Nonalcoholic fatty liver disease (NAFLD) includes a wide spectrum of progressive liver disorders ranging from simple steatosis to nonalcoholic steatohepatitis (NASH), which may lead to liver cirrhosis and even hepatocellular carcinoma [
Glycated hemoglobin (HbA1c) represents a 2-3- month average of blood glucose concentration [
This cross-sectional analysis was performed in the First Affiliated Hospital of Wenzhou Medical University (Zhejiang, China) from July 2014 to August 2017. Initial data were obtained from 13,399 subjects volunteered for a comprehensive health checkup. Finally, 5,903 subjects were excluded from the initial study population according to the following exclusion criteria: (1) positive serologic markers for hepatitis B (
Flowchart of subjects across the study.
Patients’ heights (measured to the nearest 0.1 cm) and weights (measured to the nearest 0.1 kg) were measured with light clothing by well-trained nurses in the morning. Body mass index (BMI, kg/m2) was calculated as weight divided by the height. Blood pressure was measured in the right arm at the same level as the left atrium in a seated state with a standard automatic sphygmomanometer (Omron, model 705 cp, Kyoto, Japan). After 12 h of overnight fasting, blood samples were collected from the antecubital vein by experienced nurses. Blood routine including hemoglobin, platelets, white blood cells (WBC), and biochemical markers such as albumin, HbA1c, FPG, total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglyceride (TG), gamma glutamyltransferase (GGT), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) were subsequently analyzed by an automated analyzer (Abbott AxSYM, Park, IL). All the participants routinely underwent the abdominal ultrasonography scanning (Siemens, Munich, Germany) by experienced radiologists, all of whom were blinded to the clinical status of the subjects. And, the fatty liver was diagnosed based on the guidelines for the diagnosis and treatment of nonalcoholic fatty liver disease, including (1) the near field echo of the liver is diffusely enhanced and stronger than the echo of the kidney; (2) liver brightness; and (3) vascular blurring [
We randomly selected 1,967 subjects in the enrollment, and by inserting FPG into a population regression equation expressing the linear association between HbA1c and FPG; we drove the predicted HbA1c, which is defined as follows:
The Hepatic Steatosis Index (HSI) is a preliminary screening tool for NAFLD, which uses a formula based on BMI, ALT (IU/L), and AST (IU/L), and the presence or absence of diabetes as follows:
The NAFLD Fibrosis Score (NFS) is an invasive approach to detect liver fibrosis as follows:
Data for continuous variables were expressed as mean values with standard deviation and as percentages for categorical variables. One-way analysis of variance with Bonferroni’s method, post hoc analysis, and Chi-squared test were used to compare statistical differences in the characteristics of the study participants at baseline among the groups. Univariate logistic regression analysis was acted out to explore the risk factors of NAFLD by ultrasonography. Multiple logistic regression analysis was executed to estimate the OR and 95% CI for the association of categorized HGI levels with the risk of NAFLD. The area under the curve of the receiving operating characteristic (ROC) was used to evaluate the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of the proposed risk score for the prediction of NAFLD. Statistical data analysis was performed by using the SPSS version 25.0 for Windows. And, statistical significance was set at <0.05.
Study population comprised 3,936 participants, of whom 2,349 (59.7%) were male. Of the subjects at baseline, the participants older than 50 years old and with a BMI ≥25 were 1,437 (36.5%) and 1,278 (32.5%), respectively. Overall, the participants were relatively young and not obese, with the mean age of 47.7 ± 11.0 years old and mean BMI 23.8 ± 3.2 kg/m2. The baseline of anthropometric features and biochemical findings of the study population was stratified according to quartiles of HGI value showed in Table
Baseline characteristics of study participants according to the HGI.
Variables | Total |
Quartile 1 (≤−0.22) |
Quartile 2 (−0.21∼0.02) |
Quartile 3 (0.03∼0.28) |
Quartile 4 (≥0.29) |
|
|
---|---|---|---|---|---|---|---|
Gender (male/female) | 1587/2349 | 443/555 | 388/597 | 378/597 | 378/600 | 9.33 |
|
Age (years) | 47.74 ± 11.01 | 42.81 ± 9.46 | 46.29 ± 10.59a | 48.95 ± 10.96ab | 53.01 ± 10.41abc | 170.37 |
|
SBP (mmHg) | 123.69 ± 18.56 | 120.85 ± 17.43 | 122.32 ± 17.89 | 124.91 ± 19.24ab | 126.74 ± 19.12ab | 20.09 |
|
DBP (mmHg) | 73.12 ± 12.63 | 71.91 ± 12.49 | 72.86 ± 12.63 | 73.52 ± 12.78a | 74.22 ± 12.5ab | 6.02 |
|
BMI (kg/m2) | 23.76 ± 3.16 | 23.17 ± 3.08 | 23.54 ± 3.21a | 23.91 ± 3.12ab | 24.43 ± 3.08abc | 29.34 |
|
WBC (×109/L) | 6.04 ± 1.61 | 5.79 ± 1.53 | 6.02 ± 1.55a | 6.07 ± 1.66a | 6.30 ± 1.67abc | 16.88 |
|
HB (g/L) | 145.19 ± 14.90 | 146.44 ± 15.21 | 145.97 ± 15.07 | 144.69 ± 14.64a | 143.64 ± 14.51ab | 7.15 |
|
PLT (×109/l) | 231.71 ± 53.25 | 226.23 ± 51.44 | 232.01 ± 53.57 | 231.61 ± 51.16 | 237.07 ± 56.23a | 6.86 |
|
ALB (g/L) | 44.93 ± 3.05 | 45.52 ± 3.07 | 45.16 ± 3.03a | 44.59 ± 3.07ab | 44.44 ± 2.89ab | 27.48 |
|
ALT (U/L) | 27.68 ± 33.13 | 26.09 ± 20.5 | 26.05 ± 18.57 | 26.81 ± 19.29 | 31.82 ± 57ab | 6.91 |
|
AST (U/L) | 27.28 ± 16.74 | 25.16 ± 14.01 | 25.17 ± 10.37 | 26.07 ± 9.81 | 28.74 ± 26.75abc | 10.09 |
|
GGT (U/L) | 44.94 ± 66.78 | 42.46 ± 88.12 | 42.72 ± 62.15 | 44.37 ± 51.89 | 50.27 ± 58.5ab | 2.92 |
|
FPG (mmol/L) | 4.67 ± 0.61 | 4.65 ± 0.53 | 4.63 ± 0.55 | 4.64 ± 0.62 | 4.76 ± 0.72abc | 8.92 |
|
Creatinine (umol/L) | 67.55 ± 14.15 | 66.16 ± 13.73 | 67.58 ± 14.84a | 68.23 ± 14.12a | 68.25 ± 13.83a | 4.76 |
|
TC (mmol/L) | 5.29 ± 1.05 | 5.09 ± 0.97 | 5.26 ± 1.01a | 5.35 ± 1.04a | 5.46 ± 1.13ab | 23.39 |
|
TG (mmol/L) | 1.74 ± 1.23 | 1.62 ± 1.21 | 1.70 ± 1.21 | 1.74 ± 1.23a | 1.93 ± 1.26abc | 11.05 |
|
HDL-C (mmol/L) | 1.28 ± 0.33 | 1.32 ± 0.34 | 1.29 ± 0.33 | 1.29 ± 0.33 | 1.24 ± 0.31abc | 9.93 |
|
LDL-C (mmol/L) | 3.16 ± 0.85 | 3.00 ± 0.76 | 3.13 ± 0.81a | 3.22 ± 0.86a | 3.31 ± 0.92ab | 24.29 |
|
HSI | 32.49 ± 4.84 | 31.89 ± 4.76 | 32.22 ± 5.04a | 32.50 ± 4.75ab | 33.36 ± 4.70ab | 27.62 |
|
NFS | −2.52 ± 1.02 | −2.71 ± 0.97 | −2.60 ± 1.04 | −2.42 ± 1.02a | −2.34 ± 1.02abc | 17.00 |
|
NAFLD (%) | |||||||
Hepatic US | 815 (20.7) | 164 (16.4) | 165 (16.8) | 210 (21.5) | 276 (28.2) | 54.53 |
|
HSI | 875 (22.2) | 174 (17.4) | 221 (22.4) | 207 (21.2) | 273 (27.9) | 32.14 |
|
HbA1c (%) | 5.38 ± 0.40 | 4.89 ± 0.22 | 5.25 ± 0.11a | 5.50 ± 0.12ab | 5.88 ± 0.22abc | 5599.99 |
A relationship between HGI and HSI (a) as well as HGI and NFS (b).
The proportion of subjects with NAFLD by HGI category within the nondiabetic range by US (a, b) and HSI (c, d).
To investigate the potential interactions affecting the prevalence of NAFLD, univariate logistic regression analysis was performed as shown in Table
Univariate logistic regression analysis for the risk of NAFLD by ultrasonography.
Parameters | Odds ratio | Lower 95% CI | Upper 95% CI |
|
---|---|---|---|---|
Gender (male/female) | 3.155 | 2.627 | 3.789 | <0.001 |
Age (years) | 1.004 | 0.997 | 1.011 | 0.230 |
BMI (kg/m2) | 1.406 | 1.364 | 1.451 | <0.001 |
Hypertension (%) | 1.801 | 1.514 | 2.413 | <0.001 |
Alcohol intake | 1.627 | 1.392 | 1.901 | <0.001 |
Smoking | 1.527 | 1.292 | 1.804 | <0.001 |
FPG (mmol/L) | 1.716 | 1.509 | 1.952 | <0.001 |
TG (mmol/L) | 1.837 | 1.709 | 1.975 | <0.001 |
TC (mmol/L) | 1.307 | 1.216 | 1.404 | <0.001 |
HDL-C (mmol/L) | 0.131 | 0.097 | 0.176 | <0.001 |
LDL-C (mmol/L) | 1.357 | 1.241 | 1.484 | <0.001 |
ALT (U/L) | 1.037 | 1.032 | 1.041 | <0.001 |
AST (U/L) | 1.038 | 1.031 | 1.045 | <0.001 |
GGT (U/L) | 1.009 | 1.007 | 1.010 | <0.001 |
WBC (×109/L) | 1.300 | 1.241 | 1.302 | <0.001 |
HB (g/L) | 1.047 | 1.041 | 1.053 | <0.001 |
PLT (×109/l) | 1.002 | 1.001 | 1.004 | <0.001 |
ALB (g/L) | 1.116 | 1.087 | 1.145 | <0.001 |
HGI | 2.196 | 1.787 | 2.698 | <0.001 |
BMI, body mass index; FPG, fasting blood glucose; TC, total cholesterol; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; ALT, alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma glutamyltransferase; WBC, white blood cell; HB, hemoglobin; PLT, blood platelet; ALB, albumin; HGI, hemoglobin glycation index.
Risk of NAFLD by ultrasonography according to HGI quartile and other variables.
Variables | Odds ratio (95% CI) | |||
---|---|---|---|---|
Model 1 | Model 2 | Model 3 | Model 4 | |
HGI | 1.281 (1.194∼1.375) | 1.282 (1.189–1.384) | 1.284 (1.184∼1.392) | 1.172 (1.074∼1.279) |
Gender | 3.114 (2.591∼3.743) | 2.316 (1.845∼2.906) | 1.315 (1.018∼1.699) | |
Age (years) | 0.997 (0.990∼1.005) | 0.993 (0.984∼1.002) | 1.005 (0.995∼1.015) | |
TG (mmol/L) | 1.562 (1.449∼1.684) | 1.226 (1.131∼1.331) | ||
Hypertension | 1.396 (1.147∼1.700) | 1.042 (0.842∼1.288) | ||
Alcohol | 0.840 (0.691∼1.021) | 0.831 (0.673∼1.026) | ||
Smoking | 0.817 (0.665∼1.004) | 0.759 (0.606∼0.951) | ||
Alb (g/L) | 1.061 (1.029∼1.095) | 1.110 (1.071∼1.149) | ||
FPG (mmol/L) | 1.437 (1.247∼1.657) | 1.326 (1.138∼1.544) | ||
GGT (U/L) | 1.003 (1.001∼1.004) | 1.002 (1.001∼1.004) | ||
BMI (kg/m2) | 1.284 (1.240∼1.329) | |||
AST (U/L) | 0.987 (0.973∼1.000) | |||
HDL-C (mmol/L) | 0.385 (0.262∼0.568) | |||
ALT (U/L) | 1.021 (1.013∼1.029) | |||
WBC (×109/L) | 1.162 (1.098∼1.231) |
Model 1: no adjustment; model 2: adjusted for age and sex; model 3: model 2 + TG, hypertension, alcohol, smoking, albumin, FPG, and GGT; model 4: model 3 + BMI, AST, ALT, HDL-C, and WBC. TG, triglyceride; FPG, fasting blood glucose; GGT, gamma glutamyltransferase; BMI, body mass index; AST, aspartate aminotransferase; HDL-C, high-density lipoprotein cholesterol; ALT, alanine aminotransferase; WBC, white blood cell; Alb, albumin; NAFLD, nonalcoholic fatty liver disease; HGI, hemoglobin glycation index.
According to the results, the variables such as HGI, TG, FPG, BMI, ALT, HDL-C, and WBC were the key risk factors (
Receiver operating characteristics curves of NAFLD risk score and HSI for diagnosing the risk of NAFLD. AUC, area under the curve; HSI, hepatic steatosis index.
Previous studies have reported that HGI represents the degree of nonenzymatic hemoglobin glycation, and it has been found to promote the development of microvascular and macrovascular complications in diabetic subjects. Using HGI analysis to the Diabetes Control and Complications Trial (DCCT), it has been proved that the higher the HGI, the greater the risk of retinopathy and nephropathy in the patients with type 1 diabetes [
The first study demonstrated the relationship between NAFLD and HGI in 1,120 white individuals without diabetes was in 2017. When adjusted for age, gender, and BMI, individuals in the highest quartile of HGI exhibited a 1.6-fold increased odd of having hepatic steatosis compared with subjects in the lowest. Another one published recently of 14,465 nondiabetic, Korea subjects showed that a high HGI was associated with a 1.56-fold increase in the risk of hepatic steatosis after adjusting more factors than the previous one. However, they did not further construct a risk score based on HGI to predict a person’s risk of NAFLD, which can be useful for the clinician.
Our study demonstrated that HGI was one of the independent risk factors for incident NAFLD among nondiabetic Chinese medical examination population. The result that 20.7% of the nondiabetic population who was affected by NAFLD was equivalent to the ultrasound results of one in every four adults with NAFLD in China [
Nowadays, not only the association between HGI and NAFLD has not been fully evaluated but also NAFLD pathogenesis is still unclear, so the pathophysiological mechanisms involved in the associations between each other are still undefined. Through this retrospective cross-sectional study, there are three suggested mechanisms to explain the association between HGI and NAFLD. First, we found a positive association between HGI and inflammatory biomarkers including WBC and platelet count. Obviously, chronic inflammation plays a key role in the progress of NAFLD, and insulin resistance is thought to be a core component of NAFLD. Impaired mitochondria can result in incomplete fat oxidation and generation of toxic lipid intermediates, which generate a large amount of reactive oxygen species (ROS) and reactive nitrogen free radicals (RNS) and lead to inflammation. Then, inflammation may impair insulin signaling and exacerbate liver fatty infiltration, even fuel the transition from NAFLD to NASH, and liver cirrhosis, even hepatocellular carcinoma [
Our study had some limitations that should be taken into account. First, our study subjects were derived from a single center; for each population, new regression models should be derived, and multicentered research should be performed to further confirm the association in the next step. Second, our biochemical parameters, such as plasma glucose and HbA1c, were measured once. Even though this approach is commonly used in clinical research, between-individual variability of glucose homeostasis parameters may have led to some imprecisions in the stratification of study population into HGI quartiles. Third, the diagnosis of liver steatosis was performed by ultrasound scanning rather than by invasive methods such as liver biopsy or expensive noninvasive approaches such as proton magnetic resonance spectroscopy or computed tomographic scanning. However, ultrasonography is the most commonly used method to diagnose hepatic steatosis in clinical practice and epidemiological studies. Although liver biopsy is a standard criterion for NAFLD diagnosis, the diagnosis of fatty liver was based on ultrasound imaging with 90% sensitivity and 80% specificity [
In conclusion, we further confirmed that NAFLD had an association with HGI level in nondiabetic individuals, and these associations were independent of obesity and other metabolic components. And, NAFLD risk score could be used as one of the risk predictors of NAFLD in nondiabetic population. But the causal relationship between HGI level and NAFLD in nondiabetic individuals is not undefined, while further perspective study should be taken.
The data used to support the findings of this study are available from the corresponding author upon request.
This study was approved by the institutional review board at The First Affiliated Hospital of Wenzhou Medical University. All procedures were in accordance with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.
The authors declare that there are no conflicts of interest regarding the publication of this paper.
Hu DS, Zhu SH, Chen QF, Lin CJ, Fang DH, and Wu JS designed the study; Hu DS, Zhu SH, and Chen QF collected the data; Hu DS and Zhu SH performed the statistical analyses; Hu DS, Zhu SH, Chen QF, and Wu JS reviewed the results, interpreted the data, and wrote the manuscript; all authors have made an intellectual contribution to the manuscript and approved the submission.
The authors thank the Medical and Health Care Center of the First Affiliated Hospital of Wenzhou Medical University for offering data. This work was financially supported by the Wenzhou Science and Technology Bureau (grant no. Y20180062).