American Journal of Gastroenterology ISSN 0002-9270
C©
2008 by Am. Coll. of Gastroenterology doi: 10.1111/j.1572-0241.2008.01933.x
Published by Blackwell Publishing
Validation of the NAFLD Fibrosis Score in a Chinese
Population With Low Prevalence of Advanced Fibrosis
Vincent Wai-Sun Wong, M.B.Ch.B., M.R.C.P.,
1,2
Grace Lai-Hung Wong, M.B.Ch.B., M.R.C.P.,
1,2
Angel Mei-Ling Chim, M.P.H.,
1,2
Ada Mei-Ling Tse, B.Sc.,
1,2
Steven Woon-Choy Tsang, M.B.Ch.B.,
M.R.C.P.,
3
Alex Yui Hui, M.B.B.S., M.R.C.P.,
1,2
Paul Cheung-Lung Choi, F.R.C.Path.,
4
Anthony Wing-Hung
Chan, M.B.Ch.B., M.R.C.Path.,
4
Wing-Yee So, M.B.Ch.B., M.R.C.P.,
2
Francis Ka-Leung Chan, M.D.,
F.R.C.P.,
1,2
Joseph Jao-Yao Sung, M.D., Ph.D.,
1,2
and Henry Lik-Yuen Chan, M.D., F.R.C.P.,
1,2
1
Institute of Digestive Disease and
2
Department of Medicine and Therapeutics, The Chinese University of Hong
Kong, Hong Kong SAR, China;
3
Department of Medicine, Tseung Kwan O Hospital, Hong Kong SAR, China;
and
4
Department of Anatomical and Cellular Pathology, The Chinese University of Hong Kong,
Hong Kong SAR, China
OBJECTIVES: Nonalcoholic fatty liver disease (NAFLD) is increasing in prevalence worldwide. This study aimed to
validate the NAFLD fibrosis score in the Chinese population.
METHODS: NAFLD patients were prospectively recruited for liver biopsy and blood tests. The NAFLD fibrosis
score was calculated as −1.675 + 0.037 × age (yr) + 0.094 × BMI (kg/m
2
) + 1.13 × impaired
fasting glucose/diabetes (yes = 1, no = 0) + 0.99 × AST/ALT ratio–0.013 × platelet
(×10
9
/L)−0.66 × albumin (g/dL). Advanced fibrosis was defined as stage 3 to 4 fibrosis.
RESULTS: One hundred sixty-two patients (age 46 ± 10 yr, male 59%) were included in the study. Advanced
fibrosis was found in 18 (11%) patients. Only 11 of 128 patients with the NAFLD fibrosis score
below the proposed low cutoff point (<−1.455) were under-staged, resulting in a high negative
predictive value of 91%. Only two patients exceeded the proposed high cutoff point (>0.676), but
neither had advanced fibrosis. If the NAFLD fibrosis score was implemented in the Chinese
population, 79% of liver biopsies could be avoided.
CONCLUSIONS: The NAFLD fibrosis score has high negative predictive value in excluding advanced fibrosis in the
Chinese population, and can reduce the burden of liver biopsy in the vast majority of cases. Since
there were few cases of advanced fibrosis in this cohort, this study had limited power in validating
the high cutoff point.
(Am J Gastroenterol 2008;103:1682–1688)
INTRODUCTION
Nonalcoholic fatty liver disease (NAFLD) is becoming the
most common chronic liver disease (1). It increases overall
mortality and it is associated with a higher cardiovascular risk
(2, 3). Previously thought to be a benign condition, NAFLD
is now believed to be the most important cause of cryptogenic
cirrhosis, and it may even progress to liver failure and hepato-
cellular carcinoma in some cases (4). In a long-term follow-up
study in America, up to 13% of NAFLD patients died at a
mean of 7.6 yr, and liver-related complications were an im-
portant cause of death (5). Alarmingly, developing countries
are not immune to this problem. For example, approximately
15% of the population in China was found to suffer from
NAFLD according to an ultrasonographic survey (6). Among
Asians with NAFLD, significant hepatic necroinflammation
is common and liver fibrosis may worsen with time (7, 8).
While patients with significant fibrosis or cirrhosis tend
to develop complications, those with simple steatosis run a
benign course and do not appear to have increased mortality
compared to the general population (5, 9). Therefore, it is
important to assess the degree of liver fibrosis in NAFLD
patients for prognostic purposes. Currently, the gold standard
to assess liver fibrosis is liver biopsy. According to the Asia-
Pacific Working Party on NAFLD, liver biopsies should be
considered in patients with risk factors for advanced fibrosis
(10). Thrombocytopenia, low serum albumin, elevated serum
bilirubin, and prolonged prothrombin time were features of
cirrhosis, but none of these are sensitive enough to detect early
cirrhosis. Although major complications due to liver biopsy
are rare, it is expensive and may induce pain and anxiety in
patients.
In view of the limitation of liver biopsy, there has been con-
certed effort to develop noninvasive tests for the assessment
1682

NAFLD Fibrosis Score in Chinese 1683
of liver fibrosis (11). Recently, an NAFLD fibrosis score was
derived using data from a number of major centers in the
United States, Europe, and Australia (12). With a formula
including age, body mass index (BMI), impaired fasting glu-
cose (IFG) or diabetes, aspartate aminotransferase-to-alanine
aminotransferase (AST/ALT) ratio, platelet count, and
albumin, the score had an overall accuracy of 80–90% in
predicting stage 3 or 4 liver fibrosis. Compared to previously
reported serum markers, the main advantage of this score
is that it involves parameters that are widely available and
almost routinely measured. Besides, relatively few patients
were in the indeterminate zone, in which the score would be
unable to predict whether there was a significant fibrosis.
Nevertheless, over 90% of the patients used to derive and
validate the NAFLD fibrosis score were whites (12). It needs
to be validated in other ethnic groups before it can be ap-
plied globally. In Asian patients, steatohepatitis and other
metabolic complications tend to develop at a lower BMI,
which is one of the factors in the equation of the NAFLD fi-
brosis score (13). It is therefore uncertain if ethnicity affects
the accuracy of the score. In this study, we aimed to assess the
accuracy of the NAFLD fibrosis score to predict significant
liver fibrosis in a prospective cohort of Chinese patients with
NAFLD.
METHODS
Patients
From December 2004 to May 2007, we prospectively re-
cruited patients in the liver and general medical clinics of
the Prince of Wales Hospital and Tseung Kwan O Hospital,
Hong Kong. Inclusion criteria included the presence of fatty
liver on imaging studies plus (1) persistent elevation of ALT
above the upper limit of normal for two consecutive visits
at least 12 wk apart, or (2) risk factors for advanced fibrosis
(e.g., obesity and diabetes) (10). We excluded male patients
who consumed more than 30 g of alcohol per day and female
patients who consumed more than 20 g per day. Patients with
coexisting liver disease, namely chronic viral hepatitis, au-
toimmune hepatitis, primary biliary cirrhosis, primary scle-
rosing cholangitis, Wilson’s disease, hemochromatosis, α1-
antitrypsin deficiency, biliary obstruction, and drug-induced
liver disease were excluded. Secondary causes (e.g., corti-
costeroid use, gastric bypass surgery) of liver steatosis were
also excluded.
Patients meeting the inclusion and exclusion criteria under-
went standard clinical evaluation, laboratory tests, followed
by liver biopsies. All patients gave informed written consent.
The study protocol was approved by the Ethics Committee of
the Chinese University of Hong Kong.
Clinical Evaluation and Laboratory Tests
Comprehensive clinical assessment was performed. Comor-
bid illness and drug/herb intake were recorded with a standard
questionnaire. Anthropometric tests included body weight,
body height, hip and waist circumference measurements.
BMI was calculated as weight (kg) divided by height (m)
squared. Waist–hip ratio (WHR) was defined as waist cir-
cumference (cm) divided by hip circumference (cm). Waist
circumference was measured at a level midway between the
lower rib margin and iliac crest with the tape all around the
body in the horizontal position.
On the day of liver biopsy, a fasting venous blood sam-
ple was taken for albumin, bilirubin, AST, ALT, glucose,
glycosylated hemoglobin (HbA
1c
), total cholesterol, HDL-
cholesterol, LDL-cholesterol, triglycerides, and insulin. The
upper limits of normal for AST and ALT at our labora-
tory were 64 IU/L and 58 IU/L, respectively. Insulin resis-
tance was calculated by the homeostasis model (HOMA-IR),
which was equal to fasting insulin (mU/L)× fasting glucose
(mmol/L)/22.5. Insulin level (Dako, Ely, UK) was measured
by commercial ELISA kits.
Patients with fasting plasma glucose below 7.0 mmol/L
underwent a 75-g oral glucose tolerance test. A patient was
diagnosed to have diabetes mellitus if the fasting glucose was
above 7.0 mmol/L or 2-h postglucose load plasma glucose
was above 11.1 mmol/L (14). IFG was defined as a fasting
glucose between 5.6 and 7.0 mmol/L.
The NAFLD fibrosis score was calculated according to the
following formula: −1.675 + 0.037 × age (yr) + 0.094 ×
BMI (kg/m
2
) + 1.13 × IFG/diabetes (yes = 1, no = 0) +
0.99 × AST/ALT ratio–0.013 × platelet (×10
9
/L)–0.66 ×
albumin (g/dL) (12).
AST/ALT ratio was calculated as AST (IU/L) divided by
ALT (IU/L). The AST to platelet ratio index (APRI) was
calculated as AST (IU/L)× upper limit of normal for AST×
100 / platelet (×10
9
/L) (15). The HAIR score was calculated
by summation of the scores of hypertension = 1, ALT > 40
IU/L= 1, and insulin resistance (IR) index> 5= 1. IR index
was calculated as log insulin (µU/mL) + log fasting plasma
glucose (mg/dL) (16).
Histological Assessment
Percutaneous liver biopsy was performed using the 16G
Temno needle. Liver histology was assessed by histopatholo-
gists (P.C.C., A.W.C.) who were blinded to the clinical data. A
sample was considered adequate if it was longer than 1.5 cm
and contained five portal tracts or more. Liver biopsy spec-
imens were prepared with hematoxylin and eosin stain,
Masson trichrome stain, Prussian blue stain, reticulin stain,
orcein stain, and periodic acid Schiff stain. The histological
grading and staging of NAFLD followed the Brunt’s crite-
ria (17). Macrovesicular steatosis was graded from 0 to 3,
necroinflammatory activity was graded from 0 to 3, and fibro-
sis was staged from 0 to 4. Stage 0= absence of fibrosis; stage
1 = perisinusoidal or portal; stage 2 = perisinusoidal and
portal/periportal; stage 3 = septal or bridging fibrosis; and
stage 4= cirrhosis. Significant fibrosis was defined as stage 2
fibrosis or above. Advanced fibrosis was defined as stage 3
or 4 fibrosis.
Data Analysis and Statistics
Statistical analysis was performed by Statistical Package
for Social Science (SPSS) version 11.5 (Chicago, IL).