Correspondence: Serpil Erdogan, Ankara Numune E g itim ve Aras t ı rma Hastanesi Klinik Biyokimya B ö l ü m ü 06100 Samanpazar ı , Ankara, Turkey. Tel: � 903 1250 84000. Fax: � 903 1250 84910. E-mail: drsturhan@yahoo.com

(Received 7 August 2012 ; accepted 3 February 2013 )

ORIGINAL ARTICLE

The diagnostic value of non-invasive tests for the evaluation of liver fi brosis in chronic hepatitis B patients

SERPIL ERDOGAN 1 , HALEF OKAN DOGAN 1 , SEVILAY SEZER 1 , SEMA UYSAL 1 , ESRA OZHAMAM 2 , SERRA KAYACETIN 2 & YUKSEL KOCA 1

Departments of 1 Clinical Biochemistry, and 2 Clinical Pathology, Ankara Numune Training and Research Hospital , Ankara , Turkey

Abstract Background . Liver biopsy, which is considered the gold standard for the evaluation of hepatic fi brosis in patients with chronic hepatitis B (CHB), has certain limitations. The aim of this study was to investigate the diagnostic performance of non-invasive markers of hepatic fi brosis as potential alternatives to liver biopsy. Methods . The medical records of 221 patients with a diagnosis of CHB who underwent a liver biopsy were reviewed. Indirect indicators of fi brosis were calculated for each patient based on previously described formulas [Aspartate aminotransferase (AST)/alanine aminotransferase (ALT) ratio (AAR), age-platelet index (API), cirrhosis discriminant score (CDS), AST-platelet ratio index (APRI), Forns index, FIB-4, Pohl score, AAR-platelet score (AARP), fi bro-quotient (FibroQ), AST/platelet/Gammaglutamyl transpeptidase (GGT)/Alphafetoprotein (AFP) (APGA) index, Platelet/Age/Phosphatase (ALP)/AFP/AST (PAPAS) index, Lok ’ s model, Goteborg University Cirrhosis Index (GUCI)]. Diagnostic adequacy of these indices was evaluated by receiver operating characteristic curve analysis. Results . Area under the receiver operating characteristic curves for the FIB-4, Forns, GUCI, APRI, PAPAS, APGA and FibroQ indices were 0.701, 0.680, 0.670, 0.670, 0.639, 0.638 and 0.588, respectively. The AAR, API, CDS and AARP indices, Pohl score and Lok ’ s model were all deemed diagnostically inadequate. FIB-4 had the best diagnostic adequacy whereas AAR had the worst. Conclusions . Our results suggest that out of the 13 indices evaluated, only FIB-4 index may be useful in estimating the extent of fi brosis in patients with CHB. There is a need for more comprehensive prospective studies to help determine the diagnostic value of non-invasive tests for liver fi brosis.

Key Words: Alanine aminotransferase , aspartate aminotransferase , chronic hepatitis B , liver fi brosis , platelet count

Introduction

Chronic liver diseases (CLD) are a signifi cant cause of morbidity and mortality worldwide [1], with hep-atitis B virus (HBV) infections comprising nearly one-tenth of all CLDs [2]. It is estimated that 2 billion individuals are infected with HBV worldwide, with 350 million developing chronic infections. Although the majority of infected patients do not develop complications, approximately 15 – 40% eventually progress to cirrhosis, liver failure and hepatocellular carcinoma (HCC) [3 – 5]. Cirrhosis and HCC are the leading causes of death in CLD patients [6].

Three stages of CHB infection have been described, namely the immune tolerant phase, immune active phase and inactive phase, the distinc-tion of which is based on the presence of hepatitis B

‘ e ’ antigen (HBeAg) and serum ALT concentration and HBV DNA viral load, as well as the extent of infl ammation and fi brosis in the liver [7]. Although the dynamic nature of CHB infection means that patients may alternate back and forth between stages, disease activation results in the progression of liver fi brosis and liver injury which eventually lead to the development of cirrhosis and HCC [8].

Liver biopsy remains the gold standard for the evaluation of liver fi brosis [9]. Nevertheless, a liver biopsy does not provide information regarding the balance between production and destruction of the extracellular matrix (ECM) or the rate of progression to cirrhosis; in other words liver biopsy is a static test [10]. It is an invasive procedure with major complica-tions or even death occurring in 0.5% of patients [11], and is contraindicated in disorders of hemostasis

Scandinavian Journal of Clinical & Laboratory Investigation, 2013; 73: 300–308

ISSN 0036-5513 print/ISSN 1502-7686 online © 2013 Informa HealthcareDOI: 10.3109/00365513.2013.773592

Scan

d J

Clin

Lab

Inv

est D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Lav

al o

n 06

/23/

14Fo

r pe

rson

al u

se o

nly.

Liver fi brosis in chronic hepatitis B 301

[12]. A typical biopsy specimen represents only 1/50,000 of the liver which along with non-negligible inter-observer variability are other outstanding issues regarding the diagnostic accuracy of liver biopsy [11,13]. Furthermore, some patients may refuse a liver biopsy which creates a demand for non-invasive markers [9].

Several non-invasive methods and scores have been described to evaluate the degree of hepatic fi brosis with varying diagnostic yield and confl icting results.

The aim of this study was to determine the diag-nostic validity of several indirect measures of fi brosis (AAR, API, CDS, APRI, Forns index, FIB-4, Pohl score, AARP, FibroQ, APGA index, PAPAS index, Lok ’ s model, GUCI) in patients with CHB.

Materials and methods

Study design and patient population

This retrospective study was undertaken in Ankara Numune Training and Research Hospital with the approval of the local ethics committee. The medical records of patients with a diagnosis of CHB, defi ned by hepatitis B ‘ s ’ antigen positivity of more than 6 months duration with a negative anti-HBs antibody, who underwent a liver biopsy between March 2010 and June 2012 were systematically reviewed and screened for eligibility. Patients with evidence of a concurrent liver disorder such as autoimmune hepa-titis, primary biliary cirrhosis, HCC, hepatitis C and D co-infection, and Wilson ’ s disease were excluded from the study. Other exclusion criteria included a daily alcohol intake of more than 20 grams, and

elevated urea nitrogen and creatinine concentrations more than 1.5 times the upper reference limit.

Laboratory assays

The date of liver biopsy was identifi ed for each patient, and coinciding results of laboratory tests such as serum concentrations of glucose, alpha-feto-protein (AFP), Aspartate aminotransferase (AST), ALT, gamma-glutamyl transferase, alkaline phos-phatase (ALP), total bilirubin, albumin, globulin, total cholesterol (TC), urea nitrogen and creatinine, along with results of complete blood count and coag-ulation parameters (prothrombin time and INR) were recorded. Results of the viral serological tests HBsAg, anti-HBs antibody, HBeAg, anti-HBe anti-body, HBV-DNA, delta antigen, anti-D, anti-HIV and anti-HCV antibody were also noted.

Histopathological evaluation

All liver biopsies in our centre are obtained percuta-neously under ultrasound guidance with a trucut needle (14 – 16 Ch), after which they are fi xed and then embedded in paraffi n, followed by staining with haematoxylin-eosin, Masson ’ s trichrome and reticu-lin. The specimens of the selected subjects were retrieved from the archives and re-evaluated by two independent pathologists with experience in liver his-tology who were blinded to patient details. The pathologists were asked to provide a fi brosis stage for each specimen based on the Ishak scoring system (0 – 6) [12]. Patients with a score of 3 more were classifi ed as having ‘ signifi cant ’ fi brosis.

Table I. The formula of indirect fi brosis markers.

Parameters Formula

AAR [13] AST/ALTAPI [14] Age (yr): � 30 � 0; 30 – 39 � 1; 40 – 49 � 2; 50 – 59 � 3; 60 – 69 � 4; � 70 � 5 PLT count (10 9 /L): � 225 � 0;

200 – 224 � 1; 175 – 199 � 2; 150 – 174 � 3; 125 – 149 � 4; � 125 � 5 API is the sum of the aboveCDS [15] PLT count (10 9 /L): � 340 � 0; 280 – 339 � 1; 220 – 279 � 2; 160 – 219 � 3; 100 – 159 � 4; 40 – 99 � 5;

� 40 � 6 ALT/AST ratio: � 1.7 � 0; 1.2 – 1.7 � 1; 0.6 – 1.19 � 2; � 0.6 � 3 INR: � 1.1 � 0; 1.1 – 1.4 � 1; � 1.4 � 2 CDS is the sum of the above

APRI [16] [(AST/ULN)/PLT (10 3 /L)] � 100FIB-4 [18] [Age (yr) � AST (U/L)]/[PLT (10 9 /L) � (ALT (U/L)) 1/2 ]FibroQ [21] (10 � age � AST � PT INR)/(PLT � ALT)Pohl score [19] Positive: AAR � 1 and PLT � 150 � 10 9 /LLok ’ s model [23] Log odds (predicting cirrhosis) � 5.56 – 0.0089 � Platelet count (10 9 /L) � 1.26 � AST/ALT

ratio � 5.27 � PT INR Predicted probability � exp (log odds)/[1 � exp (log odds)]GUCI [24] [(AST/ULN) � prothrombin-INR] � 100/ PLT (10 9 /L)AARP [20] Positive: AAR � 1 or PLT � 150 � 10 9 /LForns index [17] 7.811 – 3.131 � ln(PLT) � 0.781 � ln(GGT) � 3.467 � ln(age) 0.014 � (cholesterol)APGA index [5] Log (index) � 1.44 � 0.1490 � log (GGT) � 0.3308 � log (AST) 0.5846 � log (PLT) � 0.1148 � log (AFP � 1)PAPAS index [22] Log (index � 1) � 0.0255 � 0.0031 � age (years) � 0.1483 � log[ALP(U/L)] � 0.004 � log[AST(U/L)]

� 0.0908 � log[AFP(ng/mL) � 1] 0.028 � log[PLT (10 9 /L)]

AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; GGT, γ -glutamyl transpeptidase; PLT, Platelet; PT-INR, prothrombin time international normalized ratio; AAR, aspartate aminotransferase/alanine aminotransferase ratio; API, age/platelet index; CDS, cirrhosis discriminant score; APRI, AST/platelet ratio index; FibroQ, fi bro-quotient; GUCI, Goteburg University Cirrhosis Index; AARP, AAR-platelet score; APGA, AST/platelet/GGT/AFP index; PAPAS, Platelet/Age/Phosphatase/AFP/AST index.

Scan

d J

Clin

Lab

Inv

est D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Lav

al o

n 06

/23/

14Fo

r pe

rson

al u

se o

nly.

302 S. Erdogan et al.

Table II. Characteristics of CHB patients.

Total ( n � 221)

Ishak fi brosis Score 0-1-2

( n � 153)

Ishak fi brosis Score 3-4-5-6

( n � 68) P value

Age (years) (mean SD) 43.68 12.56 42.10 12.58 47.22 11.87 0.005Sex (Number %)

Male Female

140 (63.3%) 81 (36.7%)

90 (40.7%) 63 (28.5%)

50 (22.6%) 18 (8.1%)

0.037

HBeAg (Number %) HBeAg-negative HBeAg-positive

187 (84.6%) 34 (15.4%)

127 (57.5%) 26 (11.8%)

60 (27.1%) 8 (3.6%)

ns

PLT ( � 10 9 /L) (mean SD) 212.14 68.42 216.16 71.12 203.07 61.44 nsPT-INR [median (min-max)] 1.03 (0.14 – 1.58) 1.01 (0.84 – 1.58) 1.05 (0.94 – 1.31) 0.003AST (IU/L) [median

(min-max)]28 (11 – 950) 26.5 (11 – 950) 32 (15 – 263) 0.000

ALT (IU/L) [median (min-max)]

32 (10 – 637) 34 (10 – 510) 39 (17 – 637) 0.001

ALP (IU/L) [median (min-max)]

63 (8 – 652) 58.5 (8 – 652) 66 (36 – 113) ns

GGT (IU/L) [median (min-max)]

20 (2 – 384) 18 (2 – 153) 20 (9 – 384) 0.014

Bilirubin ( μ mol/L) [median (min-max)]

12.65 (1.03 – 39.67) 13.68 (1.03 – 37.62) 13.68 (6.84 – 39.67) ns

TC (mmol/L) (mean SD) 4.43 1.03 4.57 1.04 4.15 0.95 0.021Albumin (g/L) (mean SD) 41.40 5.88 41.56 4.96 41.06 7.58 nsGlobulin (g/L) [median

(min-max)]29 (17 – 46) 29 (17 – 42) 29.5 (24 – 46) 0.005

Urea nitrogen (mmol/L) (mean SD)

4.49 0.83 4.41 1.36 4.69 1.34 ns

Creatinine ( μ mol/L) (mean SD)

68.95 15.03 68.07 15.91 69.84 13.26 ns

WBC ( � 10 9 /L) [median (min-max)]

6.60 (2 – 71) 6.22 (3.3 – 11.5) 6.75 (2 – 71) ns

RBC ( � 10 12 /L) (mean SD) 4.86 0.49 4.82 0.51 4.93 0.42 nsHb (g/dL) (mean SD) 14.43 1.74 14.23 1.82 14.88 1.47 0.010AAR [median (min-max)] 0.85 (0.20 – 4.96) 0.84 (0.41 – 1.86) 0.82 (0.41 – 1.38) nsAPI [median (min-max)] 6 (5 – 10) 6 (5 – 10) 6 (5 – 10) nsCDS [median (min-max)] 4 (1 – 8) 4 (2 – 8) 4 (1 – 7) nsAPRI [median (min-max)] 0.32 (0.09 – 12.26) 0.30 (0.09 – 12.26) 0.33 (0.17 – 3.08) 0.000FIB-4 [median (min-max)] 1.04 (0.24 – 10.83) 0.93 (0.24 – 10.83) 1.20 (0.39 – 3.09) 0.000FibroQ [median (min-max)] 1.74 (0.30 – 19.15) 1.62 (0.30 – 19.15) 2.02 (0.50 – 6.49) 0.037Pohl [median (min-max)] 1 (1 – 2) 1 (1 – 2) 1 (1 – 2) nsLok ’ s model [median

(min-max)]0.29 (0.01 – 0.99) 0.27 (0.03 – 0.99) 0.28 (0.01 – 0.75) 0.055

GUCI [median (min-max)] 0.32 (0.04 – 13.97) 0.30 (0.09 – 13.97) 0.41 (0.16 – 3.23) 0.000AARP [median (min-max)] 1.5 (1 – 2) 1.5 (1 – 2) 1 (1 – 2) nsForns (mean SD) 4.33 1.77 3.99 1.81 5.00 1.48 0.001APGA [median (min-max)] 6.74 (3.08 – 36.92) 6.77 (3.08 – 36.92) 7.70 (4.02 – 15.13) 0.004PAPAS (mean SD) 1.67 0.33 1.62 0.32 1.79 0.35 0.002

CHB, chronic hepatitis B; SD, Standard deviation; ns, not signifi cant; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; GGT, γ -glutamyl transpeptidase; PLT, Platelet; PT-INR, prothrombin time international normalized ratio; TC, total cholesterol; WBC, white blood cell; RBC, red blood cell; Hb, Hemoglobin; AAR, aspartate aminotransferase/alanine aminotransferase ratio; API, age/platelet index; CDS, cirrhosis discriminant score; APRI, AST/platelet ratio index; FibroQ, fi bro-quotient; GUCI, Goteburg University Cirrhosis Index; AARP, AAR-platelet score; APGA, AST/platelet/GGT/AFP index; PAPAS, Platelet/Age/Phosphatase/AFP/AST index.

Indirect markers of fi brosis

Indirect markers of fi brosis were calculated for each patient based on previously described formulas which have been summarized in Table I (AAR [13], API [14], CDS [15], APRI [16], Forns index [17], FIB-4 [18], Pohl score [19], AARP [20], FibroQ [21], APGA index [5], PAPAS index [22], Lok ’ s model [23], GUCI [24]).

Statistical analysis

Statistical analyses were performed using the PASW Statistics 18 software. Normality of distribution was evaluated using the Kolmogorov-Smirnov test. Com-parisons of variables with a normal distribution were made using Student ’ s t -test, and values were provided as mean standard deviation (SD). For parameters with an abnormal distribution, the Mann-Whitney U

Scan

d J

Clin

Lab

Inv

est D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Lav

al o

n 06

/23/

14Fo

r pe

rson

al u

se o

nly.

Liver fi brosis in chronic hepatitis B 303

test was used for comparisons, and values were given as median (minimum-maximum). Correlation with histopathological fi brosis stage was evaluated with either the Pearson or Spearman-rank test, depending on normality of distribution of the parameter in question. Receiver operating characteristic (ROC)

curves and area under the ROC (AUROC) curves were plotted for each indirect indicator of fi brosis to evaluate and compare their diagnostic value [25]. Following determination of optimal cut-off values, sensitivity, specifi city, positive predictive value (PPV), negative predictive value (NPV), likelihood ratio � (LR � ) and likelihood ratio (LR ) were calculated for each parameter. A P -value of less than 0.05 was considered indicative of statistical signifi cance.

Results

Patient characteristics

Of all the medical records screened, 221 treatment-naive CHB patients were deemed eligible for inclu-sion in the fi nal analysis. Sixty-eight out of the 221 (30.8%) patients had signifi cant fi brosis with an Ishak fi brosis score of � 3 (F0, 15.8%; F1, 29.4%; F2, 24%; F3, 16.7%; F4, 7.7%; F5, 5%; F6, 1.4%). Demographic, clinical and histopathological features of the study population are summarized in Table II.

Patients were divided into two groups based on Ishak fi brosis score (F0 – 2 and F3 – 6). Although the F3 – 6 group had signifi cantly more male patients compared to the F0 – 2 group, there was no difference between the groups in terms of HBeAg positivity rate, platelet-, red blood cell- and white blood cell counts, serum concentrations of albumin, ALP, total biliru-bin, urea nitrogen, creatinine, AAR, API, CDS, Pohl score, AARP index, Lok ’ s model. TC concentrations

ROC Curve

AARAPICDSAARPPohlAPRIFB4FibroQLoksGUCIReference Line

0.00.0

0.2

0.4Sens

itivi

ty 0.6

0.8

1.0

0.2 0.41 - Specificity

0.6 0.8 1.0

ROC Curve

0.00.0

0.2

0.4

Sens

itivi

ty 0.6

0.8

1.0

0.2 0.41 - Specificity

0.6 0.8 1.0

ROC Curve

0.00.0

0.2

0.4

Sens

itivi

ty 0.6

0.8

1.0

0.2 0.41 - Specificity

0.6 0.8 1.0

ROC Curve

0.00.0

0.2

0.4

Sens

itivi

ty 0.6

0.8

1.0

0.2 0.41 - Specificity

0.6 0.8 1.0

(B)

(C)

(A)

(D)

Source of theCurve

Figure 1. Receiver operating characteristic curves of (A) AAR, API, CDS, AARP, Pohl score, APRI, FIB-4, FibroQ, Lok ’ s model, GUCI; (B) Forns index; (C) APGA index; (D) PAPAS index for prediction of signifi cant fi brosis. AAR, aspartate aminotransferase/alanine aminotransferase ratio; API, age/platelet index; CDS, cirrhosis discriminant score; APRI, AST/platelet ratio index; FibroQ, fi bro-quotient; GUCI, Goteburg University Cirrhosis Index; AARP, AAR-platelet score; APGA, AST/platelet/GGT/AFP index; PAPAS, Platelet/Age/Phosphatase/AFP/AST index.

Table III. Correlation of histological fi brosis severity with the variables.

Correlation coeffi cient P

Age * 0.227 0.001PT-INR * * 0.184 0.006AST * * 0.305 0.000ALT * * 0.195 0.004GGT * * 0.173 0.011TC * 0.176 0.038Globulin * * 0.229 0.001APRI * * 0.252 0.000FIB-4 * * 0.281 0.000FibroQ * * 0.145 0.031GUCI * * 0.253 0.000Forns * 0.266 0.002APGA * * 0.173 0.022PAPAS * 0.274 0.000

PT-INR, prothrombin time international normalized ratio; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; GGT, γ -glutamyl transpeptidase; TC, total cholesterol; APRI, AST/platelet ratio index; FibroQ, fi bro-quotient; GUCI, Goteburg University Cirrhosis Index; APGA, AST/platelet/GGT/AFP index; PAPAS, Platelet/Age/Phosphatase/AFP/AST index. * Pearson correlation; * * Spearman ’ s rank correlation.

Scan

d J

Clin

Lab

Inv

est D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Lav

al o

n 06

/23/

14Fo

r pe

rson

al u

se o

nly.

304 S. Erdogan et al.

were signifi cantly lower in the F3 – 6 group, whereas mean values for PT/INR, AST, ALT, GGT, globulin, and Hb were signifi cantly higher. Mean values obtained for all the non-invasive markers of fi brosis (APRI, FIB-4, FibroQ, GUCI, Forns, APGA and PAPAS indices) were higher in the F3 – 6 group (Table II).

Results of correlation analysis between Ishak fi brosis stage and indirect markers of fi brosis

Of all the parameters evaluated, AST concentrations were found to have the strongest positive correlation with histopathological fi brosis stage. Other signifi -cant positive correlations were observed with age and PT/INR, as well as with serum concentrations of ALT, GGT and globulin. A negative correlation was observed between TC concentrations and fi brosis score. Although all the indirect markers of fi brosis evaluated showed a signifi cant positive correlation with Ishak fi brosis score, the FIB-4 index was found to have the most signifi cant correlation ( r � 0.281; P � 0.001). Results of correlation analyses are summarized in Table III.

ROC curve analysis

The ROC curves for the indirect markers of fi brosis (AAR, API, CDS, AARP, Pohl score, APRI, FIB-4, FibroQ, Lok ’ s model, GUCI, Forns, APGA and PAPAS indices) have been plotted in Figure 1. The diagnostic values of the AAR, API, CDS and AARP indices, as well as the Pohl score and Lok ’ s model were deemed insuffi cient ( P � 0.05). AUROC values for the FIB-4, Forns, GUCI, APRI, PAPAS, APGA and FibroQ were 0.701, 0.680, 0.670, 0.670, 0.639, 0.638 and 0.588, respectively (Table IV).

Sensitivity, specifi city, positive predictive, nega-tive predictive, LR � and LR values were calculated based on the cut-off points determined for the 13 indirect indices, a summary of which is provided in Table IV.

Discussion

Although hepatic fi brosis may be considered a regen-erative response to liver injury, it is the main factor responsible for the development of advanced CLDs such as cirrhosis [26]. The prognosis of CLDs is

Table IV. Performance of indirect fi brosis markers for signifi cant fi brosis, and sensitivity, specifi city, predictive values and likelihood ratios of indexes according to different cut-offs for predicting signifi cant fi brosis.

AUROC (95% CI)

Optimal cut-off

Sensitivity %

Specifi city %

PPV %

NPV % LR � LR

AAR 0.509 (0.422 – 0.595)

0.930 41.18 60.13 31.46 69.70 1.03 0.981.102 26.47 83.66 41.86 71.91 1.62 0.88

API 0.529 (0.444 – 0.613)

5.50 64.71 38.56 31.88 71.08 1.05 0.928.50 20.59 88.24 43.75 71.43 1.75 0.90

CDS 0.547 (0.463 – 0.631)

3.50 75.00 31.37 32.69 73.85 1.09 0.806.50 10.29 95.42 50.00 70.53 2.25 0.94

APRI 0.670 * (0.593 – 0.748)

0.21 86.76 27.45 34.71 82.35 1.20 0.481.02 14.71 94.77 55.56 71.43 2.81 0.90

FIB-4 0.701 * (0.627 – 0.775)

1.02 77.94 58.82 45.69 85.71 1.89 0.382.01 25.00 92.81 60.71 73.58 3.48 0.81

FibroQ 0.588 * (0.503 – 0.674)

1.43 73.53 37.25 34.25 76.00 1.17 0.712.67 30.88 81.05 42.00 72.51 1.63 0.85

Pohl score 0.516 (0.433 – 0.600)

Positive 5.88 97.39 50.00 69.95 2.25 0.97

Lok ’ s model 0.581 (0.499 – 0.663)

0.23 73.53 35.95 33.78 75.34 1.15 0.740.35 36.76 70.59 35.71 71.52 1.25 0.90

GUCI 0.670 * (0.591 – 0.749)

0.31 69.12 55.56 40.87 80.19 1.56 0.560.57 44.12 85.62 57.69 77.51 3.07 0.65

AARP 0.547 (0.464 – 0.629)

Positive 48.53 60.78 35.48 72.66 1.24 0.85

Forns 0.680 * (0.590 – 0.770)

3.11 91.49 31.52 40.57 87.88 1.34 0.275.11 42.55 75.00 46.51 71.88 1.70 0.77

APGA 0.638 * (0.548 – 0.728)

6.12 69.81 42.15 34.58 76.12 1.21 0.728.37 43.40 80.99 50.00 76.56 2.28 0.70

PAPAS 0.639 * (0.548 – 0.730)

1.36 92.16 19.01 32.41 85.19 1.14 0.412.02 27.45 90.91 56.00 74.83 3.02 0.80

PPV, positive predictive value; NPV, negative predictive value; LR � , positive likelihood ratio; LR , negative likelihood ratio; AUROC, area under the receiver operating characteristic curve; 95% CI, 95% confi dence interval; AAR, aspartate aminotransferase/alanine aminotransferase ratio; API, age/platelet index; CDS, Cirrhosis discriminant score; APRI, AST/platelet ratio index; FibroQ, fi bro-quotient; GUCI, Goteburg University Cirrhosis Index; AARP, AAR-platelet score; APGA, AST/platelet/GGT/AFP index; PAPAS, Platelet/Age/Phosphatase/AFP/AST index. * P � 0.05 vs. AUROC 0.5.

Scan

d J

Clin

Lab

Inv

est D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Lav

al o

n 06

/23/

14Fo

r pe

rson

al u

se o

nly.

Liver fi brosis in chronic hepatitis B 305

directly linked to the extent of hepatic fi brosis [27]. To date, liver biopsy is considered the most reliable method to evaluate the presence of advanced fi brosis and cirrhosis [28]. Evaluation by a liver biopsy is recommended for HBeAg-negative CHB patients with an HBV-DNA viral load of greater than 2000 IU/mL and mildly elevated ALT levels [20]. How-ever, liver biopsy is an invasive procedure [11]. More-over, heterogeneous distribution of fi brosis often means that a liver biopsy may not truly be represen-tative of the whole liver [29], and convincing a patient to undergo a repeat biopsy is not always possible [27]. All these factors create a demand/need for accurate and reliable non-invasive methods for the evaluation of hepatic fi brosis.

Fibrillar ECM accumulation associated with ECM degradation and remodelling is a dynamic pro-cess that eventually results in liver fi brosis [10]. Indi-cators of fi brosis may be classifi ed as direct markers refl ecting ECM metabolism, such as hyaluronic acid, amino-terminal propeptide of type III collagen, type I and type IV collagens, laminin, matrix metallopro-teinases, tissue inhibitors of metalloproteinases, transforming growth factor- β , and tumor necrosis factor- α , and indirect markers depicting changes in liver function, including several indices such as the AAR, CDS, API and ARPI as well as the Pohl score [20,30]. Although direct markers have not yet become part of routine practice, these markers recently have attracted attention and researchers showed that they might be useful [31 – 33]. Although the diagnostic value of indirect markers as predictors of fi brosis has been demonstrated in numerous studies on patients with chronic hepatitis C (CHC), data on CHB is limited. CHB and CHC differ greatly in terms of virological characteristics, histological changes in the liver and mechanisms that trigger fi brosis [20]. It is for this reason that parameters used in CHC patients need to be validated for patients with CHB.

In our study, we managed to demonstrate that the AAR, API, CDS and AARP indices as well as Pohl score and Lok ’ s model were inadequate as indicators of liver fi brosis in patients with CHB, and that the FIB-4 index is the most useful marker in this regard. Furthermore, the Forns, GUCI, APRI (equivalent to GUCI), PAPAS, APGA and FibroQ indices all showed diagnostic adequacy as predictors of signifi -cant fi brosis. Among all the indices evaluated, AUROC analysis showed FIB-4 to have the best diagnostic accuracy, with a cut-off value of 1.02 hav-ing a sensitivity of 77.9% and a specifi city of 58.2%. Respective PPV and NPV for a cut-off value of 2.01 were 60.71% and 73.58% (Table IV).

Varying results, particularly calculated values for AUROC, have been reported in different studies evaluating the diagnostic adequacy of indirect mark-ers of fi brosis (Table V). For example, following AUROC analysis, the cut-off value we determined for APRI was 0.670, compared to a value of 0.86 in a

study by Shin et al. in which APRI was reported to have the best diagnostic adequacy among the param-eters studied. In the original article where APRI was fi rst proposed as a marker of fi brosis, cut-off values of 0.80 (training set) and 0.88 (validation set) were calculated [16]. Values reported in subsequent studies range between 0.541 and 0.83 [11,22,24,26,34 – 36]. It has been postulated that the diagnostic adequacy of APRI as a marker of liver fi brosis in CHB patients is lower than that for patients with CHC [35].

Sterling et al. [18] originally reported on an AUROC value of 0.737 for FIB-4, and in subsequent studies, values of 0.785 [34], 0.764 [11] and 0.723 [22] have been calculated. In our study, the FIB-4 index was proven to be the most successful test for the detection of signifi cant fi brosis.

The reported AUROC value for the Forns index is 0.795 [11], while in our study this index was the second best test in terms of diagnostic adequacy with an AUROC value of 0.680.

In our study group, the GUCI index was the third most valuable test for the detection of signifi cant fi brosis, with an AUROC value of 0.670. Islam et al. [24], who were the fi rst to describe this marker, reported an AUROC value of 0.85. However, in a study with a similar design, investigators concluded that the GUCI index may not be very useful as a predictor of signifi cant fi brosis [37].

The AUROC values for the FibroQ index which was fi rst developed by Hsieh et al. were 0.783 [21] and 0.789 [34], whereas our calculated value was 0.588. In a previous study by Cheung et al. [36], the reported AUROC value for the Pohl score was 0.5, demonstrating its inadequacy as a marker of fi brosis.

These contrasting results may be attributed to differences in the study populations investigated. A study group with a high proportion of patients with advanced fi brosis would result in a deceivingly high positive predictive value. Moreover, the implementa-tion of certain exclusion criteria may especially affect indices which rely on AST concentration and platelet count [30].

Earlier studies evaluating the predictive value of indirect indices of hepatic fi brosis generally involved patients with CHC. Recently, two new indices have been developed to help determine the extent of hepatic fi brosis in CHB patients, namely the APGA [5] and PAPAS [22] indices. Although experts agree that both indices need validation before being applied in patients with CHB [5,22], we did not encounter any such pub-lication in the literature. Fung et al. [5] reported on an AUROC value for APGA of 0.85. However, inves-tigators in this study based their diagnostic adequacy calculations on measurements of liver stiffness. In a study by Seto et al. [22], calculated AUROC values for the APGA and PAPAS indices were 0.757 and 0.776, respectively. In this study, approximately 80% of patients with an APGA index score of less than 8.37 did not have signifi cant fi brosis, whereas 92% of

Scan

d J

Clin

Lab

Inv

est D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Lav

al o

n 06

/23/

14Fo

r pe

rson

al u

se o

nly.

306 S. Erdogan et al.

mitochondria and decreased AST clearance [30]. Other parameters that showed a statistically signifi cant correlation in our study include age, PT/INR, ALT, GGT, TC and globulin, fi ndings which are consistent with previous studies [5,34].

In our study, we did not observe any difference between patients with mild and advanced fi brosis in terms of platelet count, or a signifi cant correlation between platelet count and fi brosis stage. Previous studies have demonstrated that decreased platelet counts in patients with advanced fi brosis may be due to reduced thrombopoietin production and increased platelet destruction [16,18,23,30,35]. Nevertheless, our fi ndings regarding platelet count are supported by some studies [9,30,38].

patients with signifi cant fi brosis had a PAPAS index score of more than 1.36. Based on our results, the PAPAS index was fi fth (AUROC � 0.639), and the APGA index was sixth (AUROC � 0.638) in terms of diagnostic adequacy ( P � 0.05).

In our study, we evaluated the correlation of param-eters such as age, gender, platelet count, PT/INR, AFP, AST, ALT, ALP, GGT, total bilirubin, TC, HBeAg, HBeAb, albumin, globulin, blood urea nitrogen, crea-tinine, WBC count, RBC count and Hb, as well as several fi brosis indices with the degree of hepatic fi bro-sis. Although low correlation coeffi cient values were estimated via this evaluation, we observed that AST showed the best positive correlation with hepatic fi bro-sis, probably due to increased AST secretion from

Table V. Published sensitivity, specifi city and predictive values of signifi cant fi brosis in CHC and CHB patients.

Study Patient Index Cut-offSensitivity

%Specifi city

%PPV %

NPV %

Hsieh et al. [34] CHC

AAR 0.4 96.4 7.1 47.8 69.2API 4 87.4 34.1 53.9 75.4CDS 5 92.8 15.1 49.0 70.4APRI 0.5 99.1 7.1 48.5 90.0FIB-4 1.45 87.4 34.9 54.2 75.9FibroQ 0.6 97.3 7.14 48.0 75.0Pohl score Positive 7.2 99.2 88.9 54.8

Bonacini et al. [15] CHC CDS 8 46 98 – –

Yilmaz et al. [26]CHC APRI 0.44 72.7 62.4 – –CHB APRI 0.36 55.0 75.4 – –

Lin et al. [35]CHB APRI 0.8 43.2 81.8 – –CHC 0.8 81.3 87.5 – –

Guzelbulut et al [11] CHCAPRI 0.5 84.34 44.78 65.42 69.77FIB-4 0.6 100 10.45 58.04 100Forns 4.2 93.98 34.33 63.93 82.14

Fung et al. [5] CHB APGA 6.9 82.4 68.5 49.1 91.3

Seto et al. [22] CHB

PAPAS 1.662 73.3 78.2 56.4 88.4APGA 6.687 16.9 98.1 81.3 71.0FIB-4 1.45 51.9 74.4 49.4 76.3APRI 0.5 89.6 40.6 42.1 89.0

Forns et al. [17] CHC Forns 6.9 30 95 66 80

Hsieh et al. [21] CHB and CHCAAR 0.54 77 63 91 36APRI 1.50 54 58 86 21FibroQ 1.6 79 71 93 41

Koda et al. [39] CHCAPRI 0.85 34.1 95.7 89.4 58.0Forns 8.7 24.3 96.6 88.2 54.9Fibroindex 2.25 35.8 97.4 94.3 59.1

Sterling et al. [18] HIV/HCV coinfection

FIB-4 3.25 23.0 96.6 64.5 82.6

Wai et al. [16] CHC APRI 0.50 91 47 61 86Lok et al. * * [23] CHC Lok ’ s model 0.9 98 6 41 83

Cheung et al. [36] CHC

AAR 1.0 21.4 81.9 42.1 62.8Pohl score Positive 9.1 97.7 70.8 63.5APRI 0.5 11.2 62.7 15.7 53.4Lok ’ s model 0.8 93.1 31 45.4 87.9

Islam et al. * * [24] CHC GUCI 1.0 80 78 31 97

Lee et al. [20] CHBCDS 4 34.8 87.5 73.8 57.0AARP Positive 39.3 85.2 72.9 58.1Lok ’ s model 0.87 50.6 85.2 77.6 63.0

Shin et al. [30] CHB APRI 1.4 79 83 84 77

* * for predicting cirrhosis. CHC, chronic hepatitis C; CHB, chronic hepatitis B; NPV, negative predictive value; PPV, positive predictive value, AAR, aspartate aminotransferase/alanine aminotransferase ratio; API, age/plate-let index; CDS, Cirrhosis discriminant score; APRI, AST/platelet ratio index; FibroQ, fi bro-quotient; GUCI, Goteburg University Cirrhosis Index; AARP, AAR-platelet score; APGA, AST/platelet/GGT/AFP index; PAPAS, Platelet/Age/Phosphatase/AFP/AST index. Sc

and

J C

lin L

ab I

nves

t Dow

nloa

ded

from

info

rmah

ealth

care

.com

by

Uni

vers

ity o

f L

aval

on

06/2

3/14

For

pers

onal

use

onl

y.

Liver fi brosis in chronic hepatitis B 307

Lavanchy D . Hepatitis B virus epidemiology, disease burden, [3] treatment, and current and emerging prevention and control measures . J Viral Hepat 2004 ; 11 : 97 – 107 . Lok AS , McMahon BJ . Chronic hepatitis B . Hepatology [4] 2007 ; 45 : 507 – 39 . Fung J , Lai CL , Fong DY , Yuen JC , Wong DK , Yuen MF . [5] Correlation of liver biochemistry with liver stiffness in chronic hepatitis B and development of a predictive model for liver fi brosis . Liver Int 2008 ; 28 : 1408 – 16 . Dufour DR , Lott JA , Nolte FS , Gretch DR , Koff RS , Seeff [6] LB . Diagnosis and monitoring of hepatic injury . II. Recom-mendations for use of laboratory tests in screening, diagnosis, and monitoring. Clin Chem 2000 ; 46 : 2050 – 68 . McMahon BJ . The natural history of chronic hepatitis B [7] virus infection . Hepatology 2009 ; 49 : S45 – 55 . Rotman Y , Brown TA , Hoofnagle JH . Evaluation of the [8] patient with hepatitis B . Hepatology 2009 ; 49 : S22 – 7 . Zeng MD , Lu LG , Mao YM , Qiu DK , Li JQ , Wan MB , Chen [9] CW , Wang JY , Cai X , Gao CF , Zhou XQ . Prediction of sig-nifi cant fi brosis in HBeAg-positive patients with chronic hepatitis B by a noninvasive model . Hepatology 2005 ; 42 : 1437 – 45 . Pinzani M , Rombouts K , Colagrande S . Fibrosis in chronic [10] liver diseases: diagnosis and management . J Hepatol 2005 ; 42 : S22 – 36 . Guzelbulut F , Cetinkaya ZA , Sezikli M , Yasar B , Ozkara S , [11] Ovunc AO . AST-platelet ratio index, Forns index and FIB-4 in the prediction of signifi cant fi brosis and cirrhosis in patients with chronic hepatitis C . Turk J Gastroenterol 2011 ; 22 : 279 – 85 . Ishak K , Baptista A , Bianchi L , Callea F , De Groote J , Gudat [12] F , Denk H , Desmet V , Korb G , MacSween RNM , Philips MJ , Portmann BG , Poulsen H , Scheuer PJ , Schmid M , Thaler H . Histological grading and staging of chronic hepatitis . J Hepatol 1995 ; 22 : 696 – 9 . Giannini E , Botta F , Fasoli A , Ceppa P , Risso D , Lantieri [13] PB , Celle G , Testa R . Progressive liver functional impair-ment is associated with an increase in AST/ALT ratio . Dig Dis Sci 1999 ; 44 : 1249 – 53 . Poynard T , Bedossa P . Age and platelet count: a simple index [14] for predicting the presence of histological lesions in patients with antibodies to hepatitis C virus . METAVIR and CLINIVIR Cooperative Study Groups. J Viral Hepat 1997 ; 4 : 199 – 208 . Bonacini M , Hadi G , Govindarajan S , Lindsay KL . Utility [15] of a discriminant score for diagnosing advanced fi brosis or cirrhosis in patients with chronic hepatitis C virus infection . Am J Gastroenterol 1997 ; 92 : 1302 – 4 . Wai CT , Greenson JK , Fontana RJ , Kalbfl eisch JD , Marrero JA , [16] Conjeevaram HS , Lok AS . A simple noninvasive index can predict both signifi cant fi brosis and cirrhosis in patients with chronic hepatitis C . Hepatology 2003 ; 38 : 518 – 26 . Forns X , Ampurdanes S , Llovet JM , Aponte J , Quinto L , [17] Martinez-Bauer E , Bruguera M , Sanchez-Tapias JM , Rodes J . Identifi cation of chronic hepatitis C patients without hepatic fi brosis by a simple predictive model . Hepatology 2002 ; 36 : 986 – 92 . Sterling RK , Lissen E , Clumeck N , Sola R , Correa MC , [18] Montaner J , M SS , Torriani FJ , Dieterich DT , Thomas DL , Messinger D , Nelson M . Development of a simple noninvasive index to predict signifi cant fi brosis in patients with HIV/HCV coinfection . Hepatology 2006 ; 43 : 1317 – 25 . Pohl A , Behling C , Oliver D , Kilani M , Monson P , Hassanein [19] T . Serum aminotransferase levels and platelet counts as predictors of degree of fi brosis in chronic hepatitis C virus infection . Am J Gastroenterol 2001 ; 96 : 3142 – 6 . Lee IC , Chan CC , Huang YH , Huo TI , Chu CJ , Lai CR , [20] Lee PC , Su CW , Hung HH , Wu JC , Lin HC , Lee SD . Com-parative analysis of noninvasive models to predict early liver fi brosis in hepatitis B e Antigen-negative Chronic Hepatitis B . J Clin Gastroenterol 2011 ; 45 : 278 – 85 .

Our study has some limitations, the most impor-tant being the majority of cases having mild hepatic fi brosis with a limited number of patients having advanced fi brosis (approximately one-third). Another limitation was the retrospective nature of the study which allowed for determination of the diagnostic adequacy of most of the parameters studied in all 221 patients, with the exception of the Forns, APGA and PAPAS indices which could only be calculated in 139, 180 and 179 patients, respectively, again only one-third of which had signifi cant fi brosis. Evalua-tions of the parameters GGT, ALP and TC also could not be made in all patients, again due to miss-ing data ( n � 216, n � 215 and n � 140, respectively). One other limitation was the fact that HBeAg-posi-tive patients only compromised 15.4% of the study population.

There is an urgent need for more reliable meth-ods for evaluating progression of fi brosis in patients with CLD to help elucidate the mechanisms behind fi brogenesis and therefore help in the development of effective treatment strategies [10]. A liver biopsy only provides static information regarding fi brosis stage at the time the biopsy is performed [10]. More-over, histopathological fi ndings may vary depending on whether the liver biopsy is performed from the left or right lobe of the liver [34]. Additionally, spec-imens obtained from cirrhotic patients are frequently fragmented [34] which may lead to patients being diagnosed as having a lower than expected fi brosis stage [23].

Our results suggested that out of the 13 indices evaluated; only FIB-4 index could show diagnostic adequacy. Instead of indirect fi brosis indices, direct markers refl ecting ECM metabolism, such as hyaluronic acid and amino-terminal propeptide of type III collagen, and imaging modalities for the evaluation and follow-up of histological changes in patients with CHB need further investigation. There is a need for more comprehensive prospective studies to help determine the diagnostic value of non-invasive tests for liver fi brosis.

Declaration of interest : The authors report no confl ict of interest. The authors alone are responsible for the content and writing of the article.

References

Perz JF , Armstrong GL , Farrington LA , Hutin YJ , Bell BP . [1] The contributions of hepatitis B virus and hepatitis C virus infections to cirrhosis and primary liver cancer worldwide . J Hepatol 2006 ; 45 : 529 – 38 . Zani C , Pasquale L , Bressanelli M , Puoti M , Paris B , [2] Coccaglio R , Lascioli I , Pieriacci G , Donato F . The epide-miological pattern of chronic liver diseases in a community undergoing voluntary screening for hepatitis B and C . Dig Liver Dis 2011 ; 43 : 653 – 8 .

Scan

d J

Clin

Lab

Inv

est D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Lav

al o

n 06

/23/

14Fo

r pe

rson

al u

se o

nly.

308 S. Erdogan et al.

Hsieh YY , Tung SY , Lee IL , Lee K , Shen CH , Wei KL , [21] Chang TS , Chuang CS , Wu CS , Lin YH . FibroQ: an easy and useful noninvasive test for predicting liver fi brosis in patients with chronic viral hepatitis . Chang Gung Med J 2009 ; 32 : 614 – 22 . Seto WK , Lee CF , Lai CL , Ip PP , Fong DY , Fung J , Wong [22] DK , Yuen MF . A new model using routinely available clini-cal parameters to predict signifi cant liver fi brosis in chronic hepatitis B . PLoS One 2011 ; 6 : e23077 . Lok AS , Ghany MG , Goodman ZD , Wright EC , Everson [23] GT , Sterling RK , Everhart JE , Lindsay KL , Bonkovsky HL , Di Bisceglie AM , Lee WM , Morgan TR , Dienstag JL , Morishima C . Predicting cirrhosis in patients with hepatitis C based on standard laboratory tests: results of the HALT-C cohort . Hepatology 2005 ; 42 : 282 – 92 . Islam S , Antonsson L , Westin J , Lagging M . Cirrhosis [24] in hepatitis C virus-infected patients can be excluded using an index of standard biochemical serum markers . Scand J Gastroenterol 2005 ; 40 : 867 – 72 . Hanley JA , McNeil BJ . A method of comparing the areas [25] under receiver operating characteristic curves derived from the same cases . Radiology 1983 ; 148 : 839 – 43 . Yilmaz Y , Yonal O , Kurt R , Bayrak M , Aktas B , Ozdogan O . [26] Noninvasive assessment of liver fi brosis with the aspartate transaminase to platelet ratio index (APRI): usefulness in patients with chronic liver disease: APRI in chronic liver disease . Hepat Mon 2011 ; 11 : 103 – 6 . Salles N , Dussarat P , Foucher J , Villars S , de Ledinghen V . [27] Non-invasive evaluation of liver fi brosis by transient elastog-raphy and biochemical markers in elderly inpatients . Gas-troenterol Clin Biol 2009 ; 33 : 126 – 32 . Kopec KL , Burns D . Nonalcoholic fatty liver disease: a [28] review of the spectrum of disease, diagnosis, and therapy . Nutr Clin Pract 2011 ; 26 : 565 – 76 . Bedossa P , Dargere D , Paradis V . Sampling variability of [29] liver fi brosis in chronic hepatitis C . Hepatology 2003 ; 38 : 1449 – 57 . Shin WG , Park SH , Jang MK , Hahn TH , Kim JB , Lee MS , [30] Kim DJ , Jun SY , Park CK . Aspartate aminotransferase to

platelet ratio index (APRI) can predict liver fi brosis in chronic hepatitis B . Dig Liver Dis 2008 ; 40 : 267 – 74 . El-Mezayen HA , Toson el SA , Shiha GE . Role of hyaluronic [31] acid, its degrading enzymes, degradation products, and fer-ritin in the assessment of fi brosis stage in Egyptian patients with chronic hepatitis C . Eur J Gastroenterol Hepatol 2013 ; 25 : 69 – 76 . Kim BK , Kim HS , Park JY , Kim do Y , Ahn SH , Chon CY , [32] Park YN , Han KH , Kim SU . Prospective validation of ELF test in comparison with Fibroscan and FibroTest to predict liver fi brosis in Asian subjects with chronic hepatitis B . PLoS One 2012 ; 7 : e41964 . Li F , Zhu CL , Zhang H , Huang H , Wei Q , Zhu X , Cheng [33] XY . Role of hyaluronic acid and laminin as serum markers for predicting signifi cant fi brosis in patients with chronic hepatitis B . Braz J Infect Dis 2012 ; 16 : 9 – 14 . Hsieh YY , Tung SY , Lee K , Wu CS , Wei KL , Shen CH , Chang [34] TS , Lin YH . Routine blood tests to predict liver fi brosis in chronic hepatitis C . World J Gastroenterol 2012 ; 18 : 746 – 53 . Lin CS , Chang CS , Yang SS , Yeh HZ , Lin CW . Retrospective [35] evaluation of serum markers APRI and AST/ALT for assessing liver fi brosis and cirrhosis in chronic hepatitis B and C patients with hepatocellular carcinoma . Intern Med 2008 ; 47 : 569 – 75 . Cheung RC , Currie S , Shen H , Bini EJ , Ho SB , Anand BS , Hu [36] KQ , Wright TL , Morgan TR . Can we predict the degree of fi brosis in chronic hepatitis C patients using routine blood tests in our daily practice? J Clin Gastroenterol 2008 ; 42 : 827 – 34 . Usluer G , Erben N , Aykin N , Dagli O , Aydogdu O , Barut [37] S , Cevik F , Ormen B . Comparison of non-invasive fi brosis markers and classical liver biopsy in chronic hepatitis C . Eur J Clin Microbiol Infect Dis 2012 ; 31 : 1873–1878 . Yang HR , Kim HR , Kim MJ , Ko JS , Seo JK . Noninvasive [38] parameters and hepatic fi brosis scores in children with non-alcoholic fatty liver disease . World J Gastroenterol 2012 ; 18 : 1525 – 30 . Koda M , Matunaga Y , Kawakami M , Kishimoto Y , Suou T , [39] Murawaki Y . FibroIndex, a practical index for predicting signifi cant fi brosis in patients with chronic hepatitis C . Hepatology 2007 ; 45 : 297 – 306 .

Scan

d J

Clin

Lab

Inv

est D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Lav

al o

n 06

/23/

14Fo

r pe

rson

al u

se o

nly.