Visceral adipose tissue area is an independent risk factor for hepatic steatosis

Authors


Yoon Jun Kim, Department of Internal Medicine, Seoul National University Hospital, 28 Yongon-dong, Chongno-gu, Seoul 110-744, Korea.
Email: yoonjun@snu.ac.kr

Abstract

Background and Aim:  Recent data indicate that hepatic steatosis is associated with insulin resistance, dyslipidemia and obesity (especially central body fat distribution). There have been few studies on the correlation between biopsy-proven hepatic steatosis and the above factors in a disease-free population. The aim of the present study was to evaluate the relation between hepatic steatosis assessed by biopsy and clinical characteristics including regional fat distribution measured by computed tomography (CT) in living liver donors.

Methods:  Laboratory data, liver/spleen Hounsfield ratio (L/S ratio), regional fat distribution by CT and liver status by biopsy were evaluated retrospectively in a total of 177 living liver donors without a history of alcohol intake.

Results:  The unpaired t-test showed that age, triglycerides (TG), high density lipoprotein, total cholesterol, alanine aminotransferase, body mass index, L/S ratio, visceral adipose tissue area (VAT) and subcutaneous adipose tissue area (SAT) were associated with hepatic steatosis. In the multiple logistic regression analysis, VAT (odds ratio 1.031, 95% CI 1.013–1.048, P < 0.01) and TG (odds ratio 1.012, 95% CI 1.004–1.020, P < 0.01) were independent risk factors of hepatic steatosis. Subgroup analysis also showed that VAT was an independent risk factor in men (odds ratio 1.022, 95% CI 1.003–1.041, P < 0.05) and women (odds ratio 1.086, 95% CI 1.010–1.168, P < 0.05).

Conclusion:  Our results suggest that visceral abdominal adiposity is correlated with hepatic steatosis in healthy living liver donors.

Introduction

Non-alcoholic fatty liver disease (NAFLD) was once considered a benign condition; however, recent data indicate that the manifestations of NAFLD are similar to those observed in patients with alcoholic liver disease, and that these range from simple steatosis to advanced fibrosis, cirrhosis, and even hepatocellular carcinoma.1,2 Non-alcoholic fatty liver disease is recognized the most common hepatic lesion with an estimated prevalence of 15–39% in the Occident and 9–13% in the Orient.3 Many potential risk factors of NAFLD have been identified and many studies have reported that NAFLD is closely related with obesity, insulin resistance, hyperlipidemia and diabetes.4 It has also been proposed that visceral adiposity is a major contributor to fatty liver in the insulin-resistant state.5,6 In fact, evidence suggests that visceral adiposity is more influential than body mass in terms of predicting the presence of fatty liver.7

To assess abdominal visceral adipose tissue, waist–hip ratio (WHR) is the most widely used index of regional adipose tissue distribution. Computed tomography (CT) is considered the gold standard not only for adipose tissue evaluation but also for multicompartment body measurement.8 Previous studies have reported that the accumulation of risk factors for metabolic syndrome is concomitant with an increase in the visceral fat area.9 Fatty infiltration of liver can be determined by ultrasonography (US), CT, magnetic resonance imaging (MRI), 1H magnetic resonance spectroscopy (1HMRS) and by liver biopsy.10,11 The sensitivities and specificities of radiological methods are unsatisfactory, however; previous studies showed that the sensitivity and the specificity of these imaging techniques for fatty infiltration is only 60% and 82%, respectively.12,13 Direct hepatic fat determination in tissues obtained by biopsy is considered the gold standard, although it is sometimes invasive because of incidental peritoneal bleeding or duct injury.14

Because of ethical considerations in obtaining tissue specimens from a disease-free population, to the best of our knowledge, no study has been undertaken to determine the nature of the correlation between regional adipose tissue distribution assessed by CT and hepatic steatosis assessed by liver biopsy in disease-free subjects. Moreover, the relatively low sensitivity and specificity of imaging techniques for NAFLD may have influenced the results of the previous studies.

The purpose of our study was to evaluate the relation between hepatic steatosis assessed by biopsy and risk factors of hepatic steatosis, including regional adipose tissue distribution as assessed by CT, in potential for living liver donors.

Methods

Subjects

A total 177 consecutive living liver donors, from whom liver biopsy samples were obtained between January 1998 and December 2005 at the Liver Transplantation Center of Seoul National University Hospital, all without history of alcohol intake, were included in the present study. Exclusion criteria were a history of cardiovascular disease, diabetes, gall bladder stone, hepatobiliary tumor, or positivity for serum HBsAg, or serum anti-HCV. Patients with another liver disease (e.g. autoimmune hepatitis, toxic hepatitis, primary biliary cirrhosis or Budd–Chiari syndrome) were also excluded. Histories of alcohol consumption were obtained, and based on the type of alcoholic beverage consumed, the frequency of drinking per week, and the amount of alcohol consumed per day, the term ‘nonalcoholic’ was applied to men who consumed less than 30 g alcohol/day and to women who consumed less than 20 g alcohol/day.

Informed consent was obtained from all subjects, and the Institutional Review Board of Human Research at Seoul National University Hospital approved the study protocol.

Histological analysis

All the histological specimens were obtained during the preoperative evaluation or at the time of operation. The biopsies were immediately put in formalin and processed according to the routines of the Pathology Department of Seoul National University Hospital, with conventional hematoxylin and eosin stains. All the slides were evaluated independently by one pathologist who was unaware of radiologic and surgical findings. Normal liver is composed of about 5% fat by weight and hepatic steatosis refers to more than 5% fat content.15 In present study, the degrees of hepatic steatosis were quantified on a percent scale, with estimation of the amount of liver parenchyma that was replaced by hepatocytes containing cytoplasmic fat droplets; hepatic steatosis was defined as an accumulation of fat in excess of 5% of liver.16 Further, specimens were designated 0 when the percentage of hepatocytes showing fatty change was <5%, 1+ for 5–10%, 2+ for ≥10% and <20%, 3+ for ≥20% and <30%, 4+ for ≥30% and <60%, 5+ for ≥60%. Non-alcoholic steatohepatitis (NASH) was defined as steatosis with lobular inflammation, hepatocellular ballooning or steatosis with fibrosis.

Assessment of abdominal visceral fat area and calculation of liver–spleen Hounsfield ratio

To determine the visceral adipose tissue (VAT) and subcutaneous adipose tissue area (SAT), a simple CT scan was taken at the level of the umbilicus, with an attenuation range of −50 to −250 Hounsfield units.17 Subjects were examined in a supine position. VAT was defined as intra-abdominal fat bound by parietal peritoneum or transversalis fascia, excluding the vertebral column and the paraspinal muscles; and SAT was defined as fat superficial to the abdominal and back muscles. Using a cursor, VAT was then measured around the inner boundary of the abdominal wall muscles. A region of interest (ROI) drawn around the external margin of the dermis was used to calculate the area of the total adipose tissue (TAT) area. The SAT was obtained by subtracting the VAT from the TAT (Fig. 1).

Figure 1.

Calculation of abdominal adipose tissue distribution using single slice CT scans. (a) Total adipose tissue areas (TAT) were obtained by applying an adipose tissue threshold to a region of interest (ROI) traced around the dermis (green in a). (b) An ROI was traced around the inner margin of abdominal wall muscles, and an adipose tissue threshold was applied to determine the area of visceral adipose tissue area (VAT) in the ROI (green in b). Subcutaneous adipose tissue area (SAT) can then obtained by subtracting VAT from TAT.

Hepatic and splenic attenuation values were measured on unenhanced CT scans using four circular ROI in the liver and spleen and the mean numbers were used to determine the liver–spleen Hounsfield ratio (L/S ratio).

Anthropometric data and laboratory investigations

Clinical characteristics were evaluated retrospectively, and included sex, age, systolic blood pressure (SBP), diastolic blood pressure (DBP), body mass index (BMI), aspartate aminotransferase (AST), alanine aminotransferase (ALT), C-reactive protein (CRP), fasting blood sugar (FBS), total cholesterol, high density lipoprotein (HDL) cholesterol and triglyceride (TG). These parameters were measured within 30 days of CT examination. Body mass index was calculated as follows: body weight (kg)/height2 (m2).

Statistical analysis

Data were expressed as mean ± SD. The unpaired t-test was used to compare mean values of the various characteristics of non-hepatic steatosis (N = 125) and hepatic steatosis groups (N = 52). A χ2 test was used for nominal variables. Multivariate analysis was performed using forward conditional logistic regression analysis. Correlations (Pearson's correlation coefficient) and multiple linear regression analyses were also performed. All data analyses were performed using spss for Windows version 12.0 (SPSS, Chicago, IL, USA). A P-value less than 0.05 was considered significant.

Results

Subject characteristics and the presence of hepatic steatosis

The total number of study subjects was 177 (123 males and 54 females), of mean age 32.6 ± 9.6 years (range 16–57 years), mean BMI 24.05 ± 3.19 (range 18–34), mean VAT 69.19 ± 41.47 cm2 (range 6.73–208.99 cm2) and mean SAT 129.61 ± 65.35 cm2 (range 6.68–330.68 cm2). Of them, 125 (70.6%) cases were included in 0, 17 (9.6%) cases in 1+, 16 (9.0%) cases in 2+, five (2.8%) cases in 3+, 10 (5.6%) cases in 4+ and four (2.3%) cases in 5+. Non-alcoholic steatohepatitis was diagnosed in one (0.56%) case. Men tended to have a higher VAT than women (70.31 ± 43.96 cm2vs 66.65 ± 35.43 cm2, P > 0.05) and women had a higher SAT than men (173.01 ± 54.27 cm2vs 116.23 ± 62.80 cm2, P < 0.01) (Table 1).

Table 1.  Subject characteristics
FeatureTotal (n = 177)Men (n = 123)Women (n = 54)P-value
  1. Data are expressed as mean ± SD, except for hepatic steatosis, which is given as the number of subjects.

  2. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CRP, C-reactive protein; DBP, diastolic blood pressure; FBS, fasting blood sugar; HDL, high density lipoprotein cholesterol; L/S ratio, liver/spleen Hounsfield ratio; NS, not significant; SAT, subcutaneous adipose tissue; SBP, systolic blood pressure; TChol, total cholesterol; TG, triglyceride; VAT, visceral adipose tissue.

Age (years)32.62 ± 9.5630.86 ± 9.3236.63 ± 8.960.000
Hepatic steatosis (yes : no)52:12535:8817:37NS
BMI (kg/m2)24.05 ± 3.1923.80 ± 3.0424.61 ± 3.49NS
SBP (mmHg)124.53 ± 15.65127.11 ± 15.47118.65 ± 14.530.001
DBP (mmHg)75.27 ± 10.2675.91 ± 10.5573.81 ± 9.49NS
FBS (mg/L)96.61 ± 11.1996.68 ± 11.8196.44 ± 9.75NS
AST (IU/L)20.20 ± 8.2920.83 ± 7.7218.78 ± 9.36NS
ALT (IU/L)22.68 ± 14.9125.93 ± 15.9315.30 ± 8.610.000
TChol (mg/dL)173.85 ± 31.00169.97 ± 29.55182.69 ± 32.660.014
TG (mg/dL)114.81 ± 74.51123.70 ± 81.2294.06 ± 51.000.018
HDL (mg/dL)50.25 ± 13.0647.90 ± 11.8655.72 ± 14.230.002
CRP (mg/dL)0.086 ± 0.1990.096 ± 0.2280.057 ± 0.075NS
VAT (cm2)69.19 ± 41.4770.31 ± 43.9666.65 ± 35.43NS
SAT (cm2)129.61 ± 65.35116.23 ± 62.80173.01 ± 54.270.000
VAT/SAT0.552 ± 0.3760.629 ± 0.3910.303 ± 0.1460.000
L/S ratio1.101 ± 0.1641.079 ± 0.1701.150 ± 0.1370.009

The hepatic steatosis and non-hepatic steatosis groups were comparable with respect to sex, FBS, CRP, AST, SBP, DBP and VAT/SAT ratio. Hepatic steatosis group members were older (34.9 ± 9.48 years vs 31.69 ± 9.48 years, P < 0.05) and had hypercholesterolemia (184.42 ± 34.61 mg/dL vs 170.28 ± 28.77 mg/dL, P < 0.05) and hypertriglyceridemia (165.14 ± 94.08 mg/dL vs 93.24 ± 51.56 mg/dL, P < 0.01), and a lower mean HDL concentration (41.67 ± 9.22 mg/dL vs 53.93 ± 12.77 mg/dL, P < 0.01), a higher mean ALT (29.88 ± 19.66 IU/L vs 19.69 ± 1.22 IU/L, P < 0.01), a higher mean BMI (25.91 ± 3.26 vs 23.27 ± 2.83, P < 0.01), a lower L/S ratio (0.977 ± 0.21 vs 1.150 ± 0.107, P < 0.01). In terms of body composition, the hepatic steatosis group had a higher mean VAT (100.37 ± 41.25 cm2vs 56.20 ± 34.10 cm2, P < 0.01), and a higher mean SAT (157.68 ± 56.99 cm2vs 118.77 ± 65.39 cm2, P < 0.01), but VAT/SAT ratios were comparable in the two groups (Table 2). Multivariate logistic regression analysis showed that VAT (odds ratio 1.031, 95% CI 1.013–1.048, P < 0.01) and TG (odds ratio 1.012, 95% CI 1.004–1.020, P < 0.01) were independent risk factors of hepatic steatosis.

Table 2.  Clinical and laboratory features of the hepatic and non-hepatic steatosis groups
FeatureNon-hepatic steatosis (n = 125)Hepatic steatosis (n = 52)P-value
  1. Data are expressed as mean ± SD, except for Male/Female, which is given as the number of subjects.

  2. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CRP, C-reactive protein; DBP, diastolic blood pressure; FBS, fasting blood sugar; HDL, high density lipoprotein cholesterol; L/S ratio, liver/spleen Hounsfield ratio; NS, not significant; SAT, subcutaneous adipose tissue; SBP, systolic blood pressure; TChol, total cholesterol; TG, triglyceride; VAT, visceral adipose tissue.

Male/Female88/3735/17NS
Age (years)31.69 ± 9.4834.90 ± 9.480.044
BMI (kg/m2)23.27 ± 2.8325.91 ± 3.260.000
SBP (mmHg)123.46 ± 14.99127.10 ± 16.99NS
DBP (mmHg)74.42 ± 9.877.30 ± 11.06NS
FBS (mg/dL)95.40 ± 9.2899.52 ± 14.51NS
AST (IU/L)19.54 ± 8.5121.81 ± 7.56NS
ALT (IU/L)19.69 ± 1.2229.88 ± 19.660.001
TChol (mg/dL)170.28 ± 28.77184.42 ± 34.610.017
TG (mg/dL)93.24 ± 51.56165.14 ± 94.080.000
HDL (mg/dL)53.93 ± 12.7741.67 ± 9.220.000
CRP (mg/dL)0.060 ± 0.0940.152 ± 0.341NS
VAT (cm2)56.20 ± 34.10100.37 ± 41.250.000
SAT (cm2)118.77 ± 65.39157.68 ± 56.990.001
VAT/SAT0.533 ± 0.4080.603 ± 0.272NS
L/S ratio1.150 ± 0.1070.977 ± 0.2100.000

Comparison of the hepatic steatosis and non-hepatic steatosis groups according to sex

Differences between the two groups with respect to gender are shown in Table 3. In terms of body composition, the men in the hepatic steatosis group had a higher mean VAT (102.02 ± 43.20 cm2vs 57.48 ± 37.46 cm2, P < 0.01) and a higher mean SAT (151.99 ± 57.09 cm2vs 102.93 ± 59.86 cm2, P < 0.01) than men in the non-steatosis group, but the VAT/SAT ratios of men were comparable in the two groups. Women in the hepatic steatosis group had a higher mean VAT (96.86 ± 37.84 cm2vs 53.23 ± 24.76 cm2, P < 0.01) than women in the non-steatosis group, but SAT levels were comparable. Mean VAT/SAT ratio was higher in women in the steatosis group (0.400 ± 0.190 vs 0.260 ± 0.1000, P < 0.05). The multivariate logistic regression analysis showed that TG (odds ratio 1.014, 95% CI 1.005–1.023, P < 0.01), ALT (odds ratio 1.062, 95% CI 1.008–1.120, P < 0.01) and VAT (odds ratio 1.022, 95% CI 1.003–1.041, P < 0.05) were independent risk factors of hepatic steatosis in men, and that VAT (odds ratio 1.086, 95% CI 1.010–1.168, P < 0.05) was an independent risk factor for hepatic steatosis in women (Table 4).

Table 3.  Clinical and laboratory features of the hepatic and non-hepatic steatosis groups subdivided by gender
FeatureMen (n = 123)Women (n = 54)
Non-hepatic steatosisHepatic steatosisP-valueNon-hepatic steatosisHepatic steatosisP-value
  1. Data are expressed as mean ± SD.

  2. ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CRP, C-reactive protein; DBP, diastolic blood pressure; FBS, fasting blood sugar; HDL, high density lipoprotein cholesterol; L/S ratio, liver/spleen Hounsfield ratio; NS, not significant; SAT, subcutaneous adipose tissue; SBP, systolic blood pressure; TChol, total cholesterol; TG, triglyceride; VAT, visceral adipose tissue.

Age (years)29.68 ± 8.833.83 ± 10.150.02536.46 ± 9.4237.00 ± 8.13NS
BMI (kg/m2)23.10 ± 2.7925.61 ± 2.950.00023.57 ± 2.8726.62 ± 3.870.002
SBP (mmHg)125.18 ± 15.04132.62 ± 15.440.027119.38 ± 14.24117.50 ± 15.86NS
DBP (mmHg)74.36 ± 9.9780.08 ± 11.110.00974.57 ± 9.5871.75 ± 9.50NS
FBS (mg/dL)95.63 ± 9.4399.34 ± 16.19NS94.86 ± 9.0099.88 ± 10.68NS
AST (IU/L)19.82 ± 7.2923.37 ± 8.270.02118.86 ± 10.9518.59 ± 4.53NS
ALT (IU/L)22.36 ± 11.6734.89 ± 21.560.00213.32 ± 7.9119.59 ± 8.730.012
TChol (mg/dL)166.89 ± 27.89177.71 ± 32.55NS178.35 ± 29.62192.12 ± 37.69NS
TG (mg/dL)96.19 ± 52.26181.78 ± 100.080.00087.00 ± 50.46115.22 ± 49.34NS
HDL (mg/dL)51.44 ± 11.4640.44 ± 8.990.00059.19 ± 13.9845.33 ± 9.450.009
CRP (mg/dL)0.065 ± 0.1000.180 ± 0.396NS0.049 ± 0.0720.078 ± 0.082NS
VAT (cm2)57.48 ± 37.46102.02 ± 43.200.00053.23 ± 24.7696.86 ± 37.840.000
SAT (cm2)102.93 ± 59.86151.99 ± 57.090.000172.50 ± 54.65174.16 ± 56.29NS
VAT/SAT0.613 ± 0.4300.673 ± 0.263NS0.260 ± 0.1000.400 ± 0.1900.01
L/S ratio1.135 ± 0.1060.932 ± 0.2190.0001.189 ± 0.1041.066 ± 0.1630.012
Table 4.  Odds ratios (OR) of hepatic steatosis by multivariate logistic regression analysis
VariableTotal (n = 94)Men (n = 75)Women (n = 19)
OR (95% CI)POR (95% CI)POR (95% CI)P
  • Systolic/diastolic blood pressure ≥ 130/85 mmHg.

  • ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CRP, C-reactive protein; DBP, diastolic blood pressure; FBS, fasting blood sugar; HDL, high density lipoprotein cholesterol; NS, not significant; SAT, subcutaneous adipose tissue; SBP, systolic blood pressure; TChol, total cholesterol; TG, triglyceride; VAT, visceral adipose tissue.

Sex 0.439    
Age 0.938 0.446 0.472
BMI 0.153 0.323 0.587
High blood pressure 0.476 0.384 0.719
FBS 0.901 0.944 0.932
AST 0.341 0.679 0.515
ALT 0.1031.062 (1.008–1.120)0.004 0.880
TChol 0.941 0.846 0.650
TG1.012 (1.004–1.020)0.0041.014 (1.005–1.023)0.004 0.369
HDL 0.051 0.162 0.154
CRP 0.467 0.495 0.088
VAT1.031 (1.013–1.048)0.0001.022 (1.003–1.041)0.0291.086 (1.010–1.168)0.026
SAT 0.291 0.480 0.223
VAT/SAT 0.199 0.433 0.270

Correlations of degree of hepatic steatosis, abdominal fat distribution and risk factors

When the degree of hepatic steatosis was divided into six groups (0, 1+, 2+, 3+, 4+, 5+), Pearson correlations between degree of hepatic steatosis, indices of abdominal fat distribution and risk factors are shown in Table 5.

Table 5.  Correlations of the degree of hepatic steatosis, abdominal fat distribution and risk factors
VariableTotalMenWomen
  • *

    P = 0.05,

  • **

    P = 0.01,

  • ***

    P = 0.001.

  • ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; CRP, C-reactive protein; DBP, diastolic blood pressure; FBS, fasting blood sugar; HDL, high density lipoprotein cholesterol; L/S ratio, liver/spleen Hounsfield ratio; NS, not significant; SAT, subcutaneous adipose tissue; SBP, systolic blood pressure; TChol, total cholesterol; TG, triglyceride; VAT, visceral adipose tissue.

Age0.0770.1040.046
VAT0.474***0.412***0.666***
SAT0.310***0.398***0.162
VAT/SAT0.0700.0160.475**
BMI0.360***0.391***0.318*
L/S ratio−0.692***−0.728***−0.625***
SBP0.186*0.236**0.046
DBP0.185*0.257**−0.026
FBS0.1280.0740.299
AST0.162*0.246**−0.19
ALT0.406***0.426***0.438***
TChol0.203*0.1690.318
TG0.325***0.334**0.292
HDL−0.351***−0.376***−0.303
CRP0.1280.1250.239

In terms of body composition, degree of hepatic steatosis was correlated with VAT, SAT and BMI. In men, degree of hepatic steatosis was also correlated with VAT, SAT and BMI. In women, degree of hepatic steatosis was correlated with VAT, V/S ratio and BMI but not SAT. The coefficient of VAT with degree of hepatic steatosis (r = 0.474, P < 0.001) was somewhat higher than SAT (r = 0.310, P < 0.001) and BMI (r = 0.360, P < 0.001). Multiple linear regression analysis showed that VAT (β = 0.429, P < 0.001) only, when adjusted by age, SAT and BMI, was correlated with degree of hepatic steatosis, and that VAT was also consistently correlated with degree of hepatic steatosis in men (β = 0.355, P < 0.05) and women (β = 0.823, P < 0.01).

Discussion

The results of the present study show that VAT is an independent risk factor for hepatic steatosis irrespective of BMI, and VAT is well correlated with the degree of hepatic steatosis irrespective of SAT and BMI. Subgroup analysis also shows that VAT is an independent risk factor in men and in women.

Obesity is known to be accompanied by several metabolic complications and is being increasingly recognized as a risk factor for non insulin dependent diabetes mellitus (NIDDM), dyslipidemia, and atherosclerotic and cardiovascular disease. In general, the risk of developing complications is proportional to the total amount of excess fat present in the body, but there is growing evidence that the regional distribution of adipose tissue appears to be an important indicator of metabolic and cardiovascular alterations since an inconstant correlation between BMI and these disturbances has been found.18

Hepatic steatosis is usually prevalent in obese subjects and several studies have reported that regional fat distribution associated with insulin resistance is an important factor for hepatic steatosis.19,20 Kral et al. used WHR and found that abdominal fat distribution is a predictor of hepatic steatosis independent of body weight or body fat.21 Eguchi et al. reported that fatty liver severity as diagnosed by ultrasonography is positively correlated with visceral fat accumulation regardless of BMI.22 In the present study, we showed that the VAT using CT and TG are independent risk factors of histologically proven hepatic steatosis, whereas the SAT is not, and that VAT is correlated with degree of hepatic steatosis. To explain these relationships, it has been proposed that excess visceral fat, which has higher rate of lipolysis, can increase the flux of free fatty acids into the liver via the portal vein and increase circulating plasma TG levels, and that both may induce hepatic insulin resistance and hepatic steatosis; however, this mechanism remains to be proven.23,24

Seppala-Lindroos et al., using in vivo MRS, reported no correlation between abdominal adiposity and hepatic fat in 30 healthy male subjects;25 however, in the report by Seppala-Lindroos et al. the value of intra-abdominal fat was higher in high liver fat group than low liver fat group, although it was not statistically significant (3208 ± 262 cm3vs 3075 ± 340 cm3, NS). Grag and Misra commented that this may be related to the small number of subjects.26 The majority of studies have utilized US, CT or MRI to diagnose fatty liver, and of these, US is the least expensive, and thus the most widely used. The sensitivity and specificity of US for detecting hepatic steatosis range from 60% to 94% and 88% to 95%, respectively;27 however, fat-associated US changes appear at a fat change of hepatocytes of more than 15–20% and US sensitivity decrease with lower degrees of fatty infiltration.28 Computed tomography is also frequently used in hepatic steatosis studies. Saadeh et al. showed CT examination without contrast has a sensitivity of 93% and a positive predictive value of 76% when one-third of liver is fatty, and that the presence of >33% fat on liver biopsy is optimal for detecting steatosis on radiological imaging.29 Although new MRI techniques have resulted in considerable improvement in detecting steatosis, these methods measure hepatic fat indirectly, and thus, the majority of comparative studies have used biopsy results.30–33 Due to the above-mentioned disadvantages and insensitivities of imaging-based modalities, we used liver biopsy results to diagnose hepatic steatosis. It should be noted that liver biopsies have limitations, for example sampling error in the case of inhomogeneous fat distribution and variability of interpretation,34,35 however, liver biopsies uniquely provide a direct measure of hepatic steatosis, and have sensitivity and specificity advantages over imaging studies. Thus, liver biopsy is the accepted gold standard for the diagnosis of NAFLD.36 The present study has an advantage in that NAFLD was diagnosed by liver biopsy (the accepted gold standard) in a disease-free population and regional fat distribution was measured using CT; also an accepted gold standard.

Hyperlipidemia is considered a risk factor for fatty infiltration of the liver; however, whether hyperlipidemia is the cause of or a consequence of NAFLD is uncertain. It is also unclear whether certain lipid profiles are more associated with NAFLD than others. Assy et al. reported a majority of hypercholesterolemic patients did not have NAFLD, whereas severe hypertriglyceridemia increased the risk of NAFLD by five- to six-fold, and hypertriglyceridemia was most often associated with NAFLD.37 Toledo et al. suggested that the association between serum TG and hepatic steatosis is largely due to high triglyceride levels in very low density lipoprotein (VLDL) particles in type 2 diabetes.38 The observation in the present study that TG is an independent risk factor for hepatic steatosis is accord with the above findings.

Recently, Lee et al. reported that age, BMI and TG are the independent risk factors of severe hepatic steatosis in living liver donors.39 In our study, TG is also a risk factor for hepatic steatosis, but BMI is not. When the risk factors were analyzed to the exclusion of VAT, SAT and V/S ratio, BMI was also selected as a risk factor (odds ratio 1.332, 95% CI 1.089–1.628, P < 0.05) in our study. We expect that VAT, rather than BMI, would be selected as one of the independent risk factors if Lee et al. included VAT, SAT and V/S ratio in their analyses. Therefore, our study strongly suggests that visceral adipose tissue plays an important role in pathophysiology in NAFLD, rather than BMI or obesity itself.

It is well known that abdominal fat distribution is different according to gender and it has been suggested that this is important factor to explain the gender differences.40,41 Enzi et al. reported that subcutaneous fat areas are significantly greater in women than in men at the abdominal level.42 Lemieux et al. concluded that gender-associated differences in body fat distribution could account for gender-specific cardiovascular risk profiles,43 and Tanaka et al. reported that the relationships of abdominal visceral fat area to cardiovascular disease were different between men and women.44 In the present study, although no statistically significant difference was observed between the BMI of men and women, men had a significantly higher VAT/SAT ratio (0.629 ± 0.391 vs 0.303 ± 0.146, P < 0.01). Subgroup analysis showed that VAT is an independent risk factor in both sexes, despite the different body composition between sexes, which strongly suggests the pathophysiological role of VAT.

Alanine aminotransferase, which has been used as a surrogate marker for liver fat accumulation, and TG were identified as independent risk factors of hepatic steatosis only in men in the present study.45,46 Several studies have indicated that ALT is related to the features of metabolic syndrome and type 2 diabetes mellitus.47,48 Ruhl and Everhart reported that abnormal ALT activity was most strongly associated with a higher WHR.49 In the present study, although the mean value of FBS, TG, blood pressure and HDL were in the normal range for both sexes, men had higher TG levels, blood pressure and VAT/SAT ratios, but lower HDL levels than women. In view of the fact that metabolic syndrome is more prevalent in men than women up to 60 years old, and that it is closely related to hepatic steatosis, our results may be reflected by such a profile.50,51

In summary, VAT was found to be an independent risk factor for hepatic steatosis in both sexes, despite their different body compositions. We concluded that VAT is more important than BMI or obesity itself for hepatic steatosis in men and women. We think the results in our study, which was performed in living liver donors, can be applied to a healthy general population. Subsequent studies are needed to clarify the pathophysiology and relationship between simple hepatic steatosis, NASH and abdominal fat distribution.

Acknowledgment

This research was supported by The GlaxoSmithKline Research Fund of the Korean Association for the Study of the Liver.

Ancillary