Remnant cholesterol has a nonlinear association with non-alcoholic fatty liver disease: A secondary retrospective cohort study in non-obese Chinese adults

doi:10.21203/rs.3.rs-4957397/v1

Download PDF

Article

Remnant cholesterol has a nonlinear association with non-alcoholic fatty liver disease: A secondary retrospective cohort study in non-obese Chinese adults

https://doi.org/10.21203/rs.3.rs-4957397/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Objective

Evidence linking non-obese non-alcoholic fatty liver disease (NAFLD) with residual cholesterol (RC) is weak. In this study, Chinese adults who were not obese were examined for a relationship between NAFLD and RC levels.

Methods

16,004 non-obese volunteers were included in a retrospective cohort study at a Chinese clinic between the start of 2010 and the end of 2014. The correlation between initial RC levels and the likelihood of developing NAFLD was investigated using the proportional hazards regression (Cox) model. Using cubic spline functions and smooth curve fitting technique, a two-piecewise proportional hazards regression (Cox) method was used to investigate nonlinear associations between RC and NAFLD. There were numerous sensitivity assessments carried out. The information was posted on the DATADRYAD website.

Results

The results showed a positive correlation between RC and incidence of NAFLD after controlling for variables (HR = 1.061, 95% CI 1.045–1.078). Between RC and NAFLD, a nonlinear connection was found, with a turning point at 98.29 mg/dL. The effect sizes (HR) were 1.150 (95% CI 1.106, 1.194) and 1.009 (95% CI 0.982, 1.037) on either side of the turning point, respectively. A sensitivity examination further supported the stability of the findings.

Conclusion

In a Chinese sample that is not obese, this research emphasizes a potentially favorable nonlinear connection between RC and NAFLD risk. When RC was below 98.29 mg/dL, RC was substantially associated with the risk of NAFLD. Thus, from a therapeutic standpoint, keeping RC levels below this cutoff would be advantageous.

Health sciences/Medical research/Outcomes research

Health sciences/Biomarkers/Predictive markers

Health sciences/Gastroenterology/Hepatology/Liver diseases

Health sciences/Endocrinology/Endocrine system and metabolic diseases/Dyslipidaemias

Health sciences/Health care/Nutrition

Health sciences/Health care/Disease prevention/Lifestyle modification

Non-alcoholic fatty liver disease

Nonlinear

Residual lipids

smooth curve fitting

Predictors

Non-obese Chinese adults

A major contributor to chronic liver disease, non-alcoholic fatty liver disease (NAFLD) affects 25–38% of people worldwide ^[1–3]. Moreover, it is anticipated that the incidence of NAFLD will reach 22.2% in China and 28.4% in the United States by 2023 ^[4]. NAFLD has a major negative influence on life quality, involvement in the work force, and use of healthcare resources. It also places a heavy financial strain on society ^[2]. Over the next few decades, it is expected that the incidence and economic burden of NAFLD would rise in tandem with the rising prevalence of obesity and type 2 diabetic mellitus (T2DM) ^{[5, 6]}. NAFLD, defined as an excessive buildup of liver triglycerides over the 5% threshold ^[7], includes a variety of chronic conditions, from non-alcoholic steatohepatitis to simple steatosis. It can result in cirrhosis, hepatocellular carcinoma, and liver fibrosis ^[1–3], which raises the incidence, mortality, and need for liver transplantation rates ^{[8, 9]}. Furthermore, there exists a reciprocal causal relationship between NAFLD, metabolic syndrome, and T2DM. This association serves to increase the development of obesity, hypertension, atherosclerotic cardiovascular disease, and chronic renal disease ^[10–13]. To stop NAFLD from progressing or becoming comorbid with the previously listed negative outcomes, it is imperative to research, track, and evaluate the prevalence of NAFLD as well as its predictive indicators, and develop early intervention strategies.

NAFLD diagnosis requires evidence of hepatic steatosis through biopsy, imaging, or biomarkers along with overweight/obesity, T2DM, or metabolic dysregulation ^[14], it seems NAFLD predominantly occurs in individuals with obesity. Surprisingly, approximately 40% of global NAFLD patients are non-obese with a healthy BMI (< 25 kg/m² for Caucasians and < 23 kg/m² for Asians), with 20% being lean ^{[6, 15]}. Furthermore, in the general population, approximately 12.1% have non-obese NAFLD and 5.1% have lean NAFLD ^[15]. These individuals are at risk of central obesity, diabetes, cardiovascular diseases, hypertension, and increased mortality rates ^{[6, 15]}. Therefore, it is crucial to identify overlooked and at-risk non-obese NAFLD individuals.

NAFLD is affected by various factors including familial predisposition, age, sex, lipid profiles, body mass index (BMI), waist-to-hip ratio, sedentary lifestyle, and diabetes ^[16–18], with lipids playing a crucial role ^[19–21]. However, the lipid-NAFLD link is complex and controversial; some studies indicate dyslipidemia as a contributing factor, whereas others question the direct association between low-density lipoprotein cholesterol or total cholesterol and NAFLD outcomes ^[22–24]. Remnant cholesterol (RC), a serum marker for cardiovascular disease and mortality, represents cholesterol bound to remnant lipoproteins after triglyceride clearance from triglyceride-rich lipoproteins, encompassing fasting very low-density lipoprotein, intermediate-density lipoprotein, and non-fasting coeliac residues ^{[25, 26]}. Cardiovascular disease is a primary cause of death for NAFLD patients ^[12], which raises the question of whether there is a relationship between remnant cholesterol (RC) and NAFLD. Limited evidence currently available suggests that RC may be a therapeutic target and may be associated with the development and progression of NAFLD ^[26–28]. Specifically, RC levels may outweigh other lipid metrics as a separate risk factor for NAFLD ^[29–31]. But current research mostly focuses on cross-sectional data, leaving out important aspects of the link between RC and NAFLD in individuals who are not obese ^{[29, 32, 33]}.

As a result, a more thorough study is desperately needed to determine how RC and NAFLD are related in non-obese people and to create early intervention strategies to prevent disease progression. In this particular context, we investigated whether RC levels are independently linked to NAFLD in non-obese individuals by performing a secondary analysis of published data.

2.1. Study design

Using data from a longitudinal, non-interventional study spanning a demographically varied sample, which was started by the Wenzhou Healthcare Center in China, which is connected to the Wenzhou People's Hospital, this research employed a retrospective analysis framework ^[34]. The key independent metric that was evaluated was the initial RC level. The dependent variable was the presence of NAFLD, which was categorized as a binary outcome: absence (coded 0) or presence (coded 1).

2.2. Data source

The dataset was accessed without charge from the DATADRYAD repository, as contributed by the research team led by Sun D Q and others, and their full research design and data are accessible online at https://doi.org/10.5061/dryad.1n6c4 in the Dryad Digital Repository ^[34]. Following the repository's usage policies, the dataset is freely available for further analytical purposes, thereby respecting the original contributors' intellectual property rights.

2.3. Study population

The demographics under scrutiny included Chinese adults who were not classified as overweight. Participants in the foundational study were chosen sequentially for routine health examinations at the previously described Wenzhou Medical Center and were clear of NAFLD ^[34]. The aforementioned hospital's ethics board accepted the study, and each participant gave their informed permission ^[34]. Therefore, subsequent analyses did not require additional ethical review ^[34]. The initial research followed the parameters set forth in the Helsinki Declaration, and all processes were carried out in compliance with the pertinent rules and regulations. The same standards were applied for the secondary analysis.

Initially, the study cohort comprised 33,135 individuals who showed up for the center's yearly health examination; however, 16,962 individuals were subsequently excluded. The remaining 16,173 individuals, who were enrolled and completed the period of observation between the first of 2010 and the close of 2014, were eligible for inclusion in this investigation (Fig. 1) ^[34]. Chinese people without NAFLD, participants in longitudinal studies, and those who underwent health evaluations between the beginning of 2010 and the end of 2014 were the inclusion criteria for the initial research. The following were the defined exclusion criteria: (i) those who use large amounts of alcohol (more than 140 g for men and more than 70 g for women per week); (ii) identified chronic liver disease causes, such as viral hepatitis, autoimmune hepatitis, or NAFLD; (iii) those with LDL-c levels exceeding 3.12 mmol/L and a BMI of ≥ 25 kg/m²; (iv) participants on medication for hypertension, hyperlipidemia, or diabetes; and (v) those who were untraceable or had incomplete data ^[34]. For the current study, additional exclusions were made for participants with abnormal or extreme RC values (n = 169). The final study sample comprised 16,004 individuals.

2.4. Independent and outcome variables

The main independent variable of interest was RC. Milligrams per deciliter, or mg/dL, was the unit of measurement used to quantify RC as a continuous metric. The precise computation of RC was performed using the formula RC (mg/dL) = total cholesterol [TC/(mg/dL)] minus high-density lipoprotein cholesterol [HDL-c (mg/dL)] minus low-density lipid cholesterol [LDL-c (mg/dL)], as described previously ^[35]. NAFLD was designated as the dependent outcome and categorized binarily. The absence of NAFLD was denoted as 0, whereas its presence was indicated by 1.

2.5. Diagnosis of NAFLD

In alignment with the original study design, employing the following standards, NAFLD was diagnosed by ultrasonography: a) generalized increase in near-field echoes and a progressive weakening of far-field sounds in the liver area, b) poor definition of intrahepatic ductal structures, c) round-edged, slight to moderate hepatomegaly, d) lowered signs of hepatic circulation, and e) inadequate or insufficient visibility of the diaphragm's dome and the lobe of the right liver ^[34]. Ultrasonographic examination of the liver was performed without bias, maintaining the same conditions as those at the onset of the investigation, to determine the incidence of NAFLD. Subjects were assessed either upon receiving an NAFLD diagnosis or at their concluding consultation, prioritizing earlier diagnosis.

2.6. Covariates

Covariates in our study were chosen based on our medical experience, previous literature, and the original research ^[34–36]. Covariates included the subsequent elements: (1) constant variables: height, weight, BMI, age, diastolic blood pressure (DBP), aspartate aminotransferase (AST), systolic blood pressure (SBP), triglyceride (TG), serum creatinine (Scr), γ-glutamyl transpeptidase (GGT), albumin (ALB), serum urea nitrogen (BUN), globulin (GLB), LDL-c, alanine aminotransferase (ALT), HDL-c, TC, total bilirubin (TBIL), uric acid (UA), and fasting plasma glucose (FPG), and (2) categorical variable: sex. The initial study delineated variable definitions and assessments ^[34].

2.7. Handling of missing data

The quantity of participants in this study lacks data for BUN, UA, Scr, DBP, ALB, SBP, GLB, ALT, AST, GGT, and TBIL were 1 (0.006%), 1 (0.006%), 1 (0.006%), 1 (0.006%), 19 (0.119%), 19 (0.119%), 1363 (8.517%), 1363 (8.517%), 4019 (25.112%), 4019 (25.112%), 4021 (25.125%), 5629 (35.172%), respectively. Multiple imputations were employed in the current research to mitigate biases caused by missing data that could make it difficult to accurately depict the target sample's statistical power during the modeling stage ^{[37, 38]}. The model of imputation comprised variables such as age, TG, sex, height, Scr, BMI, BUN, Glu, GLB, DBP, SBP, UA, ALB, ALT, AST, GGT, with five iterations, linear regression type, and missing data assumed to be missing at random (MAR).

2.8. Analytical statistics

We used the R programming language (version 4.2.0, https://www.R-project.org, R Foundation) and the statistical application EmpowerStats (version 4.1, https://www.empowerstats.com, XY Solutions, Inc., Boston, MA) to conduct data analysis. In a two-sided analysis, P values less than 0.05 were considered statistically significant.

The subjects were stratified using RC quartiles. To ascertain whether there were significant differences between groups, continuous variables are reported as medians and quartiles for skewed distributions or as means and standard deviations for normal distributions. One-way ANOVA and the Kruskal-Wallis H-test are the analysis methods used for each case. The frequencies and percentages of categorical data are displayed, and the χ² test was applied to ascertain the statistical significance among the groups. The association between RC and NAFLD risk was examined using models for Cox proportional hazards regression, both univariate and multivariate. Additionally, in order to represent the intricate link between RC and NAFLD, we used smooth curve fitting. Furthermore, a segmented Cox regression model was utilized to analyze the nuances of this nonlinear connection. Lastly, a crucial step in our analytical process was to use the log-likelihood ratio test to determine which model best captured the risk associated with RC in the setting of NAFLD. To ensure the robustness of our findings, we conducted several sensitivity analyses, categorizing RC by quartiles and assessing its trends while excluding individuals with DBP ≥ 90 mmHg, SBP ≥ 140 mmHg, or FPG > 7.0 mmol/L ^[39] because of the significant association of hypertension and diabetes with NAFLD ^[11]. All results complied with the STROBE guidelines ^{[40, 41]}.

3.1. Features of the participants

The individuals' clinical and demographic details are shown in Table 1. 43.22 ± 14.96 years was the mean age, with 7,615 males (47.58%). During the median follow-up period, 2,246 people (14.03%) developed NAFLD over 2.98 years. Based on the quartiles of RC (Q1: <73.74, Q2: 73.74–90.30, Q3: 90.30-109.10, Q4: ≥109.10 mg/dL), the participants were divided into subgroups. The Q4 group (≥ 109.10 mg/dL) showed higher levels for various covariates compared to the Q1 group (< 73.74 mg/dL), such as UA, age, BUN, TG, ALT, Scr, AST, FPG, TC, weight, BMI, GGT, SBP, LDL-c, and DBP, while HDL-c values and the male proportion were lower. Additionally, the proportion of females was slightly higher than that of males in each RC subgroup, particularly in Q4 compared to Q1. Figure 2 displayed the skewed distribution of RC values with a median of 32.10 mg/dL (23.59–42.92 mg/dL).

Table 1

Baseline characteristics of the participants.
RC quartiles	Q1(< 73.74)	Q2(73.74–90.30)	Q3(90.30-109.10)	Q4(≥ 109.10)	P-value
participants	4001	4001	4001	4001
Age(years)	42.32 ± 14.68	42.18 ± 14.62	43.36 ± 14.88	45.01 ± 15.47	< 0.001
GGT(U/L)	18.00 (14.00–26.00)	20.00 (15.00–31.00)	23.00 (16.00–35.00)	26.00 (18.00–42.00)	< 0.001
ALT(U/L)	15.00 (11.00-21.34)	16.00 (11.00–23.00)	17.00 (12.00–25.00)	18.00 (13.00-26.87)	< 0.001
AST(U/L)	21.82 ± 9.73	22.33 ± 8.90	23.11 ± 8.99	23.64 ± 10.17	< 0.001
ALB(g/L)	44.19 ± 2.84	44.32 ± 2.67	44.42 ± 2.74	44.66 ± 2.55	< 0.001
GLB(g/L)	29.35 ± 4.05	29.46 ± 3.86	29.58 ± 3.85	29.56 ± 3.68	0.025
TBIL(umol/L)	12.25 ± 5.43	12.03 ± 4.81	12.21 ± 4.87	12.12 ± 4.78	0.205
BUN (umol/L)	4.39 ± 1.39	4.44 ± 1.32	4.58 ± 1.31	4.85 ± 1.39	< 0.001
Scr (umol/L)	76.93 ± 24.56	78.45 ± 27.89	79.20 ± 23.75	78.95 ± 25.53	< 0.001
UA (umol/L)	256.60 ± 83.19	269.55 ± 80.35	284.98 ± 82.88	304.75 ± 87.42	< 0.001
FPG(mmol/L)	5.03 ± 0.67	5.10 ± 0.67	5.17 ± 0.79	5.26 ± 0.91	< 0.001
HDL-c(mg/dL)	90.31 ± 19.75	86.38 ± 20.04	82.92 ± 20.89	75.66 ± 19.07	< 0.001
TC(mg/dL)	222.32 ± 29.03	253.08 ± 26.84	276.02 ± 28.98	303.62 ± 29.40	< 0.001
TG(mg/dL)	77.78 ± 30.47	96.24 ± 39.50	116.25 ± 51.81	157.77 ± 86.99	< 0.001
LDL-c(mg/dL)	70.98 ± 14.98	84.63 ± 13.84	93.96 ± 14.03	100.96 ± 13.13	< 0.001
Height(cm)	163.79 ± 7.44	164.26 ± 7.78	165.03 ± 7.86	164.84 ± 8.03	< 0.001
Weight(kg)	55.75 ± 8.03	57.51 ± 8.48	59.14 ± 8.54	59.77 ± 8.49	< 0.001
BMI(kg/m²)	20.71 ± 2.05	21.23 ± 2.04	21.63 ± 1.98	21.91 ± 1.92	< 0.001
SBP(mmHg)	116.53 ± 15.79	119.15 ± 16.48	121.77 ± 16.23	125.14 ± 16.94	< 0.001
DBP(mmHg)	70.27 ± 9.65	71.83 ± 10.08	73.61 ± 10.28	75.34 ± 10.58	< 0.001
Sex					0.051
Male	1972 (49.29%)	1898 (47.44%)	1896 (47.39%)	1849 (46.21%)
Female	2029 (50.71%)	2103 (52.56%)	2105 (52.61%)	2152 (53.79%)

Values are mean ± SD or median (quartile) or n (%)

RC, Remnant cholesterol; GGT, γ-glutamyl transpeptidase; ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; ALB, albumin; GLB, globulin; TBIL, total bilirubin; BUN, Serum urea nitrogen; Scr, Serum creatinine; UA, uric acid; FPG, fasting plasma glucose; HDL-c, High-density lipoprotein cholesterol; TC, Total cholesterol; TG, Triglyceride; LDL-c, Low-density lipoprotein cholesterol; BMI, Body mass index; SBP, Systolic blood pressure; DBP, Diastolic blood pressure.

3.2. The incidence rate of NAFLD

In the course of a median follow-up period of 2.98 years, Table 2 reveals that 2246 clients (14.03%) received a diagnosis of NAFLD. 500.78 cases overall, cumulatively, were reported for every 10,000 person-years. For the four subgroups with RC values, the overall prevalence rates were 236.74, 397.71, 578.55, and 766.31 instances per 10,000 person-years, respectively. NAFLD's overall rates of incidence were 14.03% (13.50-14.57%), 6.42% (5.66–7.18%), 10.72% (9.76–11.68%), 15.97% (14.84–17.11%), and 23.02% (21.716–24.32%) for each of the RC subgroups.

Table 2

Incidence of NAFLD (%).
RC (mg/dL)	Participants(n)		Incidence rate (95% CI) (%)	Per 10000 person-year
Total	16004	2246	14.03(13.50-14.57)	500.78
Q1(< 73.74)	4001	257	6.42 (5.66–7.18)	236.74
Q2(73.74–90.30)	4001	429	10.72 (9.76–11.68)	397.71
Q3(90.30-109.10)	4001	639	15.97(14.84–17.11)	578.55
Q4(≥ 109.10)	4001	921	23.02 (21.716–24.32)	766.31
P for trend			< 0.001	<0.001

RC, Remnant cholesterol; NAFLD, non-alcoholic fatty liver disease; CI, confidence interval.

3.3. Single-variable analysis

In the univariate assessment utilizing Cox proportional-hazards regression modeling, it was discovered that NAFLD exhibited no significant link to GLB, ALB, or TBIL, with all P-values exceeding the 0.05 threshold. However, a substantial correlation was identified between NAFLD and various factors, including age, with a hazard ratio (HR) of 1.007 and a 95% confidence interval (CI) ranging from 1.004 to 1.009. Additionally, significant associations were observed with GGT, with an HR of 1.007 and a 95% CI of 1.006 to 1.008; ALT, with an HR of 1.008 and a 95% CI from 1.007 to 1.008; and AST, with an HR of 1.011 and a 95% CI between 1.009 and 1.013. The hazard ratios for BUN, Scr, and UA were 0.932, 1.005, and 1.005, respectively, with a 95% CI that did not overlap with the null value of 1. FPG showed a notably elevated HR of 1.296, with a 95% CI from 1.266 to 1.326. HDL-c was inversely related, with an HR of 0.979 and a 95% CI of 0.977 to 0.981. TC, TG, and LDL-c exhibited positive associations with HRs of 1.004, 1.006, and 1.019, respectively, each accompanied by a 95% CI that confirmed their significance. BMI was strongly correlated with an HR of 1.815 and a 95% CI from 1.765 to 1.867. SBP and DBP also demonstrated significant relationships with HRs of 1.022 and 1.045, respectively, both with 95% CIs that were well below the significance threshold (all P-values were less than 0.05; refer to Table 3 for detailed data).

Table 3

Results of univariate Cox proportional hazards model.
Exposure	Statistics	HR (95%CI)	P-value
Sex
Male	7615 (47.582%)	1.0
Female	8389 (52.418%)	1.178 (1.084, 1.281)	< 0.001
Age(years)	43.217 ± 14.958	1.007 (1.004, 1.009)	< 0.001
GGT(U/L)	27.626 ± 30.423	1.007 (1.006, 1.008)	< 0.001
ALT(U/L)	19.242 ± 16.515	1.008 (1.007, 1.008)	< 0.001
AST(U/L)	22.723 ± 9.486	1.011 (1.009, 1.013)	< 0.001
ALB(g/L)	44.399 ± 2.707	1.009 (0.993, 1.025)	0.262
GLB(g/L	29.491 ± 3.866	1.004 (0.993, 1.015)	0.471
TBIL(umol/L)	12.153 ± 4.977	1.001 (0.993, 1.009)	0.843
BUN (umol/L)	4.564 ± 1.364	0.932 (0.902, 0.963)	< 0.001
Scr (umol/L)	78.384 ± 25.492	1.005 (1.004, 1.005)	< 0.001
UA (umol/L)	278.968 ± 85.400	1.005 (1.005, 1.006)	< 0.001
FPG(mmol/L)	5.138 ± 0.769	1.296 (1.266, 1.326)	< 0.001
HDL-c(mg/dL)	83.818 ± 20.658	0.979 (0.977, 0.981)	< 0.001
TC(mg/dL)	263.758 ± 41.347	1.004 (1.003, 1.005)	< 0.001
TG(mg/dL)	112.009 ± 63.778	1.006 (1.006, 1.006)	< 0.001
LDL-c(mg/dL)	87.632 ± 17.952	1.019 (1.016, 1.021)	< 0.001
HEIGHT	164.480 ± 7.794	1.048 (1.043, 1.054)	< 0.001
WEIGHT1	58.041 ± 8.530	1.104 (1.099, 1.110)	< 0.001
BMI(kg/m²)	21.371 ± 2.047	1.815 (1.765, 1.867)	< 0.001
SBP(mmHg)	120.647 ± 16.673	1.022 (1.020, 1.024)	< 0.001
DBP(mmHg)	72.764 ± 10.328	1.045 (1.041, 1.049)	< 0.001

Values are presented as the mean ± SD or n (%).

GGT, γ-glutamyl transpeptidase; ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; ALB, albumin; GLB, globulin; TBIL, total bilirubin; BUN, Serum urea nitrogen; Scr, Serum creatinine; UA, uric acid; FPG, fasting plasma glucose; HDL-c, High-density lipoprotein cholesterol; TC, Total cholesterol; TG, Triglyceride; LDL-c, Low-density lipid cholesterol; BMI, Body mass index; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; HR, Hazard ratios; CI, confidence interval.

Figure 3 displays Kaplan-Meier survival curves that illustrate the likelihood of NAFLD-free survival categorized by RC subgroups. P < 0.0001 for the log-rank test indicated significant differences in the NAFLD-free survival probabilities across the RC subgroups. As the RC level increased, the probability of NAFLD-free survival progressively declined, suggesting that the group with the highest RC level also had the highest risk of NAFLD development.

3.4. Multivariate analysis

We developed a trio of Cox proportional hazards regression models to scrutinize the correlation between RC levels and the incidence of NAFLD, as detailed in Table 4. The initial, unadjusted model (referred to as Model I) revealed that for every 10 mg/dL elevation in RC, there was a corresponding 14.2% escalation in NAFLD risk, with an HR of 1.142, a 95% CI ranging from 1.125 to 1.159, and a P-value well below 0.001. Upon introducing minimal adjustments for demographic factors—namely age and sex—in Model II, the risk increment associated with a 10 mg/dL rise in RC was slightly reduced to 13.9%, yielding an HR of 1.139 and a 95% CI from 1.122 to 1.156, with statistical significance persisting (P < 0.001). In the comprehensively adjusted model, Model III, which accounted for a broader spectrum of variables including SBP, AST, sex, BMI, Scr, GGT, ALB, ALT, FPG, UA, age, and DBP, each 10 mg/dL increment in RC corresponded to a 6.1% heightened risk of NAFLD. This was represented by an HR of 1.061 and a 95% CI spanning 1.045 to 1.078, with the relationship maintaining robust statistical support (P < 0.001). The consistency of the confidence intervals across models suggests that the observed association between RC and NAFLD is statistically robust.

Table 4

Association between RC and incident NAFLD in different models.
Exposure	Model I (HR,95%CI, P)	Model II (HR,95%CI, P)	Model III (HR,95%CI, P)	Model IV (HR,95%CI, P)
RC (per 10mg/dL)	1.142 (1.125, 1.159) < 0.001	1.139 (1.122, 1.156) < 0.001	1.061 (1.045, 1.078) < 0.001	1.058 (1.041, 1.076) < 0.001
RC quartiles
Q1	Ref	Ref	Ref	1.0
Q2	1.635 (1.401, 1.908) < 0.001	1.634 (1.400, 1.907) < 0.001	1.292 (1.106, 1.508) 0.001	1.197 (1.024, 1.398) 0.024
Q3	2.320 (2.007, 2.682) < 0.001	2.305 (1.994, 2.665) < 0.001	1.576 (1.362, 1.823) < 0.001	1.408 (1.216, 1.631) < 0.001
Q4	2.882 (2.507, 3.312) < 0.001	2.840 (2.471, 3.265) < 0.001	1.672 (1.452, 1.926) < 0.001	1.560 (1.352, 1.799) < 0.001
P for trend	< 0.001	< 0.001	< 0.001	< 0.001

Model I: We did not adjust for the other covariates.

Model II: Adjusted for age and sex.

Model III: We adjusted for age, sex, SBP, BMI, DBP, ALT, GGT, ALB, FPG, AST, UA, and Scr levels.

Model IV: Adjusted for age(smooth), sex, SBP (smooth), BMI(smooth), DBP(smooth), ALT(smooth), GGT(smooth), ALB(smooth), FPG(smooth), AST(smooth), UA(smooth), and Scr (smooth).

HR, Hazard ratio; CI, confidence interval; Ref, reference; NAFLD, non-alcoholic fatty liver disease.

3.5. Sensitivity analysis

To validate the robustness of our findings, sensitivity analyses were conducted. Initially, the RC levels were divided into quartiles and reintegrated into the regression model. The analysis confirmed that HRs remained stable across quartiles, mirroring the impact observed with the continuous RC variable, and the trend's P-value was congruent with the original analysis. As depicted in Model IV of Table 4, the results were in line with the fully adjusted model, which incorporated smooth terms for Scr, UA, sex, SBP, AST, FPG, DBP, GGT, ALT, ALB, BMI, and age, yielding an HR of 1.058 with a 95% CI of 1.041–1.076 and a P-value less than 0.001. Further sensitivity analyses excluded individuals with FPG exceeding 7.0 mmol/L. Even after controlling for potential confounding factors—comprising BMI, age, ALT, UA, SBP, AST, DBP, sex, GGT, ALB, FPG, and Scr—the relationship between RC and the risk of NAFLD was still positive, with an HR of 1.061 and a 95% CI of 1.044–1.078, achieving statistical significance (P < 0.001) as documented in Table 5. Parallel sensitivity analyses, which omitted participants diagnosed with hypertension (defined by SBP of 140 mmHg or higher or DBP of 90 mmHg or higher), retained the significant association between RC and NAFLD risk after adjustment for the same set of covariates, with an HR of 1.062 and a 95% CI of 1.043–1.082 and a P-value below 0.001. The uniformity of these outcomes across all sensitivity analyses in Table 5 reinforces the reliability of our conclusions.

Table 5

Relationship between RC and NAFLD in different sensitivity analyses
Exposure	Model V (HR,95%CI, P)	Model VI (HR,95%CI, P)
RC (per 10mg/dL)	1.061 (1.044, 1.078) < 0.001	1.062 (1.043, 1.082) < 0.001
RC quartiles
Q1	Ref	Ref
Q2	1.300 (1.109, 1.524) 0.001	1.279 (1.071, 1.528) 0.007
Q3	1.562 (1.344, 1.814) < 0.001	1.528 (1.292, 1.808) < 0.001
Q4	1.664 (1.439, 1.925) < 0.001	1.672 (1.418, 1.973) < 0.001
P for trend	< 0.001	< 0.001

Model V was a sensitivity analysis for participants without FPG > 7.0 mmol/L (N = 15683). We adjusted for age, sex, SBP, BMI, DBP, ALT, GGT, ALB, FPG, AST, UA, and Scr.

Model VI was a sensitivity analysis in participants without hypertension, SBP ≥ 180 mmHg, or DBP ≥ 90 mmHg (N = 13622). We adjusted for age, sex, SBP, BMI, DBP, ALT, GGT, ALB, FPG, AST, UA, and Scr.

HR, Hazard ratio; CI, confidence interval; Ref, reference; eGFR, evaluated glomerular filtration rate (mL/min·1.73 m²); NAFLD, non-alcoholic fatty liver disease.

3.6. Nonlinear relationship analysis

Utilizing a Cox proportional hazards regression model complemented by cubic spline functions, we discerned a nonlinear association between RC levels and the prevalence of NAFLD, as depicted in Fig. 4. To dissect this nonlinearity, a piecewise binary logistic regression modeling was engaged to delineate two separate gradients. This approach was juxtaposed with a conventional binary logistic regression model through the log-likelihood ratio test, as presented in Table 6, with the outcome indicating statistical significance (P < 0.001). Employing a recursive strategy, the critical inflection point of RC within the data was pinpointed at 98.29 mg/dL. Subsequently, a two-piecewise Cox proportional hazards regression modeling was applied, taking into account covariates such as UA, DBP, age, AST, SBP, sex, ALT, BMI, ALB, GGT, FPG, and Scr. This model was utilized to gauge the HR and its CI on either side of the identified inflection point. Our findings revealed a positive link between RC and NAFLD on the lower side of the curve's inflection point, as illustrated in Fig. 4. Specifically, the HR was 1.150 on the left (95% CI: 1.106–1.194), contrasting with an HR of 1.009 on the right side (95% CI: 0.982–1.037), showcasing a differential impact of RC levels on NAFLD risk across various thresholds.

Table 6

The result of the two-piecewise Cox regression model
Incident NAFLD	HR (95%CI) P-value
Fitting model by standard Cox regression	1.061 (1.045, 1.078) < 0.001
Fitting model by two-piecewise Cox regression
The inflection point of RC	98.29
< 98.29(per 10mg/dL)	1.150 (1.106, 1.194) < 0.001
≥ 98.29 (per 10mg/dL)	1.009 (0.982, 1.037) 0.521
P for log-likelihood ratio test	< 0.001

HR, Hazard ratio; CI, confidence interval; RC, remnant cholesterol. We adjusted for age, sex, SBP, BMI, DBP, ALT, GGT, ALB, FPG, AST, UA, and Scr levels.

The study, a retrospective cohort analysis, aimed to explore the link between RC levels and the likelihood of NAFLD emergence in individuals not classified as obese. Our investigation uncovered a positive, yet nonlinear, correlation between RC and the chance of NAFLD among non-obese Chinese adults, leading to the creation of a threshold effect curve. On the lower side of the curve's inflection point, identified at an RC level of 98.29 mg/dL, there existed a positive association between RC and NAFLD risk.

In prior investigations, researchers, employing ROC curves and logistic regression equations to analyze data from 14,251 adults undergoing routine health assessments, discerned that, following thorough confounding variable adjustment, individuals exhibiting elevated values of RC were connected to a significantly increased likelihood of NAFLD development (odds ratio (OR) 1.77 per SD increase, 95% confidence interval (CI): 1.64–1.91, P < 0.001), thereby establishing RC as a separate determinant of NAFLD ^[30]. Subsequent analysis of the NHANES database, encompassing 3,370 participants from 2017 to 2020, uncovered an independent positive nonlinear correlation between RC and NAFLD prevalence ^[36]. Nonetheless, current research does not provide an in-depth examination of the connection between RC and NAFLD in individuals who are not overweight. As far as we are aware, our investigation stands as a pioneering effort in substantiating the existence of a positive and nonlinear association between RC and NAFLD risk within a non-obese Asian demographic, presenting several distinctive merits: (a) The study population is distinct; while prior research has concentrated on the general populace undergoing periodic health assessments, the current investigation is centered on non-obese individuals receiving health evaluations in a hospital context. (b) The study's design and methodology diverge; this research is a retrospective cohort study, contrasting with the predominantly cross-sectional nature of previous studies. (c) The current investigation is among the rare studies to dissect the nonlinear relationship between RBP and NAFLD. (d) Adjustments were made for different variables. (e) A more extensive sample size was employed. Sensitivity analysis, conducted among participants with DBP ≥ 90 mmHg, SBP ≥ 140 mmHg, or FPG ≥ 7.0 mmol/L, substantiated the robustness of the RC-NAFLD risk correlation. Moreover, our findings indicate that for non-obese Chinese adults, the risk of NAFLD escalates by 14.2% and 6.1% for each 10 mg/dL increment in RC in unadjusted and fully adjusted Cox proportional hazards regression models, respectively. This suggests that mitigating RC levels could potentially curtail the risk of NAFLD, with elevated RC serving as a harbinger for future NAFLD risk, prompting preemptive lifestyle adjustments to curtail its prevalence. Consequently, the regulation of RC levels may emerge as a pivotal strategy in the prophylaxis and management of NAFLD, offering insights for the refinement of NAFLD prevention strategies and enhancing clinical diagnostics and therapeutics.

In addition to RC, we have identified a spectrum of independent risk factors for NAFLD, including FPG, age, BUN, GGT, Scr, LDL-c, ALT, AST, HDL-c, TC, TG, UA, BMI, SBP, and DBP. Clinically, the mitigation of NAFLD risk necessitates a multifaceted strategy that encompasses weight management and dietary fat regulation. Furthermore, an integrated approach, addressing age, hepatic and renal function, hyperuricemia, insulin resistance, diabetes, hyperlipidemia, and hypertension, can synergize to diminish the aggregate risk of NAFLD. The confluence of these interventions is anticipated to engender more holistic and efficacious NAFLD prevention protocols within clinical realms.

In the present study, the incidence of NAFLD among non-obese Chinese adults was determined to be 14.03%, a figure that closely mirrors the general Chinese population's rate of approximately 17.6% during the corresponding period^[4]. This congruence in NAFLD prevalence between non-obese individuals and the overall population is corroborated by studies from select Western nations, where a non-obese NAFLD prevalence of 51.3% has been documented in Europe^[15]. Such findings underscore the significant occurrence of NAFLD within the non-obese demographic, a phenomenon that, despite its prevalence, is often eclipsed by the more extensively studied obesity-related NAFLD^[1]. The etiology of non-obese NAFLD remains elusive and is frequently underappreciated in the medical community^[15]. Proposed contributing factors encompass genetic predisposition, the adoption of Western dietary and lifestyle practices, and the aging of the population^{[5, 42, 43]}.

The liver's pivotal role in the de novo synthesis of fatty acids (FA) is well-established; however, unchecked FA synthesis can precipitate the irregular accumulation of triglycerides, precipitating a metabolic imbalance. This metabolic disarray can engender lipotoxicity, and endoplasmic reticulum (ER) stress, and incite inflammatory pathways, culminating in the spectrum of conditions collectively termed NAFLD. Characterized by inflammation, hepatic steatosis, ballooning degeneration, and fibrosis, NAFLD represents a complex of chronic liver pathologies^[44]. The aberrant activity of genes implicated in hepatic FA synthesis, such as methylation-controlled J protein (MCJ), fatty acid synthase (Fasn), acetyl-CoA carboxylase 1 (Acc1), TANK-binding kinase 1 (TBK1), and ATP citrate lyase (Acly), is implicated in the excessive FA production and the consequent steatosis^[44–46]. Additionally, the regulatory function of Coenzyme A (CoA) and acyl-CoA oxidase 1 (Acox1) in the lipolysis and autophagy of the liver’s lipid droplets (LDs) is critical; dysregulation can result in the proliferation of smaller, yet abnormal LDs including very low-density lipoproteins (VLDL), inter-density lipoproteins (IDL), and chylomicrons (CM), ^[47]. In summary, the dysregulation of hepatic gene expression precipitates lipid accumulation anomalies. Studies have further indicated that conditions such as hyperlipidemia, and inflammation can perturb the ER's metabolic equilibrium within hepatocytes^[48]. The ER, a key modulator of liver protein and lipid homeostasis, when compromised, can incite chronic stress and hepatocyte dysfunction, thereby triggering inflammatory and cell death pathways associated with NAFLD^[49]. Additionally, specific modes of hepatocyte death, notably ferroptosis, have been shown to either initiate or exacerbate NAFLD pathogenesis^[50–53]. The interplay of these elements results in hepatic lipid metabolic disorders, instigating a cascade of lipotoxic and inflammatory events that may precipitate or intensify NAFLD.

RC is the metabolic residue generated from the metabolism of triglyceride-rich lipoproteins (TRLs), which include VLDL and IDL in the fasting state and CM in the non-fasting state. These particles are hydrolyzed by lipoprotein lipase (LPL), which loses TG and produces metabolic residues rich in cholesterol esters ^{[24, 54]}. The concept and calculation method of RC were introduced in the 2019 guidelines for managing dyslipidemia from the European Atherosclerosis Society ^[55]. Subsequent studies have shown that RC reflects the state of lipid metabolism and the level of inflammation due to lipotoxicity in those suffering from chronic kidney disease (CKD), cardiovascular disease, and stroke more effectively than traditional lipid indicators ^{[12, 24, 56, 57]}. Similarly, our research indicated that RC can represent the body's lipid levels and predict the occurrence of non-obese NAFLD.

A few restrictions apply to the current investigation. Firstly, the research design is retrospective, and because of the way the retrospective study is set up, it may lead to selection bias, incomplete information, difficulty in determining causality, and difficulty in controlling confounding factors, thereby affecting the reliability and generalizability of the study results. Second, the participants of this study were non-obese Chinese adults, and additional studies are needed on people of other races, regions, or age groups to verify the generalizability and representativeness of our study results. Third, the current study could not distinguish between various classifications of NAFLD, including non-alcoholic fatty liver (NAFL) and non-alcoholic steatohepatitis (NASH), so it is unclear what the association between RC and each type of NAFLD is in non-obese populations. Fourth, certain unmeasured confounding factors, such as genetic susceptibility, dietary habits, and lifestyle, may have affected how RC and NAFLD are related. Fifth, NAFLD within this research was determined using abdominal ultrasound, not the most accurate liver biopsy. Future studies that require more accurate diagnostic methods to improve the accuracy of NAFLD diagnosis are needed, along with further large-scale, multicenter, high-quality studies to verify whether there is also a positive nonlinear relationship in non-obese populations across all age groups, regions, and races.

Our research has uncovered a correlation that is both positive and nonlinear, linking RC levels to the prevalence of NAFLD among Chinese adults who are not classified as obese. The association in question is marked by a saturation effect, denoting a notably positive statistical link between NAFLD and RC levels that are beneath the threshold of 98.29 mg/dL. From a therapeutic perspective, reducing RC levels to below 98.29 mg/dL may significantly decrease the risk of NAFLD. Clinicians may use this finding as a strategy to predict and treat NAFLD.

ALB, Albumin

ALT, Alanine aminotransferase

AST, Aspartate aminotransferase

BMI, Body mass index

BUN, Serum urea nitrogen

CI, Confidence intervals

CKD, Chronic kidney disease

CM, Chylomicrons

DBP, Diastolic blood pressure

FPG; Fasting plasma glucose

GGT, Gamma-glutamyl transferase

GLB, Globulin

HDL-c, High-density lipoprotein cholesterol

HR, Hazard ratio

IDL, Inter-density lipoproteins

LDL-c, Low-density lipoprotein cholesterol

LPL, Lipoprotein lipase

NASH, Non-alcoholic steatohepatitis

NAFL, Non-alcoholic fatty liver

NAFLD, Non-alcoholic fatty liver disease

NHANES, National Health and Nutrition Examination Survey

OR, Odds ratio

RC, Remnant cholesterol

Ref, Reference

SBP, Systolic blood pressure

Scr, Serum creatinine

SD, Standard deviation

T2DM, Type 2 diabetes mellitus

TBIL, Total bilirubin

TC, Total cholesterol

TG, Triglyceride

TRLs, Triglyceride-rich lipoproteins

UA, Uric acid

VLDL, Very low-density lipoproteins

Authors’ contributions

Yan Zhou initiated the concept for this research and was responsible for the initial manuscript draft. Yong Han contributed by conducting the statistical analysis and also made subsequent revisions to the manuscript. Qing Shu played a role in both the editing of the manuscript and the formulation of the study's design. Ultimately, all authors participated in reviewing the manuscript and gave their approval to the final version.

Acknowledgments

This work presents a secondary analysis, with the foundational data and methodology primarily sourced from a preceding study conducted by D. Q. Sun, S. J. Wu, W. Y. Liu, et al. Their research, titled "Association of low-density lipoprotein cholesterol within the normal range and NAFLD in the non-obese Chinese population: a cross-sectional and longitudinal study," was published in BMJ Open, 6(12), e013781 (2016) and is accessible via doi:10.1136/bmjopen-2016-013781. Our appreciation is extended to the collective of researchers and contributors involved in the original study for their valuable contributions.

Data availability statement

The dataset supporting the findings of this study is accessible for download on the DATADRYAD repository, which can be visited at [www.Datadryad.org](www.Datadryad.org).

Publication consent

This section is not applicable to the present study.

Conflict of interest statement

The authors confirm that there are no conflicts of interest to report in relation to this research.

Ethical Considerations and Informed Consent

Adhering to the ethical standards set forth in the Declaration of Helsinki, this study was executed. Informed consent was obtained from all subjects. All personal identifiers were removed to ensure participant anonymity. The research protocol received approval from the Ethics Committee of Wenzhou People's Hospital ^[34].

Funding

The authors have reported the following financial interests in relation to the research, authorship, and publication of this manuscript: Financial support for this study was provided by a grant awarded by the Shenzhen Science and Technology Program, with the grant number being KCXFZ20201221173208023. It is important to note that the funding organizations did not participate in the study's design, data collection, analysis, interpretation, or writing of the manuscript.

Younossi, Z. et al. Global Perspectives on Nonalcoholic Fatty Liver Disease and Nonalcoholic Steatohepatitis[J]. (Baltimore, Md.), 69(6): 2672–2682. DOI: (2019). 10.1002/hep.30251
Golabi, P., Owrangi, S. & Younossi, Z. M. Global perspective on nonalcoholic fatty liver disease and nonalcoholic steatohepatitis - prevalence, clinical impact, economic implications and management strategies[J]59S1–S9 (Alimentary Pharmacology & Therapeutics, 2024). Suppl 110.1111/apt.17833
Li, T. et al. Empagliflozin Alleviates Hepatic Steatosis by Activating the AMPK-TET2-Autophagy Pathway in vivo and in vitro[J]. Front. Pharmacol. 11, 622153. 10.3389/fphar.2020.622153 (2021).
Estes, C. et al. Modeling NAFLD disease burden in China, France, Germany, Italy, Japan, Spain, United Kingdom, and United States for the period 2016–2030[J]. J. Hepatol. 69 (4), 896–904. 10.1016/j.jhep.2018.05.036 (2018).
Estes, C. et al. Modelling NAFLD disease burden in four Asian regions—2019-2030[J]. Aliment. Pharmacol. Ther. 51 (8), 801–811. 10.1111/apt.15673 (2020).
Powell, E. E., Wong, V. W. S. & Rinella, M. Non-alcoholic fatty liver disease[J]. Lancet (London England). 397 (10290), 2212–2224. 10.1016/S0140-6736(20)32511-3 (2021).
Amorim, R. et al. Exploratory Data Analysis of Cell and Mitochondrial High-Fat, High-Sugar Toxicity on Human HepG2 Cells[J]. Nutrients, 13(5): 1723. DOI:10.3390/nu13051723. (2021).
Abdelhameed, F. et al. Non-invasive Scores and Serum Biomarkers for Fatty Liver in the Era of Metabolic Dysfunction-associated Steatotic Liver Disease (MASLD): A Comprehensive Review From NAFLD to MAFLD and MASLD[J]. Curr. Obes. Rep. 10.1007/s13679-024-00574-z (2024).
Li, S., Mehal, W. Z. & Ouyang, X. RNA modifications in the progression of liver diseases: from fatty liver to cancer[J]. Sci. China Life Sci. 10.1007/s11427-023-2494-x (2024).
Lange, N. F. et al. PPAR-Targeted Therapies in the Treatment of Non-Alcoholic Fatty Liver Disease in Diabetic Patients[J]. Int. J. Mol. Sci. 23 (8), 4305. 10.3390/ijms23084305 (2022).
Rinella, M. E. et al. AASLD Practice Guidance on the clinical assessment and management of nonalcoholic fatty liver disease[J]. Hepatol. (Baltimore Md). 77 (5), 1797–1835. 10.1097/HEP.0000000000000323 (2023).
Duell, P. B. et al. Nonalcoholic Fatty Liver Disease and Cardiovascular Risk: A Scientific Statement From the American Heart Association[J]. Thromb. Vascular Biology, 42(6): e168–e185. DOI:10.1161/ATV.0000000000000153. (2022).
Sun, D. Q. et al. An international Delphi consensus statement on metabolic dysfunction-associated fatty liver disease and risk of chronic kidney disease[J]. Hepatobiliary Surg. Nutr. 12 (3), 386–403. 10.21037/hbsn-22-421 (2023).
Eslam, M. et al. A new definition for metabolic dysfunction-associated fatty liver disease: An international expert consensus statement[J]. J. Hepatol. 73 (1), 202–209. 10.1016/j.jhep.2020.03.039 (2020).
Ye, Q. et al. Global prevalence, incidence, and outcomes of non-obese or lean non-alcoholic fatty liver disease: a systematic review and meta-analysis[J]. Lancet Gastroenterol. Hepatol. 5 (8), 739–752. 10.1016/S2468-1253(20)30077-7 (2020).
Rinaldi, R. et al. Gender Differences in Liver Steatosis and Fibrosis in Overweight and Obese Patients with Metabolic Dysfunction-Associated Steatotic Liver Disease before and after 8 Weeks of Very Low-Calorie Ketogenic Diet[J]. Nutrients. 16 (10), 1408. 10.3390/nu16101408 (2024).
Dallio, M. et al. The influence of acute lifestyle changes on NAFLD evolution in a multicentre cohort: a matter of body composition[J]. Nutr. Diabetes. 14 (1), 33. 10.1038/s41387-024-00294-2 (2024).
Franco, I. et al. Leisure-Time Physical Activity, Time Spent Sitting and Risk of Non-alcoholic Fatty Liver Disease: A Cross-Sectional Study in Puglia[J]. J. Gen. Intern. Med. 10.1007/s11606-024-08804-9 (2024).
Neshatbini Tehrani, A. et al. The effect of soy isoflavones supplementation on metabolic status in patients with non-alcoholic fatty liver disease: a randomized placebo controlled clinical trial[J]. BMC Public. Health. 24, 1362. 10.1186/s12889-024-18812-3 (2024).
Hou, A. et al. Excessive fatty acids activate PRMT5/MDM2/Drosha pathway to regulate miRNA biogenesis and lipid metabolism[J]. Liver International: Official J. Int. Association Study Liver. 10.1111/liv.15906 (2024).
Lopez-Yus, M. et al. Unraveling Adipose Tissue Dysfunction: Molecular Mechanisms, Novel Biomarkers, and Therapeutic Targets for Liver Fat Deposition[J]. Cells. 13 (5), 380. 10.3390/cells13050380 (2024).
National Workshop on Fatty Liver and Alcoholic Liver Disease, Chinese Society of Hepatology, Chinese Medical Association, Fatty Liver Expert Committee, Chinese Medical Doctor Association. [Guidelines of prevention and treatment for nonalcoholic fatty liver disease: a 2018 update][J]. Zhonghua Gan Zang Bing Za Zhi = Zhonghua Ganzangbing Zazhi = Chin. J. Hepatol., 26(3): 195–203. DOI:10.3760/cma.j.issn.1007-3418.2018.03.008. (2018).
Bedogni, G., Palmese, F. & Foschi, F. G. Nonalcoholic fatty liver disease: an update[J]. Curr. Opin. Lipidol. 34 (3), 114. 10.1097/MOL.0000000000000874 (2023).
Cheng, Y. et al. Remnant cholesterol, stronger than triglycerides, is associated with incident non-alcoholic fatty liver disease[J]. Front. Endocrinol. 14, 1098078. 10.3389/fendo.2023.1098078 (2023).
Ginsberg, H. N. et al. Triglyceride-rich lipoproteins and their remnants: metabolic insights, role in atherosclerotic cardiovascular disease, and emerging therapeutic strategies—a consensus statement from the European Atherosclerosis Society[J]. Eur. Heart J. 42 (47), 4791–4806. 10.1093/eurheartj/ehab551 (2021).
Liu, Y. & Chen, M. Liver-adipose tissue crosstalk in non-alcoholic fatty liver disease: The emerging role of remnant cholesterol[J]. J. Hepatol. 80 (3), e111–e112. 10.1016/j.jhep.2023.06.006 (2024).
Castañer, O. et al. Remnant Cholesterol, Not LDL Cholesterol, Is Associated With Incident Cardiovascular Disease[J]. J. Am. Coll. Cardiol. 76 (23), 2712–2724. 10.1016/j.jacc.2020.10.008 (2020).
Huang, H. et al. Remnant Cholesterol Predicts Long-term Mortality of Patients With Metabolic Dysfunction-associated Fatty Liver Disease[J]. J. Clin. Endocrinol. Metab. 107 (8), e3295–e3303. 10.1210/clinem/dgac283 (2022).
Jia, Z., Pan, X. & Chen, S. The Combination of Remnant Cholesterol and Waist Circumference Offers Predictive Value for Nonalcoholic Fatty Liver Disease: A Cross-Sectional Study from a Medical Examination Center[J]. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy, 16: 3459–3469. DOI: (2023). 10.2147/DMSO.S429963
Zou, Y. et al. Association of remnant cholesterol with nonalcoholic fatty liver disease: a general population-based study[J]. Lipids Health Dis. 20 (1), 139. 10.1186/s12944-021-01573-y (2021).
Lu, S. et al. Lipid metabolism, BMI and the risk of nonalcoholic fatty liver disease in the general population: evidence from a mediation analysis[J]. J. Translational Med. 21 (1), 192. 10.1186/s12967-023-04047-0 (2023).
Huang, H. et al. Remnant cholesterol and severity of nonalcoholic fatty liver disease[J]. Diabetol. Metab. Syndr. 15 (1), 238. 10.1186/s13098-023-01220-9 (2023).
Wang, Y. et al. Serum remnant cholesterol is a potential predictor of liver stiffness in patients with nonalcoholic fatty liver disease[J]. Scand. J. Gastroenterol. 59 (4), 469–479. 10.1080/00365521.2023.2294693 (2024).
Sun, D. Q. et al. Association of low-density lipoprotein cholesterol within the normal range and NAFLD in the non-obese Chinese population: a cross-sectional and longitudinal study[J]. BMJ open. 6 (12), e013781. 10.1136/bmjopen-2016-013781 (2016).
Gao, Y., Hu, Y. & Xiang, L. Remnant cholesterol, but not other cholesterol parameters, is associated with gestational diabetes mellitus in pregnant women: a prospective cohort study[J]. J. Translational Med. 21 (1), 531. 10.1186/s12967-023-04322-0 (2023).
Chen, J. et al. Remnant cholesterol has a non-linear association with non-alcoholic fatty liver disease[J]. Diabetes Res. Clin. Pract. 201, 110733. 10.1016/j.diabres.2023.110733 (2023).
White, I. R., Royston, P. & Wood, A. M. Multiple imputation using chained equations: Issues and guidance for practice[J]. Stat. Med. 30 (4), 377–399. 10.1002/sim.4067 (2011).
Groenwold, R. H. H. et al. Missing covariate data in clinical research: when and when not to use the missing-indicator method for analysis[J]. CMAJ: Canadian Medical Association journal = journal de l’Association medicale canadienne, 184(11): 1265–1269. DOI: (2012). 10.1503/cmaj.110977
American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes-2024[J]. Diabetes Care. 47 (Suppl 1), S20–S42. 10.2337/dc24-S002 (2024).
Haneuse, S., VanderWeele, T. J. & Arterburn, D. Using the E-Value to Assess the Potential Effect of Unmeasured Confounding in Observational Studies[J]. JAMA. 321 (6), 602–603. 10.1001/jama.2018.21554 (2019).
Von Elm, E. et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: guidelines for reporting observational studies[J]. Int. J. Surg. (London England). 12 (12), 1495–1499. 10.1016/j.ijsu.2014.07.013 (2014).
Wong, V. W. S. et al. Asia-Pacific Working Party on Non-alcoholic Fatty Liver Disease guidelines 2017-Part 1: Definition, risk factors and assessment[J]. J. Gastroenterol. Hepatol. 33 (1), 70–85. 10.1111/jgh.13857 (2018).
Chan, W. K. et al. Clinical features and treatment of nonalcoholic fatty liver disease across the Asia Pacific region-the GO ASIA initiative[J]47816–825 (Alimentary Pharmacology & Therapeutics, 2018). 610.1111/apt.14506
Fang, T. et al. Mouse models of nonalcoholic fatty liver disease (NAFLD): pathomechanisms and pharmacotherapies[J]. Int. J. Biol. Sci. 18 (15), 5681–5697. 10.7150/ijbs.65044 (2022).
Huh, J. Y. et al. TANK-Binding Kinase 1 Regulates the Localization of Acyl-CoA Synthetase ACSL1 to Control Hepatic Fatty Acid Oxidation[J]. Cell Metabol., 32(6): 1012–1027e7. DOI:10.1016/j.cmet.2020.10.010. (2020).
Barbier-Torres, L. et al. Silencing hepatic MCJ attenuates non-alcoholic fatty liver disease (NAFLD) by increasing mitochondrial fatty acid oxidation[J]. Nat. Commun. 11 (1), 3360. 10.1038/s41467-020-16991-2 (2020).
He, A. et al. Acetyl-CoA Derived from Hepatic Peroxisomal β-Oxidation Inhibits Autophagy and Promotes Steatosis via mTORC1 Activation[J]. Mol. Cell. 79 (1), 30–42e4. 10.1016/j.molcel.2020.05.007 (2020).
Lebeaupin, C. et al. Endoplasmic reticulum stress signalling and the pathogenesis of non-alcoholic fatty liver disease[J]. J. Hepatol. 69 (4), 927–947. 10.1016/j.jhep.2018.06.008 (2018).
Walter, P. & Ron, D. The unfolded protein response: from stress pathway to homeostatic regulation[J]3341081–1086 (Science, 2011). (New York, N.Y.)605910.1126/science.1209038
Rostoker, G. et al. Liver Iron Load Influences Hepatic Fat Fraction in End-Stage Renal Disease Patients on Dialysis: A Proof of Concept Study[J]. EBioMedicine. 39, 461–471. 10.1016/j.ebiom.2018.11.020 (2019).
Tsurusaki, S. et al. Hepatic ferroptosis plays an important role as the trigger for initiating inflammation in nonalcoholic steatohepatitis[J]. Cell Death Dis. 10 (6), 449. 10.1038/s41419-019-1678-y (2019).
Li, X. et al. Targeting ferroptosis alleviates methionine-choline deficient (MCD)-diet induced NASH by suppressing liver lipotoxicity[J]. Liver International: Official J. Int. Association Study Liver. 40 (6), 1378–1394. 10.1111/liv.14428 (2020).
Qi, J. et al. Ferroptosis Affects the Progression of Nonalcoholic Steatohepatitis via the Modulation of Lipid Peroxidation-Mediated Cell Death in Mice[J]. Am. J. Pathol. 190 (1), 68–81. 10.1016/j.ajpath.2019.09.011 (2020).
Campanella, A. et al. The Effect of Three Mediterranean Diets on Remnant Cholesterol and Non-Alcoholic Fatty Liver Disease: A Secondary Analysis[J]. Nutrients, 12(6)[2024-06-28]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7352824/. DOI:10.3390/nu12061674. (2020).
Mach, F. et al. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: lipid modification to reduce cardiovascular risk[J]. Eur. Heart J. 41 (1), 111–188. 10.1093/eurheartj/ehz455 (2020).
Yuan, Y. et al. Remnant cholesterol, preinflammatory state and chronic kidney disease: association and mediation analyses[J]. Ren. Fail. 46 (2), 2361094. 10.1080/0886022X.2024.2361094 (2024).
Xiong, C. C. et al. Investigating the impact of remnant cholesterol on new-onset stroke across diverse inflammation levels: Insights from the China Health and Retirement Longitudinal Study (CHARLS)[J]. Int. J. Cardiol. 405, 131946. 10.1016/j.ijcard.2024.131946 (2024).

No competing interests reported.

Download PDF

Editor invited by journal
05 Sep, 2024
Submission checks completed at journal
05 Sep, 2024
First submitted to journal
22 Aug, 2024

You are reading this latest preprint version

Remnant cholesterol has a nonlinear association with non-alcoholic fatty liver disease: A secondary retrospective cohort study in non-obese Chinese adults

Status:

Version 1

Abstract

Objective

Methods

Results

Conclusion

Figures

1. Introduction

2. Method

2.1. Study design

2.2. Data source

2.3. Study population

2.4. Independent and outcome variables

2.5. Diagnosis of NAFLD

2.6. Covariates

2.7. Handling of missing data

2.8. Analytical statistics

3. Results

3.1. Features of the participants

3.2. The incidence rate of NAFLD

3.3. Single-variable analysis

3.4. Multivariate analysis

3.5. Sensitivity analysis

3.6. Nonlinear relationship analysis

4. Discussion

5. Limitations

6. Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1