Incidence and Characteristics of Autoimmune Hepatitis Carolina Jiménez-Rivera, MD a , Simon C. Ling, MD b , Najma Ahmed, MD c , Jason Yap, MD d , Mary Aglipay, MSc a , Nick Barrowman, PhD a , Samantha Graitson, BSc a , Jeff Critch, MD e , Mohsin Rashid, MD f , Vicky L. Ng, MD b , Eve A. Roberts, MD b , Herbert Brill, MD g , Jenna K. Dowhaniuk, MD g , Garth Bruce, MD h , Kevin Bax, MD i , Mark Deneau, MD j , Orlee R. Guttman, MD k , Richard A. Schreiber, MD k , Steven Martin, MD l , Fernando Alvarez, MD m abstract BACKGROUND AND OBJECTIVES: Autoimmune hepatitis (AIH) is a progressive inflammatory liver disease of unknown etiology, with limited population-based estimates of pediatric incidence. We reported the incidence of pediatric AIH in Canada and described its clinical characteristics. METHODS: We conducted a retrospective cohort study of patients aged ,18 years diagnosed with AIH between 2000–2009 at all pediatric centers in Canada. RESULTS: A total of 159 children with AIH (60.3% female, 13.2% type 2 AIH) were identified. Annual incidence was 0.23 per 100000 children. Median age at presentation for type 1 was 12 years (interquartile range: 11– 14) versus 10 years for type 2 (interquartile range: 4.5– 13) (P 5 .03). Fatigue (58%), jaundice (54%), and abdominal pain (49%) were the most common presenting symptoms. Serum albumin (33 vs 38 g/L; P 5 .03) and platelet count (187 000 vs 249 000; P ,.001) were significantly lower and the international normalized ratio (1.4 vs 1.2; P ,.001) was higher in cirrhotic versus noncirrhotic patients. Initial treatment included corticosteroids (80%), azathioprine (32%), and/or cyclosporine (13%). Response to treatment at 1 year was complete in 90%, and partial in 3%. 3% of patients had no response, and 3% responded and later relapsed. Nine patients underwent liver transplantation, and 4 patients died at a mean follow-up of 4 years. CONCLUSIONS: AIH is uncommon in children and adolescents in Canada. Type 1 AIH was diagnosed 5.5 times more frequently than type 2 AIH. Most patients respond well to conventional therapy, diminishing the need for liver transplantation. WHAT’S KNOWN ON THIS SUBJECT: Pediatric autoimmune hepatitis is an uncommon condition; children and youth can present with a diverse and insidious clinical course and biochemical features. Response to treatment is generally good, and transplantation is rarely needed. WHAT THIS STUDY ADDS: This population-based study adds knowledge regarding the incidence of pediatric autoimmune hepatitis in Canada, as well as a description of diagnostic and therapeutic approaches among centers. Long- term outcomes are also described. a University of Ottawa and Children’ s Hospital of Eastern Ontario, Ottawa, Ontario, Canada; b University of Toronto and The Hospital for Sick Children, Toronto, Ontario, Canada; c McGill University and Montreal Children’ s Hospital, Montreal, Quebec, Canada; d University of Alberta and Stollery Children’ s Hospital, Edmonton, Alberta, Canada; e Memorial University of Newfoundland and Janeway Children’ s Hospital, St. John’ s, Newfoundland, Canada; f Dalhousie University and IWK Health Center, Halifax, Nova Scotia, Canada; g McMaster University and McMaster Children’ s Hospital, Hamilton, Ontario, Canada; h University of Saskatchewan and Children’ s Hospital of Saskatchewan, Saskatoon, Saskatchewan, Canada; i University of Western Ontario and London Health Science Center, London, Ontario, Canada; j University of Manitoba and The Children’ s Hospital of Winnipeg, Winnipeg, Manitoba, Canada; k University of British Columbia and BC Children’ s Hospital, Vancouver, British Columbia, Canada; l University of Calgary and Alberta Children’ s Hospital Calgary, Alberta, Canada; and m University of Montreal and Centre Hospitalier Universitaire Sainte-Justine, Montreal, Quebec, Canada Dr Jimenez-Rivera was the lead investigator, designed the study, collected data, assisted with analyses of the data, and drafted the initial manuscript. Drs Ling, Ahmed, and Yap were site investigators and formed the writing group, with significant intellectual input to the manuscript; they also critically revised the manuscript. Dr Barrowman, Ms Aglipay, and Ms Graitson conducted the statistical analysis and contributed intellectually to the manuscript. Drs Critch, Rashid, Ng, Roberts, Brill, Dowhaniuk, Bruce, Bax, Deneau, Guttman, Schreiber, and Matin collected data, critically revised and reviewed the manuscript, and contributed with intellectual ideas. Dr Alvarez conceptualized the project, significantly contributed to the content of the manuscript, and critically revised the manuscript. All authors approved the final manuscript as submitted. www.pediatrics.org/cgi/doi/10.1542/peds.2015-0578 DOI: 10.1542/peds.2015-0578 Accepted for publication Aug 20, 2015 PEDIATRICS Volume 136, number 5, November 2015 ARTICLE
Transcript
Incidence and Characteristics ofAutoimmune HepatitisCarolina Jiménez-Rivera, MDa, Simon C. Ling, MDb, Najma Ahmed, MDc, Jason Yap, MDd, Mary Aglipay, MSca,Nick Barrowman, PhDa, Samantha Graitson, BSca, Jeff Critch, MDe, Mohsin Rashid, MDf, Vicky L. Ng, MDb, Eve A. Roberts, MDb,Herbert Brill, MDg, Jenna K. Dowhaniuk, MDg, Garth Bruce, MDh, Kevin Bax, MDi, Mark Deneau, MDj, Orlee R. Guttman, MDk,Richard A. Schreiber, MDk, Steven Martin, MDl, Fernando Alvarez, MDm
abstractBACKGROUND AND OBJECTIVES: Autoimmune hepatitis (AIH) is a progressive inflammatory liver diseaseof unknown etiology, with limited population-based estimates of pediatric incidence. Wereported the incidence of pediatric AIH in Canada and described its clinical characteristics.
METHODS: We conducted a retrospective cohort study of patients aged ,18 years diagnosed withAIH between 2000–2009 at all pediatric centers in Canada.
RESULTS: A total of 159 children with AIH (60.3% female, 13.2% type 2 AIH) were identified. Annualincidence was 0.23 per 100000 children. Median age at presentation for type 1 was 12 years(interquartile range: 11–14) versus 10 years for type 2 (interquartile range: 4.5–13) (P 5 .03).Fatigue (58%), jaundice (54%), and abdominal pain (49%) were the most common presentingsymptoms. Serum albumin (33 vs 38 g/L; P 5 .03) and platelet count (187 000 vs 249 000;P,.001) were significantly lower and the international normalized ratio (1.4 vs 1.2; P ,.001) washigher in cirrhotic versus noncirrhotic patients. Initial treatment included corticosteroids (80%),azathioprine (32%), and/or cyclosporine (13%). Response to treatment at 1 year was complete in90%, and partial in 3%. 3% of patients had no response, and 3% responded and later relapsed. Ninepatients underwent liver transplantation, and 4 patients died at a mean follow-up of 4 years.
CONCLUSIONS: AIH is uncommon in children and adolescents in Canada. Type 1 AIH wasdiagnosed 5.5 times more frequently than type 2 AIH. Most patients respond well toconventional therapy, diminishing the need for liver transplantation.
WHAT’S KNOWN ON THIS SUBJECT: Pediatricautoimmune hepatitis is an uncommoncondition; children and youth can present witha diverse and insidious clinical course andbiochemical features. Response to treatment isgenerally good, and transplantation is rarelyneeded.
WHAT THIS STUDY ADDS: This population-basedstudy adds knowledge regarding the incidence ofpediatric autoimmune hepatitis in Canada, aswell as a description of diagnostic andtherapeutic approaches among centers. Long-term outcomes are also described.
aUniversity of Ottawa and Children’s Hospital of Eastern Ontario, Ottawa, Ontario, Canada; bUniversity of Torontoand The Hospital for Sick Children, Toronto, Ontario, Canada; cMcGill University and Montreal Children’s Hospital,Montreal, Quebec, Canada; dUniversity of Alberta and Stollery Children’s Hospital, Edmonton, Alberta, Canada;eMemorial University of Newfoundland and Janeway Children’s Hospital, St. John’s, Newfoundland, Canada;fDalhousie University and IWK Health Center, Halifax, Nova Scotia, Canada; gMcMaster University and McMasterChildren’s Hospital, Hamilton, Ontario, Canada; hUniversity of Saskatchewan and Children’s Hospital ofSaskatchewan, Saskatoon, Saskatchewan, Canada; iUniversity of Western Ontario and London Health ScienceCenter, London, Ontario, Canada; jUniversity of Manitoba and The Children’s Hospital of Winnipeg, Winnipeg,Manitoba, Canada; kUniversity of British Columbia and BC Children’s Hospital, Vancouver, British Columbia,Canada; lUniversity of Calgary and Alberta Children’s Hospital Calgary, Alberta, Canada; and mUniversity ofMontreal and Centre Hospitalier Universitaire Sainte-Justine, Montreal, Quebec, Canada
Dr Jimenez-Rivera was the lead investigator, designed the study, collected data, assisted withanalyses of the data, and drafted the initial manuscript. Drs Ling, Ahmed, and Yap were siteinvestigators and formed the writing group, with significant intellectual input to the manuscript;they also critically revised the manuscript. Dr Barrowman, Ms Aglipay, and Ms Graitson conductedthe statistical analysis and contributed intellectually to the manuscript. Drs Critch, Rashid, Ng,Roberts, Brill, Dowhaniuk, Bruce, Bax, Deneau, Guttman, Schreiber, and Matin collected data,critically revised and reviewed the manuscript, and contributed with intellectual ideas. Dr Alvarezconceptualized the project, significantly contributed to the content of the manuscript, and criticallyrevised the manuscript. All authors approved the final manuscript as submitted.
www.pediatrics.org/cgi/doi/10.1542/peds.2015-0578
DOI: 10.1542/peds.2015-0578
Accepted for publication Aug 20, 2015
PEDIATRICS Volume 136, number 5, November 2015 ARTICLE
Autoimmune hepatitis (AIH) isa progressive inflammatory liverdisease of uncertain etiology. Itspathogenesis seems to bemultifactorial, including geneticsusceptibility, abnormal regulation ofthe immune response, andenvironmental triggers. Two subtypeshave been recognized based on thepresence of serum autoantibodies.Patients with type 1 AIH haveanti–smooth muscle and/orantinuclear autoantibodies, whereastype 2 is characterized by thepresence of anti–liver kidneymicrosomal and/or anti–liver cytosoltype 1 (LC1) autoantibodies.1–3
Autoantibodies can be positive inother hepatic conditions such as viralhepatitis4,5 and Wilson’s disease,6 aswell as sclerosing cholangitis; hightiters favor the diagnosis of AIH. Thepresence of combined antibodiesincreases the accuracy of an AIHdiagnosis.7
Both types of AIH may be diagnosedin childhood, are more common infemale subjects, and may beassociated with an extrahepaticautoimmune disease such as arthritis,inflammatory bowel disease (IBD),thyroiditis, or diabetes.8,9 Risk factorsfor AIH include celiac disease,10
immune dysregulation,polyendocrinopathy, enteropathy, andX-linked syndrome.11 Drugsassociated with AIH-likehepatotoxicity include diclofenac,methyldopa, hydralazine,nitrofurantoin, minocycline, and,more recently, statins and anti–tumornecrosis factor-a agents.12
A recent study in pediatric patientsreported an incidence of 0.4 case per100 000 children.13 Another studyfrom Poland reported an incidence of3 to 4 per 100 000 children.14
Available adult data indicate anincidence varying from 0.67 to 2cases per 100 000 people.15–17 Ethnicbackground has been reported asa key factor in clinical presentationand outcomes for patients withAIH18,19; however, this finding is not
universal.20 In the present nationalstudy, our primary aim was todetermine the incidence of AIH inchildren and adolescents in Canadaand to describe its clinicalcharacteristics and natural history. Inaddition, we attempted to identifyrisk factors for poor outcomes.
METHODS
Under the auspices of the CanadianPediatric Hepatology Research Group,we performed a multicenterretrospective review of all new casesof AIH diagnosed in children aged,18 years in Canada betweenJanuary 1, 2000, and December 31,2009, and who were followed up inan academic pediatric center.Representatives of each center wereinvited to participate. Theserepresentatives identified all eligiblechildren by using health recorddepartments as well as diagnosissupport offices and personal databases.
Patients were considered to have AIHwhen there was evidence of abnormalliver test results without any otheridentifiable liver disease. Althoughthe presence of autoantibodies and/or elevation of immunoglobulin Gsupported the diagnosis, theirabsence was not an exclusioncriterion. When possible, histologicabnormalities confirmed thediagnosis of AIH. Patients also hadnegative findings when undergoingimaging of the biliary tree. Responseto treatment was reassuring ofa diagnosis of AIH retrospectively.AIH scores at diagnosis werecalculated according to the studygroup based on the revised diagnosticcriteria.21
Patients were excluded if they hadevidence of any other liver disease.Investigations included viral serology(hepatitis B virus and hepatitis Cvirus), a1-antitrypsin level, serumcopper and ceruloplasmin, andmetabolic evaluation were performedat each center if clinically indicated.Fatty liver was not a reason for
exclusion; however, this factor wasnot captured in our database. Cases ofautoimmune sclerosing cholangitis(defined by the presence of specificlesions of the intrahepatic orextrahepatic bile ducts [chronicinflammation, fibrosis, stenosis, anddilations] evident in histology orimaging and associated with featuresof AIH22) and AIH secondary tominocycline use were excluded fromthe study. Patients with suspecteddrug liver injury were also excluded.Research ethics board approval wasobtained at each center, and datawere abstracted from the charts ofthe identified patients and enteredinto a standardized case report form.
Demographic data, symptoms,physical examination findings,biochemical results, diagnosticimaging, and liver biopsy reportswere collected at the time ofdiagnosis. Autoantibodies werereported as positive or negativebecause the values and methods usedin the different centers varied.Biochemical parameters weresubsequently recorded at 3, 6, and 12months after the initial diagnosis.
Initial therapy was recorded andincluded corticosteroids,azathioprine, and cyclosporineaccording to the preference by eachcenter. Data on treatment over time at3, 6, 12, and 18 months and at the lastclinic visit were captured; however,indications for changes orintroduction of new drugs (eg,ursodeoxycholic acid) were notavailable. Supplemental Table 5describes definitions of response totherapy.21 Ultrasonographic findingsof portal hypertension included anyof the following: ascites,splenomegaly, reverse flow within theportal vein, portal vein dilation, andportosystemic collateral veins.Histologic findings were recorded ateach center and included thepresence of interface hepatitis,lymphoplasmacytic infiltrates,rosetting of hepatocytes, biliarychanges, fibrosis, and cirrhosis.
AIH scores were calculatedretrospectively by the study groupbased on the revised criteria21 whendata were available. The scoresassigned to the presence ofautoantibodies depended on availabletiters; if the center reporteda qualitative result as “positive,” thescore assigned was 1; if it wasquantitative, the score was givenaccording to the aforementionedrecommendations.
Population estimates for children andyouth aged ,18 years in eachCanadian province and territory foreach year from 2000 to 2009 wereobtained from Statistics Canada.23
These estimates were summedaccording to province/territory, aswell as overall, to obtain the totalnumber of person-years during thestudy period. The incidence of AIHwas estimated by dividing thenumber of cases by the number ofperson-years. Exact Poisson 95%confidence intervals (CIs) forincidence rates were computed byusing the EpiTools package in Rversion 3.1.0 (R Foundation forStatistical Computing, Vienna,Austria).24,25
The remaining analyses wereperformed by using SPSS version 22(IBM SPSS Statistics, IBMCorporation, Armonk, NY).26 Twoseparate analyses were performed toidentify predictive factors for patientoutcome. A series of univariatelogistic regression analyses wasperformed to examine potentialpredictors of a favorable response totreatment at 12 months’postdiagnosis. Two-sided P values,.05 were considered statisticallysignificant. A repeated measuresanalysis of variance was used tocompare the difference in BMI zscores between participants whoused corticosteroids and those whoused cyclosporine. Favorable long-term outcome was defined astransplant-free survival. Time totransplant or death was computed,and loss to follow-up and transfer to
adult care were treated as censoringevents. Kaplan-Meier curves wereused to display time to outcome indifferent groups. Log-rank tests wereused for statistical comparison ofsurvival times.
RESULTS
Incidence
A total of 200 children from the 13participating centers were included.Forty-one patients were excluded dueto the following: diagnosis ofautoimmune sclerosing cholangitis(n = 19); positive history ofminocycline use (n = 10) before thediagnosis of AIH; and diagnosisoutside of the study period (n = 12).The total population of children andyouth aged #18 years in Canadaaveraged 70 277 839. The incidencewas thus ∼0.23 per 100 000 children(95% CI: 0.19–0.26). There werestatistically significant differencesamong provinces (P = .01); whenprovinces were grouped (Atlantic,Central, West, and North), theincidence rates between regions didnot vary significantly (P = .10).Table 1 displays the incidence of AIHaccording to province, and Fig 1shows the geographical distribution.
Patient Characteristics
Median age at presentation was 12years (interquartile range: 11–14) intype 1 AIH versus 10 years(interquartile range: 4.5–13) in type 2AIH (P = .03); 60% were female, andthere was no significant difference ingender between type 1 and 2. Meantime from onset of symptoms todiagnosis was 4.7 6 7 months. Fivepatients presented with hematemesis,4 of 5 had evidence of fibrosis, and 1of 5 had cirrhosis according to liverbiopsy results; only 1 hadsplenomegaly, and 2 hadthrombocytopenia. One patientreported alcohol consumption .60g/d. Less than 3% of patients were onother medications, including protonpump inhibitors, 5-aminosalicyclicacid, and vitamin supplementation;
no patients were takingnitrofurantoin. Patient characteristicsat diagnosis are described in Table 2.
Patient history of previous orconcomitant autoimmune diseaseswas positive in 32% (49 of 154) andincluded the following: ulcerativecolitis, 12% (n = 18); Crohn’s disease,5% (n = 8); rheumatoid arthritis, 4%(n = 6); thyroid disease, 4% (n = 6);and celiac disease, 2% (n = 3). Onepatient had concomitant systemiclupus erythematosus. The timing ofoccurrence of such autoimmunediseases was not uniformly available.Family history of autoimmune diseasein first-degree relatives was reportedin 25% (38 of 154). Personal orfamily history of autoimmunediseases did not predict outcome.
Biochemistry
Liver biochemistry findings wereelevated in all children at diagnosis(Table 2). Serum albumin levels werebelow normal limits in 43% (62 of144). An increased internationalnormalized ratio was reported in45% (67 of 148), andimmunoglobulin G levels wereabnormal in 84% (125 of 148).
Autoantibodies
Data regarding autoantibodies atdiagnosis, including antinuclearantibody, smooth muscle antibody,liver/kidney microsomal, and LC1antibodies, are provided in Table 2.
Imaging
Abdominal sonography at diagnosisrevealed hepatomegaly in 26% (37 of141), splenomegaly in 50% (72 of144), ascites in 9% (12 of 139), andsigns of portal hypertension in 11%(15 of 135). Presence of portal veinthrombosis was not captured in thedata set. Imaging of the biliary treewas undertaken in 59 children; 50 ofthese underwent magnetic resonancecholangiopancreatography (MRCP),and 4 of 50 underwent endoscopicretrograde cholangiopancreatography(ERCP) as well. ERCP alone wasperformed in 9 patients, and these
PEDIATRICS Volume 136, number 5, November 2015 e1239
results were all normal. Indicationsfor ERCP were not recorded. MRCPwas performed in 58% (15 of 26) ofpatients with a concomitant diagnosisof IBD; 1 of these patients underwentERCP as well. ERCP alone wasperformed in 2 patients with IBD.
In 7 children, the initial MRCPrevealed either mild dilation orprominence of bile ducts, including 3with concomitant IBD, but no othersigns consistent with sclerosingcholangitis. Initial liver biopsy resultsof these 7 children exhibitedhistologic evidence of biliary changes
in 5 not consistent with sclerosingcholangitis. These children wereconsidered to have AIH and weretreated and followed as per usuallocal clinic routine. Of these, 6 hadnormalized liver function at 1 year,and 1 had normalized liver functionat 18 months on immunosuppression.Follow-up imaging was not availablefor analysis. One child hada previously surgically repairedcholedochal cyst and was diagnosedwith AIH at 13 years of age. TheMRCP showed no intrahepatic bileduct involvement, and the child wastherefore included in the analysis.
Histology
Initial liver biopsy reports (n = 152)noted interface hepatitis in 86% (127of 147), lymphoplasmacytic infiltratesin 97% (148 of 152), and rosetting ofhepatocytes in 18% (24 of 135).Biliary changes were noted in 39.8%(55 of 148), most commonly bileductular proliferation in 65% (35 of55); none had features of sclerosingcholangitis.
Cirrhosis at the time of diagnosis wasreported in 30 (20.0%) of 150patients; mean age was 12 6 3.3years, and 59% were female subjects.In this group, 67% were type 1 AIH(20 of 30), and 10% (3 of 30) weretype 2; 13% were seronegative (4 of30). Hepatomegaly was reported in60% (18 of 30) of these cirrhoticcases, and 67% (20 of 30) hadsplenomegaly identified either onphysical examination or according toultrasound. Ascites was reported in20% of the abdominal ultrasounds,and other signs of portalhypertension were found in 27% (8of 30). Among children with noevidence of cirrhosis in the liverbiopsy specimen (n = 120), 4% hadsigns of portal hypertension, 3% hadascites, and 40% had some degree ofsplenomegaly according toultrasound. Serum albumin andplatelet counts were significantlylower and the internationalnormalized ratio was significantlyhigher in cirrhotic patientscompared with noncirrhotic patients(Table 3).
AIH Scoring
Patients had a mean score of 15.9 63.3 at diagnosis based on theinternational AIH scoring system;56% (88 of 156) of the cases scoredas definite AIH, and 44% (68 of 156)were probable AIH. The score couldnot be calculated in 3 patients (1.9%)due to missing data. Two of thesepatients responded well to therapy,and the other patient remained onlow-dose prednisone with a partialresponse at the end of the study.
Incidence rates varied significantly between provinces (P = .01).23
FIGURE 1Annual incidence of AIH in Canada.
e1240 JIMÉNEZ-RIVERA et al
Treatment
Of the 159 participants, 147 hadcomplete information on initialtreatment. Initial treatment consistedof prednisone (unspecified route ofadministration) at a mean dose of 1.060.5 mg/kg/d (range: 0.2–2 mg/kg/d)in 87% of children; 54% of patientsreceived prednisone alone, and
33% received both prednisoneand azathioprine (mean dose: 1.2 6
0.6 mg/kg/d). Cyclosporine wasused as initial therapy in a singlecenter in 13% of the cases at a meandose of 4 6 0.8 mg/kg/d; 12% usedcyclosporine alone, and 1% usedboth cyclosporine and prednisone.Cyclosporine trough levels were
followed according to a previouslydescribed protocol.27 Levels ofazathioprine metabolites werenot consistently measured orrecorded. No other drugs were usedfor induction of remission.
At 1 year after diagnosis, 61% ofchildren were on low-doseprednisone (mean dose: 0.2 6 0.2mg/kg/d). Cyclosporine was used in2.5% (mean dose: 4.96 1.8 mg/kg/d)and mycophenolate mofetil in 5% atvariable doses (500–750 mg twicea day). Those patients takingprednisone had a mean 6 SD BMIof 1.1 6 1.0, and those not onprednisone had a mean BMI of 0.8460.87 (P = .08). At the last clinic visit(mean time: 4 6 2 years), 67.3% (99of 147) patients were onazathioprine, 31.3% (46 of 147)were on low-dose prednisone, 8.2%(12 of 147) were on mycophenolatemofetil, 4% were on tacrolimus(6 of 147), 1.3% (2 of 147) wereon cyclosporine, 16.3% (24 of 147)were on adjuvant therapy withursodeoxycholic acid, and 7%(10 of 147) were receiving notreatment. Figure 2 illustratespatients’ medication use overtime.
Response to Therapy and Outcomes
At 1 year after diagnosis of AIH, 90%of patients had complete response totherapy, 3.2% had a partial response,3.8% did not respond, and 3.2%experienced a relapse. Nine (5.7%)patients underwent livertransplantation. Trends regardingaminotransferase levels over time arerepresented in Fig 3. Mean follow upwas 4 6 2 years. Of the 157 patients,10 were lost to follow-up, 69 hadbeen transferred to adult care, and 74were still followed up in therespective centers; data were missingin 2 patients. Indications of livertransplantation included fulminanthepatic failure, “chronic” hepaticfailure (patients with cirrhosis and noimprovement of the liver failuredespite appropriate treatment), andlate liver failure in patients with
ANA, antinuclear antibody; ALT, alanine aminotransferase; AST, aspartate aminotransferase; ESR, erythrocyte sedimenta-tion rate; GGT, g-glutamyl transferase; IgG, immunoglobulin G; INR, international normalized ratio; IQR, interquartile range;LKM, liver kidney microsomal antibody; SMA, smooth muscle antibody.a IgG type 1 versus type 2, P , .01.
PEDIATRICS Volume 136, number 5, November 2015 e1241
partial response to therapy.Characteristics of patients with pooroutcomes are reported in Table 4.
Figure 4 shows the results ofunivariate logistic regression analysis,which failed to identify any significantassociations between potentialpredictive factors and poor responseto therapy at 12 months afterdiagnosis.
The BMI z score was significantlyhigher at 3 months after startingcorticosteroid therapy compared withthose patients receiving cyclosporine(1.16 [95% CI: 0.99–1.35] vs 0.64[95% CI: 0.17–1.11]; P = .04);however, this difference was notmaintained, and at 1 year there wasno significant difference (prednisone1.08 [95% CI: 0.85–1.25] vs
cyclosporine 1.25 [95% CI:0.67–1.83], P = .58) (Fig 5). Patientsurvival curve in the presence ofcirrhosis is reported in Fig 6.
DISCUSSION
In this nationwide Canadian study ofAIH in children, we identified anincidence of AIH of at least 0.23 per100 000 children per year. Previousreports of pediatric prevalence arescarce and vary from 2.2 in non–FirstNation (indigenous native Canadian)communities to 9.9 in First Nationcommunities per 100 000 children inBritish Columbia.28 A recent report byDeneau et al13 in Utah described anincidence of 0.4 case per 100 000children and a prevalence of 3 per100 000. The previously reportedincidence in adults ranged from 0.1 to1.9 per 100 000 per year (includingsome adolescents).17,29 There weresignificant differences in incidenceamong provinces in Canada, with theAtlantic provinces having the highestrates. Benchimol et al30 reporteda higher incidence of IBD in easternCanada. We speculate that uniqueenvironmental and/or genetic factorsmight play a role in the increasedrates of immune-mediated disease inthe region, but we did not capturedata regarding ethnic background orother possible epidemiologic riskfactors. Although AIH (mainly type 1)occurs in both genders, our studyconfirms the female predominancenoted in most previousreports.1,16,22,31
The number of patients in our cohortmeeting the criteria for definite AIHproposed by the InternationalAutoimmune Hepatitis Group21 was56%, with 44% falling into thecategory of probable AIH. This findingdiffers from other studies in whichapplication of the scoring systemclassified 82% as definite AIHpretreatment in the pediatric group.22
We recognize that human leukocyteantigen status was not tested in mostof our study centers, and thusadditional points could not be added
TABLE 3 Biochemical Parameters for Cirrhotic and Noncirrhotic Patients
Parameter Cirrhosis P
No (n = 120) Yes (n = 30)
Type, n (%)Type I 84 (70) 20 (67) .19Type II 14 (12) 3 (10)Seronegative for AIH 5 (4) 4 (13)Unknown 17 (14) 3 (10)
FIGURE 2Patients’ medication use over time. MMF, mycophenolate mofetil.
e1242 JIMÉNEZ-RIVERA et al
for this variable. In addition, someautoantibodies were reported aspositive or negative only, withouta titer. We therefore assigned thelowest points when a positive resultwas given, which could have resultedin a decrease of the total score and
lower number of patients classified ashaving “definite AIH.” Nine percent ofpatients were seronegative; however,measurement of other autoantibodies(eg, LC1, perinuclear antineutrophilcytoplasmic antibodies, soluble liverantigen), which may have helped
classify such patients, was notroutinely performed. We could notcalculate the AIH score posttreatmentdue to missing data.
The clinical presentation in ourcohort was variable and perhapsreflected the prolonged time todiagnosis of ∼5 months. However,many children with AIH have aninsidious pattern of symptoms andsigns and a fluctuating course of thedisease that may also lead to a delayin diagnosis. Not surprisingly, fatiguewas the most common complaint,followed by jaundice. Clinicalevidence of chronic disease with thepresence of portal hypertension wasobserved in approximately one-fifthof the patients, as indicated by thepresence of splenomegaly andthrombocytopenia. This finding ishigher than other reports in whichsplenomegaly was found in only 6%of children at onset of symptoms.32
Similar to other reports, one-thirdof the children in our cohort hada concomitant autoimmunedisease.22,32 IBD was the predominantcomorbidity in 17% of our patientpopulation, which is similar to thereport by Gregorio et al22 that
FIGURE 3Trend of aminotransferase levels over time.
TABLE 4 Characteristics of Patients Who Died or Required Liver Transplantation
Patient No. Age atDiagnosis, y
Gender AIH Type Liver Histology Initial Therapy Time toTransplant/Death, mo
Outcome
1 11 F 1 IH, LPI, bridging fibrosis Corticosteroids 62 A2 10 M 1 IH, LPI, fibrosis Corticosteroids 12 A3 13 M 1 IH, LPI, bile duct proliferation,
minimal fibrosisCorticosteroids 30 A
4 10 M 1 IH, LPI, cirrhosis Cyclosporine 19 A5 1 M Uncertain
(LKM notdone)
IH, LPI, cirrhotic liver withsubmassive necrosis of80% of parenchyma
Cyclosporine 1 D (died 4 d posttransplant)
6 13 F 2 IH, LPI, cirrhosis Cyclosporine 8 A7 16 M 1 Submassive necrosis Cyclosporine 1 D (died 10 mo after first
transplant due toacute liver failure)
8 13 M 1 IH, LPI, bile duct proliferation,periductal inflammation,fibrosis
Corticosteroids 60 A
9 15 M 1 IH, LPI, bile duct proliferation,mild fibrosis
Corticosteroids,azathioprine
17 A
10 11 M 1 IH, LPI, cirrhosis Corticosteroids,azathioprine
– D (died of lymphoma 8 yafter diagnosis of AIH)
11 16 F 1 IH, LPI, moderate fibrosis Corticosteroids,azathioprine
– D (died of meningococcalmeningitis 2 y afterdiagnosis of AIH)
A, alive; D, deceased; F, female; IH, interface hepatitis; LKM, liver kidney microsomal antibody; LPI, lymphoplasmacytic infiltrate; M, male; NA, not available.
PEDIATRICS Volume 136, number 5, November 2015 e1243
described 18% of children havingconcomitant AIH/IBD. Otherinvestigators1,32 have reported lower
rates of ∼6% of AIH/IBD. Thesedisparities could potentially beexplained by the different ethnic
background and perhaps by theoverall higher incidence of IBD inCanada.33,34 Family history of a first-degree relative with autoimmunedisease occurred in 25% of our cases;however, ethnic background was notrecorded in our study. In BritishColumbia, Chung et al28 reported notonly a higher prevalence of AIHamong First Nations children but alsoa positive history of concomitantautoimmune disease in 50% ofpatients and a positive family historyof autoimmune diseases in up to100%.
Serum aminotransferase levels wereuniversally elevated. Interestingly,a large number of children hadabnormal g-glutamyl transferaseand/or elevated alkaline phosphataselevels at diagnosis although they didnot fulfill criteria for diagnosis ofautoimmune sclerosing cholangitis intheir histology or imaging. Evidenceof synthetic dysfunction by anincreased international normalizedratio and hypoalbuminemia was seenin ∼40% to 45% of cases, whichcorrelates with histologic findings ofcirrhosis in the study group.
The type of AIH (1 or 2) has beendescribed as a factor that influencesclinical presentation, treatmentdecisions, and outcomes. In thiscohort, children with type 2 AIH weresignificantly younger, but they did nothave a higher incidence of pooroutcomes such as the need for livertransplantation or death. In fact, thetype of AIH did not predict responseto therapy, as both groups had similarresponse rates.
We found that only approximatelyone-third of the patients underwentinitial imaging of the biliary tree,which was not a standard practiceduring the study period. This patterncould potentially be currentlydifferent after the AmericanAssociation for the Study of LiverDiseases guidelines were published in2010 recommending that all childrenwith the diagnosis of AIH undergocholangiographic studies at
FIGURE 4Logistic regression analysis factors affecting response to therapy at 12 months. ANA, antinuclearantibody; LKM1, liver kidney microsomal antibody type 1; OR, odds ratio; SMA, smooth muscle antibody.
FIGURE 5BMI z score (6SE) for cyclosporine versus prednisone. *P = .04.
e1244 JIMÉNEZ-RIVERA et al
diagnosis.35 This practice will helpdifferentiate from other autoimmuneliver diseases such as primary orautoimmune sclerosing cholangitis.
Initial liver histology revealed bileduct involvement in one-half of theliver biopsy specimens, including bileductular proliferation and ductal/periductal inflammation, which ishigher compared with other studiesin which biliary changes werereported in 24% of adult cases36 andup to ∼20% of pediatric cases.1 Thenegative impact of bile duct changeson the liver biopsy findings couldhave been a factor in the lower scoresof patients in the study. Thepredominant histologic features,however, were interface hepatitis andlymphoplasmacytic infiltrates, aspreviously described.37,38 One couldhypothesize that patients with bileduct involvement could evolve to anoverlap syndrome if followed up longterm39; however, histologic evidenceof biliary damage is a well-recognizedfeature of AIH.40 A small number ofpatients in our study did not havea liver biopsy performed, presumably
due to coagulopathy; however, thesepatients had clinical and biochemicalfeatures of AIH, were treated as such,and exhibited good response. Ina study in adults, Björnsson et al41
concluded that liver biopsies are notneeded in most cases when patientshave typical features of AIH. Thisconclusion could have someapplicability in the pediatricpopulation.
We calculated the AIH scoreretrospectively and found thatapproximately one-half of the patientsmet the criteria for definite AIH,contrary to the study by Mileti et al42
in which 94% were considered tohave definite AIH. This discrepancycould be explained by the lack ofreports of quantified autoantibodiesproviding a lower score in our cohort.For a similar reason, we could notassess the accuracy of the simplifiedscore specifically developed for adultsby Hennes et al43 in our pediatricpatients. It is important that the AIHscore be calculated because it is a toolfor clinicians in the management ofthis condition.
Treatment of AIH typically consists ofimmunosuppressive drugs such ascorticosteroids and azathioprine; inour series, however, 13% of caseswere initially treated withcyclosporine at a single center.Importantly, steroid-induced weightgain had resolved by 1 year. Aninteresting study by Woynarowskiet al44 reported significant weightgain with the use of prednisonecompared with budesonide, witha similar response rate; however, inthat study, the number of patientswas too small to strongly recommendits use. Response to therapy at 1 yearwas similar in both groups (89.6%with corticosteroids vs 91.7% withcyclosporine; P = not significant).Other series have reported goodoutcomes and few adverse effectswhen short-term cyclosporine wasused.1,27 At the time of their last clinicvisit, two-thirds of the patients wereoff corticosteroids and were inremission, a rate comparable to thereport by Dumortier et al45 in which78% of patients at 5 years werecompletely off corticosteroids and inremission. Nonetheless, the use ofsteroids may still be necessary,particularly in those havingassociated extrahepatic autoimmunediseases such as IBD or rheumatoidarthritis.
Other therapeutic options such asmycophenolate mofetil andtacrolimus were not widely used andwere limited to those children withpoor response or poor tolerance toazathioprine. No patient receivedanti-CD20 therapy. It was not clearwhy some children receivedursodeoxycholic acid at their lastclinic visit. We speculate that in somecases, ursodeoxycholic acid was usedas an adjuvant AIH treatment, basedon reports of its immunosuppressiveproperties, or there may have beenbiochemical, histologic, or imagingchanges prompting its use.
The retrospective nature of thepresent study is a limitation becausethe data were collected by each
FIGURE 6Patient survival in the presence of cirrhosis.
PEDIATRICS Volume 136, number 5, November 2015 e1245
individual center and abstracted frommedical records. Some data were notavailable for analysis, such as timingof onset of other autoimmunecomorbidities (eg, IBD, thyroiddisease), or information oncorticosteroid-related events apartfrom weight gain. Imaging and liverbiopsy results were not reportedsystematically; in addition,biochemical assays and normal valuesvaried from site to site.
A major strength of this study is thatit is a population-based multicentertrial that likely captures data on thevast majority of patients aged #18years with AIH, given that suchpatients are almost universally caredfor at academic health centers inCanada, thus allowing for estimationof incidence of pediatric AIH. It is ourexperience that the majority of thosediagnosed as teenagers living inremote geographical areas of Canadaand with serious liver diseases are
referred to specialized pediatricacademic centers. Our data alsorevealed a wide variety of diagnosticand therapeutic approaches amongcenters; nevertheless, outcomeswere favorable. Our findings suggestthe need for instituting a morecollaborative care approach model,with standardized protocols aimedat improving outcomes. Indeed, theestablishment of disease registries,particularly for rare orphanpediatric conditions, is increasinglyrecognized as having a valuablerole in assuring optimal pediatrichealth and high quality of care.46
CONCLUSIONS
AIH is an uncommon condition in thepediatric age range (∼2 per 1 millionchildren aged ,18 years); clinicalpresentation is variable, althoughwith excellent response toimmunosuppressive therapy. Few
patients will need livertransplantation, and the long-termsurvival rate is favorable. Thefindings would serve as a basefor future prospective researchaimed at characterizing thesepatients with better and improvedoutcomes.
ACKNOWLEDGMENT
The authors thank Peter Gozdyra forthe creation of the incidence map.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
REFERENCES
1. Gregorio GV, Portmann B, Reid F, et al.Autoimmune hepatitis in childhood: a 20-year experience. Hepatology. 1997;25(3):541–547
2. Maggiore G, Veber F, Bernard O, et al.Autoimmune hepatitis associatedwith anti-actin antibodies inchildren and adolescents. J PediatrGastroenterol Nutr. 1993;17(4):376–381
3. Odièvre M, Maggiore G, Homberg JC,et al. Seroimmunologic classificationof chronic hepatitis in 57 children.Hepatology. 1983;3(3):407–409
4. Czaja AJ. Autoimmune hepatitis and viralinfection. Gastroenterol Clin North Am.1994;23(3):547–566
5. Pop TL, Stef�anescu A, Samasca G, Miu N.Clinical significance of theantinuclear antibodies in chronic viralhepatitis B in children. Clin Lab. 2014;60(6):931–939
6. Dhawan A, Taylor RM, Cheeseman P, DeSilva P, Katsiyiannakis L, Mieli-Vergani G.Wilson’s disease in children: 37-yearexperience and revised King’s score forliver transplantation. Liver Transpl. 2005;11(4):441–448
7. Czaja AJ. Performance parameters of theconventional serological markers forautoimmune hepatitis. Dig Dis Sci. 2011;56(2):545–554
8. Herzog D, Rasquin-Weber AM, Debray D,Alvarez F. Subfulminant hepatic failure in
autoimmune hepatitis type 1: an unusualform of presentation. J Hepatol. 1997;27(3):578–582
9. Porta G, Gayotto LC, Alvarez F. Anti-liver-kidney microsome antibody-positiveautoimmune hepatitis presenting asfulminant liver failure. J PediatrGastroenterol Nutr. 1990;11(1):138–140
13. Deneau M, Jensen MK, Holmen J,Williams MS, Book LS, Guthery SL.Primary sclerosing cholangitis,autoimmune hepatitis, and overlap inUtah children: epidemiology and naturalhistory. Hepatology. 2013;58(4):1392–1400
14. Woynarowski M, Wo�zniak M, PawlowskaM, Lebensztejn D, Socha J, Chlebcewicz-Szuba W, et al. Autoimmune hepatitis inPolish children: healthcare facilities,epidemiology, and standards of careassessed by a pediatric autoimmunehepatitis group. Exp Clin Hepatol. 2008;4(2):66–70
15. Delgado JS, Vodonos A, Malnick S, et al.Autoimmune hepatitis in southern Israel:a 15-year multicenter study. J Dig Dis.2013;14(11):611–618
16. Ngu JH, Bechly K, Chapman BA, et al.Population-based epidemiology study ofautoimmune hepatitis: a disease of olderwomen? J Gastroenterol Hepatol. 2010;25(10):1681–1686
17. Primo J, Merino C, Fernández J, MolésJR, Llorca P, Hinojosa J. Incidenceand prevalence of autoimmunehepatitis in the area of the Hospital deSagunto (Spain) [in Spanish].Gastroenterol Hepatol. 2004;27(4):239–243
19. Wong RJ, Gish R, Frederick T, Bzowej N,Frenette C. The impact of race/ethnicityon the clinical epidemiology ofautoimmune hepatitis. J ClinGastroenterol. 2012;46(2):155–161
20. van Gerven NM, Verwer BJ, Witte BI, et al;Dutch Autoimmune Hepatitis STUDYGroup. Epidemiology and clinicalcharacteristics of autoimmune hepatitisin the Netherlands. Scand JGastroenterol. 2014;49(10):1245–1254
21. Alvarez F, Berg PA, Bianchi FB, et al.International Autoimmune HepatitisGroup Report: review of criteria fordiagnosis of autoimmune hepatitis. JHepatol. 1999;31(5):929–938
22. Gregorio GV, Portmann B, Karani J, et al.Autoimmune hepatitis/sclerosingcholangitis overlap syndrome inchildhood: a 16-year prospective study.Hepatology. 2001;33(3):544–553
23. Statistics Canada. Estimates ofPopulation by Age and Sex for Canada,Provinces and Territories. http://www23.statcan.gc.ca/imdb/p2SV.pl?Function5getSurvey&SDDS53604.Updated September 26, 2014. AccessedSeptember 11, 2015
24. R Development Core Team (2011). ALanguage and Environment forStatistical Computing. Vienna, Austria: RFoundation for Statistical Computing.ISBN: 3-900051-07-0
25. Tomas J, Fay MP, Wollschlaeger D.EpiTools: R Package for EpidemiologicData and Graphics. R package version0.5-7. Available at: https://cran.r-project.org/web/packages/epitools/index.html.Updated September 30, 2012. AccessedSeptember 11, 2015
26. IBM SPSS Statistics for Windows, Version22.0. Armonk, NY: IBM Corporation; 2013
27. Alvarez F, Ciocca M, Cañero-Velasco C,et al. Short-term cyclosporine inducesa remission of autoimmune hepatitis inchildren. J Hepatol. 1999;30(2):222–227
28. Chung HV, Riley M, Ho JK, et al.Retrospective review of pediatric andadult autoimmune hepatitis in twoquaternary care centres in BritishColumbia: increased prevalence seen inBritish Columbia’s First Nationscommunity. Can J Gastroenterol. 2007;21(9):565–568
29. Boberg KM. Prevalence and epidemiologyof autoimmune hepatitis. Clin Liver Dis.2002;6(3):635–647
30. Benchimol EI, Fortinsky KJ, Gozdyra P,Van den Heuvel M, Van Limbergen J,Griffiths AM. Epidemiology of pediatricinflammatory bowel disease:a systematic review of internationaltrends. Inflamm Bowel Dis. 2011;17(1):423–439
31. Gheorghe L, Iacob S, Gheorghe C, et al.Frequency and predictive factors foroverlap syndrome between autoimmunehepatitis and primary cholestatic liverdisease. Eur J Gastroenterol Hepatol.2004;16(6):585–592
32. Vitfell-Pedersen J, Jørgensen MH, MüllerK, Heilmann C. Autoimmune hepatitis in
children in eastern Denmark. J PediatrGastroenterol Nutr. 2012;55(4):376–379
33. Benchimol EI, Manuel DG, Guttmann A,et al. Changing age demographicsof inflammatory bowel disease inOntario, Canada: a population-basedcohort study of epidemiology trends.Inflamm Bowel Dis. 2014;20(10):1761–1769
34. El-Matary W, Moroz SP, Bernstein CN.Inflammatory bowel disease inchildren of Manitoba: 30 years’experience of a tertiary center.J Pediatr Gastroenterol Nutr. 2014;59(6):763–766
35. Manns MP, Czaja AJ, Gorham JD, et al;American Association for the Study ofLiver Diseases. Diagnosis andmanagement of autoimmune hepatitis.Hepatology. 2010;51(6):2193–2213
36. Czaja AJ, Carpenter HA. Autoimmunehepatitis with incidental histologicfeatures of bile duct injury. Hepatology.2001;34(4 pt 1):659–665
37. Alvarez F. Autoimmune hepatitis andprimary sclerosing cholangitis. Clin LiverDis. 2006;10(1):89–107, vi
38. Ovchinsky N, Moreira RK, Lefkowitch JH,Lavine JE. Liver biopsy in modern clinicalpractice: a pediatric point-of-view. AdvAnat Pathol. 2012;19(4):250–262
39. Gohlke F, Lohse AW, Dienes HP, et al.Evidence for an overlap syndrome ofautoimmune hepatitis and primarysclerosing cholangitis. J Hepatol. 1996;24(6):699–705
40. de Boer YS, van Nieuwkerk CM, Witte BI,Mulder CJ, Bouma G, Bloemena E.Assessment of the histopathological keyfeatures in autoimmune hepatitis.Histopathology. 2015;66(3):351–362
41. Björnsson E, Talwalkar J, TreeprasertsukS, Neuhauser M, Lindor K. Patients withtypical laboratory features ofautoimmune hepatitis rarely need a liverbiopsy for diagnosis. Clin GastroenterolHepatol. 2011;9(1):57–63
42. Mileti E, Rosenthal P, Peters MG.Validation and modification of simplifieddiagnostic criteria for autoimmunehepatitis in children. Clin GastroenterolHepatol. 2012;10(4):417–21.e1, 2
43. Hennes EM, Zeniya M, Czaja AJ, et al;International AutoimmuneHepatitis Group. Simplified criteriafor the diagnosis of autoimmune
PEDIATRICS Volume 136, number 5, November 2015 e1247
44. Woynarowski M, Nemeth A, Baruch Y,et al; European Autoimmune Hepatitis-Budesonide Study Group. Budesonideversus prednisone with azathioprine forthe treatment of autoimmune hepatitis
in children and adolescents. J Pediatr.2013;163(5):1347–53.e1
45. Dumortier J, Arita CT, Rivet C, et al. Long-term treatment reduction andsteroids withdrawal in children withautoimmune hepatitis: a single centreexperience on 55 children. Eur J
Gastroenterol Hepatol. 2009;21(12):1413–1418
46. Jones S, James E, Prasad S. Diseaseregistries and outcomes research inchildren: focus on lysosomal storagedisorders. Paediatr Drugs. 2011;13(1):33–47
