Hepatobiliary & Pancreatic Diseases International
study was to investigate the prevalence and risk factors alcohol consumption more than 40 g (male) or 20 g
of FLD in Chengdu, Southwest China, and to provide a (female) of alcohol per day for over 5 years; NAFLD,
relevant basis for the management of FLD in China. non-drinkers or alcohol consumption less than 20
g (male) or 10 g (female) of alcohol per day for more
than 1 year; and suspected AFLD, intermediate
Methods alcohol consumption and duration which fell between
Subjects the other two subtypes.
Gallstones were diagnosed on the basis of their
In all, 11 045 subjects of over 18 years old who were
distinct ultrasonographic features, including echo
working in Chengdu and had a medical checkup at
density, acoustic shadowing, and gravitational
the Physical Examination Center of the West China
dependence. Body mass index (BMI) was calculated
Hospital of Sichuan University between January and
as a subject's weight in kg divided by the square
December 2007 were investigated. Subjects without
of their height in meters. Obesity was defined as a
complete laboratory data were excluded from this
BMI ≥25 kg/m2 in both male and female, according
investigation. At last, a total of 9094 subjects were
to the redefined WHO criteria in the Asia Pacific
included for the final analysis. This study was
Region. Hypertension was diagnosed as a systolic
approved by the Ethics Committee of the West China
blood pressure ≥140 mmHg or a diastolic blood
Hospital of Sichuan University (Chengdu, China).
pressure ≥90 mmHg, according to the WHO criteria.
Hyperlipidemia was defined as a total cholesterol level
Methods of examination
≥5.2 mmol/L or a triglyceride level ≥1.7 mmol/L.
For each subject, a comprehensive medical history Fasting hyperglycemia was defined as fasting plasma
was obtained by experienced medical staff members glucose ≥6.1 mmol/L. ALT abnormalities were
from the Physical Examination Center. The history defined as ALT ≥55 IU/L for males and ≥38 IU/L for
included alcohol consumption, smoking, and a females. Diagnoses of diabetes mellitus were based on
detailed history of viral hepatitis, gallstone disease, the WHO 1999 criteria. Participants who reported
previous diagnosis of diabetes, and hypertension. current use of anti-hypertension or anti-diabetes
Body weight, height, and blood pressure were medications were regarded as having hypertension or
measured during the examination. Liver, gallbladder, diabetes, respectively.
and spleen were examined by ultrasonography using
a Philips HD11XE (Philips Medical Systems, Bothell, Statistical analysis
USA) with a 2-5 MHz probe. After overnight fasting,
Statistical analyses were made using SPSS version
fasting plasma glucose (FPG), triglycerides (TG), total
13.0 software. The descriptive results of continuous
cholesterol (TCh), alanine aminotransferase (ALT),
variables were expressed as the mean±standard
high-density lipoprotein cholesterol (HDL-C), and
deviation (SD). Differences in numerical data
low-density lipoprotein cholesterol (LDL-C) were
were assessed using Student's t test and Wilcoxon's
measured using a Hitachi Modular analysis system
rank-sum test. Differences in categorical data were
(Roche Modular DPP, Hitachi Ltd., Tokyo, Japan).
assessed using the Chi-square test. In the analysis
The presence of hepatitis B surface antigen (HBsAg)
was tested using a diagnostic kit for HBsAg (Intec of continuous variables, data were categorized
Products Inc., Xiamen, China). according to cut-off values and analyzed using
the Chi-square test or Fisher's exact test. Logistic
regression analysis was used to identify risk factors for
FLD. Odds ratios (OR) and 95% confidence intervals
Diagnosis of FLD was based on the presence of (CI) were estimated when appropriate. All statistical
an ultrasonographic pattern which met the criteria comparisons were two-tailed. P values less 0.05 were
for FLD as established by the Chinese Society for considered statistically significant.
Liver Disease.[11, 12] The determination of the etiology
of FLD was based on the guidelines for diagnosis
and treatment of non-alcoholic and alcoholic FLD
issued by the FLD and Alcoholic Liver Disease Study
Group of the Chinese Society for Liver Disease.[11, 12] Prevalence of FLD and gender differences
According to these guidelines, FLD was divided into A total of 4721 males and 4373 females were included
three subtypes: alcoholic fatty liver disease (AFLD), in this study. The overall mean age was 43.93±13.47
378 • Hepatobiliary Pancreat Dis Int，Vol 8，No 4 • August 15，2009 • www.hbpdint.com
Prevalence and risk factors of fatty liver disease
years, with no significant difference between males For analysis of FLD risk factors, the 9094 subjects
and females. Of the 9094 subjects, 1140 (12.5%) were divided into a FLD group (n=1140) and a non-
were diagnosed as having FLD, with a 3-fold higher FLD group (n=7954). Univariate analysis showed that
prevalence in males than in females (18.9% vs. 5.7%, age, BMI, FPG, blood pressure, TG, TCh, LDL-C, and
χ2=359.624, P<0.001). ALT were all significantly higher in the FLD group
The prevalence of FLD was 9.4% in males aged than in the non-FLD group, whereas HDL-C was
less than 30 years and increased gradually to 24.4% lower in the FLD group (Table 1). Differences in the
in males aged 40-49 years, but then decreased in prevalence of the features of metabolic syndrome and
progression after 50 years of age. In females, the of several other characteristics between the FLD and
prevalence of FLD increased gradually from 0.4% in non-FLD groups were significant (P<0.001) (Table 2).
those aged less than 30 years to 18.6% in those aged In order to identify FLD risk factors, stepwise
more than 70 years (Fig.). The prevalence of FLD in logistic regression analysis was performed using a
males of less than 60 years old was higher than that probability for entry of 0.05 and for removal of 0.1. The
in females of similar age (19.8% vs. 3.9%, χ2=459.233, results revealed that FLD was significantly associated
P<0.001), but the prevalence was similar in males and with male sex, age, BMI, FPG, hypertension, TG, TCh,
females aged 60-69 years (13.8% vs. 15.0%, χ2=0.212, HDL-C, LDL-C, and ALT abnormalities (Table 3).
P=0.645). However, in subjects of more than 70 years
old, the prevalence of FLD in females was higher than
Etiological constituent ratios of FLD and their
in males (18.6% vs. 11.9%, χ2=4.155, P<0.05).
Analysis of risk factors for FLD Among the 1140 subjects with FLD, 575 (50.4%)
Table 2. Prevalence of features of metabolic syndrome and other
characteristics of subjects in the FLD and non-FLD groups
FLD group non-FLD P value
(%) group (%) (<)
Obesity 784 (68.8) 1539 (19.3) 1280.574 0.001
Fasting hyperglycemia 182 (16.0) 241 (3.0) 376.156 0.001
Diabetes mellitus 84 (7.4) 190 (2.4) 84.612 0.001
Hypertension 431 (37.8) 1323 (16.6) 220.417 0.001
Hyperlipidemia 852 (74.7) 2839 (35.7) 630.344 0.001
Gallstone 148 (13.0) 640 (8.0) 30.698 0.001
HBsAg (-) 56 (4.9) 651 (8.2) 14.894 0.001
ALT abnormalities 366 (32.1) 600 (7.5) 633.590 0.001
Fig. Prevalence of FLD in 9094 Chinese adults by age.
Alcohol-drinking 565 (49.6) 2540 (31.9) 137.792 0.001
Smoking 455 (39.9) 1959 (24.6) 119.442 0.001
Table 1. Characteristics of subjects in the FLD and non-FLD
Table 3. Multiple logistic regression of factors associated with FLD
FLD group testing variables
group P value
Characteristics (n=1140, t/Z B SE Wald P OR 95% CI
mean±SD) Male 0.262 0.98 7.097 0.008 1.299 (1.072-1.576)
Age (years) 46.81±16.62 29.25±14.40 -7.742 0.001 Age 0.011 0.004 10.392 0.001 1.011 (1.004-1.018)
BMI (kg/m ) 26.63±2.90 22.48±2.99 -43.570 0.001 BMI 0.324 0.015 459.222 <0.001 1.386 (1.343-1.425)
FPG (mg/dl) 5.41±1.68 4.74±0.95 -19.795 0.001 FPG 0.206 0.028 56.012 <0.001 1.229 (1.164-1.297)
SBP (mmHg) 126.83±34.51 116.30±16.87 -16.134 0.001 Hypertension 0.343 0.092 13.940 <0.001 1.409 (1.177-1.687)
DBP (mmHg) 8359±10.46 76.53±10.08 -21.325 0.001 TG 0.469 0.054 76.476 <0.001 1.598 (1.439-1.775)
TG (mmol/L) 2.74±1.73 1.44±1.03 -35.589 0.001 TCh -0.650 0.155 17.699 <0.001 0.522 (0.386-0.707)
TCh (mmol/L) 4.85±0.87 4.51±0.87 -12.216 0.001 HDL-C -1.074 0.197 29.550 <0.001 0.342 (0.232-0.503)
ALT (IU/L) 45.68±25.89 25.54±17.54 -30.926 0.001 LDL-C 0.925 0.162 32.490 <0.001 2.521 (1.835-3.465)
HDL-C (mmol/L) 1.21±0.29 1.56±0.40 28.589 0.001 ALT 1.041 0.098 112.161 <0.001 2.833 (2.336-3.435)
LDL-C (mmol/L) 3.03±0.75 2.75±0.75 -11.802 0.001 abnormalities
SBP: systolic blood pressure; DBP: diastolic blood pressure. Constant -11.074 0.520 452.980 <0.001 0.000
Hepatobiliary Pancreat Dis Int，Vol 8，No 4 • August 15，2009 • www.hbpdint.com • 379
Hepatobiliary & Pancreatic Diseases International
were defined as NAFLD, 235 (20.6%) as AFLD, and 330 different from these in other areas of China, all of
(29.0%) as suspected AFLD. The overall prevalence of which may be related to the development of FLD.
AFLD, suspected AFLD, and NAFLD in the entire study In this study, the prevalence of FLD was higher in
population was 2.6%, 3.6%, and 6.3%, respectively. males than in females of less than 60 years old, was
Based on BMI and alcohol consumption, 1630 of similar in males and females aged 60-69 years, and was
the 9094 enrolled subject, were in the control group higher in females after the age of 70 years. A similar
(BMI <23 kg/m2, non-drinkers or alcohol consumption phenomenon has been noted in several previous
of less than 20 g (male) or 10 g (female) alcohol per day studies.[8, 10, 18] These age-related gender differences may
for more than 1 year), 197 in the excessive drinking be related to reduced androsterone in males and low
group (BMI <23 kg/m2 and alcohol consumption of estrogen levels and relatively increased androsterone
more than 40 g (male) or 20 g (female) alcohol per day after menopause in females of more than 60 years
for over 5 years), 680 in the obese group (BMI ≥25 old.[17, 19] This possibility implies that female hormones
kg/m2, non-drinkers or alcohol consumption less than might have favorable effects on lipid metabolism in the
20 g (male) or 10 g (female) alcohol per day for more liver, while androsterone may have the opposite effect.
than 1 year), and 297 in the obese excessive drinking The prevalence of AFLD was 2.6% in our study
group (BMI ≥25 kg/m2 and alcohol consumption population, which was much higher than the 0.79%
more than 40 g (male) or 20 g (female) alcohol per and 0.94% reported in Shanghai and Zhejiang province,
day for over 5 years). The prevalence rates of FLD respectively.[8, 20] However, this finding was consistent
in the control, excessive drinking, obese, and obese with the high alcohol consumption rate (34.14%) in
excessive drinking groups were 1.35%, 3.05%, 27.21%, the subjects evaluated in the present study. Similarly,
and 42.76%, respectively. Compared with the control compared with Shanghai and Zhejiang province,
group, the odds ratios (95% CI) for FLD in the other AFLD in Chengdu comprised a larger constituent
groups were 2.30 (0.92-5.73), 27.32 (17.36-42.99), and ratio of FLD (20.6%). However, the etiological
54.60 (33.81-88.20), respectively. Among the excessive constituent ratio of FLD needs to be further studied
drinkers, obesity increased the risk for FLD by in North China, where more heavy drinkers have been
23.78-fold (10.22-55.33). However, excessive drinking reported. Compared with the controls, the risk for
was associated with only a 2-fold (1.50-2.66) increased FLD was 2.30-fold higher in heavy drinkers, 27.32-fold
risk in subjects with obesity, whereas the risk of FLD in higher in subjects with obesity, and 54.60-fold higher
subjects with obesity without excessive drinking was in obese heavy drinkers. In heavy drinkers, obesity
11.90-fold (5.19-27.30) higher than that of excessive increased the risk for FLD by 23.78-fold, whereas
drinkers without obesity. heavy drinking was associated with only a 2.00-fold
increased risk in obese subjects, indicating that FLD is
more strongly associated with obesity than with heavy
Discussion drinking, and that the prevalence of FLD dramatically
FLD is a common chronic liver disease with genetic, increased when both conditions were present.
environmental, metabolic, and stress-related com- Our data demonstrated that FLD was mainly
ponents. The natural history of FLD ranges from associated with obesity, hyperglycemia, dyslipidemia,
asymptomatic indolent to the end stages of liver and hypertension, which comprise the main features
disease. The mean prevalence of FLD in western of metabolic syndrome. People with metabolic
countries, as measured by ultrasonography, ranges syndrome are at increased risk for developing diabetes
from 20% to 60%, with 21.8% in Japan and 24.3% mellitus and cardiovascular disease.[23, 24] FLD is
in Korea.[16, 17] Our study showed that approximately believed to be an additional feature of metabolic
12.5% of Chengdu adults had FLD, which was much syndrome and is regarded as a common "burden of
lower than the 20.8% prevalence reported in Eastern disease" in the Chinese population.
China, 20.7% in Southern China, or 24.5% in Central China is an endemic area for HBV infection, with
China.[8-10] This discrepancy with previous studies a nationwide survey conducted in 2006 showing that
may be the result of differences in the methods of the prevalence of HBsAg carriers was approximately
subject selection and possible regional differences. 7.18% in the nationwide population between the
Our study included apparently healthy Chinese people ages of 1 and 59 years. In our study, the prevalence
who underwent a routine health checkup at a hospital of HBsAg carriers was 7.8%, which may indicate the
in Chengdu. Also, the relative economic conditions, prevalence of HBsAg in Chengdu is somewhat higher
age stratification, and dietary habits in Chengdu are than the mean prevalence in China. Consistent with
380 • Hepatobiliary Pancreat Dis Int，Vol 8，No 4 • August 15，2009 • www.hbpdint.com
Prevalence and risk factors of fatty liver disease
several reports showing that the development of FLD is and interpretation of the study and to further drafts. TH is the
not associated with HBV infection,[26-28] monovariant guarantor.
Competing interest: No benefits in any form have been received
regression analysis in our study showed a negative
or will be received from a commercial party related directly or
correlation between positive HBsAg and FLD. This indirectly to the subject of this article.
result suggests that chronic hepatitis B infection does
not contribute to the development of FLD.
Concern about FLD is growing, not only because References
it is a common liver disorder, but also because it is one
1 ellentani S, Bedogni G, Miglioli L, Tiribelli C. The
of the leading causes of abnormal liver function.[29, 30] epidemiology of fatty liver. Eur J Gastroenterol Hepatol
Unexplained ALT abnormalities are strongly associated 2004;16:1087-1093.
with adiposity and thus may represent NFLD. Our 2 ngelico F, Del Ben M, Conti R, Francioso S, Feole K,
study showed that the prevalence of ALT abnormalities Maccioni D, et al. Non-alcoholic fatty liver syndrome: a
hepatic consequence of common metabolic diseases. J
was 13.0%, of whom 37.9% had FLD. The percentage
Gastroenterol Hepatol 2003;18:588-594.
of ALT abnormalities was much higher in the FLD 3 ark SH, Kim BI, Yun JW, Kim JW, Park DI, Cho YK, et al.
group than in the non-FLD group (32.1% vs. 7.5%, Insulin resistance and C-reactive protein as independent
χ2=633.590, P<0.001), which suggests that the increase risk factors for non-alcoholic fatty liver disease in non-
of ALT may be a consequence of FLD and can thus be obese Asian men. J Gastroenterol Hepatol 2004;19:694-698.
regarded as a predicative indicator for FLD. 4 ngulo P. Nonalcoholic fatty liver disease. N Engl J Med
Since the present study was conducted in apparently 5 eddy JK, Rao MS. Lipid metabolism and liver
healthy adults in Chengdu who underwent health inflammation. II. Fatty liver disease and fatty acid oxidation.
checkups at a hospital, results of the study cannot be Am J Physiol Gastrointest Liver Physiol 2006;290:G852-858.
generalized to other groups and contexts in Chengdu. 6 Clark JM. The epidemiology of nonalcoholic fatty liver
In future studies, it would be of interest to investigate disease in adults. J Clin Gastroenterol 2006;40:S5-10.
7 Björnsson E. The clinical aspects of non-alcoholic fatty
the possible relationship of FLD to additional factors, liver disease. Minerva Gastroenterol Dietol 2008;54:7-18.
such as the use of various drugs, physical exercise, 8 an JG, Zhu J, Li XJ, Chen L, Li L, Dai F, et al. Prevalence
educational level, and HCV. Our diagnoses of FLD were of and risk factors for fatty liver in a general population of
based on ultrasonography instead of liver biopsy, since Shanghai, China. J Hepatol 2005;43:508-514.
liver biopsy is an invasive examination with increased 9 hao GX, Zhang XG, Huang ZP, Zhu QY, Tang H.
Discussion on body mass index and fatty liver distribution
risk and high cost. Ultrasonography is widely used for
from 28,384 patients in Shenzhen area. Zhonghua Gan
the detection of FLD with high sensitivity (up to 89%) Zang Bing Za Zhi 2003;11:372-373.
and specificity (up to 93%). Thus, we considered this 10 ang Z, Xia B, Ma C, Hu Z, Chen X, Cao P. Prevalence and
technique to be suitable for use in the diagnosis of FLD risk factors of fatty liver disease in the Shuiguohu district of
in the present study. Wuhan city, central China. Postgrad Med J 2007;83:192-195.
11 atty Liver and Alcoholic Liver Disease Study Group of
In conclusion, this study demonstrated that FLD is
the Chinese Liver Disease Association. Guidelines for
closely associated with male sex, old age, obesity, and diagnosis and treatment of nonalcoholic fatty liver diseases.
other features of metabolic syndrome. Although the Zhonghua Gan Zang Bing Za Zhi 2006;14:161-163.
prevalence of FLD in Chengdu adults was 12.5%, this 12 atty Liver and Alcoholic Liver Disease Study Group
disease must be regarded as potentially serious because of the Chinese Liver Disease Association. Guidelines
of its possible evolution to end-stage liver disease. In for diagnosis and treatment of alcoholic liver diseases.
Zhonghua Gan Zang Bing Za Zhi 2006;14:164-166.
particular, subjects with obesity and unexplained 13 an JG, Zhu J, Li XJ, Chen L, Lu YS, Li L, et al. Fatty liver
abnormal liver function tests should be screened with and the metabolic syndrome among Shanghai adults. J
abdominal ultrasonography. Specific strategies for the Gastroenterol Hepatol 2005;20:1825-1832.
prevention of FLD in Southwestern China also need to 14 nuurad E, Shiwaku K, Nogi A, Kitajima K, Enkhmaa B,
be pursued. Shimono K, et al. The new BMI criteria for asians by the
regional office for the western pacific region of WHO are
suitable for screening of overweight to prevent metabolic
Funding: The study was supported by a grant from the National syndrome in elder Japanese workers. J Occup Health 2003;
Natural Science Foundation of China (No. 30571640). 45:335-343.
Ethical approval: This study was approved by the Ethics 15 haw JE, de Courten M, Boyko EJ, Zimmet PZ. Impact
Committee of the West China Hospital of Sichuan University of new diagnostic criteria for diabetes on different
(Chengdu, China). populations. Diabetes Care 1999;22:762-766.
Contributors: TH proposed the study. LH wrote the first draft. 16 magari K, Kadokawa Y, Masuda J, Egawa I, Sawa
TK analyzed the data. All authors contributed to the design T, Hazama H, et al. Fatty liver in non-alcoholic non-
Hepatobiliary Pancreat Dis Int，Vol 8，No 4 • August 15，2009 • www.hbpdint.com • 381
Hepatobiliary & Pancreatic Diseases International
overweight Japanese adults: incidence and clinical 26 homopoulos KC, Arvaniti V, Tsamantas AC,
characteristics. J Gastroenterol Hepatol 2002;17:1098-1105. Dimitropoulou D, Gogos CA, Siagris D, et al. Prevalence
17 ark SH, Jeon WK, Kim SH, Kim HJ, Park DI, Cho YK, et
P of liver steatosis in patients with chronic hepatitis B: a
al. Prevalence and risk factors of non-alcoholic fatty liver study of associated factors and of relationship with fibrosis.
disease among Korean adults. J Gastroenterol Hepatol Eur J Gastroenterol Hepatol 2006;18:233-237.
2006;21:138-143. 27 hi JP, Fan JG, Wu R, Gao XQ, Zhang L, Wang H, et al.
18 hen L, Fan JG, Shao Y, Zeng MD, Wang JR, Luo GH,
S Prevalence and risk factors of hepatic steatosis and its
et al. Prevalence of nonalcoholic fatty liver among impact on liver injury in Chinese patients with chronic
administrative officers in Shanghai: an epidemiological hepatitis B infection. J Gastroenterol Hepatol 2008;23:1419-
survey. World J Gastroenterol 2003;9:1106-1110. 1425.
19 ishizawa H, Shimomura I, Kishida K, Maeda N,
N 28 eng D, Han Y, Ding H, Wei L. Hepatic steatosis in chronic
Kuriyama H, Nagaretani H, et al. Androgens decrease hepatitis B patients is associated with metabolic factors
plasma adiponectin, an insulin-sensitizing adipocyte- more than viral factors. J Gastroenterol Hepatol 2008;23:
derived protein. Diabetes 2002;51:2734-2741. 1082-1088.
20 i YM, Chen WX, Yu CH, Yue M, Liu YS, Xu GY, et al.
L 29 hen CH, Huang MH, Yang JC, Nien CK, Yang CC, Yeh
An epidemiological survey of alcoholic liver disease in YH, et al. Prevalence and etiology of elevated serum
Zhejiang province. Zhonghua Gan Zang Bing Za Zhi 2003; alanine aminotransferase level in an adult population in
11:647-649. Taiwan. J Gastroenterol Hepatol 2007;22:1482-1489.
21 u XL, Luo JY, Tao M, Zhao P, Zhao HL, Zhang XD, et al.
L 30 amali R, Khonsari M, Merat S, Khoshnia M, Jafari E,
Analysis of dangerous factors for alcoholic liver disease. Bahram Kalhori A, et al. Persistent alanine aminotransferase
Zhonghua Gan Zang Bing Za Zhi 2004;12:442-443. elevation among the general Iranian population: prevalence
22 lark JM, Diehl AM. Defining nonalcoholic fatty
C and causes. World J Gastroenterol 2008;14:2867-2871.
liver disease: implications for epidemiologic studies. 31 lark JM, Brancati FL, Diehl AM. The prevalence and
Gastroenterology 2003;124:248-250. etiology of elevated aminotransferase levels in the United
23 ord ES, Giles WH, Dietz WH. Prevalence of the
F States. Am J Gastroenterol 2003;98:960-967.
metabolic syndrome among US adults: findings from the 32 onardo A, Bellini M, Tartoni P, Tondelli E. The bright
third National Health and Nutrition Examination Survey. liver syndrome. Prevalence and determinants of a
JAMA 2002;287:356-359. "bright" liver echopattern. Ital J Gastroenterol Hepatol
24 archesini G, Bugianesi E, Forlani G, Cerrelli F, Lenzi M,
Manini R, et al. Nonalcoholic fatty liver, steatohepatitis,
and the metabolic syndrome. Hepatology 2003;37:917-923. Received February 18, 2009
25 vailable from: http://www.moh.gov.cn/newshtm1/21624.
A Accepted after revision June 21, 2009
382 • Hepatobiliary Pancreat Dis Int，Vol 8，No 4 • August 15，2009 • www.hbpdint.com