Increased incidence of EBV-related disease following paediatric stem cell transplantation with reduced-intensity conditioning

Increased incidence of EBV-related disease following paediatricstem cell transplantation with reduced-intensity conditioning

Epstein–Barr virus (EBV) is a c-herpes virus with B lympho-

cyte-specific tropism that causes significant clinical problems

in immunosuppressed patients of all ages. Following haema-

topoietic stem cell transplantation (HSCT), EBV reactivation

and related disease are a recognized complication, with an

incidence of 0Æ6–26%, which is higher in the context of

selective T-cell depletion (TCD; Antin et al, 1991; Lucas et al,

1998; Micallef et al, 1998; van Esser et al, 2001a).

EBV may be associated with a spectrum of clinical presen-

tations, from fever to lymphoproliferative disease (LPD),

which arise as a consequence of an outgrowth of B cells latently

infected with EBV in the setting of loss or suppression of

normal cytotoxic T-cell surveillance. LPD itself has a heterog-

enous clinico-pathological presentation, ranging from poly-

morphic hyperplasia to diffuse large B-cell lymphoma.

Established LPD post-HSCT is associated with a high mortality

and morbidity (Gross et al, 1999). With improved monitoring

and detection of EBV viraemia using polymerase chain

reaction (PCR)-based techniques, more accurate analysis of

the risk factors associated with EBV-related disease following

HSCT is now possible (Rooney et al, 1998; Milpied et al, 2000;

van Esser et al, 2002).

Several reports have attempted to define HSCT risk factors

that predispose to the development of EBV-LPD. A multicen-

tre retrospective analysis of over 18 000 transplants identified

the use of an unrelated donor, greater than single human

leucocyte antigen (HLA)-mismatch, TCD of the graft, use of

antithymocyte globulin (ATG) or anti-CD3 in vivo or the

presence of graft-versus-host disease (GvHD) as independent

risk factors (Curtis et al, 1999). In another single-centre

retrospective analysis of 2150 transplants, TCD, HLA-mis-

match, the use of ATG in vivo and primary immunodeficiency

(PID) as an underlying disease, were identified as independent

risk factors (Bhatia et al, 1996). Several other reports have

Jonathan Cohen, 1* Minal Gandhi, 1*

Paru Naik,1 David Cubitt,2 Kanchan

Rao,1 Urvashi Thaker,2 E. Graham

Davies,3 H. Bobby Gaspar, 3,4 Persis J.

Amrolia1,4 and Paul Veys1

Departments of 1Bone Marrow Transplantation,2Virology and 3Clinical Immunology, Great

Ormond Street Hospital NHS Trust, London, and4Molecular Immunology Unit, Institute of Child

Health, University College London, London, UK

Received 18 November 2004; accepted for

publication 26 January 2005

*These authors contributed equally to this study.

Correspondence: Dr Persis J. Amrolia,

Department of Bone Marrow Transplantation,

Great Ormond St Hospital, Great Ormond St,

London WC1N 3JH, UK.

E-mail: [email protected]

Summary

The incidence of Epstein–Barr virus (EBV) viraemia and lymphoproliferative

disease (LPD) was studied in a consecutive cohort of 128 paediatric patients

undergoing stem cell transplantation (SCT) with reduced-intensity

conditioning (RIC; n ¼ 65) or conventional-intensity conditioning (CIC;

n ¼ 68). Following CIC, six of 68 (8%) developed viraemia; all remained

asymptomatic. EBV viraemia (23 of 65 patients ¼ 35%, P < 0Æ001) and LPD

(10 of 65 ¼ 15%, P < 0Æ001) were significantly more frequent following RIC.

Of the 23 RIC patients who developed viraemia, eight remained

asymptomatic, five had symptomatic viraemia (fever ± rash), and 10

patients developed LPD, two of whom died. An absolute lymphocyte count

of <0Æ3 · 109/l at the time of onset of viraemia was strongly predictive of

development of LPD (P < 0Æ05) in this group. The incidence of viraemia was

significantly higher in patients receiving serotherapy with antithymocyte

globulin (ATG; 15 of 43, 35%) than Campath (12 of 73, 16Æ4%, P < 0Æ05).

Primary immunodeficiency and acute graft-versus-host disease were

associated with EBV viraemia in univariate analysis, but were not

independent risk factors. In conclusion, EBV viraemia and LPD appear to

be significantly more common in children following RIC SCT, particularly

with selective depletion of recipient T cells relative to B cells following the use

of ATG. This probably reflects the profound immunosuppression following

RIC SCT, together with the incomplete ablation of recipient-derived B cells.

Keywords: Epstein–Barr virus, lymphoproliferative disease, reduced-inten-

sity conditioning, stem cell transplant, viraemia.

research paper

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 129, 229–239 doi:10.1111/j.1365-2141.2005.05439.x

confirmed that selective depletion of T cells, rather than both

T- and B-cell populations, from the graft and HLA-mismatch

increase the risk of EBV-LPD (Hale & Waldmann, 1998; Meijer

et al, 2002). In these reports, the majority of patients treated

were adults and only EBV-LPD rather than other EBV-

associated morbidity was documented. In all these studies

conventional myeloablative conditioning regimes (CIC) were

used for cytoreduction.

Increasing numbers of SCTs are now performed using non-

myeloablative or reduced-intensity conditioning (RIC)

regimes, which are designed to produce a state of profound

immunosuppression rather than myeloablation. RIC has been

used with good results for the treatment of both malignant and

non-malignant haematological conditions in adults and chil-

dren (Slavin et al, 1998; Gross et al, 1999; Amrolia et al, 2000).

To date, very little data has been reported regarding the

incidence of EBV-related disease following RIC, although

evidence suggests that LPD maybe problematic in this setting.

A single, fatal case LPD following RIC (Milpied et al, 1999) and

a series of 30 RIC transplants, one of which led to LPD, have

recently been described (Ho et al, 2002). Furthermore, the

delayed recovery of EBV-specific cytotoxic T-lymphocytes

following a RIC fludarabine/Campath 1H regime when com-

pared with conventional conditioning regimes, may predispose

to higher incidence of EBV disease (Chakrabarti et al, 2003).

Serial PCR-based monitoring for EBV in the blood of

patients following HSCT (Lucas et al, 1998; Hoshino et al,

2001) has enabled accurate determination of EBV viral load

and prediction of patients at risk of progression to overt EBV-

LPD. Pre-emptive antiviral or anti-B-cell therapies are there-

fore now regularly used and have been shown to be successful

in preventing LPD (van Esser et al, 2001b, 2002; Lankester

et al, 2002). As a result, it is increasingly difficult to use EBV-

LPD incidence as the outcome measure in studies looking at

risk factors for its development. However, demonstrating risk

factors that predispose to EBV viraemia following transplant

will be of great use in identifying patients at increased risk of

EBV-related disease.

This series is the largest reported to date, of EBV-related

disease following HSCT in paediatric patients. We compared the

incidence of EBV viraemia in a cohort of 128 patients undergoing

allogeneic SCT with CIC or RIC at our institution. In addition,

we examined the role of potential risk factors, including PID as

an underlying diagnosis, TCD and serotherapy and degree of

HLA-mismatch, in the development of EBV-related disease. We

demonstrated for the first time that EBV-related disease is

significantly higher following RIC compared with CIC.

Methods

Patients

From January 1999, all children undergoing allogeneic SCT at

Great Ormond Street Hospital, London, were screened for

EBV viraemia by DNA-PCR weekly until the circulating CD4

count was > 0Æ3 · 109/l, and were monitored for clinical

features attributable to EBV. A total of 133 allogeneic

transplants were performed on 128 patients between January

1999 and June 2002, with five patients undergoing two

transplants during this period. Their details are shown in

Table I. Ages ranged from 0 to 17Æ7 years (median: 4Æ1). About

42 patients (32Æ8%) were transplanted for haematological

Table I. Details of 133 allogeneic HSCT between January 1999 and

June 2002, by intensity of conditioning.

RIC CIC Total P-value

N 65 68 133

Female (%) 49Æ2 50 NS*

Age range,

years (median)

0Æ19–17Æ7(5Æ17)

0Æ23–13Æ4(3Æ58)

0Æ19–17Æ7(4Æ08)

0Æ149**

Disease

MH 5 37 42 <0Æ001*

NMH 15 14 29

PID 41 8 49

Other 4 9 13

Donor

Haplo 4 11 15 NS*

MMUD 14 11 25

MMFD 1 0 1

MMSD 1 0 1

MUD 25 20 45

MFD 7 7 14

MSD 13 19 32

Cells

BM 58 54 112 NS*

PBSC 6 13 19

BM + PBSC 1 0 1

Cord 0 1 1

TCD

Yes 4 29 33 <0Æ001*

No 61 39 100

Serotherapy

ATG 30 13 43 <0Æ001*

Campath 1G 0 15 15

Campath 1H 34 24 58

Neither 1 16 17

Mixed chimaeras (%)

At 6 weeks 37 14Æ5 24Æ8 <0Æ05*

At 12 weeks 37Æ3 28Æ3 33 <0Æ05*

CIC, conventional-intensity conditioning; RIC, reduced-intensity

conditioning; MH, malignant haematology; NMH, non-malignant

haematology; PID, primary immunodeficiency; MMUD, HLA-mis-

matched unrelated donor; MMFD, HLA-mismatched family donor;

MMSD, HLA-mismatched sibling donor; MUD, HLA-matched unre-

lated donor; MFD, HLA-matched family donor; MSD, HLA-matched

sibling donor; BM, bone marrow; PBSC, peripheral blood stem cells;

TCD, ex vivo T-cell depletion; ATG, antithymocyte globulin; NS, no

significant difference at 5% level; HSCT, haematopoietic stem cell

transplantation.

*Comparison between RIC and CIC groups made by chi-squared or

Fisher’s exact test, with P-value shown.

**P-value from unpaired t-test.

J. Cohen et al

230 ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 129, 229–239

malignancy, 27 (21Æ1%) for non-malignant haematological

disease, 47 (36Æ7%) for PID and 12 (9Æ4%) for metabolic and

other disorders.

Conditioning and matching of RIC/CIC groups

The conditioning regimes were categorized as myeloablative

CIC (n ¼ 68) or non-myeloablative RIC (n ¼ 65). Of the CIC

patients, 30 received conditioning incorporating total body

irradiation, whereas 38 received chemotherapy-alone regimens

(busulphan 16 mg/kg + cyclophosphamide 200 mg/kg,

n ¼ 22; busulphan 16 mg/kg + cyclophosphamide

120 mg/kg + melphalan 140 mg/m2, n ¼ 9; fludarabine

150 mg/m2 + busulphan 16 mg/kg + melphalan 140 mg/m2,

n ¼ 7). Fifty-nine of the 65 RIC transplants used fludarabine-

based regimens and the majority of these were with fludarabine

150 mg/m2 + melphalan 140 mg/m2. Only four of 65 RIC

transplants used an in vitro TCD graft, compared with 29 of

the 68 CIC transplants. The CIC and RIC groups were matched

for sex, donor type and cell source. There was a slight

difference in age between the two groups (median 5Æ17 years

for RIC, 3Æ58 years for CIC). The CIC group contained a larger

proportion of patients treated for malignant haematological

conditions than the RIC group (54Æ4% vs. 7Æ7%), whereas the

RIC group contained more patients with primary deficiency

than the CIC group (63Æ1% vs. 11Æ8%). Differences between

the groups with regard to in vitro TCD, in vivo serotherapy and

the presence of mixed chimaerism are discussed further below.

Grafts

Of the 133 allogeneic transplants, 91 involved a fully HLA-

matched donor: matched sibling donor (n ¼ 45); matched

family donor (n ¼ 14); matched unrelated donor (n ¼ 32).

Forty-two were HLA-mismatched: 1-antigen (Ag) mismatched

family donor (n ¼ 2); 1-Ag mismatched unrelated donor

(MUD; n ¼ 25); haplo-identical donor (n ¼ 15). Bone mar-

row was used as the stem cell source in 112 transplants and 19

patients received peripheral blood stem cells. One patient

received both bone marrow and peripheral blood stem cells.

One patient received cord blood. In vitro TCD was performed

in 33 grafts (Campath 1M + T-cell addback, n ¼ 18; Miltenyi,

n ¼ 15). Only four of 65 (6Æ2%) RIC transplants involved

in vitro TCD of the graft compared with 29 of 68 (46Æ2%) CIC

transplants.

Serotherapy

One hundred and sixteen patients of the 133 also received

serotherapy in vivo with conditioning to reduce rejection and

GvHD: 43 with ATG, 73 with Campath (Campath 1G, n ¼ 15;

Campath 1H, n ¼ 58). In 1999 and early 2000, MUD patients

received Campath 1G. Because of observed high incidence of

EBV in RIC patients, Campath 1H 0Æ2 mg/kg · 5 doses was

used as serotherapy in RIC patients from December 1999, and

from June 2000 onwards, all patients receiving in vivo

Campath, received Campath 1H. As a result of this, RIC

patients were more likely to have received ATG than CIC

patients (30 of 65, 46Æ2% vs. 13 of 68, 19Æ1%), less likely to

have received Campath 1G (zero of 65, 0% vs. 15 of 68, 22Æ1%)

and more likely to have received Campath 1H (34 of 65, 52Æ3%

vs. 24 of 68, 35Æ3%). Sixteen of 68 (23Æ5%) CIC patients

received no serotherapy compared with only one of 65 (1Æ5%)

RIC patients.

Chimaerism

Thirty-seven per cent RIC patients had mixed chimaerism at

both 6- and 12-weeks post-transplant, compared with 14Æ5 and

28Æ3% in CIC patients.

EBV detection

All patients had weekly DNA-PCR monitoring for EBV,

cytomegalovirus (CMV) and adenovirus. This was continued

until CD4+ cell counts were >0Æ3 · 109/l. Semi-quantitative

analysis for EBV viral load was also performed if weekly

monitoring proved persistently positive in both blood and

plasma. DNA was extracted from 200 ll of whole blood in

ethylenediaminetetraacetic acid (EDTA) or plasma using the

QIAamp DNA blood minikit (Qiagen Ltd, Crawley, UK)

according to the manufacturer’s instructions.

For EBV detection, a PCR using primers IR1a-5¢ and IR1a-

3¢, which amplify a specific region of 288 bp within the

internal repeat (IR), was performed as previously described

(Shroff et al, 2002). Negative and positive controls were

included in each run. If EBV-DNA was detected in whole

blood, a semi-quantitative assay to determine viral load was

performed the following day on plasma using log-dilution,

comparing with known dilutions of a plasmid containing the

IR region. The threshold of detection of the assay was found to

be approximately 100 DNA copies/ml when samples were run

in parallel with an external laboratory that uses a light cycler

for quantification and EBV-infected cells (10 genomic EBV-

DNA copies/cell). The presence of EBV-DNA was regarded as

significant if it was detected in plasma as well as whole blood,

indicating reactivation and probable lytic replication of latent

virus within B cells. This is referred to as EBV viraemia, below.

Disease classification and treatment

All patients were closely monitored for clinical features

attributable to EBV infection. Those with EBV viraemia were

categorized as either (i) asymptomatic, (ii) symptomatic

viraemia (culture-negative fever with or without rash) or

(iii) LPD. EBV-LPD was subclassified clinically as either

localized or disseminated and lymphadenopathic or lympho-

matous, according to the Pittsburgh classification (Nalesnik,

1998). Biopsy specimens were classified according to the

World Health Organization (WHO) classification (Harris et al,

EBV Disease after Reduced-intensity Transplantation

ª 2005 Blackwell Publishing Ltd, British Journal of Haematology, 129, 229–239 231

1999). The presence of EBV in biopsy samples was investigated

using in situ hybridization for EBV-encoded RNAs (EBERs).

Patients with viraemia or LPD were treated initially with

withdrawal of immunosuppression ± antivirals. If there was

no response, patients were treated with rituximab 375 mg/m2

weekly for one to four doses until resolution of viraemia and/

or complete resolution of LPD.

Statistical analysis

In the comparison of baseline characteristics of the RIC and

CIC groups, chi-squared or Fisher’s exact tests were used for

non-parametric variables. Student’s t-test was used for para-

metric variables.

Overall survival was analysed using a Cox proportional

hazards model. Survival in patients with EBV viraemia post-

transplant was compared to those without viraemia. A further

comparison was made between those who developed LPD and

those who had EBV viraemia without developing LPD.

A univariate analysis was performed looking at the odds of

developing EBV viraemia or LPD with respect to the following

factors: RIC versus CIC; presence or absence of PID; use of

Campath or ATG or neither (comparing ATG versus nothing,

Campath versus nothing and ATG versus Campath); HLA-

mismatched donor (yes/no); unrelated donor (yes/no); in vitro

TCD (yes/no); acute GvHD (aGvHD; yes/no); chronic GvHD

(cGvHD; yes/no); chimaerism at 6-weeks post-transplant

(mixed/full) and chimaerism at 12-week post-transplant

(mixed/full). Using chi-square or Fisher’s exact test, factors

that showed a statistically significant effect on the odds of EBV

infection or LPD at the 10% level, were considered for

inclusion in a bivariate logistic regression analysis that

included RIC versus CIC. Each factor was considered in turn

and the likelihood ratio test was used to determine if the factor

was statistically significant in the model. If it was statistically

significant at the 5% level, it was considered for inclusion in a

multivariate model.

Results

EBV viraemia and disease manifestations following HSCT

The incidence of EBV viraemia and EBV disease was greater

after RIC HSCT than CIC HSCT. Only six patients (8Æ8%) in

the CIC HSCT cohort developed viraemia. They all remained

asymptomatic and none progressed to EBV disease. Of the 65

patients receiving RIC HSCT, 23 (35%) had EBV viraemia at a

median time of 2 months following transplant (range: 4 d to

13 months). Their median age was 6Æ3 years (range: 0Æ46–17Æ7)

with a similar male/female ratio (11 male, 12 female). The

median time to neutrophil engraftment did not differ between

those who had EBV detected (10–31 d, median 14) and those

who did not (10–21 d, median 14).

Of the 23 patients in whom EBV viraemia was detected

following RIC, eight patients remained asymptomatic with low

maximal viral titres (median of maximal viral

load ¼ 10 000 copies/ml plasma). The maximal viral titres in

all RIC patients are shown in Fig 1A. Five patients developed

symptomatic viraemia (median of maximum viral

load ¼ 100 000 copies/ml plasma) and 10 patients developed

LPD (median of maximal viral load ¼ 400 000 copies/ml

plasma). Their times to development of EBV viraemia are

shown in Fig 1B.

Of the eight asymptomatic patients, five were treated by

antivirals (ganciclovir, n ¼ 4; foscarnet, n ¼ 1, with concur-

rent CMV infection) and reduction of immunosuppression.

Ribavirin/cidofovir were used in two patients with concurrent

adenoviral infection, and one of these in whom EBV viraemia

was still not controlled, was given two doses of rituximab,

administered on a weekly basis at a dose of 375 mg/m2.

Viraemia cleared in this patient after 10 days. Two asympto-

matic patients received no specific therapy for EBV. The

viraemia cleared with immune reconstitution at 3 and

Fig 1. (A) Maximal Epstein–Barr virus (EBV) viral load in reduced-

intensity conditioning (RIC) patients with asymptomatic and symp-

tomatic viraemia and lymphoproliferative disease (LPD). Maximal

viral load (log10 copies/ml plasma) of each patient is shown, showing

values alongside mean and 95% confidence interval for each group. (B)

Time post-stem cell transplantation (SCT) to development of EBV

viraemia by clinical group in RIC patients.

J. Cohen et al


6 months respectively. The final patient, who had no symp-

toms of EBV disease, died of overwhelming pulmonary

aspergillosis with EBV-PCR detected in plasma prior to

specific therapy being instituted.

One patient with symptomatic viraemia responded to

treatment with ganciclovir and reduction of immunosuppres-

sion. The remaining four patients with symptomatic viraemia

did not respond to this, and were treated with rituximab as

described above (one dose, n ¼ 4; four doses, n ¼ 1), which

led to resolution of symptoms after a median of 13 d (range:

8–20) and clearance of viraemia after a median of 11 d (range:

4–29).

Ten patients developed LPD according to the Pittsburgh

criteria (Nalesnik, 1998). In two of the 10 patients, this was

limited to localized disease (central nervous system, n ¼ 1;

dacroadenitis, n ¼ 1). Both of these patients survived. The

remaining eight patients all had disseminated lymphomatous

disease. Viraemia commenced within 2 months of transplan-

tation in nine of the 10 patients and within 3 months for

patient 10. Two of the 10 patients responded to a reduction of

immunosuppression and treatment with antivirals (ganciclovir

alone in one patient; ganciclovir with foscarnet in one patient

with concurrent CMV viraemia). The other eight patients with

LPD were treated with rituximab. Six of the eight responded

with clinical resolution and reduction in viral load after 1–

4 weekly doses (median time to clinical complete response

23 d, range: 11–77), administered as described. There was

complete clearance of viraemia (<100 copies/ml plasma) in

these patients after a median time of 21 d (range: 9–36). One of

these patients, who was successfully treated with rituximab for

LPD, died of overwhelming GvHD 7 months after transplant.

The remaining two patients had progressive LPD that did not

respond to rituximab (one dose and three doses respectively) or

donor lymphocyte infusion (in one patient), with pulmonary/

airway involvement progressing to multiorgan failure and

death. They had both received RIC with fludarabine/melpha-

lan/ATG to treat myelodysplastic syndromes.

Rituximab was largely well tolerated, but noted to cause a

tumour lysis syndrome in one patient who had very high EBV

load (1 000 000 copies/ml) and bulky disease prior to therapy.

Eleven of 13 patients cleared viraemia, the patients who did

not died of progressive LPD. The median time to complete

clinical response was 20 d (range: 8–77) and to clearing

viraemia was 13 d (range: 4–36).

Two of the eight patients with disseminated lymphomatous

disease had a biopsy as part of their diagnostic work-up, one of

thoracic lymph node tissue, one a liver biopsy. Both demon-

strated monomorphic large B cell, EBV+, lymphomatous

disease according to the WHO criteria (Harris et al, 1999).

Both of these patients subsequently died. The source of the

proliferative B-cell clone was investigated in the two patients

with localized LPD by fluorescent in situ hybridization for sex

chromosomes (XY-FISH) and shown to be of recipient origin

in both cases. Further details of the 10 patients with LPD are

summarized in Table II.

In the 23 patients who developed viraemia following RIC, an

absolute lymphocyte count (ALC) of below 0Æ3 · 109/l at the

time of onset of viraemia was strongly predictive of develop-

ment of LPD [odds ratio (OR) ¼ 12Æ8; P < 0Æ05]. Further

details are shown in Table III.

There was a high incidence of co-infections at the time of

EBV viraemia. About 17 of the 23 patients had other viruses

detected in the same samples by PCR (EBV + CMV,

n ¼ 8; EBV + adenovirus, n ¼ 4; EBV + CMV + adenovirus,

n ¼ 5).

Survival analysis

Of the 29 patients who developed viraemia (RIC, n ¼ 23; CIC,

n ¼ 6), five had died at the time of analysis (82Æ8% survival).

Of the 104 patients without viraemia (RIC, n ¼ 42; CIC,

n ¼ 62), 33 had died (68% survival). Of the 10 patients within

the viraemic group who developed LPD, three died. When

analysed using a Cox proportional hazards model, there was

no statistically significant difference in overall survival between

those who developed viraemia post-transplant and those who

did not. Similarly there was no significant difference between

those who developed LPD, those with viraemia who did not

develop LPD and those who did not develop viraemia. Results

of this analysis are shown in Fig 2.

Risk factors associated with EBV detection or LPD

Univariate analysis. The incidences of EBV viraemia and LPD

in various groups and the unadjusted ORs for risk factors are

presented with 95% confidence intervals and chi-squared

statistics in Tables IV and V respectively, and are discussed in

further detail below.

Conditioning: RIC versus CICThe RIC was the risk factor

most strongly associated with EBV viraemia. Twenty-three of

65 patients (35Æ4%) with RIC developed EBV viraemia versus

six of 68 (8Æ8%) of those conventionally conditioned, with an

OR of 5Æ66 (P < 0Æ001). All of the patients that developed LPD

had received RIC (10 of 65, 15Æ4%, vs. zero of 68, 0%;

P < 0Æ001).

In vitro TCD. In vitro TCD using Campath 1M or CD34-

selection was associated with a lower incidence of EBV

viraemia (two of 33, 6Æ1% for TCD versus 27 of 100, 27Æ0%

for others; OR 0Æ17; P < 0Æ012) and none of these patients

developed LPD. Two of 15 patients whose grafts were prepared

by CD34 selection using the Miltenyi system developed

viraemia. Of the 18 patients whose grafts were treated with

ex vivo Campath 1M and T-cell addback, none developed

viraemia.

Serotherapy: Campath versus ATG Use of serotherapy was

associated with a higher incidence of EBV viraemia than in

those patients in whom serotherapy was not given (27 of 116,

23Æ3% vs. two of 17, 11Æ8%; not significant, NS). Patients

whose grafts were not TCD in vitro and did not receive in vivo

serotherapy had a lower incidence of EBV viraemia than those



Table

II.

Det

ails

of

10p

atie

nts

wh

od

evel

op

edly

mp

ho

pro

life

rati

ved

isea

sefo

llo

win

gre

du

ced

-in

ten

sity

con

dit

ion

edH

SCT

.

Dis

ease

Do

no

rC

on

dit

ion

ing

Tim

eto

EB

V

(mo

nth

s)

Du

rati

on

(mo

nth

s)R

x

Vir

allo

adat

star

t

of

trea

tmen

t

(co

pie

s/m

lp

lasm

a)aG

vHD

Org

an

invo

lvem

ent

Cli

nic

al

clas

sifi

cati

on

Bio

psy

Ou

tco

me

JMM

LM

SDF

lu/M

el/A

TG

21

Rit

uxi

mab

(fo

ur

do

ses)

DL

I

1000

0–10

000

0N

oG

ener

aliz

edo

rgan

infi

ltra

tio

n

and

lym

ph

no

des

Dis

sem

inat

ed

lym

ph

om

ato

us

Th

ora

cic

lym

ph

no

de:

larg

e

EB

V+

B-c

ell

NH

L

Die

dL

PD

MD

SM

MU

D

1CW

Flu

/Mel

ph

/AT

G2

1R

itu

xim

ab

(on

ed

ose

)

100

000–

100

000

0Y

esG

ener

aliz

edo

rgan

infi

ltra

tio

n

and

lym

ph

no

des

Dis

sem

inat

ed

lym

ph

om

ato

us

Liv

er:

larg

e

EB

V+

B-c

ell

NH

L

Die

dL

PD

CID

MM

UD

1A

Flu

/Mel

ph

/AT

G1

2R

itu

xim

ab

(fo

ur

do

ses)

1000

000

0–10

000

000

0Y

esL

iver

and

gen

eral

ized

lym

ph

no

des

Dis

sem

inat

ed

lym

ph

om

ato

us

ND

Die

do

fcG

vHD

7-m

on

ths

po

st-t

ran

spla

nt

Par

tial

LA

DM

UD

Flu

/Mel

ph

/Cam

pat

h1

2G

anci

clo

vir

Fo

scar

net

Val

acic

lovi

r

Rit

uxi

mab

(fo

ur

do

ses)

1000

0–10

000

0N

oC

entr

aln

ervo

us

syst

eman

dly

mp

h

no

des

EB

V-t

ran

sfo

rmed

reci

pie

nt

B-c

ell

clo

ne

iden

tifi

edin

cere

bro

spin

alfl

uid

Lo

cali

zed

ND

Surv

ived

Mar

atea

u-L

amy

MU

D

(2N

D)

Flu

/Mel

ph

/Cam

pat

h1

1R

itu

xim

ab

(tw

od

ose

s)

Dev

elo

ped

tum

ou

rly

sis

100

000

0–10

000

000

Yes

Lu

ngs

and

lym

ph

no

des

Dis

sem

inat

ed

lym

ph

om

ato

us

ND

Surv

ived

SCID

MF

DF

lu/M

elp

h/C

amp

ath

36

Rit

uxi

mab

(th

ree

do

ses)

100

000

0–10

000

000

Yes

Lu

ngs

and

lym

ph

no

des

Dis

sem

inat

ed

lym

ph

om

ato

us

ND

Surv

ived

CID

+ch

ron

icC

MV

MM

UD

1CW

Flu

/Mel

ph

/AT

G2

8G

anci

clo

vir

Fo

scar

net

No

ritu

xim

ab

100

000

0–10

000

000

Yes

Liv

er,

sple

enan

d

lym

ph

no

des

Dis

sem

inat

ed

lym

ph

om

ato

us

ND

Surv

ived

CID

/AIH

AM

MU

D

1DQ

Flu

/Mel

ph

/AT

G2

3F

osc

arn

et

Rit

uxi

mab

(on

ed

ose

)

100

000–

100

000

0Y

esE

yes

–ac

ute

dac

ryo

aden

itis

Lo

cali

zed

ND

Surv

ived

CV

ID

EB

V+

Ho

dgk

ins

MF

DF

lu/M

elp

h/C

amp

ath

11

Rit

uxi

mab

(th

ree

do

ses)

100

000–

100

000

0Y

esC

ervi

cal

lym

ph

no

des

,

live

r,fe

ver

and

rash

Dis

sem

inat

ed

lym

ph

om

ato

us

ND

Surv

ived

WA

SM

UD

Flu

/Mel

ph

/Cam

pat

h2

8G

anci

clo

vir

No

ritu

xim

ab

100

000

0–10

000

000

Yes

Cer

vica

lly

mp

hn

od

es,

live

r,fe

ver

and

rash

Dis

sem

inat

ed

lym

ph

om

ato

us

ND

Surv

ived

JMM

L,

juve

nil

em

yelo

mo

no

cyti

cle

uka

emia

;M

DS,

mye

lod

ysp

last

icsy

nd

rom

e;C

ID,

com

bin

edim

mu

no

defi

cien

cy;

LA

D,

leu

cocy

tead

hes

ion

defi

cien

cy;

SCID

,se

vere

com

bin

edim

mu

no

defi

cien

cy;

CM

V,

cyto

meg

alo

viru

s;A

IHA

,au

toim

mu

ne

hae

mo

lyti

can

aem

ia;

CV

ID,

com

mo

nva

riab

leim

mu

no

defi

cien

cy;

WA

S,W

isko

tt-A

ldri

chsy

nd

rom

e;N

HL

,n

on

-Ho

dgk

in’s

lym

ph

om

a;M

MU

D,

HL

A-m

ism

atch

ed

un

rela

ted

do

no

r;M

UD

,H

LA

-mat

ched

un

rela

ted

do

no

r;M

FD

,H

LA

-mat

ched

fam

ily

do

no

r;M

SD,

HL

A-m

atch

edsi

bli

ng

do

no

r;D

LI,

do

no

rly

mp

ho

cyte

infu

sio

n;

ND

,n

ot

do

ne;

EB

V,

Ep

stei

n–

Bar

rvi

rus;

LP

D,

lym

ph

op

roli

fera

tive

dis

ease

;H

SCT

,h

aem

ato

po

ieti

cst

emce

lltr

ansp

lan

tati

on

;H

LA

,h

um

anle

uco

cyte

anti

gen

.

J. Cohen et al


who received grafts without in vitro TCD but with in vivo

serotherapy (two of 17, 11Æ8% vs. 25 of 83, 30Æ1%; NS). In

those patients who did receive serotherapy, use of ATG was

strongly related to the development of EBV viraemia. Fifteen of

43 patients (34Æ9%) who received ATG developed viraemia,

compared with 12 of 73 (16Æ4%) who received Campath and

two of 17 patients (11Æ8%) who received neither ATG nor

Campath. When comparing ATG with Campath, the OR for

development of viraemia was 2Æ72 (P < 0Æ05). Patients who

received ATG as part of their RIC had the highest incidence of

viraemia (13 of 30, 43Æ3%). Patients that had RIC with

Campath were three times as likely to have EBV viraemia than

those who received CIC with Campath (nine of 34, 26Æ5% vs.

three of 39, 7Æ69%; P < 0Æ05).

LPD was only seen amongst patients who received sero-

therapy (ATG, five of 43; Campath, five of 73; neither zero of

17; NS).

Primary immunodeficiency. Patients that underwent trans-

plantation for PID were twice as likely to have EBV viraemia

following transplantation than those transplanted for other

diseases (16 of 50, 32Æ0% vs. 13 of 83, 15Æ7%; OR 2Æ53;

P < 0Æ05). Of the 16 PID patients with viraemia, six developed

LPD. Forty-one of 50 patients with PID received RIC. The

higher incidence of EBV viraemia in PID patients was

restricted to those who had received RIC. In patients who

received CIC, no significant difference was found in the

incidence of EBV viraemia in the PID and non-PID groups.

Acute GvHD. Sixteen of 51 (31Æ4%) patients who developed

aGvHD by day 100 became viremic compared with 13 of 82

(15Æ9%) without aGvHD (OR 2Æ42; P < 0Æ05). Acute GvHD

was also strongly associated with LPD (eight of 50, 16% vs. two

of 83, 9Æ64%; OR 7Æ71; P < 0Æ01).

Other variables. In this series, none of the following factors

was associated with a statistically significant increase in

incidence of EBV viraemia: use of unrelated donor or HLA-

mismatched graft, cGvHD, presence of mixed chimaerism at 6-

or 12-weeks following transplant. None of these factors was

associated with a statistically significant increase in incidence

of EBV-LPD: use of serotherapy, PID, use of unrelated donor

or HLA-mismatched graft, cGvHD, presence of mixed chima-

erism at 6- or 12-weeks following transplant. When RIC

transplants were looked at alone, mixed chimaerism at 6- or

12-weeks following transplant was again not associated with a

statistically significant increase in either of EBV viraemia or

development of LPD.

Bivariate and multivariate analysis

Bivariate analysis was performed as described above, combi-

ning RIC with each of the other significant risk factors in the

univariate model, investigating the effect of each factor on the

OR of EBV infection for RIC versus CIC. The results of this are

shown in Table VI, and indicate that all factors were non-

significant at the 5% level. Models were also developed that

included other combinations of the above factors. All factors

remained non-significant in these models.

Discussion

This study represents the largest examination of the impact of

RIC on the development of post-transplant EBV-related

disease. In our cohort of paediatric patients, RIC was

associated with a significantly higher incidence of both EBV

viraemia and LPD post-SCT compared with CIC. While the

Fig 2. Overall survival of patients following bone marrow transplan-

tation (BMT), according to clinical group. Kaplan–Meier curves gen-

erated using Cox proportional hazards model are shown. (A) Patients

with Epstein–Barr virus (EBV) viraemia post-BMT versus patients

without EBV viraemia; (B) patients with EBV-lymphoproliferative

(LPD) disease post-BMT versus patients with EBV viraemia without

LPD versus patients without EBV viraemia.

Table III. Incidence of LPD in patients with EBV viraemia following

RIC according to ALC at time of onset of viraemia.

ALC (·109/l) Incidence of LPD

<0Æ3 7/9 (77Æ8)

>0Æ3 3/14 (21Æ4)*

*P ¼ 0Æ013; odds ratio: 12Æ8 (95% confidence interval: 1Æ69–97Æ2).

Values within parentheses represent percentage.

EBV, Epstein–Barr virus; LPD, lymphoproliferative disease; ALC,

absolute lymphocyte count .



higher incidence of PID and ATG usage, both of which are

established risk factors for LPD, may partially explain this, our

univariate analysis suggests that RIC was the risk factor most

strongly associated with EBV viraemia and LPD and bivariate

analysis suggests an effect of RIC independent of these factors.

The use of in vivo ATG as part of RIC significantly increased

the likelihood of EBV viraemia, with an incidence in this

subgroup of 43%. Both patients who died of LPD had received

RIC incorporating ATG. Whilst EBV viraemia and LPD were

frequent following the use of Campath in the RIC setting,

viraemia was significantly less common than when ATG was

used. Acute GvHD was associated with EBV viraemia and LPD

in univariate analysis, but was not a risk factor independent of

RIC.

In vitro TCD of the graft has been previously shown to

significantly increase the risk of LPD (Antin et al, 1991;

Micallef et al, 1998). In this study, in vitro TCD was associated

with less EBV viraemia than in those cases where the graft was

not TCD in vitro. This is likely to reflect the fact that, in the

majority of the latter group, patients received in vivo

Table IV. Univariate analysis of risk factors for EBV viraemia following allogeneic HSCT, showing incidence of viraemia by risk factor group,

unadjusted OR with 95% CI, chi-squared (v2) and P-value for significance of difference obtained from chi-squared test.

Group Incidence Group Incidence OR 95% CI v2 P-value

RIC 23/65 (35Æ4) CIC 6/68 (8Æ82) 5Æ66 2Æ00–15Æ99 13Æ65 <0Æ001Serotherapy 27/116 (23Æ3) No serotherapy 2/17 (11Æ8) 2Æ27 0Æ49–10Æ58 1Æ15 0Æ283

ATG 15/43 (34Æ9) No serotherapy 2/17 (11Æ8) 4Æ02 0Æ76–21Æ17 3Æ15 0Æ076

Campath 12/73 (16Æ4) No serotherapy 2/17 (11Æ8) 1Æ48 0Æ29–7Æ39 0Æ23 0Æ630

ATG 15/43 (34Æ9) Campath 12/73 (16Æ4) 2Æ72 1Æ10–6Æ73 5Æ11 0Æ024TCD 2/33 (6Æ10) NotTCD 27/100 (27Æ0) The 0Æ04–0Æ78 6Æ38 0Æ012PID 16/50 (32) NotPID 13/83 (15Æ7) 2Æ53 1Æ07–5Æ97 4Æ85 0Æ027aGvHD 16/50 (32Æ0) None 13/83 (15Æ7) 2Æ53 1Æ07–5Æ97 4Æ85 0Æ027cGvHD 3/11 (27Æ3) None 26/122 (21Æ3) 1Æ38 0Æ34–5Æ63 0Æ21 0Æ647

Unrelated 17/70 (24Æ3) Related 12/63 (19Æ0) 1Æ36 0Æ59–3Æ15 0Æ53 0Æ465

HLA-mismatch 12/42 (28Æ6) HLA-match 17/91 (18Æ7) 1Æ74 0Æ74–4Æ12 1Æ64 0Æ199

6-week MC 6/25 (24Æ0) 6-week FC 15/76 (19Æ7) 1Æ28 0Æ43–3Æ80 0Æ21 0Æ649


RIC, reduced-intensity conditioning; CIC, conventional-intensity conditioning; ATG, in vivo antithymocyte globulin; Campath, in vivo Campath;

TCD, ex vivo T-cell depleted; NotTCD, not ex vivo T-cell depleted; PID, primary immunodeficiency; NotPID, not primary immunodeficiency;

aGvHD, acute graft-versus-host disease; cGvHD, chronic graft-versus-host disease; MC, mixed chimaeras; FC, full chimaeras; EBV, Epstein–Barr

virus; HSCT, haematopoietic stem cell transplantation; HLA, human leucocyte antigen; OR, odds ratio; CI, confidence interval.


Table V. Univariate analysis of risk factors for EBV-LPD following allogeneic HSCT, showing incidence of LPD by risk factor group, unadjusted OR

with 95% CI, chi-squared (v2) and P-value for significance of difference obtained from chi-squared test.

Group Incidence Group Incidence OR 95% CI v2 P-value

RIC 10/65 (15Æ4) CIC 0/68 (0) – – 11Æ3 0Æ001Serotherapy 10/116 (8Æ62) No serotherapy 0/17 (0) – – 1Æ59 0Æ208

ATG 5/43 (11Æ6) No serotherapy 0/17 (0) – – 2Æ16 0Æ142

Campath 5/73 (6Æ8) No serotherapy 0/17 (0) – – 1Æ23 0Æ267

ATG 5/43 (11Æ6) Campath 5/73 (6Æ8) 0Æ56 0Æ15–2Æ05 0Æ78 0Æ376

TCD 0/33 (0) NotTCD 10/100 (10) – – 3Æ57 0Æ059

PID 6/50 (12) NotPID 4/83 (4Æ82) 2Æ69 0Æ72–10Æ1 2Æ31 0Æ128

aGvHD 8/50 (16Æ0) None 2/83 (9Æ64) 7Æ714 1Æ57–38Æ0 8Æ29 0Æ004cGvHD 0/11 (0) None 10/122 (8Æ2) – – 0Æ975 0Æ323

Unrelated 7/70 (10) Related 3/63 (4Æ76) 2Æ22 0Æ55–8Æ99 1Æ31 0Æ253

HLA-mismatch 4/42 (9Æ52) HLA-match 6/91 (6Æ59) 1Æ49 0Æ39–5Æ59 0Æ355 0Æ551



RIC, reduced-intensity conditioning; CIC, conventional-intensity conditioning; ATG, in vivo antithymocyte globulin; Campath, in vivo Campath;

TCD, ex vivo T-cell depleted; NotTCD, not ex vivo T-cell depleted; PID, primary immunodeficiency; NotPID, not primary immunodeficiency;

aGvHD, acute graft-versus-host disease; cGvHD, chronic graft-versus-host disease; MC, mixed chimaeras; FC, full chimaeras; EBV, Epstein–Barr

virus; HSCT, haematopoietic stem cell transplantation; HLA, human leucocyte antigen; OR, odds ratio; CI, confidence interval.


J. Cohen et al


serotherapy (83 of 101 patients), because of the high

proportion of patients receiving grafts from unrelated donors.

Additionally, in our cohort, the methods of in vitro TCD were

not T-cell specific and would deplete B cells as well. The

selective depletion of T cells without depletion of B cells

appears critical for the pathogenesis of PTLD. In the CIC

setting, depletion of both T and B cells from the graft has been

associated with a very low incidence of LPD (Hale &

Waldmann, 1998). Our data supports this finding: only two

of 15 patients transplanted with grafts that were CD34-selected

using the Miltenyi system developed EBV viraemia. Similarly,

where TCD was performed with Campath 1M with subsequent

T-cell addback, again no patient developed EBV viraemia. The

use of ATG in vivo was significantly associated with higher EBV

viraemia than with Campath 1H.

The incidence of LPD in our RIC patients who received

Campath 1H in vivo was higher than that seen in previous

cohorts of patients conditioned with CIC regimes incorporat-

ing this antibody (Hale & Waldmann, 1998). Peggs et al (2003)

reported four cases of LPD in a cohort of 100 adult patients

allografted using a RIC protocol similar to ours, two of which

were associated with the use of ATG for treatment of GvHD.

The incidence of EBV viraemia was not studied, but the lower

incidence of LPD observed in this study compared with our

cohort may reflect the higher proportion of transplants from

alternative donors in our study, an increased recognition of

atypical clinical presentations of LPD with prospective mon-

itoring of viraemia or differences between paediatric and adult

populations. Whether our finding of increased EBV viraemia

and LPD after RIC transplantation is specific to non-myelo-

ablative protocols utilizing serotherapy is unclear and this

needs to be addressed in further studies. The increased

incidence of EBV-associated disease in the RIC setting

observed in our study may relate to the persistence of EBV

in residual recipient B cells after non-myeloablative condi-

tioning. Indeed, in the two patients who were investigated for

the origin of the proliferative B-cell clones, both were

identified by FISH as recipient-derived. This suggests that it

is the relative preservation of recipient B cells in a profoundly

immunosuppressed setting which increases the chances of

post-transplant EBV proliferation.

Reduced Intensity Conditioning produces a state of profound

functional immunosuppression for at least 6–9 months follow-

ing transplant (Amrolia et al, 2000; Chakrabarti et al, 2002).

While recent work has demonstrated that, following an

unrelated donor transplant for PID, T- and B-cell reconstitution

occurs with similar kinetics in the RIC and CIC settings (Rao

et al, 2005), the number of circulating CD4+ and CD8+ T cells at

the time of the majority of EBV reactivations (2–3 months post-

SCT) with either conditioning regimen was profoundly

depressed. We have demonstrated that an ALC < 0Æ3 · 109/l

at the time of onset of viraemia was strongly predictive of

development of LPD, highlighting the need to take into account

immune reconstitution as well as viral load in estimating the risk

of development of LPD. We have previously observed a higher

incidence of viral reactivations (EBV, CMV and adenovirus) in

PID patients after RIC than CIC (Rao et al, 2005) and the

coincident appearance of other viruses around the same time as

EBV observed in this study is highly suggestive of a reduction in

early viral-specific immunity with RIC. Reduced numbers of

EBV-specific cytotoxic T-lymphocytes have been shown to be

associated with the development of EBV post-SCT (Meij et al,

2003). Indeed, in vitro studies of EBV-specific immune

responses using enzyme-linked immunospot assays have shown

markedly delayed recovery of EBV-specific immunity with a

fludarabine/melphalan/Campath regime similar to ours, when

compared with a CIC regime (Chakrabarti et al, 2003). We are

now prospectively studying EBV-specific T-cell immune recon-

stitution, comparing RIC and CIC in a paediatric setting.

The significance of EBV viraemia post-SCT should be

interpreted in the context of both the type of transplant and

the extent of immune reconstitution. A number of groups

have demonstrated that high viral loads may be predictive of

development of LPD following TCD myeloablative trans-

plants (Rooney et al, 1995; van Esser et al, 2001a; Sirvent-

von Bueltzingsloewen et al, 2002; Wagner et al, 2004), but in

recipients of unmanipulated grafts this is less predictive.

Even in the former group of patients, while EBV viraemia

has a high negative predictive value (NPV), the positive

predictive value (PPV) for developing LPD is generally only

25–40%. Given the efficacy of pre-emptive rituximab in

preventing LPD in patients with viral loads (van Esser et al,

2002), it is now ethically difficult to perform similar studies

to determine the true predictive value of high EBV loads in

patients undergoing transplant with RIC. In our study, the

majority of patients with asymptomatic or symptomatic

viraemia were treated with modulation of immunosuppres-

sion ± pre-emptive rituximab. It is likely that without such

interventions our incidence of LPD may have been even

higher. Nonetheless, the fact that 10 of 23 viraemic patients

in our RIC cohort went on to develop LPD (PPV: 43%),

demonstrates that EBV viraemia is frequently clinically

significant after RIC.

Table VI. Bivariate analysis of risk factors for EBV viraemia following

allogeneic HSCT, showing adjusted OR for risk factor with 95% CI,

chi-squared (v2) and P-value for significance of difference obtained

from chi-squared test, after accounting for the use of RIC.

Factor adjusted for Adjusted OR 95% CI v2 P-value

PID (versus NotPID) 1Æ19 0Æ45–3Æ12 0Æ12 0Æ730

ATG (versus Campath) 2Æ09 0Æ83–5Æ29 2Æ49 0Æ288

TCD (versus NotTCD) 0Æ40 0Æ08–2Æ01 1Æ42 0Æ234

aGvHD (versus none) 2Æ20 2Æ12–15Æ08 3Æ07 0Æ080

PID, primary immunodeficiency; NotPID, not primary immunodefi-

ciency; ATG, in vivo antithymocyte globulin; TCD, ex vivo T-cell

depleted; NotTCD, not ex vivo T-cell depleted; aGvHD, acute graft-

versus-host disease; EBV, Epstein–Barr virus; HSCT, haematopoietic

stem cell transplantation; RIC, reduced-intensity conditioning; OR,

odds ratio, CI, confidence interval.



Surveillance studies published subsequent to the start of this

study have suggested cut-off values for EBV viral loads, above

which anti-EBV therapeutic manoeuvres are suggested to pre-

empt and thereby prevent the progression to LPD (van Esser

et al, 2001b, 2002; Hoshino et al, 2001) Such studies used

differing laboratory methodologies to quantify the viral load

on differing patient samples, and hence suggested different

thresholds, but the available evidence suggests this approach

may be effective in preventing LPD (Lankester et al, 2002).

Because of the semi-quantitative methodology used and the

fact that the threshold for initiation of anti-EBV therapies

varied between the patients reported in this study, we are

unable to state whether or not such threshold values are

supported by our data. In eight of the 10 patients who

developed LPD in this study, viral load exceeded 105 copies/ml

at the start of treatment. Additionally, even without an

absolute threshold value for initiation of therapy, the rate of

rise in viral load may be significant. There is clearly a need for

large, prospective studies to determine appropriate thresholds

for, and the efficacy of, pre-emptive reduction in immuno-

suppression/rituximab in patients undergoing TCD SCT who

develop high-level EBV viraemia. Such studies should incor-

porate standardized thresholds and utilize accurate quantifi-

cation of viral load with real-time PCR.

The risk of progression to LPD appears to depend on the

balance between EBV reactivation and recovery of EBV-

specific immunity: indeed, combining tetramer assays of

recovery of circulating EBV-specific CD8+ T cells with high-

level viraemia increases the PPV markedly (Meij et al, 2003).

However, such assays are not widely available and what is

needed is a simple method for identifying those viraemic

patients at high risk of LPD. In our patient cohort, EBV

viraemia in patients with a lymphocyte count of <0Æ3 · 109/l

was strongly predictive of LPD.

The majority (six of eight) of patients with established LPD

went into durable complete response after rituximab, demon-

strating that this is an effective therapy in the SCT setting,

avoiding the need to use myelosuppressive chemotherapy.

Indeed, in our cohort, survival in patients with either viraemia

or LPD was equivalent to survival in those with no viraemia

[the statistically non-significant increased survival in patients

with viraemia may reflect the lower treatment-related mortal-

ity in the RIC group (Rao et al, 2005)]. The response rate to

rituximab is similar to that previously reported (Faye et al,

2001) and this is sufficient therapy in the majority of cases.

However, our experience suggests that in some patients with

extremely aggressive LPD, particularly those with airway/

pulmonary involvement, chemotherapy may be indicated for

more rapid disease control.

In conclusion, in our cohort of paediatric patients, RIC

appears to be a significant risk factor for developing EBV-LPD

post-transplant, particularly when ATG is incorporated into

conditioning. Further studies are needed to determine if this is

also true for other patient groups, such as adults with

haematological malignancies. This group of patients needs

careful observation to detect early EBV viraemia through PCR

monitoring of the peripheral blood, together with a high

clinical index of suspicion. The recent introduction of real-

time PCR for monitoring viraemia (Jabs et al, 2001; Wagner

et al, 2004), combined with studies of reconstitution of EBV-

specific immunity and pre-emptive treatment with rituximab,

should lead to the development of more rational strategies to

prevent and treat LPD post-SCT.

References

Amrolia, P., Gaspar, H.B., Hassan, A., Webb, D., Jones, A., Sturt, N.,

Mieli-Vergani, G., Pagliuca, A., Mufti, G., Hadzic, N., Davies, G. &

Veys, P. (2000) Nonmyeloablative stem cell transplantation for

congenital immunodeficiencies. Blood, 96, 1239–1246.

Antin, J.H., Bierer, B.E., Smith, B.R., Ferrara, J., Guinan, E.C., Sieff, C.,

Golan, D.E., Macklis, R.M., Tarbell, N.J. & Lynch, E. (1991) Selective

depletion of bone marrow T lymphocytes with anti-CD5 mono-

clonal antibodies: effective prophylaxis for graft-versus-host disease

in patients with hematologic malignancies. Blood, 78, 2139–2149.

Bhatia, S., Ramsay, N.K., Steinbuch, M., Dusenbery, K.E., Shapiro,

R.S., Weisdorf, D.J., Robison, L.L., Miller, J.S. & Neglia, J.P. (1996)

Malignant neoplasms following bone marrow transplantation.

Blood, 87, 3633–3639.

Chakrabarti, S., Mackinnon, S., Chopra, R., Kottaridis, P.D., Peggs, K.,

O’Gorman, P., Chakraverty, R., Marshall, T., Osman, H., Mahendra,

P., Craddock, C., Waldmann, H., Hale, G., Fegan, C.D., Yong, K.,

Goldstone, A.H., Linch, D.C. & Milligan, D.W. (2002) High

incidence of cytomegalovirus infection after nonmyeloablative stem

cell transplantation: potential role of Campath-1H in delaying im-

mune reconstitution. Blood, 99, 4357–4363.

Chakrabarti, S., Milligan, D.W., Pillay, D., Mackinnon, S., Holder, K.,

Kaur, N., McDonald, D., Fegan, C.D., Waldmann, H., Hale, G.,

Rickinson, A. & Steven, N. (2003) Reconstitution of the Epstein-

Barr virus-specific cytotoxic T-lymphocyte response following T-

cell-depleted myeloablative and nonmyeloablative allogeneic stem

cell transplantation. Blood, 102, 839–842.

Curtis, R.E., Travis, L.B., Rowlings, P.A., Socie, G., Kingma, D.W.,

Banks, P.M., Jaffe, E.S., Sale, G.E., Horowitz, M.M., Witherspoon,

R.P., Shriner, D.A., Weisdorf, D.J., Kolb, H.J., Sullivan, K.M.,

Sobocinski, K.A., Gale, R.P., Hoover, R.N., Fraumeni, J.F. Jr & Deeg,

H.J. (1999) Risk of lymphoproliferative disorders after bone marrow

transplantation: a multi-institutional study. Blood, 94, 2208–2216.

van Esser, J.W., van der Holtm, B., Meijer, E., Niesters, H.G.,

Trenschel, R., Thijsen, S.F., van Loon, A.M., Frassoni, F., Baciga-

lupo, A., Schaefer, U.W., Osterhaus, A.D., Gratama, J.W.,

Lowenberg, B., Verdonck, L.F. & Cornelissen, J.J. (2001a) Epstein-

Barr virus (EBV) reactivation is a frequent event after allogeneic

stem cell transplantation (SCT) and quantitatively predicts EBV-

lymphoproliferative disease following T-cell-depleted SCT. Blood,

98, 972–978.

van Esser, J.W., Niesters, H.G., Thijsen, S.F., Meijer, E., Osterhaus,

A.D., Wolthers, K.C., Boucher, C.A., Gratama, J.W., Budel, L.M.,

van der Holt, B., van Loon, A.M., Lowenberg, B., Verdonck, L.F. &

Cornelissen, J.J. (2001b) Molecular quantification of viral load in

plasma allows for fast and accurate prediction of response to therapy

of Epstein-Barr virus-associated lymphoproliferative disease after

allogeneic stem cell transplantation. British Journal of Haematology,

113, 814–821.

J. Cohen et al


van Esser, J.W.J., Niesters, H.G.M., van der Holt, B., Meijer, E., Os-

terhaus, A.D.M.E., Gratama, J.W., Verdonck, L.F., Lowenberg, B. &

Cornelissen, J.J. (2002) Prevention of Epstein-Barr virus-lympho-

proliferative disease by molecular monitoring and preemptive ri-

tuximab in high-risk patients after allogeneic stem cell

transplantation. Blood, 99, 4364–4369.

Faye, A., Quartier, P., Reguerre, Y., Lutz, P., Carret, A.S., Dehee, A.,

Rohrlich, P., Peuchmaur, M., Matthieu-Boue, A., Fischer, A. &

Vilmer, E. (2001) Chimaeric anti-CD20 monoclonal antibody (ri-

tuximab) in post-transplant B-lymphoproliferative disorder follow-

ing stem cell transplantation in children. British Journal of

Haematology, 115, 112–118.

Gross, T.G., Steinbuch, M., DeFor, T., Shapiro, R.S., McGlave, P.,

Ramsay, N.K., Wagner, J.E. & Filipovich, A.H. (1999) B cell lym-

phoproliferative disorders following hematopoietic stem cell trans-

plantation: risk factors, treatment and outcome. Bone Marrow

Transplantation, 23, 251–258.

Hale, G. & Waldmann, H. (1998) Risks of developing Epstein-Barr

virus-related lymphoproliferative disorders after T-cell-depleted

marrow transplants. CAMPATH users. Blood, 91, 3079–3083.

Harris, N.L., Jaffe, E.S., Diebold, J., Flandrin, G., Muller-Hermelink,

H.K., Vardiman, J., Lister, T.A. & Bloomfield, C.D. (1999) World

Health Organization classification of neoplastic diseases of the he-

matopoietic and lymphoid tissues: Report of the Clinical Advisory

Committee Meeting – Airlie House, Virginia, November 1997.

Journal of Clinical Oncology, 17, 3835–3849.

Ho, A.Y., Adams, S., Shaikh, H., Pagliuca, A., Devereux, S. & Mufti,

G.J. (2002) Fatal donor-derived Epstein-Barr virus-associated post-

transplant lymphoproliferative disorder following reduced intensity

volunteer-unrelated bone marrow transplant for myelodysplastic

syndrome. Bone Marrow Transplantation, 29, 867–869.

Hoshino, Y., Kimura, H., Tanaka, N., Tsuge, I., Kudo, K., Horibe, K.,

Kato, K., Matsuyama, T., Kikuta, A., Kojima, S. & Morishima, T.

(2001) Prospective monitoring of the Epstein-Barr virus DNA by a

real-time quantitative polymerase chain reaction after allogenic stem

cell transplantation. British Journal of Haematology, 115, 105–111.

Jabs, W.J., Hennig, H., Kittel, M., Pethig, K., Smets, F., Bucsky, P.,

Kirchner, H. & Wagner, H.J. (2001) Normalized quantification by

real-time PCR of Epstein-Barr virus load in patients at risk for

posttransplant lymphoproliferative disorders. Journal of Clinical

Microbiology, 39, 564–569.

Lankester, A.C., van Tol, M.J., Vossen, J.M., Kroes, A.C. & Claas, E.

(2002) Epstein-Barr virus (EBV)-DNA quantification in paediatric

allogenic stem cell recipients: prediction of EBV-associated lym-

phoproliferative disease. Blood, 99, 2630–2631.

Lucas, K.G., Burton, R.L., Zimmerman, S.E., Wang, J.H., Cornetta, K.G.,

Robertson, K.A., Lee, C.H. & Emanuel, D.J. (1998) Semiquantitative

Epstein-Barr virus (EBV) polymerase chain reaction for the

determination of patients at risk for EBV-induced lymphoprolifera-

tive disease after stem cell transplantation. Blood, 91, 3654–3661.

Meij, P., van Esser, J.W., Niesters, H.G., van Baarle, D., Miedema, F.,

Blake, N., Rickinson, A.B., Leiner, I., Pamer, E., Lowenberg, B.,

Cornelissen, J.J. & Gratama, J.W. (2003) Impaired recovery of Ep-

stein-Barr virus (EBV)-specific CD8+ T lymphocytes after partially

T-depleted allogeneic stem cell transplantation may identify patients

at very high risk for progressive EBV reactivation and lymphopro-

liferative disease. Blood, 101, 4290–4297.

Meijer, E., Slaper-Cortenbach, I.C., Thijsen, S.F., Dekker, A.W. &

Verdonck, L.F. (2002) Increased incidence of EBV-associated

lymphoproliferative disorders after allogeneic stem cell transplan-

tation from matched unrelated donors due to a change of T cell

depletion technique. Bone Marrow Transplantation, 29, 335–339.

Micallef, I.N.M., Chhanabhai, M., Gascoyne, R.D., Shepherd, J.D.,

Fung, H.C., Nantel, S.H., Toze, C.L., Klingemann, H.G., Sutherland,

H.J., Hogge, D.E., Nevill, T.J., Le, A. & Barnett, M.J. (1998) Lym-

phoproliferative disorders following allogeneic bone marrow trans-

plantation: the Vancouver experience. Bone Marrow

Transplantation, 22, 981–987.

Milpied, N., Coste-Burel, M., Accard, F., Moreau, A., Moreau, P., Ga-

rand, R. & Harousseau, J.L. (1999) Epstein-Barr virus-associated B cell

lymphoproliferative disease after non-myeloablative allogeneic stem

cell transplantation. Bone Marrow Transplantation, 23, 629–630.

Milpied, N., Vasseur, B., Parquet, N., Garnier, J.L., Antoine, C.,

Quartier, P., Carret, A.S., Bouscary, D., Faye, A., Bourbigot, B.,

Reguerre, Y., Stoppa, A.M., Bourquard, P., de Ligny, B.H., Dubief,

F., Mathieu-Boue, A. & Leblond, V. (2000) Humanized anti-CD20

monoclonal antibody (Rituximab) in post transplant B-lympho-

proliferative disorder: a retrospective analysis on 32 patients. Annals

of Oncology, 11, 113–116.

Nalesnik, M.A. (1998) Clinical and pathological features of post-

transplant lymphoproliferative disorders (PTLD). Springer Seminars

in Immunopathology, 20, 325–342.

Peggs, K.S., Banerjee, L., Thomson, K. & Mackinnon, S. (2003) Post

transplant lymphoproliferative disorders following reduced intensity

conditioning with in vivo T cell depletion. Bone Marrow Trans-

plantation, 31, 725–726.

Rao, K., Amrolia, P.J., Jones, A., Cale, C.M., Naik, P., King, D., Davies,

G.E., Gaspar, H.B. & Veys, P.A. (2005) Improved survival after

unrelated donor bone marrow transplant in children with primary

immunodeficiency using a reduced intensity conditioning regimen.

Blood, 105, 879–885.

Rooney, C.M., Loftin, S.K., Holladay, M.S., Brenner, M.K., Krance,

R.A. & Heslop, H.E. (1995) Early identification of Epstein-Barr

virus-associated post-transplantation lymphoproliferative disease.

British Journal of Haematology, 89, 98–103.

Rooney, C.M., Smith, C.A., Ng, C.Y.C., Loftin, S.K., Sixbey, J.W., Gan,

Y.J., Srivastava, D.K., Bowman, L.C., Krance, R.A., Brenner, M.K. &

Heslop, H.E. (1998) Infusion of cytotoxic T cells for the prevention

and treatment of Epstein-Barr virus-induced lymphoma in allo-

geneic transplant recipients. Blood, 92, 1549–1555.

Shroff, R., Trompeter, R., Cubitt, D., Thaker, U. & Rees, L. (2002)

Epstein-Barr virus monitoring in paediatric renal transplant

recipients. Paediatric Nephrology, 17, 770–775.

Sirvent-von Bueltzingsloewen, A., Morand, P., Buisson, M., Souillet,

G., Chambost, H., Bosson, J.L. & Bordigoni, P. (2002) A prospective

study of Epstein-Barr virus load in 85 hematopoietic stem cell

transplants. Bone Marrow Transplantation, 29, 21–28.

Slavin, S., Nagler, A., Naparstek, E., Kapelushnik, Y., Aker, M., Civi-

dalli, G., Varadi, G., Kirschbaum, M., Ackerstein, A., Samuel, S.,

Amar, A., Brautbar, C., Ben Tal, O., Eldor, A. & Or, R. (1998)

Nonmyeloablative stem cell transplantation and cell therapy as an

alternative to conventional bone marrow transplantation with lethal

cytoreduction for the treatment of malignant and nonmalignant

hematologic diseases. Blood, 91, 756–763.

Wagner, H.J., Cheng, Y.C., Huls, M.H., Gee, A.P., Kuehnle, I., Krance,

R.A., Brenner, M.K., Rooney, C.M. & Heslop, H.E. (2004) Prompt

versus preemptive intervention for EBV lymphoproliferative disease.

Blood, 103, 3979–3981.



Documents

Increased incidence of EBV-related disease following paediatric stem cell transplantation with reduced-intensity conditioning