Guangdong Province with a population of about 85 millions (mainly Han nationality people) is located in an economically advanced region in South China. Its capital city, Guangzhou, has 8-million residents. The heterogeneity of economy, geography, and culture in different areas inside the province influences the life styles of the people.

Between April and November of 2005, a cross-sectional survey with multiple-stage stratified cluster and random sampling was performed on the basis of a previous analysis of the population distribution in this province[6,7]. In order to represent the data in general population, sampling clusters were randomly drawn from urban and rural areas all over the province. Six areas with 4365 residents over 7 years of age were selected, i.e., 2 communities from downtown of Guangzhou city (Yuanzhongyuan community of the Liwan district and the Xinyi community of the Baiyun district, with a total of 1436 urban residents) and four units within the province, i.e., a village in central area of the province (Huadu county, 822 rural residents), a village in southern area (Huizhou county, 1037 rural residents), a village in western area (Zhanjiang county, 351 rural residents) and a community in northern area (Shaoguan city, 719 urban residents).

Of the 4365 potential participants, 3903 agreed to take part in this investigation with a responder rate of 89.6%. Among them, 3543 subjects (90.6%, 1311 males and 2232 females, at the age of 7-100 years) provided their complete information, accounting for approximately 44.3/100000 of the total population of Guangdong Province[7]. Interview, physical examination, laboratory assessment and ultrasonographic examination were performed for each subject on the same day at a mobile examination center. The study was approved and supported by Guangzhou Health Bureau. Written consent was obtained from each participant.

A face-to-face interview was carried out by specially trained post-graduate students of Guangzhou Medical College and supervised by experienced investigators. Standard questionnaires, designed by co-working of epidemiologists and hepatologists, included the following items: demographic characteristics, current medication use, medical history and health-relevant behaviors, i.e., alcohol consumption, smoking habits, dietary habits and physical activities. Physical examination covered body height, weight, blood pressure, waist circumference (WC), waist-to-hip ratio (WHR) and routine anthropometric parameters in healthy check-up.

Intravenous blood samples were collected from fasting subjects by routine methods. Fasting serum glucose levels and lipid profiles were measured with an automatic chemistry- immuno-analyzer (Olympus Corporation, Tokyo, Japan). Serum asparatate aminotransferase (AST), and alanine aminotransferase (ALT), serum bilirubin (BIL) and albumin levels were measured by standard laboratory methods. Tests for the serum makers of hepatitis A virus (HAV), hepatitis B virus (HBV) and hepatitis C virus (HCV) were also carried out.

Real-time ultrasonography (US) of the upper abdominal organs was performed for each subject by 2 experienced physicians using a scanner equipped with a 3.5 mmHz transducer (Siemens Adama, German). The physicians performing the US were unaware of the clinical and laboratory results.

FLD, NAFLD and ALD were diagnosed according to the guidelines for diagnosis and treatment of nonalcoholoic and alcoholoic fatty liver diseases issued by Fatty Liver and Alcoholic Liver Disease Study Group of the Chinese Liver Disease Association[8,9], which were adapted from the American Gastroenterological Association[10]. Briefly, the diagnosis of FLD, NAFLD and ALD was based on the combination of medical history, clinical symptoms, laboratory and US findings. Viral hepatitis and other chronic liver diseases needed to be ruled out. Liver biopsy was taken when the diagnosis was suspected. In this epidemiological study, only 12 subjects underwent biopsy. In most cases FLD was diagnosed and staged by US findings. ALD was diagnosed when a subject fulfilling the FLD criteria drank more than 40 g (male) or 20 g (female) alcohol per day over 5 years. Suspected ALD was defined when the alcohol consumption was 20-40 g (male) or 10-20 g (female). NAFLD was diagnosed when the alcohol consumption was less than these amounts.

The US diagnostic patterns of FLD were based on the presence of a ‘bright’ liver (brightness and posterior attenuation) with stronger echoes in the hepatic parenchyma than in the renal parenchyma, vessel blurring and narrowing of the lumen of hepatic veins in the absence of findings suggestive of other chronic liver diseases. The degree of FLD evaluated by US reflected the degree of steatosis. A liver displaying increased but homogenous dot reflection, blurred but still discernible vascular profiles and weak portal vein echogenicity was defined as early-stage. A medium-stage FLD was characterized by disappearance of the portal vascular wall and slightly increased liver volume. Patients with nebulous optical appearance in the liver, little portal vein echogenicity in more than 1/3-1/2 of the liver area and increased liver volume were considered to be in the advanced stage[11-13].

Obesity was categorized according to the body mass index (BMI) criteria for Asians issued by the Regional Office for Western Pacific Region of the World Health Organization[14]. Subjects with BMI ≥ 25 were considered as obese, BMI ≥ 23 but < 25 as overweight, BMI ≥ 18.5 but < 23 as normal, BMI < 18.5 as underweight. Central (or abdominal) obesity was estimated by waist circumference (WC) and waist-to-hip ratio (WHR). A WHR value ≥ 0.9 (male) or ≥ 0.8 (female) was considered as central obesity. Dyslipidemia was considered when having one of following serum lipid profiles: total cholesterol (TC) ≥ 5.72 mmol/L, triglycerides (TG) ≥ 1.70 mmol/L, high-density lipoprotein cholesterol (HDL-C) < 0.91 mmol/L, and low-density lipoprotein cholesterol (LDL-C) ≥ 3.64 mmol/L. Subjects with their fasting serum glucose (FSG) value of ≥ 7.0 mmol/L and/or with a history of diabetes were considered to have diabetes mellitus. Hypertension was defined as systolic or diastolic blood pressure above 140 mmHg or 90 mmHg, respectively. Smokers in this study were defined as those smoking more than one cigarette each day for at least one year or more than 20 packages of cigarettes in total. Alcohol drinkers were defined as those drinking more than 2 times each month for more than 5 years. Heavy tastes meant the habits of preferring salted and spiced foods. Extraverted character meant the active personal characters of subjects. The definitions of urban and rural areas were in accordance with the publication of the Fifth Chinese National Population Census in 2000[7].

Data were examined with the SPSS 13.0 for Windows. P < 0.05 was considered statistically significant. Exposure ratio comparison and multivariate regression analyses were performed to evaluate the risk factors.

Data are shown in Tables 1 and 2. Of the 3543 subjects, 609 (17.2%) were diagnosed having FLD, among them 367 (10.4%), 218 (6.1%) and 24 (0.7%) fell into early, medium and advanced degrees by US staging. The overall prevalence of FLD increased to the peak of 27.4% in the group at the age of 60- years and then decreased in both genders (P < 0.01). There was no prevalent difference between males and females (18.0% vs 16.7%, P > 0.05). However, FLD was more common in males than in females under 50 years of age. An opposite trend was noted over 50 years of age (P < 0.05). The prevalence of FLD in children at the age of 7-18 years (5/379, 1.3%) differed significantly from that in adults (604/3164, 19.1%, P < 0.001). The overall prevalence of ALD, suspected ALD and NAFLD was 0.4%, 1.8% and 15%, respectively, accounting for 2.3%, 10.7% and 87.0% of the 609 FLD subjects, respectively. In subgroup comparison, the prevalence of confirmed ALD and suspected ALD was significantly higher in males (0.9% and 4.0%) than in females (0.1% and 0.5%, P < 0.001). Nevertheless, the prevalence of NAFLD was lower in males than in females (13.0% vs 16.1%, P < 0.05).

The prevalence of FLD was significantly lower in rural areas than in urban areas (12.9% vs 23.0%, P < 0.001), mainly due to the difference in NAFLD (11.1% vs 20.3%, P < 0.001). A slight difference was observed in rural areas between ALD and suspected ALD subjects compared with urban areas, but it was not significant (P > 0.05) (Table 3).

High serum levels of TG, TC, LDL-C and FSG were all positively associated with FLD (P < 0.01), but HDL-C did not reach significance (P > 0.05) (Table 4). The prevalence of FLD in dyslipidemia subjects was 3.65 (95% CI = 2.87- 4.65) folds higher than that in normal controls. High TG levels were the most prominent factor for dyslipidemia with an odd ratio (OR) of 4.49 (95% CI = 3.56-5.66). The OR of high TC was 1.95 (95% CI = 1.56-2.44). The OR of elevated FSG and combined dyslipidemia was 4.93 (95% CI = 3.55-6.84) and 6.64 (95% CI = 4.83-9.12), respectively.

After adjustment for gender, age and residence area according to the demographic characteristics of the Fifth Chinese Census[7], the age-, gender- and residency- specific standardized prevalence of FLD was 14.5%. Confirmed ALD, suspected ALD and NAFLD were 0.5%, 2.3% and 11.7% respectively in adults (over 18 years of age). The overall prevalence of FLD was 11.3% and confirmed ALD, suspected ALD and NAFLD were 0.4%, 1.8% and 9.1% respectively, in all people over 7 years of age.

The exposure ratio (%) analysis of anthropometric and clinical parameters in adults (above 18 years of age, n = 3164) with and without FLD was carried out to screen for the relevant factors for FLD (Table 5). The ratios of all variables for FLD and controls were significantly different (P < 0.01). In order to identify the risk factors, we further performed a multivariate regression logistic analysis with probability for entry 0.05 and removal 0.1 (Table 6). Male gender, urban inhabitation, hypertension, high BMI, WC, WHR, serum TG, and FSG were found to be independent risk factors for FLD, and education was a protective factor for FLD. Alcoholic consumption and heavy tastes in food were marginally but not significantly related to FLD. Smoking and extraverted character were excluded from the risk factors for FLD (P > 0.05).

Fatty liver disease (FLD) has become a common problem in both developed and developing countries. In recent years, due to alterations in life style and dietary habits, the incidence of FLD has increased dramatically in China. It is thus of importance to assess the epidemiological features of FLD in this country in order to facilitate its prevention and treatment. Although some epidemiological studies concerning FLD have been reported in China, most of them were based on clinical settings or on health check-up groups, which do not represent the population data. So far only one population-based study from Shanghai, an eastern city of China, has been published in English. The data from rural areas and other parts of this country are still lacking.

FLD encompasses a morphological spectrum consisting of hepatic steatosis and steatohepatitis which can progress to cirrhosis and hepatocellular carcinoma. In the past, excess alcohol consumption accounted for most cases of FLD, but recently nonalcoholic causes for fatty liver have attracted considerable attention. In addition to alcohol consumption, factors such as insulin resistance (IR), oxidative stress, mitochondrial dysfunction, immune deregulation, and adipokines play an important role in the pathogenesis of FLD. Although FLD has been reported worldwide, it is difficult to determine the true prevalence because of problems in interpreting data from various studies due to referral bias, population heterogeneity, study design, imaging modalities, and liver biopsies. These studies on the FLD epidemiology have improved the detailed knowledge about the prevalence of FLD and the metabolic risk factors associated regional (e.g., urban vs rural) and ethnic differences, as they may provide clues to pathogenesis and individual risk factors for liver disease and metabolic complications. Further studies are also required to confirm the impact of fatty liver on chronic liver diseases.

The incidence of FLD is likely to rise steadily in China in urban areas owing to the westernization of the diet, excessive food intake, increased elderly population, changes in life style and lack of exercise. Because the majority of people live in rural areas of China, it is necessary to investigate the prevalence of FLD in China. In this study, since the random sampling method and high response rate limited the selection and responder biases, our results reflect the population-based data and the major risk factors for FLD in both urban and rural areas. This study assessed the prevalence of and risk factors for NAFLD in China.

With regards to increasing obesity, FLD, especially NAFLD, is believed to be the most common form of chronic liver diseases, which can progress to cirrhosis. Because there is no proven management for NAFLD, identification of important risk factors for this condition will provide valuable information on both risk stratification and development of risk-reduction strategies.

Fatty liver disease (FLD), defined as lipid accumulation exceeding the normal range of 5% of liver wet weight, is a kind of chronic disorders, including alcoholic liver disease (ALD) and nonalcoholic fatty liver diseases (NAFLD). Briefly, the diagnosis of FLD, NAFLD and ALD is based on the combination of medical history, clinical symptoms, laboratory and imaging modalities (ultrasonography, computed tomography scans, and magnetic resonance imaging). Viral hepatitis and other chronic liver diseases need to be ruled out. ALD is diagnosed when a subject fulfilling the FLD criteria drinks more than 40 g (male) or 20 g (female) alcohol per day over 5 years. NAFLD is diagnosed when alcohol consumption is less than 40 g (male) or 20 g (female).

This is an interesting, well written and generally well designed study. Although it does not identify fundamentally new or unexpected findings with respect to the incidence of FLD, the data highlighting the nonalcohol - related incidence of FLD in developing areas are important.