Autoimmune, Cholestatic and Biliary Disease
Prevalence of sclerosing cholangitis in adults with autoimmune hepatitis: A prospective magnetic resonance imaging and histological study
Article first published online: 17 APR 2009
Copyright © 2009 American Association for the Study of Liver Diseases
Volume 50, Issue 2, pages 528–537, August 2009
How to Cite
Lewin, M., Vilgrain, V., Ozenne, V., Lemoine, M., Wendum, D., Paradis, V., Ziol, M., Arrivé, L., Beaugrand, M., Poupon, R., Valla, D., Chazouillères, O. and Corpechot, C. (2009), Prevalence of sclerosing cholangitis in adults with autoimmune hepatitis: A prospective magnetic resonance imaging and histological study. Hepatology, 50: 528–537. doi: 10.1002/hep.23024
- Issue published online: 29 JUL 2009
- Article first published online: 17 APR 2009
- Accepted manuscript online: 17 APR 2009 12:00AM EST
- Manuscript Accepted: 8 APR 2009
- Manuscript Received: 2 FEB 2009
The development of sclerosing cholangitis (SC) is observed in up to 50% of children followed up for autoimmune hepatitis (AIH). In adults, the prevalence is less known, although a recent study found evidence of SC in 10% of AIH patients using magnetic resonance cholangiopancreatography (MRCP). The aim of this study was to assess prospectively the prevalence of SC in adults with AIH. Fifty-nine consecutive patients with AIH diagnosed according to International Autoimmune Hepatitis Group score (women, 71%; mean age, 48 years; cirrhosis, 23%) underwent both MRCP and percutaneous liver biopsy. Twenty-seven patients with cirrhosis of nonbiliary or non-autoimmune etiology served as controls. Fourteen AIH patients (24%) showed mild MRCP abnormalities of intrahepatic bile ducts (IHBDs). None had abnormal common bile duct or convincing evidence of SC on MRCP or biopsy. A diagnosis of overlapping SC was nevertheless retained in one patient with MRCP abnormalities who subsequently developed symptomatic cholestasis despite corticosteroid therapy. Fibrosis score was the only independent parameter associated with bile duct abnormalities on MRCP (odds ratio 2.4; 95% confidence interval 1.4-4.7) and the percentage of patients with IHBD MRCP abnormalities was not different among F3-F4 AIH patients (n = 24) and cirrhotic controls (46% versus 59%; NS). Conclusion: In this cohort of adult patients with AIH, the prevalence of SC was much lower than previously reported (1.7%). Mild MRCP abnormalities of IHBD were seen in a quarter of patients, but these abnormalities resulted from hepatic fibrosis and not SC. In the absence of cholestatic presentation, MRCP screening does not seem justified in adult-onset AIH. (HEPATOLOGY 2009.)
Autoimmune hepatitis (AIH) is an idiopathic chronic hepatitis affecting predominantly women at any age, characterized by liver necroinflammation and presence of autoantibodies as well as elevated immunoglobulin levels in the serum. Because specific markers are lacking, the diagnosis must be based on a combination of suggestive features and the exclusion of other causes of chronic liver disease.1 AIH has a very wide spectrum. It is now well-established that a few patients can exhibit, either simultaneously or sequentially, some of the features of primary biliary cirrhosis (PBC) or of primary sclerosing cholangitis (PSC), two chronic cholestatic conditions of presumed autoimmune origin.2 The early recognition of these overlap syndromes is of importance because it may lead to appropriate treatment and follow-up. Whereas PBC diagnosis is facilitated by the detection of antimitochondrial antibodies in the serum, that of PSC largely depends on cholangiographic evidence of multiple strictures of the biliary tract.3–8
In children, it is known that AIH commonly exists concurrently with sclerosing cholangitis (SC). In a large pediatric series, half of the children with a definite diagnosis of AIH demonstrated cholangiographic features of SC at presentation.9 Sequential evolution of AIH toward SC has also been reported in children and young adults, mainly those who became resistant to immunosuppressive treatment or developed cholestasis.9, 10 This so-called autoimmune sclerosing cholangitis affects boys and girls equally and is more likely to progress to cirrhosis, liver transplantation, or death rather than AIH.11 In marked contrast, the prevalence of underlying SC in adult-onset AIH is poorly understood. Indeed, there is no available series of endoscopic retrograde cholangiopancreatography (ERCP) in adults with AIH. This gap should be bridged shortly now that magnetic resonance cholangiopancreatography (MRCP) has emerged as an alternative, noninvasive, and reliable method for the diagnosis of SC.12, 13 Abdalian et al.14 have recently reported the results of a routine screening by way of MRCP of 79 adult patients with AIH. In this study, evidence of large duct SC on MRCP was found in 10% of patients. However, this is the only such study so far, and these data have to be confirmed to state whether regular evaluation of biliary tract with MRCP should be performed in the management of adult-onset AIH.
In this context, we report here the results of a prospective evaluation of bile duct and liver abnormalities in unselected adult patients with AIH using both liver magnetic resonance imaging (MRI) with MRCP and histological examination of the liver. Because SC-like intrahepatic cholangiographic appearance can be seen in fibrotic livers in the absence of any biliary disease,15, 16 we included control cirrhotic patients with nonbiliary, non-autoimmune diseases.
Patients and Methods
A total of 59 consecutive patients with a diagnosis of AIH, including 40 women and 19 men (mean age 48 ± 19 years [range, 15-84 years]) seen from 2006 to 2008 at three University Hospitals of Ile-de-France (Saint-Antoine Hospital, Paris [n = 28]; Beaujon Hospital, Clichy [n = 19]; Jean Verdier Hospital, Bondy [n = 12]), were enrolled in the study. Inclusion criteria were a diagnosis of AIH defined on classical criteria and the updated International Autoimmune Hepatitis Group score ≥10 for corticosteroid-naïve patients and/or ≥12 for patients already treated17 and a liver biopsy and MRI of the liver and biliary tract performed at the time of inclusion. A maximum 3-month interval between the biopsy and MRI was tolerated. Exclusion criteria were fulminant form of AIH, established diagnosis of PSC (prior to the study) or PBC, history of liver transplantation, seropositivity for antimitochondrial antibodies, hepatitis B or hepatitis C virus infection, excessive alcohol consumption, decompensated cirrhosis, pregnancy, and absolute contraindication for MRI (pacemaker, metallic implants, claustrophobia).
An additional group of 27 patients with compensated and histologically proven cirrhosis of nonbiliary and non-autoimmune etiology (hepatitis C virus [n = 12]; alcohol [n = 9]; hepatitis B virus [n = 4]; nonalcoholic steatohepatitis [n = 2]) served as a control group. These patients, consisting of 12 women and 15 men (mean age, 58 ± 11 years [range, 40-79 years]), underwent an MRI of the liver and MRCP under the same conditions as the cases.
The study was approved by the institutional review board. All patients provided written informed consent to participate.
Biochemical and Immunological Markers.
Biochemical and immunological markers were recorded at the time of AIH diagnosis. The biochemical markers included serum activities of aspartate aminotransferase, alanine aminotransferase, gamma glutamyl transpeptidases, and alkaline phosphatase; serum levels of total bilirubin and albumin; platelet count; and prothrombin index. The immunological markers included serum levels of gamma globulins, immunoglobulin G and M, and antitissue antibody profile, including indirect immunofluorescence titers on rat tissue sections of antinuclear antibodies, anti–smooth muscle antibodies, M2 antimitochondrial antibodies, type 1 liver-kidney microsome antibodies, and antineutrophil cytoplasmic antibodies. Sera were considered positive at a dilution of ≥1:80 (antinuclear antibodies, anti–smooth muscle antibodies), ≥1:40 (antimitochondrial antibodies, type 1 liver-kidney microsome antibodies), or ≥1:20 (antineutrophil cytoplasmic antibodies). Soluble liver antigen antibodies were detected using an inhibition capture enzyme-linked immunosorbent assay based on the binding inhibition of labeled serum-positive for anti–soluble liver antigen.18
All liver biopsies were obtained by way of percutaneous puncture and were analyzed at each study center by an experienced liver pathologist blinded to the clinical and radiological data (D. W., Saint-Antoine; V. P, Beaujon; M. Z., Jean Verdier). All the specimens were studied according to a pre-established standardized procedure. Biopsies of <8 mm were excluded. Microscopic examination was performed on formalin-fixed, paraffin-embedded, 4-μm-thick liver sections that were stained with hematoxylin-eosin and Sirius red. A minimum of three sections per biopsy was examined. Cytokeratin 7 immunostaining was performed in all cases. Fibrosis and necroinflammatory activity indices were semiquantitatively evaluated according to the METAVIR scoring system.19 Briefly, fibrosis was staged on a 0-4 scale as follows: F0, no fibrosis; F1, portal fibrosis without septa; F2, portal fibrosis and few septa; F3, numerous septa without cirrhosis; F4, cirrhosis. Necroinflammatory activity was graded on a 0-3 scale as follows: A0, no activity, A1, mild activity, A2, moderate activity, A3, marked activity. The degree of activity was assessed through integration of the intensity of both piecemeal necrosis and lobular necrosis as described. The presence of plasmocyte infiltrate, rosette-forming hepatocytes, bridging necrosis, ductular proliferation, ductopenia, fibrous obliterative cholangitis, granulomatous cholangitis, and destructive/nondestructive lymphocytic cholangitis were noted. A dense periductal fibrosis associated with compression and atrophy of the biliary epithelium or a nodular scar replacing bile duct were designated fibrous obliterative cholangitis. A granulomatous inflammatory infiltrate associated with disruption of the biliary epithelium or destruction of the basement membrane or wall was designated granulomatous cholangitis. A mononuclear inflammatory infiltrate surrounding and penetrating the bile duct epithelium with or without associated epithelial damage and/or destruction of the basement membrane or wall was designated destructive/nondestructive lymphocytic cholangitis. The proportion of portal tracts without identifiable bile ducts after cytokeratin 7 immunostaining was calculated. A minimum of 10 portal tracts was required for the estimation of ductopenia. Interlobular bile duct paucity was defined as a ductopenia of > 50%.20
Magnetic Resonance Imaging.
Magnetic resonance examinations were obtained with a 1.5-T system (1.5-T Magnetom Maestro Class Symphony, Siemens Medical Systems, Erlangen, Germany and Gyroscan Intera, Philips Medical Systems, Best, the Netherlands) using a phased array abdominal coil for signal reception. MRCP included an axial breath-hold or free-breathing, half-Fourier acquisition single-shot turbo spin-echo (HASTE) sequence without fat suppression and a coronal, free-breathing, three-dimensional (3D) high-spatial-resolution fast spin-echo sequence. Postprocessing of the image data was performed to obtain maximum intensity projection images and multiplanar reformatted images. Before the 3D sequence was available, and in cases where the patient's respiratory pattern was too irregular, a coronal breath-hold single-shot rapid acquisition with relaxation enhancement sequence was used.
Liver examination included a respiratory-triggered fat-suppressed T2-weighted fast spin-echo sequence, T1-weighted in-phase and opposed-phase imaging, and a 3D fat-suppressed breath-hold T1-weighted gradient-echo sequence. In addition, a fat-suppressed gradient-echo T1-weighted MRI sequence was performed during the hepatic arterial, portal venous, and delayed phases (at 20 seconds, 60 seconds, and 5 minutes, respectively) after administration of gadolinium-based contrast medium at a dose of 0.1 mmol/kg.
The reading of MRI studies was centralized and performed by two experienced abdominal radiologists (M. L. and V. V.) who were blinded to the clinical and histological features of the patients. Definitive decisions were taken by consensus. Magnetic resonance images as well as the 3D reconstruction of cholangiograms were reviewed on an interactive workstation. The MRCP were evaluated for increased bile duct visualization, bile duct irregularities, bile duct strictures, and bile duct dilatation. Intrahepatic biliary duct dilatation was considered present if intrahepatic ducts were of greater diameter than more central ducts or if they were greater than 3 mm. Sclerosing cholangitis was defined by the presence of multiple and diffuse intrahepatic and/or extrahepatic bile duct strictures with or without associated biliary dilatation in the absence of apparent cholangiocarcinoma.12 A normal cholangiogram was defined by the absence of apparent stricture, irregularity, and bile duct dilatation. Evaluation of liver MRI abnormalities included changes in morphologic features (segmental enlargement or atrophy, irregularity of the liver contour), presence of confluent fibrosis, and existence of portal hypertension (splenomegaly and ascites or portosystemic collateral vessels). Confluent fibrosis was defined as a region of amorphous fibrosis greater than 2 cm in diameter that had a hyperintense signal on T2-weighted imaging, a hypointense signal on T1-weighted imaging, and showed enhancement on postcontrast images. Additional criteria were based on evaluation of lymphadenopathy and hepatocellular carcinoma nodules.
The data are expressed as the mean ± standard deviation and percentages. The AIH patients were discriminated into two groups according to the presence or absence of MRCP abnormalities. The clinical, biochemical, and histological characteristics of the patients were compared between the two groups using the Pearson's chi-square test (or the Fisher's exact test when appropriate) for categorical variables and the nonparametric Wilcoxon rank-sum test for continuous variables. A P value of <0.05 was considered statistically significant. The variables found to be individually associated with MRCP abnormalities were subsequently included in a multiple regression logistic model. A backward selection procedure was used to select the variables independently linked to biliary abnormalities. Finally, the prevalence of MRCP abnormalities in the subgroup of AIH patients with mutilating hepatic fibrosis or cirrhosis (F3-F4 score on biopsy) was compared with that observed in the cirrhotic control group using the chi-square test.
The main characteristics of the patients are shown in Table 1. All patients were graded using the revised International Autoimmune Hepatitis Group scoring system. The mean score of the 59 AIH patients at diagnosis was 17.1 ± 2.2 (range, 14-23). A total of 46 patients (78%) fulfilled pretreatment criteria for a definite diagnosis (mean score, 17.9 ± 1.7) and 13 patients (22%) for a probable diagnosis (mean score, 14.3 ± 0.5). More than half of the patients (n = 35 [59%]) were enrolled at the time of diagnosis. The others (n = 24) were diagnosed an average of 8.6 years (range, 1.2-24.5 years) before enrollment. Among those patients, 11 (46%) were treated with corticosteroids in combination with azathioprine (or mycophenolate mofetil), nine were treated with azathioprine (or mycophenolate mofetil) alone, two (8%) were treated with corticosteroids alone, and two were followed up without treatment with a sustained remission after withdrawal from drugs. Five patients (8%) were diagnosed under the age of 18 years, but only two were minors (both 15 years) at the time of study. No patients had previous history of pruritus, infectious cholangitis, or chronic inflammatory intestinal disease. Biliary tracts were considered normal by way of ultrasonography in all cases.
|Age (years)||47.7 ± 18.5 (15.4–84.0)|
|Female sex||42 (71%)|
|Follow-up (years)||3.4 ± 6.1 (0–24.5)|
|Type 1 AIH||51 (87%)|
|AIH score at diagnosis||17.1 ± 2.2 (14–23)|
|Bilirubin (mg/dL)||3.7 ± 6.3 (0.3–30.1)|
|ALT (ULN)||14.2 ± 19.6 (0.3–72.1)|
|ALP (ULN)||1.3 ± 0.7 (0.3–4.1)|
|Albumin (g/L)||39.5 ± 6.1 (23–48)|
|Gamma globulins (g/L)||19.6 ± 9.6 (7.6–53.0)|
|IgG (ULN)||1.3 ± 0.4 (0.4–2.3)|
|IgM (ULN)||0.7 ± 0.4 (0.2–2.3)|
|Prothrombin index (%)||84 ± 18 (39–129)|
|Platelet count (per mm3)||218 ± 82 (71–491)|
|ANA positivity||34 (58%)|
|ASMA positivity||40 (68%)|
|LKM1 positivity||2 (3%)|
|ANCA positivity*||9 (29%)|
|SLA positivity†||2 (6%)|
|No autoantibody||5 (8%)|
The histological characteristics are outlined in Table 2. The mean length of biopsy specimens was 19 ± 7 mm (range, 8-43 mm), and the mean number of portal tracts per specimen was 15 ± 11 (range, 4-59). Twenty-two percent of the patients had cirrhosis, and half had extensive hepatic fibrosis (F3-F4). On average, the activity grade was 1.9 ± 1.0 and the fibrosis stage was 2.1 ± 1.4. No case of bile duct paucity was registered. One fifth of the patients (n = 12) had evidence of destructive/nondestructive lymphocytic cholangitis. In these patients, a maximum of two such lesions was observed per biopsy, affecting 10% of portal tracts on average. No significant differences in the clinical and biochemical profiles (especially serum bilirubin, alkaline phosphatase, and gamma glutamyl transpeptidase) or in therapeutic response were shown between patients with and without this histological feature. Bile duct changes compatible with a fibrous obliterative cholangitis were observed in a 62-year old woman with precirrhotic fibrosis. In this patient, a single lesion of concentric periductal fibrosis was seen among 21 portal tracts with identifiable bile duct.
|Absent or mild||16 (27%)|
|Ductular proliferation||16 (27%)|
|Fibrous obliterative cholangitis||1 (1.7%)|
|Destructive/nondestructive lymphocytic cholangitis||12 (20%)|
|Interlobular bile duct paucity||0|
The MRI characteristics of the AIH patients are shown in Table 3. The mean interval between biopsy and the realization of MRI was 1.8 ± 1.6 months (range, 0-6 months). No patient showed abnormality of the common bile duct. Significant changes in intrahepatic bile duct (IHBD) were observed in 14 patients (24%). In all cases, the changes were mild and limited, as shown in Fig. 1. The proportion of these changes was higher in patients diagnosed before enrollment than in those enrolled at the time of diagnosis, but the difference was not significant (33% versus 17% [P = 0.15]). None of the patients had definite evidence of SC on imaging. However, a single patient showed dilatation of bile ducts upstream of a long stricture (Fig. 2). This 28-year-old man, who initially presented with typical autoimmune cirrhosis without features of SC, subsequently developed pruritus and cholestasis despite corticosteroid and azathioprine treatment. He had no histological evidence of SC on biopsy and no sign of chronic intestinal inflammatory disease. A diagnosis of AIH/SC overlap syndrome was nonetheless considered probable, and ursodeoxycholic acid therapy was added, leading to a significant improvement in clinical and biochemical features. The patient who had a single histological lesion compatible with fibrous obliterative cholangitis on his biopsy showed mild bilateral irregularities of IHBD concurrently with marked liver fibrosis and morphologic changes of the liver on MRI. This patient had no clinical or biological signs of cholestasis and was not antineutrophil cytoplasmic antibody–positive.
|Intrahepatic bile duct changes||14 (24%)|
|Increase bile duct visualization||4 (7%)|
|Common bile duct change||0|
|Definite diagnosis of sclerosing cholangitis||0|
|Liver abnormalities||22 (37%)|
|Liver morphology changes||22 (37%)|
|Confluent fibrosis||13 (22%)|
|Portal hypertension||4 (7%)|
One-third of the patients (n = 22) showed MRI hepatic changes suggestive of chronic liver disease. Confluent fibrosis was observed in 13 of the 22 patients with liver morphologic change (Fig. 3). Evidence of portal hypertension was found in 7% of patients (n = 4). Lymphadenopathy was not observed. No case of hepatocellular carcinoma was registered.
The MRI characteristics of the cirrhotic control group in comparison with AIH patients are shown in Table 4. In the control group, significant IHBD changes and signs of cirrhosis were observed in 16 (59%) and 21 (78%) cases, respectively (Fig. 4). Confluent fibrosis was observed in three patients (12%), and evidence of portal hypertension was found in 10 patients (37%).
|F0-F2 AIH (n = 35)||F3-F4 AIH (n = 24)||Cirrhotic controls (n = 27)|
|IHBD changes*||3 (9%)||11 (46%)||16 (59%)|
|Increase bile duct visualization||1 (3%)||3 (12%)||4 (15%)|
|Irregularities||1 (3%)||10 (42%)||16 (59%)|
|Strictures||1 (3%)||4 (17%)||3 (11%)|
|Dilatation||1 (3%)||1 (4%)||3 (11%)|
|Liver MRI abnormalities||3 (9%)||19 (79%)||21 (78%)|
|Liver morphology changes||3 (9%)||19 (79%)||21 (78%)|
|Confluent fibrosis||1 (3%)||12 (50%)||3 (12%)|
|Portal hypertension||0||4 (17%)||10 (37%)|
Variables Associated with Bile Duct Changes on MRI.
The AIH patients were divided into two groups according to the presence or absence of bile duct changes on MRCP. The comparison of these two groups is shown in Table 5. In univariate analysis, the features significantly associated with MRCP IHBD abnormalities were longer duration of follow-up, lower prothrombin index, lower platelet count, higher serum immunoglobulin G level, and higher fibrosis stage. The mean AIH score in patients with bile duct abnormalities on MRCP did not differ from those with normal MRCP (17.2 ± 0.6 versus 17.1 ± 0.3). On multivariate analysis, fibrosis stage remained the only independently associated factor (odds ratio [adjusted for age and duration of follow-up] 2.4; 95% confidence interval 1.4-4.7 [P < 0.005]). This finding suggests a close relationship between the amount of hepatic fibrosis and IHBD changes on MRI. The MRCP findings in F3-F4 AIH patients were compared with those observed in cirrhotic controls. The percentages of IHBD changes did not differ between the two groups (Table 4). Three patients had MRCP abnormalities without F3-F4 score of fibrosis at histology: one with increased bile duct visualization, one with bile duct irregularities, and one with bile duct strictures. None of these patients showed histological features of SC on biopsy or biochemical cholestasis. Of these three patients, liver morphologic changes suggestive of extensive fibrosis or cirrhosis were found in one patient on MRI, but not on liver biopsy.
|Variable||IHBD Changes on MRI (n = 14)||Normal IHBD on MRI (n = 45)|
|Age (years)||47.7 ± 19.6||47.6 ± 18.4|
|Age at diagnosis (years)||41.6 ± 20.1||45.4 ± 19.5|
|Male sex||5 (36%)||12 (27%)|
|Follow-up (years)||6.3 ± 8.1*||2.6 ± 5.1|
|AIH score at diagnosis||17.2 ± 0.6||17.1 ± 0.3|
|Bilirubin (mg/dL)||3.2 ± 5.7||3.9 ± 6.5|
|ALT (ULN)||12.5 ± 20.4||14.7 ± 19.5|
|ALP (ULN)||1.3 ± 0.5||1.2 ± 0.8|
|ALP/ALT||0.6 ± 0.6||0.5 ± 0.6|
|Albumin (g/L)||36.2 ± 8.0||40.6 ± 5.0|
|Gammaglobulins (g/L)||24.9 ± 15.1||17.7 ± 6.0|
|IgG (ULN)||1.5 ± 0.4 *||1.2 ± 0.4|
|IgM (ULN)||0.7 ± 0.1||0.6 ± 0.1|
|Prothrombin index (%)||74.4 ± 16.8 *||87.1 ± 17.9|
|Platelet count (per mm3)||169 ± 59 *||231 ± 83|
|ANA positivity||7 (50%)||27 (60%)|
|ASMA positivity||8 (57%)||32 (71%)|
|Anti-LKM1 positivity||1 (7%)||1 (2%)|
|ANCA positivity||3 (50%)||6 (24%)|
|Activity grade||2.1 ± 1.0||1.8 ± 1.1|
|Fibrosis stage||3.1 ± 0.9 *,†||1.8 ± 1.3|
|Lymphocytic cholangitis||3 (21%)||9 (20%)|
|Fibro-obliterative cholangitis||1 (7%)||0|
|Ductular proliferation||4 (29%)||12 (27%)|
Bile duct abnormalities were found by way of MRCP in nearly a quarter of our AIH patients. The most frequent abnormality was the presence of mild IHBDs. Most patients presenting IHBD MRI changes had extensive fibrosis on biopsy, and fibrosis stage was the only parameter independently linked to these changes on MRCP. Of note, there was no significant difference in the percentage of IHBD MRI changes between F3-F4 AIH patients and cirrhotic controls. This suggests that most of the observed IHBD changes on MRCP were nonspecific and the result of fibrosis deposition and remodeling of the liver. The hypothesis that underlying SC may explain more advanced fibrosis in these patients was unlikely, because none of them, except one established case of AIH/SC association, showed concurrent features of chronic cholestasis.
All our patients had histological examination of the liver at enrollment. The reading of biopsies was performed in each center by one experienced liver pathologist using a pre-established, standardized scoring procedure in order to limit misinterpretations and interobserver variations. The sensitivity of the liver biopsy for the diagnosis of SC is known to be poor. In a study carried out by Wiesner et al.21 only seven (12%) of 60 biopsies obtained from SC patients had diagnostic features of fibrous cholangitis. Thus, the possibility of undiagnosed small duct SC in our study is not completely excluded even if unlikely. A single lesion of concentric periductal fibrosis among 21 portal tracts was observed in one patient. There was no associated feature of ductular proliferation, cholate stasis, or bile duct loss. This patient had evidence of precirrhotic fibrosis on biopsy and mild peripheral IHBD irregularities on MRCP, but no biochemical signs of cholestasis. In this context, the diagnostic value of this histological lesion remained uncertain, and longer follow-up is required to make a definite diagnosis. It must be noted that a recent retrospective study of 16 AIH patients who had been confirmed with AIH/PSC overlap syndrome at some stage of their medical history has shown that histological re-examination of initial biopsies revealed PSC features in all biopsies.22 Histological evidence of destructive/nondestructive cholangitis was found in 12 (20%) patients. This finding is in keeping with the results of Czaja and Carpenter,20 who found such bile duct lesions in 16 of 84 biopsies obtained from patients with AIH. No case of bile duct loss was observed.
According to our data, the estimated prevalence of unknown SC in adult-onset AIH appears to be low (1.7%). A single definite case of AIH/SC overlap syndrome was indeed diagnosed among 59 consecutive AIH patients. In this patient, the diagnosis was not initially made on MRCP findings, which were not typical of SC, but only after he had developed pruritus and marked cholestasis while he was already on immunosuppressive drugs. Our findings do not support the conclusions of Abdalian et al.,14 who reported a rather high prevalence of SC of 10% in adults with AIH. We do not think that this difference might result from a discrepancy in the sensitivity of imaging techniques, because similar 1.5-T MRI systems were used in both studies. In contrast, discrepancies in the diagnostic criteria may explain the observed differences, at least in part. Indeed, beside the definite cases of SC, Abdalian et al. considered SC as probable when IHBD irregularities were observed in multiple segments, even in the absence of any stricture. In our study, the diagnosis of SC was defined by multiple intrahepatic and/or extrahepatic bile duct strictures. It is noteworthy that three of the eight patients diagnosed with SC in the Abdalian et al. study were labeled as having probable (that is, uncertain) disease, and none was reported to have concomitant extrahepatic cholangitis. Furthermore, histological data were available in only a few patients, and no case of fibrous obliterative cholangitis was observed in the subgroup of patients analyzed with an MRCP diagnosis of SC. Another explanation for such differences may lie in the methodological approaches. The Canadian study was based on a retrospective design, which implies some selection bias, as to propose MRCP evaluation more likely in patients with clinically suspected SC. In our study, this bias was avoided by the prospective enrollment of the patients and the exclusion of patients with typical features of AIH/SC overlap syndrome at inclusion. It is possible that the requirement of histological examination as an eligibility criterion for our study might have generated a selection bias, resulting in an overrepresentation of patients with more severe or atypical disease. However, it seems unlikely that this could have strongly influenced our results, because most of the biopsies have been performed as routine procedures, either in the context of diagnosis or in the monitoring protocol of patients. Furthermore, if this bias had been significant, the prevalence of SC/AIH overlap syndrome would have been, in all probability, higher than that actually observed.
One limitation of our study was the lack of follow-up of AIH patients by MRCP, which could have underestimated the actual frequency of SC. However, it must be noted that 24 (41%) of our patients were followed up for an average period of 8.6 years (range, 1-24 years) before enrollment. Interestingly, although these patients were more frequently found to have IHBD MRCP abnormalities compared with those enrolled at the time of diagnosis, they were also more likely to have extensive fibrosis and cirrhosis and marked morphologic changes of the liver. Indeed, if longer follow-up was individually associated with MRCP abnormalities in univariate analysis, hepatic fibrosis remained the only factor independently linked to these changes in a multiple regression analysis. Thus, duration of follow-up was a confounding variable interfering with the probability of MRCP bile duct abnormalities through its strong positive relationship with fibrosis stage. These data render the hypothesis of a sequential progression of adult-onset AIH to SC unlikely. However, a prospective monitoring through MRCP of AIH patients and fibrosis-matched controls is the only way to determine the actual incidence of SC in this disease setting.
It is known that hepatic fibrosis, especially cirrhosis, may result in a distortion of IHBDs with cholangiographic features mimicking SC. In a retrospective ERCP study, a high incidence of abnormal findings with abnormalities of ductal caliber, multiple focal stenosis (56%), and decreased arborization or pruning of ductular branches (64%) were found in cirrhotic patients.15 In PBC, which is known to spare large bile ducts and to be virtually never associated with SC, irregularities of the intrahepatic biliary tree were found to be tightly associated with histologically proven cirrhosis.23 In a postmortem cholangiographic and histopathological study, Terada and Nakanuma16 reported intrahepatic cholangiographic appearance simulating SC in two-thirds of the cirrhotic livers. Although these results were not reproduced in living patients, minor ductal irregularities, crowding, pruning, and quick tapering of the intrahepatic biliary tree on ERCP were confirmed in 70% of cirrhotic patients by Altraif and Lewall.24 More recently, Fulcher et al12 pointed out that false-positive diagnoses of SC may occur on MRCP in patients with cirrhosis. The suggested mechanism for these biliary changes is the imprinting on the bile duct lumen by the distorted hepatic architecture due to mutilating fibrosis.
The spatial resolution of MRCP is still a technical limitation, even in the 3D imaging technique, which has the advantages over two-dimensional imaging to obtain thinner sections with no gap and a higher signal-to-noise ratio.12 Indeed, the changes predominating in the third- and fourth-order bile ducts (diameter ranging from 500 to 390 μm) are not always adequately analyzable, making it difficult to interpret peripheral duct abnormalities. These difficulties might explain in part the mild bile duct changes described in our study in two of the three AIH patients with no histological evidence of cirrhosis or advanced fibrosis, while the remaining patient, who had obvious features of cirrhosis on MRI, was likely to be underestimated by histological staging. Conversely, the negative findings on MRCP might not completely exclude early-stage SC, because MRCP is considered by some authors of lower sensitivity than ERCP for the diagnosis of proximal bile duct lesions, even if this has not been clearly demonstrated in the field of PSC.25, 26
Bilaj et al.27 have recently documented the MRI appearance of AIH, and this report is consistent with our data. The authors observed mild IHBD dilatation in 12% of AIH patients (4/32 patients). Interestingly, a common finding was the presence of extensive reticular and/or confluent fibrosis (6/32 patients) that may help to establish diagnosis of AIH. However, such areas of confluent fibrosis are also a recognized feature of SC. They were indeed observed in more than half of SC patients who had imaging findings suggestive of cirrhosis.28
In conclusion, our data do not support the need for routine radiological evaluation of the biliary tree in adults with a diagnosis of AIH. Indeed, the prevalence of overlapping SC in these patients appears very low—below 2% in the present study. Mild abnormalities of IHBD on MRCP are present in one quarter of patients, but these abnormalities are linked to the extent of hepatic fibrosis and not to SC itself. Therefore, we recommend investigating adult-onset AIH by way of MRCP only in cases of cholestatic manifestations or poor response to corticosteroid treatment.
- 4Overlap of autoimmune hepatitis and primary biliary cirrhosis: an evaluation of a modified scoring system. Am J Gastroenterol 2002; 97: 1191–1197., , , .Direct Link:
- 6Features of autoimmune hepatitis in primary sclerosing cholangitis: an evaluation of 114 primary sclerosing cholangitis patients according to a scoring system for the diagnosis of autoimmune hepatitis. HEPATOLOGY 1996; 23: 1369–1376., , , , , , et al.
- 11Outcome and survival in childhood onset autoimmune sclerosing cholangitis and autoimmune hepatitis: a 13 year follow-up study [Abstract]. HEPATOLOGY 2007; 46( Suppl): 555A., , , , .
- 16Intrahepatic cholangiographic appearance simulating primary sclerosing cholangitis in several hepatobiliary diseases: a postmortem cholangiographic and histopathological study in 154 livers at autopsy. HEPATOLOGY 1995; 22: 75–81., .