Journal of Hepatology
Volume 38, Issue 3 , Pages 357-360, March 2003

Rating fibrosis progression in chronic liver diseases

Liver Group, University of Southampton, Southampton SO16 6YD, UK

See Article, pages 257–265

Article Outline

 

Much of the morbidity and almost all of the mortality due to chronic liver disease can be attributed to advanced liver fibrosis, i.e. cirrhosis. A wide range of hepatic insults are associated with the accumulation of new extra-cellular matrix frequently accompanied by infiltration of the liver by immunocytes. Progressive accumulation of matrix causes changes in the cellular morphology of all components of the hepatic acinus, disruption of their inter-cellular interactions, impairment of blood flow and a loss of functioning liver cell mass. Eventually this process gives rise to cirrhosis. It has become apparent that liver fibrosis results from a dynamic process in which the accumulation of scar tissue is accompanied by matrix degradation and remodelling [1] in a finely balanced process that can be driven towards cirrhosis or reversed to restore normal hepatic structure and function. It is now recognized that even cirrhotic changes can be remodelled in response to removal of an hepatic insult [2], [3], treatment of immune mediated liver diseases [4], [5], [6] or antiviral therapy [7], [8], [9]. Despite increasing knowledge of how fibrosis develops and regresses surprisingly little is know about the kinetics of this process, particularly in human disease.

The concept of the ‘rate of fibrosis progression’ has arisen from recognition of the variability in the outcome of chronic hepatitis C (CHC) [10]. Patients assessed for fibrosis at the same time after infection with HCV were found to have developed differing degrees of liver fibrosis. Poynard et al. suggested that the histological fibrosis score divided by the time taken to reach that stage of fibrosis could be used to calculate the rate of fibrosis progression. Once established as a concept it was possible to search for factors that might account for differences in outcome in CHC. Such studies required knowledge of the time of infection, a point that may be estimated from knowledge of the time of exposure to a risk factor such as blood products or injection paraphernalia.

This approach, based on observation at a single time-point and estimation of duration of infection, can provide a velocity of fibrosis progression and, a priori, assumes a linear rate of fibrosis progression. It cannot identify changes in the velocity of fibrosis progression (acceleration or deceleration) but is nevertheless useful. Adopting such an approach, Poynard and others have identified alcohol, male sex and older age as factors that appear to cause more rapid progression of fibrosis [10]. Identification of even moderate alcohol consumption as a modifiable adverse prognostic factor in hepatitis C has been of great benefit to patients and clinicians.

When applied to the analysis of the results of antiviral therapy trials, changes in the rate of fibrosis progression have been cited as evidence of the beneficial effects of interferon alpha and ribavirin and the superiority of pegylated interferon over standard interferon [9]. Thus, the concept of a fibrosis progression rate has had implications for risk stratification, prognostication, risk modification and therapy for hepatitis C.

Wider application of the concept of the rate of fibrosis progression would be of considerable value in extending understanding of basic patho-biology, the ability to prognosticate, and the evaluation of the impact of risk modification and therapeutic interventions.

In this edition of the Journal, Poynard et al. [11] have applied the concept of fibrosis progression to a range of chronic liver diseases. The authors have extracted cross-sectional data, collected at single time-points, concerning 4852 patients registered in 13 cohorts spanning eight different disease processes. The single metric of liver fibrosis used in this study was severity of fibrosis as detected in a liver biopsy. As different histological staging systems had been used to score the liver biopsies from patients with different diseases, adjustments were made to extrapolate the scores to a uniform system. A total of 504 representative biopsies were then re-scored by a central pathologist using a five point staging system analogous to the widely used METAVIR system. The level of agreement between the central pathologist's scores and the extrapolated or original scores was ‘substantial’ or better, only when biopsies were categorized as binary outcomes (F0-F1 vs. F2-F4, and no cirrhosis vs. cirrhosis). On the basis of this finding the subsequent analyses were based on patients with septal fibrosis (F2) or cirrhosis (F4). By restricting the analyses to patients reaching these end-points the data are based on 876 cases with septal fibrosis and 681 cirrhotics.

Having fixed the severity of fibrosis at the time of the biopsy as a defined outcome the next step required estimation of the duration of disease up to this point in order to determine the rate of fibrosis progression. Although the time of infection can be identified in some cases of viral hepatitis, in the vast majority of cases of chronic liver disease the duration of disease cannot be reliably known. The authors have addressed this problem by taking two approaches to estimating the duration of disease, both of which are problematic. One approach is to use the patient's age at the time of biopsy (life modelling), and the other is to estimate the length of time that the patient has been exposed to the insult causing the disease (exposure modelling). Life modelling can be used to determine the probability of fibrosis according to age and gender, irrespective of disease duration. Exposure modelling is better suited to comparisons of rates of fibrosis progression in different disorders. In common with many other studies for HCV infection, dates of transfusion or first episode of injecting-drug-use can be employed as reasonable surrogates for the date of infection. The authors have chosen to consider primary biliary cirrhosis (PBC) and hereditary haemochromatosis (HHC) as genetic disorders and so have used the date of birth as the start of exposure to the disease. This is inappropriate for PBC in which fibrosis develops in late middle life in the majority of cases, and in HHC, in which fibrosis only develops after the hepatic stores have become saturated with iron, usually in middle age in men and later in women. Employing date of birth as the onset of disease in these conditions will have the effect of greatly underestimating the rate of fibrosis progression.

Data concerning modifying factors including diagnosis, age, sex, and alcohol consumption were considered. These parameters can be used to identify factors that might influence the rate of fibrosis progression, their relative influence in different diseases and in the case of alcohol, the likely benefit of modifying exposure. In the analyses alcohol consumption was considered as a dichotomous outcome (< or ≥50 g alcohol per day). This level of alcohol consumption may be regarded as rather high and when counselling patients it would be useful to know whether lower levels of consumption produce adverse outcomes.

Comparison of the rates of progression of fibrosis in subjects with differing diseases may answer a number of fundamental questions. To what extent does the rate of fibrosis differ between diseases? Does fibrosis progress to cirrhosis in all patients given sufficient time? What are the host factors that determine the rate of fibrosis in different individuals with the same disease? The authors present the numbers of years taken to reach septal fibrosis or cirrhosis in different diseases in Table 2 [11]. The 50th and 90th percentiles are shown for both the life and exposure modelling. It is apparent that the diseases can be ranked with HCV-HIV co-infection causing the most rapid fibrosis and PBC causing the slowest.

Because of the extreme uncertainty surrounding the duration of exposure for PBC and haemochromatosis and the relatively small number of cases, the estimates of the rate of fibrosis progression for these conditions must be viewed with caution. When these caveats are considered it is apparent that fibrosis progression is similar in HBV and HCV, more rapid in alcohol and fastest in HCV-HIV co-infection. Fibrosis progression curves for each disease are generated from the cumulative cross-sectional data and are presented in Figs. 1 and 2, broken down by sex [11]. The gradients of the curves are similar. It must be recognized that the curves are the product of modelling from cross-sectional data and do not reflect true rates of change in real patients.

Sub-group analyses in specific diseases revealed some interesting findings. Modelling suggests that patients with HCV-HIV co-infection will die from liver cirrhosis rather than HIV infection in the era of highly active anti-retroviral therapy, predicted by studies of co-infected haemophiliacs [12].

Earlier studies by the same group [10] have failed to demonstrate a relationship between hepatic inflammation and fibrosis in viral hepatitis. In the present work the authors report that patients with moderate or severe necro-inflammatory activity had more rapid fibrosis progression than those with no or minimal activity when infected with HCV-HIV, HCV alone or HBV.

Among those patients with detectable HBV DNA, patients with HBeAg negative chronic hepatitis B progressed to fibrosis faster than patients who were HBeAg positive suggesting that the presence of pre-core mutants is associated with more aggressive disease, an observation consistent with the high prevalence of advanced disease at presentation in many patients carrying pre-core mutants [13].

Comparison of fibrosis rates can be used to shed light on the effect of additional pathologies on a primary disorder. Poynard et al. have analyzed the effect of heavy (≥50 g per day) alcohol consumption on a range of disorders and found that it increases the life-time risk of progression to cirrhosis in all cases. The relative risk varies between disorders from 6.5 in HCV infection to 1.7 in HBV. As the risks attributable to the primary pathology vary it would be interesting to know if the absolute risk attributable to alcohol varies depending on the primary pathology. Further work might focus on whether pathologies such as alcohol and non-alcoholic fatty liver disease are synergistic or merely additive to other primary chronic liver diseases.

A major reservation about Poynard et al.'s study must be the use of cross-sectional data to model longitudinal changes. Once the concept of a fibrosis progression rate has been established it follows that such a rate may be subject to acceleration or deceleration. At its simplest level the rate of progression may be assumed to be linear. Care must be taken not to confuse more rapid development of fibrosis in certain circumstances with accelerating fibrosis rates. Evidence emerging in hepatitis C suggests that the rate of progression of fibrosis may increase with the duration of infection and this may apply to all chronic liver diseases. Prospective data acquired over an appropriate period of time and more complex modelling will be required to address this issue but such studies will take a long time and may never be completed. In the meantime the approach used by Poynard et al. is a reasonable substitute that has precedence in studies of cancer survival [14].

Concerns over the research methodology detract from the comparison of fibrosis rates in different diseases. Of more interest are the analyses of the impact of host factors on fibrosis progression in individual diseases. Here the errors in estimating the duration of disease may be assumed to apply to all patients in any given disease category making comparisons within disease categories far more likely to be robust than comparisons between diseases. Older age was associated with faster fibrosis progression in almost all diseases. The authors erroneously suggest that these data imply acceleration in the fibrosis rate with time. This interpretation is likely to be correct but it cannot be made from modelling based on cross-sectional data that assumes linearity of fibrosis progression. Further prospective studies with repetitive biopsies are required.

The authors found that male gender adversely affects progression to cirrhosis in viral hepatitis but not in alcoholic liver disease when duration of exposure and age at onset of drinking are considered. The reasons behind these findings warrant further study. It may be that female sex hormones are generally antifibrotic but that the greater dose per kilogram of body weight of alcohol in females causes more damage, cancelling the protective effects of oestrogens. Alternatively there may be a behavioural explanation. Sex differences in fibrosis certainly warrant further research.

One of the more subtle consequences of a better understanding of the rate of fibrosis progression is the ability to identify points in the natural history of a disorder at which intervention is likely to carry the maximum benefit. Such information can be used to design a screening or detection strategy that can be used to detect disease at a point when prognosis is clear-cut, when morbidity is at a minimum and when intervention is likely to produce greatest benefit. Furthermore, better understanding of the factors that influence the rate of fibrosis progression would allow stratification of individuals for risk of end-stage disease. This approach could be used to allocate costly or toxic therapies to those at greatest risk or to determine the optimal time for an intervention.

Poynard et al. attempt to draw upon their findings to suggest ages at which patients might be screened for hepatic fibrosis. This may be over-ambitious due to the considerable uncertainty that must surround the estimates of duration of exposure and thus fibrosis progression rates in the diseases studied. It is an important and sufficient contribution to show that evidence of fibrosis should be sought in middle-aged men and women and that heavy alcohol consumption by patients with hepatic co-morbidity will accelerate the progression of fibrosis. No conclusions should be drawn about HHC or PBC on the basis of the exposure modelling in this study.

Attempts such as those of Poynard et al. to develop more accurate models of prognosis in chronic liver disease are of great importance. Further studies employing more appropriate methodology must be conducted in order to answer the remaining questions relating to basic patho-biology. Better understanding of prognosis would benefit patients and those responsible for resource allocation. Measurement of the rate of fibrosis progression can be used in the evaluation of interventions and to assess modification of risk factors.

Having made the case for better knowledge of the rate of fibrosis progression in chronic liver diseases how might such knowledge be derived?

The ideal research methodology that would generate this knowledge is prospective cohort studies. A group of subjects with liver disease of interest should be assembled early in the course of their disease at a defined point in the natural history of the condition and followed prospectively. Follow-up should be as complete as possible and for a sufficient period of time to allow the evolution of fibrosis to cirrhosis and end-stage complications. Information about predetermined outcome measures that reflect hepatic fibrosis should be recorded at regular and frequent intervals. The accuracy with which these outcome measures reflect hepatic fibrosis should be known and should be high.

Examination of this list of methodological requirements reveals why determining rates of fibrosis progression is so difficult in chronic liver diseases. Careful thought will be required to select appropriate time-points from which to calculate the duration of exposure to disease, particularly for immune and genetic liver diseases. The slow progression of fibrosis will require follow-up to be conducted over decades.

The most significant barrier to accurate studies of fibrosis progression is the need for an accurate measure of fibrosis. Any study of fibrosis progression is dependent on a reliable and reproducible measure of fibrosis.

There are numerous reasons why histological examination of a liver biopsy is a sub-optimal way to measure liver fibrosis. Obtaining a liver biopsy is painful for the patient, hazardous, time consuming and costly. Sampling error is a distinct problem particularly in some disorders. Interpretation of biopsies using fibrosis stage scoring systems is problematic. The stages are assigned ordinal scores but these numbers are ciphers for qualitative descriptors of histology rather than quantitative measures of fibrosis. Progression from one stage to another does not necessarily represent an ordinal progression in matrix accumulation. For example, a biopsy with a fibrosis score of 4 may contain many times more than double the amount of matrix that is found in a biopsy with a score of 2. Inter-observer variation in interpretation introduces quantifiable errors that are well documented and compare unfavourably with automated pathology assays. If liver biopsy is to remain the means by which fibrosis is assessed, other approaches incorporating histological image analysis to quantify fibrosis may provide more accurate measures of matrix accumulation. The tools currently available for the assessment of liver fibrosis are inadequate particularly in advanced disease. None of the existing fibrosis staging systems distinguishes between differing degrees of severity of cirrhosis. Clinical scoring systems have been developed to stage the progressive deterioration of liver function in cirrhotic patients and at least two of these have been validated as prognostic scoring systems [15], [16], [17]. Thus, a comprehensive study of liver fibrosis from initiation to end-stage cirrhosis would require the combination of both histological and clinical scoring systems incorporating clinical end-points as the metric of fibrosis such as the onset of complications of cirrhosis, portal pressure measurements and death.

Inability to perform frequent and repeated liver biopsies and errors in interpretation confound attempts to answer many of the questions about the progression of chronic liver diseases.

Ultimately reliable, reproducible, non-invasive surrogate indicators would be most useful. Recent interest has focussed on the use of surrogate serum markers of liver fibrosis to monitor fibrosis progression. This approach has the advantages that serum samples can be obtained frequently with minimum risk and inconvenience to the patient. Studies have shown that algorithms combining these markers can reflect the severity of fibrosis in a concurrent liver biopsy with reasonable accuracy [18], [19]. Their use in monitoring fibrosis progression and as prognostic markers remains to be evaluated. Serum markers of liver fibrosis may hold great promise but prospective validation studies in which irrefutable outcomes are assessed will be needed.

Further studies of rates of fibrosis progression and the factors that influence them will undoubtedly provide important insights into chronic liver diseases and liver fibrosis. Careful selection of research methods and data interpretation will be vital.

Back to Article Outline

References 

  1. Friedman SL. Molecular regulation of hepatic fibrosis, an integrated cellular response to tissue injury. J Biol Chem. 2000;275:2247–2250
  2. Powell LW, Kerr JF. Reversal of “cirrhosis” in idiopathic haemochromatosis following long-term intensive venesection therapy. Australas Ann Med. 1970;19:54–57
  3. Hammel P, Couvelard A, O'Toole D, Ratouis A, Sauvanet A, Flejou JF, et al.  Regression of liver fibrosis after biliary drainage in patients with chronic pancreatitis and stenosis of the common bile duct. N Engl J Med. 2001;344:418–423
  4. Dufour JF, DeLellis R, Kaplan MM. Reversibility of hepatic fibrosis in autoimmune hepatitis. Ann Intern Med. 1997;127:981–985
  5. Wanless IR. Use of corticosteroid therapy in autoimmune hepatitis resulting in the resolution of cirrhosis. J Clin Gastroenterol. 2001;32:371–372
  6. Kaplan MM, DeLellis RA, Wolfe HJ. Sustained biochemical and histologic remission of primary biliary cirrhosis in response to medical treatment. Ann Intern Med. 1997;126:682–688
  7. Kweon YO, Goodman ZD, Dienstag JL, Schiff ER, Brown NA, Burkhardt E, et al.  Decreasing fibrogenesis: an immunohistochemical study of paired liver biopsies following lamivudine therapy for chronic hepatitis B. J Hepatol. 2001;35:749–755
  8. Lau DT, Kleiner DE, Park Y, Di Bisceglie AM, Hoofnagle JH. Resolution of chronic delta hepatitis after 12 years of interferon alfa therapy. Gastroenterology. 1999;117:1229–1233
  9. Poynard T, McHutchison J, Manns MP, Trepo C, Lindsay KL, Goodman Z, et al.  Impact of pegylated interferon alfa-2b and ribavirin on liver fibrosis in patients with chronic hepatitis C. Gastroenterology. 2002;122(5):1303–1313
  10. Poynard T, Bedossa P, Opolon P. Natural history of liver fibrosis progression in patients with chronic hepatitis C. The OBSVIRC, METAVIR, CLINIVIR, and DOSVIRC groups. Lancet. 1997;349:825–832
  11. Poynard T, Mathurin P, Lai C-L, Guyader D, Poupon R, Tainturier M-H, et al.  A comparison of fibrosis progression in chronic liver diseases. J Hepatol. 2003;38:257–265
  12. Darby SC, Ewart DW, Giangrande PL, Dolin PJ, Spooner RJ, Rizza CR. Mortality before and after HIV infection in the complete UK population of haemophiliacs. UK Haemophilia Centre Directors' Organisation. Nature. 1995;377:79–82
  13. Hadziyannis SJ, Vassilopoulos D. Hepatitis B e antigen-negative chronic hepatitis B. Hepatology. 2001;34:617–624
  14. Brenner H, Soderman B, Hakulinen T. Use of period analysis for providing more up-to-date estimates of long-term survival rates: empirical evaluation among 370,000 cancer patients in Finland. Int J Epidemiol. 2002;31(2):456–462
  15. Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg. 1973;60:646–649
  16. Forman LM, Lucey MR. Predicting the prognosis of chronic liver disease: an evolution from child to MELD. Mayo end-stage liver disease. Hepatology. 2001;33:473–475
  17. Kamath PS, Wiesner RH, Malinchoc M, Kremers W, Therneau TM, Kosberg CL, et al.  A model to predict survival in patients with end-stage liver disease. Hepatology. 2001;33:464–470
  18. Rosenberg WM, Burt AD, Becka M, Voelker M, Arthur MJ. Automated assays of serum markers of liver fibrosis predict histological hepatic fibrosis. Hepatology. 2000;32:183A
  19. Imbert-Bismut F, Ratziu V, Pieroni L, Charlotte F, Benhamou Y, Poynard T. Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: a prospective study. Lancet. 2001;357:1069–1075

PII: S0168-8278(03)00010-2

doi:10.1016/S0168-8278(03)00010-2

Journal of Hepatology
Volume 38, Issue 3 , Pages 357-360, March 2003

Article Tools