Metabolic-associated fatty liver disease (MAFLD) is becoming increasingly prevalent in non-obese patients with type 2 diabetes mellitus (T2DM) and leads to serious liver damage in this population. The study aims to develop and validate a nomogram to predict the risk of MAFLD in non-overweight individuals with newly diagnosed T2DM.
Methods
A total of 2372 non-obese patients with newly diagnosed T2DM and MAFLD were enrolled and randomly assigned to the training and validation sets in a ratio of 7:3. The independent risk factors associated with MAFLD were screened by univariate and multivariate logistic regression, and a nomogram was constructed to predict the risk of MAFLD. Receiver operating characteristic curve (ROC), calibration curves, and decision curve analysis (DCA) were used to verify the performance and clinical utility of the model.
Results
Seven predictors, namely body mass index (BMI), alanine aminotransferase/aspartate aminotransferase (ALT/AST), triglyceride (TG), high-density lipoprotein-cholesterol (HDL-C), fasting blood glucose (FBG), creatinine (Cr) and serum uric acid (SUA), were identified by multivariate logistic regression analysis from a total of 14 variables studied. The nomogram built using these seven predictors showed good prediction ability (AUC: 0.815 in the training cohort; AUC: 0.787 in the validation cohort), along with favorable calibration and clinical utility.
Conclusion
The nomogram demonstrated effectiveness as a screening tool for evaluating the risk of MAFLD in T2DM individuals without obesity, facilitating early identification and supporting enhanced management strategies for MAFLD.
metabolic-associated fatty liver diseasetype 2 diabetes mellituswithout obesitynomogramrisk predictionThe author(s) declare that financial support was received for the research and/or publication of this article. This study was supported by Shandong Provincial Natural Science Foundation, China (ZR2021QH181).section-at-acceptanceClinical DiabetesIntroduction
Metabolic-associated fatty liver disease (MAFLD), formerly known as non-alcoholic fatty liver disease (NAFLD), is widely recognized as the most prevalent chronic liver disease that develops from excessive hepatic lipid accumulation and metabolic syndromes. MAFLD is defined by the presence of steatosis in more than 5% of hepatocytes, regardless of alcohol consumption or other concomitant liver diseases, and is strongly associated with obesity, type 2 diabetes mellitus (T2DM), and other metabolic disorders (1). The pathogenesis of MAFLD, according to the “multiple-hit” theory, involves various factors including insulin resistance, lipid accumulation, oxidative stress, endoplasmic reticulum stress, lipotoxicity, adipokines secreted from adipose tissue, nutritional factors, gut microbiota, and genetic and epigenetic influences (2, 3). It encompasses a wide spectrum of hepatic conditions, ranging from simple steatosis to non-alcoholic steatohepatitis (NASH), fibrosis and hepatocellular carcinoma (HCC) (4). In addition, MAFLD increases the occurrence and progression of extrahepatic diseases, such as cardiovascular and chronic kidney disease (5). A significant correlation between T2DM and NAFLD has been established, with more than 50% of individuals with T2DM diagnosed with NAFLD (6, 7). Insulin resistance (IR) appears to be a central pathophysiological mechanism shared by both conditions (8). Notably, IR may precede the diagnosis of T2DM and contribute to the development of various complications, including asymptomatic NAFLD in the early stage of T2DM (9). This mechanism helps explain the high prevalence of NAFLD in T2DM. Moreover, the presence of NAFLD is associated with glucose metabolism disorders (8) and a higher risk of advanced fibrosis (10–12) in patients with diabetes mellitus. Hence, it’s of great significance to identify MAFLD in T2DM populations.
Currently, MAFLD is diagnosed based on imaging evidence, assessment of liver histology and measurement of non-invasive biomarkers (13). Liver biopsy is the most accurate diagnostic technique for MAFLD, but is unsuitable for routine screening because it is invasive and challenging to perform (14, 15). Although ultrasonography is noninvasive, the subjectivity of visual assessments of fatty liver on gray-scale images leads to significant interobserver variability (16) and reduced sensitivity in detecting mild fatty liver (17). Moreover, it may not be routinely conducted in primary or secondary medical centers (18). Therefore, there is a significant need to develop a simple, non-invasive, and highly accurate predictive model for the rapid screening of MAFLD.
Nomograms have been widely regarded as a valuable tool for creating a simple and intuitive graph of a statistical predictive model that quantifies the risk of various diseases (19, 20), including MAFLD or NAFLD. For instance, a nomogram established in a Chinese population with T2DM could screen for NAFLD well but has unclear applicability to non-obese diabetes patients who exhibit unique metabolic profiles (21). Several studies have confirmed that a considerable number of non-overweight patients with T2DM suffer from NAFLD (22, 23). However, in clinical practice, NAFLD/MAFLD in this population is easily overlooked due to the absence of obesity as a conventional risk factor, leading to missed opportunities for early intervention. Furthermore, to the best of our knowledge, no studies have yet developed nomograms for predicting the risk of MAFLD specifically in non-obese patients with T2DM.
With this background, the present study aims to develop a nomogram-based, non-invasive model for quantitatively evaluating the risk of MAFLD in non-obese patients with newly diagnosed type 2 diabetes in a Chinese population.
Materials and methodsStudy design and participants
The retrospective cross-sectional study was conducted on non-obese patients with newly-diagnosed T2DM who visited the Department of Endocrinology at Qilu Hospital of Shandong University Dezhou Hospital between 2020 and 2024. The inclusion criteria were as follows: 1) Age ≥ 18 years, 2) BMI < 25 kg/m2 (24), 3) Patients with newly diagnosed T2DM according to the 1999 WHO criteria (25), who had not received treatment through exercise, diet, or medications before hospital admission, 4) ultrasound examination indicating fatty liver (13). 5) Patients without severe heart or kidney dysfunction, infections, other liver diseases such as drug-induced, viral or autoimmune hepatitis, and mental health issues. There was no stress hyperglycemia or transient hyperglycemia caused by other reasons. Specifically, the exclusion criteria for mental health issues were: 1) Patients with a confirmed psychiatric disorder, including depression, anxiety, bipolar disorder, schizophrenia, etc., 2) Individuals who have received pharmacological treatment or psychological interventions for mental health issues within the past 6 months, including the use of antidepressants (e.g., SSRIs), anxiolytics (e.g., benzodiazepines), or antipsychotic medications, 3) For participants without a clear diagnosis, the Patient Health Questionnaire-9 (PHQ-9) was used for screening, with a cutoff score of ≥10 indicating moderate to severe depressive symptoms, ensuring exclusion of individuals with significant symptoms but no formal diagnosis (26). The exclusion criteria for infections included: 1) Patients presenting with symptoms of acute respiratory, gastrointestinal, or urinary tract infections, such as fever (body temperature >37.3°C), cough, sputum production, abdominal pain, diarrhea, frequent urination, urgency, dysuria, or those with infections of the skin, soft tissues, or joints, 2) Patients with a white blood cell count >10×109/L or C-reactive protein (CRP) >10 mg/L, 3) Patients who have used antibiotics (e.g., cephalosporins, quinolones) or antiviral medications (e.g., oseltamivir) within the past 4 weeks.
The study was carried out in accordance with the Declaration of Helsinki and approved by the Ethical Committee of Qilu Hospital of Shandong University Dezhou Hospital (Ethical approval number: 2024123). All participants in this study provided their informed consent.
Data collection and definitions
Predictor variables were chosen based on their clinical importance and evidence related to MAFLD. The collected data contained demographic information (sex, age, course of T2DM and history of alcohol intake), anthropometric parameters (height, weight, systolic blood pressure [SBP] and diastolic blood pressure [DBP]). blood biochemical indexes (alanine aminotransferase/aspartate aminotransferase [ALT/AST], gamma-glutamyl transpeptidase [GGT], triglyceride [TG], high-density lipoprotein cholesterol [HDL-C], low-density lipoprotein cholesterol [LDL-C], fasting blood glucose [FBG], blood urea nitrogen [BUN], creatinine [Cr] and serum uric acid [SUA] (Table 1). In total, 14 variables were collected.
Measurement methods and reagents for biochemical parameters.
Biochemical parameters
The principles and methodologies of measurement
Reagents
ALT
Alanine aminotransferase activity was measured using the Alanine Aminotransferase Assay Kit (Lactate Dehydrogenase Method), following the IFCC-recommended protocol for enzymatic activity measurement. Absorbance at 340 nm was monitored, and the rate of change was used to determine alanine aminotransferase activity in the sample.
AST activity was measured by reacting L-aspartate with α-ketoglutarate to produce L-glutamate and oxaloacetate. Oxaloacetate was converted to L-malate by Malate Dehydrogenase (MDH), and the absorbance change at 340 nm was recorded. The rate of absorbance change was used to calculate AST activity.
GGT activity was measured using L-γ-glutamyl-3-carboxy-4-nitroanilide as the substrate, which reacts with glycylglycine to form 5-amino-2-nitrobenzoic acid. Absorbance at 410/480 nm was recorded, and the rate of change was used to calculate GGT activity.
Uric acid concentration was measured using a uricase-based assay. The formed hydrogen peroxide reacted with 4-aminophenazone and MADB in the presence of peroxidase to produce a chromophore. Absorbance at 660/800 nm was recorded, and the increase in absorbance was used to calculate the SUA concentration.
Glucose concentration was measured using a hexokinase-based assay. In the presence of ATP and magnesium ions, glucose was phosphorylated to glucose-6-phosphate. Glucose-6-phosphate was then oxidized by glucose-6-phosphate dehydrogenase (G6PDH), with the concurrent reduction of NAD+ to NADH. The increase in absorbance at 340 nm was measured to determine glucose concentration.
Triglyceride concentration was measured using an enzymatic assay. The sample was hydrolyzed by microbial lipases to release glycerol. Glycerol was phosphorylated by glycerol kinase (GK) to produce glycerol-3-phosphate, which was oxidized by glycerol phosphate oxidase (GPO). The hydrogen peroxide formed reacted with 4-aminophenazone and MADB to produce a chromophore. Absorbance at 660/800 nm was recorded, and the increase in absorbance was used to calculate triglyceride concentration.
Total cholesterol concentration was measured using an enzymatic assay. Cholesteryl esters were hydrolyzed by cholesterol esterase (CHE) to produce free cholesterol. The free cholesterol was oxidized by cholesterol oxidase (CHO) to form cholestene-3-one and hydrogen peroxide (H2O2). In the presence of peroxidase (POD), H2O2 reacted with 4-aminoantipyrine and phenol to form a chromophore. Absorbance at 540/600 nm was recorded, and the increase in absorbance was used to calculate total cholesterol concentration.
LDL concentration was measured using the CHO/PAP system. Reagent 1 protected LDL from enzymatic reactions, while non-LDL lipoproteins (HDL, VLDL, and CM) were decomposed by cholesterol oxidase (CHO) and cholesterol esterase (CHE). The hydrogen peroxide produced was broken down by catalase in reagent 1. Upon adding reagent 2, the protecting agent was released from LDL, and catalase was inactivated by sodium azide, allowing for the quantitative determination of LDL cholesterol.
HDL cholesterol was measured using an enzyme chromogen system. Reagent 1 contained anti-human-β-lipoprotein antibody, which combined with lipoproteins other than HDL (LDL, VLDL, and chylomicrons) to form insoluble antigen-antibody complexes. Upon adding reagent 2, these complexes blocked enzyme reactions, allowing for the quantitative determination of HDL cholesterol.
Urea concentration was measured using an enzymatic assay. Urease hydrolyzed urea to produce ammonia and carbon dioxide. Ammonia then reacted with NADH and 2-oxoglutarate in the presence of glutamate-dehydrogenase (GLDH) to form NAD+ and glutamate. The reduction in absorbance of NADH was measured to calculate urea concentration.
Creatinine concentration was measured using an enzymatic assay. Creatinine was hydrolyzed to creatine by creatininase, which was then hydrolyzed by creatinase to sarcosine and urea. Sarcosine oxidase catalyzed the conversion of sarcosine to glycine, formaldehyde, and hydrogen peroxide. The hydrogen peroxide reacted with 4-aminoantipyrine and HMMPS in the presence of peroxidase (POD) to form a blue pigment. Absorbance at 600/700 nm was measured to calculate creatinine concentration.
The quality of data collection was rigorously controlled by the following procedures. Blood pressure, including SBP and DBP, was measured on the right arm after the participants had rested in a seated position for 10 min. The serology indicators were evaluated in the morning after an overnight fast using an automatic biochemical analyzer (Type 7600, Hitachi Ltd. Tokyo, Japan). The body mass index (BMI) was calculated as weight (kg) divided by the square of height (m2).
The latest diagnostic criteria for MAFLD, described by Eslam et al. (13), were based on ultrasonographically confirmed hepatic steatosis along with one of the following conditions: overweight/obesity, T2DM, and metabolic dysregulation. Metabolic dysfunction was further defined as the presence of ≥ 2 of the following criteria (13): [i] waist circumference [WC] ≥ 90/80 cm in Asian men and women, respectively; [ii] blood pressure ≥ 130/85 mmHg or receiving specific medications; [iii] TG ≥ 1.7 mmol/L or receiving specific drug treatment; [iv] HDL-C < 1.0 mmol/L in men and < 1.3 mmol/L in women; [v] prediabetes (FBG of 5.6-6.9 mmol/L or 2-hour postload glucose level of 7.8-11.0 mmol/L or glycated hemoglobin A1c [HbA1c] of 5.7%-6.4%); [vi] insulin resistance index based on the steady-state model ≥ 2.5; and [vii] blood hypersensitive C-reactive protein (hsCRP) > 2 mg/L. The diagnosis of MAFLD in our study was based on ultrasonically confirmed steatosis of the liver in non-obese patients with T2DM.
All data analyses were performed with SPSS 27.0 (IBM Corp., Armonk, NY, USA) and R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria). P < 0.05 was considered statistically significant. The normality of continuous data was evaluated using the Kolmogorov-Smirnov test. Normally distributed variables were presented as mean ± standard deviation, and t-test was used for comparisons between two groups. For data that did not follow a normal distribution, values were expressed as median (25-75%), and group comparisons were performed using the Mann-Whitney U test.
The sample size was determined based on the Riley principle (27), using an adjusted Cox-Snell R² of 0.1 to ensure a sufficient sample size, reduce overfitting, and improve model robustness and generalizability. Accordingly, the dataset of 2372 patients with T2DM were randomly divided into a training set (1660 subjects) and a validation set (712 subjects) in a 7:3 ratio using the R caret package. We first conducted univariate logistic regression analyses to examine the crude associations between each candidate variable and MAFLD. Variables with a significance level of P < 0.05 were selected for further multivariable analysis to minimize the exclusion of potentially important predictors. We assessed multicollinearity among the variables using variance inflation factors (VIF), excluding those with VIF values greater than 5 in an iterative manner to ensure model stability. The results showed that all VIF values were <5, indicating no multicollinearity issues. We then applied multivariable logistic regression analysis to identify the most parsimonious set of independent predictors. Variables were retained in the final model if they achieved a significance level of P < 0.05. The goodness-of-fit of the final model was confirmed by the non-significant Hosmer-Lemeshow test (P = 0.25), a Nagelkerke R²value of 0.32, and a comprehensive test of model coefficients (P < 0.001), indicating adequate calibration. Finally, a nomogram based on the multivariate model incorporating the optimal predictors was developed to predict the risk of MAFLD. Additionally, we used the R pROC package to plot receiver operating characteristic (ROC) curves, with the area under the curve (AUC) applied to evaluate discrimination performance (28). Calibration curves were drawn using the R rms package to assess the concordance between the practical results and the predicted probabilities. Decision curve analysis (DCA) was conducted using the R rmda package to evaluate and compare predictive models, as well as to calculate the net benefits across threshold probabilities (29).
ResultsClinical characteristics of subjects
In total, 2372 participants with T2DM, including 1550 men (65.3%) and 822 women (34.7%), were finally enrolled. There were 1141 cases (48.1%) that had MAFLD among these patients based on the novel MAFLD diagnostic criteria. In our study, participants were randomly assigned to the training dataset (n = 1660) and the validation dataset (n = 712) in a 7:3 ratio. The basic characteristics of the two datasets were shown in Table 2. No significant differences were observed in any characteristics between the two datasets, indicating that the random grouping did not introduce bias. As shown in Table 3, in the training dataset, participants with MAFLD had higher BMI, SBP, DBP, ALT/AST, GGT, TG, LDL-C, FBG, SUA levels, and lower HDL-C, BUN, Cr concentrations than those without MAFLD (P < 0.05). Moreover, patients with MAFLD were younger than those without MAFLD (P < 0.05). No statistical differences were observed between genders.
Characteristics of participants in the training and validation datasets.
Variables
All patients (n = 2372)
Training dataset (n = 1660)
Validation dataset (n = 712)
P value
Male (%)
1550 (65.3%)
1074 (64.7%)
476 (66.85)
0.323
Age (yr)
58.55 ± 11.66
58.8 ± 11.59
57.99 ± 11.81
0.127
BMI (kg/m2)
23.01 ± 1.62
23.03 ± 1.63
22.96 ± 1.58
0.337
SBP (mmHg)
136.09 ± 19.8
136.04 ± 19.8
136.23 ± 19.81
0.833
DBP (mmHg)
80.21 ± 11.61
80.01 ± 11.76
80.66 ± 11.24
0.211
ALT/AST
0.89 (0.73-1.1)
0.89 (0.72-1.1)
0.88 (0.73-1.11)
0.886
GGT (IU/L)
24 (18-35)
24 (18-35)
24 (18-35)
0.826
TG (mmol/L)
1.37 (0.99-2.05)
1.37 (0.99-2.1)
1.36 (0.98-1.95)
0.288
HDL-C (mmol/L)
1.36 ± 0.33
1.36 ± 0.33
1.37 ± 0.32
0.767
LDL-C (mmol/L)
3.22 ± 0.88
3.23 ± 0.86
3.2 ± 0.89
0.622
FBG (mmol/L)
8.3 (7.4-10)
8.29 (7.4-9.9)
8.3 (7.4-10.1)
0.671
BUN (mmol/L)
5.52 ± 1.52
5.51 ± 1.53
5.53 ± 1.51
0.777
Cr (μmol/L)
64 (54-73)
64 (55-73.35)
64 (54-73)
0.28
SUA (μmol/L)
312 (260.25-368)
311 (260-370)
312.5 (262.3-364)
0.563
MAFLD (%)
1141 (48.1%)
805 (48.5%)
336 (4 7.2%)
0.591
Values in table are presented as the mean with the standard deviation (variables with a normal distribution) or median (25-75%) (variables with a non-normal distribution). BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; BUN, blood urea nitrogen; Cr, creatinine; SUA, serum uric acid; MAFLD, metabolic-associated fatty liver disease.
Baseline characteristics of MAFLD and without MAFLD patients in the training dataset.
Variables
Without MAFLD (n = 855)
With MAFLD (n = 805)
P value
Gender, male (%)
560 (65.49%)
500 (62.11%)
0.184
Age (yr)
60.63 ± 11.8
56.85 ± 11.04
<0.001
BMI (kg/m2)
22.57 ± 1.82
23.52 ± 1.23
<0.001
SBP (mmHg)
134.89 ± 19.73
137.24 ± 19.81
0.016
DBP (mmHg)
78.4 ± 11.42
81.73 ± 11.89
<0.001
ALT/AST
0.83 (0.68-1.0)
0.97 (0.81-1.21)
<0.001
GGT (IU/L)
21 (16-28)
28 (21-43)
<0.001
TG (mmol/L)
1.14 (0.84-1.62)
1.74 (1.24-2.66)
<0.001
HDL-C (mmol/L)
1.42 ± 0.33
1.3 ± 0.32
<0.001
LDL-C (mmol/L)
3.1 ± 0.84
3.35 ± 0.89
<0.001
FBG (mmol/L)
8.1 (7.3-9.5)
8.5 (7.5-10.7)
<0.001
BUN (mmol/L)
5.59 ± 1.61
5.42 ± 1.44
0.028
Cr (μmol/L)
65 (55-75)
63 (53-72)
0.001
SUA (μmol/L)
300 (250-355)
324 (275-383)
<0.001
Values in table are presented as the mean with the standard deviation (variables with a normal distribution) or median (25-75%) (variables with a non-normal distribution). The bold font indicates that P values are statistically significant. BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; BUN, blood urea nitrogen; Cr, creatinine; SUA, serum uric acid.
Values in table are presented as the mean with the standard deviation (variables with a normal distribution) or median (25-75%) (variables with a non-normal distribution). BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; BUN, blood urea nitrogen; Cr, creatinine; SUA, serum uric acid; MAFLD, metabolic-associated fatty liver disease.
Values in table are presented as the mean with the standard deviation (variables with a normal distribution) or median (25-75%) (variables with a non-normal distribution). The bold font indicates that P values are statistically significant. BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; BUN, blood urea nitrogen; Cr, creatinine; SUA, serum uric acid.
Identifying predictors and constructing a nomogram for MAFLD
Based on the univariate analysis (Table 4), we selected candidate variables with p < 0.05 for inclusion in the multivariate logistic regression analysis. As shown in Table 4, seven of the original 14 variables, namely BMI, ALT/AST, TG, HDL-C, FBG, Cr and SUA, showed significant statistical differences. These variables were identified as independent risk factors for MAFLD among the patients with T2DM and were introduced into the predictive model to develop a MAFLD risk nomogram (Figure 1).
Univariate and multivariate analysis for the prediction of MAFLD.
Variables
Univariate logistic regression analysis
Multivariate logistic regression analysis
OR
95% CI
P
OR
95% CI
P
Gender (male/female)
0.892
0.729-1.091
0.266
Age (yr)
0.972
0.963-0.98
<0.001
BMI (kg/m2)
1.517
1.41-1.632
<0.001
1.468
1.348-1.598
<0.001
SBP (mmHg)
1.006
1.001-1.011
0.016
DBP (mmHg)
1.025
1.016-1.034
<0.001
ALT/AST
5.996
4.201-8.559
<0.001
3.571
2.381-5.357
<0.001
GGT (IU/L)
1.014
1.01-1.019
<0.001
TG (mmol/L)
2.034
1.804-2.293
<0.001
1.516
1.332-1.752
<0.001
HDL-C (mmol/L)
0.364
0.268-0.494
<0.001
0.491
0.324-0.743
0.001
LDL-C (mmol/L)
1.385
1.237-1.551
<0.001
1.249
1.071-1.456
0.05
FBG (mmol/L)
1.093
1.055-1.134
<0.001
1.059
1.012-1.109
0.016
BUN (mmol/L)
0.931
0.874-0.993
0.029
Cr (μmol/L)
0.99
0.985-0.996
0.002
0.98
0.971-0.988
<0.001
SUA (μmol/L)
1.004
1.002-1.005
<0.001
1.003
1.001-1.005
<0.001
The bold font indicates that P values are statistically significant. BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; ALT/AST, the ratio of alanine aminotransferase to aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; BUN, blood urea nitrogen; Cr, creatinine; SUA, serum uric acid.
Nomogram for predicting MAFLD in non-obese patients with T2DM. Using the nomogram, the corresponding points for each variable are added to obtain the total score. A vertical line is then drawn from the total points axis to the MAFLD risk axis to determine the predicted risk value.
Each risk factor corresponded to a score on the first row of the scale. The scores of all factors were summed to obtain a total score. The higher the total score, the greater the probability of developing MAFLD for an individual. For example, using the nomogram model, a 43-year-old male patient with T2DM, BMI of 24.82 kg/m2, ALT/AST of 1.25, TG of 2.99 mmol/L, HDL-C of 1.21 mmol/L, FBG of 9 mmol/L, Cr of 63 μmol/L and SUA of 285 μmol/L has an estimated 80% probability of developing MAFLD.
The bold font indicates that P values are statistically significant. BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; ALT/AST, the ratio of alanine aminotransferase to aspartate aminotransferase; GGT, gamma-glutamyl transpeptidase; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; FBG, fasting blood glucose; BUN, blood urea nitrogen; Cr, creatinine; SUA, serum uric acid.
Validation of the nomogram
The ROC curve was used to evaluate the predictive accuracy of the model. The results showed that the area under the ROC curve (AUC) for the training and validation groups were 0.815 (95% confidence interval 0.772-0.858) and 0.787 (95% confidence interval 0.754-0.820), respectively (Figure 2), indicating a moderately good performance. Next, a calibration curve was employed to assess the deviation between the predicted and actual values. The predicted results indicated that there was good agreement between the training and validation cohorts (Figure 3). The DCA curve demonstrated that this model provided greater net benefits for predicting MAFLD risk compared to the “all” or “none” strategies within a threshold probability range of almost 0.1 to 1.0 in both the training and validation sets (Figure 4).
ROC curve of the predictive model and in the training cohort (left) and validation cohort (right).
Calibration curve of the predictive model in the training cohort (left) and validation cohort (right). The X-axis represents the predicted risk of MAFLD in non-obese populations with T2DM. The Y-axis represents the actual occurrence rate of MAFLD in non-obese populations with T2DM.
DCA of the predictive model in the training cohort (left) and validation cohort (right). The y-axis measures the net benefit. The thick solid line represents the assumption that all patients have no MAFLD, the thin solid line represents the assumption that all patients have MAFLD, the red line represents the risk nomogram.
Discussion
In our study, the incidence rate of MAFLD in non-obese T2DM patients was 48.1%, which was higher than the reported 25.5% prevalence in hospitalized Chinese counterparts (30). The association between T2DM and NAFLD is primarily mediated by IR and the resulting hyperinsulinemia, which arise from β-cell dysfunction in T2DM (31). IR, a key pathophysiological feature of T2DM, often precedes the clinical diagnosis, with newly diagnosed individuals frequently presenting with IR-related complications (32). Moreover, IR contributes to hepatic steatosis by promoting de novo lipogenesis, enhancing fat accumulation, and amplifying hepatic oxidative stress and inflammation (33). As such, it is unsurprising that a substantial proportion of T2DM patients also present with NAFLD.
We identified seven independent risk variables for MAFLD in non-obese patients with T2DM, namely BMI, ALT/AST, TG, HDL-C, FBG, Cr, and SUA. In our study, non-obese participants with T2DM and MAFLD had significantly higher BMI than non-MAFLD controls. BMI has been identified as an independent risk factor for MAFLD in non-obese individuals with T2DM (30), which is consistent with the findings of our study. The increases in ALT and AST, which are liver enzymes, vary with the degree and duration of liver diseases, so the ALT/AST ratio has considerable clinical significance in diagnosing liver disease (34, 35). The ALT/AST ratio is related to metabolic syndrome and can better reflect the fat accumulation in the liver than the traditional liver enzyme index (36, 37). Large amounts of studies showed that a high ALT/AST ratio was a significant risk factor for the development and severity of NAFLD (35, 38–40). Our research also revealed that ALT/AST was an independent risk factor for MAFLD in T2DM patients without obesity, and further studies are needed to explore the underlying mechanisms in this population.
Our results indicated that non-obese individuals with T2DM and MAFLD had higher TG and lower HDL-C levels than those without MAFLD as Dang et al. presented (30). Furthermore, some studies have proved that the ratio of TG to HDL-C is independently related to insulin resistance, metabolic syndrome and NAFLD (41, 42). In this context, the prevalence rate of NAFLD was 33.41% among patients with the lowest TG/HDL-C ratios, compared to 78.04% in those with the highest ratios (7). In addition, low levels of HDL-C were associated with an increased risk of T2DM (43, 44), potentially due to the role of HDL-C in influencing pancreatic β-cell function and glucose metabolism (45). An increasing body of data highlighted a link between SUA and NAFLD (46–48). Previous studies have suggested SUA is an independent risk factor for NAFLD in non-obese subjects. One retrospective cohort study performed by Eshraghian et al. indicated a positive association between SUA and NAFLD in lean Iranian population (49). In a study involving 95924 subjects from a population in China, they found elevated SUA levels were related to increased risk of NAFLD, independent of other metabolic factors (50). Similarly, our previous finding revealed that enhanced SUA was significantly linked to greater risk of NAFLD in non-obese patients with T2DM (24). This is consistent with our present observations. The mechanisms underlying the positive relationship between uric acid and NAFLD remain poorly understood. Elevated SUA levels can lead to the development of insulin resistance and promote triglyceride accumulation in hepatocytes (51). Additionally, SUA is associated with mitochondrial oxidative stress, playing a crucial role in hepatic steatosis induced by uric acid (52). Hence, all suspected or diagnosed non-obese MAFLD patients, especially those with T2DM, should be tested for SUA and provided with appropriate management for elevated levels. Besides, low Cr levels are common in patients with MAFLD/NAFLD. The SUA/Cr ratio was significantly elevated in subjects with MAFLD/NAFLD, and it was independently associated with the risk of MAFLD/NAFLD development (53, 54). Our study also confirmed a clear association between FBG and the high risk of MAFLD. Some researchers have proposed that FBG is a risk factor of liver fibrosis in MAFLD patients (55).
Due to the lack of specific clinical symptoms for MAFLD, it is often discovered incidentally during tests for other diseases or annual physical examinations. Our model can be easily developed into a web page or electronically on computer to help clinicians quickly assess MAFLD risk in patients. As detailed in the results section, if a patient has an 80% probability of MAFLD, which is above 46.7% (the ROC curve cut-off value in our study), they should be classified as high-risk, prompting immediate imaging and pharmacological intervention. Thus, using this novel approach, clinical workers can quickly and accurately identify subjects potentially at risk for MAFLD.
Several advantages of our study are worth mentioning. First, although several models for diagnosing NAFLD, such as the Fatty Liver Index (FLI) and Hepatic Steatosis Index (HSI), have been developed, they are invalid predictors of steatosis in patients with T2DM (56). Moreover, these models involved complicated formulas that limited their practical application and ease of use in clinical settings. However, the nomogram provides a simple, visual tool for estimating the risk of MAFLD based on specific variables, making it both effective and easy to use. It also shows how changes in risk factor values can affect the prevalence of MAFLD. In addition, our nomogram is the first to predict MAFLD risk in non-obese individuals with T2DM and may compensate for some limitations of previous MAFLD screening tools. For example, the nomograms developed by Song et al. (57) for predicting MAFLD risk in overweight and obese populations, by Zhu et al. (58) for lean populations, and by Xue et al. (59) for T2DM populations, are not applicable to non-obese individuals with T2DM. The FLI index has been proposed for predicting fatty liver in lean individuals, but it is not only unsuitable for T2DM patients (60). Furthermore, the clinical and laboratory nomogram (CLN) model for predicting NAFLD required improvements in sensitivity and specificity, and it was not applicable to the newly defined condition of MAFLD.
There are some limitations in our study. First, the cross-sectional nature of our study prevents us from establishing a causal relationship between risk factors and MAFLD. Second, the diagnosis of MAFLD was based on steatosis detected by liver ultrasonography rather than biopsy, as performing liver biopsies on every patients is impractical. However, ultrasound diagnosis of MAFLD has some shortcomings, particularly its relatively low sensitivity for mild steatosis when fat accumulation is below 30%. Furthermore, the diagnostic accuracy of ultrasound can be influenced by the technical levels of operators, potentially leading to undiagnosed cases of mild fatty liver. We recognize that advanced techniques like transient elastography, MRI proton density fat fraction (PDFF), and liver biopsy offer higher accuracy, and we plan to incorporate them in future follow-up studies to enhance diagnostic accuracy and validate our findings. Third, data of waist circumference, dietary habits, physical activity status and alcohol intake were not included in this analysis. The impacts of these characteristics on the development of MAFLD cannot be assessed. Fourth, this study had a retrospective design, and the sample size was limited due to strict inclusion and exclusion criteria, which inevitably made it susceptible to selection bias. Finally, this study used an internal dataset for model validation, which may limit the generalizability of the findings, highlighting the need for future multi-center studies to externally validate the model across diverse cohorts and settings.
Conclusion
The seven indicators confirmed by the nomogram in this analysis—BMI, ALT/AST, TG, HDL-C, FBG, Cr and SUA—are important for assessing MAFLD risk in non-obese patients with T2DM. These indicators also contribute to early screening and the prevention of related complications. Thus, introducing them in the risk nomogram is valuable for predicting MAFLD risk in non-obese individuals with T2DM.
Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Ethics statement
The studies involving humans were approved by the Ethical Committee of Qilu Hospital of Shandong University Dezhou Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declare that no Generative AI was used in the creation of this manuscript.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
ReferencesEslamMSanyalAJGeorgeJ.
MAFLD: A consensus-driven proposed nomenclature for metabolic associated fatty liver disease. Gastroenterology+. (2020) 158:1999–2014. doi: 10.1053/j.gastro.2019.11.312GuoXYinXLiuZWangJ.
Non-alcoholic fatty liver disease (NAFLD) pathogenesis and natural products for prevention and treatment. Int J Mol Sci. (2022) 23(24):15489. doi: 10.3390/ijms232415489PafiliKRodenM.
Nonalcoholic fatty liver disease (NAFLD) from pathogenesis to treatment concepts in humans. Mol Metab. (2021) 50:101122. doi: 10.1016/j.molmet.2020.101122EslamMSarinSKWongVWFanJGKawaguchiTAhnSH.
The Asian Pacific Association for the Study of the Liver clinical practice guidelines for the diagnosis and management of metabolic associated fatty liver disease. Hepatol Int. (2020) 14:889–919. doi: 10.1007/s12072-020-10094-2LeeHLeeYHKimSUKimHC.
Metabolic dysfunction-associated fatty liver disease and incident cardiovascular disease risk: A nationwide cohort study. Clin Gastroenterol Hepatol. (2021) 19:2138–2147.e10. doi: 10.1016/j.cgh.2020.12.022PowellEEWongVWRinellaM.
Non-alcoholic fatty liver disease. Lancet+. (2021) 397:2212–24. doi: 10.1016/s0140-6736(20)32511-3YounossiZMGolabiPde AvilaLPaikJMSrishordMFukuiN.
The global epidemiology of NAFLD and NASH in patients with type 2 diabetes: A systematic review and meta-analysis. J Hepatol. (2019) 71:793–801. doi: 10.1016/j.jhep.2019.06.021TilgHMoschenARRodenM.
NAFLD and diabetes mellitus. Nat Rev Gastroenterol Hepatol. (2017) 14:32–42. doi: 10.1038/nrgastro.2016.147SaponaroCGagginiMGastaldelliA.
Nonalcoholic fatty liver disease and type 2 diabetes: common pathophysiologic mechanisms. Curr Diabetes Rep. (2015) 15:607. doi: 10.1007/s11892-015-0607-4LiangYChenHLiuYHouXWeiLBaoY.
Association of MAFLD with diabetes, chronic kidney disease, and cardiovascular disease: A 4.6-year cohort study in China. J Clin Endocrinol Metab. (2022) 107:88–97. doi: 10.1210/clinem/dgab641PuchakayalaBKVermaSKanwarPHartJSanivarapuRRMohantySR.
Histopathological differences utilizing the nonalcoholic fatty liver disease activity score criteria in diabetic (type 2 diabetes mellitus) and non-diabetic patients with nonalcoholic fatty liver disease. World J Hepatol. (2015) 7:2610–8. doi: 10.4254/wjh.v7.i25.2610KwokRChoiKCWongGLZhangYChanHLLukAO.
Screening diabetic patients for non-alcoholic fatty liver disease with controlled attenuation parameter and liver stiffness measurements: a prospective cohort study. Gut+. (2016) 65:1359–68. doi: 10.1136/gutjnl-2015-309265EslamMNewsomePNSarinSKAnsteeQMTargherGRomero-GomezM.
A new definition for metabolic dysfunction-associated fatty liver disease: An international expert consensus statement. J Hepatol. (2020) 73:202–9. doi: 10.1016/j.jhep.2020.03.039ChalasaniNYounossiZLavineJEDiehlAMBruntEMCusiK.
The diagnosis and management of non-alcoholic fatty liver disease: practice guideline by the American Gastroenterological Association, American Association for the Study of Liver Diseases, and American College of Gastroenterology. Gastroenterology+. (2012) 142:1592–609. doi: 10.1053/j.gastro.2012.04.001ZhouXLinXChenJPuJWuWWuZ.
Clinical spectrum transition and prediction model of nonalcoholic fatty liver disease in children with obesity. Front Endocrinol (Lausanne). (2022) 13:986841. doi: 10.3389/fendo.2022.986841StraussSGavishEGottliebPKatsnelsonL.
Interobserver and intraobserver variability in the sonographic assessment of fatty liver. AJR Am J Roentgenol. (2007) 189:W320–3. doi: 10.2214/ajr.07.2123TanCHVenkateshSK.
Magnetic resonance elastography and other magnetic resonance imaging techniques in chronic liver disease: current status and future directions. Gut Liver. (2016) 10:672–86. doi: 10.5009/gnl15492HernaezRLazoMBonekampSKamelIBrancatiFLGuallarE.
Diagnostic accuracy and reliability of ultrasonography for the detection of fatty liver: a meta-analysis. Hepatology+. (2011) 54:1082–90. doi: 10.1002/hep.24452ZhangYSunYXiangJZhangYHuHChenH.
A clinicopathologic prediction model for postoperative recurrence in stage Ia non-small cell lung cancer. J Thorac Cardiovasc Surg. (2014) 148:1193–9. doi: 10.1016/j.jtcvs.2014.02.064LiangWZhangLJiangGWangQLiuLLiuD.
Development and validation of a nomogram for predicting survival in patients with resected non-small-cell lung cancer. J Clin Oncol. (2015) 33:861–9. doi: 10.1200/jco.2014.56.6661ZhangYShiRYuLJiLLiMHuF.
Establishment of a risk prediction model for non-alcoholic fatty liver disease in type 2 diabetes. Diabetes Ther. (2020) 11:2057–73. doi: 10.1007/s13300-020-00893-zEASL-EASD-EASO. Clinical practice guidelines for the management of non-alcoholic fatty liver disease. Obes Facts. (2016) 9:65–90. doi: 10.1159/000443344ChalasaniNYounossiZLavineJECharltonMCusiKRinellaM.
The diagnosis and management of nonalcoholic fatty liver disease: Practice guidance from the American Association for the Study of Liver Diseases. Hepatology+. (2018) 67:328–57. doi: 10.1002/hep.29367CuiYLiuJShiHHuWSongLZhaoQ.
Serum uric acid is positively associated with the prevalence of nonalcoholic fatty liver in non-obese type 2 diabetes patients in a Chinese population. J Diabetes Complications. (2021) 35:107874. doi: 10.1016/j.jdiacomp.2021.107874AlbertiKGZimmetPZ.
Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus. Diabet Med. (1998) 15(7):539–53. doi: 10.1002/(SICI)1096-9136(199807)15:7<539::AID-DIA668>3.0.CO;2-SKroenkeKSpitzerRLWilliamsJB.
The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. (2001) 16:606–13. doi: 10.1046/j.1525-1497.2001.016009606.xRileyRD.
Correction to: Minimum sample size for developing a multivariable prediction model: Part II-binary and time-to-event outcomes by Riley RD, Snell KI, Ensor J, et al. Stat Med. (2019) 38:5672. doi: 10.1002/sim.8409LättiSNiinivehmasSPentikäinenOT.
Rocker: Open source, easy-to-use tool for AUC and enrichment calculations and ROC visualization. J Cheminform. (2016) 8:45. doi: 10.1186/s13321-016-0158-yVickersAJElkinEB.
Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making. (2006) 26:565–74. doi: 10.1177/0272989x06295361DangSWGaoLLiYJZhangRXuJ.
Metabolic characteristics of non-obese and obese metabolic dysfunction-associated fatty liver disease in type 2 diabetes mellitus and its association with diabetic peripheral neuropathy and diabetic retinopathy. Front Med (Lausanne). (2023) 10:1216412. doi: 10.3389/fmed.2023.1216412TanaseDMGosavEMCosteaCFCiocoiuMLacatusuCMMaranducaMA.
The intricate relationship between type 2 diabetes mellitus (T2DM), insulin resistance (IR), and nonalcoholic fatty liver disease (NAFLD). J Diabetes Res. (2020) 2020:3920196. doi: 10.1155/2020/3920196Galicia-GarciaUBenito-VicenteAJebariSLarrea-SebalASiddiqiHUribeKB.
Pathophysiology of type 2 diabetes mellitus. Int J Mol Sci. (2020) 21(17):6275. doi: 10.3390/ijms21176275SakuraiYKubotaNYamauchiTKadowakiT.
Role of insulin resistance in MAFLD. Int J Mol Sci. (2021) 22(8):4156. doi: 10.3390/ijms22084156McPhersonSStewartSFHendersonEBurtADDayCP.
Simple non-invasive fibrosis scoring systems can reliably exclude advanced fibrosis in patients with non-alcoholic fatty liver disease. Gut+. (2010) 59:1265–9. doi: 10.1136/gut.2010.216077ZouYZhongLHuCShengG.
Association between the alanine aminotransferase/aspartate aminotransferase ratio and new-onset non-alcoholic fatty liver disease in a nonobese Chinese population: a population-based longitudinal study. Lipids Health Dis. (2020) 19:245. doi: 10.1186/s12944-020-01419-zRinellaME.
Nonalcoholic fatty liver disease: a systematic review. Jama+. (2015) 313:2263–73. doi: 10.1001/jama.2015.5370LinMSLinHSChungCMLinYSChenMYChenPH.
Serum aminotransferase ratio is independently correlated with hepatosteatosis in patients with HCV: a cross-sectional observational study. BMJ Open. (2015) 5:e008797. doi: 10.1136/bmjopen-2015-008797LeeJHKimDKimHJLeeCHYangJIKimW.
Hepatic steatosis index: a simple screening tool reflecting nonalcoholic fatty liver disease. Dig Liver Dis. (2010) 42:503–8. doi: 10.1016/j.dld.2009.08.002FuyanSJingLWenjunCZhijunTWeijingMSuzhenW.
Fatty liver disease index: a simple screening tool to facilitate diagnosis of nonalcoholic fatty liver disease in the Chinese population. Dig Dis Sci. (2013) 58:3326–34. doi: 10.1007/s10620-013-2774-yGuanHShaoGChengFNiPWuM.
Risk factors of nonalcoholic fatty liver disease in healthy women. Med (Baltimore). (2023) 102:e34437. doi: 10.1097/md.0000000000034437LiNTanHXieALiCFuXXangW.
Value of the triglyceride glucose index combined with body mass index in identifying non-alcoholic fatty liver disease in patients with type 2 diabetes. BMC Endocr Disord. (2022) 22:101. doi: 10.1186/s12902-022-00993-wBabicNValjevacAZaciragicAAvdagicNZukicSHasicS.
The triglyceride/HDL ratio and triglyceride glucose index as predictors of glycemic control in patients with diabetes mellitus type 2. Med Arch. (2019) 73:163–8. doi: 10.5455/medarh.2019.73.163-168SchmidtMIDuncanBBBangHPankowJSBallantyneCMGoldenSH.
Identifying individuals at high risk for diabetes: The Atherosclerosis Risk in Communities study. Diabetes Care. (2005) 28:2013–8. doi: 10.2337/diacare.28.8.2013WilsonPWMeigsJBSullivanLFoxCSNathanDMD’AgostinoRB C.OMMAS.R.X.X.X.Prediction of incident diabetes mellitus in middle-aged adults: the Framingham Offspring Study. Arch Intern Med. (2007) 167:1068–74. doi: 10.1001/archinte.167.10.1068FemlakMGluba-BrzózkaACiałkowska-RyszARyszJ.
The role and function of HDL in patients with diabetes mellitus and the related cardiovascular risk. Lipids Health Dis. (2017) 16:207. doi: 10.1186/s12944-017-0594-3LombardiRPisanoGFargionS.
Role of Serum Uric Acid and Ferritin in the Development and Progression of NAFLD. Int J Mol Sci. (2016) 17(4):548. doi: 10.3390/ijms17040548YangHLiDSongXLiuFWangXMaQ.
Joint associations of serum uric acid and ALT with NAFLD in elderly men and women: a Chinese cross-sectional study. J Transl Med. (2018) 16:285. doi: 10.1186/s12967-018-1657-6Zelber-SagiSBen-AssuliORabinowichLShalevVShiboletOChodickG.
Response to The relationship between serum uric acid levels and NAFLD. Liver Int. (2016) 36:769–70. doi: 10.1111/liv.13101EshraghianANikeghbalianSGeramizadehBKazemiKShamsaeefarAMalek-HosseiniSA.
Characterization of biopsy proven non-alcoholic fatty liver disease in healthy non-obese and lean population of living liver donors: The impact of uric acid. Clin Res Hepatol Gastroenterol. (2020) 44:572–8. doi: 10.1016/j.clinre.2019.09.002ZhengXGongLLuoRChenHPengBRenW.
Serum uric acid and non-alcoholic fatty liver disease in non-obesity Chinese adults. Lipids Health Dis. (2017) 16:202. doi: 10.1186/s12944-017-0531-5ChoiYJShinHSChoiHSParkJWJoIOhES.
Uric acid induces fat accumulation via generation of endoplasmic reticulum stress and SREBP-1c activation in hepatocytes. Lab Invest. (2014) 94:1114–25. doi: 10.1038/labinvest.2014.98LanaspaMASanchez-LozadaLGChoiYJCicerchiCKanbayMRoncal-JimenezCA.
Uric acid induces hepatic steatosis by generation of mitochondrial oxidative stress: potential role in fructose-dependent and -independent fatty liver. J Biol Chem. (2012) 287:40732–44. doi: 10.1074/jbc.M112.399899XingYChenJLiuJSongGMaH.
Relationship between serum uric acid-to-creatinine ratio and the risk of metabolic-associated fatty liver disease in patients with type 2 diabetes mellitus. Diabetes Metab Syndr Obes. (2022) 15:257–67. doi: 10.2147/dmso.S350468WangXHanYLiuYHuH.
Association between serum uric acid-to-creatinine ratio and non-alcoholic fatty liver disease: a cross-sectional study in Chinese non-obese people with a normal range of low-density lipoprotein cholesterol. BMC Gastroenterol. (2022) 22:419. doi: 10.1186/s12876-022-02500-wChaoGZhuYBaoY.
A screening study of high-risk groups for liver fibrosis in patients with metabolic dysfunction-associated fatty liver disease. Sci Rep. (2024) 14:23714. doi: 10.1038/s41598-024-74792-9GuiuBCrevisy-GirodEBinquetCDuvillardLMassonDLepageC.
Prediction for steatosis in type-2 diabetes: clinico-biological markers versus 1H-MR spectroscopy. Eur Radiol. (2012) 22:855–63. doi: 10.1007/s00330-011-2326-9SongDGeQChenMBaiSLaiXHuangG.
Development and validation of a nomogram for prediction of the risk of MAFLD in an overweight and obese population. J Clin Transl Hepatol. (2022) 10:1027–33. doi: 10.14218/jcth.2021.00317ZhuRXuCJiangSXiaJWuBZhangS.
Risk factor analysis and predictive model construction of lean MAFLD: a cross-sectional study of a health check-up population in China. Eur J Med Res. (2025) 30:137. doi: 10.1186/s40001-025-02373-1XueMYangXZouYLiuTSuYLiC.
A non-invasive prediction model for non-alcoholic fatty liver disease in adults with type 2 diabetes based on the population of Northern Urumqi, China. Diabetes Metab Syndr Obes. (2021) 14:443–54. doi: 10.2147/dmso.S271882LiCGuoPZhangRZhangMLiYHuangM.
Both WHR and FLI as better algorithms for both lean and overweight/obese NAFLD in a chinese population. J Clin Gastroenterol. (2019) 53:e253–60. doi: 10.1097/mcg.0000000000001089
Edited by: Charles W. Putnam, University of Arizona, United States
Reviewed by: Hua Xiang Zhuang, Shandong University, China
Mostafa Vaghari-Tabari, Tabriz University of Medical Sciences, Iran
Kengo Moriyama, Tokai University Hachioji Hospital, Japan