There is no specific diagnostic test for autoimmune

Transcription

1 Long-Term Follow-Up of Antimitochondrial Antibody Positive Autoimmune Hepatitis Conor O Brien, 1 Supriya Joshi, 1 Jordan J. Feld, 2 Maha Guindi, 3 Hans P. Dienes, 4 and E. Jenny Heathcote 1 Antimitochondrial antibodies (AMAs) are the serological hallmark for primary biliary cirrhosis (PBC). When AMAs are detected in patients with chronic hepatitis, they negatively impact on the autoimmune hepatitis (AIH) scoring system. The purpose of this study was to determine if AMAs detected in the sera of patients with overt AIH have clinical or pathological significance. All patients with a clinicopathologic diagnosis of AIH from one center were reviewed. Only those meeting the criteria for probable or definite AIH according to the International Autoimmune Hepatitis Group were included. Patients found to be consistently AMA-positive via enzyme-linked immunosorbent assay (ELISA) for pyruvate dehydrogenase complex E2 subunit were reviewed in detail. Fifteen of 126 patients with typical features of AIH (pretreatment AIH score >10) had detectable AMAs in serum. None had any histologic features suggestive of PBC. None had detectable anti-liver kidney microsomal antibodies. Of these 15 patients, all have remained persistently AMA-positive via ELISA. All 15 patients have been followed long-term, and their clinical course remained typical for AIH. No bile duct damage typical of PBC was seen on initial or follow-up liver biopsies. Conclusion: Patients with overt AIH who test positive for AMAs at initial presentation and are treated with corticosteroid therapy have shown no clinical or histologic evidence of PBC despite the continued detection of AMAs over a follow-up of up to 27 years. (HEPATOLOGY 2008;48: ) There is no specific diagnostic test for autoimmune hepatitis (AIH). A scoring system was devised by the International Autoimmune Hepatitis Group (IAHG), which, although originally designed for case definition in clinical trials, may also be used clinically to help establish a diagnosis, particularly in patients with atypical features. Points are added for features considered typical of AIH and subtracted for those associated with other Abbreviations: AIH, autoimmune hepatitis; ALP, alkaline phosphatase; ALT, alanine aminotransferase; AMA, antimitochondrial antibody; ANA, antinuclear antibody; ELISA, enzyme-linked immunosorbent assay; IAHG, International Autoimmune Hepatitis Group; IgG, immunoglobulin G; IgM, immunoglobulin M; LKM, liver kidney microsomal; PBC, primary biliary cirrhosis. From the 1 Department of Medicine, University Health Network, Toronto Western Hospital, University of Toronto, Toronto, Canada; 2 Liver Diseases Branch, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD; 3 Department of Pathology, University Health Network, Toronto General Hospital, University of Toronto, Toronto, Canada; and 4 Institute of Pathology, University of Cologne, Cologne, Germany. Received January 30, 2008; accepted April 4, Address reprint requests to: E. Jenny Heathcote, M.D., Department of Medicine, University Health Network, Toronto Western Hospital, 399 Bathurst Street, 6B Fell Pavilion-154, M5T 2S8 Toronto, Ontario, Canada. jenny. heathcote@utoronto.ca; fax: Copyright 2008 by the American Association for the Study of Liver Diseases. Published online in Wiley InterScience ( DOI /hep Potential conflict of interest: Dr. Heathcote is a consultant for, advises, is on the speakers bureau of, and received grants from Axcan-Pharma. chronic liver diseases. The scoring system was revised in Items with a positive score include female sex; a high aminotransferase/alkaline phosphatase ratio; high levels of immunoglobulin G (IgG); compatible liver histology; the presence of antinuclear antibodies (ANAs), anti smooth muscle antibodies, and liver kidney microsomal (LKM) antibodies; and a good response to immunosuppressive therapy. Points are subtracted for items such as regular alcohol consumption, markers of viral hepatitis, a positive medication history, and the presence of antimitochondrial antibodies (AMAs). In the setting of chronic liver disease, serum AMAs are considered the serological hallmark of primary biliary cirrhosis (PBC); however, when they are present in patients with other diseases, including AIH, their significance is unknown. AMAs have been detected by using the lessspecific immunofluorescent method, albeit rarely, in patients with collagen vascular diseases, drug-induced liver injury, syphilis, tuberculosis, heart disease, and other chronic liver diseases. 2 In individuals who appear to have PBC (AMA-positive) and yet have features of AIH, it has been proposed that they have a PBC/AIH overlap syndrome. 3-5 Such patients may respond well to standard therapy for PBC, namely ursodeoxycholic acid, 6 but Chazouilleres et al

2 HEPATOLOGY, Vol. 48, No. 2, 2008 O BRIEN ET AL. 551 suggested such patients have a more aggressive disease course and should have corticosteroid therapy added to their regimen. Other patients have been described who lose all features of PBC, including AMA and subsequently develop classical AIH. 7 However, there are very few reports of patients with fully documented AIH who were treated with immunosuppressive therapy and subsequently develop clinical and histologic evidence of PBC. Suzuki et al. 8 described a female patient with typical AIH who had no biliary pathology on initial biopsy and was AMA-negative at diagnosis who subsequently developed AMA-positive stage IV PBC 17 years after the initial diagnosis of AIH. Poupon et al. 9 reported 5 patients with AIH who where subsequently found to also have PBC, but no details are available. The significance of finding isolated AMAs in the setting of clear-cut AIH remains unclear. Clarification of this issue may have important relevance to both prognosis and management. If both conditions coexist, treatment for both may be deemed necessary 5,9 ; however, if only one disease is truly present, redundant therapies may lead to unnecessary side effects. Thus, our aim was to evaluate the significance of the presence of persistent AMAs in patients presenting with typical AIH to determine if clinical, biochemical, or histologic features of PBC develop during long-term follow-up. Patients and Methods Patients. Patients with a clinicopathologic diagnosis of AIH (pretreatment AIH score 10) from the liver clinic at Toronto Western Hospital were reviewed. Patient records were reviewed for clinical, biochemical, serological, and histological characteristics at the time of diagnosis and at follow-up. Response to therapy was also assessed. This study was approved by the research ethics board of the University Health Network. Serological and Immunological Tests. At the time of initial presentation, patients were tested for IgG and immunoglobulin M (IgM), ANAs, smooth muscle antibodies (SMAs), and when applicable, LKM antibodies. The upper limit of normal values were 13.3 g/l for IgG and 2.3 g/l for IgM. IgG and IgM levels were measured using standard techniques and anti SMAs, ANAs, and AMAs were determined via indirect immunofluorescence with use of mouse kidney and stomach as substrates. All positive AMA results were confirmed via ELISA, employing the specific autoantigen pyruvate dehydrogenase complex E2 subunit (Varelisa M2 antibody kit, Pharmacia Diagnostics). Serum autoantibodies and immunoglobulin levels were measured on a yearly basis as part of routine patient follow-up. Histological Evaluation. At the time of initial presentation, liver biopsies were performed on all patients. Biopsies were repeated in six patients to re-evaluate disease status. Two hepatopathologists reviewed all archived biopsies from patients with AIH in whom AMAs were detected at initial presentation. Liver biopsies were stained for hematoxylin-eosin, Masson s trichrome, reticulin, and PASD using standard protocols. Biopsies from these patients were specifically evaluated for features of both AIH and PBC. Biopsies were evaluated for hepatitic activity, including features of AIH according to the criteria of the IAHG. 1 Features examined included periportal/ periseptal interface hepatitis, confluent necrosis and lobular necroinflammation, rosetting, and portal lymphoplasmacytic infiltrate. Biopsies were examined for features of PBC, including duct damage, florid duct lesions, ductopenia, ductular reaction, cholestasis, and granulomata. Assessment of fibrosis was performed according to Desmet et al. 10 CK-7 staining was employed in most biopsies to closely examine the bile ducts to identify subtle, early changes of PBC. Each biopsy was assigned a total score according to the histologic variables included in the IAHG scoring system. 1 These variables are as follows: interface hepatitis 3, predominantly lymphoplasmacytic infiltrate 1, resetting of liver cells 1, none of the above 5, biliary changes 3, other changes 3. Fibrosis was staged according to Desmet et al. 10 Results Patients. Fifteen patients with AIH were found to be positive for AMAs via ELISA on at least three occasions during their follow-up. Investigations for hepatitis A, hepatitis B, hepatitis C, Wilson disease, hemochromatosis, alcoholic and nonalcoholic fatty liver disease, and gallstone disease were negative. All 15 patients were female. The median age at diagnosis of AIH was 47 years. Overt jaundice was the presenting symptom in seven patients, whereas five were asymptomatic. The median alanine aminotransferase (ALT) level at presentation was 373 U/L, and the median IgG level was elevated at 21 g/l (Table 1). The median alkaline phosphatase (ALP) level was 125 U/L with an elevated median IgM at 3.15 g/l. Six patients had an ANA titer of 1:80 or more (two speckled, two multiple nuclear dot, one homogeneous, and one centromeric), two had an SMA titer of over 1:40, whereas the 11 patients tested for LKM antibodies were all negative (Table 1). In seven patients, AMAs represented the only positive autoantibody. Four patients who initially tested AMA-positive via indirect immunofluorescence (median titer 1:160, range, 1:160 1:320) were subsequently confirmed via ELISA. All patients were tested for AMAs via ELISA on at

3 552 O BRIEN ET AL. HEPATOLOGY, August 2008 Table 1. Characteristics at Presentation in 15 Patients with Classical AIH and AMA Positivity via ELISA Number of patients 15 Female 15 Median age at diagnosis (range), years 47 (27 65) Median follow-up (range), years 8 (1 27) Presenting symptom Jaundice 7 Fatigue 3 Asymptomatic 5 Median (range) AIH score (pretreatment) 14 (11 15) AIH score (excluding AMA) 18 (15 19) Median (range) AST (0 35 U/L) 324 ( ) ALT (0 40 U/L) 373 ( ) ALP (0 110 U/L) 125 ( ) GGT (0 40 U/L) 110 ( ) Bilirubin (1 22 mol/l) 14 (11 300) IgG ( g/l) 21 ( ) IgM ( g/l) 3.15 ( ) AMA ( IU/mL), ELISA 79 (23 200) ANA-positive, 1:80 6 SMA-positive, 1:40 2 LKM-positive 0 (n 11) Median number of AMA tests/patient (range) 3 (3 6) Abbreviations: AST, aspartate aminotransferase; GGT, gamma glutamyl transpeptidase. least three occasions; the median AMA titer (via ELISA) was 79 IU/mL (range, 23 to 200 IU/mL) (Table 1). The median pretreatment AIH score was 14 (range, 11 15), and with elimination of the deduction for AMAs, this median score rose to 18 and all patients scored as definite cases of AIH. Fourteen patients underwent liver biopsy at the time of initial presentation, one of whom had their first liver biopsy 6 years after initial diagnosis. Treatment. Ten patients were treated with a combination of prednisone and azathioprine, one with prednisone alone (intolerant of azathioprine), and one with ursodeoxycholic acid alone (commenced prior to referral for presumed PBC) (Table 2). One patient received a combination of ursodeoxycholic acid, prednisone, and azathioprine because her biopsy demonstrated some biliary features not diagnostic of PBC but more often seen in Table 2. Initial Treatment and Response Prednisone alone 1 Prednisone and azathioprine 10 UDCA alone 1 UDCA/azathioprine/prednisone 1 No treatment 2 Median time to normalize ALT (range), n 10 6 months (1 24 months) Median time to reduce ALT 2 ULN (range), n 2 6 months (4 6 months) Median no. of relapses, ALT 2 ULN (range) 0 (0 3) Abbreviation: UDCA, ursodeoxycholic acid; ULN, upper limit of normal. Table 3. Current Treatment and Biochemistry (n 14) Prednisone alone 1 Prednisone and azathioprine 3 Azathioprine alone 5 UDCA alone 1 UDCA/prednisone/azathioprine 1 No treatment 3 Median AST (0 35 U/L) 34 (23 129) ALT (0 40 U/L) 23 (12 116) ALP (0 110 U/L) 93 ( ) GGT (0 40 U/L) 38 (15 134) Bilirubin (1 22 mol/l) 13 (5 60) Abbreviations: AST, aspartate aminotransferase; GGT, gamma glutamyl transpeptidase; UDCA, ursodeoxycholic acid. AIH. One patient did not fulfill AIH treatment criteria (AASLD Guidelines 2002), and one declined all medications. Ten of 12 patients given immunosuppressive therapy had a complete response to therapy and developed normal ALT at a median time interval of 6 months (range, 1 24), whereas in two patients therapy only reduced ALT levels to 2 times the upper limit of normal (Table 2). Follow-up ranged from 1 to 27 years. During this time, five patients demonstrated at least one flare typical of AIH, and all five responded to reinstitution of immunosuppressive therapy. No patient developed pruritus or fatigue, symptoms typical of PBC, during the follow-up period. Currently, nine patients still require immunosuppressive therapy with either azathioprine or prednisone alone or azathioprine in combination with low-dose prednisone; in two cases, prednisone is required for nonhepatic reasons (renal disease and temporal arteritis) (Table 3). The two patients commenced on ursodeoxycholic acid continue to take it currently. In one patient it was discontinued; however, she was observed to have an elevated ALP once again, so it was reinstituted. The other patient has normal liver enzymes on a combination of ursodeoxycholic acid, azathioprine, and prednisone. A summary of clinical and biochemical parameters at last assessment are shown in Table 3. Histological Examination. Nine patients had only one biopsy. Of the six patients who had repeat biopsies, four had two biopsies and two had three biopsies. Biopsies were repeated to confirm a flare of AIH or prior to immunosuppressant withdrawal. Overall, 23 biopsies were performed on the 15 patients with persistently AMA-positive AIH. Of the 18 biopsies available for study review, 16 were performed at Toronto Western Hospital. Five biopsies from four patients were not available for review. In these four patients, the unavailable biopsies were the initial biopsies, but the subsequent repeat biopsies were reviewed in each case. Histologic data in these instances

4 HEPATOLOGY, Vol. 48, No. 2, 2008 O BRIEN ET AL. 553 Table 4. Summary of Histological Assessment in Biopsies of AMA-Positive Patients with AIH Number of Biopsies Histological Score Comment 13 5 No points deducted for one biopsy with moderate steatosis 1 3 Deduction for absence of rosetting and interface activity (posttreatment) 2 4 Deduction for absence of rosetting in each (posttreatment) 1 4 Deduction for lack of interface activity in incidentally found asymptomatic AIH, mildly active with severe fibrosis 1 1 A positive score of 2 for hepatitic features was offset by a score of 3 due to the presence of some biliary changes (this is a repeat biopsy post-aih treatment) were obtained from the initial pathology report (also performed by a hepatopathologist), but they were not assigned a score upon review. The reports of these five biopsies indicate absence of duct injury or biliary changes. All reviewed biopsies met the IAHG criteria for AIH on their initial biopsy, and none had definitive features of PBC (Table 4). Thirteen of the 18 reviewed biopsies had a total score of 5/5 for the histologic criteria in the IAHG scoring system. One biopsy had large droplet steatosis in approximately 50% of hepatocytes without evidence of active steatohepatitis; therefore, no points were deducted from its score, because steatosis did not impact on histologic evaluation of hepatitic activity and bile ducts. One biopsy achieved a score of 3/5, three biopsies achieved a score of 4/5, and one biopsy had a score of 1. The biopsy with a score of 3/5 and two of the biopsies with a Fig. 2. Mild bile duct injury. Photomicrograph shows portal tracts with dense inflammation by lymphocytes and plasma cells in liver biopsy obtained during a flare of AIH. The bile duct epithelium (arrow) is minimally invaded by lymphocytes. (Hematoxylin-eosin stain, magnification 400.) score of 4/5 were related to absence of rosetting and/or interface activity. However, these were posttreatment biopsies, and we postulated that response to treatment may have caused these changes to disappear. We acknowledge that the IAHG score, including its histologic components, is intended and validated for AIH diagnosis pretreatment and thus may not be well suited to evaluation of posttreatment biopsies. The remaining biopsy with a score of 4/5 was at initial diagnosis of asymptomatic AIH with severe fibrosis. The patient had been found quite incidentally to have abnormal liver enzymes. The biopsy with a total score of 1 was a repeat biopsy after treat- Fig. 1. Active necroinflammation during a flare of AIH. Photomicrograph shows parenchymal inflammation by lymphoplasmacytic infiltrate, portal inflammation, and interface activity. The bile ducts appear intact. (Hematoxylin-eosin stain, magnification 200.) Fig. 3. (A) Repeat liver biopsy of one of the patients showing cirrhosis but with normal ducts present, indicating the fibrosis is not related to duct loss (Masson s trichrome stain). (B) The same biopsy stained with cytokeratin 7 immunostain highlighting the presence of bile ducts.

5 554 O BRIEN ET AL. HEPATOLOGY, August 2008 ment for AIH. Interface hepatitis was absent. Two points gained for the lymphoplasmacytic composition of the portal infiltrate and the presence of rosetting were offset by deduction of 3 points due to the presence of two small ducts with degenerative appearing epithelium, ductular reaction, and intermediate cells on CK7 immunostaining biliary features beyond what may be usually expected in AIH but that are insufficient for a specific diagnosis of biliary disease. Three patients with dense portal infiltration had the occasional lymphocyte in duct epithelium without biliary epithelial necrosis or disruption on basement membrane. This was interpreted as spillover of inflammatory cells in bile duct epithelium. The time between the initial and the most recent liver biopsy ranged from 4 to 24 years. In the follow-up biopsies, no duct paucity, duct injury, ductular proliferation, or cholestasis was evident apart from the one patient who developed mild biliary changes as discussed above (Figs. 1). Outcome. These patients have been followed for a median of 8 years (range, 1 27 years). One patient who had superimposed hepatic steatosis on her liver biopsy died due to complications of type 2 diabetes mellitus and cardiac disease 10 years after she first presented with AIH. Nine patients had cirrhosis on biopsy at diagnosis, and one patient developed cirrhosis 8 years after initial diagnosis. Four of these patients with cirrhosis have esophageal varices, and two have had variceal bleeding. Two patients have decompensated with ascites; one as mentioned above died of nonhepatic causes, and the other is currently being evaluated for liver transplantation. Twelve patients have undergone magnetic resonance cholangiopancreatography with no evidence of large bile duct disease. Discussion Although the presence of AMAs has been described in patients with AIH, 11 their follow-up has not been reported; consequently, the long-term significance of AMA-positive AIH remains unclear. Fifteen patients who presented with clinical, biochemical, and histological features compatible with or diagnostic of AIH alone are described. All 15 had repeatedly detectable AMA confirmed via ELISA. However, during up to 27 years of follow-up, none developed clinical or histological features consistent with the development of PBC. Over 95% of patients given a diagnosis of PBC have AMAs in serum; however, whether all individuals found to test positive for AMA have or will develop PBC is less clear. Population-based epidemiological studies document AMA prevalence rates of approximately 0.5%, yet the highest reported prevalence of PBC is 940 per million (0.09%) in females over 40 years old in northeast England, 12 suggesting that only 1% to 20% of AMA-positive individuals are, at the time of initial detection, diagnosed with AMA-positive PBC In contrast, a study that followed a small cohort of AMA-positive individuals paints a different picture. Mitchison et al. 16 and Metcalf et al. 17 reported the histological findings of asymptomatic individuals with normal liver biochemistry who underwent liver biopsy after testing positive for AMAs. Of 29 patients, 24 (83%) had liver histology typical or compatible with PBC. Ten-year follow-up of the 24 surviving patients demonstrated that all eventually developed abnormal liver biochemical tests (21/24 had elevated ALP), and 22 developed symptoms attributable to PBC (fatigue, pruritus). 17 Repeat liver biopsy also documented mild progression of disease over a 10-year period in some. This small study would suggest that in individuals with AMAs, nearly all have underlying PBC, even if their liver biochemical tests are normal and that progression to biochemical and/or symptomatic PBC is typical. Whether this is also the case in the setting of underlying AIH is not known. The presence of AMAs in patients with AIH was described in 1986 by Kenny et al. 18 in a report of 37 patients with AIH who had detectable AMAs via immunofluorescence. Subsequently, many of these patients tested positive for LKM antibodies and it was evident that the reported AMA results were false positives due to confusion between mitochondrial and microsomal staining. 19 With the recognition of pyruvate dehydrogenase complex E2 subunit as the specific antigen for AMA, ELISA and immunoblotting became available, both of which have much higher specificity than IF. The reported frequency of AMA detection in patients with definite AIH ranges from 5% to 34%. Czaja et al. 19 demonstrated that AMA was detectable in 8 of 162 (5%) patients with AIH using ELISA and in 5 of 62 (8%) AIH patients in a separate study. 20 Farias et al. 21 reported AMA positivity in 18 of 206 (7.5%) patients with definite AIH using immunoblotting. In a recent report, Montano-Loza et al. stated that AMAs were detected in the sera of 31 of 240 (13%) AIH cases, 22 a rate similar to that reported in our study. In a smaller study in Japan, AMAs were detected via immunoblotting in 14 of 41 (34%) patients with AIH. 23 A recent study demonstrated the development of AMAs to one or more of the mitochondrial autoantigens in PBC in 28 of 69 (40.6%) patients with acute liver failure of varying etiologies. 24 This supports the hypothesis that oxidative stress is a potential inducer of AMAs and therefore may play a role in the pathogenesis of PBC in genetically predisposed individuals. AMA positivity in this cohort of acute liver failure patients was of lower titer

6 HEPATOLOGY, Vol. 48, No. 2, 2008 O BRIEN ET AL. 555 than traditionally seen in PBC and was also short-lived; only one subject remained AMA-positive 24 months after the initial liver insult. In contrast, our cohort of patients remained persistently AMA-positive over a median of 8 years of follow-up, even during disease remission, indicating that acute liver injury alone is unlikely to be the stimulus for AMA production in these patients. Also, only 5 of 15 patients in our study had severe hepatitis on biopsy at diagnosis, yet all 15 tested positive for AMAs, again minimizing the role of severe oxidative stress in production of these autoantibodies in our patient population. A case report from 1997 described a 12-year-old girl with otherwise typical features of AIH with detectable AMAs confirmed via ELISA and immunoblotting. Despite 12 years of follow-up, six liver biopsies, and a postmortem examination, no clinical, biochemical, or histological features of PBC developed. 25 Treatment with prednisone and azathioprine resulted in resolution of the patient s jaundice and normalization of serum aminotransferases. Czaja et al. 20 and Farias et al. 21 have reported patients with AIH who were AMA-positive, and both found that none of their patients had any features of PBC and all patients responded well to standard immunosuppressive therapy for AIH. Similarly, in the 15 patients we describe, all had clear clinical, laboratory, and histological evidence of AIH, and aside from testing AMA-positive had no suggestion of concomitant PBC either at presentation or at follow-up. One subject with a suggestion of PBC was a patient with a sustained mild elevation in ALP and mild duct injury on a recent biopsy, 26 years after her original diagnosis of AIH. However, this elevation in ALP was noted when the patient was first found to have cirrhosis in 1978, and there was no evidence of duct loss or other changes of PBC reported on this or subsequent liver biopsies that were 2 cm in length. At the time of her initial presentation, she was jaundiced, and this improved with steroid therapy. She remains alive and well (anicteric) 27 years later. This pattern of disease presentation with such longevity with icteric stage IV PBC has not been described. One patient with markedly cholestatic liver enzymes (ALP 1112) had no clinical or histological changes suggestive of PBC, scored 5 in the IAHG scoring system for liver histology, and had a normal magnetic resonance cholangiopancreatogram. This patient has declined all treatment to date, and her enzymes remained stable during a follow-up period of 4 years. Although AIH and PBC may share some common histological features, the patterns seen on liver biopsy typically differ. In the setting of acute inflammation, as may be seen in AIH, ductular reaction may be apparent; however, the ductular lesions typical of PBC are not present, even in those patients who are AMA positive. 11 To exclude the possibility of subtle early changes of PBC, the biopsy specimens from all our patients were stained for CK-7 to highlight the bile ducts. To ensure that PBC had not developed over time, repeat biopsies were performed in six patients, with up to 26 years separating their first and last biopsy. The fact that no duct lesions were seen during follow-up suggests that there was no progression to PBC in these patients, despite the persistence of detectable AMAs. Hence, in the absence of any histological features of PBC, both at the time of initial diagnosis and in subsequent follow-up, it appears that PBC did not coexist at any time in these six patients with AIH who persistently tested positive for AMA. The nine patients with a single biopsy had no definitive histological evidence of PBC, and their subsequent clinical course and response to immunosuppressive therapy to date is typical of that seen in classic AIH. There have been several case reports of patients with clear-cut PBC who have subsequently lost all features (clinical, biochemical, serological, and histological) of PBC and have developed typical AIH. 7,26,27 In addition, patients with AMA-positive PBC with biochemical, serological, and histological features of AIH have been described. 3-5,9 Whether liver disease in such patients truly represents an overlap syndrome is controversial. We previously reported that patients with a primary diagnosis of PBC with the biochemical, serological, and histological features of AIH described by Chazouilleres 5 respond well to treatment with ursodeoxycholic acid. 6 In a recently published series by Poupon et al., 9 4% of patients with an initial diagnosis of PBC developed classical AIH when followed long-term. What we describe in this report is very different. In this same report by Poupon, 9 brief mention is made of five patients with AIH whom we are told also had PBC, but no details of their presentation, histology, or follow-up is provided. There are three possible explanations for the phenomena we describe. This group of patients may represent another subtype of AIH (that is, AMA-positive AIH), which nevertheless responds well to corticosteroids and is not associated with features of PBC. An alternative hypothesis is that early treatment with high-dose prednisone given to individuals with circulating AMAs prevents the development of subsequent PBC. Trials of corticosteroids in patients with PBC have shown significant improvement in histological findings in those with early stage PBC It is possible that the introduction of immunosuppressive therapy in our treated patients at the time they presented with AIH may have prevented any subsequent progression to PBC. Unfortunately, this hypothesis cannot be evaluated easily, because withholding immunosup-

7 556 O BRIEN ET AL. HEPATOLOGY, August 2008 pressive treatment from a patient with symptomatic AIH and positive AMA could be life-threatening. Two of our patients who did not receive immunosuppressive therapy have remained clinically and biochemically stable after 3 to 5 years of follow-up, and no features of a cholestatic syndrome have developed. This concept may provide support for consideration of corticosteroid therapy in very early PBC, at least in those patients whose liver biochemistry does not return to normal on ursodeoxycholic acid monotherapy. Short-term trials offer some support for this approach, 5,28-30 but long-term data are lacking. Finally, it is possible that our patients will develop PBC in the future; however, given the long duration of follow-up in most cases to date, this seems unlikely. In conclusion, the long-term follow-up of these 15 cases suggests that patients with overt AIH who also have circulating AMA may not have coexistent PBC and should be managed as is appropriate for AIH. References 1. Alvarez F, Berg, PA, Bianchi FB, Burroughs AK, Cancado EL, Chapman RW, et al. International Autoimmune Hepatitis Group Report: review of criteria for diagnosis of autoimmune hepatitis. J. Hepatol 1999;31: Dienes HP, Erberich H, Dries V, Schirmacher P, Lohse A. Autoimmune hepatitis and overlap syndromes. Clin Liver Dis 2002;6: Duclos-Vallee JC, Hadengue A, Ganne-Carrie N, Ganne-Carrie N, Robin E, Degott C, et al. Primary biliary cirrhosis-autoimmune hepatitis overlap syndrome. Dig Dis Sci 1995;40: Lohse A, Meyer zum Buschenfelde KH, Franz B, Kanzler S, Gerkin G, Dienes HP. Characterization of the overlap syndrome of primary biliary cirrhosis (PBC) and autoimmune hepatitis: evidence for it being a hepatitic form of PBC in genetically susceptible individuals. HEPATOLOGY 1999;29: Chazouilleres O, Wendum D, Serfaty L, Montembault S, Rosmorduc O, Poupon R. Long term outcome and response to therapy of primary biliary cirrhosis autoimmune hepatitis overlap syndrome. J Hepatol 2006;44: Joshi S, Cauch-Dudek K, Wanless IR, Lindor KD, Jorgensen R, Batts K, et al. Primary biliary cirrhosis with additional features of autoimmune hepatitis: response to therapy with ursodeoxycholic acid. HEPATOLOGY 2002; 35: Colombato LA, Alvarez F, Cote J, Huet PM. Autoimmune cholangiopathy: the result of consecutive primary biliary cirrhosis and autoimmune hepatitis? Gastroenterology 1994;107: Suzuki Y, Arase Y, Ikeda K, Saitoh S, Tsubota A, Suzuki F, et al. Clinical and pathological characteristics of the autoimmune hepatitis and primary biliary cirrhosis overlap syndrome. J Gastroenterol Hepatol 2004;19: Poupon R, Chazouilleres O, Corpechot C, Chretien Y. Development of autoimmune hepatitis in patients with typical primary biliary cirrhosis. HEPATOLOGY 2006;44: Desmet VJ, Gerber M, Hoofnagle JH, Manns M, Scheuer PJ. Classification of chronic hepatitis: diagnosis, grading and staging. HEPATOLOGY 1994;19: Czaja AJ, Carpenter HA. Autoimmune hepatitis with incidental histologic features of bile duct injury. HEPATOLOGY 2001;34: James OF, Bhopal R, Howel D, Gray J, Burt AD, Metcalf JD. Primary biliary cirrhosis once rare, now common in the United Kingdom? HEPA- TOLOGY 1999;30: Mattalia A, Quaranta S, Leung PSC, Bauducci M, Van de Water J, Calvo PL, et al. Characterization of antimitochondrial antibodies in healthy adults. HEPATOLOGY 1998;27: Ohba K, Omagari K, Kinoshita H, Soda H, Masuda J, Hazama H, et al. Primary biliary cirrhosis among atomic bomb survivors in Nagasaki, Japan. J Clin Epidemiol 2001;54: Shibata M, Onozuka Y, Morizane T, Koizumi H, Kawaguchi N, Miyakawa H, et al. Prevalence of antimitochondrial antibody in Japanese corporate workers in Kanagawa prefecture. J Gastroenterol 2004;39: Mitchison HC, Bassendine MF, Hendrick A, Bennett MK, Bird G, Watson AJ, et al. Positive antimitochondrial antibody but normal alkaline phosphatase: is this primary biliary cirrhosis? HEPATOLOGY 1986;6: Metcalf, JV, Mitchison HC, Palmer JM, Jones DE, Bassendine MF, James OF. Natural history of early primary biliary cirrhosis. Lancet 1996;348: Kenny RP, Czaja AJ, Ludwig J, Dickson ER. Frequency and significance of antimitochondrial antibodies in severe chronic active hepatitis. Dig Dis Sci 1986;31: Czaja AJ, Carpenter HA, Manns MP. Antibodies to soluble liver antigen, P450IID6, and mitochondrial complexes in chronic hepatitis. Gastroenterology 1993;105: Czaja AJ. Frequency and nature of the variant syndromes of autoimmune liver disease. HEPATOLOGY 1998;28: Farias AQ, Goncalves LL, Bittencourt PL, De Melo ES, Abrantes-Lemos CP, Porta G, et al. Applicability of the IAIHG scoring system to the diagnosis of anitmitochondrial/anti-m2 seropositive variant form of autoimmune hepatitis. J Gastroenterol Hepatol 2006;21: Montano-Loza AJ, Carpenter HA, Czaja AJ. Frequency, behaviour and prognostic implications of antimitochondrial antibodies in type 1 autoimmune hepatitis. [Abstract 166]. Gastroenterology 2007;132(Suppl): A Nezu S, Tanaka A, Yasui H, Imamura M, Nakajima H, Ishida H, et al. Presence of antimitochondrial autoantibodies in patients with autoimmune hepatitis. J Gastroenterol Hepatol 2006;21: Leung PS, Rossaro L, Davis PA, Park O, Tanaka A, Kikuchi K, et al. Acute Liver Failure Study Group. Antimitochondrial antibodies in acute liver failure: implications for primary biliary cirrhosis. HEPATOLOGY 2007;46: Gregorio GV, Portmann B, Mowat AP, Vergani D, Mieli-Vergagni G. A 12-year-old girl with anitmitochondrial antibody positive autoimmune hepatitis. J Hepatol 1997;27: Weyman RL and M Voigt. Consecutive occurrence of primary biliary cirrhosis and autoimmune hepatitis: a case report and review of the literature. Am J Gastroenterol 2001;96: Poupon R, Chazouilleres O, Poupon RE. Chronic cholestatic diseases. J Hepatol 2000;32(Suppl 1): Mitchison HC, Palmer JM, Bassendine MF, Watson AJ, Record CO, James OF. A controlled trial of prednisolone in primary biliary cirrhosis: three year results. J Hepatol 1992;15: Leuschner M, Maier KP, Schlichting J, Strahl S, Herrmann G, Dahm HH, et al. Oral budesonide and ursodeoxycholic acid for treatment of primary biliary cirrhosis: results of a prospective double blind trial. Gastroenterology 1999;117: Rautiainen H, Karkkainen P, Karvonen AL, Nurmi H, Pikkarainen P, Nuutinen H, et al. Budesonide combined with UDCA to improve liver histology in primary biliary cirrhosis: a three year randomized trial. HEPA- TOLOGY 2005;41: