1. Top of page
  2. Abstract

Recurrent hepatitis C (RHCV) after liver transplantation is almost universal, and occasional patients will have an aggressive course characterized histologically by pericellular/sinusoidal fibrosis and cholestasis, known as fibrosing cholestatic hepatitis (FCH). The early stages and evolution of this disease have not been well characterized. A total of 77 liver biopsies performed for indication (nonprotocol) were evaluated for necroinflammation, rejection, cholestasis, and fibrosis. Control groups were composed of protocol biopsies from HCV transplant patients (10 biopsies) as well as non–HCV transplant patients (6 biopsies). Scoring for necroinflammation, rejection, and fibrosis were compiled using standard criteria (hepatic activity index, Banff, Ishak, METAVIR). Pericellular fibrosis was staged with a novel “sinusoidal” system. A cholestasis scoring system was developed to quantitate parenchymal and portal features of cholestasis. Biopsies were categorized as rejection, RHCV, FCH, and stable based on histology and clinical information. FCH was found to have a higher fibrosis stage overall when compared to most diagnostic groups, regardless of the staging system used. Additionally, sinusoidal fibrosis was significantly higher in the FCH diagnosis group. Cholestasis was more prominent in biopsies of FCH in all comparisons. In conclusion, the presence of cholestasis and fibrosis with mild to moderate RHCV should raise the suspicion of FCH. When studying the evolution of these cases, the first abnormality to appear is RHCV and cholestasis, fibrosis develops soon after, and both continue to worsen until the point of allograft failure or patient death. Liver Transpl, 2006. © 2006 AASLD.

Recurrent hepatitis C (RHCV) in the posttransplantation period is increasing in frequency and there is often an accompanying increase in severity.1 Recurrence is virtually universal and is a significant cause of severe liver allograft damage and subsequent decompensation,2 especially with genotype 1b,2, 3 which is one of the most common genotypes in North America and Europe. Management of these patients is delicate, and care must be taken to balance recurrent disease and allograft rejection by finding the appropriate combination of immunosuppression and antiviral therapy. RHCV is especially important in the setting of retransplantation. Increasing numbers of hepatitis C virus (HCV)-infected recipients in need of a second allograft are confronted with in an increasingly tight organ supply pool. Often, the second graft can be expected to have similar severity of recurrent disease as was experienced in the first,4 raising questions regarding the advisability of retransplantation in these cases.

Occasional transplant patients will have RHCV that takes an aggressive course with rapid deterioration. Fibrosing cholestatic hepatitis (FCH) is an uncommon but well-documented complication effecting transplant recipients with hepatitis B virus or HCV. This entity was originally described in the early 1990s; characterized histologically by extensive dense portal fibrosis with immature fibrous bands extending into sinusoidal spaces, ductular proliferation and hypercellularity, marked canalicular and cellular cholestasis, and moderate mononuclear inflammation.5 Although first described as a complication of hepatitis B, this phenomenon has been well documented in association with HCV in liver transplantation6–8 as well as HCV in the setting of bone marrow transplant,9 heart transplantation,10 and end-stage human immunodeficiency virus infection.11

Although FCH is a known complication of liver transplantation, the early stages and evolution have not been well described. Predicting which of these patients will have an aggressive course is difficult, and although biopsy staining with alpha smooth muscle actin has shown promise in predicting at risk patients,12, 13 this is not routinely performed. Early control of this entity is imperative if the allograft is to be salvaged. In the transplantation setting the most common diagnostic considerations are acute cellular rejection (ACR) and a more routine recurrent RHCV. It is important to be able to distinguish FCH from other, possible concurrent, complications in order to ensure proper therapy.

Recognition of FCH is imperative due to the rapid progression of this disease. This study compares the histology of FCH as opposed to other common diagnoses such as RHCV and ACR to delineate histologic characteristics that will aid in early diagnosis.


  1. Top of page
  2. Abstract

Archived biopsy specimens were obtained for review under approval of our institution's institutional review committee. A patient pool was drawn from patients who had undergone prior liver transplantation and who underwent biopsy during the period of January 2004 to December 2004. All HCV patients who had underwent biopsy based on clinical indications, usually elevated liver function tests, were selected for study. Control groups were drawn from the same patient pool and included HCV as well as non-HCV transplant patients who had undergone protocol biopsies (nonindication). Pretransplantation diagnoses for nonhepatitis patients included primary sclerosing cholangitis, primary biliary cirrhosis, and autoimmune hepatitis. A total of 59 patients (including 16 controls) were included in the study, producing 88 biopsies to review. Biopsies were randomized and hematoxylin and eosin stained slides were blindly reviewed, without background clinical information, and graded for necroinflammation, rejection, and cholestasis. Masson trichrome blue stained slides were reviewed to assess fibrosis staging.

Necroinflammatory score was graded using the hepatitis activity index (HAI), adapted from Ishak et al.14 Scores were assessed based on the presence or and degree of periportal or interface hepatitis, confluent necrosis, focal necrosis, and inflammation as well as portal inflammation. The Banff rejection consensus criteria were used to grade for acute cellular rejection.15 The presence or absence of mixed portal inflammation, bile duct damage, and endotheliitis were assessed and a grade was assigned.

A methodology of scoring cholestasis was developed that incorporated parenchymal and portal features. Parenchymal features noted were hepatocellular bile pigment, canalicular bile plugs, Kupffer cell bile pigment, and Zone 3 hepatocellular swelling and degeneration (Fig. 1A). Portal features were bile ductular proliferation, inspissated bile in ductules, and inspissated bile in ducts (Fig. 1B). The presence of these features and the severity of the findings were used to generate a score (Table 1).

thumbnail image

Figure 1. (A) Parenchymal features of cholestasis in FCH: hepatocellular bile pigment, canalicular bile plugs, and Kupffer cell bile pigment (hematoxylin-eosin, original magnification 20×). (B) Portal features of cholestasis in FCH: bile ductular proliferation (hematoxylin-eosin, original magnification 20×).

Download figure to PowerPoint

Table 1. Cholestasis Scoring System
Parenchymal features 
 Hepatocellular bile pigment 
 Canalicular bile plugs 
 Kupffer cell bile pigment 
 Zone 3 hepatocyte swelling/degeneration 
Portal tract features 
 Bile ductular proliferation 
 Inspissated bile in bile ductules 
 Inspissated bile in bile ducts 
Negative for cholestasis0
1 or more parenchymal features of cholestasis, but mild in extent1
2 or more parenchymal features of cholestasis, moderate in extent2
3 or more parenchymal features and 1 or more portal tract features, mild to moderate in extent3
3 or more parenchymal features and 1 or more portal tract features, severe in extent4

The presence and degree of fibrosis was assessed using standard criteria, namely the modified Ishak et al.14 and the METAVIR staging systems.16 Neither of the current systems takes into consideration the pericellular pattern of fibrosis that is often described in FCH (Fig. 2). Thus, an additional fibrosis staging system was developed that would incorporate these features. This system is based the number and frequency of sinusoidal “rays” as well as the eventual bridging and subdivision of the parenchyma (Table 2).

thumbnail image

Figure 2. Pericellular pattern of fibrosis in FCH (Masson trichrome blue, original magnification 20×).

Download figure to PowerPoint

Table 2. Fibrosis Scoring Systems
Modified Knodell/Ishak14 
 No fibrosis0
 Fibrous expansion of some portal tracts, with or without short fibrous septa1
 Fibrous expansion of most portal tracts, with or without short fibrous septa2
 Fibrous expansion of portal tracts with occasional portal to portal bridging3
 Fibrous expansion of portal tracts with marked portal to portal bridging as well as portal to central4
 Marked bridging portal to portal and/or portal to central with occasional nodules5
 Cirrhosis, probable or definite6
 No fibrosis0
 Portal fibrosis1
 Periportal fibrosis2
 Bridging fibrosis3
Sinusoidal fibrosis system (current paper) 
 No sinusoidal fibrosis0
 3 or fewer “rays” of sinusoidal fibrosis around portal tracts or central veins, nonbridging1
 Consistent 4 or greater “rays” of sinusoidal fibrosis, nonbridging2
 Bridging (portal-to-central) rays of sinusoidal fibrosis3
 Bridging rays of sinusoidal fibrosis with extensive subdivision of the parenchyma4

Once each was blindly assessed and scored, the biopsies were unrandomized and divided into diagnostic categories that were determined using clinical assessment (such as viral load, clinical outcome, and response to treatment) and original histologic interpretation. Biliary imaging was inconsistently available, and was not included in the assessment. Based on this information, biopsies were categorized as ACR, RHCV, FCH, and HCV with histologically stable allograft (stable). Clinically, biopsies categorized as ACR were responsive to steroid therapy and had low viral loads, whereas RHCV patients responded to antiviral therapy and often had high viral loads. Biopsies categorized as FCH had rapid clinical deterioration, with worsening liver function and very high viral loads. HCV infected patients with stable disease had low viral loads and mild liver test abnormalities. Control groups, which consisted of protocol biopsies (nonindication), were all clinically stable, and categorized as “control HCV”, and “control non-HCV” based on their pretransplantation diagnosis.

A total of 13 biopsies (representing 5 patients) were categorized as having complications other than those listed above, and were not included in the statistical analysis. Reasons for exclusion included surgical complications such as postoperative bleeding, bile leakage, and biliary stricture. Other complications necessitating exclusion were renal complications resulting in multiorgan failure, posttransplantation autoimmune hepatitis, recurrent lymphoma, and sepsis.

Of the remaining 77 biopsies eligible for this study, 32 were classified as RHCV, 6 as ACR, 11 as FCH, and 12 as stable. Of the control biopsies, the 10 HCV biopsies and 6 non-HCV biopsies were clinically and histologically stable (Table 3). Patient demographics and time status posttransplantation varied, and is reported in Table 3.

Table 3. Clinical Diagnostic Categories
Clinical diagnostic categoryNumber of biopsiesAge* (years)Male (%)s/p OLTX* (months)
  • *

    Values reported as median (range).

Recurrent hepatitis C (RHCV)3253 (±9.3)8124 (±30)
Acute cellular rejection (ACR)645 (±3.7)504.8 (±3.8)
Fibrosing cholestatic hepatitis (FCH)1152 (±3.1)759.8 (±8.9)
Histologically stable hepatitis C (stable)1250 (±14)7541 (±38)
Control HCV1054 (±6.4)7040 (±37)
Control non-HCV651 (±8.3)1756 (±21)

Necroinflammatory, rejection, fibrosis, and cholestasis scores for each diagnosis category were compared using the Kruskal-Wallis statistical test. Each diagnosis category was compared to FCH to determine if statistically significant distinguishing factors emerged.


  1. Top of page
  2. Abstract

Necroinflammatory Score

The individual components of the HAI grade, as well as the total score, were compared between FCH and each diagnostic category (Table 4). RHCV and FCH both tended to have higher HAI grades (median values 5.0 and 7.0, respectively). FCH was found to have a consistently higher HAI grade than RHCV (P = 0.042). Strict application of the HAI criteria to biopsies of ACR resulted in a high median (5.0), although this is likely an artificially inflated value. The high grade of ACR was largely due to higher scores in periportal/periseptal hepatitis (Table 4, A) and portal inflammation (Table 4, D), features seen in both HCV and ACR. Histologically characteristic features of rejection were present in the biopsies of ACR, and were reflected in the Banff score discussed below. Essentially, the histologic diagnosis of ACR was not an issue and these “HAI grade” values are reported for the purposes of comparison.

Table 4. Comparison of the Average HAI Grade and Components14 of Each Diagnostic Category to Fibrosing Cholestatic Hepatitis
 FCH (n = 11)RHCV (n = 32)ACR (n = 6)Stable (n = 12)Control HCV (n = 10)Control non-HCV (n = 6)
  • *

    Values reported as median (range).

HAI grade*
  P = 0.042P = 0.12P = 0.00010P = 0.00020P = 0.00020
A. Periportal or periseptal interface hepatitis*
  P = 0.017P = 0.22P = 0.00020P = 0.00080P = 0.0030
B. Confluent necrosis*000000
  P = 0.96P = 0.65P = 0.30P = 0.34P = 0.46
C. Focal (spotty) lytic necrosis, apoptosis and focal inflammation*
  P = 0.88P = 0.40P = 0.038P = 0.022P = 0.024
D. Portal inflammation*
  P = 0.0061P = 0.080P = 0.00010P = 0.00020P = 0.013

When considering individual criteria of the HAI grade, FCH showed a consistently higher amount of interface activity (Table 4, A) over RHCV (median values 3.0 and 2.0, respectively, with P = 0.017). Portal inflammation (Table 4, D) was also notably higher in FCH over RHCV (median values 3.0 and 2.0 respectively, with P = 0.0061).

Rejection Score

The Banff criteria were used to establish a rejection score in accordance with the published algorithm.15 The individual components (portal inflammation, bile duct damage, and venous endotheliitis) as well as the overall score were compared between FCH and the other diagnostic categories (Table 5). FCH was found to be significantly different from the HCV control group and the non-HCV control group on this parameter (P = 0.0090 and 0.018, respectively), recognizing that the Banff scoring system was not established for evaluation of HCV infection. FCH tended to have a higher score than recurrent HCV (median value 0, P = 0.0040). Reassuringly, ACR was found to have a higher score on average than FCH, as would be expected (median value 4.0, P = 0.0080).

Table 5. Comparison of the Banff Grade and Components15 of Each Diagnostic Category to Fibrosing Cholestatic Hepatitis
 FCH (n = 11)RHC (n = 32)ACR (n = 6)Stable (n = 12)Control HCV (n = 10)Control non-HCV (n = 6)
  • *

    Values reported as median (range).

  P = 0.0040P = 0.0080P = 0.12P = 0.0090P = 0.018
Portal inflammation*
  P = 0.0040P = 0.20P = 0.10P = 0.010P = 0.017
Bile duct damage*001.5000
  P = 0.012P = 0.010P = 0.26P = 0.040P = 0.10
Venous endothelium*001.0000
  P = 0.095P = 0.0033P = 0.49P = 0.17P = 0.28

Given that ACR vs. FCH is a common diagnostic dilemma, the rejection score was broken down into individual criteria and compared. FCH was found to have a relatively low overall Banff score (median value 2.0), with portal inflammation as the major contributor. Mixed inflammation was not uncommon in biopsies of FCH, and interestingly differences in score for portal inflammation between FCH and ACR were not found to be statistically significant (P = 0.2). As alluded to in the discussion of HAI grading, portal inflammation is a common finding in both of these entities. Rather, other features of rejection, such as bile duct damage and venous endotheliitis, were found to be better at differentiating the diagnoses. Bile duct damage scoring was consistently low in FCH biopsies, with a median value of 0 vs. an ACR median value of 1.5 (P = 0.010). Moreover, venous endotheliitis was virtually absent in FCH but was a frequent finding in the biopsies of ACR (median values 0 and 1.0, P =0.0033).

Fibrosis Staging

Fibrosis staging with the commonly utilized methods, the Ishak and METAVIR systems, yielded comparable results. FCH was found to have a statistically significant higher stage of fibrosis using the modified Ishak system when compared to all the different diagnostic groups, although when using METAVIR a statistical difference did not emerge when compared to recurrent HCV (Table 6).

Table 6. Comparisons of Fibrosis Staging Systems of Each Diagnostic Category to Fibrosing Cholestatic Hepatitis
 FCH (n = 11)RHCV (n = 32)ACR (n = 6)Stable (n = 12)Control HCV (n = 10)Control non-HCV (n = 6)
  • *

    Values reported as median (range).

  • See Table 2 for sinusoidal fibrosis system.

Ishak et al.14*
  P = 0.018P = 0.011P = 0.021P = 0.0020P = 0.0028
  P = 0.075P = 0.011P = 0.042P = 0.0020P = 0.0029
  P = 0.0016P = 0.0092P = 0.0066P = 0.0012P = 0.0036

We devised a system of grading fibrosis that attempted to quantitate features of fibrosis that are historically associated with FCH, specifically the sinusoidal pattern (Table 2). Traditional methods of fibrosis staging are based on a portal to portal or portal to central bridging process and do not take into account the sinusoidal fibrosis pattern present in FCH. This scoring system enabled us to semiquantitate such a pattern. Statistically significant differences between FCH and all of the other diagnostic categories were identified (Table 6).


A standardized system of grading cholestasis was also developed, using histologic findings and severity (Table 1). Cholestasis was a consistent finding in biopsies of FCH, and results of this grading system were compared between FCH and each of the diagnostic groups (Table 7). Cholestatic features were not infrequent in ACR (median value 2.0), although the features were more prominent in FCH (median value 4.0, P = 0.037). Cholestatic features were infrequent findings in the other diagnostic entities.

Table 7. Cholestasis
 FCH (n = 11)RHCV (n = 32)ACR (n = 6)Stable (n = 12)Control HCV (n = 10)Control non-HCV (n = 6)
  • *

    Values reported as median (range).

  • See Table 1 for cholestasis scoring system.

Cholestasis score*,
  P < 0.00010P = 0.037P = 0.00020P < 0.00010P = 0.00070

Observations Regarding Disease Evolution

The 11 biopsies of FCH represented 4 patients with sequential biopsies, allowing a careful study of the disease progression. In the earliest biopsies, the first abnormality to appear is mild recurrent HCV and cholestasis. Not infrequently, the significance of the cholestasis is missed. Cholestasis is not commonly seen in standard RHCV, and the presence of cholestasis can help to distinguish FCH from RHCV (Table 7). However, cholestasis is not uncommon in ACR (Table 7), and can confuse the diagnosis. The cholestasis, however, continues to progress over several weeks (Fig. 3). Fibrosis seems to lag behind cholestasis, usually by several weeks. This feature is uncommon in RHCV and ACR, and can help making the diagnosis. Early fibrosis consists of delicate pericellular bands that gradually extend until the classic fibrosis pattern becomes obvious (Fig. 4). FCH is usually evident in the fulminant, late-stage, form, but the early subtle features of cholestasis and early sinusoidal fibrosis can easily be overlooked or misinterpreted. Of note, all 4 of our cases of FCH resulted in patient death.

thumbnail image

Figure 3. Features of cholestasis in 1 patient, progressing over several months. (A) Early FCH showing cholestasis and ballooning (hematoxylin-eosin, original magnification 20×). (B) At 1 month later with more extensive ballooning (hematoxylin-eosin, original magnification 20×). (C) Approximately 1.5 months later, 11 weeks from initial biopsy; portal changes appear such as bile ductular proliferation (hematoxylin-eosin, original magnification 20×). (D) At 5.5 weeks later, 4 months after initial biopsy, severe ballooning and cholestasis (hematoxylin-eosin, original magnification 20×).

Download figure to PowerPoint

thumbnail image

Figure 4. Progressive fibrosis over several months, same patient as in Figure 3. (A) Initial biopsy, no significant pericellular fibrosis (Masson trichrome blue, original magnification 20×). (B) At 1 month after first biopsy, very early pericellular fibrosis (Masson trichrome blue, original magnification 20×). (C) Approximately 1.5 months later, 11 weeks after initial biopsy, showing more extensive fibrosis (Masson trichrome blue, original magnification 20×). (D) At 5.5 weeks later, 4 months after initial biopsy; extensive, bridging pericellular fibrosis (Masson trichrome blue, original magnification 20×).

Download figure to PowerPoint


  1. Top of page
  2. Abstract

Recognizing FCH is imperative due to the rapid progression and devastating consequences of this disease. This study compares the histology of FCH to other common diagnoses in this clinical setting to characterize features that will aid in the diagnosis.

Fibrosis is a well-known complication of RHCV as well as FCH. The sinusoidal fibrosis system is appealing as it addressed the pattern that is considered to be characteristic for FCH. However, a higher fibrosis stage, regardless of the system used, is associated with FCH. More important than the staging system is recognition of the sinusoidal pattern, and its implications on the differential diagnosis. Biopsies of HCV with this fibrosis pattern should be considered as possible cases of FCH. Fibrosis is a less common finding in the other major diagnostic consideration, ACR, and the presence of fibrosis, regardless of degree, can aid in this differential diagnosis. An assessment of these biopsies for hepatic stellate cell activation (via immunostaining with smooth muscle actin) may also help to elucidate cases at risk of FCH.

The systematic application of a cholestasis scoring system, though unnecessary in standard practice, was undertaken to offer a semiquantitative measure of this feature for our study. Cholestasis was found to be significantly higher in biopsies of FCH in all comparisons. Cholestasis, in addition to fibrosis, is one of the prominent and characteristic findings in FCH and is not common in standard RHCV. ACR may also have cholestasis; however, the cholestasis of FCH tends to be more severe, and the other findings of ACR such as bile duct damage and endotheliitis are not seen in FCH.

Important diagnostic clues to the diagnosis of FCH can be surmised from this study. In any case of graft dysfunction in the post-transplant HCV patient, FCH should be a diagnostic consideration. Although FCH is not a common diagnosis, the clinical implications are important. Without aggressive therapy, the patient will continue to deteriorate, and in cases of misdiagnosed ACR, more aggressive immunosuppression can be devastating. Our study helped to highlight the features of FCH, and showed specific features, such as fibrosis (especially sinusoidal) and cholestasis, can reliably be used to aid in the differential diagnosis.


  1. Top of page
  2. Abstract
  • 1
    Alonso O, Loinaz C, Abradelo M, Perez B, Manrique A, Gomez T, Jimenez C, et al. Changes in the incidence and severity of recurrent hepatitis C after liver transplantation over 1990–1999. Transplant Proc 2003; 35: 18361837.
  • 2
    Sánchez-Fueyo A, Restrepo JC, Quintó L, Bruguera M, Grande L, Sanchez-Tapias JM, et al. Impact of the recurrence of hepatitis C virus infection after liver transplantation on the long-term viability of the graft. Transplantation 2002; 73: 5663.
  • 3
    Prieto M, Berenguer M, Rayón J, Córdoba J, Argüello L, Carrasco D, et al. High incidence of allograft cirrhosis in hepatitis C virus genotype 1b infection following transplantation: Relationship with rejection episodes. Hepatology 1999; 29: 250256.
  • 4
    Berenguer M, Prieto M, Palau A, Rayón J, Carrasco D, San Juan F, et al. Severe recurrent hepatitis C after liver retransplantation for hepatitis C virus-related graft cirrhosis. Liver Transpl 2003; 9: 228235.
  • 5
    Davies S, Portmann B, O'Grady J, Aldis P, Chaggar K, Alexander G, Williams R. Hepatic histological findings after transplantation for chronic hepatitis B infection, including a unique pattern of fibrosing cholestatic hepatitis. Hepatology 1991; 13: 150157.
  • 6
    Schluger LK, Sheiner PA, Thung SN, Lau JY, Min A, Wolf DC, Fiel I, et al. Severe recurrent cholestatic hepatitis C following orthotopic liver transplantation. Hepatology 1996; 23: 971976.
  • 7
    Dickson RC, Caldwell SH, Ishitani MB, Lau JY, Driscoll CJ, Stevenson WC, et al. Clinical and histologic patterns of early graft failure due to recurrent hepatitis C in four patients after liver transplantation. Transplantation 1996; 61: 701705.
  • 8
    Gane EJ, Portmann BC, Naoumov NV, Smith HM, Underhill JA, Donalson PT, et al. Long-term outcome of hepatitis C infection after liver transplantation. N Engl J Med 1996; 334: 815820.
  • 9
    Cooksley W, McIvor C. Fibrosing cholestatic hepatitis and HBV after bone marrow transplantation. Biomed Pharmacother 1995; 49: 117124.
  • 10
    Delgado J, Munoz de Bustillo E, Ibarrola C, Colina F, Morales JM, Rodriguez E, et al. Hepatitis C virus-related fibrosing cholestatic hepatitis after cardiac transplantation: is azathioprine a contributory factor? J Heart Lung Transplant 1999; 18: 607610.
  • 11
    Poulet B, Chapel F, Deny P, Schneider MC, Jeantils V, Beaugrand M, Guettier C. Fibrosing cholestatic hepatitis by B virus reactivation in AIDS. Ann Pathol 1996; 16: 188191.
  • 12
    Russo M, Firpi R, Nelson D, Schoonhoven R, Shretha R, Fried M. Early hepatic stellate cell activation is associated with advance fibrosis after liver transplantation in recipients with hepatitis C. Liver Transpl 2005; 11: 12351241.
  • 13
    Gawrieh S, Papouchado B, Burgart L, Kobayashi S, Charlton M, Gores G. Early Hepatic stellate cell activation predicts severe hepatitis C recurrence after liver transplantation. Liver Transpl 2005; 11: 12071213.
  • 14
    Ishak K, Baptista A, Bianchi L, Callea F, De Groote J, Gudat F, et al. Histological grading and staging of chronic hepatitis. J Hepatol 1995; 22: 696699.
  • 15
    Banff Schema for grading liver allograft rejection: an international consensus document. Hepatology 1997; 25: 65863.
  • 16
    French METAVIR Cooperative Study Group. Intraobserver and interobserver variations in liver biopsy interpretation in patient's with chronic hepatitis C. Hepatology 1994; 20: 1520.