- Email: [email protected]
A histopathological study of the effects of 6-month versus 12-month interferon α-2b therapy in chronic hepatitis C
Journal of Hepatology 1996; 25:833-841 Printed in Denmark • All rights reserved Munksgaard. Copenhagen
Copyright © European Association for the Study of the Liver 1996 Journal of Hepatology ISSN 0168-8278
A histopathological study of the effects of 6-month versus 12-month interferon -2b therapy in chronic hepatitis C Marianne Ziol, Jeanne Tran Van Nhieu, Franqoise Roudot-Thoraval, Jean-Michel M6treau, Yves Deugnier, Daniel Dhumeaux and Elie Serge Zafrani Services d'Anatomie et de Cytologie Pathologiques et d'H~patologie, H~pital Henri Mondor, Crdteil, France
Background~Aims: Interferon therapy has been shown to have beneficial effects in chronic hepatitis C, but the optimal duration of treatment has not been clearly defined. The aims of this study were: (a) to perform a detailed histological comparison of the effects of a 6-month and a 12-month treatment using the Knodeil score as well as a recently proposed grid of analysis, (b) to determine possible histological predictive factors of response to therapy, and (c) to attempt to relate histological and biochemical modifications. Methods: Liver biopsies obtained before and 18 months after beginning of treatment were therefore compared in 26 patients treated for 6 months, and in 34 patients treated for 12 months. Results: Six months of treatment induced a significant decrease in periportal (p=0.02) and intralobular (p=0.004) hepatocyte necrosis. The same items were improved in the 12-month-treated patients but, in addition, portal inflammation (p=0.01), bile duct lesions (p=0.03), lymphoid aggregates
(p=0.002) and fibrosis (p=0.008) were also improved, according to the Knodell score. Low scores for fibrosis, steatosis and cholangiolar proliferation on the pretreatment liver biopsy could be considered predictive factors for alanine aminotransferase normalization at 6 months. There was no relationship between biochemical response and modification of fibrosis. Conclusion: Our results suggest that: (a) a decrease in fibrosis might be detected only after a 12month interferon treatment, and (b) initial fibrosis, cholangiolar proliferation and steatosis are predictive of a lack of biochemical response. The absence of a relation between biochemical response and evolution of fibrosis implies that the evaluation of treatments in chronic hepatitis C should always include a detailed histopathological study.
EPATmS C virus (HCV) infection leads to the development of chronic hepatitis in approximately 70% of cases (1). In patients with chronic hepatitis, cirrhosis occurs in 20 to 30% of the cases (1), and is associated with a significant risk of hepatocellular carcinoma development (2). Several multicenter randomized trials have demonstrated the beneficial effects of ix-interferon (IFN) therapy in chronic hepatitis C (3). Six months of ther-
apy has been shown to induce both reduction in alanine aminotransferase (ALT) levels and improvement of histological periportal and lobular necroinfiammatory lesions in half of the cases (4). However, the high prevalence of post-treatment relapse led to studies of the effects of more prolonged or higher doses of IFN therapy. In a recent study (5,6), it has been shown that 12 months of IFN treatment induced a trend to a decrease in post-treatment biochemical relapse, in comparison with a 6-month treatment. Other studies comparing 6-month and 12-month IFN treatment have reported the same results for biochemical response (7,8). In addition, a 12-month treatment compared to a placebo treatment has been reported to decrease all Knodell's histological items including
H
Received 27 September 1995; revised 20 January," accepted 25 February 1996
Correspondence: E. S. Zafrani, Service d'Anatomie et de Cytologie Pathologiques, H6pital Henri Mondor, 51 Avenue du Mar6chal de Lattre de Tassigny, 94 010 Cr6teil cedex, France. Tel. 33-1-49 81 27 32. Fax 33-1-49 81 27 33.
Key words: Chronic hepatitis C; Fibrosis; Hepatitis C virus; Histological score; Interferon ~ therapy.
833
M. Ziol et al.
TABLE 1 Semi-quantitative grading system of liver lesions in chronic hepatitis C (derived from the METAVIR Study Group propositions) Fibrosis (evaluated on picro-sirius-red-stained sections): 0: no fibrosis; 1: stellate enlargement of portal tracts without septa formation; 2: enlargement of portal tracts with rare septa formation; 3: numerous septa without cirrhosis; 4: cirrhosis. Portal inflammation and periportal hepatocyte necrosis: 0: absent; 1: mild; 2: moderate; 3: severe. Lymphoid aggregates in portal tracts: 0: none; 1: present in less than one third of portal tracts; 2: present in one third to two thirds of portal tracts; 3: present in more than two thirds of portal tracts. Lesions of inteflobular bile ducts: 0: none or rare; 1: presence of lymphocytes within duct epithelium or slight degenerative changes; 2: presence of primary biliary cirrhosis-like lesions. Cholangiolar proliferation: 0: none; 1: mild; 2: moderate; 3: marked. Intrasinusoidal lymphocyte aggregates: 0: absent or rare; 1: numerous; Intralobular hepatocyte necrosis: 0: absent or mild; 1: moderate; 2: severe. Steatosis: 0: absent; 1: involving less than 10% of the hepatocytes; 2: involving 10 to 30% of the hepatocytes; 3: involving more than 30% of the hepatocytes. Iron storage (evaluated on Perls' stained sections): 0: absent; 1: mild accumulation in hepatocytes or Kupffer cells; 2: moderate or marked accumulation in hepatocytes or Kupffer cells. Global activity: 1: no activity; 2: minimal activity; 3: moderate activity; 4: severe activity. Cirrhosis: 0: absent; 1: potentially present; 2: certain.
fibrosis score (9), and discrepancies between biochemical and histological responses have been underlined (10-12). The main purpose of the present work was to perform a detailed histological study of the effects of the two types of treatment in the trial cited above (6). The other aims of the work were also: (a) to report the initial lesions on pretreatment biopsies and to study their possible relationships; (b) to relate initial lesions to ALT modifications in order to determine possible histological predictive factors for response to IFN therapy; (c) to try to relate biochemical and histological modification after treatment. To address these questions, we used a recently published and evaluated grid of liver biopsy analysis proposed by the METAVIR cooperative study group (13). This grid semiquantitatively grades all elementary lesions possibly encountered in chronic hepatitis C, and not only necroinflammatory changes and fibrosis as in the 834
Knodell index (14). It also gives a global score of hepatitis activity separated from fibrosis score. This grid has the advantage of being more precise than the Knodell index (14) for detailed histopathological studies. However, we also used the Knodell index to compare our findings with those published in other reports.
Patients and Methods The 60 patients who had sufficient pre- and post-treatment biopsy material for detailed histopathological study were selected from the 108 patients with chronic active hepatitis C, from 10 French hospitals enrolled in the randomized control trial comparing a 6-month and a 12-month IFN regimen (6). The semiquantitative liver biopsy analysis according to the METAVIR grid using 13 parameters required large biopsy samples. This condition was achieved in both pre- and post-treatment biopsies in only 60 patients.
Histological evaluation of lFN therapy TABLE 2 Distribution of initial lesions according to the METAVIR scoring system on pre-treatment biopsies Score for each item n=60
0
1
2
3
4
Fibrosis Portal inflammation Periportal hepatocyte necrosis Portal lymphoid aggregates Bile duct epithelium lesions Cholangiolar proliferation Intrasinusoidal lymphocyte aggregates Intralobular hepatocyte necrosis Steatosis Iron storage* Global activity Cirrhosis
0 0 0 17 30 9 8 10 15 39 0 40
21 9 16 25 25 32 44 41 13 10 16 3
15 46 37 16 4 17 8 9 15 5 37 17
7 5 7 2
17
2
17 7
* n=54 because of 6 deficient Perls' staining.
The details of the trial have previously been reported (6). All patients gave written formal consent and the study was approved by the ethics committee, at Hospital Henri Mondor, Creteil, France. In the 60 patients, diagnosis was based on an elevation of ALT level for more than 6 months, and on positivity for hepatitis C virus antibodies. Chronic active hepatitis with or without cirrhosis was histologically proven on a liver biopsy done within 6 months before inclusion. Non-inclusion criteria were: (a) signs of severe liver failure or portal hypertension; (b) presence of antiHIV antibodies; (c) antiviral or corticosteroid therapy within 6 months prior to inclusion; (d) alcoholism or drug addiction within 2 years before inclusion; (e) hemoglobin rate below 11 g/l, platelet count below 70000/ktl, leucocytes count below 3000/ktl. Determination of gamma-glutamyl transpeptidase (?GT) and alkaline phosphatase (AP) levels was performed within the 2 weeks before inclusion. The patients were randomly assigned to either 6month or 12-month IFN (Intron-A; Schering-Plough) therapy. Fifty-two patients received 3 MIU IFN subcutaneously three times a week for 6 months, whereas 56 were given 12-month IFN therapy consisting of 3 MIU three times a week for 6 months, 2 MIU three times a week for the following 3 months, and 1 MIU three times a week for the last 3 months. Groups I and II were similar in terms of clinical, biological and histological data at the time of starting therapy. In addition, there was no significant difference between the two groups of patients as far as dropout rate and side effects were concerned. All patients were followed up every month for 18 months after the beginning of therapy. The biochemical response to IFN was defined as complete when the serum ALT level was normal at the end of ther-
apy. It was defined as sustained when serum ALT level was still normal 6 months after the end of therapy. Other patients were considered non-responders. All patients underwent two liver biopsies: the first one was performed during the 6 months before beginning therapy, and the second, 18 months after beginning therapy. Liver biopsies were fixed in Bouin's liquid, then embedded in paraffin. Five-micrometerthick sections were stained with hematoxylin and eosin, picro-sirius red for collagen and Perls' technique for hemosiderin. Sections were examined under code without knowledge of the type of treatment or of the chronologic order of the biopsies. Twenty-six of the 60 patients selected in the present study received a 6-month treatment (Group I) and the 34 others received a 12-month treatment (Group II). The 13 items that were graded are derived from the METAVIR Cooperative Study Group proposition (13) and are characterized by either continuous or categorical variables. This liver biopsy grading system is described in Table 1. Conclusive items include global evaluation of hepatitis activity and of fibrosis. All biopsy specimens were also graded according to the Knodell score. Biopsies were separately examined and graded by two pathologists (E.S.Z and M.Z), and then simultaneously reviewed using a double-pipe microscope in order to analyze discrepancies and to reach a consensus.
Statistical analysis Relationships between the initial histological lesions were evaluated by the non-parametric Spearman rank correlation for each item. The relationship between initial bile duct lesions and presence of follicles in portal tracts on one hand, and pretreatment AP and 835
M. Ziol et al.
TABLE 3 Changes in histological score before and 18 months after the beginning of therapy in 6-month-treated patients (group I) and 12-month-treated patients (group II) Group I: n=26 METAVIR grid items: Fibrosis Portal inflammation Periportal bepatocyte necrosis Portal lymphoid aggregates Bile duct lesions Cholangiolar proliferation Intrasinusoidal lymphocytes Intralobular hepatocyte necrosis Steatosis Iron storage Global activity Cirrhosis Knodell items: Fibrosis Periportal necrosis Intralobular necrosis Portal inflammation Group II: n=34 METAVIR grid items: Fibrosis Portal inflammation Periportal hepatocytes necrosis Portal lymphoid aggregate Bile duct lesions Cholangiolar proliferation Intrasinusoidal lymphocytes Intralobular hepatocyte necrosis Steatosis Iron storage Global activity Cirrhosis Knodell items: Fibrosis Periportal necrosis Intralobular necrosis Portal inflammation
Worse (%)
No change (%)
Better (%)
p*
11 15 12 19 27 8 8 15 8 5 12 12
62 46 46 42 42 69 38 46 77 77 42 81
27 39 42 39 31 23 54 39 15 18 46 7
NS NS p=0.02 NS NS NS p--0.002 NS (p--0.07) NS NS p--0.01 NS
8 12 4 12
77 50 54 61
15 38 42 27
NS NS p=0.004 NS
Worse(%)
No change (%)
Better (%)
p*
12 9 12 6 12 18 18 9 21 17 9 9
53 53 41 53 47 68 41 50 44 69 41 79
35 38 47 41 41 14 41 41 35 14 50 12
NS (p=0.1) p=0.01 p=0.006 p=0.002 p=0.03 NS p=0.03 p=0.006 NS NS p=0.001 NS
6 12 12 9
62 41 36 54
32 47 52 37
p=0.008 p=0.0037 p=0.005 p=0.007
*: Wilcoxon signed-rank test.
qa3T levels on the other hand, were analyzed using the Chi-square test or Fisher exact test. Pre- and posttreatment grades for each histological item, including those of the Knodell score, were compared using the Wilcoxon non-parametric test for paired data in groups I and II. Histological and biochemical modifications, 18 months after the beginning of therapy, were compared using a Chi-square test in all patients. Lesions were considered improved when the difference between pre- and post-treatment scores was superior or equal to 1. In order to find histological predictive factors for biochemical response, histological scores on pretreatment biopsies were compared by Chi-square test in patients who either completely responded or failed to respond 6 months after the
836
beginning of therapy. At that time, patients had received the same IFN doses. A multivariate analysis using a stepwise logistic regression model was further performed in order to find factors independently related to biochemical response. Assessment of histological criteria predictive of amelioration of activity and fibrosis was done using univariate and multivariate analyses.
Results Pre-treatment histological lesions Distribution of the histological lesions on initial biopsies, graded according to the METAVIR evaluation system, is reported in Table 2. In summary, cirrhosis was observed in 17 patients (28%) and global evalua-
Histological evaluation of lFN therapy TABLE 4 Histological response in relation to sustained biochemical response 18 months after beginning of therapy in 6-month and 12-month treated patients Histological improvement METAVIR grid items: Fibrosis Portal inflammation Portal lymphoid aggregates Bile duct lesions** Periportal hepatocyte necrosis Intralobular necrosis Intrasinusoidal lymphocytes Global activity Knodell items: Fibrosis Periportal hepatocyte necrosis Intralobular hepatocyte necrosis Portal inflammation
Sustained responders n=20
Non-sustained responders n--40
Number of patients n=60
Significance*
8(40%) 14(70%) 13(65 %) 10(50%) 15(75%) 13(65 %) 11(55%) 16(80%)
11(28%) 9(23%) 11(28%) 10(25%) 12(30%) 11(28%) 17(43 %) 13(33%)
19(32%) 23(38%) 24(40%) 20(33%) 27(45%) 24(40%) 28(47%) 29(48%)
NS p=0.0004 p=0.005 p=0.04 p=0.01 p--0.005 NS p=0.0005
5(25%) 15(75%) 15(75%) 13(65%)
10(25%) 11(28%) 14(35%) 7(18%)
15(25%) 26(43%) 29(48%) 20(33%)
NS p=0.0005 p=0.0035 p--0.0002
*: Chi-square test. **: Sustained responders n=19, non-sustained responders n=39.
tion of hepatitis activity showed mild or moderate activity in 53 patients (88%). Portal lymphoid aggregates were found in 43 patients (72%), bile duct damage in 29 (49%), with primary biliary cirrhosis-like lesions of the interlobular bile ducts in only four patients (6%). Steatosis was found in 45 patients (75%). The grades of each item on initial biopsies were interrelated, excluding the global hepatitis activity score and the presence or not of cirrhosis, which are already integrated items. Fibrosis was strongly related to periportal necrosis (p=0.0001) and to cholangiolar proliferation (p=0.0001). Portal inflammation, intralobular and periportal necrosis, and intrasinusoidal lymphocytes were all related (p<0.001 for each item to the other). Bile duct damage was related to portal inflammation (p=0.0001), to intralobular necrosis (p=0.001) and to the presence of portal lymphoid aggregates (p=0.001). Other items, such as steatosis and iron storage, did not appear to be related to other lesions. Study of the relationships between interlobular bile duct damage and AP and ~GT levels showed that primary biliary cirrhosis-like lesions (grade 2) of the interlobular bile ducts were associated with elevated ~3T (p=0.052). Grade 0 or 1 lesions of the interlobular bile ducts were not related to AP or ~3T levels.
between the two groups. Six of the 13 responders (46%) in group I had sustained response, whereas 14 of the 18 responders (78%) had sustained response in group II. The difference between the two groups approached but did not reach significance (p=0.07).
Biochemical response to treatment Thirteen of the 26 patients (50%) had a complete response in the 6-month treated group and 18 of 34 (53%) had a complete response in the 12-month treated group. There was no significant difference
Histological predictive factors for short-term biochemical response A relationship between initial histological lesions and biochemical response 6 months after beginning therapy was also searched for. At 6 months, group I and
Histological response to treatment The main results of the histological evaluation performed on the 60 pre- and post-treatment biopsies are indicated in Table 3. Comparison of pre- and posttreatment scores in the 6-month-treated group showed a significant improvement for periportal hepatocyte necrosis evaluated by the METAVIR scoring system, for intralobular necrosis evaluated by the Knodell score, for intrasinusoidal lymphocyte score and for global hepatitis activity index according to METAVIR. In the 12-month-treated group, the same items improved significantly. In addition, portal inflammation evaluated by the two grading systems as well as bile duct lesions and lymphoid aggregates were also significantly decreased. It must be pointed out that patients who were treated for 12 months had a significant improvement in their fibrosis score according to Knodell; when the METAVIR scoring system was used, there was a trend to decreased fibrosis, but this decrease did not reach significance (Table 3).
837
M. Ziol et al.
group II patients had received the same IFN doses. High scores for fibrosis and cholangiolar proliferation appeared to be predictive of persistence of elevated ALT at 6 months. Indeed, 16 of 34 (47%) patients who initially had a high score (grade 3 or 4 according to Knodell) of fibrosis were responders vs 20 of 26 (77%) who initially had a low score (grade 0 or 1 according to Knodell) of fibrosis (p=0.02); 8 of 19 patients (42%) who had a high score (grades 2 or 3 according to METAVIR) of cholangiolar proliferation were responders vs 28 of 41 (68%) who initially had a low score (grades 0 or 1) of cholangiolar proliferation (p=0.05). In addition, the presence of steatosis was also predictive for a lack of short-term biochemical response, since 23 of 45 patients (51%) who had minimal, moderate or marked steatosis were responders vs 13 of 15 (87%) who had no steatosis (p=0.01). Multivariate analysis, using a stepwise logistic regression model, showed that steatosis (p=0,01) and cholangiolar proliferation (p=0,05) were the two factors independently related to short-term biochemical response.
Histological predictive factors for activity and fibrosis amelioration Univariate analysis showed that: (a) an initial high score for periportal hepatocyte necrosis (p=0.01) and a low score for steatosis (p=0.02) were significantly related to amelioration of activity according to METAVIR; there was also a trend for an initial high score for portal lymphoid aggregates (p=0.07); (b) an initial low score for cholangiolar proliferation (p=0.01) and a high score for fibrosis (p=0.03) were predictive of fibrosis amelioration. Multivariate analysis showed that: (a) an initial high score for periportal hepatocyte necrosis (p=0.03) and a low score for fibrosis (p=0.01) were independently and significantly related to amelioration of activity according to METAVIR; (b) an initial low score for cholangiolar proliferation (p=0.01) and a high score for fibrosis (p<0.001) were predictive of fibrosis amelioration.
Relationships between biochemical and histological response Relationships between histological improvement and sustained biochemical response at month 18 are shown in Table 4, only for histological items improved by therapy. Rates of histological improvement were significantly higher in patients who were responders at 18 months by comparison with nonresponders, for portal inflammation, periportal and intralobular hepatocyte necrosis, bile duct lesions and
838
global evaluation of hepatitis activity score. In contrast, improvement in fibrosis according either to METAVIR or to Knodell, did not appear to be related to sustained biochemical response.
Discussion In order to evaluate pre- and post-treatment liver biopsies in chronic hepatitis C treated by interferon, two scoring systems (i.e., the Knodell score and the METAVIR grid) were used in the present work. The Knodell score was used to compare our findings to other published reports in which it is usually employed (9,11). However, the Knodell score only evaluates hepatocyte necrosis, portal inflammation and fibrosis. The METAVIR scoring system, for which an analysis of intra- and inter-observer variations has been recently published (13), has the advantage of semi-quantitatively grading all elementary lesions possibly encountered in chronic hepatitis C (e.g., bile duct lesions, presence of portal lymphoid aggregates, steatosis) and giving a global score of hepatitis activity separated from fibrosis score. Using this METAVIR evaluation system, the interrelationships between numerous histological lesions of chronic hepatitis C and the relationships between histological and biochemical responses can be studied. In addition, modifications of liver biopsy findings, when different modalities of treatment are compared, can be evaluated on more histological items. Finally, possible histological predictive factors of biochemical response may be determined. The prevalences of each histopathological lesion on pre-treatment biopsies are close to those previously reported in chronic hepatitis C (15-18). In summary, chronic hepatitis is mildly or moderately active in 88% of the cases, steatosis of more than 10% of the hepatocytes is observed in 75% of the liver biopsies, the presence of at least one lymphoid aggregate is noted in 72% of the cases, and mild or severe bile duct damage can be seen in 49% of the cases. The discrepancies in the determination of the prevalence of bile duct damage, varying from 91% (16) to 31% (17), are probably due to an absence of consensus for the definition of this lesion. High prevalences could be explained by the fact that the presence of occasional inflammatory cells in the interlobular bile duct seems to be considered indicative of bile duct damage by some authors (16). In our study, bile duct lesions were observed in 49% of the cases but only four of the 60 patients (6%) had severe damage resembling the florid lesion of primary biliary cirrhosis. The study of the interrelationships between pretreatment histological lesions showed that periportal
Histological evaluation of lFN therapy
and intralobular hepatocyte necrosis were related to portal and intralobular lymphocytic infiltration and that periportal necrosis was strongly linked to fibrosis. These relationships may be explained by events responsible for liver fibrogenesis, e.g., release of mitogens by necrotic hepatocytes, subendothelial matrix degradation, production of cytokines by platelets and macrophages, and autocrine and paracrine stimulation of Ito cells that produce fibrosis components (19). We have also noted that bile duct damage was related to portal lymphoid aggregates and that these two lesions were also related to intralobular hepatocyte necrosis. Lau et al. (20) reported similar findings and observed a relationship between such lesions and high levels of viremia. Since HCV virus cannot be immunohistochemically detected in the bile duct of portal tracts (21), an immune-mediated mechanism has been suggested for such lesions. In addition, the relation of lymphoid aggregates to bile duct damage is of interest, since these lesions may also be associated with the genotype lb of virus C (22). The present histological evaluation demonstrated better effects of a 12-month than a 6-month IFN treatment on some histological liver lesions. In the 6month-treated patients, periportal necrosis, intrasinusoidal lymphocytic infiltration and global activity score were significantly decreased. In the 12-monthtreated patients, the same histological lesions were also improved, and, in addition, portal inflammation, bile duct lesions, and fibrosis evaluated by the Knodell score were significantly reduced. Changes in fibrosis evaluated by the METAVIR grid showed a trend to decrease that did not reach significance. Such a difference in statistical significance when the two systems of histological evaluation are used may be explained by the fact that the Knodell score could be more sensitive in revealing improvement in fibrosis after IFN treatment: indeed, in the Knodell index there are less categories for fibrosis than in the METAVIR grid, and the presence of rare fibrous septa is not differentiated from the presence of numerous septa as in METAVIR scoring system. It is therefore possible to hypothesize that a 12month IFN treatment is probably able to slightly ameliorate fibrosis, but that this amelioration, which does not appear significant when multiple categories are defined in a scoring system, can be statistically revealed only when a more discontinuous score is used. Improvement of necrosis and, with a lower prevalence, of portal inflammation, is a well-described beneficial effect of IFN therapy in the literature (11,
23-26). The results of the present study, in the 6month as well as in the 12-month treated patients, are in agreement with these findings. However, significant reduction of fibrosis has rarely been reported. One report studying 59 patients with chronic non-A, non-B hepatitis receiving IFN or placebo for 6 months (27) has shown a significant decrease of total liver collagen in IFN-treated patients using a quantitative histochemical and morphometrical evaluation of collagen in liver biopsies, but this decrease of liver collagen was not reflected by a reduction of the fibrosis score evaluated according to Knodell. In two other series of patients treated with IFN for 12 months, an improvement in the total Knodell score and in each of its components, including fibrosis, has been reported on liver biopsies performed immediately at the end of the treatment (9,10). Our results suggest that histological improvement can be sustained 6 months after the end of treatment. An attempt to determine which histological lesions might or might not be related to ALT normalization was made in this study. Whereas amelioration of necroinflammatory changes as well as bile duct damage appeared to be related to ALT normalization at 18 months after beginning of therapy, no relationship was found between fibrosis and ALT levels. Indeed, 10 of 40 non-responder patients (25%) and 5 of 20 responders (25%) showed improvement in the fibrosis item of the Knodell score; 8 of 20 responders (40%) vs 11 of 40 non-responders (28%) had amelioration of fibrosis according to the METAVIR grading system. Such a finding might suggest that the effect of IFN on liver fibrosis could be independent of its antiviral activity that probably reduces hepatocyte necrosis. This hypothesis is supported by recent studies that postulate a direct inhibitory effect of IFN on fibrogenesis. Indeed, Mallat et al. (28) demonstrated a growth inhibition effect and antifibrogenic effects of tx- and T-interferon on cultured human Ito cells, and Castilla et al. (29) showed normalization of serum procollagen type III peptide and of TGF [31mRNA levels in liver tissue in eight patients treated with IFN for 1 year. High scores of fibrosis and cholangiolar proliferation, as well as steatosis appeared to be predictive histological factors for a lack of ALT normalization. Fibrosis has already been identified as a predictive factor for biochemical response (30). It is not surprising that cholangiolar proliferation was also predictive of such a lack of biochemical response, since we showed that this item was strongly related to fibrosis. Steatosis, which is not significantly modified by IFN therapy, has already been implicated in biochemical 839
M. Ziol et al.
response by Schvarcz et al. (31), who suggested that steatosis could mask biochemical response by being directly responsible for raised ALT levels. The search for histological predictive factors for amelioration of fibrosis and activity suggests that modifications under therapy of these two major parameters might be independent, since initial low scores of fibrosis were related to activity amelioration, whereas initial high scores of fibrosis were related to fibrosis amelioration. In conclusion, the present histopathological study, using a detailed scoring system such as proposed by the METAVIR Group, shows that: (a) in HCV-infected patients, there are strong relationships between initial elementary lesions, such as between nodular lymphoid aggregates and bile duct damage, which might be related to HCV viremia levels and to HCV genttypes; in addition, high scores of fibrosis, cholangiolar proliferation and steatosis seem to be predictive for absence of ALT normalization; (b) a 12-month treatment by IFN not only allows the maintenance of the beneficial histological effects noted at 6 months, but might also induce a slight improvement in fibrosis; it also seems to ameliorate portal inflammation, portal lymphoid aggregates, bile duct lesions, and intrasinusoidal lymphocytes scores; (c) the absence of relationship between biochemical response and modification of fibrosis implies that the evaluation of IFN effects should always include, not only the determination of biochemical parameters, but also a detailed study of comparative histopathological findings.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Acknowledgements The authors thank the members of the French Group for the Study of Chronic Hepatitis NANB/C Treatment: Y. Calmus, R. Poupon (H6pital Saint Antoine, Paris), H. Jouanolle (H6pital Pontchaillou, Rennes), J.P. Miguet (H6pital Jean Minjoz, Besanqon), C Br6chot (H6pital La6nnec, Paris), J.L. Payen, P. Cal6s (H6pital de Purpan, Toulouse), A. Tran (H6pital de l'Archet, Nice), M. Beaugrand (H6pital Jean Verdier, Bondy), J.P. Capron (CHU Nord, Amiens), J.P. Zarski (H6pital de la Tronche, Grenoble), M. Doffo~l (Hotel Dieu, Strasbourg), E. A. Pariente (H6pital la Source, Orl6ans), and C. Lemonnier (Schering Plough, Levallois-Perret).
12.
13.
14.
15.
16.
References 1. Giral P, Serfaty L, Loria A, Poupon R. Hepatite C. Rev Med Intern 1994; 15: 487-93. 2. Tsukuma H, Hiyama T, Tanaka S, Nakao M, Yabunchi T, Kitamira T, Nakanishi K, Fujimoto I, Inoue A, Yamasaki H, Kawashima T. Risk factors for hepatocellular carcinoma
840
17.
among patients with chronic liver disease. N Engl J Med 1993; 328: 1797-801. Pawlotsky JM, Marcellin P, Trepo C, Dhumeaux D. Traitement des hepatites virales C. Gastroenterol Clin Biol 1994; 18: 362-73. Hoofnagle JH. Therapy of chronic hepatitis B and C. In: Miguet JP, Dhumeaux D, eds. Progress in Hepatology 93. Paris: John Libbey Eurotext, 1993: 89-97. Metreau JM and the French Group for the study of NANB/C chronic hepatitis treatment. Results of long-term interferon treatment in non-A, non-B/C chronic active hepatitis. Gut 1993; suppl: S112-S113. Jouet P, Roudot-Thoraval F, Dhumeaux D, Metreau JM, and the French Group for the Study of NANB/C Chronic Hepatitis Treatment. Comparative efficacy of Interferon alpha in cirrhotic and noncirrhotic patients with non-A non-B, C hepatitis. Gastroenterology 1994; 106: 686-90. Milella M, Santantonio M, Monno L, Pastore G. Comparison of short vs long term treatment with recombinant alpha interferon in chronic hepatitis C (Abstract). In: Abstracts of the 1993 International Symposium on Viral Hepatitis and Liver Disease, Tokyo 1993: 245. Chemello L, Pontisso P, Rose KA, Simmonds P, Boneui P, Cavalletto L. The long-term response (LTR) to interferon alpha-2a(IFN-2a) in chronic hepatitis C is influenced by dose and duration of treatment and by the HCV serotype (abstract). J Hepatol 1993; 18(suppl 1): SI0-S11. Camps J, Castilla A, Ruiz J, Civeira MP, Prieto J. Randomised trial of lymphoblastoid ct-interferon in chronic hepatitis C. J Hepatol 1993; 17: 390-6. Reichard O, Glaumann H, Norkrans G, Wejstal R, Fryden A, Schvarcz R, Weiland O. Histological outcome in patients with chronic hepatitis C given a 60-week interferon alfa-2b treatment course. Liver 1994; 14: 169-74. Causse X, Godinot H, Chevallier M, Chossegros P, Zoulim E Ouzan D, Heyraud JP, Fontanges T, Albrecht J, Meschievitz C, Trepo C. Comparison of 1 or 3 MU of interferon alfa2b and placebo in patients with chronic non-A, non-B hepatitis. Gastroenterology 1991:101: 497-502. Marcellin P, Pouteau M, Duchatelle V, Degott C, Degos F, Boyer N, Erlinger S, Benhamou JP. Post-treatment histologic outcome in patients with chronic hepatitis C treated with recombinant alpha interferon. In: AASLD Abstracts. Hepatology 1993; 18: 243A. The French METAVIR Cooperative Study Group. Intra- and interobserver variations in liver biopsy interpretation in patients with chronic hepatitis C. Hepatology 1994; 20: 15-20. Knodell RG, Ishak KG, Black WC, Chen TS, Craig R, Kaplowitz N, Kiernan TW, Wollman J. Formulation and application of a numerical scoring system for assessing histological activity in asymptomatic chronic active hepatitis. Hepatology 1981; 1: 431-5. Czaja AJ, Carpenter HA. Sensitivity, specificity and predictability of biopsy interpretations in chronic hepatitis. Gastroenterology 1993; 105: 1824-32. Bach N, Thung SN, Schaffner E The histological features of chronic hepatitis C and autoimmune chronic hepatitis: a comparative analysis. Hepatology 1992; 15: 572-7. Lefkowitch JH, Schiff ER, Davis GL, Perillo RP, Lindsay K, Bodenheimer HC, Balart LA, Ortego TJ, Payne J, Dienstag JL, Gibas A, Jacobson IM, Tamburro CH, Carey W, O'Brien C, Sampliner R, Van Thiel DH, Feit D, Albrecht J, Me-
Histological evaluation of lFN therapy
18.
19.
20.
21.
22.
23.
24.
schievitz C, Sabghvi B, Vaughan RD. Pathological diagnosis of chronic hepatitis C. A multicenter comparative study with chronic hepatitis B. Gastroenterology 1993; 104: 595-603. Scheuer PJ, Ashrafzadeh P, Sherlock S, Brown D, Dusheiko GM. The pathology of hepatitis C. Hepatology 1992; 15: 567-71. Friedman SL. The cellular basis of hepatic fibrosis. Mechanisms and treatment strategies. N Engl J Med 1993; 328: 1828-35. Lau JYN, Davis GL, Kniffen J, Qian KP, Urdea MS, Chan CS, Mizokami M, Neuwald PD, Wilber CW. Significance of serum hepatitis C virus RNA levels on chronic hepatitis C. Lancet 1993: 341: 1501--4. Krawczynski K, Beach MJ, Bradley DW. Hepatitis C virus antigen in hepatocytes: immunomorphologic detection and identification. Gastroenterology 1992; 103: 622-9. Tran Van Nhieu J, Roudot-Thoraval F, Pawlotsky JM, Pellet C, Dhumeaux D, Zafrani ES. Relationship between genttypes of C virus and histologic liver lesions in chronic hepatitis C infection. In: AASLD Abstracts. Hepatology 1994; 20: 247A. David E, Puci A, Palladin D, Saracco G, Garello E, Pintus C, Roca G, Chiandussi L, Solinas A, Mollo E Rizetto M, Mazzuco G, Verme G. Histologic changes in liver biopsy specimens produced by recombinant interferon u-2b therapy for chronic non-A, non-B viral hepatitis. Am J Clin Pathol 1992; 98: 397--401. Davis GL, Balart LA, Schiff ER, Lindsay K, Bodenheimer HC, Perrillo RP, Garey W, Jacobson IM, Payne J, Dienstag JL, VanThiel DH, Tamburro C, Lefkowitch J, Albrecht J, Meschievitz C, Ortego TJ, Gibas A, and The Hepatitis Inter-
25.
26.
27.
28.
29.
30.
31.
national Therapy Group. Treatment of chronic hepatitis C recombinant interferon alfa. N Engl J Med 1989; 321: 1501-6. Di Bisceglie AM, Martin P, Kassianides C, Lisker-Melman M, Murray L, Waggoner J, Goodman Z, Banks SM, Hoofnagle JH. Recombinant interferon alfa therapy for chronic hepatitis C. N Engl J Med 1989; 321: 1506-10. Omata M, Ito Y, Yokosuka O, Imazeki F, Tagawa M, Takano S, Hosoda K, Tada M, Ohto M, Ito K, Okuda K. Randomized, double-blind, placebo-controlled trial of eight-week course of recombinant ct-interferon for chronic non-A, non-B hepatitis. Dig Dis Sci 1991; 36: 1217-22. Manabe N, Chevallier M, Chossegros P, Causse X, Guerret S, Trepo C, Grimaud JA. Interferon-~2b therapy reduces liver fibrosis in chronic non-A, non-B hepatitis. A quantitative histological evaluation. Hepatology 1993; 18: 1344-9. Mallat A, Preaux AM, Blazejewski S, Rosenbaum J, Dhumeaux D, Mavier P. Interferon alpha and gamma inhibit proliferation and collagen synthesis of human Ito cells in culture. Hepatology 1995; 21: 1003-10. Castilla A, Prieto J, Fausto N. Transforming growth factors B1 and o~ in chronic liver disease. N Engl J Med 1991; 324: 933--40. Pagliario L, Craxi A, Camaa C, Tine F, Di Marco V, Lo Iacono O, Almaso P. Interferon-or for chronic hepatitis C: an analysis of pretreatment clinical predictors of response. Hepatology 1994; 19: 820-28. Schvarcz R, Glauman H, Weiland O, Norkrans G, Wejstal R, Fryden A. Interferon alpha2b treatment of chronic posttransfusion non-A, non-B/C hepatitis: long-term outcome and effect of increased interferon doses in non-responders. Scand J Infect Dis 1991; 23: 413-20.
841
Copyright © European Association for the Study of the Liver 1996 Journal of Hepatology ISSN 0168-8278
A histopathological study of the effects of 6-month versus 12-month interferon -2b therapy in chronic hepatitis C Marianne Ziol, Jeanne Tran Van Nhieu, Franqoise Roudot-Thoraval, Jean-Michel M6treau, Yves Deugnier, Daniel Dhumeaux and Elie Serge Zafrani Services d'Anatomie et de Cytologie Pathologiques et d'H~patologie, H~pital Henri Mondor, Crdteil, France
Background~Aims: Interferon therapy has been shown to have beneficial effects in chronic hepatitis C, but the optimal duration of treatment has not been clearly defined. The aims of this study were: (a) to perform a detailed histological comparison of the effects of a 6-month and a 12-month treatment using the Knodeil score as well as a recently proposed grid of analysis, (b) to determine possible histological predictive factors of response to therapy, and (c) to attempt to relate histological and biochemical modifications. Methods: Liver biopsies obtained before and 18 months after beginning of treatment were therefore compared in 26 patients treated for 6 months, and in 34 patients treated for 12 months. Results: Six months of treatment induced a significant decrease in periportal (p=0.02) and intralobular (p=0.004) hepatocyte necrosis. The same items were improved in the 12-month-treated patients but, in addition, portal inflammation (p=0.01), bile duct lesions (p=0.03), lymphoid aggregates
(p=0.002) and fibrosis (p=0.008) were also improved, according to the Knodell score. Low scores for fibrosis, steatosis and cholangiolar proliferation on the pretreatment liver biopsy could be considered predictive factors for alanine aminotransferase normalization at 6 months. There was no relationship between biochemical response and modification of fibrosis. Conclusion: Our results suggest that: (a) a decrease in fibrosis might be detected only after a 12month interferon treatment, and (b) initial fibrosis, cholangiolar proliferation and steatosis are predictive of a lack of biochemical response. The absence of a relation between biochemical response and evolution of fibrosis implies that the evaluation of treatments in chronic hepatitis C should always include a detailed histopathological study.
EPATmS C virus (HCV) infection leads to the development of chronic hepatitis in approximately 70% of cases (1). In patients with chronic hepatitis, cirrhosis occurs in 20 to 30% of the cases (1), and is associated with a significant risk of hepatocellular carcinoma development (2). Several multicenter randomized trials have demonstrated the beneficial effects of ix-interferon (IFN) therapy in chronic hepatitis C (3). Six months of ther-
apy has been shown to induce both reduction in alanine aminotransferase (ALT) levels and improvement of histological periportal and lobular necroinfiammatory lesions in half of the cases (4). However, the high prevalence of post-treatment relapse led to studies of the effects of more prolonged or higher doses of IFN therapy. In a recent study (5,6), it has been shown that 12 months of IFN treatment induced a trend to a decrease in post-treatment biochemical relapse, in comparison with a 6-month treatment. Other studies comparing 6-month and 12-month IFN treatment have reported the same results for biochemical response (7,8). In addition, a 12-month treatment compared to a placebo treatment has been reported to decrease all Knodell's histological items including
H
Received 27 September 1995; revised 20 January," accepted 25 February 1996
Correspondence: E. S. Zafrani, Service d'Anatomie et de Cytologie Pathologiques, H6pital Henri Mondor, 51 Avenue du Mar6chal de Lattre de Tassigny, 94 010 Cr6teil cedex, France. Tel. 33-1-49 81 27 32. Fax 33-1-49 81 27 33.
Key words: Chronic hepatitis C; Fibrosis; Hepatitis C virus; Histological score; Interferon ~ therapy.
833
M. Ziol et al.
TABLE 1 Semi-quantitative grading system of liver lesions in chronic hepatitis C (derived from the METAVIR Study Group propositions) Fibrosis (evaluated on picro-sirius-red-stained sections): 0: no fibrosis; 1: stellate enlargement of portal tracts without septa formation; 2: enlargement of portal tracts with rare septa formation; 3: numerous septa without cirrhosis; 4: cirrhosis. Portal inflammation and periportal hepatocyte necrosis: 0: absent; 1: mild; 2: moderate; 3: severe. Lymphoid aggregates in portal tracts: 0: none; 1: present in less than one third of portal tracts; 2: present in one third to two thirds of portal tracts; 3: present in more than two thirds of portal tracts. Lesions of inteflobular bile ducts: 0: none or rare; 1: presence of lymphocytes within duct epithelium or slight degenerative changes; 2: presence of primary biliary cirrhosis-like lesions. Cholangiolar proliferation: 0: none; 1: mild; 2: moderate; 3: marked. Intrasinusoidal lymphocyte aggregates: 0: absent or rare; 1: numerous; Intralobular hepatocyte necrosis: 0: absent or mild; 1: moderate; 2: severe. Steatosis: 0: absent; 1: involving less than 10% of the hepatocytes; 2: involving 10 to 30% of the hepatocytes; 3: involving more than 30% of the hepatocytes. Iron storage (evaluated on Perls' stained sections): 0: absent; 1: mild accumulation in hepatocytes or Kupffer cells; 2: moderate or marked accumulation in hepatocytes or Kupffer cells. Global activity: 1: no activity; 2: minimal activity; 3: moderate activity; 4: severe activity. Cirrhosis: 0: absent; 1: potentially present; 2: certain.
fibrosis score (9), and discrepancies between biochemical and histological responses have been underlined (10-12). The main purpose of the present work was to perform a detailed histological study of the effects of the two types of treatment in the trial cited above (6). The other aims of the work were also: (a) to report the initial lesions on pretreatment biopsies and to study their possible relationships; (b) to relate initial lesions to ALT modifications in order to determine possible histological predictive factors for response to IFN therapy; (c) to try to relate biochemical and histological modification after treatment. To address these questions, we used a recently published and evaluated grid of liver biopsy analysis proposed by the METAVIR cooperative study group (13). This grid semiquantitatively grades all elementary lesions possibly encountered in chronic hepatitis C, and not only necroinflammatory changes and fibrosis as in the 834
Knodell index (14). It also gives a global score of hepatitis activity separated from fibrosis score. This grid has the advantage of being more precise than the Knodell index (14) for detailed histopathological studies. However, we also used the Knodell index to compare our findings with those published in other reports.
Patients and Methods The 60 patients who had sufficient pre- and post-treatment biopsy material for detailed histopathological study were selected from the 108 patients with chronic active hepatitis C, from 10 French hospitals enrolled in the randomized control trial comparing a 6-month and a 12-month IFN regimen (6). The semiquantitative liver biopsy analysis according to the METAVIR grid using 13 parameters required large biopsy samples. This condition was achieved in both pre- and post-treatment biopsies in only 60 patients.
Histological evaluation of lFN therapy TABLE 2 Distribution of initial lesions according to the METAVIR scoring system on pre-treatment biopsies Score for each item n=60
0
1
2
3
4
Fibrosis Portal inflammation Periportal hepatocyte necrosis Portal lymphoid aggregates Bile duct epithelium lesions Cholangiolar proliferation Intrasinusoidal lymphocyte aggregates Intralobular hepatocyte necrosis Steatosis Iron storage* Global activity Cirrhosis
0 0 0 17 30 9 8 10 15 39 0 40
21 9 16 25 25 32 44 41 13 10 16 3
15 46 37 16 4 17 8 9 15 5 37 17
7 5 7 2
17
2
17 7
* n=54 because of 6 deficient Perls' staining.
The details of the trial have previously been reported (6). All patients gave written formal consent and the study was approved by the ethics committee, at Hospital Henri Mondor, Creteil, France. In the 60 patients, diagnosis was based on an elevation of ALT level for more than 6 months, and on positivity for hepatitis C virus antibodies. Chronic active hepatitis with or without cirrhosis was histologically proven on a liver biopsy done within 6 months before inclusion. Non-inclusion criteria were: (a) signs of severe liver failure or portal hypertension; (b) presence of antiHIV antibodies; (c) antiviral or corticosteroid therapy within 6 months prior to inclusion; (d) alcoholism or drug addiction within 2 years before inclusion; (e) hemoglobin rate below 11 g/l, platelet count below 70000/ktl, leucocytes count below 3000/ktl. Determination of gamma-glutamyl transpeptidase (?GT) and alkaline phosphatase (AP) levels was performed within the 2 weeks before inclusion. The patients were randomly assigned to either 6month or 12-month IFN (Intron-A; Schering-Plough) therapy. Fifty-two patients received 3 MIU IFN subcutaneously three times a week for 6 months, whereas 56 were given 12-month IFN therapy consisting of 3 MIU three times a week for 6 months, 2 MIU three times a week for the following 3 months, and 1 MIU three times a week for the last 3 months. Groups I and II were similar in terms of clinical, biological and histological data at the time of starting therapy. In addition, there was no significant difference between the two groups of patients as far as dropout rate and side effects were concerned. All patients were followed up every month for 18 months after the beginning of therapy. The biochemical response to IFN was defined as complete when the serum ALT level was normal at the end of ther-
apy. It was defined as sustained when serum ALT level was still normal 6 months after the end of therapy. Other patients were considered non-responders. All patients underwent two liver biopsies: the first one was performed during the 6 months before beginning therapy, and the second, 18 months after beginning therapy. Liver biopsies were fixed in Bouin's liquid, then embedded in paraffin. Five-micrometerthick sections were stained with hematoxylin and eosin, picro-sirius red for collagen and Perls' technique for hemosiderin. Sections were examined under code without knowledge of the type of treatment or of the chronologic order of the biopsies. Twenty-six of the 60 patients selected in the present study received a 6-month treatment (Group I) and the 34 others received a 12-month treatment (Group II). The 13 items that were graded are derived from the METAVIR Cooperative Study Group proposition (13) and are characterized by either continuous or categorical variables. This liver biopsy grading system is described in Table 1. Conclusive items include global evaluation of hepatitis activity and of fibrosis. All biopsy specimens were also graded according to the Knodell score. Biopsies were separately examined and graded by two pathologists (E.S.Z and M.Z), and then simultaneously reviewed using a double-pipe microscope in order to analyze discrepancies and to reach a consensus.
Statistical analysis Relationships between the initial histological lesions were evaluated by the non-parametric Spearman rank correlation for each item. The relationship between initial bile duct lesions and presence of follicles in portal tracts on one hand, and pretreatment AP and 835
M. Ziol et al.
TABLE 3 Changes in histological score before and 18 months after the beginning of therapy in 6-month-treated patients (group I) and 12-month-treated patients (group II) Group I: n=26 METAVIR grid items: Fibrosis Portal inflammation Periportal bepatocyte necrosis Portal lymphoid aggregates Bile duct lesions Cholangiolar proliferation Intrasinusoidal lymphocytes Intralobular hepatocyte necrosis Steatosis Iron storage Global activity Cirrhosis Knodell items: Fibrosis Periportal necrosis Intralobular necrosis Portal inflammation Group II: n=34 METAVIR grid items: Fibrosis Portal inflammation Periportal hepatocytes necrosis Portal lymphoid aggregate Bile duct lesions Cholangiolar proliferation Intrasinusoidal lymphocytes Intralobular hepatocyte necrosis Steatosis Iron storage Global activity Cirrhosis Knodell items: Fibrosis Periportal necrosis Intralobular necrosis Portal inflammation
Worse (%)
No change (%)
Better (%)
p*
11 15 12 19 27 8 8 15 8 5 12 12
62 46 46 42 42 69 38 46 77 77 42 81
27 39 42 39 31 23 54 39 15 18 46 7
NS NS p=0.02 NS NS NS p--0.002 NS (p--0.07) NS NS p--0.01 NS
8 12 4 12
77 50 54 61
15 38 42 27
NS NS p=0.004 NS
Worse(%)
No change (%)
Better (%)
p*
12 9 12 6 12 18 18 9 21 17 9 9
53 53 41 53 47 68 41 50 44 69 41 79
35 38 47 41 41 14 41 41 35 14 50 12
NS (p=0.1) p=0.01 p=0.006 p=0.002 p=0.03 NS p=0.03 p=0.006 NS NS p=0.001 NS
6 12 12 9
62 41 36 54
32 47 52 37
p=0.008 p=0.0037 p=0.005 p=0.007
*: Wilcoxon signed-rank test.
qa3T levels on the other hand, were analyzed using the Chi-square test or Fisher exact test. Pre- and posttreatment grades for each histological item, including those of the Knodell score, were compared using the Wilcoxon non-parametric test for paired data in groups I and II. Histological and biochemical modifications, 18 months after the beginning of therapy, were compared using a Chi-square test in all patients. Lesions were considered improved when the difference between pre- and post-treatment scores was superior or equal to 1. In order to find histological predictive factors for biochemical response, histological scores on pretreatment biopsies were compared by Chi-square test in patients who either completely responded or failed to respond 6 months after the
836
beginning of therapy. At that time, patients had received the same IFN doses. A multivariate analysis using a stepwise logistic regression model was further performed in order to find factors independently related to biochemical response. Assessment of histological criteria predictive of amelioration of activity and fibrosis was done using univariate and multivariate analyses.
Results Pre-treatment histological lesions Distribution of the histological lesions on initial biopsies, graded according to the METAVIR evaluation system, is reported in Table 2. In summary, cirrhosis was observed in 17 patients (28%) and global evalua-
Histological evaluation of lFN therapy TABLE 4 Histological response in relation to sustained biochemical response 18 months after beginning of therapy in 6-month and 12-month treated patients Histological improvement METAVIR grid items: Fibrosis Portal inflammation Portal lymphoid aggregates Bile duct lesions** Periportal hepatocyte necrosis Intralobular necrosis Intrasinusoidal lymphocytes Global activity Knodell items: Fibrosis Periportal hepatocyte necrosis Intralobular hepatocyte necrosis Portal inflammation
Sustained responders n=20
Non-sustained responders n--40
Number of patients n=60
Significance*
8(40%) 14(70%) 13(65 %) 10(50%) 15(75%) 13(65 %) 11(55%) 16(80%)
11(28%) 9(23%) 11(28%) 10(25%) 12(30%) 11(28%) 17(43 %) 13(33%)
19(32%) 23(38%) 24(40%) 20(33%) 27(45%) 24(40%) 28(47%) 29(48%)
NS p=0.0004 p=0.005 p=0.04 p=0.01 p--0.005 NS p=0.0005
5(25%) 15(75%) 15(75%) 13(65%)
10(25%) 11(28%) 14(35%) 7(18%)
15(25%) 26(43%) 29(48%) 20(33%)
NS p=0.0005 p=0.0035 p--0.0002
*: Chi-square test. **: Sustained responders n=19, non-sustained responders n=39.
tion of hepatitis activity showed mild or moderate activity in 53 patients (88%). Portal lymphoid aggregates were found in 43 patients (72%), bile duct damage in 29 (49%), with primary biliary cirrhosis-like lesions of the interlobular bile ducts in only four patients (6%). Steatosis was found in 45 patients (75%). The grades of each item on initial biopsies were interrelated, excluding the global hepatitis activity score and the presence or not of cirrhosis, which are already integrated items. Fibrosis was strongly related to periportal necrosis (p=0.0001) and to cholangiolar proliferation (p=0.0001). Portal inflammation, intralobular and periportal necrosis, and intrasinusoidal lymphocytes were all related (p<0.001 for each item to the other). Bile duct damage was related to portal inflammation (p=0.0001), to intralobular necrosis (p=0.001) and to the presence of portal lymphoid aggregates (p=0.001). Other items, such as steatosis and iron storage, did not appear to be related to other lesions. Study of the relationships between interlobular bile duct damage and AP and ~GT levels showed that primary biliary cirrhosis-like lesions (grade 2) of the interlobular bile ducts were associated with elevated ~3T (p=0.052). Grade 0 or 1 lesions of the interlobular bile ducts were not related to AP or ~3T levels.
between the two groups. Six of the 13 responders (46%) in group I had sustained response, whereas 14 of the 18 responders (78%) had sustained response in group II. The difference between the two groups approached but did not reach significance (p=0.07).
Biochemical response to treatment Thirteen of the 26 patients (50%) had a complete response in the 6-month treated group and 18 of 34 (53%) had a complete response in the 12-month treated group. There was no significant difference
Histological predictive factors for short-term biochemical response A relationship between initial histological lesions and biochemical response 6 months after beginning therapy was also searched for. At 6 months, group I and
Histological response to treatment The main results of the histological evaluation performed on the 60 pre- and post-treatment biopsies are indicated in Table 3. Comparison of pre- and posttreatment scores in the 6-month-treated group showed a significant improvement for periportal hepatocyte necrosis evaluated by the METAVIR scoring system, for intralobular necrosis evaluated by the Knodell score, for intrasinusoidal lymphocyte score and for global hepatitis activity index according to METAVIR. In the 12-month-treated group, the same items improved significantly. In addition, portal inflammation evaluated by the two grading systems as well as bile duct lesions and lymphoid aggregates were also significantly decreased. It must be pointed out that patients who were treated for 12 months had a significant improvement in their fibrosis score according to Knodell; when the METAVIR scoring system was used, there was a trend to decreased fibrosis, but this decrease did not reach significance (Table 3).
837
M. Ziol et al.
group II patients had received the same IFN doses. High scores for fibrosis and cholangiolar proliferation appeared to be predictive of persistence of elevated ALT at 6 months. Indeed, 16 of 34 (47%) patients who initially had a high score (grade 3 or 4 according to Knodell) of fibrosis were responders vs 20 of 26 (77%) who initially had a low score (grade 0 or 1 according to Knodell) of fibrosis (p=0.02); 8 of 19 patients (42%) who had a high score (grades 2 or 3 according to METAVIR) of cholangiolar proliferation were responders vs 28 of 41 (68%) who initially had a low score (grades 0 or 1) of cholangiolar proliferation (p=0.05). In addition, the presence of steatosis was also predictive for a lack of short-term biochemical response, since 23 of 45 patients (51%) who had minimal, moderate or marked steatosis were responders vs 13 of 15 (87%) who had no steatosis (p=0.01). Multivariate analysis, using a stepwise logistic regression model, showed that steatosis (p=0,01) and cholangiolar proliferation (p=0,05) were the two factors independently related to short-term biochemical response.
Histological predictive factors for activity and fibrosis amelioration Univariate analysis showed that: (a) an initial high score for periportal hepatocyte necrosis (p=0.01) and a low score for steatosis (p=0.02) were significantly related to amelioration of activity according to METAVIR; there was also a trend for an initial high score for portal lymphoid aggregates (p=0.07); (b) an initial low score for cholangiolar proliferation (p=0.01) and a high score for fibrosis (p=0.03) were predictive of fibrosis amelioration. Multivariate analysis showed that: (a) an initial high score for periportal hepatocyte necrosis (p=0.03) and a low score for fibrosis (p=0.01) were independently and significantly related to amelioration of activity according to METAVIR; (b) an initial low score for cholangiolar proliferation (p=0.01) and a high score for fibrosis (p<0.001) were predictive of fibrosis amelioration.
Relationships between biochemical and histological response Relationships between histological improvement and sustained biochemical response at month 18 are shown in Table 4, only for histological items improved by therapy. Rates of histological improvement were significantly higher in patients who were responders at 18 months by comparison with nonresponders, for portal inflammation, periportal and intralobular hepatocyte necrosis, bile duct lesions and
838
global evaluation of hepatitis activity score. In contrast, improvement in fibrosis according either to METAVIR or to Knodell, did not appear to be related to sustained biochemical response.
Discussion In order to evaluate pre- and post-treatment liver biopsies in chronic hepatitis C treated by interferon, two scoring systems (i.e., the Knodell score and the METAVIR grid) were used in the present work. The Knodell score was used to compare our findings to other published reports in which it is usually employed (9,11). However, the Knodell score only evaluates hepatocyte necrosis, portal inflammation and fibrosis. The METAVIR scoring system, for which an analysis of intra- and inter-observer variations has been recently published (13), has the advantage of semi-quantitatively grading all elementary lesions possibly encountered in chronic hepatitis C (e.g., bile duct lesions, presence of portal lymphoid aggregates, steatosis) and giving a global score of hepatitis activity separated from fibrosis score. Using this METAVIR evaluation system, the interrelationships between numerous histological lesions of chronic hepatitis C and the relationships between histological and biochemical responses can be studied. In addition, modifications of liver biopsy findings, when different modalities of treatment are compared, can be evaluated on more histological items. Finally, possible histological predictive factors of biochemical response may be determined. The prevalences of each histopathological lesion on pre-treatment biopsies are close to those previously reported in chronic hepatitis C (15-18). In summary, chronic hepatitis is mildly or moderately active in 88% of the cases, steatosis of more than 10% of the hepatocytes is observed in 75% of the liver biopsies, the presence of at least one lymphoid aggregate is noted in 72% of the cases, and mild or severe bile duct damage can be seen in 49% of the cases. The discrepancies in the determination of the prevalence of bile duct damage, varying from 91% (16) to 31% (17), are probably due to an absence of consensus for the definition of this lesion. High prevalences could be explained by the fact that the presence of occasional inflammatory cells in the interlobular bile duct seems to be considered indicative of bile duct damage by some authors (16). In our study, bile duct lesions were observed in 49% of the cases but only four of the 60 patients (6%) had severe damage resembling the florid lesion of primary biliary cirrhosis. The study of the interrelationships between pretreatment histological lesions showed that periportal
Histological evaluation of lFN therapy
and intralobular hepatocyte necrosis were related to portal and intralobular lymphocytic infiltration and that periportal necrosis was strongly linked to fibrosis. These relationships may be explained by events responsible for liver fibrogenesis, e.g., release of mitogens by necrotic hepatocytes, subendothelial matrix degradation, production of cytokines by platelets and macrophages, and autocrine and paracrine stimulation of Ito cells that produce fibrosis components (19). We have also noted that bile duct damage was related to portal lymphoid aggregates and that these two lesions were also related to intralobular hepatocyte necrosis. Lau et al. (20) reported similar findings and observed a relationship between such lesions and high levels of viremia. Since HCV virus cannot be immunohistochemically detected in the bile duct of portal tracts (21), an immune-mediated mechanism has been suggested for such lesions. In addition, the relation of lymphoid aggregates to bile duct damage is of interest, since these lesions may also be associated with the genotype lb of virus C (22). The present histological evaluation demonstrated better effects of a 12-month than a 6-month IFN treatment on some histological liver lesions. In the 6month-treated patients, periportal necrosis, intrasinusoidal lymphocytic infiltration and global activity score were significantly decreased. In the 12-monthtreated patients, the same histological lesions were also improved, and, in addition, portal inflammation, bile duct lesions, and fibrosis evaluated by the Knodell score were significantly reduced. Changes in fibrosis evaluated by the METAVIR grid showed a trend to decrease that did not reach significance. Such a difference in statistical significance when the two systems of histological evaluation are used may be explained by the fact that the Knodell score could be more sensitive in revealing improvement in fibrosis after IFN treatment: indeed, in the Knodell index there are less categories for fibrosis than in the METAVIR grid, and the presence of rare fibrous septa is not differentiated from the presence of numerous septa as in METAVIR scoring system. It is therefore possible to hypothesize that a 12month IFN treatment is probably able to slightly ameliorate fibrosis, but that this amelioration, which does not appear significant when multiple categories are defined in a scoring system, can be statistically revealed only when a more discontinuous score is used. Improvement of necrosis and, with a lower prevalence, of portal inflammation, is a well-described beneficial effect of IFN therapy in the literature (11,
23-26). The results of the present study, in the 6month as well as in the 12-month treated patients, are in agreement with these findings. However, significant reduction of fibrosis has rarely been reported. One report studying 59 patients with chronic non-A, non-B hepatitis receiving IFN or placebo for 6 months (27) has shown a significant decrease of total liver collagen in IFN-treated patients using a quantitative histochemical and morphometrical evaluation of collagen in liver biopsies, but this decrease of liver collagen was not reflected by a reduction of the fibrosis score evaluated according to Knodell. In two other series of patients treated with IFN for 12 months, an improvement in the total Knodell score and in each of its components, including fibrosis, has been reported on liver biopsies performed immediately at the end of the treatment (9,10). Our results suggest that histological improvement can be sustained 6 months after the end of treatment. An attempt to determine which histological lesions might or might not be related to ALT normalization was made in this study. Whereas amelioration of necroinflammatory changes as well as bile duct damage appeared to be related to ALT normalization at 18 months after beginning of therapy, no relationship was found between fibrosis and ALT levels. Indeed, 10 of 40 non-responder patients (25%) and 5 of 20 responders (25%) showed improvement in the fibrosis item of the Knodell score; 8 of 20 responders (40%) vs 11 of 40 non-responders (28%) had amelioration of fibrosis according to the METAVIR grading system. Such a finding might suggest that the effect of IFN on liver fibrosis could be independent of its antiviral activity that probably reduces hepatocyte necrosis. This hypothesis is supported by recent studies that postulate a direct inhibitory effect of IFN on fibrogenesis. Indeed, Mallat et al. (28) demonstrated a growth inhibition effect and antifibrogenic effects of tx- and T-interferon on cultured human Ito cells, and Castilla et al. (29) showed normalization of serum procollagen type III peptide and of TGF [31mRNA levels in liver tissue in eight patients treated with IFN for 1 year. High scores of fibrosis and cholangiolar proliferation, as well as steatosis appeared to be predictive histological factors for a lack of ALT normalization. Fibrosis has already been identified as a predictive factor for biochemical response (30). It is not surprising that cholangiolar proliferation was also predictive of such a lack of biochemical response, since we showed that this item was strongly related to fibrosis. Steatosis, which is not significantly modified by IFN therapy, has already been implicated in biochemical 839
M. Ziol et al.
response by Schvarcz et al. (31), who suggested that steatosis could mask biochemical response by being directly responsible for raised ALT levels. The search for histological predictive factors for amelioration of fibrosis and activity suggests that modifications under therapy of these two major parameters might be independent, since initial low scores of fibrosis were related to activity amelioration, whereas initial high scores of fibrosis were related to fibrosis amelioration. In conclusion, the present histopathological study, using a detailed scoring system such as proposed by the METAVIR Group, shows that: (a) in HCV-infected patients, there are strong relationships between initial elementary lesions, such as between nodular lymphoid aggregates and bile duct damage, which might be related to HCV viremia levels and to HCV genttypes; in addition, high scores of fibrosis, cholangiolar proliferation and steatosis seem to be predictive for absence of ALT normalization; (b) a 12-month treatment by IFN not only allows the maintenance of the beneficial histological effects noted at 6 months, but might also induce a slight improvement in fibrosis; it also seems to ameliorate portal inflammation, portal lymphoid aggregates, bile duct lesions, and intrasinusoidal lymphocytes scores; (c) the absence of relationship between biochemical response and modification of fibrosis implies that the evaluation of IFN effects should always include, not only the determination of biochemical parameters, but also a detailed study of comparative histopathological findings.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Acknowledgements The authors thank the members of the French Group for the Study of Chronic Hepatitis NANB/C Treatment: Y. Calmus, R. Poupon (H6pital Saint Antoine, Paris), H. Jouanolle (H6pital Pontchaillou, Rennes), J.P. Miguet (H6pital Jean Minjoz, Besanqon), C Br6chot (H6pital La6nnec, Paris), J.L. Payen, P. Cal6s (H6pital de Purpan, Toulouse), A. Tran (H6pital de l'Archet, Nice), M. Beaugrand (H6pital Jean Verdier, Bondy), J.P. Capron (CHU Nord, Amiens), J.P. Zarski (H6pital de la Tronche, Grenoble), M. Doffo~l (Hotel Dieu, Strasbourg), E. A. Pariente (H6pital la Source, Orl6ans), and C. Lemonnier (Schering Plough, Levallois-Perret).
12.
13.
14.
15.
16.
References 1. Giral P, Serfaty L, Loria A, Poupon R. Hepatite C. Rev Med Intern 1994; 15: 487-93. 2. Tsukuma H, Hiyama T, Tanaka S, Nakao M, Yabunchi T, Kitamira T, Nakanishi K, Fujimoto I, Inoue A, Yamasaki H, Kawashima T. Risk factors for hepatocellular carcinoma
840
17.
among patients with chronic liver disease. N Engl J Med 1993; 328: 1797-801. Pawlotsky JM, Marcellin P, Trepo C, Dhumeaux D. Traitement des hepatites virales C. Gastroenterol Clin Biol 1994; 18: 362-73. Hoofnagle JH. Therapy of chronic hepatitis B and C. In: Miguet JP, Dhumeaux D, eds. Progress in Hepatology 93. Paris: John Libbey Eurotext, 1993: 89-97. Metreau JM and the French Group for the study of NANB/C chronic hepatitis treatment. Results of long-term interferon treatment in non-A, non-B/C chronic active hepatitis. Gut 1993; suppl: S112-S113. Jouet P, Roudot-Thoraval F, Dhumeaux D, Metreau JM, and the French Group for the Study of NANB/C Chronic Hepatitis Treatment. Comparative efficacy of Interferon alpha in cirrhotic and noncirrhotic patients with non-A non-B, C hepatitis. Gastroenterology 1994; 106: 686-90. Milella M, Santantonio M, Monno L, Pastore G. Comparison of short vs long term treatment with recombinant alpha interferon in chronic hepatitis C (Abstract). In: Abstracts of the 1993 International Symposium on Viral Hepatitis and Liver Disease, Tokyo 1993: 245. Chemello L, Pontisso P, Rose KA, Simmonds P, Boneui P, Cavalletto L. The long-term response (LTR) to interferon alpha-2a(IFN-2a) in chronic hepatitis C is influenced by dose and duration of treatment and by the HCV serotype (abstract). J Hepatol 1993; 18(suppl 1): SI0-S11. Camps J, Castilla A, Ruiz J, Civeira MP, Prieto J. Randomised trial of lymphoblastoid ct-interferon in chronic hepatitis C. J Hepatol 1993; 17: 390-6. Reichard O, Glaumann H, Norkrans G, Wejstal R, Fryden A, Schvarcz R, Weiland O. Histological outcome in patients with chronic hepatitis C given a 60-week interferon alfa-2b treatment course. Liver 1994; 14: 169-74. Causse X, Godinot H, Chevallier M, Chossegros P, Zoulim E Ouzan D, Heyraud JP, Fontanges T, Albrecht J, Meschievitz C, Trepo C. Comparison of 1 or 3 MU of interferon alfa2b and placebo in patients with chronic non-A, non-B hepatitis. Gastroenterology 1991:101: 497-502. Marcellin P, Pouteau M, Duchatelle V, Degott C, Degos F, Boyer N, Erlinger S, Benhamou JP. Post-treatment histologic outcome in patients with chronic hepatitis C treated with recombinant alpha interferon. In: AASLD Abstracts. Hepatology 1993; 18: 243A. The French METAVIR Cooperative Study Group. Intra- and interobserver variations in liver biopsy interpretation in patients with chronic hepatitis C. Hepatology 1994; 20: 15-20. Knodell RG, Ishak KG, Black WC, Chen TS, Craig R, Kaplowitz N, Kiernan TW, Wollman J. Formulation and application of a numerical scoring system for assessing histological activity in asymptomatic chronic active hepatitis. Hepatology 1981; 1: 431-5. Czaja AJ, Carpenter HA. Sensitivity, specificity and predictability of biopsy interpretations in chronic hepatitis. Gastroenterology 1993; 105: 1824-32. Bach N, Thung SN, Schaffner E The histological features of chronic hepatitis C and autoimmune chronic hepatitis: a comparative analysis. Hepatology 1992; 15: 572-7. Lefkowitch JH, Schiff ER, Davis GL, Perillo RP, Lindsay K, Bodenheimer HC, Balart LA, Ortego TJ, Payne J, Dienstag JL, Gibas A, Jacobson IM, Tamburro CH, Carey W, O'Brien C, Sampliner R, Van Thiel DH, Feit D, Albrecht J, Me-
Histological evaluation of lFN therapy
18.
19.
20.
21.
22.
23.
24.
schievitz C, Sabghvi B, Vaughan RD. Pathological diagnosis of chronic hepatitis C. A multicenter comparative study with chronic hepatitis B. Gastroenterology 1993; 104: 595-603. Scheuer PJ, Ashrafzadeh P, Sherlock S, Brown D, Dusheiko GM. The pathology of hepatitis C. Hepatology 1992; 15: 567-71. Friedman SL. The cellular basis of hepatic fibrosis. Mechanisms and treatment strategies. N Engl J Med 1993; 328: 1828-35. Lau JYN, Davis GL, Kniffen J, Qian KP, Urdea MS, Chan CS, Mizokami M, Neuwald PD, Wilber CW. Significance of serum hepatitis C virus RNA levels on chronic hepatitis C. Lancet 1993: 341: 1501--4. Krawczynski K, Beach MJ, Bradley DW. Hepatitis C virus antigen in hepatocytes: immunomorphologic detection and identification. Gastroenterology 1992; 103: 622-9. Tran Van Nhieu J, Roudot-Thoraval F, Pawlotsky JM, Pellet C, Dhumeaux D, Zafrani ES. Relationship between genttypes of C virus and histologic liver lesions in chronic hepatitis C infection. In: AASLD Abstracts. Hepatology 1994; 20: 247A. David E, Puci A, Palladin D, Saracco G, Garello E, Pintus C, Roca G, Chiandussi L, Solinas A, Mollo E Rizetto M, Mazzuco G, Verme G. Histologic changes in liver biopsy specimens produced by recombinant interferon u-2b therapy for chronic non-A, non-B viral hepatitis. Am J Clin Pathol 1992; 98: 397--401. Davis GL, Balart LA, Schiff ER, Lindsay K, Bodenheimer HC, Perrillo RP, Garey W, Jacobson IM, Payne J, Dienstag JL, VanThiel DH, Tamburro C, Lefkowitch J, Albrecht J, Meschievitz C, Ortego TJ, Gibas A, and The Hepatitis Inter-
25.
26.
27.
28.
29.
30.
31.
national Therapy Group. Treatment of chronic hepatitis C recombinant interferon alfa. N Engl J Med 1989; 321: 1501-6. Di Bisceglie AM, Martin P, Kassianides C, Lisker-Melman M, Murray L, Waggoner J, Goodman Z, Banks SM, Hoofnagle JH. Recombinant interferon alfa therapy for chronic hepatitis C. N Engl J Med 1989; 321: 1506-10. Omata M, Ito Y, Yokosuka O, Imazeki F, Tagawa M, Takano S, Hosoda K, Tada M, Ohto M, Ito K, Okuda K. Randomized, double-blind, placebo-controlled trial of eight-week course of recombinant ct-interferon for chronic non-A, non-B hepatitis. Dig Dis Sci 1991; 36: 1217-22. Manabe N, Chevallier M, Chossegros P, Causse X, Guerret S, Trepo C, Grimaud JA. Interferon-~2b therapy reduces liver fibrosis in chronic non-A, non-B hepatitis. A quantitative histological evaluation. Hepatology 1993; 18: 1344-9. Mallat A, Preaux AM, Blazejewski S, Rosenbaum J, Dhumeaux D, Mavier P. Interferon alpha and gamma inhibit proliferation and collagen synthesis of human Ito cells in culture. Hepatology 1995; 21: 1003-10. Castilla A, Prieto J, Fausto N. Transforming growth factors B1 and o~ in chronic liver disease. N Engl J Med 1991; 324: 933--40. Pagliario L, Craxi A, Camaa C, Tine F, Di Marco V, Lo Iacono O, Almaso P. Interferon-or for chronic hepatitis C: an analysis of pretreatment clinical predictors of response. Hepatology 1994; 19: 820-28. Schvarcz R, Glauman H, Weiland O, Norkrans G, Wejstal R, Fryden A. Interferon alpha2b treatment of chronic posttransfusion non-A, non-B/C hepatitis: long-term outcome and effect of increased interferon doses in non-responders. Scand J Infect Dis 1991; 23: 413-20.
841