Evaluation of Liver Fibrosis and Activity in Chronic Hepatitis C Significance of Biochemical Markers

Transcription

1 Thesis for the degree of PhD in Pathology Evaluation of Liver Fibrosis and Activity in Chronic Hepatitis C Significance of Biochemical Markers Shoukat Ali Arain Ziauddin University Karachi Pakistan 2010

2 SUPERVISOR S CERTIFICATE This is to certify that the research work presented in this thesis has been carried out under my supervision. This thesis has been prepared according to the guidelines set out by Ziauddin University. I have specifically checked the: 1. Introduction 2. Methodology 3. Results 4. Statistical analysis 5. Discussion and conclusion 6. Bibliography 7. Grammar, punctuation and spelling Prof. Dr. Qamar Jamal M.B.B.S., M. Phil., Ph. D. Professor of Pathology Ziauddin University

3 CERTIFICATE THIS THESIS IS SUBMITTED BY DR. SHOUKAT ALI ARAIN IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY IN PATHOLOGY GRANTED BY THE ZIAUDDIN UNIVERSITY, KARACHI, PAKISTAN. EXAMINATION COMMITTEE CHAIRPERSON: Prof. Dr. Qamar Jamal Date approved Karachi

4 Dedicated to Late Dr. Sarwar Jehan Zuberi in the memory of her enthusiastic guidance and kind support

5 Table of Contents S. NO. DESCRIPTION PAGE A. List of abbreviations (a) B. List of tables (c) C. List of figures (e) D. Acknowledgements (g) 1. INTRODUCTION Clinical context Biology of the virus HCV associated disease Assessment of liver histology Potential biomarkers Selection of markers for this study Purpose of study METHODS Study design Liver biopsy Biochemical markers Statistical analysis RESULTS Baseline characteristics of participants Univriate analysis Evaluation of single markers Formulation and assessment of indexes DISCUSSION SUMMARY AND CONCLUSION BIBLIOGRAPHY APPENDICIES i-xiii

6 List of Abbreviations A2M α-2 Macroglobulin ALT Alanine Transaminase ALP Alkaline Phosphatase Apo A-1 Apolipoprotein A-1 APRI AST to Platelet Ratio Index AST Aspartate Transaminase AUROC Area Under Receiver Operating Characteristic CHB Chronic Hepatitis B CHC Chronic Hepatitis C CLD Chronic Liver Disease EB Elution Buffer ECM Extracellular Matrix ELISA Enzyme Linked Immunosorbent Assay GGT Gamma-Glutamyl Transferase HA Hyaluronic Acid HABP Hyaluronic Acid Binding Protein HAI Histological Activity Index HBV Hepatitis B Virus HBsAg Hepatitis B Surface Antigen HCC Hepatocellular Carcinoma HCV Hepatitis C Virus HIV Human Immunodeficiency Virus HPLC High Performance Liquid Chromatography HYP Hydroxyproline TFCC International Federation of Clinical Chemistry a

7 LN Lobular Necrosis MMP Metalloproteinase mrna Messenger RNA NPV Negative Predictive Value PIIINP N-terminal propeptide of type III collagen PCR Polymerase Chain Reaction PITC Phenyl Isothiocyanate PMN Piecemeal Necrosis PPV Positive Predictive Value RNA Ribonucleic Acid ROC Receiver Operating Characteristic TGF Transforming Growth Factor TIMP Tissue Inhibitor of Metalloproteinase TNF Tumor Necrosis Factor trna Transfer RNA ULN Upper Limit of Normal UTR Untranslated Region b

8 List of Tables TABLE NO. DESCRIPTION PAGE 1.1. Comparison of commonly used fibrosis staging systems Characteristics of an ideal non-invasive marker of liver disease Some indirect marker assays for assessment of liver histology Direct markers of liver fibrosis Indexes with combined direct and indirect markers Distribution of patients in different grades and stages Comparison of histopathological characteristics among minimal and significant fibrosis stage categories Biochemical features of all patients Univariate analysis of demographic and laboratory features for minimal and significant fibrosis Univariate analysis of demographic and laboratory features for minimal and significant necroinflammatory grade categories Univariate analysis of demographic and laboratory features for overall minimal and significant disease Univariate analysis of demographic and laboratory features for mild and advanced/widespread fibrosis 60 c

9 3.8. Diagnostic value of serum ALT for different histopathological categories Diagnostic value of ALT at different cut-off levels for overall significant disease Diagnostic value of serum hyaluronic acid for different histopathological categories Diagnostic value of hyaluronic acid at different cut-off levels for overall significant disease Diagnostic performance of indexes for histopathological categories in terms of AUROC Scores of 6 marker index with their coordinate Liverscores Diagnostic performance of Liverscore at different cutoff points for overall disease Diagnostic performance of different non-invasive indexes in terms of AUROC 78 d

10 List of Figures FIGURE NO. DESCRIPTION PAGE 1.1. Schematic diagram showing life cycle of HCV Natural history of HCV infection Inflammatory cycle in HCV infection Algorithm for METAVIR activity grading Distribution of the patients in age groups Distribution of the patients in grades on liver biopsy Distribution of the patients in stages on liver biopsy Portal inflammation and foci of interface hepatitis (H&E x200) Severe portal inflammations with diffuse interface hepatitis (H&E x 200) Several foci of lobular inflammation (H&E x 400) Lobular necro-inflammation and grade 2 steatosis (H&E x200) Fibrosis limited to portal tract (Masson s trichrome x 200) Fibrosis stage 2 A fibrous septum with portal inflammation and severe steatosis (Masson s trichrome x 200) A case of F3 fibrosis This liver biopsy showed numerous septa without cirrhosis (Masson s trichrome x 200) 52 e

11 3.11. Cirrhosis Numerous fibrous septa with nodule formation (Masson s trichrome x 100) Chromatogram of standard material Chromatogram of a sample ROC curve showing the sensitivity and specificity of ALT for the prediction of overall disease on biopsy ROC curve showing the sensitivity and specificity of HA for the prediction of overall liver disease on biopsy ROC curve showing the sensitivity and specificity of Liverscore for the prediction of overall disease on biopsy 71 f

12 Acknowledgements I offer my thanks and gratitude to Almighty Allah, who guides us in darkness and helps us in difficulties. All respect for His Last Prophet (PBUH). I am grateful to my supervisor Professor Dr. Qamar Jamal for her continued encouragement and guidance throughout my studies. She worked really hard during all that period especially in histopathological evaluation of the liver biopsies and manuscript writing. My special thanks are to Late Dr. Sarwar Jehan Zuberi. She was the main inspiration for the initiation of this research project. She guided me a lot in synopsis writing and retrieval of literature during earlier part of my project. I am obliged to Dr. Waqaruddin Ahmed (PMRC, Karachi) and Dr. Nasir Laique (ZH, North Nazimabad Campus) for helping me in specimen collection. I am grateful to Dr. Adnan Zubairi and Dr. Arif Hussain, Chemical Pathologists (ZH, North Nazimabad Campus), and to Mr. Sohail Shaukat, Senior Scientific Officer (PCSIR), for helping me in chemical analysis. I am also thankful to Drs. Aamir Omair and M. Athar Khan of CHS department for helping me in statistical analysis. My special thanks are to Dr. Sheikh Yaeesh for his continued encouragement and critical assessment of my work. I am indebted to all senior faculty members, colleagues and friends in Clifton and North Nazimabad campuses. I acknowledge with gratitude the advice, invaluable support and encouraging critique afforded to me by Dr. NA Jafarey, Dr. HR Ahmad, Dr. Serajuddaula Syed, Dr. Anila Jaleel and Dr. Nasir Ali Afsar. I am also thankful to the entire library and research laboratory staff. In particular, Syed Ahmed Naqvi, the chief librarian, for his advice, encouragement and support and Ms. Shazia for her help in retrieval of literature. I offer sincere thanks to Mr. Shakaib and Moazzam Ali Shahid for providing me technical support in the research laboratory, and data entry. Finally, I am grateful to Pakistan Medical Research Council, Islamabad and Ziauddin University, Karachi, for providing me financial support to complete this project. g

13 1. INTRODUCTION 1

14 2 1.1 Clinical Context Around million individuals have chronic hepatitis C (CHC) viral infection worldwide and 3 to 4 million new infections occur each year (Sy and Jamal, 2006). A characteristic feature of hepatitis C viral (HCV) infection is its propensity to evolve into a chronic infection. Chronic infection with HCV induces injury and inflammation of the liver, which appears to be responsible for the associated fibrogenesis. Morbidity and mortality associated with CHC results mainly from the development of liver fibrosis and hepatocellular carcinoma (Seeff et al., 2001). The rate of fibrosis progression varies markedly from person to person and may not be linear over time. In many patients it may not progress to end stage liver disease and its complications. Therefore, assessment of the fibrosis stage and rapidity of progression of fibrosis (necro-inflammation) may help in determining the prognosis and the need of therapy in an individual patient (Marcellin et al., 2002). Liver biopsy, the gold standard for assessment of fibrosis and necroinflammatory activity, is an invasive procedure associated with complications and is not always possible (Poynard et al., 2000). Alternative strategies, like imaging and non-invasive biochemical monitoring of liver disease are being actively evaluated. Some of the biochemical markers, especially panels of multiple markers in the form of indexes are promising and may reduce the number of liver biopsies for assessment of liver disease (Reiss and Keeffe, 2005).

15 3 1.2 Biology of the virus Structure Hepatitis C virus, first detected in 1989 as a new viral agent of non-a, non-b blood-born hepatitis (Choo et al., 1989), and was later identified as a major cause of blood borne hepatitis worldwide (Lauer and Walker, 2001). It belongs to flaviviridae family and is member of the Hepacivirus genus. It is a small enveloped, single-stranded RNA virus with a genome of approximately 9.6 kb, flanked by 5 and 3 untranslated regions (UTR). It has one large open reading frame that translates into a single polyprotein of approximately 3000 amino acids. The polyprotein is subsequently cleaved into four structural and six nonstructural proteins. A highly conserved nucleocapsid protein is coded on the 5 end, followed by envelop proteins E1 and E2, and a protein p7. Six less conserved nonstructural proteins are NS2, NS3, NS4A, NS4B, NS5A and NS5B (Moradpour and Blum, 2004). Due to inherent low fidelity of viral RNA polymerase, HCV genome replication leads to genetic variability resulting in multiple genotypes (1 6), subgenotypes and quasispecies. Even in a given patient HCV may circulate as a divergent genomic population (Alexopoulou et al., 2005; Pawlotsky, 2006). Transmission HCV is primarily transmitted through exposure to contaminated blood and blood products. Blood transfusions and therapeutic use of blood products were the major source of infection until In socio-economically developed countries modes of transmission have changed. In United States, more than

16 4 50% transmissions occur through illegal drug use, by sharing syringes or intranasal use of cocaine (Alter et al., 1999; Armstrong et al., 2006). In lesser developed countries, inoculations during medical procedures like blood transfusions, injection needle reuse and hemodialysis are the major source of infection (Prati et al., 1997; Frank et al., 2000; Qureshi et al., 2009). Sexual and perinatal transmission is low. In cross sectional studies, 1-3% sexual partners of HCV positive persons are found to be infected. Co-infection with other sexually transmitted infections, duration of the relationship and chronic liver disease are the possible independent cofactors for risk of transmission (Rooney et al., 1998). The risk of transmission through needle stick injuries is about six times higher in HCV than HIV and perinatal transmission occurs in 6% of births to infected mothers, as compared to 20% - 60% in hepatitis B (Kim et al., 2000). Life cycle Steps involved in the life cycle of HCV include its attachment and entry into the host cell, uncoating of the viral genome, translation of viral proteins and replication within the cell, assembly and release of the virus particles (Figure 1.1).

17 5 Figure1.1 Schematic diagram showing life cycle of HCV (reproduced from Tellinghuisen et al., 2007).

18 6 Envelope glycolproteins E1 and E2 are essential for viral entry into the cell. These bind to CD81 receptors on host cell surface (Drummer et al., 2005). CD81 is expressed on many cells and may not explain the liver tropism of HCV. It is thought to modulate entry (Grove et al., 2007). Additional factors identified to facilitate entry include the human scavenger receptors class B type 1 (SR-B1) and the tight-junction proteins claudin 1 and occludin. In the liver cells, claudin 1 is expressed at high levels and might explain liver tropism of HCV (Evans et al., 2007; Benedicto et al., 2009). Entry is thought to occur through endosomal route followed by acid-triggered fusion (Lavillette et al., 2006). Once inside the host cell, viral genome uncoats and through its internal ribosomal entry site (IRES) binds to host machinery for translation. Viral polyprotein formed during this process is cleaved to form different proteins including proteins required to establish viral replication machinery. Viral RNA replication begins along the intracellular membranes. The details of HCV viral replication are not completely understood, it is suggested that viral genome serves as a template for negative strand RNA and two combine to form a double stranded RNA. This is used for semi conservative replication to generate multiple copies of progeny, positive strand RNA genomes (Targett- Adams et al., 2008). Little is known about the later stages of cell cycle; genome assembly and secretion. With the development of tractable HCV cell culture systems the kinetics of assembly and secretion has started to unleash. Secreted HCV has a

19 7 low density relating its release with lipoproteins which might protect the virion from host immune systems (Gastaminza et al., 2008). 1.3 HCV associated disease Disease Burden Around million individuals have chronic hepatitis C viral (HCV) infection worldwide and 3 to 4 million new infections occur each year and is a common cause of end stage liver disease and hepatocellular carcinoma (HCC). It is the most important cause of liver transplantation in the United States (Kim et al., 2002). In Pakistan around 10 million people are living with HCV infection, with a prevalence of about 5% in the general population (PMRC National Survey, 2008; Waheed et al., 2009). Prevalence is even higher for certain high risk groups like health care workers and those suffering from thalassaemia (Aziz et al., 2002; Hussain et al., 2008). CHC is found to be one of the major cause of chronic liver disease (CLD) associated morbidity (Bukhtiari et al., 2003), and mortality (Khokhar and Niazi, 2003). Course of the disease Incubation period for hepatitis C varies from 2 to 26 weeks. Acute hepatitis C infection, refers to first six months, remains asymptomatic in about 85% of the patients and goes undetected. Minority of the patients experience mild, nonspecific symptoms. Only rarely symptoms approach to those of acute hepatitis A and B. A characteristic feature of acute hepatitis C is to evolve into chronic infection. The actual chronicity rate is not well established. There have been a

20 8 few large scale prospective studies of unselected patients with acute infection to document the actual frequency of chronicity (Figure 1.2). Retrospective analyses of post-transfusion hepatitis cases suggest that 75% of the patients develop chronic disease (Alter and Seeff, 2000). Progression to persistent infection seems to be multifactorial and a result of complicated host viral interactions. It is estimated that 10% - 20% of CHC patient progress to cirrhosis in 20 years. In cirrhotic patients, the 5-year probability of clinical decompensation is 18% and in these decompensated patients 5-year survival rate falls to 50%. Once cirrhosis establishes, the 5-year risk for the development of HCC is 7% (Fattovich et al., 1997; Chen and Morgan, 2006). Figure 1.2 Natural history of HCV infection (reproduced from Chen and Morgan, 2006).

21 9 Pathogenesis of Cell Injury and Fibrosis During chronic infection, persistent intrahepatic HCV replication establishes an inflammatory cycle through secretion of chemokines. Recruitment of inflammatory cells with secretion of cytokines and generation of free radicals cause cell injury leading to fibrogenesis (Figure 1.3). Increased free radical mediated oxidative stress with decreased anti-oxidative capacity has been observed in CHC (Venturini et al., 2010). Figure 1.3 Inflammatory cycle in HCV infection. In 20% patients, vigorous inflammatory response resolves infection. In chronic hepatitis, intrahepatic persistence of the virus stimulates secretion of chemokines. Recruited inflammatory cells in turn produce cytokines. Inflammatory cycle causes liver cell injury that culminates in fibrosis (Reproduced from Zeremski et al., 2007).

22 10 Fibrogenesis is a dynamic process characterized by the synthesis of constituents of the extra cellular matrix (ECM), which is a complex mixture of glycoproteins and proteoglycans organized in a three-dimensional network (Schuppan et al., 2001). It is a non-specific mechanism, which lasts as long as injury persists in the liver and is believed to help limit the extension of the inflammatory reaction. In all forms of chronic hepatitis, active fibrosis begins around the portal areas (periportal or zone 1 fibrosis) and gradually extends out into the lobules towards the central veins (zone 3), with septa formation and then bridging fibrosis. The final stage of fibrosis constitutes cirrhosis with extensive fibrosis linking portal and central areas and nodular regeneration of liver parenchyma (Goodman and Ishak, 1995). Collagen and matrix proteins of fibrosis are largely produced by activated hepatic stellate cells. The stellate cells are activated from quiescent lipocyte phenotype to fibroblastic phenotype. The activation occurs in two phases: initially activation of stellate cells by cytokines, chemokines and other signaling molecules induced by the inflammatory process followed by transformation into myofibroblastic phenotype. These myofibroblasts, in turn, can proliferate, attract leucocytes and produce extracellular collagen and matrix proteins. The steps in stellate cell activation have been demonstrated to occur in chronic hepatitis C (Eng and Friedman, 2000). The major fibrogenic cytokine, transforming growth factor (TGF)-β, and mrna levels for this cytokine are found to be increased in the liver in CHC. TGF-β is also increased in the serum. Although the role of HCV core protein remains poorly

23 11 understood, it may contribute to fibrogenesis via up regulation of connective tissue growth factor (CTGF) and TGF-β1 (Shin et al., 2005). During fibrogenesis a basal membrane appears separating the hepatocytes from sinusoidal blood and perturbing the exchange of nutrients between blood and hepatocytes, a process known as capillarization of sinusoid. In addition to quantitative increase, during fibrogenesis there is a qualitative change, transformation from normal ECM into a reticulated and fibrillar type dense matrix (Paradis et al., 2002). In CHC, the rate of fibrogenesis has shown no association with viral factors like genotype and viral load. Host factors associated with rapid fibrosis include excessive consumption of alcohol, male gender, obesity, age more than 40 years at the time of infection, diabetes, persistent elevation of serum ALT, degree of necro-inflammation and fibrosis on liver biopsy and co-infection with HBV and HIV (NIH Consensus Statement, 2002; Ryder et al., 2006; Chen and Morgan, 2006). 1.4 Assessment of liver histology Rationale As mentioned earlier, morbidity and mortality of CHC is associated with the development of cirrhosis and its complications. Natural history studies have revealed that fibrosis progression varies tremendously between patients and may vary overtime. The risk of developing cirrhosis varies from 10% to 20% over a period of 20 years (Chen and Morgan, 2006). Therefore, the future course of the disease is difficult to predict in an individual.

24 12 Treatment of CHC is complex, costly and associated with side effects that are difficult to accept especially in a predominantly asymptomatic population. Furthermore, about half of the patients with genotype 1 and slightly lesser than that in other genotypes, fail to respond to anti-viral therapy (Fried et al., 2002; Idrees and Riazuddin, 2009). Of the several prognostic variables, inflammatory grade and fibrosis stage appear to be most reliable. Treatment decisions are recommended to be individualized on the basis of severity of the liver disease, treatment response rates, co-morbid conditions and the readiness of the patient for treatment. Treatment is recommended for the patients with bridging fibrosis and compensated cirrhosis, if liver histology is known (Ghany et al., 2009). Assessment of fibrosis and fibrosis progression can thus be valuable in CHC for the following reasons. In majority of the patients fibrosis progression is so slow that only 25% will progress to cirrhosis. Assessment helps in estimation of the time required to develop cirrhosis and need of therapy in an individual patient (Friedman, 2003). Stage of fibrosis indicates the likelihood of response to antiviral therapy; advanced stages are generally associated with lower response rates (Kau et al., 2008). The long term effect of anti-viral and antifibrotic therapies on liver morphology can only be monitored through assessment of fibrosis (Friedman, 2003).

25 13 Liver biopsy Although, liver biopsy is thought to be the gold standard for assessment of liver histology, its diagnostic value is limited by substantial sampling error and inter-observer subjectivity. Furthermore, it is associated with potential morbidity and mortality, and has several other limitations. Average size of a liver biopsy represents 1/50,000 th the size of entire liver. Although, fibrosis is thought to be a diffuse process, it may be variable. Regev et al. (2002) carried out a study on 124 HCV patients to assess the discordance between two liver biopsies, one from the left and other from the right lobe. Only non-fragmented biopsies larger than 15 mm and with at least five portal spaces, figures which eliminate half of liver biopsies done in daily practice, were used. Despite these precautions, there was a 33% discordance of at least one stage and 24% discordance for the necro-inflammatory activity grade. Studies involving expert histopathologists showed an inter-observer agreement in fibrosis staging ranging from 70% to 90% and an intra-observer agreement of 60% to 90%. The levels of concordance for necro-inflammatory grade was even lower (Bedossa et al., 1994; Westin et al., 1999). Several semi-quantitative histological systems have been proposed to minimize inter-observer variability and are commonly applied for scoring liver lesions in hepatitis B and C. The commonly used systems score for two separate items; fibrosis stage and necro-inflammatory grade. The Ishak s Histological Activity Index (HAI) is a modification of the Knodell HAI (Ishak, 1994). It classifies fibrosis from stage 0 6, permitting physicians to assess the effect of therapy

26 14 in better way (Table 1.1). METAVIR system is simpler and has been used for scoring large number of biopsies especially in hepatitis C associated studies. It scores fibrosis from stage 0 4 and has shown a better interobserver concordance (Bedossa et al., 1994). Table 1.1 Comparison of commonly used fibrosis staging systems Fibrosis METAVIR Ishak No fibrosis 0 0 Fibrosis in some portal tracts 1 1 Fibrosis in most portal tracts 1 2 Portal fibrosis with rare septa 2 3 Portal fibrosis with numerous septa 3 4 Marked bridging without cirrhosis 3 5 Cirrhosis 4 6

27 15 For grading necroinfalmmatory activity Ishak s modified HAI scores portal inflammation (0 4), interface hepatitis (0 4), focal lobular necrosis and inflammation (0 4) and confluent necrosis (0 6). Grading score is obtained by adding scores of all the categories. Thus maximum possible score for grading is 18. METAVIR grading score for chronic hepatitis C (A0 A3) is obtained through an algorithm (Fig. 2.1). This algorithm includes only lobular necrosis (0 2) and interface hepatitis (0 3). Portal inflammation is recorded as a prerequisite for the diagnosis of chronic hepatitis but is not included in grading algorithm ((Bedossa and Poynard, 1996). These histological scoring systems are discontinuous and semi-quantitative. This limitation is significant in clinical practice since many infected patients have intermediate stages of disease. Moreover, a liver biopsy only provides static data, not dynamic findings reflecting the ongoing pathogenetic mechanisms. Cost of liver biopsy and expertise required in the face of large number of patients needing assessment are the other concerns. Liver biopsy is not risk free and can be experienced by some patients as aggressive procedure, becoming a potential obstacle to the effective management of hepatitis C. After biopsy 30% patients feel pain, 0.3% have severe complications and 0.03% die (Poynard et al., 2000). Serum markers With the growing sense about the limitations of liver biopsy, efforts have quickened to develop non-invasive markers of liver fibrosis and necro-

28 16 inflammation. Non-invasive markers of liver disease are needed for several reasons: Although biopsy remains the gold standard for the assessment of liver histology, it is associated with complications and has several limitations. Out of millions of patients infected with HCV, only a minority is likely to develop significant fibrosis or cirrhosis, with prevalence peak between the years 2010 and Thus, increasing number of patients will require assessment. This will expose these patients to the potential risks, inconvenience and cost of liver biopsy (Friedman, 2003). Many studies have shown that fibrosis is a dynamic and reversible process (Sarmento-Castro et al., 2007). Currently the greatest hurdle in the testing and development of anti-fibrotic therapy is thought to be a lack of a robust and convenient fibrosis marker. With the development of anti-fibrotic therapies, there will be a need for regular and more frequent monitoring. The need for such frequent monitoring will exceed what is possible with liver biopsy (Friedman, 2003). The human development index published by United Nations ranks Pakistan 141 st in 182 countries (Human Development Report, 2009). Prevalence of Anti HCV antibodies in Pakistani population is 4.9% (PMRC National Survey, 2008). It seems unlikely, either to assess liver histology through biopsy or to offer treatment, to all the patients. There

29 17 is a need to identify patient categories to rationalize the need for therapy. Patients showing minimal fibrosis and slow progression of the fibrosis over a long period of time may not require urgent therapy; they may defer treatment until safer, better tolerated and more cost effective therapy is available. Pakistan Society of Gastroenterology recommends treatment for the patients having serum ALT levels above 1.5 to 2 times the upper limit of normal (ULN) for six month or more (Hamid et al., 2004). Serum ALT level is the most commonly investigated marker to evaluate its predictive value for the assessment of liver histology. Although, in most of the studies it was found to be associated with fibrosis stage, it has only modest value in predicting fibrosis stage on liver biopsy in an individual. Use of ALT only may misclassify many patients (Gebo et al., 2002). 1.5 Potential biomarkers Noninvasive means of assessing fibrosis include novel imaging techniques and serum markers. There are several features required for an ideal serum biochemical marker (Table 1.2).

30 18 Table 1.2 Characteristics of an ideal non-invasive marker of liver disease Specific Simple and inexpensive Reliable and reproducible Independent of metabolic alterations and drugs Sensitive enough to discriminate different disease categories Informative for disease progression and related to outcome Applicable to different causes of chronic liver disease In the last decade several biochemical markers have been proposed and evaluated. These markers can generally be classified into two categories: markers related to alterations in hepatic function and those related to extracellular matrix (ECM). As the markers related to alterations in hepatic function do not directly reflect ECM metabolism, these are known as indirect markers. On the other hand, markers related to ECM molecules and modifying enzymes and cytokines are directly linked to modifications in ECM. Thus, these are labeled as direct markers. The indirect markers include laboratory tests that are part of routine evaluation of CHC and some specific proteins related to hepatic function. Although, these markers are indirect expressions of liver injury and fibrosis, they are simple, easy to perform, widely available and show a statistical association with

31 19 hepatic damage. Routine laboratory tests evaluated in different assays are: albumin, serum bilirubin, alanine transaminase (ALT), aspartate transaminase (AST), alkaline phosphatase (ALP), γ-glutamyl transferase (GGT), prothrombin time, serum cholesterol and platelet count. These tests have been proposed alone and in different combinations as markers of injury and fibrosis. Specific proteins like haptoglobin, α-2 macroglobulin (A2M), apolipoprotein A-1 (Apo A-1) and γ-globulin have also been included in different assays combinations (Table 1.3). ALT is the most commonly investigated marker. AST to ALT ratio is one of the first non-invasive markers reported. AST to platelet ratio index (APRI) is another simple, cheap and easily available ratio. Results of several validation studies suggest that these routine tests may be used in combined indices, but generally there predictive value is suboptimal (Fontana and Lok, 2003; Lackner et al., 2005; Sebastiani et al., 2008). Current consensus seems that these simple markers of fibrosis render liver biopsy unnecessary in only a minority of patients. Meta-analysis of these studies showed a modest value of these ratios in predicting fibrosis. Efforts to assay several markers from the same patient promise a greater likelihood of success in discriminating minimal from severe fibrosis and in predicting cirrhosis (Gebo et al., 2002).

32 20 Table 1.3 Some indirect marker assays for assessment of liver histology Study Assay Components (Poynard et al., 1997) AP index Age, platelet count (Imperiale et al., 2000) AAR ALT/AST (Pohl et al., 2001) Pohl score AAR, platelet count (Imbert-Bismut et al., 2001) Fibrotest Age, sex, Bilirubin, GGT, haptoglobin, A2M, apo A-1 (Imbert-Bismut et al., 2001) Actitest Fibrotest components & ALT (Forns et al., 2002) Forns index Age, GGT, cholesterol, Platelets (Wai et al., 2003) APRI AST to platelet ratio index (Lok et al., 2005) Lok index Platelets, ALT/AST ratio, INR (Koda et al., 2007) Fibroindex AST, platelets, γ-globulin (Vallet-Pichard et al., 2007) Fib-4 Age, AST, ALT, platelets

33 21 Fibrotest (Table 1.3) is calculated from five biochemical markers adjusted for age and gender of the patient. An extension of this test for the assessment of inflammatory activity is Actitest which includes ALT in addition to Fibrotest components. The MULTIVIRC group evaluated the biochemical markers included in these indexes for their ability to identify patients with clinically significant fibrosis. The authors concluded that use of these indexes may reduce the number of hepatic biopsies by at least 50% (Imbert-Bismut et al., 2001). Some studies from Pakistan also have evaluated the predictive value of these markers. A study on 120 treatment naïve patients evaluated APRI and found an acceptable diagnostic accuracy of 48% for significant fibrosis and 66% for advanced fibrosis (Khan et al., 2008). A combination of AST to ALT ratio and platelets count evaluated for diagnosis of advanced fibrosis showed a sensitivity of 85.6 % and specificity of 90 % with positive predictive value (PPV) of 91.2% and negative predictive value (NPV) of 83.4 % (Khokhar, 2003). Direct markers of liver fibrosis are various serum or urinary substances including cytokines, components of ECM and molecules related to ECM metabolism (Table 1.4). These markers have a patho-physiological rationale because they are related to either deposition or removal of ECM and thus may provide information on its metabolism. Since ECM metabolism not only reflects ECM deposition and removal, but also the remodeling of established ECM, serum levels of these markers may reflect both the activity of the

34 22 process and amount of the ECM undergoing remodeling. Thus they have a potential to be informative not only for stage of fibrosis but also for rate of its progression with a more relevant prognostic value (Afdhal and Nunes, 2004). This is supported by the observations that levels are increased in conditions with rapid progressive fibrosis, levels fall after treatment of the underlying disease prior to any reduction in amount of fibrosis, and levels of some these markers correlate with stage of fibrosis and total matrix content of the liver (Leroy et al., 2001; Parés et al., 1996). There are limitations of these direct markers. Because fibrosis is a non-specific process, current serum assays of matrix molecules or their fragments are not liver specific, may reflect impaired hepatic clearance, and often underestimate quiescent cirrhosis. Many assays fail to detect liver disease until late stages and may not discriminate between different fibrosis stages. Another limitation of these direct markers is that they are not routinely available in hospital settings for clinical use. Among these markers, serum hyaluronic acid (HA) and N- terminal pro-peptide of type III collagen (PIIINP) levels have been most extensively evaluated in patients with chronic hepatitis C. One study of 486 CHC patients found that a serum HA level of less than 60 µg/l had 99% NPV for cirrhosis on liver histology, but PPV was only 30% (McHutchison et al., 2000). Serum levels of PIIINP correlate better with inflammation and are inferior to serum HA levels in predicting hepatic fibrosis (Guechot et al., 1996).

35 23 Table 1.4 Direct markers of liver fibrosis Glycoproteins Hyaluronic acid Laminin YKL-40 (Human cartilage glycoprotein 39) Collagens and their components Procollagen I Procollagen III Procollagen IV Type IV collagen Collagenases and their inhibitors Metalloproteinase (MMP) Tissue inhibitor of Metalloproteinase (TIMP) Cytokines TGF-β1 TNF-α

36 24 Hyaluronic acid and some other direct markers have been combined with indirect markers in different fibrosis indexes to increase the diagnostic accuracy (Table 1.5). Initial reports are quite encouraging for the diagnosis of significant fibrosis and excellent for the diagnosis of cirrhosis (Adams et al., 2005; Naveau et al., 2009). Large scale, independent studies are needed for their validation. Table 1.5 Indexes with combined direct and indirect markers Study Assay Components (Adams et al., 2005) Hepascore Bilirubin GGT, HA, A2M, age, gender (Rosenberg et al., 2004) ELF Age, HA, PIII NP, TIMP1 (Cales et al., 2005) Fibrometer Age, AST, platelets, HA, INR, A2M, urea (Patel et al., 2004) Fibrospect A2M, HA, TIMP1 Abbreviations: HA: Hyaluronic acid, ELF: European liver fibrosis study group, PIIINP: N-terminal propeptide of type III collagen

37 Selection of markers for this study In present study we selected indirect markers of fibrotest and actitest from MULTIVIRC group for evaluation of their ability to identify necroinflammatory grade and fibrosis stage. Along with these, direct makers of fibrosis; hyaluronic acid, hydroxyproline (HYP) and proline were also included. MULTIVIRC group index is comprehensive, including markers for both necro-inflammation and fibrosis. MULTIVIRC group used A2M, haptoglobin, apolipoprotein A1, total bilirubin and GGT, adjusted for age and gender of the patient, to predict fibrosis. Serum ALT is added for activity index along with these markers. Most of the other proposed markers have been evaluated either for fibrosis or inflammation. Both fibrosis and necro-inflammatory activity (fibrogenesis) are important for making treatment decisions and prognostication (Dienstag, 2002). To date, it is the most validated non-invasive fibrosis and activity marker in all etiologies of chronic liver disease like CHC, CHB, alcoholic liver disease, non-alcoholic fatty liver disease and viral co-infections (Sebastiani, 2009). Authors have got these indexes patented as Fibrotest and Actitest; they are commercially available in more than 50 countries and are being used in clinical practice for making treatment decisions. Their cost varies between 100 to 300 Euros in different countries and test computation is available online (Poynard et al., 2007). Evaluation of these markers in our population

38 26 may help developing a similar index for clinical use at a low cost or/and may justify the use of same Fibrotest and Actitest in our population. Hyaluronic acid: It is nonsulfated glycosaminoglycans and is major component of ECM. Among the direct markers of liver fibrosis, as mentioned earlier, HA been most extensively studied in CHC, and a few studies are also available in other etiologies of CLD. It increases in the liver during fibrogenesis and released in systemic circulation during remodeling. This marker has shown a good overall accuracy in the diagnosis of significant fibrosis. Recent indexes consisting of HA in combination with indirect markers have shown promising results (Adams et al., 2005). Also that it is cost effective and measurable by ELISA, available in routine clinical laboratories. Proline and hydroxyproline: Fibrosis is a dynamic and reversible process; even cirrhosis may regress after successful treatment of HCV (Shiratori et al., 2000; Poynard et al., 2002). The amount of collagen deposition depends upon the activation of stellate cells and net collagenase activity (Friedman, 2003). Collagen is 21% proline and hydroxyproline. Mature collagen type I contain about 1000 amino acids of which about 100 are proline and another 100 are hydroxyproline. Hydroxyproline content of the collagen type III is even higher. Hydroxyproline is present almost exclusively in collagens. These degradation amino acids; proline and hydroxyproline are released into the circulation during the process of remodeling and collagen degradation. Also that released hydroxyproline is

39 27 not reutilized since there is no trna for this amino acid (Murray and Keeley, 2003). 1.7 Purpose of study Purpose of this study was to evaluate the predictive value of selected biochemical markers separately and in combined indexes, for the diagnosis of different histological lesions in chronic hepatitis C.

40 2. METHODS 28

41 Study design It was a cross sectional study to determine the diagnostic accuracy of noninvasive biomarkers, conducted at Ziauddin University, Clifton campus, Karachi, from June 2005 to July The patients were selected from Ziauddin University Hospital North Nazimabad, Karachi and outpatient clinic of Pakistan Medical Research Council, Karachi. This study was approved by Ethics Review Committee, Ziauddin University. Patients Treatment naïve chronic hepatitis C patients, with a positive HCV PCR, 20 to 60 years of age, were included in the study. Patients with HBV co-infection (positive HBsAg) and diabetes mellitus were excluded. Patients with history of other chronic inflammatory conditions like rheumatoid arthritis, alcohol intake and blood disorders requiring frequent blood transfusions were also excluded from the study. A questionnaire was completed for every patient to document possible variables like demographic factors, and to rule out causes of exclusion. A written informed consent was taken from each patient. 2.2 Liver Biopsy A percutaneous liver biopsy was performed with gauge modified Menghini aspiration needle (Surecut TSK, Japan). Tissue was formalin fixed, paraffin-embedded and stained with standard hematoxylin and eosin (H & E)

42 30 staining procedure. Besides, slides were also stained with connective tissue stains: Masson s trichrome, Van Gieson s and reticulin for identification of collagen and reticular fibers. Sections were also stained with standard glycogen and iron stains to rule out other vauses of liver damage besides HCV infection. Liver biopsy specimens of only more than 10mm length and having not less than five portal tracts were included in the study (Bedossa and Poynard, 1996; Afdhal and Nunes, 2004). Biopsy specimens with evident pathology but with lesser number/without identification of the correct number of portal tracts were also included. Number of portal tracts was recorded in each biopsy specimen. Histological features of the liver biopsy specimens were analyzed according to METAVIR group scoring system (Bedossa et al., 1994; Bedossa and Poynard, 1996), blindly, without any knowledge about the clinical data or biochemical markers. (A) Every specimen was staged for fibrosis on a five-point scale; F0 = no fibrosis F1 = portal fibrosis without septa F2 = portal fibrosis with rare septa F3 = numerous septa without cirrhosis and F4 = cirrhosis. There was an element of subjectivity in classifying patients in different stages. In a biopsy having ten portal tracts, for example, in a biopsy having 10 portal tracts and one fibrous septum, it was very subjective to decide whether this was a case of rare fibrosis. We, however, considered any fibrosis beyond portal

43 31 tracts as significant (F2). Fibrosis was considered F3 when 50% or more portal tracts show fibrous septa extending beyond portal tracts. For statistical analysis, different stages of fibrosis were further lumped together into different clinically significant groups, henceforth; F0 or F1 = minimal fibrosis F2 to F4 = significant fibrosis F3 and F4 = widespread fibrosis (B) Necroinflammatory lesions were graded on the basis of focal lobular necro-inflammation and piecemeal necrosis. Although portal inflammation was also recorded but was not part of grading algorithm. (Bedossa and Poynard, 1996). Focal Lobular Necrosis: 0: Less than one necroinflammatory focus per lobule 1: At least one necroinflammatory focus per lobule 2: Several necroinflammtory foci per lobule or confluent or bridging necrosis Piecemeal necrosis: 0: Absent 1: Focal alteration of the periportal plate in some portal tracts 2: Diffuse in some portal tracts OR Focal lesion around all portal tracts 3: Diffuse in all portal tracts Portal inflammation: 0: Absent

44 32 1: Presence of mononuclear aggregates in some portal tracts 2: Mononuclear aggregates in all portal tracts 3: Large and dense mononuclear aggregates in all portal tracts Histological activity was graded on a four point scale on the basis of an algorithm (Fig. 2.1). Figure 2.1 Algorithm for METAVIR activity grading (Bedossa and Poynard, 1996). LN: 0 A0 PMN 0 LN: 1 A1 LN: 2 A2 PMN 1 LN: 0, 1 A1 LN: 2 A2 LN: 0, 1 PMN 2 LN: 2 A2 PMN 3 LN: 0, 1, 2 A3 Abbreviations: PMN: piecemeal necrosis/interface hepatitis, LN: lobular necrosis. A0: no histological activity, A1: mild activity, A2: moderate activity, A3: severe activity.

45 33 Histological activity grades were also lumped together into two clinically significant groups; A0 or A1 = minimal activity A2 or A3 = significant activity On the basis of fibrosis stage and necro-inflammatory grade patients were divided into two overall severity groups; Minimal disease = <F2 and <A2 Significant disease = F2 or A2 Steatosis was graded, on the basis number of hepatocytes showing steatosis, as follows (Brunt et al., 1999); Grade 0 = upto 2% Grade1 = 3 to 29% Grade 2 = 30 to 59% Grade 3 = 60% Effect of biopsy size on the assessment of liver disease was evaluated by categorizing them in two categories on the basis of number of portal tracts; < 11, and 11 or more portal tracts (Colloredo et al., 2003). 2.3 Biochemical markers On the day of biopsy, blood (10 ml) was collected by venipuncture in plain tube without any additive, for the following markers; 1. Alanine transaminase 2. Gamma-glutamyl transpeptidase

46 34 3. Alkaline phosphatase 4. Total bilirubin 5. Direct bilirubin 6. Apolipoprotein A-1 7. Haptoglobin 8. Alpha-2 macroglobulin 9. Hyaluronic acid 10. Hydroxyproline 11. Proline Serum was separated within two hours of specimen collection. Laboratory analysis for alanine transaminase (ALT), gamma-glutamyl transpeptidase (GGT), alkaline phosphatase (ALP) and bilirubin was performed on the same day. Serum was frozen and stored at C for apolipoprotein A-1 (Apo-A1), haptoglobin and alpha-2 macroglobulin (A2M), and amino acids of collagen degradation; hydroxyproline (HYP) and proline, and hyaluronic acid (HA). Assays of these markers were performed in batches. Assay Methods ALT was performed by IFCC standardized UV enzymatic method (Bergmeyer et al., 1985). Reagents from Roche diagnostics (Cat. No ) were used according to manufacturer s instructions. GGT was performed by enzymatic colorimetric method using reagents from Roche Diagnostics (Cat. No ).This method has been standardized against the original Szasz reagents (Persijn and van der Slik, 1976). ALP assay was performed by

47 35 colorimetric method (Bowers and McComb, 1975), using reagents from Roche Diagnostics (Cat No ). All enzymatic activities were measured at 37 C. Total and direct bilirubin assays were performed according to the method described by Jendrassik and Grof (Jendrassik, 1938), using reagents from Roche Diagnostics (Cat. No ). Alpha-2 macroglobulin and haptoglobin assays were performed by Immunoturbidimetric method using polyclonal rabbit anti-human antibodies (DakoCytomation Denmark Code. Nos. Q0102 & Q0330). In this method specific antibodies react with antigens in the sample to form antigen-antibody complexes which, after agglutination, are determined turbidimetrically (Dati et al., 1996). The Apo-A1 assay was also performed immune-turbidimetrically using reagents from Randox (Cat. No. LP2989). Hyaluronic Acid was measured by an enzyme-linked binding protein microplate assay (Corgenix Inc.USA). In this assay, hyaluronic acid binding protein (HABP) is used as capture molecule (Lindqvist et al., 1992). Serum was incubated in HABP coated microwells according to manufacturer s instructions. After washing, to remove the unbound serum molecules, HABP conjugated with horseradish peroxidase was added to the microwell to form complexes with bound HA. A chromogenic substrate was then added to develop a colored reaction. The intensity of the developed color was measured at 450 nm.

48 36 All the above mentioned assays, except HA, were performed on Hitachi 911. HA was performed on microplate reader ELISA system. All the assays were calibrated with relevant calibration materials. For internal quality control low and high controls and patients sera were used to determine within-run and between-day reproducibility. Results were accepted only when control values fell within ±2SD. Hydroxyproline and proline assays were carried out by high performance liquid chromatography (HPLC) method as described by Lange and Mályusz (Lange and Mályusz, 1994), with slight modifications. A HPLC unit LC-20AT was used with UV/VIZ photodiode array detector (SPD-M20A) and system controller, CBM-20A, all from Shimadzu Corporation Japan. A 6 mmid x 15 cm, C 18 (Shim-pack CLC-ODS) chromatography column was used. Briefly, 500 µl of serum was deproteinized in 5 ml of 90% aqueous ethanol. The mixture was shaken for 10 minutes and centrifuged at 3000g for 5 minutes. The removed supernatant was dried at 60 0 C under reduced pressure. The dried specimen was reconstituted, for derivatization, by adding 200 µl distilled water having 50 µl o-pthaldialdehyde (OPA) reagent (TCI, Japan Cat No. P0280). Mixture was vigorously shaken and allowed to stand for 5 minutes. To remove the primary amino acids this mixture was passed through a 400 mg apolar column (RP-18, Adsorbex Merck) after adding 800 ml of elution buffer (EB) consisting of sodium acetate (Bio Basic, Canada Cat No. SB1611) and acetonitrile (TEDIA USA. Cat No. AS ). First 100 µl of the elute was discarded. The remaining 900 µl of elute was collected for post

49 37 column derivatization with phenyl isothiocyanate (PITC) reagent (TCI, Japan Cat No ). After leaving the mixture for 20 minutes at ambient temperature, specimen was again evaporated to complete dryness, under reduced pressure. The dried specimen was reconstituted with 50 µl of doubly distilled water and 50 µl of PITC reagent. The mixture was shaken vigorously and allowed to stand for 20 minutes at room temperature followed by centrifugation at 3000g for five minutes. Supernatant was removed and 20 µl of it was injected into the HPLC system. EB was used as solvent A and acetonitrile mixed with doubly distilled water (60:40, by vol.) as solvent B. Flow rate was adjusted as 1.5 ml/minute with 95% solvent A. After a batch of 10 specimens the mobile phase was changed to solvent B for 5 minutes to clean the column. The column was then reequilibrated with solvent A for 10 minutes. The photodiode array detector was set to a wavelength of 254 nm. The retention time for HYP was 5-6 minutes and for proline minutes, depending upon the instrument pressure. Instrument was calibrated using HYP (TCI, Japan Cat No. H0296) and proline (TCI, Japan Cat No. P0481) standards. These standards were also used to prepare standard curves and evaluate the linearity of the method. Reproducibility and recovery were evaluated before analyzing patients specimens. Between-batch coefficient of variance (CV) for HYP analysis was 4.2% and for proline was found to be 4.0%. The recovery of HYP was 98% and of proline was 84%. The results of proline are reported after adjustment for low recovery (obtained value divided by 0.84).