serum-markers-of-liver-fibrosis by huanghengdong


									    B iomarkers                                                                                                                           As published in CLI December 2005

                                                            It is important to diagnose the presence of liver fibrosis and its severity for optimal patient manage-

           of the month                                     ment. While liver biopsy is the standard method of diagnosis, the procedure is invasive and subject
                                                            to diagnostic limitations. This article discusses the current role and limitations of direct and indirect
                                                            serum markers of liver fibrosis and cirrhosis, and considers the improvements which are necessary
                                                            before assays for such markers can be considered as a diagnostic alternative to liver biopsy.

                     Serum markers of
           liver fibrosis
           by Dr I. N. Guha and Prof. J. P. Iredale

Fibrosis is part of the innate wound healing response which occurs in injured
tissues and represents a final, common pathway of chronic liver disease.
Within the liver, fibrosis is characterised by the deposition of extracellular
matrix. Current evidence indicates that net deposition of matrix is the result
of a balance between synthesis and degradation and constitutes a dynamic
process. The evolution of fibrosis to cirrhosis is dependent on a variety of
mechanisms including iterative injury, inflammation, angiogenesis, architec-
tural disturbance and oxidative stress. The clinical condition of cirrhosis is
associated with decreased hepatic function, aberrant haemodynamics and a
propensity for malignant transformation.

There are a number of reasons why clinicians need to diagnose the presence
and severity of fibrosis. Firstly, it will influence if and when treatment should
be commenced, eg in chronic hepatitis C. Secondly, it has been proven that the
degree of fibrosis will directly determine future prognosis [1, 2]. Thirdly, there
is evidence from animal models and clinical studies that fibrosis may be at
least in part reversible [3, 4]. This has led to considerable interest in the
research and development of antifibrotic medication. The efficacy of this
treatment can only be determined by the ability to monitor the progression or
regression of fibrosis closely. Finally, cirrhosis is often asymptomatic but diag-
nosing this entity is vital in planning the timing for transplantation, screening
for complications, eg hepatocellular cell carcinoma (HCC), and gathering
accurate epidemiological data to allocate health care resources appropriately.

Hitherto the liver biopsy has represented the "gold standard" for diagnosing
liver fibrosis and cirrhosis. The biopsy has the advantages of providing more
information than simply the amount of fibrosis present in the liver. However,
in the assessment of fibrosis alone it has limitations. Histological analysis is
subject to sampling error, unsurprising when one considers that on average
1/50 ,000th of the organ is sampled. Liver biopsies taken at laparoscopy ( i.e.
under direct vision) show a 30% discordance in staging of fibrosis when sam-
ples are taken form right and left lobes in the context of hepatitis C (HCV) - a
disease that is believed to have a homogenous distribution in the liver [5].
Furthermore, the absolute value of liver biopsy is limited by intra-observer and
inter-observer variability and limitations of using categorical variables to
describe a continuous process. When these deficiencies are taken together with
the fact that liver biopsy is invasive, and that although it is a relatively safe pro-
cedure it is still associated with pain in 20% of patients and severe morbidity
in 0.57% [6], it becomes clear why there has been an intensive search to find
accurate non-invasive markers of liver fibrosis. Other developments of effec-
tive non-invasive markers of liver fibrosis would markedly facilitate the devel-
opment of antifibrotic therapies.
                                                                                         Table 1. Examples of direct serum markers of liver fibrosis.
Serum markers of liver fibrosis
Serum markers can be broadly divided into direct and indirect measures of                Direct serum markers
liver fibrosis. The direct markers measure constituents of extracellular matrix          Table 1 shows examples of direct markers used as surrogates of liver fibrosis.
(ECM) or their breakdown products; the indirect markers measure products                 In addition to hyaluronic acid (HA) and the N-termial peptide of type III pro-
that increase or decrease as a result of fibrosis or cirrhosis.                          collagen ( PIIINP), other examples of direct markers include YKL-40, laminin,
                                                                                                                                     As published in CLI December 2005
 B iomarkers
                                                                                      have diagnostic excellence at the extreme thresholds which encompass only 30-
                                                                                      40% of the study population [Parkes and Guha et al, submitted].

                                                                                      Limitations of serum markers
                                                                                      Some of the limitations of serum markers have been discussed above but there
                                                                                      are common reasons why both direct and indirect markers may not be achiev-
                                                                                      ing the required diagnostic excellence. It is likely that ECM products such as
                                                                                      HA are influenced by not only the production within the liver but also extrac-
                                                                                      tion by the liver. Moreover, as fibrosis represents a balance of fibrogenesis and
                                                                                      fibrolysis, the simple measurement of one aspect of this equation may not be
                                                                                      sufficient for a robust diagnostic tool. Indirect markers may rely on parame-
                                                                                      ters of functionality such as platelet count, but these may only be altered at
                                                                                      critical levels of fibrosis rather than sequential levels of fibrosis i.e. there is a
                                                                                      threshold effect. Panel markers are heavily dependent on the characteristics of
                                                                                      the derivation population which may explain the discrepancy when they are
                                                                                      tested on external cohorts.

                                                                                      One of the biggest challenges facing serum markers is however the limitations
                                                                                      of the reference standard to which they are being compared, i.e. the liver biop-
                                                                                      sy. Any inaccuracies due to the liver biopsy (see above) will compound the
                                                                                      deficiencies of the serum markers. As histology is a surrogate for clinical out-
                                                                                      come measures, it has been suggested that serum markers should be compared
                                                                                      to hard clinical outcomes such as death, bleeding and HCC. However in liver
                                                                                      disease, many of the measures take a considerable length of time to evolve
                                                                                      making these trials more difficult to undertake. Investigators have therefore
                                                                                      tried to reduce the limitations of biopsy (eg taking bigger biopsies, having one
                                                                                      central pathologist score all the biopsies etc) to improve the performance of
                                                                                      putative surrogate markers.

                                                                                      Clinical utilisation of serum markers
                                                                                      The present use of serum markers in clinical practice will depend on the dis-
                                                                                      ease and outcome measure. For example, the diagnosis of cirrhosis by HA in
                                                                                      alcoholic liver disease has an AUROC of 0.93 [10] and can be considered a
                                                                                      useful test. In more general terms, serum markers may be useful in addition to
                                                                                      biopsy rather than replacing biopsy. If patients have scores above or below
                                                                                      extreme thresholds, scores below the lower threshold suggesting no/mild
                                                                                      fibrosis and scores above the higher threshold suggesting moderate/severe
                                                                                      fibrosis, a biopsy could be potentially avoided. As the majority of the popu-
                                                                                      lation tested presently lie between these extreme thresholds, liver biopsy will
Table 2. Example of panel markers used to assess fibrosis in chronic hepatitis C.     be required in this "indeterminate group", however the total number of biop-
                                                                                      sies performed could be potentially reduced.
undulin, tenascin and collagen IV. The sensitivity and specificity is dependent
on the disease tested eg HA in general is more accurate in differentiating fibro-     Serum markers may also have a role in population studies rather than indi-
sis in alcoholic liver disease (ETOH) compared to hepatitis C (HCV).                  viduals. For example, if a potential antifibrotic improved the fibrosis scores
Furthermore, the direct markers have greater accuracy in diagnosing cirrhosis         of the cohort as a whole but not in individuals, this may demonstrate a cer-
but only have moderate accuracy in separating mild from severe disease.               tain degree of efficacy.
Moreover, they are unable to distinguish more detailed subdivisions of fibrosis.
Indirect markers                                                                      It is likely that the accuracy of serum markers will be improved by using them
The majority of literature related to serum markers has been performed on             in combination with powerful technologies such as proteomics or imaging
subjects with chronic hepatitis C. A detailed systematic review by Gebo et al         modalities such as elastography or microbubble ultrasound. To be considered
[15] suggested that panel marker tests may offer the most promise in assessing        a serious alternative to liver biopsy, they will need to be improved in accura-
fibrosis non-invasively. Panel tests consist of markers which have usually been       cy, have the ability to differentiate subtler forms of fibrosis and maintain
extrapolated by multiple logistic regression in a derivation population (training     accuracy when measured in different populations.
set) and then applied to separate a population (validation set). Table 2 shows
examples of published panel markers for assessing fibrosis in chronic hepatitis
C. The data are presented as the area under the receiving operator characteris-       References
tic curve (AUROC). This is generated by plotting the sensitivity against 1-speci-     1. Yano M et al. The long-term pathological evolution of chronic hepatitis C.
ficity at different thresholds. A value of 0.5 represents a result that could be      Hepatology 1996; 23: 1334-1340.
explained by random effects alone and 1 represents a perfect diagnostic test.         2. Harrison SA et al. The natural history of nonalcoholic fatty liver disease: a clinical
The AUROC of the panel tests in Table 2 range from 0.74 to 0.9 but in general         histopathological study. Am J Gastroenterol 2003; 98: 2042-2047.
the AUROC is lower in the validation sets compared to the training sets, espe-        3. Abdel-Aziz G et al. Reversibility of hepatic fibrosis in experimentally induced
cially if external validation is performed (e.g. external validation of the           cholestasis in rat. Am J Pathol 1990; 137: 1333-1342.
Fibrotest by Rossi produced a figure of 0.74 compared to 0.84 in the original         4. Dufour JF et al. Regression of hepatic fibrosis in hepatitis C with long-term inter-
derivation/training cohort). Similar to the direct markers, the panel tests are       feron treatment. Dig Dis Sci 1998; 43: 2573-2576.
superior in distinguishing cirrhosis but are not very effective in the diagnosis of   5. Regev A et al. Sampling error and intraobserver variation in liver biopsy in
subtler forms of fibrotic injury. Furthermore, when separating no fibrosis/mild       patients with chronic HCV infection. Am J Gastroenterol 2002; 97: 2614-2618.
fibrosis from moderate/severe fibrosis (i.e. F0/F1 vs F2/F3/F4) the panels only       6. Cadranel JF et al. Practices of liver biopsy in France: results of a prospective
                                                                                                                                      As published in CLI December 2005
 B iomarkers
nationwide survey. For the Group of Epidemiology of       14. Nojgaard C et al. Serum levels of YKL-40 and PII-        23. Thabut D et al. Noninvasive prediction of fibrosis
the French Association for the Study of the Liver         INP as prognostic markers in patients with alcoholic         in patients with chronic hepatitis C. Hepatology 2003;
(AFEF). Hepatology 2000; 32: 477-481.                     liver disease. J Hepatol 2003; 39: 179-186.                  37: 1220-1221.
7. Guechot J et al. Diagnostic accuracy of hyaluronan     15. Gebo KA et al. Role of liver biopsy in management        24. Leroy V et al. Circulating matrix metalloproteinas-
and type III procollagen amino-terminal peptide           of chronic hepatitis C: a systematic review. Hepatology      es 1, 2, 9 and their inhibitors TIMP-1 and TIMP-2 as
serum assays as markers of liver fibrosis in chronic      2002; 36: S161-S172.                                         serum markers of liver fibrosis in patients with chron-
viral hepatitis C evaluated by ROC curve analysis. Clin   16. Imbert-Bismut F et al. Biochemical markers of liver      ic hepatitis C: comparison with PIIINP and hyaluron-
Chem 1996; 42: 558-563.                                   fibrosis in patients with hepatitis C virus infection: a     ic acid. Am J Gastroenterol 2004; 99: 271-279.
8. Wong VS et al. Serum hyaluronic acid is a useful       prospective study. Lancet 2001; 357: 1069-1075.              25. Rosenberg WM et al. Serum markers detect the
marker of liver fibrosis in chronic hepatitis C virus     17. Rossi E et al. Validation of the FibroTest biochemi-     presence of liver fibrosis: a cohort study.
infection. J Viral Hepat 1998; 5: 187-192.                cal markers score in assessing liver fibrosis in hepatitis   Gastroenterology 2004; 127: 1704-1713.
9. Boeker KH et al. Diagnostic potential of circulating   C patients. Clin Chem 2003; 49: 450-454.
TIMP-1 and MMP-2 as markers of liver fibrosis in          18. Poynard T et al. Biochemical surrogate markers of        The authors
patients with chronic hepatitis C. Clin Chim Acta         liver fibrosis and activity in a randomized trial of         Dr I. N. Guha
2002; 316: 71-81.                                         peginterferon alfa-2b and ribavirin. Hepatology 2003;        Professor J.P. Iredale
10. Naveau S et al. Biomarkers for the prediction of      38: 481-492.                                                 University of Southampton Liver Unit
liver fibrosis in patients with chronic alcoholic liver   19. Wai CT et al. A simple noninvasive index can pre-        Mail Point 811
disease. Clin Gastroenterol Hepatol 2005; 3: 167-174.     dict both significant fibrosis and cirrhosis in patients     Level D
11. Oberti F et al. Noninvasive diagnosis of hepatic      with chronic hepatitis C. Hepatology 2003; 38: 518-          South Academic block
fibrosis or cirrhosis. Gastroenterology 1997; 113:        526.                                                         Southampton General Hospital
1609-1616.                                                20. Le Calvez S et al. The predictive value of Fibrotest     Trenoma Road
12. Saitou Y et al. Noninvasive estimation of liver       vs. APRI for the diagnosis of fibrosis in chronic hepa-      Southampton SO16 6YD
fibrosis and response to interferon therapy by a serum    titis C. Hepatology 2004; 39: 862-863.                       UK.
fibrogenesis marker, YKL-40, in patients with HCV-        21. Sud A et al. Improved prediction of fibrosis in
associated liver disease. World J Gastroenterol 2005;     chronic hepatitis C using measures of insulin resist-
11: 476-481.                                              ance in a probability index. Hepatology 2004; 39:
13. Murawaki Y et al. Diagnostic value of serum mark-     1239-1247.
ers of connective tissue turnover for predicting histo-   22. Forns X et al. Identification of chronic hepatitis C
logical staging and grading in patients with chronic      patients without hepatic fibrosis by a simple predictive
hepatitis C. J Gastroenterol 2001; 36: 399-406.           model. Hepatology 2002; 36: 986-992.

To top