www.elsevier.com/locate/jhep

Journal of Hepatology 48 (2008) 68–73

Evolution in the management of acute liver failure-associatedaplastic anaemia in children: A single centre experience q,qq

Nedim Hadzic1,*, Susan Height2, Sarah Ball3, Mohamed Rela1, Nigel D. Heaton1,Paul Veys4, Giorgina Mieli-Vergani1

1Institute of Liver Studies, King’s College London School of Medicine at King’s College, London, UK2Department of Haematology, King’s College London School of Medicine at King’s College, London, UK

3Department of Haematology, St. George’s Hospital, London, UK4Bone Marrow Transplant Unit, Great Ormond Street Hospital for Children, London, UK

Background/Aims: Bone marrow failure (BMF) is a potentially life-threatening complication of acute liver failure (ALF).

Methods: To investigate prevalence and evolving management of BMF associated with ALF, we reviewed all cases seen in

our centre over 17 years. BMF was classified as: (a) bone marrow hypoplasia, (b) severe aplastic anaemia (SAA) and (c) very

severe aplastic anaemia (VSAA), using standard criteria. We compared outcomes in children receiving: (1) medical treatmentonly with or without immunomodulation (anti-lymphocyte globulin, calcineurin inhibitors, G-CSF); (2) medical treatment

with or without immunomodulation plus liver transplantation (LT); (3) haematopoietic stem cell transplantation (HSCT).

Results: Of 213 patients with ALF, 20 [(9.4%); 14 (70%) boys] developed BMF after a median of 1 month (range, 0.5 to

7). Seven had VSAA, 7 SAA and 6 bone marrow hypoplasia. Five children were treated medically, including 3 by immu-

nomodulation; 10 (50%) received LT, with immunomodulation in 6; 5 (25%) received HSCT, in one after LT. Four (20%)

children died, only one as consequence of AA. There was no difference in recovery, complication rates or outcome among

the three groups.

Conclusions: Aggressive management of ALF-associated AA, including immunomodulation, HSCT and LT, is success-ful in most cases. HSCT has the advantage of removing the risk of late clonal disorders.

� 2007 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Keywords: Liver transplantation; Bone marrow failure; Aplastic anaemia; Haemopoietic stem cell transplantation

0168-8278/$32.00 � 2007 European Association for the Study of the Liver.

doi:10.1016/j.jhep.2007.08.017

Received 1 May 2007; received in revised form 22 July 2007; accepted 1

August 2007; available online 25 October 2007

Associate Editor: P.-A Clavienq The authors declare that they do not have anything to disclose

regarding funding or conflict of interest with respect to this manuscript.qq This work was presented as a poster at the Annual Meeting of theAmerican Association for the Study of Liver Disease (AASLD) inBoston, USA, 26–31 October 2006.

* Corresponding author. Address: Paediatric Liver Centre, VarietyClub Children’s Hospital, King’s College Hospital, Denmark Hill,London SE5 9RS, UK. Tel.: +44 0 2032994120; fax: +44 0 2032994228.

E-mail address: nedim.hadzic@kcl.ac.uk (N. Hadzic).Abbreviations: ALF, acute liver failure; BMF, bone marrow failure;

AA, aplastic anaemia; SAA, severe aplastic anaemia; VSAA, very severeaplastic anaemia; G-CSF, granulocyte-colony stimulating factor; ATG,anti-thymocyte globulin; ALG, anti-lymphocyte globulin; HSCT, hae-matopoietic stem cell transplantation; MOF, multiorgan failure; LT, livertransplant; GvHD, graft-versus-host disease; CR, chronic rejection.

1. Introduction

Up to 28% of patients with acute liver failure (ALF)develop bone marrow failure (BMF), its severity rangingfrom mild pancytopenia to severe aplastic anaemia (AA)[1–8]. The pathogenesis of acquired BMF has beenlinked to exposure to viruses, chemicals, including drugsand environmental xenobiotics, but also to immunolog-ically-mediated injury directed to as yet unidentifiedendogenous antigens associated with haemopoietic stemcells [9].

Viruses such as hepatitis A, adenovirus and parvovi-rus B19 have been occasionally reported in AA associ-ated to ALF [10–12]. The commonest cause of ALFand ALF-associated BMF in children, however, remains

Published by Elsevier B.V. All rights reserved.

N. Hadzic et al. / Journal of Hepatology 48 (2008) 68–73 69

idiopathic [13,14]. As the name implies, the aetiology ofthis rare, but life-threatening, condition, variably calledcryptogenic, seronegative or non-A–E hepatitis, isunknown [15]. Associated clinical features often includea febrile prodrome, mild non-specific abdominal pain,diarrhoea and/or rash, suggesting the involvement ofan infectious agent.

The management of ALF-associated AA has mirroredthe management of AA at large. Early protocols includedimmunoablation with anti-thymocyte or anti-lymphocyteglobulins (ATG/ALG), combined with steroids andcyclosporine A (CyA), and intermittent use of granulo-cyte-colony stimulating factor (G-CSF) [16–18].Recently, in haematological units managing AA, haema-topoietic stem cell transplantation (HSCT) has becomethe preferred first-line treatment for acquired severe(SAA) or very severe aplastic anaemia (VSAA) in patientsyounger than 50 years of age, who have a matched siblingdonor [19,20]. The reported success rate of allogeneicHSCT is between 75% and 90%, the best outcome beingobserved when a fully human leukocyte antigen (HLA)-matched sibling donor is available [19,20]. In view of theseresults, in recent years we have introduced HSTC in themanagement of AA associated to ALF.

The aim of this study is to review retrospectively theevolving experience in the management of BMF associ-ated with ALF in children referred to our tertiary refer-ral centre and to determine the role of HSTC.

2. Methods

The Paediatric Liver Service at King’s College Hospital, London isa supraregional referral centre. A retrospective analysis of the patientdatabase between 1989 and 2005 was undertaken to identify childrenwith ALF and BMF. The diagnostic criteria for ALF included aninternational normalized prothrombin ratio (INR) >1.5 and bilirubin>100 lmol/L, with or without the presence of encephalopathy [13].

At presentation, other causes of liver disease such as autoimmu-nity, Wilson’s disease, alpha-1-antitrypsin deficiency, other metabolicconditions and drug toxicity were excluded by clinical history andappropriate investigations. Viral studies were performed using stan-dard serological methods for hepatitis A, B, C and (when indicated)E, cytomegalovirus, Epstein–Barr virus, herpes �1 and �2 viruses,parvovirus B19, rubella and human immunodeficiency virus �1 and�2.

Bone marrow aspirate and trephine biopsy were performed in allpatients developing pancytopenia, defined as: absolute neutrophilcount <1.0 · 109/L and platelet count <30 · 109/L. The diagnosis ofAA was made in the presence of bone marrow cellularity <25%, with2 out of 3 of the following: absolute neutrophil count <0.5 · 109/L(SAA), or <0.2 · 109/L (VSAA); platelet count <20 · 109/L; reticulo-cyte count <20 · 109/L. Cytogenetic analysis of the bone marrowwas performed in all cases.

In patients with AA, additional investigations included screeningfor paroxysmal nocturnal haemoglobinuria by flow cytometric analysisof GPI-anchored protein expression (glycophorin A, CD55, CD59 onthe red cells and CD13/33, CD16, CD66b on neutrophils) and exclu-sion of Fanconi anaemia by chromosome fragility testing on peripheralblood lymphocytes with diepoxybutane (DEB).

Since 1996, once the diagnosis of SAA was confirmed, family mem-bers underwent HLA typing by standard molecular methods to assesstheir suitability as potential donors for HSCT. If sibling donors were

available, the patients were referred to the national UK HSCT centresat St. George’s Hospital and Great Ormond Street Hospital, London,whilst those with no suitable donor received standard immunosuppres-sive treatment with ATG/ALG (ATGAM, Upjohn, Kalamazoo, MI,USA & Pasteur Merieux Connaught, Lyon, France; 20 mg/kg/d · 5),calcineurin inhibitors (CyA 10-15 mg/kg/d; Tac 0.1-0.2 mg/kg/d) andG-CSF (10 lg/kg/d) at King’s College Hospital until response isobserved [18].

The HSCT conditioning regimen included: Campath 1H 0.2 mg/kgDays �9 to �5, and cyclophosphamide (50 mg/kg) (with Mesna 5 mg/kg/h) Days �5 to �2. One patient, who received a low intensity (‘‘non-myeloablative’’) HSCT, was conditioned with melphalan (140 mg/m2

at Day �2), fludarabine (30 mg/m2 at Days �7 to �3) and ATG(ATGAM, Upjohn, Kalamazoo, MI, USA; 25 mg/10 kg at Days �1to +2). No manipulation of the bone marrow graft was undertakenin any of the transplanted patients.

Graft-versus-host disease (GvHD) prophylaxis from Day �1included calcineurin inhibitors: CyA (10 mg/kg/d) or, alternatively,for a patient who had received a liver transplant (LT) prior to HSCT,tacrolimus (0.2 mg/kg/d). Antibiotic, antiviral, antifungal and intrave-nous immunoglobulin prophylaxis was given to all patients post-HSCT.CyA was discontinued after 6 months in children with satisfactory bonemarrow function who had not undergone LT.

Standard criteria for response to medical treatment of AA wereused: haemoglobin >8 g/L, neutrophil count >0.5 · 109/L and plateletcount >20 · 109/L (partial), and haemoglobin >11 g/L, neutrophilcount >1.5 · 109/L and platelet count >100 · 109/L (complete). Post-HSCT engraftment was defined as: haemoglobin >8 g/L, platelet count>20 · 109/L, and neutrophil count >0.5 · 109/L for more than 3 days.

3. Results

Over the 17-year observation period, out of 213patients with ALF, 20 [9.4%; 14 (70%) boys] developedpancytopenia. On bone marrow aspirate, 6 had onlymild hypoplasia, while 14 [6.6%; 11 (78.5%) boys] hadAA (7 VSAA, 7 SAA). The cause of ALF was idiopathicin 18 children and neonatal haemochromatosis in 2.Overall, ALF was idiopathic in 96 of the 213 (45%) chil-dren [21], BMF developing in 20.8%, and AA in 14.5%of them. In all but one patient pancytopenia followedALF, after a median interval of 1 month (range, 0.5 to7); in the remaining patient it was present at diagnosisof ALF (#4, Table 1). Eleven (55%) patients requiredemergency LT. Five children (25%) were treated byHSCT, in one after LT.

Of 5 children (all boys) who did not receive LT orHSCT (Table 1), 2 had mild transient bone marrowhypoplasia not requiring treatment, while 3 had SAA.Median age at presentation with ALF was 31 months(range 0–120) and median interval to the developmentof pancytopenia 1 month (range 0–2). One each of thethree children with AA was treated with high dose ste-roids, G-CSF, or a combination of CyA (10 mg/kg/d)and ALG (Pasteur Merieux Connaught, Lyon, France)(10 mg/kg/d for 5 days), while the two with mild bonemarrow changes received supportive treatment only.Two patients died; a patient with neonatal haemochro-matosis and mild bone marrow hypoplasia of multior-gan failure two months after admission, and thepatient treated with high dose steroids for AA from sys-temic infection four days after developing ALF. This

Table 1

Patients managed by medical support only

Patientnumber

Year ofreferral

Diagnosis Age atALF onset(months)

Gender Interval fromALF to BMF(months)

Bone marrowaspirate

BMFtreatment

Interval toBMFrecovery(months)

Follow-up(months)

Outcome

1 2000 NH 0 M 1 Mild hypoplasia Nil – 2 Death (MOF)2 2000 Idiopathic 37 M 1 Mild hypoplasia Nil 4 76 Alive & well3 2000 Idiopathic 31 M 2 SAA G-CSF 8 68 Alive & well4 1991 Idiopathic 8 M 0 SAA Steroids – 0 Death (sepsis)5 2000 Idiopathic 120 M 2 SAA CyA, ALG 6 63 Alive & well

NH, neonatal haemochromatosis; ALF, acute liver failure; BMF, bone marrow failure; SAA, severe aplastic anaemia; CyA, cyclosporine A; G-CSF,granulocyte colony stimulating factor; ALG, anti-lymphocyte globulin; MOF, multiorgan failure.

70 N. Hadzic et al. / Journal of Hepatology 48 (2008) 68–73

child had been diagnosed with dyserythropoietic anae-mia at the age of 4 months, before developing concom-itantly ALF and AA at the age of 8 months. Theremaining two children with AA achieved partialresponse at 4 and 2 months, and complete response at6 and 8 months after starting the treatment, respectively.

Of 10 children (6 boys), who received LT alone(Table 2), 4 had mild transient bone marrow hypoplasianot requiring treatment, while 6 fulfilled the AA criteria(4 SAA, 2 VSAA). Median age at presentation withALF was 71 months (range 0–177) and median intervalto the development of pancytopenia 1 month (range 0.5–7). The median period between admission to our unitand liver transplantation was 9 days (range 3–27). The6 children with AA were treated with the calcineurin

Table 2

Patients managed by liver transplantation for ALF and medical support for B

Patient

number

Year of

referral

Diagnosis Age at

ALF onset

(months)

Gender Interval

from

ALF to

LT (days)

Interval

from ALF

to BMF

(months)

Bone

aspir

1 1998 Idiopathic

(parvovirus

B19)

108 M 22 0.5 VSA

2 2000 Idiopathic 82 M 4 4 SAA

3 2003 NH 0 M 14 1 SAA

4 1995 Idiopathic 49 F 3 1 VSA

5 1997 Idiopathic 60 M 4 1 Mild

6 1997 Idiopathic 35 F 17 1 SAA

7 2002 Idiopathic 177 F 22 1 Mild

8 1999 Idiopathic 36 M 3 7 Mild

9 1999 Idiopathic 100 F 27 4 Mild

10 2001 Idiopathic 173 M 4 0.5 SAA

NH, neonatal haemochromatosis; ALF, acute liver failure; BMF, bone maranaemia; G-CSF, granulocyte-colony stimulating factor; ATG, anti-thymocyrejection; HAT, hepatic artery thrombosis; Tac, tacrolimus; CyA, cyclospor

inhibitor used as primary anti-rejection drug (tacrolimusin 5, CyA in 1) and recombinant G-CSF (5 lg/kg/d),which were given in combination with ATG (ATGAM;Upjohn, Kalamazoo, MI, USA) (10 mg/kg/d for 5 days)in 2. Two children with AA died of multiorgan failure:one with neonatal haemochromatosis 3 weeks after pre-sentation, and one with idiopathic ALF 74 months afterthe first admission and following retransplantation forchronic rejection. In this patient AA had resolved 8weeks after presentation. Median time to achieve a par-tial response to the treatment for the surviving childrenwas 2 months (range 0.5–14).

Five patients (4 boys), all with VSAA, received HSCT(Table 3). Median age at presentation of ALF was 70months (range 61–155) and median interval to the devel-

MF

marrow

ate

BMF treatment Interval to

BMF

recovery

(months)

Follow-up

(months)

Outcome

A G-CSF, ATG, (Tac

for LT)

1.5 87 Alive & well post

auxiliary LT

(no treatment)

G-CSF (Tac for

LT)

3 66 Alive & well (Tac

for LT)

G-CSF (Tac for

LT)

– 0.75 Death (MOF)

A G-CSF, ATG,

(CyA for LT)

4 132 Alive & well (CyA

for LT)

hypoplasia Nil (CyA for LT) 1 112 Alive & well after

re-LT (x2) for HAT

(Tac for LT)

G-CSF (Tac for

LT)

2 74 Death after re-LT

for CR (MOF)

hypoplasia Nil (Tac for LT) 0.5 28 Alive & well (Tac

for LT)

hypoplasia Nil (Tac for LT) 6 71 Alive & well (Tac

for LT)

hypoplasia Nil (CyA for LT) 0.5 84 Alive & well (CyA

for LT)

G-CSF (Tac for

LT)

14 55 Alive & well after

re-LT for CR (Tac

for LT)

row failure; SAA, severe aplastic anaemia; VSAA, very severe aplasticte globulin; MOF, multiorgan failure; LT, liver transplant; CR, chronicine A.

Table 3

Patients managed by haemopoietic stem cell transplantation

Patient Year of

referral

Diagnosis Age at

ALF

onset

(months)

Gender Interval

from ALF

to BMF

(months)

Bone

marrow

aspirate

Pre-HSCT

treatment

Interval from

BMF to

HSCT

(months)

Interval to

recovery

(months)

Follow-up

(months)

Outcome

1 2003 Idiopathic ALF 155 M 0.5 VSAA Nil 2 7 30 Alive & well

(no treatment)

2 2003 Idiopathic ALF 86 M 4 VSAA G-CSF 6 3 48 Alive & well

(no treatment)

3 1997 Idiopathic ALFa 64 M 1 VSAA G-CSF,

ALG

17 6 81 Alive & well

(no treatment)

4 2005 Idiopathic ALF 70 F 1 VSAA LT 2 8 10 Alive & well

(Tac for LT)

5 2005 Idiopathic ALF 61 M 3 VSAA Nil 1.5 3 15 Alive & well

ALF, acute liver failure; BMF, bone marrow failure; VSAA, very severe aplastic anaemia; HSCT, haemopoietic stem cell transplantation; G-CSF,granulocyte colony stimulating factor; ALG, anti-lymphocyte globulin; LT, liver transplantation; Tac, tacrolimus.

a Developed hyperplastic myelodysplasia with monosomy 7 after ATG treatment.

N. Hadzic et al. / Journal of Hepatology 48 (2008) 68–73 71

opment of BMF 1 month (range 0.5–4). One girl (#4)underwent LT before HSCT. Four children receivedHSCT from fully-matched sibling donors, while one(#3) did not have a sibling donor. This child had normalblood counts on admission with ALF, but subsequentlydeveloped VSAA with no cytogenetic abnormality. Hehad a poor response to treatment with ATG, CyA andG-CSF, achieving neutrophil recovery >1.0 · 109/L,but remaining transfusion-dependent. Ten months latera repeated bone marrow aspirate for persisting pancyto-penia showed a hypercellular marrow with dysplasticfeatures, 5% blasts and monosomy 7. G-CSF and CyAwere stopped, and within two months he developedacute myeloblastic leukaemia. After achieving remissionwith vincristine and prednisolone treatment, he under-went a non-myeloablative HSCT from a matched unre-lated donor.

Post-HSCT no patient developed clinical features ofGvHD, sinusoidal obstructive syndrome, drug hepato-toxicity or serious infections. The children achievedengraftment at a median of 6 weeks (range 3–8) afterHSCT. Liver function tests had returned to normalwithin 2 months after presentation with ALF and beforeHSCT in 4 patients (including the child, who received aliver transplant), while the fourth boy (#1) had mini-mally deranged serum bilirubin and aspartate amino-transferase levels until 22 months post-HSCT. Allchildren are alive and well at a median follow up of 30months (range 10–81) after HSCT.

Though the patient number is too small for statisticalanalysis, we have not observed a difference in the age atpresentation, interval to develop pancytopenia and timeto recover amongst the three groups of patients. Onlyone previously reported [6] patient (#1, Table 2), whoreceived auxiliary LT, had serological evidence of acuteparvovirus B19 infection. All other viral studies werenegative. Overall, four (20%) children died, but onlyone as a direct consequence of AA. No patient diedamong those treated by HSCT.

4. Discussion

This single centre study, based on a relatively largenumber of children with ALF, shows that 6.6% developaplastic anaemia and that an aggressive therapeuticapproach, including immunomodulation, HSCT, and,when indicated, LT leads to 80% long-term survival.

Almost four-fifths (78.5%) of our patients are boys,confirming the previously observed male preponderanceamong adults or children who develop acute hepatitis-or ALF-associated AA [7,8]. The cause of this genderdifference is unclear and unlikely to be due to hormonalfactors given the young, pre-pubertal, age of most of ourpatients.

The commonest underlying diagnosis among childrendeveloping pancytopenia in this study is idiopathicALF, also confirming previous reports [2,3,6–8]. Weshow that BMF occurs in 20.8%, and AA in 14.5% ofchildren with this condition. In idiopathic ALF an infec-tious aetiology is suspected, though not proven. Thesame putative infectious agent causing ALF could alsobe responsible for the development of BMF, as sug-gested by the short time interval between the occurrenceof hepatitis and the onset of pancytopenia. Interestingly,the only two children for whom the cause of ALF wasidentified in our series had neonatal haemochromatosis,a condition where infectious/immune mechanisms arealso invoked [22].

Only one of our patients was positive both by serol-ogy and polymerase chain reaction (PCR) for parvovi-rus B19, an ubiquitous community virus, which hasbeen suggested as cause of idiopathic ALF and AA[10–12]. Though this low detection rate of parvovirusB19 infection may be due to the use of insensitive, anti-body-based, techniques in most of our patients, also inan earlier study from our unit, in which PCR was used,only two of eight children with ALF and AA were par-vovirus B19 positive [6]. In accordance with these find-ings, a recent PCR-based controlled study was unable

72 N. Hadzic et al. / Journal of Hepatology 48 (2008) 68–73

to prove an association between parvovirus B19 infec-tion and ALF-associated AA, showing in addition thatparvovirus B19 is no more frequently present in the livertissue of patients with idiopathic ALF than in the livertissue of patients with confirmed HBV or HCV infec-tions [23].

An infection driven immunopathogenesis for AAassociated to non-A–E acute hepatitis is supported bythe demonstration of peripheral blood lymphocyte acti-vation [5,24], and liver infiltration by both clonal andnon-clonal T-cells, giving rise to a markedly skewed T-cell repertoire similar to that observed in the liver ofpatients with viral hepatitis B or C [25]. The reportedclinical response of these patients to either immunosup-pression [6,18] or bone marrow transplantation [8,19]further supports an immune-mediated pathogenesis.

The management of ALF-related AA, a rare but life-threatening condition, is not standardised. Our earlytherapeutic approach was to intensify immunosuppres-sion using immunoablation with ATG/ALG, calcineu-rin inhibitors and steroids [6,18], according tocontemporary haematological guidelines. More recently,HSCT has become our standard mode of treatment ofSAA or VSAA if a matched related donor is available.Both therapeutic approaches require long-term hospital-isation, blood product support and are associated with ahigh susceptibility to infection. Our cohort was not largeenough to determine a difference in complicationsbetween patients treated medically or with HSCT,though the only child who died of infection was treatedby immunomodulation.

An important advantage of HSCT is the eliminationof possible late haematological complications, reportedto occur in up to 10% of patients with acquired AA afterimmunosuppressive treatment [19,20]. After non-HSCTtreatment, the patients remain at risk of relapsing ordeveloping clonal bone marrow abnormalities, such asparoxysmal nocturnal haemoglobinuria and myelodys-plasia [20], which may be also associated with clonalcytogenetic abnormalities.This scenario is well illus-trated by our patient #3, who did not have a siblingdonor and failed to respond to ATG. He subsequentlydeveloped hyperplastic myelodysplasia with monosomy7, rapidly followed by acute myeloblastic leukaemia,successfully treated by matched unrelated donor HSCT.HSCT, therefore, appears to offer a cure for the acutebone marrow injury with the added benefit of removingthe risk of developing clonal disorders, whereas manip-ulation of the bone marrow by intense immunosuppres-sion and G-CSF may theoretically increase such risk.

Of note, the rate of liver retransplantation amongstthe children with pancytopenia requiring liver graftingin this series appears to be higher (3/11, 27%) than thosepreviously observed in our centre after LT performedfor other conditions (13.3%) and also after LT forALF overall (18.9%) [26], possibly indicating the ongo-

ing presence of a pathogenic process that could be atthe basis of both progressive liver injury and the devel-opment of haematological complications.

In this series, overall mortality among children withALF-related pancytopenia was 20% and in those withconfirmed AA 21.5%, but was mostly related to hepaticcomplications, with only one (7.1%) of the 14 childrenwith AA dying as a direct consequence of AA treatment.Previous small series have reported a mortality of 43%[2,4]. In a recent report, out of 9 children who developedAA after LT for ALF, 7 survived (78%), 2 being treatedby bone marrow transplantation and 5 by immunomod-ulation [8].

Post-HSCT, our patients had no serious infectious orimmunological complications. Whilst the lack of life-threatening infections could be explained by vigorousanti-microbial prophylaxis, brief periods of severe cyto-penia and prompt recovery of liver function, the absenceof clinical features of GvHD in our five patients isintriguing and different from the experience in adult withhepatitis-associated AA, where following bone marrowtransplantation from their HLA-identical siblings 10 of18 (56%) patients with sustained engraftment developedacute and/or chronic GvHD, fatal in 2 [27]. This dis-crepancy, which may be due to the small number orthe age of our patients, should be verified in a larger pae-diatric cohort.

Our current practice is to work up patients withALF-associated AA for HSCT as soon as the diagnosisof AA is made. In the absence of an HLA-matched sib-ling donor, early immunomodulatory treatment with acombination of ATG and calcineurin inhibitors isstarted. If a constitutional or clonal chromosomalabnormality is detected, matched unrelated HSCT isconsidered. In the presence of an HLA-matched siblingdonor, we proceed to early HSCT, but preferably whenliver function has returned to normal either spontane-ously or after LT.

In conclusion, this retrospective study shows thataggressive management of ALF-associated AA, includ-ing immunomodulation, HSCT and, when necessary,LT, is successful in most cases. Due to the rarity ofALF-related AA, a randomised prospective study toinvestigate the effectiveness of a conservative immuno-modulatory approach versus HSCT is unlikely to be fea-sible. Though no significant difference in outcomebetween HSCT and immunomodulation on its ownwas demonstrated in our series, the former has theadvantage of removing the risk of developing clonaldisorders.

References

[1] Stock PG, Steiner ME, Freese D, Sharp H, Ascher NL. Hepatitis-associated aplastic anemia after liver transplantation. Transplan-tation 1987;43:595–597.

N. Hadzic et al. / Journal of Hepatology 48 (2008) 68–73 73

[2] Tzakis AG, Arditi M, Whitington PF, Yanaga K, Esquivel C,Andrews WA, et al. Aplastic anemia complicating orthotopicliver transplantation for non-A, non-B hepatitis. N Engl J Med1988;319:393–396.

[3] Schlitt HJ, Ringe B, Rodeck B, Burdelski M, Kuse E, PichlmayrR. Bone marrow dysfunction after liver transplantation forfulminant non-A, non-B hepatitis. High risk for young patients..Transplantation 1992;54:936–937.

[4] Cattral MS, Langnas AN, Markin RS, Antonson DL, HeffronTG, Fox IJ, et al. Aplastic anemia after liver transplantation forfulminant liver failure. Hepatology 1994;20:813–818.

[5] Brown KE, Tisdale J, Barrett AJ, Dunbar CE, Young NS.Hepatitis-associated aplastic anemia. N Engl J Med1997;336:1059–1064.

[6] Tung J, Hadzic N, Layton M, Baker AJ, Dhawan A, Rela M,et al. Bone marrow failure in children with acute liver failure. JPediatr Gastroenterol Nutr 2000;31:557–561.

[7] Itterbeek P, Vandenberghe P, Nevens F, Fevery J, Aerts R, YapSH, et al. Aplastic anemia after transplantation for non-A, non-B, non-C fulminant hepatic failure: case report and review of theliterature. Transpl Int 2002;15:117–123.

[8] Molina RA, Katzir L, Rhee C, Ingram-Drake L, Moore T,Krogstad P, et al. Early evidence of bone marrow dysfunction inchildren with indeterminate fulminant hepatic failure who ulti-mately develop aplastic anemia. Am J Transplant2004;4:1656–1661.

[9] Young NS, Maciejewski J. The patophysiology of acquiredaplastic anaemia. N Engl J Med 1997;336:1365–1372.

[10] Yoto Y, Kudoh T, Asanuma H, Numazaki K, Tsutsumi Y,Nakata S, et al. Transient disturbance of consciousness andhepatic dysfunction associated with human parvovirus B19infection. Lancet 1994;344:624–625.

[11] Langnas AN, Markin RS, Cattral MS, Naides SJ. Parvovirus B19as a possible causative agent of fulminant liver failure andassociated aplastic anemia. Hepatology 1995;22:1661–1665.

[12] Sokal EM, Melchior M, Cornu C, Vandenbroucke AT, Buts JP,Cohen BJ, et al. Acute parvovirus B19 infection associated withfulminant hepatitis of favourable prognosis in young children.Lancet 1998;352:1739–1741.

[13] Dhawan A, Cheeseman P, Mieli-Vergani G. Approaches to acuteliver failure in children. Pediatr Transpl 2004;8:584–588.

[14] Devictor D, Desplanques L, Debray D, Ozier Y, Dubouset AM,Valayer J, et al. Emergency liver transplantation for fulminant

liver failure in infants and children. Hepatology1992;15:1156–1162.

[15] Bhaduri BR, Mieli-Vergani G. Fulminant hepatic failure: pediat-ric aspects. Semin Liver Dis 1996;16:349–355.

[16] Frickhofen N, Kaltwasser JP, Schrezenmeier H, Raghavachar A,Vogt HG, Herrmann F, et al. Treatment of aplastic anemia withantilymphocyte globulin and methylprednisolone with or withoutcyclosporine. N Engl J Med 1991;324:1297–1304.

[17] Rosenfeld SJ, Kimball J, Vining D, Young NS. Intensiveimmunosuppression with antithymocyte globulin and cyclospor-ine as treatment for severe acquired aplastic anemia. Blood1995;85:3058–3065.

[18] Hadzic N, Layton M, Heaton ND, Rela M, Baker AJ, Mieli-Vergani G. Anti-thymocyte globulin, cyclosporine A and granu-locyte colony-stimulating factor for severe aplastic anaemiacomplicating paediatric liver transplantation. Eur J Pediatr1998;157:107–108.

[19] Ball SE. The modern management of severe aplastic anaemia. Br JHematol 2000;110:41–53.

[20] Young NS. Acquired aplastic anemia. Ann Intern Med2002;136:534–546.

[21] Bansal S, Dhawan A. Acute liver failure. Indian J Pediatr2006;73:931–934.

[22] Whitington PF, Hibbard JU. High-dose immunoglobulin duringpregnancy for recurrent neonatal haemochromatosis. Lancet2004;364:1690–1698.

[23] Wong S, Young NS, Brown KE. Prevalence of parvovirus B19in liver tissue: no association with fulminant hepatitis orhepatitis-associated aplastic anemia. J Infect Dis 2003;187:1581–1586.

[24] Kojima S, Matsuyama K, Kodera Y, Okada J. Circulatingactivated suppressor T lymphocytes in hepatitis-associated aplas-tic anaemia. Br J Haematol 1989;1:147–151.

[25] Lu J, Basu A, Melenhorst J, Young NS, Brown KE. Analysis ofT-cell repertoire in hepatitis-associated aplastic anemia. Blood2004;103:4588–4593.

[26] Deshpande RR, Rela M, Girlanda R, Bowles MJ, Muiesan P,Dhawan A, et al. Long-term outcome of liver retransplantation inchildren. Transplantation 2002;74:1124–1130.

[27] Kiem HP, McDonald GB, Myerson D, Spurgeon CL, Deeg HJ,Sanders JE, et al. Marrow transplantation for hepatitis-associ-ated aplastic anemia: a follow-up of long-term survivors. BiolBlood Marrow Transpl 1996;2:93–99.