- Email: [email protected]
Prognosis of acute severe autoimmune hepatitis (AS-AIH): The role of corticosteroids in modifying outcome
Accepted Manuscript Prognosis of acute severe autoimmune hepatitis (AS-AIH): The role of corticosteroids in modifying outcome Andrew D. Yeoman, Rachel H. Westbrook, Yoh Zen, William Bernal, Thawab Al-Chalabi, Julia A. Wendon, John G. O’Grady, Michael A. Heneghan PII: DOI: Reference:
S0168-8278(14)00362-6 http://dx.doi.org/10.1016/j.jhep.2014.05.021 JHEPAT 5173
To appear in:
Journal of Hepatology
Received Date: Revised Date: Accepted Date:
24 September 2013 6 May 2014 9 May 2014
Please cite this article as: Yeoman, A.D., Westbrook, R.H., Zen, Y., Bernal, W., Al-Chalabi, T., Wendon, J.A., O’Grady, J.G., Heneghan, M.A., Prognosis of acute severe autoimmune hepatitis (AS-AIH): The role of corticosteroids in modifying outcome, Journal of Hepatology (2014), doi: http://dx.doi.org/10.1016/j.jhep. 2014.05.021
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Yeoman et al. Prognosis of acute severe autoimmune hepatitis (AS-AIH): The role of corticosteroids in modifying outcome.
Andrew D Yeoman, Rachel H Westbrook, Yoh Zen, William Bernal, Thawab AlChalabi, Julia A Wendon, John G O’Grady, Michael A Heneghan.
Corresponding author: Michael A Heneghan, Consultant Hepatologist & Reader in Hepatology, Institute of Liver Studies, Kings College Hospital, Denmark Hill, London SE5 9RS, United Kingdom. Fax: +44-203-2993167 Email: [email protected]
Institute of Liver Studies, Kings College Hospital, Denmark Hill, London SE5 9RS, United Kingdom.
Word Count: Abstract: 254 Manuscript without references, tables and figures: 3,665 References: 25 Tables: 4 Figures: 1
1
Yeoman et al. ABSTRACT: Background and aims: No standardised definition exists for acute, severe AIH (AS-AIH). However, rapid identification of AS-AIH and early corticosteroid therapy may prevent the need for liver transplantation (LT). We set out to determine the clinical outcomes of patients with AS-AIH presenting to our institution with particular focus on the role of corticosteroids. Methods: Retrospective analysis of a prospectively collated database identified patients presenting with AS-AIH from 1999-2009. We defined AS-AIH as an acute presentation with an INR of ≥1.5 at any time without histological evidence of cirrhosis. Results: 32 patients were identified with AS-AIH. Among the 32 AS-AIH patients 23 were treated with corticosteroids of whom 10 (48%) required LT, whilst all 9 untreated patients required LT (p= 0.01). Untreated patients demonstrated higher MELD scores at presentation (34 vs 28 p=0.01) and a non-significant decrease in episodes of sepsis but no difference in sepsis or mortality was observed between untreated or treated patients (11% vs 26% p=0.6 and 22% vs 17% p=0.99 respectively. Among treated patients, no difference in MELD scores was observed between responders or failures. Despite 59% undergoing LT, six deaths (19%) occurred. Conclusion: In a well characterised cohort of patients with AS-AIH, almost 60% required LT and 20% died. There was no difference in prognostic scores between steroid responders and failures and steroid exposure did not appear to jeopardise survival. Patients with AS-AIH should be considered for a trial of corticosteroids expediently whilst a thorough search for sepsis and assessment for LT should occur if clinical deterioration or encephalopathy develops.
2
Yeoman et al. INTRODUCTION: Autoimmune hepatitis (AIH) is a hepatic necro-inflammatory disorder which typically manifests as a chronic disease, although a proportion of patients may have an acute presentation. In all, approximately 20% of patients with AIH have an acute presentation which has been reported to be associated with a high incidence of acute liver failure (ALF) [1, 2]. The precise proportion of patients with AIH who develop ALF is, however, unknown. Additionally, such patients remain extremely difficult to identify from the outset of their clinical presentation. For example, in acute, severe AIH (AS-AIH), the classical autoimmune manifestations of positive anti-nuclear, anti-smooth muscle or anti liver kidney microsomal antibodies (ANA, SMA, LKM) and elevated gammaglobulins may not be apparent [3, 4]. As such several investigators have observed different histological patterns of liver injury in patients with AS-AIH compared with patients who present with chronic liver disease [1, 5]. The assessment of such patients is further hampered by the observation that some patients with an apparent denovo presentation of AIH have underlying chronic liver disease. This reinforces the critical role of early liver biopsy in the assessment of patients for whom AIH is considered as a possible diagnosis. Diagnostic criteria [6, 7] developed to facilitate clinical study of patients with AIH, were not designed for patients with acute, severe disease and have not been prospectively validated in this context. Finally, no standardised definition exists about what truly constitutes AS-AIH, representing a further barrier to clinical study of this critically important aspect of disease presentation. It has been reported that up to 20% of ALF in the United States is of indeterminate cause[8] and other investigators have demonstrated that approximately
3
Yeoman et al. 50% of “cryptogenic” ALF would meet diagnostic criteria for a probable diagnosis of AIH with detectable autoantibodies, hypergammaglobulinaemeia or characteristic patient demographics[9]. Therefore, a proportion of idiopathic ALF will have an autoimmune aetiology and these cases can frequently go unrecognised. Recently, there has been a surge of interest in this clinical entity [10, 11] culminating in an attempt to codify the diagnosis of autoimmune acute liver failure (AIALF) [12] which, historically, has been underestimated. One of the reasons for failure to define ALF due to AIH has been the lack of pathognomonic features of AIH and of a formal histological classification of ALF in this context. The US acute liver failure study group (USALF) report attempts to address this issue through the blinded review of liver tissue in 72 patients with indeterminate ALF, postulating that a significant proportion of these patients will have an autoimmune aetiology [12]. Based on this review, five patterns of massive hepatocellular necrosis (MHN) were identified. These are listed in figure 1. Two specific patterns (MHN 4 and 5) were felt to be more specific of an autoimmune aetiology. Other features such as plasma cell enrichment, central venulitis and lymphoid aggregates were also analysed to generate an assessment of overall histological appearance as being probable, compatible or not AIH and scored 2,1 or 0 as in the simplified IAIHG diagnostic criteria [7]. This report identified almost 60% of indeterminate ALF patients as having a probable autoimmune aetiology and, furthermore, discovered that these patients were more likely to have an AIH phenotype as characterised by female preponderance, presence of autoantibodies, raised immunoglobulins and development of chronic hepatitis when compared to patients without the specified autoimmune histological features [12].
4
Yeoman et al. We therefore set out to determine the frequency and clinical outcomes of patients presenting to our institution with AS-AIH and evidence of significant hepatic dysfunction, since these individuals are likely to be at very high risk of developing ALF. Furthermore, the role of corticosteroids in modifying such an acute, severe phenotype remains controversial with only limited data available from small, often poorly characterised, cohorts with varying inclusion criteria [13-15].
MATERIALS & METHODS: A prospectively collated database was retrospectively analyzed for all patients presenting with AS-AIH to our institution from the January 1999 to July 2009. This time period coincides with the introduction of an electronic patient record at our institution, thus facilitating the acquisition of complete patient datasets. Although the diagnosis of AIH was made clinically, all patients were evaluated as per the 1999 IAIHG diagnostic criteria for a diagnosis of probable or definite AIH.[6, 7] The 1999 IAIHG criteria were preferentially chosen as they are more comprehensive and previous work from this institution and others have shown they demonstrate greater sensitivity in acute presentations of AIH [16, 17]. A full serological screen was undertaken to exclude other causes of acute liver disease including assessment for viral hepatitis (A, B, C and E, Epstein Barr virus, Herpes simplex and Cytomegalovirus), Wilson’s disease. Drug induced hepatoxicity was excluded through a careful drug history. Ultrasound and/or computer tomography (CT) was performed to assess the hepatic parenchyma and vasculature and to exclude acute biliary pathology. In the absence of a standardised, accepted definition we determined AS-AIH as an acute presentation of disease (onset of symptoms to presentation of less than 26
5
Yeoman et al. weeks) in association with evidence of significant hepatic dysfunction as evidenced by an INR of ≥1.5 at any time during the index presentation, this being the same INR threshold used in the aforementioned USALF report [12]. In addition the absence of chronic liver disease on histological evaluation was a pre-requisite for a diagnosis of AS-AIH. The serum level of AST was not used as a marker of disease severity in this cohort as previous work from this institution has shown that those with a higher AST have a better prognosis [18] and also that, in acute disease, the early fall in AST post treatment does not adequately predict treatment responsiveness [19].
The Model for End Stage Liver Disease (MELD) score was originally developed to predict poor outcomes following trans-jugular intra-hepatic portosystemic shunts (TIPSS) [20]. However, its utility as a predictor of outcome has been extended across almost the whole spectrum of acute and chronic liver disease. In the United Kingdom, a similar scoring system, termed the United Kingdom End Stage Liver Disease (UKELD) [21] has been developed. In contrast to MELD, the UKELD incorporates serum sodium into its equation and reduces the weighting of creatinine in order to obtain its composite score. In AIH, MELD score has been shown to correlate with treatment responsiveness with non-responders demonstrating higher MELD scores at diagnosis than responders [22]. Furthermore, previous work from this institution has demonstrated that MELD and UKELD scores can predict corticosteroid treatment failure in icteric presentations of AIH [19]. MELD and UKELD scores were calculated on hospital admission and correlated with clinical outcomes which, for the purposes of this study, were the need for emergency LT (as a consequence of ALF), death and confirmed episodes of sepsis.
6
Yeoman et al. Patients with chronic liver disease on histology were then separately analysed to evaluate whether any significant differences exist between “apparent” AS-AIH and “true” AS-AIH. Finally, histological information, where available, was formally regraded by a single, experienced, hepatopathologist (YZ) in respect of the parameters felt to be indicative of autoimmune ALF by the recent USALF group report [12].
Statistical Analysis All data are reported as median and range or percentage with the exception of the comparison of patient characteristics between this cohort and that of the USALF group where mean and standard error of the mean (SEM) were used to facilitate direct comparison. In comparing groups, analysis of proportions was undertaken by the chisquared test, unless the number of variables in any category was less than 5, in which case Fisher’s exact test was used. Continuous variables were analysed by the MannWhitney U test on the basis that they were not normally distributed as determined by the Smirnov-Kolmogorov method. Results of statistical analysis were considered significant if the p value was equal or less than 0.05. All statistical analysis was performed using SPSS statistical package, version 15.0 (Chicago, Illinois, USA).
7
Yeoman et al. RESULTS: Based on our definition of AS-AIH, 47 patients were identified over the period 1999 to 2009. Among this cohort, 15/47 (32%) were noted to be cirrhotic on histological evaluation suggesting an acute presentation of a more chronic autoimmune disease process. The remaining 32 patients therefore comprise the ASAIH cohort. Patients with AS-AIH comprised 23 females and 9 males with a median age at presentation of 42 years. All patients met the 1999 revised IAIHG criteria for a probable (16/32, 50%) or definite (16/32, 50%) diagnosis of AIH. Eighty one percent of patients were Caucasian with the remainder being of Asian (7%) or Afro-Caribbean (13%) ethnicity.
Immunological features among patients with AS-AIH. Twenty eight patients (88%) tested positive for either ANA, SMA or LKM at titres ranging from 1:20 to 1:2,560. However, when IAIHG recommended autoantibody titre thresholds (≥1:40) were applied only 21 patients (66%), had positive autoantibodies and a lower proportion (17/32=53%) were noted to have elevated serum globulins.
Corticosteroid treated versus untreated patients in AS-AIH Twenty three patients (72%) were treated with corticosteroids, whereas 9 were untreated. In all 23 patients the interval between onset of jaundice and initiation of corticosteroids was less than 4 weeks. Oral prednisolone doses were 40 mg per day in 19 patients, 30mg in 1 and 20mg in another. Two patients were initiated on intravenous hydrocortisone 100mg tds at referring institutions. Among treated patients
8
Yeoman et al. just 10 (43%) responded to corticosteroid therapy. Consequently, 57% failed to respond to corticosteroid therapy and of these, 4 required addition of second line agents, consisting of tacrolimus in 3 patients (1mg per day in 2 patients, 1mg alternate days in another) and mycophenolate mofetil (500mg bd) in 1 patient. Half of these 4 patients responded with the other half requiring liver transplant. Conversely, all 9 untreated patients proceeded to LT. The laboratory, scoring system characteristics and clinical outcomes between corticosteroid treated and untreated patients are summarised in table 1. Those not treated had higher MELD scores at presentation (34 vs 28 p=0.01) and all untreated patients required emergency LT as compared to 43% of treated patients (p=0.004). Importantly, amongst treated patients, those that responded to corticosteroids demonstrated similar MELD and UKELD scores to those that failed corticosteroid treatment and required LT (26 vs 28 and 61 vs 63 p= 0.8 and 0.17 respectively). These results are summarised in table 2. For steroid responders, steroid tapering and introduction of steroid sparing treatment with azathioprine was undertaken according to established schedules as previously published from this institution [23]. During follow up (median duration 64 months, range 1-123) 2 patients died of complications of liver disease at 1 and 7 years after the index admission.
Acute Liver Failure in AS-AIH Since the cardinal diagnostic features of ALF necessitate the presence of hepatic encephalopathy (HE) we demonstrate that ALF, occurred in 22/32 (69%) of patients in the AS-AIH study cohort. All untreated patients developed HE with 4/9 (44%) doing so within 3 days of admission with a longest interval of 11 days. These
9
Yeoman et al. patients were considered too advanced for corticosteroid treatment and proceeded direct to LT.
In contrast, 5 patients developed HE prior to commencement of
corticosteroids, doing so at 0, 0, 2, 2 and 3 days prior to steroid commencement respectively. None of these individuals had HE greater than grade 2 based on the West-Haven criteria.[24] Of these 5 patients, 2 proceeded to LT and 3 subsequently responded to therapy without the need for LT and all 5 patients survived to hospital discharge. Comparison was also made between demographic, clinical and laboratory parameters in the AS-AIH cohort and those with indeterminate aetiology in the USALF study with the results displayed in table 3.
CLINICAL OUTCOMES IN AS-AIH Liver Transplantation Nineteen out of 32 patients (59%) with AS-AIH underwent emergency LT. The median interval between admission to our institution and LT was 5 days. There was no correlation between admission bilirubin, INR, creatinine, MELD or UKELD scores for those transplanted or not transplanted.
Episodes of Sepsis Overall, 7/32 patients (22%) developed a bacteriologically confirmed episode of sepsis. Among corticosteroid treated patients, there was an increased frequency of confirmed sepsis (6/23=26% VS 1/9=11%), but did not reach statistical significance (p=0.64).
Corticosteroid
treated
patients
prophylactically.
10
were
not
given
antibiotics
Yeoman et al. The median interval post steroid initiation was 41 days (range 23-204). Importantly all confirmed bouts of sepsis were observed in patients after the development of ALF, occurring post LT. Sepsis was due to methicillin resistant Staphylococcous aureus (MRSA) in 2 patients, methicillin sensitive Staphylococcus aureus in a further 2 patients and Escherichia coli, Stenotrophomonas maltophilia, multi-resistant Pseudomonas aeruginosa and vancomycin resistant Enterococcus faecalis (VRE) each in one patient.
Death Of 6 deaths in the cohort, all occurred post transplant at 3, 6, 14, 21, 22 and 121 days respectively. Among 4 patients who died following steroid exposure the interval between initiation and death was 23, 30, 38, and 140 days. Two deaths occurred in untreated patients 2/9 (22%) compared with 4/23 (17%) in patients treated with corticosteroids (p=0.99). Death occurred in 4 patients due to sepsis as the sole cause (2 MRSA bacteraemias, one due to Stenotrophomonas maltophilia and one due to a resistant Pseudomonas aeruginosa species). The other 2 patients died as a consequence of multi-organ failure without confirmed sepsis and a combination of severe early graft dysfunction post LT and VRE related sepsis. MELD and UKELD scores at diagnosis were higher in those patients who died than in survivors (32 vs 28, p= 0.03 and 65 vs 62, p=0.02).
Histological features among ALF patients in AS-AIH study cohort Among the AS-AIH cohort, 15 patients (47%) had a percutaneous biopsy, the remaining 17 patients either presented late in their course such that the clinical priority was assessment of LT (15 patients) or did not undergo biopsy (2). All
11
Yeoman et al. percutaneous biopsies were obtained prior to steroid treatment. Indeed, twenty two patients (69%) in the AS-AIH cohort developed ALF and 19/22 (86%) underwent emergency LT. All 22 had liver tissue available for histological review (6 percutaneous biopsies and 16 explants). Among explant specimens 7 patients had prior steroid exposure with a median interval of 5 days. Histological review of specimens in accordance with proposed histological classification for autoimmune AIALF[12] identified all as being either probable (7/22) or consistent (15/22) with AIH. Of individual histological features considered typical of AIH, plasma cell enrichment was the most consistent feature in this series, being noted in 17/22 (77%) cases. Lymphoid aggregates were observed in 16/22 (72%) and central venulitis in 15/22 (68%). However, a type 4 or 5 MHN pattern (suggested by the USALF group as being most specific for AIALF) was observed in only 5/22 cases (23%). Amongst the 22 patients with ALF, 5 had MHN type 5, 1 MHN type 4, 5MHN type 3, 10 MHN type 2, and 1 MHN type 1.
Comparison between AS-AIH and acute exacerbation presentations Median bilirubin was significantly higher in those with AS-AIH as opposed to those with an acute exacerbation type (463µmol/l vs 269 p=0.015). MELD and UKELD scores were also higher in AS-AIH (29 vs 23 p=0.005 and 62 vs 58 p=0.006 respectively). In contrast, globulin concentrations were significantly lower in AS-AIH patients (39g/dl vs 47/dl p=0.02). There were however, no significant differences between the two groups in the following parameters at diagnosis: median age - vs 40 AST, INR or creatinine. Just 1/15 (7%) of acute exacerbation type presentation
12
Yeoman et al. underwent liver transplant versus 19/32 (59%) in the AS-AIH cohort (p=0.001). These results are summarised in table 4.
DISCUSSION We report a well characterised group of patients with AIH and an acute, severe presentation in whom the development of acute liver failure is extremely high (69%). Whilst generally felt to represent an uncommon presentation of disease, AS-AIH as here defined, represented 32/289 (11%) of all AIH cases presenting to our tertiary institution over the study period. Furthermore, such a presentation was associated with the need for LT in almost 60% with an almost 20% risk of death. Whether earlier diagnosis and initiation of immunosuppressive therapy would obviate the need for LT is impossible to ascertain from this study and remains a pressing clinical question in respect of AS-AIH. Moreover, it is extremely difficult to comprehensively answer this question due to the rarity of AS-AIH and the relatively few series in the literature addressing it [13, 14]. Whilst Ichai and colleagues suggested that corticosteroids are of no benefit in severe or fulminant forms of AIH[13], this conclusion was based on a smaller series of 16 patients which were not as well characterised given neither IAIHG scores, nor liver histology appearances were reported. Of 12 corticosteroid treated patients in their series, 10 underwent LT and only 1 died of sepsis. Revealingly, of the 4 untreated patients, 3 required LT and 1 died of sepsis. Additionally, 2/12 patients had received steroids for >90 days, so would not meet accepted definitions of acute or sub-acute liver failure [25]. Consequently, their conclusions are difficult to justify on the basis of a single sepsis related death.
13
Yeoman et al. In contrast, three findings from our AS-AIH cohort support early diagnosis (utilising liver biopsy) and introduction of corticosteroids in AS-AIH. Firstly, untreated patients had significantly higher MELD scores than treated patients suggesting a later presentation and diagnosis at a more advanced stage of their clinical course. Also, all untreated patients who met the criteria of AS-AIH proceeded to develop HE and underwent LT. Secondly, analysis of corticosteroid responders reveals that MELD/UKELD scores were not significantly different between immunosuppression responders and failures who proceeded to LT. This suggests that there may be no finite point beyond which treatment can be considered futile. Support for this argument emerges from the observation that 2 patients already demonstrating HE were rescued from LT by corticosteroids. Whilst no futility threshold can be identified, previous work from this institution has shown that failure to improve prognostic scores within 7 days in icteric presentations of AIH indicates a group at high risk of ALF and should lead to the early consideration of alternative therapeutic strategies that may include LT [19]. Finally, exposure to corticosteroids did not increase mortality despite an increase in confirmed sepsis in treated patients and all 9 untreated patients undergoing LT. Thus death appears to occur due to the consequences of multi-organ dysfunction inherent upon the development ALF. The concept of early diagnosis and steroid use is also indirectly supported by the USALF study group who identified that a significant proportion of indeterminate ALF patients have autoimmune features which could potentially benefit from corticosteroids yet only 19% of this cohort were exposed to them [12]. A critical unanswered question in relation to corticosteroid use in this context is the optimal dose and route of administration. The steroid doses for
14
Yeoman et al. patients in this study were based according to standard practice within our institution or what was commenced at referring institutions. Whether weight basing and higher doses of steroids, via oral prednisolone or intravenous hydrocortisone administration, would lead to higher response rates and a reduced need for LT is unknown. Furthermore, the putative benefit of higher doses of steroids has to be balanced with the possibility of a higher risk of sepsis to the effect that a higher steroid dose strategy might actually be more harmful. However, due to the infrequency of acute severe presentations of AIH this question is likely to remain unanswered for the foreseeable future. The question of whether low grade HE represents an absolute contraindication to steroid therapy remains difficult to answer. However, among this cohort, 5 patients were treated with corticosteroids following the development of HE with a typically short interval. Although a strategy of corticosteroid treatment in low grade HE, with concurrent assessment for LT and de-listing with improvement cannot be formally recommended based on this study, clinical outcomes, albeit in n small subset, were not jeopardised by such a strategy. Indeed, 2 patients with existing HE treated with corticosteroids recovered without recourse to LT, lending further credence to the argument in favour of a corticosteroid treatment trial at the earliest opportunity. Additionally survival was not inferior in steroid treated patients and non-treatment had a high likelihood of resulting in LT. As to which features can be utilized to improve the diagnosis of AS-AIH, whilst accepting a lower titre of relevant autoantibodies, such as 1:20 would increase the sensitivity for diagnosing AIH, specificity would decrease and, importantly, such thresholds are not supported by the IAIHG [6]. Furthermore, IAIHG scores were identical between treated and non treated patients.
15
Yeoman et al. In terms of histological evaluation, plasma cell enrichment appears the most sensitive indicator of an autoimmune aetiology, (75% of cases), followed by central venulitis (64%). However, only 39% manifested MHN patterns 4 and 5 felt by the USALF as being most specific for the disease. Thus, the proposed MHN patterns appear to lack sensitivity. A significant proportion of histology assessments (72% Kings Cohort, 64% USALF cohort) were made on explant specimens and 33% of patients in this cohort and 19% of the USALF cohort were exposed to corticosteroids prior to histological evaluation. Caution must therefore be applied when extrapolating patterns observed in whole livers (following corticosteroid administration) to liver biopsy specimens in treatment naïve patients. Why so few patients in our cohort had liver biopsies is likely because many patients were referred or presented late to our institution and deemed too late for biopsy. This is suggested by the MELD, INR at admission/peak and need for LT all being significantly higher in untreated patients. However, 93% of those who had a liver biopsy as part of their diagnostic workup had a trial of steroids compared to just 53% of those not biopsied (p=0.04) reinforcing the importance of early consideration of biopsy in such patients.
Conclusion Patients with AS-AIH are at extremely high risk of progressing to ALF and requiring emergency LT. However it remains difficult to identify at diagnosis which patients will require LT and no threshold appears to exist beyond which treatment can be considered futile. This data suggests all patients should be considered for a trial of corticosteroids at the earliest opportunity. Such a strategy does not appear to
16
Yeoman et al. jeopardise patient survival but should involve a thorough search for sepsis and concurrent assessment for LT.
17
Yeoman et al.
References [1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10] [11]
[12]
[13]
[14]
Kessler, W.R., Cummings, O.W., Eckert, G., Chalasani, N., Lumeng, L., Kwo, P.Y. 2004. Fulminant hepatic failure as the initial presentation of acute autoimmune hepatitis. Clin Gastroenterol Hepatol, 2: 625-31. Ferrari, R., Pappas, G., Agostinelli, D., Muratori, P., Muratori, L., Lenzi, M., Verucchi, G., Cassani, F., Chiodo, F., Bianchi, F.B. 2004. Type 1 autoimmune hepatitis: patterns of clinical presentation and differential diagnosis of the 'acute' type. Qjm, 97: 407-12. Iwai, M., Jo, M., Ishii, M., Mori, T., Harada, Y. 2008. Comparison of clinical features and liver histology in acute and chronic autoimmune hepatitis. . Hepatol Res, 38: 784-789. Miyake, Y., Iwasaki, Y., Kobashi, H., T., Y., Ikeda, F., Takaki, A., Yamamoto, K. 2010. Autoimmune hepatitis with acute presentation in Japan. Digestive and Liver Disease, 42: 51-54. Lefkowitch, J.H., Apfelbaum, T.F., Weinberg, L., Forester, G. 1984. Acute liver biopsy lesions in early autoimmune ("lupoid") chronic active hepatitis. Liver, 4: 379-86. Alvarez, F., Berg, P.A., Bianchi, F.B., Bianchi, L., Burroughs, A.K., Cancado, E.L., Chapman, R.W., Cooksley, W.G., Czaja, A.J., Desmet, V.J. et al. 1999. International Autoimmune Hepatitis Group Report: review of criteria for diagnosis of autoimmune hepatitis. J Hepatol, 31: 929-38. Hennes, E.M., Zeniya, M., Czaja, A.J., Pares, A., Dalekos, G.N., Krawitt, E.L., Bittencourt, P.L., Porta, G., Boberg, K.M., Hofer, H. et al. 2008. Simplified criteria for the diagnosis of autoimmune hepatitis. Hepatology, 48: 169-176. Ostapowicz, G., Fontana, R.J., Schiodt, F.V., Larson, A., Davern, T.J., Han, S.H.B. 2002. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med, 137: 947-954. Bernal, W., Ma, Y., Smith, H.M., Portmann, B., Wendon, J., Vergani, D. 2007. The significance of autoantibodies and immunoglobulins in acute liver failure: a cohort study. Journal of Hepatology, 47: 664-670. Takahashi, H., Zeniya, M. 2011. Acute presntation of autoimmune hepatitis: Does it exist? A published work review. Hepatology Research, 41: 498-504. Fujiwara, K., Yasui, Y., Tawada, A., Fukuda, Y., Nakano, M.Y., O. 2011. Diagnostic value and utility of the simplified International AutoimmuneHepatitis Group criteria in acute-onset autoimmune hepatitis. Liver International, 31: 1013-1020. Stravitz, T.R., Lefkowitch, J., Fontana, R., Gershwin, M., Leung, P., Sterling, R., Manns, M., Norman, G., Lee, W. 2011. Autoimmune acute liver failure: Proposed clinical and histological criteria. Hepatology, 53: 517-526. Ichai, P., Duclos-Vallee, J.-C., Guettier, C., Hamida, S.B., Antonini, T., Delvart, V., et al. 2007. Usefulness of Corticosteroids for the Treatment of Severe and Fulminant Forms of Autoimmune Hepatitis. Liver Transplantation, 13: 996-1003. Miyake, Y., Iwasaki, Y., Terada, R., Onishi, T., Okamoto, R., Sakai, N., Sakaguchi, K., Shiratori, Y. 2006. Clinical characteristics of fulminant-type
18
Yeoman et al.
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
autoimmune hepatitis: an analysis of eleven cases. Aliment Pharmacol Ther, 23: 1347-53. Verma, S., Mahaeshwari, A., Thuluvath, P. 2009. Liver failure as initial presentation of autoimmune hepatitis: Clinical characteristics, predictors of response to corticosteroid therapy, and outcomes. Hepatology, 49: 1396-1397. Yeoman, A.D., Westbrook, R.H., Al-Chalabi, T., Carey, I., Heaton, N.D., Portmann, B.C., Heneghan, M.A. 2009. Diagnostic value and utility of the simplified International Autoimmune Hepatitis Group Criteria (IAIHG) in acute and chronic liver disease. Hepatology, 50: 538-545. Fujiwara, K., Yasui, S., Tawada, A., Fukuda, Y., Nakano, M., Yokosuka, O. 2011. Diagnostic Value and Utility of the Simplified International Autoimmune Hepatitis Group Criteria in Acute-onset Autoimmune Hepatitis. Liver International, 31: 1013-1020. Al-Chalabi, T., Underhill, J., Portmann, B.C., McFarlane, I.G., Heneghan, M.A. 2008. Effects of Serum Aspartate Transaminase Levels in Patients with Autoimmune Hepatitis Influence Disease Course and Outcome. Clin Gastroenterol and Hepatol, 6: 1389-1395. Yeoman, A.D., Westbrook, R.H., Zen, Y., Maninchedda, P., Portmann, B.C., Devlin, J., O'Grady, J.G., Harrison, P.M., Heneghan, M.A. 2011. Early predictors of corticosteroid treatment failure in icteric presentations of autoimmune hepatitis. Hepatology, 53: 926-934. Malinchoc, M., Kamath, P.S., Gordon, F.D., Peine, C.J., Rank, J., Borg, P.C. 2000. A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts. HEPATOLOGY, 31: 864-871. Neuberger, J., Gimson, A., Davies, M., Akyol, M., O’Grady, J.G., Burroughs, A., Hudson, M. 2008. Selection of patients for liver transplantation andallocation of donated livers in the UK. GUT, 57: 252-257. Montano-Loza, A.J., Carpenter, H.A., Czaja, A.J. 2007. Features Associated with Treatment Failure in Type 1 Autoimmune Hepatitis and Predictive Value of the Model of End-Stage Liver Disease. Hepatology, 46: 1138-1145. Al-Chalabi, T., Boccato, S., Portmann, B.C., McFarlane, I.G., Heneghan, M.A. 2006. Autoimmune hepatitis (AIH) in the elderly: a systematic retrospective analysis of a large group of consecutive patients with definite AIH followed at a tertiary referral centre. J Hepatol, 45: 575-83. Atterbury, C.E., Maddrey, W.C., Conn, H.O. 1978. Neomycin-sorbitol and lactulose in the treatment of acute portal-systemic encephalopathy. A controlled double blind clinical trial. Am J Dig Dis, 23: 398-406. O'Grady, J.G., Schalm, S.W., Williams, R. 1993. Acute Liver Failure: redefining the syndromes. The Lancet, 342: 273-275.
19
Yeoman et al. Figure 1. Proposed Histological Patterns Identified in Posssible Autoimmune Acute Liver Failure (AIALF)12.
Pattern of MHN
Characteristic Features
MHN 1
Panlobular necrosis and neocholangiolar proliferation
MHN 2
MHN1 plus regenerative nodules “submassive necrosis”
MHN 3
Severe acute hepatitis: Interface hepatitis, multilobular and bridging necrosis
MHN 4
MHN 1 with centrilobular haemorrhagic necrosis
MHN 5
Confluent necrosis superimposed on chronic hepatitis
Adapted from Stravitz et al12. MHN: massive hepatocellular necrosis
20
Yeoman et al. Table 3. Laboratory parameters, calculated prognostic scores, scoring system characteristics and clinical outcomes among treated and untreated patients in the study cohort. Parameter Bilirubin µmol/l (3-20) (range) AST mmol/l (10-50) (range) Globulin (g/l) (25-35) (range) IgG (g/l) Admission INR (0.9-1.2) (range) Peak INR (0.9-1.2) (range) Creatinine µmol/l (45-120) (range) MELD Score (range) UKELD Score (range) IAIHG Score (range) AutoAbs >1:80 Transplant Sepsis Death
Untreated n=9 477 (277-583) 779 (156-2690) 30 (23-47) 15.2 (14-23.8)
Treated n=23 453 (55-1208) 556 (116-2564) 41 (14-68) 27.1 (3.8-31.4)
p value
2.8 (1.8-4.7) 3.2 (2.2-5.0) 104 (78-226) 34 (25-40) 65 (59-69) 15 (11-18) 67% 100% 11% 22%
2.01 (1.5-3.5) 2.3 (1.5-4.5) 97 (43-210) 28 (16-40) 62 (56-67) 14 (10-19) 52% 43% 26% 17%
0.04
0.75 0.38 0.16 0.04
0.004 0.25 0.01 0.10 0.5 0.70 0.004 0.64 0.99
Abbreviations: AST, aspartate aminotrasnferase; INR, international normalized ratio; MELD, Model for End Stage Liver Disease; UKELD, United Kingdom End Stage Liver Disease; IAIHG, International Autoimmune Hepatitis Group; µmol/l, micromoles per litre; mmol/l, millimoles per liter; g/l, grams per litre.
21
Yeoman et al. Table 2: Laboratory parameters, prognostic indicators and outcomes among treatment responders and treatment failures Parameter
Immunosuppressive Therapy Responder n=13 447 (197-1208) 397 (173-1420) 31 (14-68) 1.9 (1.5-2.4) 104 (43-210) 26 (22-40) 61 (56-66)
Immunosuppressive Therapy Failure n=10 475 (55-898) 829 (58-2564) 42 (27-52) 2.3 (1.6-3.5) 89 (64-137) 28 (16-33) 63 (60-67)
p value
Liver transplant
0%
100%
<0.0001
Death
0%
40%
0.03
Bilirubin µmol/l (3-20) (range) AST mmol/l (10-50) (range) Globulin g/l (25-35) (range) Admission INR (0.9-1.2) (range) Creatinine µmol/l (45-120) (range) MELD Score (range) UKELD Score (range)
0.42 0.15 0.48 0.30 0.23 0.80 0.17
Abbreviations: AST, aspartate aminotrasnferase; INR, international normalized ratio; MELD, Model of End Stage Liver Disease; UKELD, United Kingdom End Stage Liver Disease; µmol/l, micromoles per liter; mmol/l, millimoles per litre; g/l, grams per litre.
22
Yeoman et al. Table 3. Comparsion of demographic, laboratory and clinical characteristics between Kings AS-AIH cohort with ALF and USALF indeterminate cohort.
Parameter Age Females White:Non White Bilirubin µmol/l AST mmol/l ALP U/l INR Creatinine µmol/l Globulins g/l Positive autoantibodies Steroid exposure Survival
King’s Cohort n= 22 41 ± 3.4 91% 77% 514 ± 48 975 ± 179 200 ± 21 2.7 ± 0.2 106 ± 9 37 ± 2.0 77% 33% 73%
USALF Cohort n=72 41 ± 1.7 58% 67% 422 ± 22 1231 ± 197 205 ± 15 3.2 ± 0.2 159 ± 18 34 ± 0.4 66% 19% 71%
Data expressed as mean ± standard error Abbreviations: AST, aspartate aminotrasnferase; ALP, alkaline phosphatise; INR, International Normalized Ratio; MELD, Model for End Stage Liver Disease; UKELD, United Kingdom End Stage Liver Disease; IAIHG, International Autoimmune Hepatitis Group; mg/dl, milligrams per decilitre; U/l, Units per litre; mmol/l, millimoles per litre; g/l, grams per litre.
23
Yeoman et al. Table 4. Variables recorded at presentation between patients with AS-AIH and acute exacerbation type presentations.
Parameter Age (range) Bilirubin µmol/l (3-20) (range) AST mmol/l (10-50) (range) INR (0.9-1.2) (range) Globulin g/l (25-35) (range) Autoantibodies (≥1:80) % Creatinine µmol/l (45-120) (range) MELD Score (range) UKELD Score (range) IAIHG Score Death Transplant
AS-AIH n= 32 42 (16-68) 463 (55-1208) 605 (58-2690) 2.2 (1.5-3.5) 39 (22-68) 56%
Acute Exacerbation Type n=15 42 (17-60) 269 (77-574) 706 (188-1712) 1.9 (1.6-2.6) 47 (24-81) 80%
p value
101 (26-226) 29 (22-40) 62 (55-69) 15 (10-19) 19%
98 (69-142) 23 (16-35) 58 (55-67) 17 (10-22) 0%
0.8
59%
7%
Abbreviations: AST, aspartate aminotrasnferase; INR, international normalized ratio; MELD, Model of End Stage Liver Disease; UKELD, United Kingdom End Stage Liver Disease; µmol/l, micromoles per litre; mmol/l, millimoles per litre; g/l, grams per litre.
24
0.30 0.015 0.85 0.13 0.02 0.19
0.005 0.006 0.06 0.15 0.001
S0168-8278(14)00362-6 http://dx.doi.org/10.1016/j.jhep.2014.05.021 JHEPAT 5173
To appear in:
Journal of Hepatology
Received Date: Revised Date: Accepted Date:
24 September 2013 6 May 2014 9 May 2014
Please cite this article as: Yeoman, A.D., Westbrook, R.H., Zen, Y., Bernal, W., Al-Chalabi, T., Wendon, J.A., O’Grady, J.G., Heneghan, M.A., Prognosis of acute severe autoimmune hepatitis (AS-AIH): The role of corticosteroids in modifying outcome, Journal of Hepatology (2014), doi: http://dx.doi.org/10.1016/j.jhep. 2014.05.021
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Yeoman et al. Prognosis of acute severe autoimmune hepatitis (AS-AIH): The role of corticosteroids in modifying outcome.
Andrew D Yeoman, Rachel H Westbrook, Yoh Zen, William Bernal, Thawab AlChalabi, Julia A Wendon, John G O’Grady, Michael A Heneghan.
Corresponding author: Michael A Heneghan, Consultant Hepatologist & Reader in Hepatology, Institute of Liver Studies, Kings College Hospital, Denmark Hill, London SE5 9RS, United Kingdom. Fax: +44-203-2993167 Email: [email protected]
Institute of Liver Studies, Kings College Hospital, Denmark Hill, London SE5 9RS, United Kingdom.
Word Count: Abstract: 254 Manuscript without references, tables and figures: 3,665 References: 25 Tables: 4 Figures: 1
1
Yeoman et al. ABSTRACT: Background and aims: No standardised definition exists for acute, severe AIH (AS-AIH). However, rapid identification of AS-AIH and early corticosteroid therapy may prevent the need for liver transplantation (LT). We set out to determine the clinical outcomes of patients with AS-AIH presenting to our institution with particular focus on the role of corticosteroids. Methods: Retrospective analysis of a prospectively collated database identified patients presenting with AS-AIH from 1999-2009. We defined AS-AIH as an acute presentation with an INR of ≥1.5 at any time without histological evidence of cirrhosis. Results: 32 patients were identified with AS-AIH. Among the 32 AS-AIH patients 23 were treated with corticosteroids of whom 10 (48%) required LT, whilst all 9 untreated patients required LT (p= 0.01). Untreated patients demonstrated higher MELD scores at presentation (34 vs 28 p=0.01) and a non-significant decrease in episodes of sepsis but no difference in sepsis or mortality was observed between untreated or treated patients (11% vs 26% p=0.6 and 22% vs 17% p=0.99 respectively. Among treated patients, no difference in MELD scores was observed between responders or failures. Despite 59% undergoing LT, six deaths (19%) occurred. Conclusion: In a well characterised cohort of patients with AS-AIH, almost 60% required LT and 20% died. There was no difference in prognostic scores between steroid responders and failures and steroid exposure did not appear to jeopardise survival. Patients with AS-AIH should be considered for a trial of corticosteroids expediently whilst a thorough search for sepsis and assessment for LT should occur if clinical deterioration or encephalopathy develops.
2
Yeoman et al. INTRODUCTION: Autoimmune hepatitis (AIH) is a hepatic necro-inflammatory disorder which typically manifests as a chronic disease, although a proportion of patients may have an acute presentation. In all, approximately 20% of patients with AIH have an acute presentation which has been reported to be associated with a high incidence of acute liver failure (ALF) [1, 2]. The precise proportion of patients with AIH who develop ALF is, however, unknown. Additionally, such patients remain extremely difficult to identify from the outset of their clinical presentation. For example, in acute, severe AIH (AS-AIH), the classical autoimmune manifestations of positive anti-nuclear, anti-smooth muscle or anti liver kidney microsomal antibodies (ANA, SMA, LKM) and elevated gammaglobulins may not be apparent [3, 4]. As such several investigators have observed different histological patterns of liver injury in patients with AS-AIH compared with patients who present with chronic liver disease [1, 5]. The assessment of such patients is further hampered by the observation that some patients with an apparent denovo presentation of AIH have underlying chronic liver disease. This reinforces the critical role of early liver biopsy in the assessment of patients for whom AIH is considered as a possible diagnosis. Diagnostic criteria [6, 7] developed to facilitate clinical study of patients with AIH, were not designed for patients with acute, severe disease and have not been prospectively validated in this context. Finally, no standardised definition exists about what truly constitutes AS-AIH, representing a further barrier to clinical study of this critically important aspect of disease presentation. It has been reported that up to 20% of ALF in the United States is of indeterminate cause[8] and other investigators have demonstrated that approximately
3
Yeoman et al. 50% of “cryptogenic” ALF would meet diagnostic criteria for a probable diagnosis of AIH with detectable autoantibodies, hypergammaglobulinaemeia or characteristic patient demographics[9]. Therefore, a proportion of idiopathic ALF will have an autoimmune aetiology and these cases can frequently go unrecognised. Recently, there has been a surge of interest in this clinical entity [10, 11] culminating in an attempt to codify the diagnosis of autoimmune acute liver failure (AIALF) [12] which, historically, has been underestimated. One of the reasons for failure to define ALF due to AIH has been the lack of pathognomonic features of AIH and of a formal histological classification of ALF in this context. The US acute liver failure study group (USALF) report attempts to address this issue through the blinded review of liver tissue in 72 patients with indeterminate ALF, postulating that a significant proportion of these patients will have an autoimmune aetiology [12]. Based on this review, five patterns of massive hepatocellular necrosis (MHN) were identified. These are listed in figure 1. Two specific patterns (MHN 4 and 5) were felt to be more specific of an autoimmune aetiology. Other features such as plasma cell enrichment, central venulitis and lymphoid aggregates were also analysed to generate an assessment of overall histological appearance as being probable, compatible or not AIH and scored 2,1 or 0 as in the simplified IAIHG diagnostic criteria [7]. This report identified almost 60% of indeterminate ALF patients as having a probable autoimmune aetiology and, furthermore, discovered that these patients were more likely to have an AIH phenotype as characterised by female preponderance, presence of autoantibodies, raised immunoglobulins and development of chronic hepatitis when compared to patients without the specified autoimmune histological features [12].
4
Yeoman et al. We therefore set out to determine the frequency and clinical outcomes of patients presenting to our institution with AS-AIH and evidence of significant hepatic dysfunction, since these individuals are likely to be at very high risk of developing ALF. Furthermore, the role of corticosteroids in modifying such an acute, severe phenotype remains controversial with only limited data available from small, often poorly characterised, cohorts with varying inclusion criteria [13-15].
MATERIALS & METHODS: A prospectively collated database was retrospectively analyzed for all patients presenting with AS-AIH to our institution from the January 1999 to July 2009. This time period coincides with the introduction of an electronic patient record at our institution, thus facilitating the acquisition of complete patient datasets. Although the diagnosis of AIH was made clinically, all patients were evaluated as per the 1999 IAIHG diagnostic criteria for a diagnosis of probable or definite AIH.[6, 7] The 1999 IAIHG criteria were preferentially chosen as they are more comprehensive and previous work from this institution and others have shown they demonstrate greater sensitivity in acute presentations of AIH [16, 17]. A full serological screen was undertaken to exclude other causes of acute liver disease including assessment for viral hepatitis (A, B, C and E, Epstein Barr virus, Herpes simplex and Cytomegalovirus), Wilson’s disease. Drug induced hepatoxicity was excluded through a careful drug history. Ultrasound and/or computer tomography (CT) was performed to assess the hepatic parenchyma and vasculature and to exclude acute biliary pathology. In the absence of a standardised, accepted definition we determined AS-AIH as an acute presentation of disease (onset of symptoms to presentation of less than 26
5
Yeoman et al. weeks) in association with evidence of significant hepatic dysfunction as evidenced by an INR of ≥1.5 at any time during the index presentation, this being the same INR threshold used in the aforementioned USALF report [12]. In addition the absence of chronic liver disease on histological evaluation was a pre-requisite for a diagnosis of AS-AIH. The serum level of AST was not used as a marker of disease severity in this cohort as previous work from this institution has shown that those with a higher AST have a better prognosis [18] and also that, in acute disease, the early fall in AST post treatment does not adequately predict treatment responsiveness [19].
The Model for End Stage Liver Disease (MELD) score was originally developed to predict poor outcomes following trans-jugular intra-hepatic portosystemic shunts (TIPSS) [20]. However, its utility as a predictor of outcome has been extended across almost the whole spectrum of acute and chronic liver disease. In the United Kingdom, a similar scoring system, termed the United Kingdom End Stage Liver Disease (UKELD) [21] has been developed. In contrast to MELD, the UKELD incorporates serum sodium into its equation and reduces the weighting of creatinine in order to obtain its composite score. In AIH, MELD score has been shown to correlate with treatment responsiveness with non-responders demonstrating higher MELD scores at diagnosis than responders [22]. Furthermore, previous work from this institution has demonstrated that MELD and UKELD scores can predict corticosteroid treatment failure in icteric presentations of AIH [19]. MELD and UKELD scores were calculated on hospital admission and correlated with clinical outcomes which, for the purposes of this study, were the need for emergency LT (as a consequence of ALF), death and confirmed episodes of sepsis.
6
Yeoman et al. Patients with chronic liver disease on histology were then separately analysed to evaluate whether any significant differences exist between “apparent” AS-AIH and “true” AS-AIH. Finally, histological information, where available, was formally regraded by a single, experienced, hepatopathologist (YZ) in respect of the parameters felt to be indicative of autoimmune ALF by the recent USALF group report [12].
Statistical Analysis All data are reported as median and range or percentage with the exception of the comparison of patient characteristics between this cohort and that of the USALF group where mean and standard error of the mean (SEM) were used to facilitate direct comparison. In comparing groups, analysis of proportions was undertaken by the chisquared test, unless the number of variables in any category was less than 5, in which case Fisher’s exact test was used. Continuous variables were analysed by the MannWhitney U test on the basis that they were not normally distributed as determined by the Smirnov-Kolmogorov method. Results of statistical analysis were considered significant if the p value was equal or less than 0.05. All statistical analysis was performed using SPSS statistical package, version 15.0 (Chicago, Illinois, USA).
7
Yeoman et al. RESULTS: Based on our definition of AS-AIH, 47 patients were identified over the period 1999 to 2009. Among this cohort, 15/47 (32%) were noted to be cirrhotic on histological evaluation suggesting an acute presentation of a more chronic autoimmune disease process. The remaining 32 patients therefore comprise the ASAIH cohort. Patients with AS-AIH comprised 23 females and 9 males with a median age at presentation of 42 years. All patients met the 1999 revised IAIHG criteria for a probable (16/32, 50%) or definite (16/32, 50%) diagnosis of AIH. Eighty one percent of patients were Caucasian with the remainder being of Asian (7%) or Afro-Caribbean (13%) ethnicity.
Immunological features among patients with AS-AIH. Twenty eight patients (88%) tested positive for either ANA, SMA or LKM at titres ranging from 1:20 to 1:2,560. However, when IAIHG recommended autoantibody titre thresholds (≥1:40) were applied only 21 patients (66%), had positive autoantibodies and a lower proportion (17/32=53%) were noted to have elevated serum globulins.
Corticosteroid treated versus untreated patients in AS-AIH Twenty three patients (72%) were treated with corticosteroids, whereas 9 were untreated. In all 23 patients the interval between onset of jaundice and initiation of corticosteroids was less than 4 weeks. Oral prednisolone doses were 40 mg per day in 19 patients, 30mg in 1 and 20mg in another. Two patients were initiated on intravenous hydrocortisone 100mg tds at referring institutions. Among treated patients
8
Yeoman et al. just 10 (43%) responded to corticosteroid therapy. Consequently, 57% failed to respond to corticosteroid therapy and of these, 4 required addition of second line agents, consisting of tacrolimus in 3 patients (1mg per day in 2 patients, 1mg alternate days in another) and mycophenolate mofetil (500mg bd) in 1 patient. Half of these 4 patients responded with the other half requiring liver transplant. Conversely, all 9 untreated patients proceeded to LT. The laboratory, scoring system characteristics and clinical outcomes between corticosteroid treated and untreated patients are summarised in table 1. Those not treated had higher MELD scores at presentation (34 vs 28 p=0.01) and all untreated patients required emergency LT as compared to 43% of treated patients (p=0.004). Importantly, amongst treated patients, those that responded to corticosteroids demonstrated similar MELD and UKELD scores to those that failed corticosteroid treatment and required LT (26 vs 28 and 61 vs 63 p= 0.8 and 0.17 respectively). These results are summarised in table 2. For steroid responders, steroid tapering and introduction of steroid sparing treatment with azathioprine was undertaken according to established schedules as previously published from this institution [23]. During follow up (median duration 64 months, range 1-123) 2 patients died of complications of liver disease at 1 and 7 years after the index admission.
Acute Liver Failure in AS-AIH Since the cardinal diagnostic features of ALF necessitate the presence of hepatic encephalopathy (HE) we demonstrate that ALF, occurred in 22/32 (69%) of patients in the AS-AIH study cohort. All untreated patients developed HE with 4/9 (44%) doing so within 3 days of admission with a longest interval of 11 days. These
9
Yeoman et al. patients were considered too advanced for corticosteroid treatment and proceeded direct to LT.
In contrast, 5 patients developed HE prior to commencement of
corticosteroids, doing so at 0, 0, 2, 2 and 3 days prior to steroid commencement respectively. None of these individuals had HE greater than grade 2 based on the West-Haven criteria.[24] Of these 5 patients, 2 proceeded to LT and 3 subsequently responded to therapy without the need for LT and all 5 patients survived to hospital discharge. Comparison was also made between demographic, clinical and laboratory parameters in the AS-AIH cohort and those with indeterminate aetiology in the USALF study with the results displayed in table 3.
CLINICAL OUTCOMES IN AS-AIH Liver Transplantation Nineteen out of 32 patients (59%) with AS-AIH underwent emergency LT. The median interval between admission to our institution and LT was 5 days. There was no correlation between admission bilirubin, INR, creatinine, MELD or UKELD scores for those transplanted or not transplanted.
Episodes of Sepsis Overall, 7/32 patients (22%) developed a bacteriologically confirmed episode of sepsis. Among corticosteroid treated patients, there was an increased frequency of confirmed sepsis (6/23=26% VS 1/9=11%), but did not reach statistical significance (p=0.64).
Corticosteroid
treated
patients
prophylactically.
10
were
not
given
antibiotics
Yeoman et al. The median interval post steroid initiation was 41 days (range 23-204). Importantly all confirmed bouts of sepsis were observed in patients after the development of ALF, occurring post LT. Sepsis was due to methicillin resistant Staphylococcous aureus (MRSA) in 2 patients, methicillin sensitive Staphylococcus aureus in a further 2 patients and Escherichia coli, Stenotrophomonas maltophilia, multi-resistant Pseudomonas aeruginosa and vancomycin resistant Enterococcus faecalis (VRE) each in one patient.
Death Of 6 deaths in the cohort, all occurred post transplant at 3, 6, 14, 21, 22 and 121 days respectively. Among 4 patients who died following steroid exposure the interval between initiation and death was 23, 30, 38, and 140 days. Two deaths occurred in untreated patients 2/9 (22%) compared with 4/23 (17%) in patients treated with corticosteroids (p=0.99). Death occurred in 4 patients due to sepsis as the sole cause (2 MRSA bacteraemias, one due to Stenotrophomonas maltophilia and one due to a resistant Pseudomonas aeruginosa species). The other 2 patients died as a consequence of multi-organ failure without confirmed sepsis and a combination of severe early graft dysfunction post LT and VRE related sepsis. MELD and UKELD scores at diagnosis were higher in those patients who died than in survivors (32 vs 28, p= 0.03 and 65 vs 62, p=0.02).
Histological features among ALF patients in AS-AIH study cohort Among the AS-AIH cohort, 15 patients (47%) had a percutaneous biopsy, the remaining 17 patients either presented late in their course such that the clinical priority was assessment of LT (15 patients) or did not undergo biopsy (2). All
11
Yeoman et al. percutaneous biopsies were obtained prior to steroid treatment. Indeed, twenty two patients (69%) in the AS-AIH cohort developed ALF and 19/22 (86%) underwent emergency LT. All 22 had liver tissue available for histological review (6 percutaneous biopsies and 16 explants). Among explant specimens 7 patients had prior steroid exposure with a median interval of 5 days. Histological review of specimens in accordance with proposed histological classification for autoimmune AIALF[12] identified all as being either probable (7/22) or consistent (15/22) with AIH. Of individual histological features considered typical of AIH, plasma cell enrichment was the most consistent feature in this series, being noted in 17/22 (77%) cases. Lymphoid aggregates were observed in 16/22 (72%) and central venulitis in 15/22 (68%). However, a type 4 or 5 MHN pattern (suggested by the USALF group as being most specific for AIALF) was observed in only 5/22 cases (23%). Amongst the 22 patients with ALF, 5 had MHN type 5, 1 MHN type 4, 5MHN type 3, 10 MHN type 2, and 1 MHN type 1.
Comparison between AS-AIH and acute exacerbation presentations Median bilirubin was significantly higher in those with AS-AIH as opposed to those with an acute exacerbation type (463µmol/l vs 269 p=0.015). MELD and UKELD scores were also higher in AS-AIH (29 vs 23 p=0.005 and 62 vs 58 p=0.006 respectively). In contrast, globulin concentrations were significantly lower in AS-AIH patients (39g/dl vs 47/dl p=0.02). There were however, no significant differences between the two groups in the following parameters at diagnosis: median age - vs 40 AST, INR or creatinine. Just 1/15 (7%) of acute exacerbation type presentation
12
Yeoman et al. underwent liver transplant versus 19/32 (59%) in the AS-AIH cohort (p=0.001). These results are summarised in table 4.
DISCUSSION We report a well characterised group of patients with AIH and an acute, severe presentation in whom the development of acute liver failure is extremely high (69%). Whilst generally felt to represent an uncommon presentation of disease, AS-AIH as here defined, represented 32/289 (11%) of all AIH cases presenting to our tertiary institution over the study period. Furthermore, such a presentation was associated with the need for LT in almost 60% with an almost 20% risk of death. Whether earlier diagnosis and initiation of immunosuppressive therapy would obviate the need for LT is impossible to ascertain from this study and remains a pressing clinical question in respect of AS-AIH. Moreover, it is extremely difficult to comprehensively answer this question due to the rarity of AS-AIH and the relatively few series in the literature addressing it [13, 14]. Whilst Ichai and colleagues suggested that corticosteroids are of no benefit in severe or fulminant forms of AIH[13], this conclusion was based on a smaller series of 16 patients which were not as well characterised given neither IAIHG scores, nor liver histology appearances were reported. Of 12 corticosteroid treated patients in their series, 10 underwent LT and only 1 died of sepsis. Revealingly, of the 4 untreated patients, 3 required LT and 1 died of sepsis. Additionally, 2/12 patients had received steroids for >90 days, so would not meet accepted definitions of acute or sub-acute liver failure [25]. Consequently, their conclusions are difficult to justify on the basis of a single sepsis related death.
13
Yeoman et al. In contrast, three findings from our AS-AIH cohort support early diagnosis (utilising liver biopsy) and introduction of corticosteroids in AS-AIH. Firstly, untreated patients had significantly higher MELD scores than treated patients suggesting a later presentation and diagnosis at a more advanced stage of their clinical course. Also, all untreated patients who met the criteria of AS-AIH proceeded to develop HE and underwent LT. Secondly, analysis of corticosteroid responders reveals that MELD/UKELD scores were not significantly different between immunosuppression responders and failures who proceeded to LT. This suggests that there may be no finite point beyond which treatment can be considered futile. Support for this argument emerges from the observation that 2 patients already demonstrating HE were rescued from LT by corticosteroids. Whilst no futility threshold can be identified, previous work from this institution has shown that failure to improve prognostic scores within 7 days in icteric presentations of AIH indicates a group at high risk of ALF and should lead to the early consideration of alternative therapeutic strategies that may include LT [19]. Finally, exposure to corticosteroids did not increase mortality despite an increase in confirmed sepsis in treated patients and all 9 untreated patients undergoing LT. Thus death appears to occur due to the consequences of multi-organ dysfunction inherent upon the development ALF. The concept of early diagnosis and steroid use is also indirectly supported by the USALF study group who identified that a significant proportion of indeterminate ALF patients have autoimmune features which could potentially benefit from corticosteroids yet only 19% of this cohort were exposed to them [12]. A critical unanswered question in relation to corticosteroid use in this context is the optimal dose and route of administration. The steroid doses for
14
Yeoman et al. patients in this study were based according to standard practice within our institution or what was commenced at referring institutions. Whether weight basing and higher doses of steroids, via oral prednisolone or intravenous hydrocortisone administration, would lead to higher response rates and a reduced need for LT is unknown. Furthermore, the putative benefit of higher doses of steroids has to be balanced with the possibility of a higher risk of sepsis to the effect that a higher steroid dose strategy might actually be more harmful. However, due to the infrequency of acute severe presentations of AIH this question is likely to remain unanswered for the foreseeable future. The question of whether low grade HE represents an absolute contraindication to steroid therapy remains difficult to answer. However, among this cohort, 5 patients were treated with corticosteroids following the development of HE with a typically short interval. Although a strategy of corticosteroid treatment in low grade HE, with concurrent assessment for LT and de-listing with improvement cannot be formally recommended based on this study, clinical outcomes, albeit in n small subset, were not jeopardised by such a strategy. Indeed, 2 patients with existing HE treated with corticosteroids recovered without recourse to LT, lending further credence to the argument in favour of a corticosteroid treatment trial at the earliest opportunity. Additionally survival was not inferior in steroid treated patients and non-treatment had a high likelihood of resulting in LT. As to which features can be utilized to improve the diagnosis of AS-AIH, whilst accepting a lower titre of relevant autoantibodies, such as 1:20 would increase the sensitivity for diagnosing AIH, specificity would decrease and, importantly, such thresholds are not supported by the IAIHG [6]. Furthermore, IAIHG scores were identical between treated and non treated patients.
15
Yeoman et al. In terms of histological evaluation, plasma cell enrichment appears the most sensitive indicator of an autoimmune aetiology, (75% of cases), followed by central venulitis (64%). However, only 39% manifested MHN patterns 4 and 5 felt by the USALF as being most specific for the disease. Thus, the proposed MHN patterns appear to lack sensitivity. A significant proportion of histology assessments (72% Kings Cohort, 64% USALF cohort) were made on explant specimens and 33% of patients in this cohort and 19% of the USALF cohort were exposed to corticosteroids prior to histological evaluation. Caution must therefore be applied when extrapolating patterns observed in whole livers (following corticosteroid administration) to liver biopsy specimens in treatment naïve patients. Why so few patients in our cohort had liver biopsies is likely because many patients were referred or presented late to our institution and deemed too late for biopsy. This is suggested by the MELD, INR at admission/peak and need for LT all being significantly higher in untreated patients. However, 93% of those who had a liver biopsy as part of their diagnostic workup had a trial of steroids compared to just 53% of those not biopsied (p=0.04) reinforcing the importance of early consideration of biopsy in such patients.
Conclusion Patients with AS-AIH are at extremely high risk of progressing to ALF and requiring emergency LT. However it remains difficult to identify at diagnosis which patients will require LT and no threshold appears to exist beyond which treatment can be considered futile. This data suggests all patients should be considered for a trial of corticosteroids at the earliest opportunity. Such a strategy does not appear to
16
Yeoman et al. jeopardise patient survival but should involve a thorough search for sepsis and concurrent assessment for LT.
17
Yeoman et al.
References [1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10] [11]
[12]
[13]
[14]
Kessler, W.R., Cummings, O.W., Eckert, G., Chalasani, N., Lumeng, L., Kwo, P.Y. 2004. Fulminant hepatic failure as the initial presentation of acute autoimmune hepatitis. Clin Gastroenterol Hepatol, 2: 625-31. Ferrari, R., Pappas, G., Agostinelli, D., Muratori, P., Muratori, L., Lenzi, M., Verucchi, G., Cassani, F., Chiodo, F., Bianchi, F.B. 2004. Type 1 autoimmune hepatitis: patterns of clinical presentation and differential diagnosis of the 'acute' type. Qjm, 97: 407-12. Iwai, M., Jo, M., Ishii, M., Mori, T., Harada, Y. 2008. Comparison of clinical features and liver histology in acute and chronic autoimmune hepatitis. . Hepatol Res, 38: 784-789. Miyake, Y., Iwasaki, Y., Kobashi, H., T., Y., Ikeda, F., Takaki, A., Yamamoto, K. 2010. Autoimmune hepatitis with acute presentation in Japan. Digestive and Liver Disease, 42: 51-54. Lefkowitch, J.H., Apfelbaum, T.F., Weinberg, L., Forester, G. 1984. Acute liver biopsy lesions in early autoimmune ("lupoid") chronic active hepatitis. Liver, 4: 379-86. Alvarez, F., Berg, P.A., Bianchi, F.B., Bianchi, L., Burroughs, A.K., Cancado, E.L., Chapman, R.W., Cooksley, W.G., Czaja, A.J., Desmet, V.J. et al. 1999. International Autoimmune Hepatitis Group Report: review of criteria for diagnosis of autoimmune hepatitis. J Hepatol, 31: 929-38. Hennes, E.M., Zeniya, M., Czaja, A.J., Pares, A., Dalekos, G.N., Krawitt, E.L., Bittencourt, P.L., Porta, G., Boberg, K.M., Hofer, H. et al. 2008. Simplified criteria for the diagnosis of autoimmune hepatitis. Hepatology, 48: 169-176. Ostapowicz, G., Fontana, R.J., Schiodt, F.V., Larson, A., Davern, T.J., Han, S.H.B. 2002. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med, 137: 947-954. Bernal, W., Ma, Y., Smith, H.M., Portmann, B., Wendon, J., Vergani, D. 2007. The significance of autoantibodies and immunoglobulins in acute liver failure: a cohort study. Journal of Hepatology, 47: 664-670. Takahashi, H., Zeniya, M. 2011. Acute presntation of autoimmune hepatitis: Does it exist? A published work review. Hepatology Research, 41: 498-504. Fujiwara, K., Yasui, Y., Tawada, A., Fukuda, Y., Nakano, M.Y., O. 2011. Diagnostic value and utility of the simplified International AutoimmuneHepatitis Group criteria in acute-onset autoimmune hepatitis. Liver International, 31: 1013-1020. Stravitz, T.R., Lefkowitch, J., Fontana, R., Gershwin, M., Leung, P., Sterling, R., Manns, M., Norman, G., Lee, W. 2011. Autoimmune acute liver failure: Proposed clinical and histological criteria. Hepatology, 53: 517-526. Ichai, P., Duclos-Vallee, J.-C., Guettier, C., Hamida, S.B., Antonini, T., Delvart, V., et al. 2007. Usefulness of Corticosteroids for the Treatment of Severe and Fulminant Forms of Autoimmune Hepatitis. Liver Transplantation, 13: 996-1003. Miyake, Y., Iwasaki, Y., Terada, R., Onishi, T., Okamoto, R., Sakai, N., Sakaguchi, K., Shiratori, Y. 2006. Clinical characteristics of fulminant-type
18
Yeoman et al.
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
autoimmune hepatitis: an analysis of eleven cases. Aliment Pharmacol Ther, 23: 1347-53. Verma, S., Mahaeshwari, A., Thuluvath, P. 2009. Liver failure as initial presentation of autoimmune hepatitis: Clinical characteristics, predictors of response to corticosteroid therapy, and outcomes. Hepatology, 49: 1396-1397. Yeoman, A.D., Westbrook, R.H., Al-Chalabi, T., Carey, I., Heaton, N.D., Portmann, B.C., Heneghan, M.A. 2009. Diagnostic value and utility of the simplified International Autoimmune Hepatitis Group Criteria (IAIHG) in acute and chronic liver disease. Hepatology, 50: 538-545. Fujiwara, K., Yasui, S., Tawada, A., Fukuda, Y., Nakano, M., Yokosuka, O. 2011. Diagnostic Value and Utility of the Simplified International Autoimmune Hepatitis Group Criteria in Acute-onset Autoimmune Hepatitis. Liver International, 31: 1013-1020. Al-Chalabi, T., Underhill, J., Portmann, B.C., McFarlane, I.G., Heneghan, M.A. 2008. Effects of Serum Aspartate Transaminase Levels in Patients with Autoimmune Hepatitis Influence Disease Course and Outcome. Clin Gastroenterol and Hepatol, 6: 1389-1395. Yeoman, A.D., Westbrook, R.H., Zen, Y., Maninchedda, P., Portmann, B.C., Devlin, J., O'Grady, J.G., Harrison, P.M., Heneghan, M.A. 2011. Early predictors of corticosteroid treatment failure in icteric presentations of autoimmune hepatitis. Hepatology, 53: 926-934. Malinchoc, M., Kamath, P.S., Gordon, F.D., Peine, C.J., Rank, J., Borg, P.C. 2000. A model to predict poor survival in patients undergoing transjugular intrahepatic portosystemic shunts. HEPATOLOGY, 31: 864-871. Neuberger, J., Gimson, A., Davies, M., Akyol, M., O’Grady, J.G., Burroughs, A., Hudson, M. 2008. Selection of patients for liver transplantation andallocation of donated livers in the UK. GUT, 57: 252-257. Montano-Loza, A.J., Carpenter, H.A., Czaja, A.J. 2007. Features Associated with Treatment Failure in Type 1 Autoimmune Hepatitis and Predictive Value of the Model of End-Stage Liver Disease. Hepatology, 46: 1138-1145. Al-Chalabi, T., Boccato, S., Portmann, B.C., McFarlane, I.G., Heneghan, M.A. 2006. Autoimmune hepatitis (AIH) in the elderly: a systematic retrospective analysis of a large group of consecutive patients with definite AIH followed at a tertiary referral centre. J Hepatol, 45: 575-83. Atterbury, C.E., Maddrey, W.C., Conn, H.O. 1978. Neomycin-sorbitol and lactulose in the treatment of acute portal-systemic encephalopathy. A controlled double blind clinical trial. Am J Dig Dis, 23: 398-406. O'Grady, J.G., Schalm, S.W., Williams, R. 1993. Acute Liver Failure: redefining the syndromes. The Lancet, 342: 273-275.
19
Yeoman et al. Figure 1. Proposed Histological Patterns Identified in Posssible Autoimmune Acute Liver Failure (AIALF)12.
Pattern of MHN
Characteristic Features
MHN 1
Panlobular necrosis and neocholangiolar proliferation
MHN 2
MHN1 plus regenerative nodules “submassive necrosis”
MHN 3
Severe acute hepatitis: Interface hepatitis, multilobular and bridging necrosis
MHN 4
MHN 1 with centrilobular haemorrhagic necrosis
MHN 5
Confluent necrosis superimposed on chronic hepatitis
Adapted from Stravitz et al12. MHN: massive hepatocellular necrosis
20
Yeoman et al. Table 3. Laboratory parameters, calculated prognostic scores, scoring system characteristics and clinical outcomes among treated and untreated patients in the study cohort. Parameter Bilirubin µmol/l (3-20) (range) AST mmol/l (10-50) (range) Globulin (g/l) (25-35) (range) IgG (g/l) Admission INR (0.9-1.2) (range) Peak INR (0.9-1.2) (range) Creatinine µmol/l (45-120) (range) MELD Score (range) UKELD Score (range) IAIHG Score (range) AutoAbs >1:80 Transplant Sepsis Death
Untreated n=9 477 (277-583) 779 (156-2690) 30 (23-47) 15.2 (14-23.8)
Treated n=23 453 (55-1208) 556 (116-2564) 41 (14-68) 27.1 (3.8-31.4)
p value
2.8 (1.8-4.7) 3.2 (2.2-5.0) 104 (78-226) 34 (25-40) 65 (59-69) 15 (11-18) 67% 100% 11% 22%
2.01 (1.5-3.5) 2.3 (1.5-4.5) 97 (43-210) 28 (16-40) 62 (56-67) 14 (10-19) 52% 43% 26% 17%
0.04
0.75 0.38 0.16 0.04
0.004 0.25 0.01 0.10 0.5 0.70 0.004 0.64 0.99
Abbreviations: AST, aspartate aminotrasnferase; INR, international normalized ratio; MELD, Model for End Stage Liver Disease; UKELD, United Kingdom End Stage Liver Disease; IAIHG, International Autoimmune Hepatitis Group; µmol/l, micromoles per litre; mmol/l, millimoles per liter; g/l, grams per litre.
21
Yeoman et al. Table 2: Laboratory parameters, prognostic indicators and outcomes among treatment responders and treatment failures Parameter
Immunosuppressive Therapy Responder n=13 447 (197-1208) 397 (173-1420) 31 (14-68) 1.9 (1.5-2.4) 104 (43-210) 26 (22-40) 61 (56-66)
Immunosuppressive Therapy Failure n=10 475 (55-898) 829 (58-2564) 42 (27-52) 2.3 (1.6-3.5) 89 (64-137) 28 (16-33) 63 (60-67)
p value
Liver transplant
0%
100%
<0.0001
Death
0%
40%
0.03
Bilirubin µmol/l (3-20) (range) AST mmol/l (10-50) (range) Globulin g/l (25-35) (range) Admission INR (0.9-1.2) (range) Creatinine µmol/l (45-120) (range) MELD Score (range) UKELD Score (range)
0.42 0.15 0.48 0.30 0.23 0.80 0.17
Abbreviations: AST, aspartate aminotrasnferase; INR, international normalized ratio; MELD, Model of End Stage Liver Disease; UKELD, United Kingdom End Stage Liver Disease; µmol/l, micromoles per liter; mmol/l, millimoles per litre; g/l, grams per litre.
22
Yeoman et al. Table 3. Comparsion of demographic, laboratory and clinical characteristics between Kings AS-AIH cohort with ALF and USALF indeterminate cohort.
Parameter Age Females White:Non White Bilirubin µmol/l AST mmol/l ALP U/l INR Creatinine µmol/l Globulins g/l Positive autoantibodies Steroid exposure Survival
King’s Cohort n= 22 41 ± 3.4 91% 77% 514 ± 48 975 ± 179 200 ± 21 2.7 ± 0.2 106 ± 9 37 ± 2.0 77% 33% 73%
USALF Cohort n=72 41 ± 1.7 58% 67% 422 ± 22 1231 ± 197 205 ± 15 3.2 ± 0.2 159 ± 18 34 ± 0.4 66% 19% 71%
Data expressed as mean ± standard error Abbreviations: AST, aspartate aminotrasnferase; ALP, alkaline phosphatise; INR, International Normalized Ratio; MELD, Model for End Stage Liver Disease; UKELD, United Kingdom End Stage Liver Disease; IAIHG, International Autoimmune Hepatitis Group; mg/dl, milligrams per decilitre; U/l, Units per litre; mmol/l, millimoles per litre; g/l, grams per litre.
23
Yeoman et al. Table 4. Variables recorded at presentation between patients with AS-AIH and acute exacerbation type presentations.
Parameter Age (range) Bilirubin µmol/l (3-20) (range) AST mmol/l (10-50) (range) INR (0.9-1.2) (range) Globulin g/l (25-35) (range) Autoantibodies (≥1:80) % Creatinine µmol/l (45-120) (range) MELD Score (range) UKELD Score (range) IAIHG Score Death Transplant
AS-AIH n= 32 42 (16-68) 463 (55-1208) 605 (58-2690) 2.2 (1.5-3.5) 39 (22-68) 56%
Acute Exacerbation Type n=15 42 (17-60) 269 (77-574) 706 (188-1712) 1.9 (1.6-2.6) 47 (24-81) 80%
p value
101 (26-226) 29 (22-40) 62 (55-69) 15 (10-19) 19%
98 (69-142) 23 (16-35) 58 (55-67) 17 (10-22) 0%
0.8
59%
7%
Abbreviations: AST, aspartate aminotrasnferase; INR, international normalized ratio; MELD, Model of End Stage Liver Disease; UKELD, United Kingdom End Stage Liver Disease; µmol/l, micromoles per litre; mmol/l, millimoles per litre; g/l, grams per litre.
24
0.30 0.015 0.85 0.13 0.02 0.19
0.005 0.006 0.06 0.15 0.001