg)78.0 (71.0–86.0)77.0 (70.0–85.0)80.0 (72.0–88.0)＜0.001Taking antihypertensive medications, n (%)13 568 (51.7)7308 (53.9)6260 (49.4)＜0.001Diabetes medications, n (%) Insulin10 773 (41.0)5318 (39.1)5455 (43.0)＜0.001 Biguanides9442 (35.9)3445 (25.3)5997 (47.2)＜0.001 Sulfonylureas1168 (4.4)513 (3.8)655 (5.2)＜0.001 TZDs2846 (10.8)753 (5.5)2093 (16.5)＜0.001 GLP-1 RA3834 (14.6)1074 (7.9)2760 (21.7)＜0.001 DPP4i4333 (16.5)2243 (16.5)2090 (16.5)＜0.001 SGLT2i6974 (26.5)2905 (21.4)4069 (32.0)＜0.001RC (mmol/L)0.6 (0.3–1.0)0.5 (0.3–0.9)0.7 (0.4–1.2)＜0.001LDL-C (mmol/L)2.6 (1.9–3.3)2.5 (1.9–3.2)2.7 (2.0–3.3)＜0.001TC (mmol/L)4.4 (3.6–5.2)4.2 (3.4–5.0)4.6 (3.9–5.3)＜0.001TG (mmol/L)1.7 (1.2–2.8)1.5 (1.0–2.2)2.1 (1.4–3.5)＜0.001HDL-C (mmol/L)1.0 (0.8–1.2)1.0 (0.8–1.2)1.0 (0.8–1.1)＜0.001Non-HDL-C (mmol/L)3.3 (2.6–4.1)3.1 (2.4–3.9)3.6 (2.9–4.3)＜0.001Insulin (mU/L)10.4 (6.1–19.4)8.6 (4.9–16.5)11.7 (7.2–21.2)＜0.001HbA1c (%)7.2 (6.3–9.2)6.9 (6.1–8.6)7.6 (6.4–9.8)＜0.001FPG (mmol/L)8.0 (6.1–12.3)7.8 (6.0–11.8)8.3 (6.2–12.8)＜0.001eGFR (mL/min/1.73 m2)91.5 (68.1–104.0)83.9 (55.3–98.9)97.3 (81.1–108.6)＜0.001Creatinine (umol/L)74.0 (60.0–94.0)79.0 (63.0–110.0)70.0 (58.0–84.0)＜0.001UACR (mg/g)22.6 (7.8–194.1)40.0 (9.3–670.9)16.3 (6.9–68.6)＜0.001hs-CRP (mg/L)2.5 (0.9–8.9)2.5 (0.8–13.0)2.4 (1.1–6.6)0.307Data were presented as median (IQR) or numbers (percentages).BMI, body mass index; DBP, diastolic pressure; DPP4i, dipeptidyl peptidase-4 inhibitors; eGFR, estimated glomerular filtration rate; FPG, fasting plasma glucose; GLP-1RA, glucagon-like peptide-1 receptor agonists; HbA1c, glycosylated hemoglobin A1c; HDL-C, high-density lipoprotein cholesterol; hs-CRP, high sensitivity C reactive protein; LDL-C, low-density lipoprotein cholesterol; MASLD, metabolic dysfunction-associated steatotic liver disease; non-HDL-C, non-high-density lipoprotein cholesterol; RC, remnant cholesterol; SBP, systolic pressure; SGLT2i, sodium-glucose cotransporter-2 inhibitors; TC, total cholesterol; TG, triglyceride; TZDs, thiazolidinediones; UACR, urinary albumin-to-creatinine ratio.The association of RC with MASLD As shown in table (additional file: online supplemental table 1), RC was significantly associated with MASLD risk after adjusting for age, sex, BMI, smoking status, drinking status, history of hypertension, antihypertensive medication use, HbA1c and hs-CRP. Using the lowest RC quantile group as the reference, the second (OR=1.12, 95% CI 1.03 to 1.22), the third (OR=1.33, 95% CI 1.22 to 1.45) and the fourth RC quantile group (OR=1.79, 95% CI 1.65 to 1.96) had significantly greater odds for MASLD. Similar overall results were observed across quantiles of LDL-C and non-HDL-C. However, RC was more closely associated with MASLD than LDL-C.SP110.1136/bmjdrc-2025-005841.supp1Supplementary dataEffects of the discordance/concordance between RC and LDL-C on MASLD risk As shown in table 2, regardless of which method of discordance analysis is used, RC was associated with MASLD risk stronger than LDL-C. For example, for the cut-off points approach, the risk of MASLD is greater in the LDL-C<2.60 mmol/L and RC ≥0.62 mmol/L group (OR=1.95, 95% CI 1.79 to 2.12) than in the LDL-C ≥2.6 mmol/L and RC <0.62 mmol/L group (OR=1.60, 95% CI 1.47 to 1.74) in model 3. For the percentile differences approach, the risk of MASLD was significantly higher in the group with discordant high RC (OR=1.22, 95% CI 1.12 to 1.33) than in the group with discordant low RC (OR=1.08, 95% CI 0.99 to 1.17 in the group with elevated LDL-C). For the residual approach, in a regression model including LDL-C and RC residual, adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, antihypertensive medication use, HbA1c and hs-CRP; RC residual was more associated with MASLD risk than LDL-C (OR=1.26, 95% CI 1.22 to 1.30 for RC residual and OR=1.04, 95% CI 1.01 to 1.07 for LDL-C). In a regression model including RC and LDL-C residual, RC was more associated with MASLD risk than LDL-C residual (OR=1.24, 95% CI 1.20 to 1.29 for RC and OR=1.09, 95% CI 1.06 to 1.12 for LDL-C residual).Table 2Effects of the discordance/concordance between RC and LDL-C on MASLD riskOR (95% CI)Model 1Model 2Model 3Clinical cut-off points: LDL-C 2.60 mmol/L, RC 0.62 mmol/L LDL-C<2.60 mmol/L and RC<0.62 mmol/LReferenceReferenceReference LDL-C<2.60 mmol/L and RC≥0.62 mmol/L2.42 (2.26 to 2.60)1.96 (1.81 to 2.12)1.95 (1.79 to 2.12) LDL-C≥2.60 mmol/L and RC<0.62 mmol/L2.05 (1.91 to 2.19)1.64 (1.51 to 1.77)1.60 (1.47 to 1.74) LDL-C≥2.60 mmol/L and RC≥0.62 mmol/L2.64 (2.45 to 2.83)1.86 (1.72 to 2.02)1.83 (1.67 to 2.00)Differences of>10 percentile points: RC percentile minus LDL-C percentile Concordant (within 10 percentiles)ReferenceReferenceReference Discordantly low RC1.13 (1.06 to 1.21)1.09 (1.01 to 1.17)1.08 (0.99 to 1.17) Discordantly high RC1.23 (1.15 to 1.32)1.21 (1.12 to 1.31)1.22 (1.12 to 1.33)Residuals LDL-C1.16 (1.13 to 1.18)1.05 (1.02 to 1.08)1.04 (1.01 to 1.07) RC residual1.44 (1.40 to 1.48)1.26 (1.22 to 1.30)1.26 (1.22 to 1.30)Residuals RC1.39 (1.35 to 1.43)1.24 (1.20 to 1.28)1.24 (1.20 to 1.29) LDL-C residual1.24 (1.21 to 1.27)1.10 (1.07 to 1.13)1.09 (1.06 to 1.12)Data were represented as the odds ratios (95% CIs).Model 1: unadjusted.Model 2: adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, anti-hypertensive medication use, and HbA1c.Model 3: adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, antihypertensive medication use, HbA1c, and hs-CRP. BMI, body mass index; HbA1c, glycosylated hemoglobin A1c; hs-CRP, high-sensitive C reactive protein; LDL-C, low density lipoprotein cholesterol; MASLD, metabolic dysfunction-associated steatotic liver disease; RC, remnant cholesterol.Effects of the discordance/concordance between RC and non-HDL-C on MASLD risk As shown in table 3, for the cut-off points approach, the risk of MASLD is greater in the non-HDL-C ≥3.37 mmol/L and RC <0.62 mmol/L group (OR=1.62, 95% CI 1.48 to 1.77) than in the non-HDL-C <3.37 mmol/L and RC ≥0.62 mmol/L group (OR=1.42, 95% CI 1.29 to 1.55) in model 3. For the percentile differences approach, in both the discordant low RC group (OR=1.04, 95% CI 0.97 to 1.12) and the discordant high RC group (OR=0.96, 95% CI 0.90 to 1.04), there was no statistically significant association with the risk of MASLD. For the residual approach, in a regression model including non-HDL-C and RC residual, non-HDL-C was more associated with MASLD risk than RC residual (OR=1.19, 95% CI 1.16 to 1.22 for non-HDL-C and OR=1.15, 95% CI 1.11 to 1.20 for RC residual). However, in a regression model including RC and non-HDL-C residual, RC was more associated with MASLD risk than non-HDL-C residual (OR=1.25, 95% CI 1.21 to 1.29 for RC and OR=1.10, 95% CI 1.07 to 1.13 for non-HDL-C residual).Table 3Effects of the discordance/concordance between RC and non-HDL-C on MASLD riskOR (95% CI)Model 1Model 2Model 3Clinical cut-off points: non-HDL-C 3.37 mmol/L, RC 0.62 mmol/L Non-HDL-C <3.37 mmol/L and RC<0.62 mmol/LReferenceReferenceReference Non-HDL-C <3.37 mmol/L and RC≥0.62 mmol/L1.50 (1.39 to 1.62)1.41 (1.30 to 1.54)1.42 (1.29 to 1.55) Non-HDL-C≥3.37 mmol/L and RC<0.62 mmol/L2.02 (1.88 to 2.17)1.64 (1.51 to 1.78)1.62 (1.48 to 1.77) Non-HDL-C≥3.37 mmol/L and RC≥0.62 mmol/L2.54 (2.39 to 2.69)1.91 (1.78 to 2.04)1.91 (1.77 to 2.05)Differences of>10 percentiles points: RC percentile minus non-HDL-C percentile Concordant (within 10 percentiles)ReferenceReferenceReference Discordantly low RC1.05 (0.98 to 1.11)1.03 (0.96 to 1.10)1.04 (0.97 to 1.12) Discordantly high RC0.82 (0.78 to 0.87)0.94 (0.88 to 1.01)0.96 (0.90 to 1.04)Residuals Non-HDL-C1.35 (1.32 to 1.38)1.20 (1.17 to 1.22)1.19 (1.16 to 1.22) RC residual1.16 (1.12 to 1.20)1.15 (1.10 to 1.19)1.15 (1.11 to 1.20)Residuals RC1.39 (1.35 to 1.43)1.25 (1.21 to 1.29)1.25 (1.21 to 1.29) Non-HDL-C residual1.24 (1.21 to 1.27)1.11 (1.08 to 1.14)1.10 (1.07 to 1.13)Data were represented as the ORs (95% CIs).Model 1: unadjusted.Model 2: adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, anti-hypertensive medication use, and HbA1c.Model 3: adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, antihypertensive medication use, HbA1c, and hs-CRP.BMI, body mass index; HbA1c, glycosylated hemoglobin A1c; hs-CRP, high-sensitive C reactive protein; LDL-C, low density lipoprotein cholesterol; MASLD, metabolic dysfunction-associated steatotic liver disease; non-HDL-C, non-high-density lipoprotein cholesterol; RC, remnant cholesterol.Effects of the discordance/concordance between LDL and non-HDL-C on MASLD risk As shown in table 4, for the cut-off points approach, the risk of MASLD is greater in the LDL-C <2.60 mmol/L and non-HDL-C ≥3.37 mmol/L group (OR=2.31, 95% CI 2.08 to 2.57) than in the LDL-C ≥2.60 mmol/L and non-HDL-C <3.37 mmol/L group (OR=1.20, 95% CI 1.08 to 1.33) in model 3. For the percentile differences approach, the risk of MASLD was significantly higher in the group with discordant high non-HDL-C (OR=1.63, 95% CI 1.50 to 1.77) than in the group with discordant low non-HDL-C (OR=0.95, 95% CI 0.89 to 1.02 in the group with elevated LDL-C). For the residual approach, in a regression model including non-HDL-C and LDL residual, non-HDL-C was more associated with MASLD risk than LDL residual (OR=1.19, 95% CI 1.16 to 1.22 for non-HDL-C and OR=0.87, 95% CI 0.83 to 0.90 for LDL residual). In a regression model including LDL and non-HDL-C residual, non-HDL-C residual was more associated with MASLD risk than LDL (OR=1.27, 95% CI 1.23 to 1.31 for non-HDL-C residual and OR=1.05, 95% CI 1.02 to 1.08 for LDL). The results indicated that non–HDL-C is a better indicator of MASLD risk than LDL-C.Table 4Effects of the discordance/concordance between LDL-C and non-HDL-C on MASLD riskOR (95% CI)Model 1Model 2Model 3Clinical cut-off points: LDL-C 2.60 mmol/L, non-HDL-C 3.37 mmol/L LDL-C<2.60 mmol/L and non-HDL-C <3.37 mmol/LReferenceReferenceReference LDL-C<2.60 mmol/L and non-HDL-C≥3.37 mmol/L3.10 (2.85 to 3.39)2.32 (2.11 to 2.56)2.31 (2.08 to 2.57) LDL-C≥2.60 mmol/L and non-HDL-C <3.37 mmol/L1.37 (1.26 to 1.48)1.22 (1.12 to 1.34)1.20 (1.08 to 1.33) LDL-C≥2.60 mmol/L and non-HDL-C≥3.37 mmol/L2.05 (1.94 to 2.17)1.58 (1.49 to 1.69)1.56 (1.46 to 1.67)Differences of>10 percentile points: non-HDL-C percentile minus LDL-C percentile Concordant (within 10 percentiles)ReferenceReferenceReference Discordantly low non-HDL-C0.94 (0.89 to 1.00)0.96 (0.90 to 1.02)0.95 (0.89 to 1.02) Discordantly high non-HDL-C1.89 (1.76 to 2.02)1.63 (1.51 to 1.76)1.63 (1.50 to 1.77)Residuals Non-HDL-C1.35 (1.32 to 1.38)1.20 (1.17 to 1.22)1.19 (1.16 to 1.22) LDL residual0.86 (0.83 to 0.89)0.87 (0.84 to 0.91)0.87 (0.83 to 0.90)Residuals LDL1.16 (1.13 to 1.18)1.06 (1.03 to 1.09)1.05 (1.02 to 1.08) Non-HDL-C residual1.44 (1.40 to 1.48)1.27 (1.23 to 1.31)1.27 (1.23 to 1.31)Data were represented as the ORs (95% CIs).Model 1: unadjusted.Model 2: adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, antihypertensive medication use, and HbA1c.Model 3: adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, antihypertensive medication use, HbA1c, and hs-CRP.BMI, body mass index; HbA1c, glycosylated hemoglobin A1c; HDL-C, high-density lipoprotein cholesterol; hs-CRP, high-sensitive C reactive protein; LDL-C, low-density lipoprotein cholesterol; MASLD, metabolic dysfunction-associated steatotic liver disease; non-HDL-C, non-high-density lipoprotein cholesterol.Subgroup analyses The effects of the discordance/concordance between RC and LDL-C on MASLD risk remained consistent when stratified by age, sex, BMI status, BP, types of antidiabetic medications and glycemia control status (additional file: online supplemental table 2). After adjusting for potential confounding variables, RC was more closely associated with MASLD risk than LDL-C, irrespective of age, sex, BMI status, BP, types of antidiabetic medications and glycemia control status. The test for interactions did not reach statistical significance for sex, types of antidiabetic medications or BMI (all p for interaction >0.05), indicating that the stronger association of RC with MASLD than LDL-C did not differ by sex, types of antidiabetic medications or BMI status. The stronger association of RC with MASLD than LDL-C was more evident in patients with age <65 years, optimal BP control and poor glycemic control (all p for interaction <0.05).Independent and combined effects of RC and hs-CRP on MASLD Table 5 demonstrates that in the unadjusted model (model 1), the risk of MASLD was highest in individuals with RC ≥0.62 mmol/L and hs-CRP of 1–3 mg/L (OR=2.88, 95% CI 2.60 to 3.18). After adjusting for age, sex, BMI, smoking status, drinking status, history of hypertension, anti-hypertensive medication use and HbA1c (model 2), the association remained for the group with RC ≥0.62 mmol/L and hs-CRP of 1–3 mg/L (OR=2.09, 95% CI 1.87 to 2.33). Moreover, the interactions between RC and hs-CRP on MASLD risk did not reach statistical significance (p=0.091), indicating that the associations of RC with MASLD did not differ by hs-CRP levels. Additive interaction analysis revealed no statistically significant additive interaction between RC and hs-CRP on the risk of developing MASLD (RERI: 0.000, 95% CI −0.001 to 0.001; AP: 0.004, 95% CI −0.005 to 0.018; synergy index: 0.088, 95% CI 0.055 to 0.119). These results collectively indicated that the associations of RC with MASLD did not differ by hs-CRP levels.Table 5The independent and combined effects of RC and hs-CRP on MASLD riskOR (95% CI)P value for interactionRERI (95% CI)AP (95% CI)SI (95% CI)Model 1Model 2RC<0.62 mmol/L and hs-CRP<1 mg/LReferenceReference0.0910.001 (−0.001 to 0.001)0.004 (-0.005 to 0.018)0.088 (0.055 to 0.119)RC<0.62 mmol/L and hs-CRP 1–3 mg/L1.72 (1.56 to 1.90)1.46 (1.31 to 1.62)RC<0.62 mmol/L and hs-CRP≥3 mg/L1.14 (1.04 to 1.25)0.96 (0.86 to 1.06)RC≥0.62 mmol/L and hs-CRP<1 mg/L1.64 (1.47 to 1.83)1.48 (1.32 to 1.67)RC≥0.62 mmol/L and hs-CRP 1–3 mg/L2.88 (2.60 to 3.18)2.09 (1.87 to 2.33)RC≥0.62 mmol/L and hs-CRP≥3 mg/L2.02 (1.84 to 2.21)1.41 (1.28 to 1.56)Data were represented as the ORs (95% CIs).P value for interaction was evaluated using the likelihood ratio test.Additive interaction was assessed by RERI, AP and SI.Model 1: unadjusted.Model 2: adjusted for age, sex, BMI, smoking status, drinking status, history of hypertension, antihypertensive medication use and HbA1c.AP, attributive ratio; BMI, body mass index; HbA1c, glycosylated hemoglobin A1c; hs-CRP, high-sensitive C reactive protein; LDL-C, low density lipoprotein cholesterol; MASLD, metabolic dysfunction-associated steatotic liver disease; RC, remnant cholesterol; RC, remnant cholesterol; RERI, relative excess risk; SI, synergy index.Sensitivity analyses We observed consistent results after excluding individuals with lipid-lowering medication use (additional files: online supplemental tables 3-6) and excluding the history of receiving novel antidiabetic medications (SGLT2i or GLP-1RA) (additional files: online supplemental tables 7,8). In addition, the results were essentially the same when we repeated the analysis in T2DM patients with TG<1.70 mmol/L and LDL-C<2.60 mmol/L (additional file: online supplemental table 9).Discussion This is, as far as we known, the first report to describe the risk of RC and effects of discordant RC with LDL-C as well as the independent/combined effects of RC and hs-CRP on MASLD risk in patients with T2DM. We found that (1) RC was positively associated with MASLD risk independent of traditional cardiovascular risk factors, (2) the odds for MASLD were significantly and consistently higher in low LDL-C/high RC group than in high LDL-C/low RC group in the discordance analyses and (3) hs-CRP did not modify the association between RC and MASLD. Our findings suggest that the measurement of RC-related residual risk in addition to LDL-C is clinically relevant, such as helping to distinguish T2DM patients who are predisposed to MASLD.The associations of RC with MASLD have been established in the general population.33 However, in patients with T2DM, the association remains less clear. Here, we supplemented the evidence about RC and MASLD risk in patients with T2DM. We confirmed the significant associations of RC with MASLD risk in T2DM patients, which is inconsistent with previous studies conducted in the general population.14 We also noted that RC was associated with an increased risk of MASLD superior to LDL-C. A recent study showed that RC was independently associated with the progression and regression of MASLD.16Since RC and LDL-C are biologically linked variables, treating RC or LDL-C separately or just considered the other cholesterol as a potential confounder when evaluating the risk factor-disease relationship fails to capture and compare the independent effects of RC and LDL-C on disease risk.11 34 Statistical attempts to strip their mutual effect are needed. Discordance analysis can address this issue.35 This analysis evaluating the effects of discordant RC with LDL-C on cardiovascular disease and stroke has received increased attention.34 Almost all these reports indicated that discordantly high RC, not discordantly high LDL-C, was associated with a higher risk of cardiovascular disease and stroke.36 Yet the discordance analysis has not been applied to MASLD field. Whether these findings for cardiovascular disease and stroke can be generalized to MASLD remains unclear. In the present study, we, for the first time, applied the discordance analyses to evaluate whether, and if so to what extent, RC and LDL-C are independent determinants of MASLD risk in patients with T2DM. We used three approaches to do the discordance analyses: clinical cut-off points, percentile differences and residuals, which is the least arbitrary and captures the maximum of the differing information between RC and LDL-C. We found that discordantly high RC was associated with higher MASLD risk than discordantly high LDL-C, irrespective of which method is used. These findings suggest that RC may provide an additional index for the risk stratification of MASLD beyond LDL-C in patients with T2DM. Our findings also extended the evidence that this association remained significant even among participants with optimal LDL-C and TG control levels. Therefore, assessment of the discordance between RC and LDL-C still contributes to the risk stratification in T2DM patients with optimal LDL-C and TG control.Since age, sex, BMI, diabetic control parameters such as HbA1c, types of antidiabetic medications and BP, have strong effects on MASLD and lipid profile,3 7 37 we further investigated whether the relationship between RC and LDL-C discordance/concordance and MASLD differed by these parameters. We found that MASLD risk is associated with RC more than LDL-C, irrespective of age, sex, BMI status, BP, types of antidiabetic medications and glycemia control status. Furthermore, we found that RC has similar effect on MASLD in men and women. The significant and independent associations of RC with MASLD in different BMI categories and types of antidiabetic medications noted in our study highlight that RC predisposes to increased MASLD risk, regardless of BMI categories and types of antidiabetic medications. Moreover, we demonstrated that patients with age <65 years, optimal BP control or poor glycemic control are more sensitive to MASLD from RC exposure. These findings suggested a more detrimental effect of high RC in these patients. Hence, it is worth noting that patients with age <65 years, optimal BP control or poor glycemic control require stricter RC management. However, estimates in subgroup analysis should be interpreted with caution due to limited sample size and inadequate statistical power.The exact mechanism for high RC having greater MASLD risk than high LDL-C remains unclear. This may be due to the following: first, unlike LDL-C, RC can be taken up by macrophages directly and thus may lead to hepatocellular injury through mitochondrial dysfunction or Kupffer cell activation.38 Second, RC, rather than LDL-C, was often accompanied by a pro-inflammatory state, which is a known pathogenetic factor for MASLD.39 40 Third, RC is more abundant, larger and carries more cholesterol than LDL-C particles, so it can be inferred that RC is more harmful to liver.41 42There is so far no evidence that supports a role for inflammation in the relationship between RC and MASLD. We addressed this fundamental knowledge gap in the present study. We found that RC remained an independent risk factor of MASLD even after additionally adjusting for hs-CRP. Furthermore, we also noted that hs-CRP did not modify the relationship between RC and MASLD. Studies indicated that inflammation, as indicated by hs-CRP, has less effect on the relationship between RC and cardiovascular disease.43 Our study together with these reports indicated that systemic proinflammatory response caused by elevated RC is a bystander.The main strength of this study is the large number of T2DM patients included from an academic hospital. We can get access to clinical, laboratory and imaging data in medical records, which provided more in-depth clinical information that are not usually available in large epidemiological surveys. In addition, the present study is the first to evaluate not only the independent effects of RC as a continuous measure but also its additional conferred risk by using a discordance analysis. We used three different approaches to do the discordance analyses to verify the robustness of our findings. Moreover, we did a sensitivity analysis after excluding individuals taking lipid-lowering medications or receiving antidiabetic medications to avoid the effect of statin use or treatment bias for diabetes on the association of the discordant/concordant LDL-C and RC with MASLD.We recognize several limitations of our study. First, the cross-sectional study design makes it difficult to infer causality between RC and MASLD. Second, information on directly measured RC or the more atherogenic lipoprotein remnant-like particle cholesterol, which were more accurate than the calculated method, were not available, which may lead to inaccurate values and measurement errors. Nevertheless, the calculated RC is closely related to RC directly measured and is widely used in studies.44 45 Third, although we adjusted for multiple potential confounding variables, residual and unmeasured confounding might not be fully addressed. For example, lack of information on residual confounding variables such as diabetes duration prevented us from being able to assess these variables as potential confounders. Fourth, our study population was mainly based on inpatients suffering from T2DM, whose health conditions might be severer than those of outpatients. Thus, our findings could not be generalized to outpatients with T2DM.Conclusion In summary, we found that RC was positively associated with the MASLD risk in patients with T2DM. Furthermore, discordance analyses showed that the discordantly high RC was associated with higher MASLD risk than discordantly high LDL-C, even among T2DM patients with optimal LDL-C and TG levels. Moreover, RC is associated with MASLD risk irrespective of hs-CRP levels. Our findings suggest that RC may serve as a potential target for prevention and intervention for MASLD. Future studies are warranted to determine whether lowering the RC level can produce benefits for MASLD.",
  "title": "Discordant high remnant cholesterol with LDL-C increases the metabolic dysfunction-associated steatotic liver disease risk in patients with type 2 diabetes mellitus",
  "uid": "7c32eb6d-a62d-55d1-8aa3-5d0aefbe7d3e"
}