Clinicopathological features and outcome of late relapses ofnatural killer cell lymphomas 10–29 years after initial remissionWing-Yan Au,1 Seok-Jin Kim,2 Harry H.Y. Yiu,3 Roger K.C. Ngan,3 Florence Loong,4

Won-Seog Kim,2 and Yok-Lam Kwong1*

Natural killer (NK)-cell lymphomas are aggressive and relapses occur

early. Late relapses are exceptional. Ten relapses 17.5 (11–29) years

after first complete remission (CR1) were analyzed. Initial diseases

were stage-I (nasal, n 5 8; tonsil, n 5 1; ileum, n 5 1), treated with

radiotherapy (n 5 6), combined radiotherapy/chemotherapy (n 5 3),

and chemotherapy (n 5 1). Relapse occurred at the same (nasal, n 5

6; tonsil, n 5 1), adjacent (initial: nasal, relapse: palate; n 5 1) or dis-

tant (n 5 2) sites. Five patients died soon afterwards. Four patients

remitted with chemotherapy, two remaining in CR2 (3, 14 years). This

series documented the rare late relapses in NK-cell lymphomas, which

may still respond to salvage therapy.

NK-cell lymphomas show a specific geographical predilection for Asian and

South American countries, and are extremely uncommon in other popula-

tions [1]. Another peculiar feature is the consistent presence of Epstein-Barr

virus (EBV) infection in the lymphoma cells [1–3]. Stage I/II NK-cell lympho-

mas respond well to combined radiotherapy and chemotherapy, with long-

term survivals of up to 50% of patients [2]. However, stage III/IV diseases

have poor prognoses with few survivors.

Most relapses of NK-cell lymphoma occur within the first two years of

remission [1]. However, an exceptional case of relapse of nasal NK-cell lym-

phoma has been described after 38 years of remission [3]. Such late relap-

ses are so rare that practically nothing is known about them.

We analyzed a group of patients with NK-cell lymphomas relapsing after

prolonged remissions to define their clinicopathological features and treat-

ment outcome. Nine cases of late relapses were identified from a cohort of

40 patients followed up long-term in Hong Kong, comprising 26 men and 14

women (Table I). Two patients had been reported briefly before [4]. One

additional case was identified from a Seoul series. There were five men and

five women, at a median presentation age of 34.5 (17–68) years. All patients

had stage I disease, with the nasal area involved in eight cases, the tonsil in

one case, and the ileum in one case. The initial treatment was radiotherapy

in six cases, combined radiotherapy and chemotherapy in three cases, and

chemotherapy alone in one case. The median duration of CR1 was 17.5

(11–29) years before disease relapsed.

In eight cases, first relapse (R1) occurred at the same (nasal, n 5 6; tonsil, n

5 1) or adjacent (initial: nasal, relapse: palate; n 5 1) sites. In two other cases,

the relapse was systemic. Five patients (including two systemic relapses) died

soon afterwards, so that the impact of treatment could not be evaluated. How-

ever, for four patients, who could complete the planned chemotherapy, CR2

was achieved. Two patients had remained in CR2 for 3 and 14 years. One

patient had R2 again at the same nasal site a year afterwards, and achieved a

durable CR3 with further radiotherapy. The remaining patient had R2 12 years

later at the age of 80 years, opted for conservative treatment and died.

This series documented the rare occurrence of late relapses in NK-cell

lymphoma. Late relapses are uncommon in aggressive lymphomas. For

high-grade B-cell lymphomas, survival curves start to plateau at 3–5

years, with relapses after 5–10 years accounting for only about 20% of

treatment failures [5,6]. Relapses, after 10 or more years, are considered

exceptionally unusual, and documented cases are few. In contrast, low-

grade B-cell lymphomas relapse continuously for decades [7]. Interest-

ingly, it has been shown that late relapses of aggressive B-cell lympho-

mas (>10 years) might in fact be due to an undiagnosed underlying low-

grade lymphoma [5].

Although repeated relapses are expected of low-grade lymphomas, the

phenomenon of late relapses in high-grade lymphomas remains ill-defined.

Rearrangement of the immunoglobulin or T-cell receptor gene is a useful

marker to determine if the initial and subsequent lymphomas belong to the

same clone. However, archival materials are not always available, particu-

larly a long time after initial remission. Limited data appear to show that

lymphomas recurring after prolonged remissions may either represent

recrudescence of the initial tumor, or development of a new disease [8,9].

Late recurrences are more likely to involve the same clone in low-grade

lymphomas, [10] but unrelated clones in high-grade lymphomas [11]. Such

findings are consistent with the indolent clinical course of low-grade lym-

phomas, and the aggressive course in high-grade lymphomas.

In our retrospective analysis, it is not possible to determine if the relapses

represented recurrences of minimal tumor cells, or emergence of a new lym-

phoma. It is interesting to note that in 70% of our patients, the relapse

occurred as in-field failures after local radiotherapy, suggesting residual

tumor cells to be present. However, NK-cell lymphomas are high-grade lym-

phomas, and for lymphoma cells to remain quiescent for decades is incon-

sistent with their aggressive nature.

Because NK-cell lymphomas are rare, emergence of a new tumor may

imply an inherent predisposition. In fact, a state of chronic active EBV infec-

tion (CAEBV) has been recognized in Asian populations [12]. In CAEBV,

EBV-infected NK-cells may evolve over years from a polyclonal to a mono-

clonal state, culminating in the development of a frank NK-cell lymphoma. It

is therefore possible that in some patients with NK-cell lymphomas, the origi-

nal tumors have arisen from premalignant EBV-infected NK-cell clones. After

eradication of the initial lymphoma, clonal evolution of the residual EBV-

infected NK-cells over time leads to a second lymphoma. The recurrence of

the lymphoma at the original site may reflect a preferential homing of NK-

cells to specific extranodal sites, which is frequently observed in primary

NK-cell lymphomas [1–3].

The distinction between these two mechanisms requires further studies.

Different from T-cells and B-cells, clonal NK-cell proliferations do not pos-

sess a molecular marker. Analysis of the terminal repeats of the EBV may

show clonal viral incorporation [13]. However, because there are a finite

number of repeats, the assay cannot be used confidently to prove clonal

identity in metachronous lymphomas. The use of more sophisticated tech-

nologies, including cytogenetic analysis or array studies, may be needed to

provide clonal-specific markers to define the relationship between the initial

and relapsed tumors. Quantification of plasma EBV DNA is an accurate sur-

rogate marker of tumor load in NK-cell lymphomas [14]. Serial prospective

monitoring may enable the time-point at which premalignant or malignant

lesions re-emerge to be defined.

In this series, half of the patients died before effective treatment could be

administered or completed. Some of these patients were quite elderly when

the late relapses happened. Two cases relapsed systemically, which imparted

a poor prognosis [1]. However, in 40% of patients who received timely treat-

ment, CR could still be achieved by salvage chemotherapy. These results

show that therapy may be effective for NK-cell lymphomas with recurrences

after prolonged remissions. The CR rate of these late relapses to chemother-

apy is largely comparable to that of de novo NK-cell lymphomas [1]. Our

results thus indicate that the prognosis of late relapses of NK-cell lymphomas

may be similar to primary NK-cell lymphomas, and not as poor as in relapses

of other high-grade lymphomas. Finally, the role of autologous hematopoietic

stem cell transplantation (HSCT) in improving survivals has not been firmly

established in NK-cell lymphomas, particularly for early-stage disease [15].

Therefore, although HSCT is usually performed in patients with relapsed lym-

phomas, the procedure may not necessarily have to be offered to late relap-

ses of NK-cell lymphomas achieving CR after salvage therapy.

Methods

Extranodal NK/T-cell lymphoma, nasal-type, was diagnosed according to

the World Health Organization classification system [3]. Consecutive case

Letters

VVC 2010 Wiley-Liss, Inc.

American Journal of Hematology 362 http://www3.interscience.wiley.com/cgi-bin/jhome/35105

records of patients from Queen Mary Hospital, Hong Kong, Queen Elizabeth

Hospital, Hong Kong, and Samsung Hospital, Seoul, from 1980 to 2009

were reviewed. Late relapses were defined as relapses occurring after a first

CR1 of �10 years. All relapses were biopsy-proven, and shown histologi-

cally to be similar to the original disease. Standard staging procedures

including computerized tomographic scan and bilateral bone marrow biop-

sies were adopted. All relapses were biopsy proven. Treatment of the

patients at presentation and relapse depended on the discretion of the

attending physician.

1Department of Medicine, Queen Mary Hospital, Hong Kong2Division of Hematology/Oncology, Department of Medicine, Samsung Medical Center,

Sungkyunkwan University School of Medicine, Seoul, Korea3Department of Clinical Oncology, Queen Elizabeth Hospital, Hong Kong

4Department of Pathology, Queen Mary Hospital, Hong Kong*Correspondence to: Y.L. Kwong, M.D., Department of Medicine, Professorial

Block, Queen Mary Hospital, Pokfulam Road, Hong KongE-mail: ylkwong@hkucc.hku.hk

Received for publication 16 November 2009; Accepted 19 January 2010Conflict of interest: Nothing to report.

Published online 27 January 2010 in Wiley InterScience(www.interscience.wiley.com).

DOI: 10.1002/ajh.21663

References1. Kwong YL. Natural killer-cell malignancies: Diagnosis and treatment. Leuke-

mia 2005;19:2186–2194.2. Kwong YL, Anderson BO, Advani R, et al. Management of T-cell and natural-

killer-cell neoplasms in Asia: Consensus statement from the Asian OncologySummit 2009. Lancet Oncol 2009;10:1093–1101.

3. Chan JK, Quintanilla-Martinez L, Ferry JA, Peh SC.Extranodal NK/T-cell lym-phoma, nasal type. In:Swerdlow SH,Campo E,Haris NL,Jaffe ES,Pileri SA,S-tein H,Thiele J,Vardiman JW, editors.WHO Classification of Tumours of Hae-matopoietic and Lymphoid Tissues. Lyon:International Agency for Researchon Cancer; 2008. pp 285–288.

4. Schrader C, Janssen D, Kneba M, Lennert K. A 38-year history of natural-

killer-cell lymphoma. N Engl J Med 2004;350:418–419.5. Lee AY, Connors JM, Klimo P, O’Reilly SE, Gascoyne RD. Late relapse in

patients with diffuse large-cell lymphoma treated with MACOP-B. J Clin Oncol

1997;15:1745–1753.6. Zinzani PL, Martelli M, Magagnoli M, et al. How do patients with aggres-

sive non-Hodgkin’s lymphoma treated with third-generation regimens

(MACOP-B and F-MACHOP) fare in the long-term? Haematologica 1999;

84:996–1001.7. Vidal L, Gafter-Gvili A, Leibovici L, et al. Rituximab maintenance for the

treatment of patients with follicular lymphoma: systematic review

and meta-analysis of randomized trials. J Natl Cancer Inst 2009;101:248–

255.8. Nishiuchi R, Yoshino T, Teramoto N, et al. Clonal analysis by polymerase

chain reaction of B-cell lymphoma with late relapse: A report of five cases.

Cancer 1996;77:757–762.9. Libra M, De Re V, Gasparotto D, et al. Differentiation between non-Hodgkin’s

lymphoma recurrence and second primary lymphoma by VDJ rearrangementanalysis. Br J Haematol 2002;118:809–812.

10. Aarts WM, Bende RJ, Bossenbroek JG, et al. Variable heavy-chain gene

analysis of follicular lymphomas: Subclone selection rather than clonal evolu-

tion over time. Blood 2001;98:238–240.11. Lossos A, Ashhab Y, Sverdlin E, et al. Late-delayed cerebral involvement in

systemic non-Hodgkin lymphoma: A second primary tumor or a tardy recur-

rence? Cancer 2004;101:1843–1849.

12. Cohen JI, Kimura H, Nakamura S, et al. Epstein-Barr virus-associated lym-

phoproliferative disease in non-immunocompromised hosts: A status report

and summary of an international meeting, 8–9 September 2008. Ann Oncol

2009;20:1472–1482.

13. Kwong YL, Lam CC, Chan TM. Post-transplantation lymphoproliferative dis-

ease of natural killer cell lineage: A clinicopathological and molecular analy-

sis. Br J Haematol 2000;110:197–202.

14. Au WY, Pang A, Choy C, et al. Quantification of circulating Epstein-Barr

virus (EBV) DNA in the diagnosis and monitoring of natural killer cell and

EBV-positive lymphomas in immunocompetent patients. Blood 2004;104:

243–249.15. Kwong YL. High dose chemotherapy and hematopoietic SCT in the manage-

ment of natural killer-cell malignancies. Bone Marrow Transplant 2009;44:

709–714.

TABLE I. Ten Cases of Late Relapses of Extranodal NK/T-cell Lymphoma, Nasal-type

Initial presentation Relapse

Case Sex/Age Site Stage Treatment (courses) Relapse Site Treatment (courses) Outcome

1 M / 50 nasal IA RT 11 years nasal CEOP (6) 1 ESHAP (1) CR2 for 1 year, second relapse;RT; CR3 for 71 years

2 M / 37 tonsil IA RT 11 years tonsil ESHAP (6) CR2. 31 years3 M / 42 nasal IA ProMACE-cytaBOM (6) 1 RT 12 years nasal – evaluated for salvage treatment4 M / 68 nasal IB RT 14 years nasal mBACOD (6) CR2 for 12 years, second relapse and died5 F / 17 nasal IA ProMACE-cytaBOM (3)

1 mBACOD (3) 1 RT17 years nasal IMVP16 (3) 1 auto-HSCT CR2, 141 years

6 F / 22 ileum IA CHOP (6) 18 years systemic IMVP16 (1) died7 F / 32 nasal IA BACOD (6) 1 RT 19 years nasal, orbit ICE (1) died8 F / 41 nasal IA RT 28 years nasal Nil died before treatment9 M / 23 nasal IA RT 28 years systemic Nil died before treatment

10 F / 31 nasal IA RT 29 years palate CHOP (1) died

M: male; F: female; RT: radiotherapy; CEOP: cyclophosphamide, epirubicin, vincristine, prednisolone; ESHAP: etoposide, cytarabine, cisplatinum, prednisolone; CR: com-plete remission; ProMACE-cytaBOM: cyclophosphamide, doxorubicin, etoposide, prednisolone, cytarabine, bleomycin, vincristine, methotrexate; mBACOD: methotrexate,bleomycin, adriamycin, vincristine, dexamethasone; IMVP16: ifosfamide, methotrexate, etoposide; auto-HSCT: autologous haematopoietic stem cell transplantation; ICE:ifosfamide, carboplatinum, etoposide CHOP: cyclophosphamide, adriamycin, vincristine, prednisolone.

A large family from Argentina with Prekallikrein deficiency dueto a compound heterozygosis (T insertion in intron 7 andAsp558Glu in exon 15): Prekallikrein CordobaAntonio Girolami,1* Sebastian Marun,2 Silvia Vettore,1 Gilda Scaliter,2

Angelica Molina,2 Pamela Scarparo,1 Aldo Tabares,2 and Anna Maria Lombardi1

A family with prekallikrein deficiency (PK) from Argentina was investi-

gated by molecular biology techniques. Parents were not consanguine-

ous. Propositus had prolonged PTT but normal PT. Degree of prolon-

gation of aPTT varied with the activator and showed shortening on

prolonged incubation. Defect was fully corrected by normal, FXII, FXI,

or high molecular weight kininogen (HMWK) deficient plasmas. Prekal-

likrein activity was 4% whereas prekallikrein antigen was normal and

had a normal electrophoretic mobility. Propositus presented a thigh

letters

American Journal of Hematology 363

deep vein thrombosis. He was treated with LMWH and coumarin

together with supportive stockings. An affected brother of the proposi-

tus had two myocardial infarctions and was on coumarin therapy.

Other family members never showed thrombotic events. No bleeding

tendency was ever noted. Children of probands showed PK activity of

50% of normal whereas antigen was normal.

Propositus and three of his siblings were found to be compound heterozy-

gotes for an insertion of a T in intron 7 and an Asp558Glu mutation in exon

15; a Asn124Ser polymorphism was also present. Children of probands

were heterozygotes for either of the two abnormalities. A brother of the pro-

positus, besides the two basic mutations, showed a synonymous coding

polymorphism Asn568Asn, which was present in his son but not in other

family members.

Patients have an Italian background, and reside near Cordoba (Argentina)

(Fig. 1).

Case 1 is the propositus, a 40-year-old man, who was first seen for the

suspicion of a deep vein thrombosis of the right thigh confirmed by sonogra-

phy. Thrombosis appeared idiopathic. Before starting anticoagulant therapy,

a routine test showed a severely prolonged PTT. Further studies demon-

strated PK. There was no bleeding history. In spite of the prolonged PTT,

the patient was treated with enoxaparin twice daily together with 5 mg of

warfarin. After a few days, when the INR had reached the value of 2.5, hep-

arin was discontinued. Coumarin medication was continued for about 1 year

together with 20–30 Hg knee-length supportive stocking. Evolution was sat-

isfactory and there were no recurrences. After 1 year, anticoagulation was

discontinued while he continued to wear stockings. There are no evident

signs of postphlebitic syndrome. Case 2 is a 57-year-old sister of the propo-

situs, who had a negative past personal history. Three deliveries were car-

ried out without any complication. Tooth extractions were uneventful. Case 3

is another 55-year-old sister of the propositus, who had also a negative per-

sonal history. Case 4 is a 53-year-old heavy smoker with hypercholesterole-

mia and hypertension. At the age of 44, he had a myocardial infarction,

which forced him to stop smoking, improve diet, and control hypertension.

Since that time he was treated with coumarin drugs. At the age of 52 he

had another myocardial infarction in spite of the persistent anticoagulation.

At present his condition is fair. All other family members never presented

bleeding or thrombosis.

The propositus and his two sisters showed a prolonged PTT together with

a normal PT and TT. The prolonged aPTT was fully corrected by the addition

of normal plasma, serum, plasma deficient in FXI, FXII, or HMWK. Long

incubation in the aPTT system yielded progressively shortened clotting times

reaching near normal values after 10 min. Bleeding time, platelet count, clot

retraction, Factors I, II, VII, IX, X, XI, and XIII were normal. Prekallikrein

level was 4% of normal. Factor XII activity and HMWK were borderline

(Table I). Prekallikrein antigen was normal and showed normal electropho-

retic mobility in the crossover electrophoresis.

A sibling of the propositus (II-7) showed PK activity level around 50% of

normal whereas PK antigen was normal. Five children of the four affected

subjects showed PK levels around 50% of normal but normal antigen.

Hereditary pattern was autosomal. Molecular biology analysis showed that

the probands were compound heterozygotes for insertion of a T in intron 7

(GeneBank NM_000892:c.830-12T[9]) and a missense Asp558Glu mutation

in exon 15. The sibling (II-7) was also double heterozygote for the same

mutations. The children of the probands were heterozygotes either for the

T insertion in intron 7 or for the Asp558Glu mutation (Supporting Information

Fig. 1). The only grandchild available and son of a heterozygote for the T

insertion mutation were normal (IV-1).

The Asn124Lys polymorphism in Exon 5 was found to segregate together

with the insertion in intron 7 indicating that the two defects were located in

the same allele. Another polymorphism, which did not cause any aminoacid

change (Asn568Asn), was found in exon 15 of the proband, who showed

only about 50% PK activity but normal antigen (II-7). His son (III-10), who

carried the T insertion in intron 7, also showed this second polymorphism.

No other of family member carried this synonymous coding polymorphism.

PK is a glycoprotein synthesized in the liver and secreted as a single poly-

peptide chain with a molecular weight of 88,000 Da. The protein circulates

mainly as a complex with HMWK. Only about 2–5% is free. The normal

plasma concentration is about 40 mg/ml. Prekallikrein is converted to plasma

kallikrein by activated factor XII (FXIIa). This results in the formation of a ser-

ine protease composed of a heavy chain (binding site) and a light chain (cata-

lytic domain) held together by disulfide bonds [1,2]. PK was first described in

1965 by Hathaway et al. [3]. Only about thirty cases were reported up to 1983

[4–5]. An approximately equal number of cases has been described after that

date [6–11]. The peculiarity of this disorder is that the severe in vitro disturb-

ance is not accompanied by hemorrhagic symptoms [2,7]. Only few families

have been studied by molecular biology techniques [12–17].

The gene controlling prekallikrein synthesis is located in chromosome 4

and consists of 14 exons and 15 introns [21]. The amino-terminal portion of

the prekallikrein molecule contains a unique structure, named the apple

domain [2,22].

The finding of a prolonged aPTT corrected by normal plasma and serum

requires differentiation of FXI deficiency from other defects of the contact

phase (Factor XII deficiency, Prekallicrein deficiency and Kininogen deficien-

cies, Williams’s defect, if both high and low molecular weight forms are

absent, and Fitzgerald’s or Flaujeac’s defect if only LMWK is deficient). The

negative clinical picture together with the results of the laboratory tests

clearly established the diagnosis in the propositi.

The shortening of aPTT seen during longer incubation confirms previous

observations, seems typical of PK deficiency and indicates that some activa-

tion of the contact phase may still occur [12,23]. The majority of patients

with prekallikrein defects are cases of true deficiency although abnormal

forms have been reported [24]. Patients with prekallikrein Long Beach and

Zurich showed 35 and 20% of normal antigen, respectively [8,9] The only

Figure 1. Family pedigree. Parents of propositi were not consanguineous. Arrow indicates propositus. and deceased male and female; and compound hetero-zygotes for T insertion (intron 7) and Asp558Glu missense mutation (exon15); heterozygotes for T insertion; Heterozygotes for Asp558Glu missense mutation. Othersymbols refer to normal subjects (open symbols) or to subjects not available for study (symbols with a vertical bar)

letters

364 American Journal of Hematology

patient with a near normal antigen and low activity was described by Kat-

suda et al. [15].

Saito reported the presence of CRM in five of 18 patients investigated.

These five patients came from Mediterranean countries whereas the CRM

negative were Afro-Americans [24]. This observation is confirmed by the

present study. However, this geographical distribution is not absolute as

there are patients with Mediterranean origin, who showed lack of CRM in

their plasma [12]. PK has been associated with hyperthyroidism [10,11]. No

history of hyperthyroidism was present in the probands. An association with

thrombotic disorders has also been occasionally reported [25,26]. The

venous thrombosis seen in the propositus was spontaneous while the MI

occurred in a subject (II-7), who was a smoker, had hypertension and hyper-

cholesterolemia.

The mutation seen in the propositus and in his siblings consisted of the

combination of a substitution Asp558Glu in exon 15 together with a T inser-

tion in intron 7 between nucleotides 220 and 212 from the start codon of

exon 8 together with a known polymorphism (Asn124Ser). This polymor-

phism has already been described at the homozygote level in prekallikrein

Saki, and it was maintained to be necessary for the appearance of full PK

Saki phenotype (low activity and near normal antigen) when coinherited with

the Gly104Arg mutation in exon 5 [15]. This polymorphism is common in the

US population but its role on the PK level is unknown [21]. One of the

molecular abnormalities seen in the propositus (Asp558Glu mutation in exon

15) involves residue 558, just beside the Ser 559 belonging to the catalytic

triade His415, Asp464, and Ser559 responsible for enzymatic activity. We

have no sure explanation for the discrepancy seen in one of the propositi (II-

7) between the genetic pattern and the clotting tests. It is possible that the

synonymous coding polymorphism Asn568Asn seen only in this patient and

in his son may exercise a mitigating or compensatory effect on one of the

two basic mutations. The Asn568Asn polymorphism is located close to the

splicing site of exon 15, which is only six nucleotides (two triplets) apart

from the Asp558Glu mutation. It is highly probable that the compensatory

effect of this polymorphism is exercised on the exon 15 mutation.

Synonimous coding polymorphisms have been recently shown, in some

noncoagulation related diseases, to influence genetic penetrance [27–30].

The effect seems on splicing and mRNA stability [30]. The present observa-

tion represents the first example of such mechanism for genes involved in

the synthesis of coagulation factors.

The probands’ children were heterozygous for either of the two mutations.

As the PK activity was always 50% of normal it may be assumed that each

mutation contributed to the same extent to the pattern seen in the probands.

The number of families with PK deficiency studied by molecular biology

techniques is so far limited (Table II).

The exons involved are 5, 8, 11, 14, and 15. All mutations were different

with the exception of two instances. The Cys529Tyr (exon 14) was described

both at the heterozygous and homozygous level [13,16]. The Gly104Glu

(exon 11) mutation was found in two distinct unrelated Japanese families

[13,17]. The mutation Lys529Tyr in exon 14 is close to the site of one of the

two mutations seen in the propositi, namely the Asp558Glu. However, phe-

notype is different since the Lys529Tyr mutation shows low activity and low

antigen levels.

The presence of the same two defects in all four patients, the compatible

hereditary pattern, the presence of either defect at the heterozygous level in

the children of the propositi and the lack of these mutations in a control pop-

TABLE II. Main Features of All Cases of PK Deficiency Who Have Been Characterized by Molecular Biology Studies

Author (year)Age and

sex Consanguinity PTTPK

activityPK

antigen Background Genotype Mutation (Exon) Comments

Lombardi et al. (2003) 14, M No 110 <1 Traces Italian Compound Het. Trp383Stop 1Lys529Tyr(11 and 14)

Shigekiyo et al. (2003) 47, M n.r. 102 <1 25 Japan Hom. Gly401Glu (11) A sister similarly affectedJones et al. (2004) 79, M Yes 125 <1 4 Caucasian Hom. Arg94Stop (5) Parents were first cousinsKatsuda et al. (2007) 53, M Yes 74 1–3 n.r. Japan Double hom. Gly104Asp 1

Asn124Ser (5)A brother and sister

similarly affectedFrancois A et al. (2007),

Case 163, M n.r. 176 <1 7 Portugal Hom. Lys529Tyr (14)

Case 2 38, M n.r. 186 <1 7.5 France Hom Lys529Tyr (14) Unrelated to previous caseNagaya et al. (2009) 69, F Yes 65 1 <10 Japan Hom. Gly401Glu (11) Unrelated to family presented

by Shigakiga et al. in 2003Present kindred 40, M No 96 4 90 Italian Comp Het T insertion 1

Asp558Glu(intr 7/ex. 15)

One brother and two sisterssimilarly affected

TABLE I. Coagulation Study in the Propositus

Patient Normal values Comments

Coagulation testsPlatelet count 2,00,000/ml 1,50,000–3,00,000/mlBleeding time Normal <5 minAPTT 96a 32–38 secPT 13.5 13–15 secTT 19 18–22 secPK activity 5 70–130 secPK antigen 95 70–130 secFactors I, II, V, VII, VIII, IX, X, XI, and XIII normal 70–130%Plasminogen 90 70–130 Carried out on frozen plasmaFactor XII 70 70–130HMWK 60 70–130 Carried out on frozen plasmaEuglobulin lysis times 25 hours 10–30 hours Carried out on frozen plasma

aPTT mixing testsPatient plasma 1 pooled normal plasma (frozen) 38Patient plasma 1 prekallikrein deficient plasma 115 Lyophilized plasma obtained from

Sigma-AldrichPatient plasma 1 prekallikrein deficient plasma 106 Frozen plasma (personal case)Patient plasma 1 Factor XI deficient plasma 41 Frozen plasma (personal case)Patient plasma 1 HMWK deficient plasma 42 Lyophilized plasma obtained from

‘‘Henry Ford Health System, (Detroit)Patient plasma 1 Factor XII deficient plasma 39 Fresh plasma (personal case)

aMean value obtained with different reagents.Similar results were obtained in his two affected sisters. The other affected member of the family showed higher PK levels (see text)

letters

American Journal of Hematology 365

ulation firmly establish the genotype-phenotype relation in the family here

presented.

Methods

Routine coagulation tests and clotting assays were carried out in Cordoba

on fresh plasma and repeated in Padua on frozen samples as previously

reported [18].

Different PTT reagents were used, namely Actin (Dade Laboratories,

Miami) Pathrontin (Berhingwerke, Marbug, Germany), Thrombosil S (Ortho

Diagnostic, Raritan). Different incubation times were used in the PTT sys-

tem, namely, 2, 4, 6, 8, and 10 min. Prekallikrein clotting assay was carried

out using an ellagic acid activated partial thromboplastin time and two pre-

kallikrein deficient plasmas (personal case and lyophilized PK free plasma

from Sigma Laboratories, St Louis). Factor XI, XII, and HMWK clotting

assays were carried out using as substrate specific factor deficient plasmas

(Factor XII or Factor XI deficient plasmas from personal patients, or lyophi-

lized plasma HMWK free plasma kindly supplied by investigators of Henry

Ford Health System, Detroit).

Cross correction studies were carried out using of either home made fresh

or deep frozen plasmas (factor XI, factor XII, and PK) or lyophilized plasma

kindly supplied to us by Henry Ford Hospital investigators (HMWK). Euglo-

bulin lysis time and plasminogen activity were carried out as previously

reported [18].

Electroimmunoassay was carried out using two rabbit raised antiserums

supplied by Behring Laboratories, Marburg, Germany or by Affinity Biochem-

icals, Ancester, Ontario, Canada) [19]. Cross over immunoelectrophoresis

was carried out as previously reported [20] using the raised rabbit antiserum

supplied by Affinity Biochemicals, Ancester, Ontario, Canada.

Genomic DNA was extracted from peripheral blood leukocytes by QIAamp

DNA blood mini kit (Qiagen, Hilden, Germany) following the manufacture’s

procedure.

The promoter, exons 1–15, and boundaries introns including the splice

junctions of Prekallikrein gene were amplified by PCR method, using, for all

members of the family, specific primers as reported by Yu et al. [21]. The

PCR method by Yu et al. was partially modified by us; in particular, the reac-

tion was performed in a total volume of 50 ml using 200 ng of extracted

DNA. The mix contained 50 mM KCL, 10 mM Tris-HCL (pH 9), 2.5 mM

MgCl2, 200 mM dNTPs (Ultrapure dNTP Set, Pharmacia Biotech, Uppsala,

Sweden), 300 ng of every primer (MGW spa Biotech.), and 1U Taq DNA

Polymerase (Promega Madison, WI) [12].

The reaction mixture was then amplified using the cycling steps: 948C, 5min for one cycle; 948C, 1 min; annealing temperature, 1 min, 728C, 1 min

for 30 cycle and 728C, 7 min for one cycle.

The PCR products were electrophoresed on a 1% agarose gel in Tris-

borate-EDTA buffer (pH 8.3) stained with ethidium bromide and purified with

filter Microcon (Millipore Corporation, Bedford, MA). The purified PCR prod-

uct was sequenced directly using a Taq Dye Deoxy Terminator Sequencing

kit (Applied Biosystems, Foster City, CA) and a ABIPRISM 3100 Genetic

DNA sequencer (Applied Biosystems, Foster City, CA).

1University of Padua Medical School, Department of Medical and SurgicalSciences, Padua, Italy

2Centro Medico de Cordoba S.A., Servicios de Bioquımica Clınica y de MedicinaVasculary Transfusion, Cordoba, Argentina

Contract grant sponsor: Associazione per l’Emofilia e le Coagulopatie delle TreVenezie

*Correspondence to: A. Girolami, Department of Medical and Surgical Sciences,Padua University, Via Ospedale, 105, 35128 Padua, Italy

E-mail: ylkwong@hkucc.hku.hkAdditional supporting information may be found in the online version of this article.

Conflict of interest: Nothing to report.Published online 20 January 2010 in Wiley InterScience

(www.interscience.wiley.com).DOI: 10.1002/ajh.21654

References1. Chung DW, Fujikawa K, McMullen BA, Davie EW. Human plasma prekallik-

rein, a zymogen to a serine protease that contains four tandem repeats. Bio-chemistry 1986;25:2410–2417.

2. Colman RW. Contact Activation (Kallikrein-Kinin) pathway: Involving inflamma-tory, fibrinolytic, anticoagulant physiologic and pathophisiologic activities. In:Colman RW, Marder VJ, Clowes AW, George JN, Goldhaber SZ, editors.Hemostasis and Thrombosis, 5th ed. Philadelphia: Lippincot & Williams;2006. pp 107–130.

3. Hathaway WE, Belhasen LP, Hathaway HS. Evidence for a new plasmathromboplastin factor I. Case report, coagulation studies and physicochemicalproperties. Blood. 1965;26:521–532.

4. Goodnough LT, Saito H, Ratnoff DD. Thrombosis or myocardial infarction incongenital clotting factor abnormalities and chronic thrombocytopenias: Areport of 21 patients and a review of 50 previously reported cases. Medicine(Baltimore) 1983;62:248–55.

5. Entes K, La Duca FM, Tourbaf KD. Fletcher factor deficiency, source of varia-tions of the actived partial thromboplastin time test. Am J Clin Pathol 1981;75:626–628.

6. DeLa Cadena RA. Fletcher factor deficiency in a 9-year-old girl: Mechanismsof the contact pathway of blood coagulation. Am J Hematol 1995;48:273–277.

7. Saito H, Kojima T. Factor XII, Prekallikrein and high molecolar weight kinino-gen. In: High KA, Roberts HR, editors. Molecular Basis of Thrombosis andHaemostasis. New York: M. Dekker Inc.; 1995. pp 269–285.

8. Wuillemin WA, Furlan M, von Felten A, Lammie B. Functional characterizationof a variant prekallikrein (PK Zurich). Thromb Haemost 1993;70:427–432.

9. Bouma BN, Kerbiriou DM, Baker J, Griffin JH. Characterization of a variantprekallikrein, prekallikrein Long Beach, from a family with mixed cross-react-ing material-positive and cross-reacting material-negative prekallikrein defi-ciency. J Clin Invest 1986;78:170–176.

10. De Stefano V, Leone G, Teofili L, et al. Association of Graves’ disease andprekallikrein congenital deficiency in a patient belonging to the first CRM1prekallikrein-deficient Italian family. Thromb Res 1990;60:397–340.

11. Kyrle PA, Niessner H, Deutsch E, et al. CRM1 severe Fletcher factor defi-ciency associated with Graves’ disease. Haemostasis 1984;14:302–306.

12. Lombardi AM, Sartori MT, Cabrio L, et al. Severe prekallikrein (Fletcher fac-tor) deficiency due to a compound heterozygosis (383 Trp stop codon andCys529Tyr). Thromb Haemost 2003;90:1040–1045.

13. Shigekiyo T, Fujino O, Kanagawa Y, Matsumoto T. Prekallikrein (PK) Tokush-ima: PK deficiency caused by a Gly401-->Glu mutation. J Thromb Haemost2003;1:1314–1316.

14. Jones DW, Russel G, Allford SL, et al. Severe prekallikrein deficiency associ-ated with homozygosity for an Arg 94 Stop nonsense mutation. Brit J Haema-tol 2004;127:220–223.

15. Katsuda I, Maruyama F, Ezaki K, et al. A new type of plasma prekallikreindeficiency associated with homozygosity for Gly104Arg and Asn124Ser inapple domain 2 of the heavy-chain region. Eur J Haematol 2007;79:59–68.

16. Francois D, Trigui N, Leterreux G, et al. Severe prekallikrein deficiencies dueto homozygous C529Y mutations. Blood Coagul Fibrinolysis 2007;18:283–286.

17. Nagaya S, Morishita E, Maruyama K, et al. An elderly case of congenital Pre-kallikrein deficiency. Nippon Ron Igak Zasshi 2009;46:148–151.

18. Girolami A, Molaro G, Lazzarin M, et al. A ‘‘new’’ congenital haemorrhagiccondition due to the presence of an abnormal factor X (factor X Friuli): Studyof a large kindred. Br J Haematol 1970;19:179–192.

19. Girolami A, Patrassi GM, De Marco L, Cappellato G. Lack of factor XII anti-gen in factor XII deficiency as de terminate by electroimmuno assay. Blut1982;45:141–142.

20. Girolami A, Sticchi A, Bareggi G. Crossover electrophoresis (electrosyneresis)visualization of the abnormal factor X (factor X Friuli). J Lab Clin Med 1972;80:740–747.

21. Yu H, Anderson PJ, Freedman BI, et al. Genomic structure of the humanplasma prekallikrein gene, identification of allelic variants, and analysis inend-stage renal disease. Genomics 2000;69:225–234.

22. McMullen BA, Fujikawa K, Davie EW. Location of the disulfide bonds inhuman plasma prekallikrein: The presence of four novel apple domainsin the amino-terminal point of the molecule. Biochemistry 1991;30:2050–2056.

23. Asmis LM, Sulrer I, Furlan M, Lammle B. Prekallikrein deficiency: The charac-teristic normalizazion of the severely prolunged aPTT following incrased pre-incubation time is due to autoactivation of factor XII. Thromb Res 2002;105:463–470.

24. Saito H, Goodenough LT, Soria J, et al. Heterogeneity of humanprekallikreindeficiency (Fletcher Factor): Evidence that five of 18 cases are positive forcross-reactive material. New Eng J Med 1981;305:910–914.

25. Currimbhoy Z, Vinciguerra V, Palakavongs P, et al. Fletcher factor deficiencyand myocardial infarction. Am J Clin Pathol 1976;65:970–973.

26. Harris MG, Exner T, Rickard KA, Kronenberg H. Multiple cerebral thrombosisin Fletcher factor (prekallikrein) deficiency: A case report. Am J Hematol1985;19:387–393.

27. Chamary JV, Parmley JL, Hurst LD. Hearing silence: Non-neutral evolution at

synonymous sites in mammals. Nat Rev Genet 2006;7:98–108.28. Narendra U, Gayle JT, Pauer GJT, Hagstrom SA. Genetic analysis of comple-

ment factor H related 5, CFHR5, in patients with age-related macular degen-

eration. Mol Vis 2009;15:731–736.

29. Siddiq A, Lepretre F, Hercberg S, et al. A synonymous coding polymorphism

in the alpha2-Heremans-schmid glycoprotein gene is associated with type 2

diabetes in French Caucasians. Diabetes 2005;54:2477–2481.

30. Nielsen KB, Sorensen S, Cartegni L, et al. Seemingly neutral polymorphic

variants may confer immunita to splicing-inactivating mutations: A synony-

mous SNP in Exon 5 of MCAD protects from Deleterious Mutations in a flank-

ing exonic Splicing Enhancer. Am J Hum Gen 2007;80:416–432.

letters

366 American Journal of Hematology

Pain as an emergent issue in thalassemiaFelicia Trachtenberg,1* Dru Foote,2Marie Martin,3 Susan Carson,4 Thomas Coates,4 Owen Beams,3

Olivia Vega,2Manuela Merelles-Pulcini,5 Patricia J. Giardina,6 Dorothy A. Kleinert,6 Janet Kwiatkowski,3

Alexis A. Thompson,7 Ellis J. Neufeld,8 Leann Schilling,9 Vivek Thayalasuthan,10 Zahra Pakbaz,2 andRobert Yamashita11, For The Thalassemia Clinical Research Network

Thalassemia is a congenital blood disorder often requiring chronic blood

transfusions and iron chelation therapy [1,2]. While advances in treatment

have resulted in increased life expectancy [3], extended life spans have

exposed previously unidentified issues, including bodily pain. The aim of

this study was to examine the prevalence, severity, predictors, and effects of

pain in 265 adults/adolescents and 103 children with thalassemia. Overall,

69% of adults/adolescents reported bodily pain on the SF-36v2 health sur-

vey, with 28% reporting at least moderate pain. Parents reported pain in

56% of children using the PF-28 child health questionnaire, with only 11%

reporting pain fairly often. There were no significant differences in pain in

children with thalassemia compared with the general population. In adults/

adolescents, pain increased significantly with age (P 5 0.005), more so than

in the general population. This study highlights the fact that children and

young adults with thalassemia experience pain comparable to the general

population, whereas older adults (aged 351) experience greater pain. Our

findings show that increased pain is associated with decreased quality of

life and increased anxiety and depression.

A recent review of the literature revealed no comprehensive assessments of pain

in the thalassemia population, although some studies included specific types of

pain, pain treatment, or pain as a component of general quality of life (QOL) [4–

15]. The Thalassemia Clinical Research Network (TCRN) is an NIH/NHLBI-

funded network composed of five core thalassemia centers in North America,

one in London and their associated satellite sites. To date, the TCRN Low Bone

Mass Cross-sectional Observational Study (LBMCOS) has published data on

bone and joint pain, with 34% of participants reporting pain [12]. We expand on

these findings using further data from the LBMCOS study, along with recently

collected baseline data from the Thalassemia Longitudinal Cohort (TLC).

LBMCOS Study

Data were collected from 371 patients (aged 6–75, mean 22.9 ± 12.1

years, 49% male, 65% b-thal major). One hundred twenty-seven (34%)

reported bone or joint pain in the last 30 days that was not related to a spe-

cific trauma or injury. These patients reported feeling pain on average 3

days a week, 7.5 hr per day, in their worst week that month. The most fre-

quent site of bone/joint pain was the back (24%), followed by knees (15%)

and head/neck (10%). Eight percent of participants experienced pain in their

shoulders, hips, ankles, and/or legs. Up to 5% of participants reported pain

in their hands, wrists, elbows, arms, feet, and/or ribs. Forty-four participants

had received transfusions while experiencing pain and 34% reported that

the transfusions helped reduce or eliminate the pain.

TLC Study

Data were collected from 265 adults/adolescents (29.4 ± 10.4 years,

range 14–58, 52% female, 77% b-thal major) and 103 children (9.8 ± 2.6

years, range 5–14, 52% female, 71% b-thal major) (Table I). Overall, 69% of

adults/adolescents reported bodily pain in the past 4 weeks, with 28%

reporting at least moderate pain. Fifty percent reported that pain interfered

with their work (both inside and outside of the home) in the last 4 weeks,

with 25% reporting at least moderate interference. As expected, pain

severity and interference were highly correlated (Spearman correlation 5

0.82; P < 0.001). Parents reported pain in 56% of children, with 11% report-

ing pain fairly often in the last 4 weeks.

Pain severity increases with age (Table II), but does not vary significantly

with sex or thalassemia diagnosis. In analysis of the SF-36 bodily pain scale,

which combines both pain severity and interference, age was a significant

independent predictor of pain, with a decrease in quality of life due to pain

with age (slope 5 20.26 ± 0.09/year; P 5 0.005). Quality of life due to pain

in thalassemia declines greatly with age, compared with only a slight decline

in the general population (Fig. 1A). Z-scores of pain in U.S. patients in relation

to the normal population decrease significantly with age (P 5 0.042), with a

noticeable gap by age 35. In children, there were no significant differences

(Fig. 1B); however, quality of life due to pain incidence decreased nonsignifi-

cantly with age in children (slope 5 21.01 ± 0.93/year; P 5 0.28). There was

no association between pain and gender, thalassemia diagnosis, bone density,

pretransfusion hemoglobin, vitamin D, or chelator choice in children.

Clinically, in adults, lower bone density was marginally significantly correlated

with increased pain (r 5 0.29 for whole body, P 5 0.065; r 5 0.15 for hip, P 5

0.079; r 5 0.11 for spine, P 5 0.095), as was use of bisphosphonates (P 5

0.14). Pretransfusion hemoglobin was not correlated with pain (r 5 20.06, P 5

0.33). Hormone replacement therapy and 1,25 vitamin D were not associated

with pain, but low 25 vitamin D was associated with increased pain (P 5

0.010). There was a marginally significant effect of chelator, with higher pain

associated with use of deferoxamine compared with deferasirox (P 5 0.11),

with a slight decrease in pain with increasing dose of deferasirox (P 5 0.074).

In adults and adolescents, bodily pain is significantly correlated with

decreased quality of life in all domains measured by the SF-36 (physical

functioning, role-physical, general health, vitality, social functioning, role-

emotional, and mental health). Pain shows a higher correlation with the

physical component summary (r 5 0.78, P < 0.001) than the mental com-

ponent summary (r 5 0.33, P < 0.001) (Fig. 2A). Increased bodily pain is

also significantly correlated with increased anxiety and depression (r 5

This is publication number 12 of the Thalassemia Clinical Research Net-

work (TCRN). The following institutions and researchers contributed to the

Thalassemia Clinical Research Network Thalassemia Longitudinal Cohort

data reported in this article.

Children’s Hospital, Boston (N 5 38): Ellis Neufeld, MD, PhD, Principal Investi-

gator, Jennifer Braunstein, NP, Research Nurse, Amber Smith, Study Coordina-

tor, Latoya Lashley, Study Coordinator; Satellite: University of Texas Southwest-

ern Medical Center at Dallas (N 5 12), Charles Quinn, MD, MS, Principal Inves-

tigator, Deborah Boger, RN, MSN, PNP, Study Coordinator, Leah Adix, Study

Coordinator, Sandra Richardson, Study Coordinator; Children’s Healthcare of

Atlanta (N 5 16), Jeanne Boudreaux, MD, Principal Investigator, Leann Hassen,

Study Coordinator; Baylor College of Medicine (N 5 6), Brigitta Mueller, MD,

Principal Investigator, Bogden Dino, Study Coordinator. Weill Medical College of

Cornell University (N 5 59): Patricia Giardina, MD, Principal Investigator, Eliza-

beth Evans, Study Coordinator; Satellite: Winthrop University Hospital (N 5 6),

Mark Weinblatt, MD, Principal Investigator, Linda Skelly, Study Coordinator. The

Children’s Hospital of Philadelphia (N 5 59): Janet Kwiatkowski, MD, Principal

Investigator, Marie Martin, RN, Research Nurse, Owen Beams, Study Coordina-

tor; Satellite: Children’s Memorial Hospital, Chicago, IL (N 5 39), Alexis Thomp-

son, MD, Principal Investigator, Janice Beatty, RN, Research Nurse, Tiffany

Drinkwater, Study Coordinator. Children’s Hospital at Oakland (N 5 52): Elliott

Vichinsky, MD, Principal Investigator, Dru Foote, NP, Research Nurse, Nancy

Sweeters, Study Coordinator, Olivia Vega, Study Coordinator; Satellites: Child-

ren’s Hospital of Los Angeles (N 5 12), Thomas Coates, MD, Principal Investi-

gator, Susan Carson, RN, Research Nurse, Eun Ha Pang, Study Coordinator,

Rachna Khanna, Study Coordinator; Stanford Hospital (N 5 5), Michael Jeng,

MD, Principal Investigator, Kokil Bakshi, Clinical Research Associate; Children’s

and Women’s Health Center of British Columbia (N 5 4), John Wu, Principal

Investigator, Heather McCartney, RN, Research Nurse, Colleen Fitzgerald, Study

Coordinator, Stephanie Badour, Study Coordinator. Toronto General Hospital,

Toronto, Ontario, Canada (N 5 5): Nancy F. Olivieri, MD, Principal Investigator,

Vivek Thayalasuthan, Study Coordinator; Satellite: Hospital for Sick Children (N

5 64), Isaac Odame, MD, Principal Investigator, Manuela Merelles-Pulcini, RN,

Study Coordinator. University College London (N 5 15), John Porter, MD, Princi-

pal Investigator, Cindy Bhagwandin, Study Coordinator; Satellite: Whittington

Hospital (N 5 24), Farrukh Shah, MD, Principal Investigator. NHLBI oversight,

Kathryn Hassell, MD. Data Coordinating Center: New England Research Insti-

tutes, Sonja McKinlay, PhD, Principal Investigator, Lisa Virzi, RN, MS, MBA, Proj-

ect Director, Felicia Trachtenberg, PhD, Senior Statistician.

letters

American Journal of Hematology 367

20.37 for anxiety, r 5 20.36 for depression, P < 0.001 for both) (Fig. 2B). In

children, the bodily pain scale is significantly correlated with the physical func-

tioning, mental health, general health, and family cohesion scales, but not the

role-emotional behavior, role-physical behavior, self esteem, parental impact-

emotional, parent impact-time, or family activities scales. Pain was significantly

correlated with the physical summary score (r 5 0.53, P < 0.001), but not the

psychosocial summary score (r 5 0.12, P 5 0.24) (Fig. 2C).

This study had several advantages compared with previous studies of

pain in thalassemia. First, we were able to assess bodily pain in a large

sample of adults and children of varying ages with thalassemia. Second,

because we used validated and well-studied quality of life instruments, we

were able to compare our findings to the general U.S. population of varying

ages. We were also able to study the impact of bodily pain on quality of life

and mood using these validated instruments.

This study identified 69% of adults/adolescents reporting bodily pain in the

past 4 weeks, which is consistent with the 62% reporting any pain in the

Scalone study [18]. This study also found 28% of adults/adolescents experi-

encing at least moderate pain. This is consistent with the Pakbaz [9] study,

which found moderate pain in 21% of transfusion-dependent and independ-

ent patients, though unlike this study, found decreased prevalence of severe

pain in transfusion-independent patients. The high prevalence of pain in the

thal intermedia patients in this study is likely due to the eligibility criteria

requiring patients who need at least annual monitoring for end-organ injury.

This study also expanded upon the Vogiatzi [12] study, which found bone

and/or joint pain in 34% of participants, compared with our finding of any

bodily pain in 68% of participants. Thus, further study of causes and loca-

tions of pain in addition to bone/joint pain are warranted.

Parents reported frequent pain in only 11% of children. This is in compari-

son to impaired QOL due to pain/discomfort in 64% of children in the Shali-

gram [10] study. This sharp difference is likely due to differences in quality

of life instruments. The EQ5D used by Shaligram asks parents (as a proxy)

to assess their child’s pain or discomfort (none, moderate, or extreme),

whereas the CHQ asks how often does the child have bodily pain or discom-

fort. Importantly, our CHQ PF28 data found some pain in 58% of children.

Consistent with the Vogiatzi [12] study of bone/joint pain, we found

increased prevalence of pain with age. Young adults with thalassemia expe-

rienced pain comparable to the general population, whereas older adults

(aged 351) experienced greater pain. This may explain why impaired QOL

due to bodily pain has not been consistently observed [13–15,18]. Given the

association of pain with low vitamin D level, care should be taken to ensure

adequate vitamin D intake. Additionally, there was a trend toward increased

pain with lower bone density and for patients on bisphsphonates, suggestive

of a possible relationship in a subset of patients. Since there have been no

previous studies of pain in thalassemia, the possible sources of pain are not

clearly apparent. Some possible mechanisms of pain in thalassemia include

the following: compression fractures, pathologic fractures, and impingement

on nerve roots by extra medullary hematopoietic masses. While case stud-

ies of these pathophysiologic mechanisms have been discussed in standard

textbooks, there has been no systematic review of the prevalence within this

population. Patients with osteopenia and osteoporosis often complain of

pain secondary to fractures, particularly back pain associated with compres-

sion fractures of the vertebrae.

Thalassemia patients as they age frequently complain of ‘‘throbbing back

pain’’ in the week before transfusion. Practitioners generally assume that

this is due to bone marrow ‘‘pressure’’ from increased marrow activity which

is subsequently suppressed when the patient is transfused. This assumption

is supported by our finding that 34% of participants reported improvement of

pain with transfusion. Marrow expansion could also be a source of pain.

As expected, pain impacts negatively on quality of life, affecting both phys-

ical functioning and mental health, including increased reports of anxiety

and depression. Increased pain was observed with use of deferoxamine

compared with deferasirox, which may be due to the route of administration.

This association may also reflect patient/physician preferences in chelator

choice and is complicated by the fact that many patients have switched che-

lators in recent years, some multiple times.

As responses to the instruments used may vary by culture, evidenced by

the large difference observed between patients in North America and the

U.K., comparison across cultures is not possible. Another study limitation is

that we have only assessed pain at one time point, although longitudinal

data collection is ongoing. Furthermore, this study was not adequately pow-

ered to detect differences in levels of pain by thalassemia diagnosis, as the

majority of participants were b-thal major. Given the emergence of pain as

an issue in older thalassemia patients, further longitudinal study of both

transfused and nontransfused patients would be advantageous to discern

timing, location, and causes of pain in thalassemia.

Methods

The TCRN LBMCOS and TLC protocols were approved by the TCRN Data

and Safety Monitoring Board and by the ethical review boards of all TCRN

institutions. Informed consent, and assent in the case of a minor, was

obtained.

The LMBCOS study recruited TCRN patients of all thalassemia syn-

dromes, age 6 and older, excluding pregnancy and pre-existing medical con-

ditions known to require chronic systemic administration of steroids. Data on

bone and joint pain was collected by patient interview.

TABLE I. Baseline Demographics and Pain Dataa for the Thalasse-

mia Clinical Research Network (TCRN) Thalassemia Longitudinal

Cohort (TLC)

Aged 141 yearsb

N 5 265Aged <14 yearsc

N 5 103

Age (years) 29.4 (10.4), 14.2–58.3 9.8 (2.6), 5.0–14.1SexMale 128 (48.3%) 49 (47.6%)Female 137 (51.7%) 54 (52.4%)

Thalassemia diagnosisB-thal transfused 81 204 (77.0%) 73 (70.9%)B-thal transfused <8 24 (9.1%) 3 (2.9%)B-thal nontransfused 2 (0.8%) 2 (1.9%)HbH 0 (0.0%) 4 (3.9%)HbH CS 4 (1.5%) 3 (2.9%)E-B-thal transfused 81 21 (7.9%) 10 (9.7%)E-B-thal transfused <8 6 (2.3%) 2 (1.9%)E-B-thal nontransfused 1 (0.4%) 0 (0.0%)alpha-thal 2 (0.8%) 3 (2.9%)Other/missing 1 (0.4%) 3 (2.9%)

CountryUS 194 (73.2%) 76 (73.8%)Canada 36 (13.6%) 27 (26.2%)UK 35 (13.2%) 0 (0.00%)

Hemoglobind (g/dL) 10.0 (1.1), 5.0–12.9 9.5 (0.9), 6.4–11.2Bone mineral density z-scoree

Spine 21.8 (1.7), 26.9–4.7 21.0 (1.3), 23.0–2.6Hip 21.3 (1.2), 23.7–2.3 21.0 (1.3), 22.6–1.0Whole body 20.5 (2.0), 23.7–3.5 21.5 (1.7), 23.2–1.5

Pain severityf

None 81 (30.6%)Very mild 65 (24.5%)Mild 46 (17.4%)Moderate 40 (15.1%)Severe 25 (9.4%)Very severe 8 (3.0%)

Pain interferenceg

Not at all 133 (50.2%)A little bit 65 (24.5%)Moderately 29 (10.9%)Quite a bit 34 (12.8%)Extremely 4 (1.5%)

Pain incidence in past 4 weeksh

None of the time 45 (43.7%)Once or twice 32 (31.1%)A few times 15 (14.6%)Fairly often 6 (5.8%)Very often, every/almostevery day

5 (4.9%)

aMean (SD), range for continuous variables; N (%) for categorical variables.bCompleted the self-report SF-36v2 quality of life health survey.cCompleted the parent-report PF-28 child health questionnaire (CHQ).dPretransfusion, for transfused patients only. N 5 239 aged 141 and N 5 92 aged <14.eFor age 141, N 5 222 for spine, 136 for hip, and 41 for whole body; for age

<14, N 5 36 for spine, 11 for hip, and 14 for whole body.f‘‘How much bodily pain have you had during the past 4 weeks?’’ (SF36v2 question 8).g‘‘During the past 4 weeks, how much did pain interfere with your normal work

(including both work outside the home and housework)’’ (SF36v2 question 9).h‘‘During the past 4 weeks, how often has your child had bodily pain or discom-

fort’’ (CHQ PF28 question 4.1).

letters

368 American Journal of Hematology

Eligibility for the TLC study included patients of all thalassemia syndromes

who required at least annual monitoring for end-organ injury related to tha-

lassemia. This manuscript reports baseline data from this ongoing study.

Patients with a prior successful hematopoietic stem cell transplant (N 5 12)

were excluded from analysis. At baseline, participants aged 14 and older

were asked to complete the SF-36v2 health survey [16] and the hospital

anxiety and depression scale (HADS) [19]. Parents of children <14 years at

baseline were asked to complete the PF-28 child health questionnaire

(CHQ) [17]. The SF-36v2 and CHQ both include core pain assessments.

For all participants, current medications, chelation regimen, most recent pre-

transfusion hemoglobin, and bone mineral density assessed by dual-energy

X-ray absorptiometry (DXA) were recorded from chart review. Urine and

serum samples from each participant were stored at 2808C and analyzed

as a batch at a central facility. 25-hydroxyvitamin D was measured by com-

petitive radioimmunoassay following extraction and 1,25-dihydroxyvitamin D

by column chromatography and radioimmunoassay

Statistical Analysis

Frequencies were calculated for all pain measures, both overall and by

age groups, sex, and thalassemia diagnosis. Age groups were chosen to be

consistent with those reported in the SF-36v2 and CHQ manuals [16,17] for

the normal population. Beta-thal major was defined as b-thalassemia requir-

ing at least eight transfusions in the last year, whereas b-thal intermedia

were homozygous or compound heterozygous patients requiring fewer than

eight transfusions in the last year.

The bodily pain scales of the SF-36 and CHQ were scored and responses

compared with the ‘‘normed’’ general adult and pediatric populations [16,17]

by computation of z-scores based on normed means and standard devia-

tions. Regression analysis of z-scores on age, controlling for gender, were

used to assess differences between pain in U.S. thalassemia patients and

U.S. norms over age.

Correlation analysis and linear and ordinal logistic regression were used

to model predictors of pain. Predictors significant in initial analysis, control-

ling for age, sex, and country, were entered into multivariate models. Poten-

tial predictors included age, sex, country, thalassemia diagnosis, regular

transfusion (y/n), bone density, pretransfusion hemoglobin level, vitamin D

level, hormone replacement therapy, bisphosphonate use, and chelator

choice/dose. Partial correlation, controlling for age and sex, was used to

TABLE II. Pain Severity in Adultsa and Childrenb by Diagnosis, Sex, and Age

Pain severity: Adults

None Very mild Mild Moderate Severe Very severe

DiagnosisB-thal major 62 (30.4%) 53 (26.0%) 32 (15.7%) 34 (16.7%) 17 (8.3%) 6 (2.9%)B-thal intermedia 6 (23.1%) 5 (19.2%) 6 (23.1%) 4 (15.4%) 4 (15.4%) 1 (3.9%)

SexMale 40 (31.3%) 40 (31.3%) 17 (13.3%) 18 (14.1%) 11 (8.6%) 2 (1.6%)Female 41 (29.9%) 25 (18.3%) 29 (21.2%) 22 (16.1%) 14 (10.2%) 6 (4.4%)

Age14–17 years 17 (47.2%) 12 (33.3%) 4 (11.1%) 0 (0.0%) 3 (8.3%) 0 (0.0%)18–24 years 25 (32.9%) 20 (26.3%) 15 (19.7%) 9 (11.8%) 7 (9.2%) 0 (0.0%)25–34 years 22 (31.0%) 16 (22.5%) 14 (19.7%) 11 (15.5%) 4 (5.6%) 4 (5.6%)35–44 years 14 (23.7%) 15 (25.4%) 6 (10.2%) 13 (22.0%) 9 (15.3%) 2 (3.4%)45–54 years 3 (15.0%) 2 (10.0%) 6 (30.0%) 6 (30.0%) 1 (5.0%) 2 (10.0%)55–64 years 0 (0.0%) 0 (0.0%) 1 (33.3%) 1 (33.3%) 1 (33.3%) 0 (0.0%)

Pain severity: Children

None of the time Once or twice A few times Fairly often Very often

DiagnosisB-thal major 30 (41.1%) 25 (34.3%) 12 (16.4%) 4 (5.5%) 2 (2.7%)B-thal intermedia 1 (20.0%) 2 (40.0%) 1 (20.0%) 1 (20.0%) 0 (0.0%)

SexMale 20 (40.8%) 15 (30.6%) 9 (18.4%) 3 (6.1%) 2 (4.1%)Female 25 (46.3%) 17 (31.5%) 6 (11.1%) 3 (5.6%) 3 (5.6%)

Age5–7 years 14 (48.3%) 8 (27.6%) 4 (13.8%) 1 (3.5%) 2 (6.9%)8–10 years 18 (50.0%) 11 (30.6%) 4 (11.1%) 2 (5.6%) 1 (2.8%)11–12 years 9 (40.9%) 6 (27.3%) 4 (18.2%) 2 (9.1%) 1 (4.6%)13–14 years 4 (25.0%) 7 (43.8%) 3 (18.8%) 1 (6.3%) 1 (6.3%)

aPain severity as self-reported on the SF-36v2 quality of life health survey by participants aged 141 years.

bPain severity as parent-reported on the PF-28 child health questionnaire (CHQ) for participants <14 years.

Figure 1. Bodily pain: thalassemia cohort by country versus U.S. norms. (A)Adults (Bodily pain (BP) scale of the SF-36v2 quality of life health survey com-pared with U.S. adult norms [16]. Higher scores indicate higher QOL (i.e., lesspain). There are no normed data available for age <18.). (B) Children (Bodily pain(BP) scale of the PF-28 child health questionnaire (CHQ) compared with U.S.pediatric norms [17]. Higher scores indicate higher QOL (i.e., less pain).). Asnorms differ by culture, only thalassemia patients in North America are comparedwith U.S. norms; age trends are similar for patients in the U.K., but scores areconsistently lower than in the North American group.

letters

American Journal of Hematology 369

assess the effect of pain (bodily pain scale of the SF-36 and CHQ) on qual-

ity of life (all other SF-36 and CHQ scales) and anxiety and depression

(HADS scales).

1New England Research Institutes, Watertown, Massachusetts2Hematology/Oncology, Children’s Hospital and

Research Center Oakland, California3Hematology, Children’s Hospital of Philadelphia, Pennsylvania

4Hematology, Children’s Hospital Los Angeles, California5Hematology/Oncology, Hospital for Sick Children, Toronto, Ontario

6Hematology/Oncology, Weill Cornell Medical College University, New York7Hematology, Children’s Memorial Hospital, Chicago, Illinois

8Hematology, Children’s Hospital Boston, Massachusetts9Hematology/Oncology, Children’s Healthcare of Atlanta at Scottish Rite, Georgia

10Hematology/Oncology, Toronto General Hospital, Canada11Liberal Studies, California State University, San Marcos, California

Contract grant sponsor: NIH-NHLBI Cooperative Agreement; Contract grantnumber: U01 HL065238

*Correspondence to: Felicia Trachtenberg, New England Research Institutes,9 Galen Street, Watertown, MA 02472E-mail: ftrachtenberg@neriscience.comConflict of interest: Nothing to report.

Published online 8 February 2010 in Wiley InterScience(www.interscience.wiley.com).

DOI: 10.1002/ajh.21670

References1. Forget B, Cohen A.Thalassemia syndromes. In:Hoffman R,Benz E,Shattil S,

editors. Hematology: Basic Principles and Practices, 4th ed. Philadelphia,PA:Elsevier, Churchill, Livingstone; 2005. pp 557–598.

2. Cohen AR, Galanello R, Pennell DJ, et al. Thalassemia. Hematology Am SocHematol Educ Program 2004;14–34.

3. Modell B, Khan M, Darlison M, et al. Improved survival of thalassaemia majorin the UK and relation to T2* cardiovascular magnetic resonance. J Cardio-vasc Magn Reson 2008;10:42.

4. Desigan S, Hall-Craggs MA, Ho CP, et al. Degenerative disc disease as acause of back pain in the thalassaemic population: A case-control study usingMRI and plain radiographs. Skeletal Radiol 2006;35:95–102.

5. Fucharoen S, Ketvichit P, Pootrakul P, et al. Clinical manifestation of beta-tha-lassemia/hemoglobin E disease. J Pediatr Hematol Oncol 2000;22:552–557.

6. Munn RK, Kramer CA, Arnold SM. Spinal cord compression due to extrame-dullary hematopoiesis in beta-thalassemia intermedia. Int J Radiat Oncol BiolPhys 1998;42:607–609.

7. Onur O, Sivri A, Gumruk F, Altay C. Beta thalassaemia: A report of 20 chil-dren. Clin Rheumatol 1999;18:42–44.

8. Otrock ZK, Azar ST, Shamseddeen WA, et al. Intravenous zoledronic acidtreatment in thalassemia-induced osteoporosis: Results of a phase II clinicaltrial. Ann Hematol 2006;85:605–609.

9. Pakbaz Z, Treadwell M, Yamashita R, et al. Quality of life in patients with tha-lassemia intermedia compared to thalassemia major. Ann NY Acad Sci 2005;1054:457–461.

10. Shaligram D, Girimaji SC, Chaturvedi SK. Psychological problems and qualityof life in children with thalassemia. Indian J Pediatr 2007;74:727–730.

11. Vichinsky EP. The morbidity of bone disease in thalassemia. Ann NY AcadSci 1998;850:344–348.

12. Vogiatzi MG, Macklin EA, Fung EB, et al. Bone disease in thalassemia: A fre-quent and still unresolved problem. J Bone Miner Res 2009;24:543–557.

13. Messina G, Colombo E, Cassinerio E, et al. Psychosocial aspects and psy-chiatric disorders in young adult with thalassemia major. Intern Emerg Med2008;3:339–343.

14. Payne KA, Desrosiers MP, Caro JJ, et al. Clinical and economic burden ofinfused iron chelation therapy in the United States. Transfusion 2007;47:1820–1829.

15. Payne KA, Rofail D, Baladi JF, et al. Iron chelation therapy: Clinical effective-ness, economic burden and quality of life in patients with iron overload. AdvTher 2008;25:725–742.

16. Ware JE, Kosinski M, Bjorner BB, et al.User’s Manual for the SF-36v2 HealthSurvey, 2nd ed. Lincoln, RI:Quality Metric Incorporated; 2007.

17. Landgraf JM, Abetz L, Ware JE.Child Health Questionnaire (CHQ): A User’sManual, Third Printing. Lincoln, RI:Quality Metric Incorporated; 2006.

18. Scalone L, Mantovani LG, Krol M, et al. Costs, quality of life, treatment satis-faction and compliance in patients with beta-thalassemia major undergoingiron chelation therapy: The ITHACA study. Curr Med Res Opin 2008;24:1905–1917.

19. Snaith R, Zigmond A.The Hospital Anxiety and Depression Scale Manual.London, UK:nferNelson; 1994.

A comprehensive, simple molecular assay of common deletionsand mutations causing a-thalassemia in Southeast Asia andsouthern ChinaJingzhong Liu,1* Xingyuan Jia,2 Ning Tang,3 Lirong Wang,2 Han Han,1 Ren Cai,3

Qingtao Wang,1 and Bai Xiao1

a-Thalassemia is the most common recessively inherited hemoglobin

disorder [1]. In Southeast Asia and southern China, most a-thalasse-

mia is caused by deletion of one (2a4.2/, 2a3.7/; termed a-thalassemia-

2) or both (22SEA, 22THAI; termed a-thalassemia-1) of the two func-

tional a-globin genes (see Fig. 1) [2,3] and by nondeletional mutations

[4]. A rapid assay is needed to address the diagnostic challenges of

identifying both deletions and mutations. Hung et al. [5] reported a

molecular assay to detect deletions in a single PCR based on DHPLC,

but the assay can distinguish only two genotypes, 2a3.7/aa and 2a4.2/

aa. Here, we show a comprehensive and simple diagnostic assay for

detecting all genotypes composed by the most common three muta-

tions (CS, QS, and WS), three deletions (2a4.2/, 2a3.7/, and 22SEA),

and wild-type allele (aa). In this assay, we combined a duplex PCR/

DHPLC with an improved version of Hung’s single-tube PCR followed

by DHPLC analysis. Using this assay, we detected 142 samples and

found that this technique was 100% concordant with results of current

standard methods. This comprehensive molecular assay can be used

to fully diagnose a-thalassemia in this geographical area.

PCR primers were designed as shown in Fig. 1. We performed a duplex

PCR with primers of P1/P2 and P3/P4. The products were analyzed by

DHPLC at 508C to detect the 22SEA/and nondeletional a-globin alleles (aa/

or aTa/). On the DHPLC profiles, an absorption peak appeared at 1.0 ± 0.1

Figure 2. Quality of Life by pain severity (excludes patients from the U.K. because of significant country effects on these scales; however, U.K. patients show similartrends). (A) PCS and MCS summary scales of the SF-36 in adults/adolescents (Physical Component Summary and Mental Component Summary scales of the SF-36v2health survey administered to participants aged 141 years. Higher scores indicate higher QOL.). (B) Anxiety and Depression in adults/adolescents (Hospital Anxiety andDepression Scale (HADS) administered to participants aged 141 years. Higher scores indicate higher levels of anxiety and depression.). (C) PhS and PsS summaryscales of the CHQ in children (Physical Summary and Psychosocial Summary scales of the PF-28 child health questionnaire (CHQ) for participants <14 years. Higherscores indicate higher QOL.).

letters

370 American Journal of Hematology

min in all samples with 22SEA deletion allele (Fig. 2A, 1–3), indicating that

they were positive for this allele. An absorption peak appeared at 5.1 ± 0.1 min

in all samples expressing aa, aCSa, or aQSa alleles (Fig. 2A, 2–7), indicating

that they were all positive for a nondeletional allele (aa/ or aTa/). Both peaks

appeared simultaneously in Samples 2 and 3, in agreement with the known

genotype (22/aa and 22/aCSa). Potential mutations in the duplex PCR prod-

uct can be rapidly detected according to peak profile differences using DHPLC

at 63.88C. The DHPLC profiles from normal DNA samples (aa/aa) showed

one peak appeared after 4.8 min (Fig. 2B, 4). Homoduplex double peaks were

seen at 4.7–4.8 min in the CS/aa sample, and heteroduplex double peaks

appeared simultaneously at 4.1–4.2 min (Fig. 2B, 2). A single peak appeared

at 4.8 ± 0.1 min in the QS/aa samples, and the heteroduplex double peak

appeared simultaneously at 4.5–4.6 min (Fig. 2B, 3). Three continuous peaks

appeared at 4.25–4.45 min in the WS/aa samples (Fig. 2B, 1).

We performed detection of 2a3.7 and 2a4.2 using an improved single-

tube PCR/DHPLC at 648C. The improvement of the single-tube PCR for

detecting the 2a3.7 and 2a4.2 was that primers were redesigned to

amplify a shorter amplicon (293 bp) than the original (353 bp) [5]. The

reverse primer P6 sequences were 50-ATC GAC TCC AGC GGG ATC-30.On the DHPLC profile for the products of the improved single-tube PCR,

wild-type samples displayed four peaks, two heteroduplex and two homodu-

plex. The heteroduplex formation of the PCR amplicons is due to the signifi-

cant similarity between a1 and a2. The stutter peaks (heteroduplex) were

eluted much earlier in the chromatographic process than the main peaks

(homoduplex). According to the stability ranking described previously [6], it

would be expected that the homoduplex with the ATC pair (a1) would have

a reduced retention time compared with the GCT pair (a2). Samples were

eluted from the column, and peaks were seen at 5.22 min for the a1 allele

and at 5.37 min for the a2 allele. These two peaks were clearly distinct at

648C. The methodology provided not only a qualitative but also a quantita-

tive comparison between the two homoduplexes, allowing for the differentia-

tion of the genotypes. Samples from wild-type, 2a3.7/aa, and 2a4.2/aa

individuals displayed different patterns and could be easily identified.

Although the aa/aa, 22/aa, and compound heterozygous 2a3.7/2a4.2

genotypes had similar patterns, the dissociation curve (DC) analysis of the

duplex PCR products could distinguish the three genotypes very clearly.

Likewise, the 2a3.7/2a3.7 and the 2a3.7/22 DNA, which had the same

DHPLC pattern, could also be distinguished by the DC analysis of the

duplex PCR products. This was also true when comparing 2a4.2/2a4.2

and 2a4.2/22 genotypes. Therefore, all of the major genotypes (2a3.7/aa,

2a4.2/aa, 2a3.7/22SEA, 2a4.2/22SEA, 22SEA/22SEA, 2a3.7/2a3.7,

2a3.7/2a4.2, 2a4.2/2a4.2, 22SEA/aa, and aa/aa) were easily distin-

guishable based on the two PCRs and their DHPLC and DC profiles.

We collected 142 DNA samples expressing various a-thalassemia-associ-

ated deletions and mutations from Liuzhou city in southern China from January

2007 to February 2009 for this study. Using this assay, we detected 34 cases

of a-thalassemia-2 carriers, 5 cases of a-thalassemia-2 homozygous, 23 cases

of a-thalassemia-1 carriers, 2 cases of a-thalassemia-1 homozygous, 9 cases

of nondeletional Hb H, 16 cases of genetic compounds of a-thalassemia-1 and

a-thalassemia-2, 43 cases of normal DNA samples, and 10 cases of nondele-

tional mutation carriers: 4 cases of aCSa/aa, 3 cases of aQSa/aa, and 3 cases

of aWSa/aa. These results were in accordance with those from the three- and

four-tube PCR/DC, the RDB, and the sequencing methods. The blinded study

results evaluating 142 DNA samples using the new assay were 100% concord-

ant with the genotypes detected using the current standard methods.

Populations in Southeast Asia and southern China have a high frequency of

a-thalassemia because of a-globin gene mutations and/or deletions. Molecular

diagnosis, prenatal diagnosis, and large-scale population screenings are very

important for decreasing the incidence of a-thalassemia. In this study, we com-

bined the duplex PCR/DHPLC with an improved version of Hung’s single-tube

PCR to achieve a complete diagnosis for a-thalassemia resulting from the three

deletions and three mutations in the geographical area. All of the major geno-

types were then easily distinguishable based on the two PCRs and their

DHPLC and DC profiles. To generate reproducibly good DHPLC patterns for

the 2a3.7 and 2a4.2 deletions, besides shortening the amplicon, we optimized

the thermocycling program, added an additional denaturing step and a renatur-

ing step following the final elongation step. More typical patterns could be

observed after the additional denature/renature steps. Integrative diagnosis can

be carried out on deletional a-thalassemia by comprehensively observing the

DC patterns of the duplex PCR products and the DHPLC patterns of the

improved single-tube PCR for the 2a3.7 and 2a4.2 deletions. Both DHPLC

and DC can be used to rapidly obtain an accurate diagnosis from the duplex

PCR and may be performed according to the equipment available to a given

laboratory. The method reported here is simple, rapid, accurate, semiautomatic,

and cost effective, which makes it suitable for large-scale screening.

Acknowledgments

The authors thank Dr. Zhang Ze-Yun for assistance with the DHPLC analysis.

Figure 1. Designing the primers. The upper part shows the location of the primersP3 and P4 in the a-gene cluster and the three deletions. The three mutations, CS,QS, and WS, are located in the 206-bp amplicon from primer pair P1 and P2, illus-trated in the middle part. The primer pair P5 and P6 was used to amplify 293-bpsegments from both the a1 and a2 Z boxes in a single-tube PCR reaction. Theamplicons contain three nucleotides, which differ from each other.

Figure 2. A: DHPLC profiles for the duplex PCR products of seven DNA samples of known genotypes at 508C. B: DHPLC profiles at 63.88C for the duplex PCR products ofDNA from four known mutation carriers and one negative control. [Color figure can be viewed in the online issue, which is available at www.interscience.wiley.com.]

letters

American Journal of Hematology 371

Primary high-grade ocular adnexal lymphoma:Clinicopathological characteristics and prognosticfactors of a single-centre seriesCristina Michaela Precupanu,1 Patricia Validire,1 Christine Levy,2 Corine Plancher,3 Anne Vincent-Salomon,4

Remi Dendale,5 Bernard Asselain,4 Vincent Servois,6 Laurence Desjardins,2Manuela Zanni,1

Rafika Sahli,1 Livia Lumbroso-le Rouic,2 and Didier Decaudin1,7*

The majority of ocular adnexal lymphomas (OAL) occurring in ophthal-

mologic sites are low-grade B-cell non-Hodgkin lymphomas, and about

15% of all cases are high-grade. Although numerous reports have

already studied the features of low-grade OAL, high-grade OAL have

been usually described as part of various published series of OAL

without specifically focusing on these histopathologic subtypes of

lymphoma. The aim of this study was therefore to retrospectively

review a single-centre series of 15 patients with high-grade B-cell OAL

treated at the Institut Curie. The main clinical and laboratory features

of our selected patients, i.e., median age (66 years), PS greater than 1

(20%), stage IV (53%), bone marrow involvement (17%), more than one

extranodal site (27%), and elevated serum LDH (14%), were not signifi-

cantly different from those of nodal non-Hodgkin’s lymphomas (NHL),

except for a lower proportion of OAL patients with elevated LDH level.

Similarly, the prognosis of these patients was not different, with a 5-

year overall survival of 54%. When high-grade OAL were compared to

a historical series of low-grade OAL, a lower rate of conjunctival

involvement was observed, but both types presented a majority of

orbital and/or lacrimal gland involvement as well as bilateral ophthal-

mological tumors. The recommended treatment for high-grade B-cell

OAL can be the same as that for nodal CD20-positive NHL, namely a

combination of rituximab and CHOP-like regimen.

OAL constitute primary or secondary manifestations of malignant lymphoid

disease arising in the orbit, salivary gland, conjunctiva, and/or eyelid. The

majority of these lymphomas occurring in ophthalmologic sites are low-

grade B-cell non-Hodgkin lymphomas (NHL)[1]. According to the World

Health Organization classification of lymphoid neoplasms [2], extranodal

marginal zone B-cell lymphomas of MALT-type ‘‘Mucosa Associated Lym-

phoid Tissue’’ represent the most common subtype of OAL [1,3,4]. Diffuse

large B-cell and follicular lymphomas constitute the second and third most

frequent subcategories of lymphomas occurring in the periocular sites,

respectively. Inversely, the endemic and nonendemic forms of Burkitt’s

lymphomas rarely involve the ocular adnexa in children [5,6] or adults [7].

Non-B-cell lymphomas are rare and mainly represent secondary manifes-

tations of a systemic T-cell lymphoproliferation such as mycosis fungoides

[8]. Finally, Hodgkin’s lymphoma occurs extremely rarely in the ocular

adnexa and, like T-cell lymphomas, is usually a secondary manifestation of

advanced systemic disease [9].

TABLE I. Patient Characteristics and Outcome

Pts Histo. Bcl-2Age

(years) Sex B PSOphth.site Bi. N BM LDH Stage NES IPI 1st TT Resp.

Relapse(months) Status (months)a

1 DBLCL NP 73 F Yes 2 LG No 0 Yes NP IV 2 3 RT P / D (lymphoma) (7)2 DBLCL NP 66 M No 0 Orbit No 0 No NP IE 1 1 CT1 RT RC No D (not lymphoma) (169)3 DBLCL NP 37 M No 0 Orbit No 0 No NP IE 1 1 CT1 RT RC 3 months D (lymphoma) (22)4 DBLCL NP 71 F No 0 LG No 0 No <N IE 1 2 CT1 RT RC No D (not lymphoma) (210)5 DBLCL NP 49 F No 1 LG No 1 No >4 N IV 1 3 CT P / D (lymphoma)(24)6 DBLCL NP 50 M No 0 Orbit No 0 No <N IE 1 0 CT 1 RT RC No Lost to FU (11)7 DBLCL 1 49 M No 0 Orbit No 0 No <N IE 1 0 CT1 RT RC No A (151)8 DBLCL NP 61 M No 0 Orbit No 0 No <N IE 1 1 CT1 RT RC No A (143)9 DBLCL 2 87 M No 2 Orbit No 0 No <N IV 4 3 CT1 RT P / D (lymphoma) (5)

10 DBLCL 1 82 F No 1 Orbit No 1 No >2 N IV 0 3 CT1 RT P / D (lymphoma) (2)11 DBLCL 1 76 M No 0 Conj. No 0 No <N IE 1 1 CT RC No A (40)12 DBLCL NP 60 F No 0 Orbit Yes 0 No <N IV 3 3 CT RC No A (35)13 DBLCL 2 81 M No 2 Orbit No 1 Yes <N IV 2 2 CT RC No A (20)14 DBLCL 1 82 M No 1 Conj. No 1 No <N IV 2 2 CT RC No D (not lymphoma) (7)15 DBLCL 2 58 F No 0 LG Yes 0 No <N IV 3 2 CT RC No A (26)

Pts, patients; Histo., histologic subtype of lymphoma; B, B-symptoms; PS, performance status; Ophth. Site, ophthalmologic site; Bi., bilateral ophthalmologic involvement;N, nodal involvement; BM, bone marrow involvement; NES, number of extranodal sites including the ophthalmologic involvement; TT, first line of treatment; Resp.,response to first line of treatment; DLBCL, diffuse large B-cell lymphoma; NP, not performed; M, male; F, female; LG, lacrimal gland; conj., conjunctival site; N, normal;RT, radiotherapy; CT, chemotherapy; CR, complete response; P, progression; A, alive; D, dead; FU, follow-up.aFrom the initial diagnosis of OAL to the date of last follow-up.

1Basic Medical Research Center, Beijing Chaoyang Hospital Affiliateof Capital Medical University, Beijing, China

2Department of Medical Genetics, Institute of Basic Medical Sciences,Chinese Academy of Medical Sciences and Peking Union Medical College,

Beijing, China3Division of Medical Genetics, Liuzhou Municipal Maternity and Child Healthcare

Hospital, Liuzhou City, Guangxi Autonomous Region, ChinaContract grant sponsor: Natural Science Foundation of Beijing, China;

Grant number: 7092034*Correspondence to: Jingzhong Liu, Basic Medical Research Center, Beijing

Chaoyang Hospital Affiliate of Capital Medical University, 8, Gongtinanlu,Chaoyang District, Beijing 100020, China

E-mail: liujingzhong090525@hotmail.com or jiaxy@hotmail.comConflict of interest: Nothing to report.

Received for publication 6 January 2010; Revised 24 January 2010;Accepted 28 January 2010

Published online 8 February 2010 in Wiley InterScience(www.interscience.wiley.com).

DOI: 10.1002/ajh.21671

References1. Higgs DR, Vickers MA, Wilkie AO, et al. A review of the molecular genetics of

the human alpha-globin gene cluster. Blood 1989;73:1081–1104.2. Chong SS, Boehm CD, Cutting GR, Higgs DR. Simplified multiplex-PCR diag-

nosis of common Southeast Asian deletional determinants of alpha-thalasse-mia. Clin Chem 2000;46:1692–1695.

3. Liu J, Ou-Yang C, Wang L, et al. Detection of three common deletional alpha-thalassemia determinants in southern China by a single-tube multiplex poly-merase chain reaction method. Hemoglobin 2004;28:39–44.

4. Cai R, Liu J,Wang L, et al. Study onmolecular epidemiology of the alpha-thalassemiain LiuzhouCity,Guangxi AutonomousRegion, China.Hemoglobin 2004;28:325–333.

5. Hung C, Lee C, Chen C, et al. Molecular assay of -alpha (3.7) and -alpha(4.2) deletions causing alpha-thalassemia by denaturing high-performanceliquid chromatography. Clin Biochem 2007;l40:817–821.

6. Xiao W, Oefner PJ. Denaturing high-performance liquid chromatography: Areview. Hum Mutat 2001;17:439–474.

letters

372 American Journal of Hematology

Although a very large majority of reports have focused on MALT-type lym-

phomas, with well defined initial histopathologic and clinical features and

prognostic factors, high-grade OAL have been usually described as part of

various published series of OAL. However, it is unknown whether these lym-

phomas constitute a peculiar entity of high-grade lymphomas in terms of ini-

tial characteristics, in addition to ophthalmologic involvement, and in terms

of prognostic factors and outcome. In particular, the differences between

these high-grade OAL and other nodal and extranodal high-grade NHL have

not been clearly defined. To address these issues, we therefore retrospec-

tively reviewed all patients treated at the Institut Curie for high-grade OAL.

The data of 18 patients were reviewed and 15 cases were included in the

study on the basis of the selection criteria. One patient was excluded from

the study because of intraocular lymphoma and, to reduce the heterogeneity

of the studied population, two cases of Burkitt’s lymphomas have not been

included in the final analyses. All patients were diagnosed as having malig-

nant non-Hodgkin’s lymphoma. Pathologic review according to the WHO

classification showed all cases were of diffuse large B-cell lymphomas. Bcl2

positivity (>50% positive tumor cells) was observed in four of the seven

patients in whom Bcl-2 protein analysis was performed (57%). The major

symptoms listed in decreasing order of frequency were exophthalmia in

eight patients (53%), ptosis in four patient (27%), conjunctival mass in four

patients (27%), decreased visual acuity in four patients (27%), diplopia in

three patients (20%), and visible orbital mass in two patients (13%). Oph-

thalmologic sites were intraorbital in eight cases (53%), lacrimal gland in

four patients (27%), conjunctival in one case (7%), and both lacrimal gland

1 conjunctival in one patient (7%). Two patients had bilateral ophthalmologic

sites (13%). Initial characteristics of the overall selected population, pre-

sented in Table I, were: median age: 66 years (range: 37–87 years); M/F

ratio: 1.5; B symptoms in one patient (7%); performance status (PS) �2 in

three patients (20%); nodal involvement in four patients (27%); stage IE and

IV in seven (47%) and 8 (53%) patients, respectively; nonophthalmologic vis-

ceral involvement was observed in four patients (27%), namely bone marrow

involvement in 2/12 patients (17%), gastric involvement in two cases (13%),

sinus involvement in two cases (13%), and splenic involvement in one patient.

Serum LDH and ß2 microglobulin levels were evaluated in 2/14 cases (14%)

and 4/8 cases (50%), respectively. According to the International Prognostic

Index (IPI), six patients (40%) were classified in the low risk group (IPI 0–1),

five cases (33%) in the low to intermediate risk group (IPI 5 2), and four

patients (27%) in the intermediate to high risk group (IPI 5 3).

All but one patient received chemotherapy (CT): nine cases received the

CHOP regimen with rituximab in five cases, six patients received the GCD2

regimen, two cases received the R-mini-CEOP regimen, and two patients

received an oral combination of etoposide and chloraminophene (Table I).

Nine patients received radiotherapy (RT) including ophthalmologic RT in

Figure 1. (A) Overall survival of the study population; (B) Overall survival of the study population according to clinical stage; (C) Overall survival of the study populationaccording to PS; (D) Overall survival of the study population according to bone marrow involvement; (E) Overall survival of the study population according to LDH level;(F) Overall survival of the study population according to IPI score.

letters

American Journal of Hematology 373

eight cases. On completion of first-line treatment, 10/15 patients were in

complete remission (67%), two were in partial response (13%), and three

patients progressed during treatment. Among the 12 patients in first com-

plete or partial remission, two cases relapsed (17%) at 3 and 6 months,

respectively. With a mean follow-up of 168 months (range 7–171 months),

and of the 12 patients in first complete or partial remission, eight patients

were still alive in CR in whom two patients were lost to follow-up and seven

patients had died in whom five from lymphoma (Table I). The 5-year overall

survival of the 15 patients evaluated was 54% (Fig. 1A).

Univariate analysis was performed to identify poor prognostic factors for

OS. All factors correlated with a lower OS were a clinical stage of III or IV

(P 5 0.002), a PS > 1 (P 5 0.02), bone marrow involvement (P 5 0.02),

and elevated serum LDH level (P 5 0.04), while other clinical and laboratory

factors (age, sex, B symptoms, and ESR) were not demonstrate to influence

overall survival (Table II and Fig. 1B–E). We have observed a trend of signif-

icance for a number of extranodal sites greater than 1 (P 5 0.08). Finally,

the IPI score was significantly prognostic for lower OS (P 5 0.02) (Fig. 1F).

Because of the small number of patients, multivariate analysis was not per-

formed in this patient population.

We report the analysis of a single-centre series of 15 patients with diffuse

large B-cell OAL in terms of initial characteristics and patient outcome. The

aim of this study was to evaluate whether or not ophthalmologic involvement

has a significant impact on the clinical presentation and prognosis of the dis-

ease, particularly compared to high-grade nodal lymphomas and low-grade

OAL. Considering the main clinical and laboratory features of this series, i.e.,

median age (66 years), PS greater than 1 (20%), stage IV (53%), bone marrow

involvement (13%), more than one extranodal site (40%), and elevated serum

LDH (17%), no significant difference was observed in comparison to the 3,273

patients reported in the International Non-Hodgkin’s Lymphoma Prognostic

Factors Project [10], except for the proportion of patients with elevated LDL

level that represented 52% of evaluable patients. Moreover, the proportions of

patients in the four prognostic groups of the IPI score were 40%, 27%, 33%,

and 0% in our series, and 35%, 27%, 22%, and 16% in the International report.

Our data therefore suggest that the initial characteristics of OAL patients are

not different from those of high-grade nodal lymphoma patients. This observa-

tion is confirmed by the fact that the 5-year overall survival in our series was

54% versus 52% in the International Non-Hodgkin’s Lymphoma Prognostic Fac-

tors Project [10]. Finally, It is noteworthy that all prognostic factors identified for

poor OS were those involved in the definition of the IPI previously defined in

high-grade NHL [10]. However, no multivariate analyses could be performed

because of the small number of patients and our conclusions should therefore

be mentioned cautiously. One large series of primary or secondarily ocular

adnexal DLBCL has been reported, showing a proportion of stage IV disease in

35% of primary OAL, and an overall survival of 56% of all cases, that results

being concordant to our data [11]. In this study, almost all patients received CT

and, as in our report, local disease correlates with better outcomes [11].

Another question was addressed by our study, i.e., are ophthalmologic

sites of high-grade OAL different from those of low-grade OAL? In a

previous report [1], we showed that conjunctival involvement was preferen-

tially observed in low-grade OAL (42% versus 7% for high-grade lymphoma),

and this was confirmed by the present series (12%). Eyelid involvement was

observed in 0–44% of patients in the various published series [4], with a

mean of about 10%. However, this marked heterogeneity is probably due to

different patient selection criteria, such as histologic subtype of lymphoma or

stage of disease. Finally, with 18% of cases with bilateral ophthalmologic

involvement, this observation is concordant with published data [4].

In most reported series, regardless of the subtype of NHL, patients were

predominantly treated with RT alone and only a small proportion was treated

by RT plus CT or CT alone. Treatment consisted of single-agent CT such as

chlorambucil or fludarabine for low-grade lymphoma, and combined CT regi-

mens such as CHOP (cyclophosphamide, doxorubicin, vincristine, and pre-

dnisone) or CHOP-like protocols for high-grade lymphomas [4]. In our study,

treatments consisted of RT alone (one patient), CHOP-like regimen in 12

patients, and oral etoposide or chlorambucil in two patients. It is unclear

whether or not specific CT is warranted, but the rituximab-CHOP regimen

appears to be the best treatment option [12].

In conclusion, this study of a homogeneous population of high-grade OAL

patients showed that the clinical and laboratory features of these lymphomas

were not different from those of high-grade nodal NHL, and that ophthalmo-

logic involvement per se does not have a prognostic impact on the outcome

of these patients. However, the management of these patients requires a

multidisciplinary approach by a team composed of hematologists, radiothera-

pists, and ophthalmologists, particularly to define the impact of OAL on the

eye(s) and visual function at diagnosis.

Patients and Methods

Patient selection. This study comprised all patients diagnosed with high-

grade ocular adnexal lymphoma and treated at the Institut Curie between July

1981 and August 2007. Patients with high-grade lymphoma and a primary

ophthalmologic site without intraocular involvement were included in the study.

The diagnosis of malignant lymphoma was established in all cases on lymph

node or extranodal (ophthalmologic) biopsy, according to the World Health

Organization (WHO) classification of hematopoietic and lymphoid tissues [2].

The tissue biopsy specimens had been fixed in 10% formaldehyde solution

and embedded in paraffin. Conventional histologic stains included hematoxylin

and eosin. Immunohistochemical analysis was performed using the peroxi-

dase-antiperoxidase (PAP) technique with various antibodies. Finally, a mini-

mum follow-up of 3 months was required for inclusion of the patient.

Staging. The staging evaluation included personal and family medical history,

description of the initial clinical symptoms, physical examination with deter-

mination of performance status (PS) according to the ECOG classification

[13], the presence of B-symptoms and the clinical stage of the disease at

diagnosis according to the Ann Arbor staging modified for extranodal dis-

ease [14]. Patients with ocular or choroidal involvement were considered to

have intraocular NHL and were not included in the study. Staging of the dis-

ease was completed by full blood count, ESR (erythrocyte sedimentation

rate), liver and renal function tests, serum protein electrophoresis, serum

lactate dehydrogenase (LDH) and ß2-microglobulin assay, bone marrow

biopsy, chest X-ray, CT scans and/or magnetic resonance imaging of the

orbit, CT scan of the thorax and abdomen, and fiberoptic gastroscopy in

most cases from 1990 onwards. The IPI was assessed for all cases included

[10]. All these elements recorded at the time of diagnosis were also deter-

mined at first relapse of the disease.

Treatment. Treatment modalities included RT alone, CT alone, immuno-

therapy alone, or combinations of immunochemotherapy and radiochemo-

therapy. The most commonly used CT was the CHOP (cyclophosphamide-

vincristine-doxorubicin-prednisone) regimen, with or without rituximab. Other

CT regimens were used, including R-mini-CEOP (rituximab-cyclophospha-

mide-epidoxorubicin-vinblastine-prednisone), GCD2 regimen [15], and oral

etoposide and chloraminophene combination. Ophthalmologic RT used a

TABLE II. Prognostic Factors of 5-years Overall Survival of All

High-Grade OAL

Parameters

5-year overallsurvival

N P

Age <60 years 6 0.58�60 years 9

Sex Male 9 0.58Female 6

B symptoms No 14 0.15Yes 1

PS 0-1 12 0.02�2 3

Clinical stage IE 7 0.002IV 8

Number of extranodal sites 0-1 11 0.08�2 4

Bone marrow involvement No 10 0.02Yes 2Unknown 3

ESR <30 7 0.81�30 2Unknown 6

Elevated LDH No 12 0.04Yes 2Unknown 1

IPI score 0-1 6 0.022 53 44-5 0

letters

374 American Journal of Hematology

combination of an anteroposterior field and an ipsilateral anterior oblique

field with complex treatment planning, doses ranged between 36 and 80 Gy

(mean 43 Gy) in 28 to 59 days (mean 41 days). Disease status was eval-

uated at the end of first-line treatment and at first relapse, according to the

international workshop to standardize response criteria for NHL [14].

Analysis of prognostic factors. Overall survival (OS) was determined for all

patients included and univariate analysis was performed to identify poor prognostic

factors for OS. OS was measured from the date of inclusion to the date of death

irrespective of the cause using the Kaplan-Meier method [16]. For prognostic fac-

tors, survival endpoints were analyzed using a two-tailed unstratified log rank test

for corresponding estimation of hazards ratios. Statistical analysis was performed

with BMDP software (version 7.1; BMDP statistical software, Los Angeles, CA) R

version 2.2 (http://cran.r-project.org) using a 5% level of significance, and 95%

confidence intervals were also presented when considered relevant.1Department of Clinical Hematology, Institut Curie, Paris

2Department of Ophthalmological Oncology, Institut Curie, Paris3Department of Biostatistics, Institut Curie, Paris

4Department of Tumor Biology, Institut Curie, Paris5Department of Radiotherapy, Institut Curie, Paris

6Department of Radiology, Institut Curie, Paris7Laboratory of preclinical investigation, Translational Research

Department, Institut Curie, Paris*Correspondence to: Didier Decaudin, Service d’Hematologie,Institute Curie, 26 rue d’Ulm, Paris 75.248, cedex 05, France

E-mail: didier.decaudin@curie.netConflict of interest: Nothing to report.

Published online 8 February 2010 in Wiley InterScience(www.interscience.wiley.com).

DOI: 10.1002/ajh.21673References1. Meunier J, Lumbroso L, Vincent-Salomon A, et al. Ophthalmologic and intraocu-

lar Non-Hodgkin’s lymphoma: A large single centre study of initial characteris-

tics, natural history, and prognostic factors. Hematol Oncol 2004;22:143–158.

2. Harris NL, Jaffe ES, Diebold J, et al. The World Health Organization classifi-

cation of neoplastic diseases of the haematopoietic and lymphoid tissues:

Report of the Clinical Advisory Committee Meeting, Airlie House, Virginia,

November 1997, Histopathology 2000;36:69–87.

3. Coupland SE, Krause L, Delecluse HJ, et al. Lymphoproliferative lesions ofthe ocular adnexa. Ophthalmology 1998;105:1430–1441.

4. Decaudin D, De Cremoux P, Vincent-Salomon A, et al. Ocular adnexal lym-phoma: A review of clinicopathological features and treatment options. Blood2006;108:1451–1460.

5. Edelstein C, Shields JA, Shields CL, et al. Non-African Burkitt lymphoma pre-senting with oral thrush and an orbital mass in a child. Am J Ophthalmol1997;124:859–861.

6. Weisenthal A, Streeten BW, Dubansky AS, et al. Burkitt lymphoma presentingas a conjunctival mass. Ophthalmology 1995;102:129–134.

7. Reifler DM, Warzynski MJ, Blount WR, et al. Orbital lymphoma associatedwith acquired immune deficiency syndrome (AIDS). Surv Ophthalmol 1994;38:371–380.

8. Coupland SE, Foss HD, Assaf C, et al. T-cell and T/natural killer-cell lympho-mas involving ocular and ocular adnexal tissues: A clinicopathologic, immuno-histochemical, and molecular study of seven cases. Ophthalmology 1999;106:2109–2120.

9. Jakobiec FA. Orbital Hodgkin’s disease: Clinicopathologic conference. N EngJ Med 1989;320:447–457.

10. Shipp MA, Harrington DP, Anderson JR, et al. A predictive model for aggres-sive non-Hodgkin’s lymphoma: The international non-Hodgkin’s lymphomaprognostic factors project. N Engl J Med 1993;329:987–994.

11. Madge SN, Mccormick A, Patel I, et al. Ocular adnexal diffuse large B-cell lymphoma:Local disease correlates with better outcomes. Eye 2009 [Epub ahead of print].

12. Coiffier B, Lepage E, Briere J, et al. CHOP chemotherapy plus rituximabcompared with CHOP alone in elderly patients with diffuse large-B-cell lym-phoma. N Engl J Med 2002;346:235–242.

13. Oken MM, Creech RH, Tormay DC, et al. Toxicity and response criteria ofthe Eastern Cooperative Oncology Group. Am J Clin Oncol 1982;5:649–655.

14. Cheson BD, Horning SJ, Coiffier B, et al. Report of an international workshopto standardize response criteria for non-Hodgkin’s lymphomas. NCI Spon-sored Working Group. J Clin Oncol 1999;42:1271–1278.

15. Dumont J, Charpy-Validire P, Mosseri V, et al. Multicentre combined chemo-therapy protocol for large cell advanced non-Hodgkin’s lymphoma. HematolOncol 1991;9:197–207.

16. Kaplan E, Meier P. Nonparametric evaluation from incomplete observation.J Am Stat Assoc 1958;53:457–458.

Imatinib dose escalation in 74 failure or suboptimal responsechronic myeloid leukaemia patients at 3-year follow-upMassimo Breccia,1* Fabio Stagno,2,3 Paolo Vigneri,2,3 Roberto Latagliata,1 Laura Cannella,1

Vittorio Del Fabro,2,3 Francesco Di Raimondo,2,3 and Giuliana Alimena1

We report here on the long-term efficacy of imatinib dose escalation in 74

patients after failure to imatinib conventional dose (13 with hematologic fail-

ure and 57 patients with cytogenetic resistance) or suboptimal response

(four patients for cytogenetic or molecular). Fifty-four patients received imati-

nib dose escalation from 400 to 600 mg and 20 patients to 800 mg. A major

cytogenetic response was achieved in 41 patients (72%) who escalated ima-

tinib dose for cytogenetic failure and in six patients (46%) with hematologic

failure (P 5 0.002). Complete cytogenetic response (CCyR) was achieved in

27 (37%) patients: 38% of the hematological failure patients and 42% of the

cytogenetic resistant patients (P 5 0.345). Cytogenetic suboptimal response

patients obtained complete molecular response, whereas the patient, who

escalated the dose for molecular suboptimal response at 18 months, did not

obtain improvement. After 3 years of follow-up all responding patients are in

sustained CCyR. The estimated 2-year PFS and OS is 87% and 85%, respec-

tively. In conclusion, imatinib dose escalation appears to induce sustained

responses in CML patients with cytogenetic resistance, in particular in those

with acquired resistance. In hematological failure patients, a rapid switch to

second generation TKI is instead recommended.

In 2006, European LeukemiaNet recommendations defined as failure

patients those who did not achieve complete hematologic response (CHR) or

cytogenetic response (CyR) at different time points or loss of CHR, CCyR or

acquired BCR-ABL mutations at any time. Suboptimal response was defined

as incomplete hematologic response or CyR at different time points, less than

major molecular response (MMR) or acquisition of cytogenetic abnormalities

in Ph1 cells, mutations of BCR-ABL or loss of MMR at any time [1].

Despite the beneficial effects of imatinib, some patients may develop pri-

mary or acquired resistance [2]. The IRIS study at 8-year follow-up provided

an indication of imatinib resistance: 31% of patients had discontinued imati-

nib with an estimated resistance of 16% [3]. Several mechanisms may con-

tribute to this phenomenon, including increased expression of BCR-ABL kin-

ase through gene amplification, decreased intracellular drug concentration

caused by drug efflux proteins (OCT1) or by p-glycoprotein (PgP), clonal

evolution or overexpression of Src kinases (Lyn, Hck) involved in BCR-ABL

independent activation of alternative pathways [4]. However, 40% of resist-

ance in CP is attributed to the emergence of clones expressing mutant

forms of BCR-ABL [5,6].

Imatinib dose escalation was proved as an effective strategy to overcome

imatinib resistance, and contrasting results have been reported to date [7–

10]. Some reports questioned on the opportunity of high-dose imatinib in

hematologic resistant patients and on the durability of responses obtained

with such a strategy in cytogenetic resistant patients [8,9]. Others reported

on the efficacy of dose escalation in patients who met the criteria for failure

to standard dose imatinib [7,10].

We report here the long-term efficacy of imatinib dose escalation in CP-

CML patients who demonstrated a poor response or a relapse after stand-

ard therapy in two different Italian Institutions.

The characteristics of patients treated with high-dose imatinib are

shown in Table I. Thirteen patients were classified as hematologic fail-

ure (10 primary and three secondary resistance) and 57 patients had

cytogenetic resistance (24 primary and 33 acquired resistance). All of

these patients received IM dose escalation from 400 to 600 mg (50

patients) or to 800 mg (20 patients). Four patients were suboptimal res-

ponder (three cytogenetic treated with 600 mg/day and one molecular

treated with 800 mg/day). Median time from diagnosis to therapy was 3

months (range 1–13) and median duration of imatinib was 36 months

(range 21–70).

letters

American Journal of Hematology 375

Overall after a median follow-up of 36 months, 68/74 patients (91.8%)

maintained or achieved CHR: all patients who escalated the dose for cytoge-

netic failure or suboptimal response maintained the state of CHR. A major

cytogenetic response (MCyR) was achieved in 41/57 patients (72%) who

escalated imatinib dose for cytogenetic failure and in 6/13 patients (46%)

who escalated imatinib for hematologic failure (P 5 0.002). CCyR were

achieved in 27 (37%) of the patients: 38% of the hematological failure

patients and 42% of the cytogenetic resistant patients (P 5 0.345). Three

additional patients escalated the dose for suboptimal CyR and obtained

CCyR and complete molecular response (CMR), whereas one patient esca-

lated the dose for molecular suboptimal response at 18 months and did not

obtain a CMR. Median time to CyR was 3.5 months.

CyR’s occurred in 40/53 resistant patients who escalated the dose to 600

mg and in 10/20 resistant patients who escalated to 800 mg daily (P 5

0.234).

Overall, CMR was obtained in 10 resistant patients: seven patients who

escalated the dose for cytogenetic failure and three patients who escalated

for suboptimal CyR.

Of the 13 hematological failure patients, only five achieved CCyR and all

patients were in the acquired resistance category. Of the 57 cytogenetic fail-

ure patients, 22 obtained CCyR: this response occurred in 27% of the pri-

mary cytogenetic resistance patients, and in 50% of the acquired cytoge-

netic resistant patients (P 5 0.02).

After 3 year of follow-up, all patients are in persistent CCyR. The estimated

2-year PFS and OS are 87% and 85%, respectively (Figs. 1 and 2).

As regards toxicity, most common adverse events were muscle cramps

and peripheral edema: in particular, we recorded 20% of muscle cramps in

patients treated with 600 mg compared with 30% in patients treated with

800 mg/day and 35% peripheral edema in patients treated with 600 mg

compared with 40% in patients treated with 800 mg/day (p 5 ns). We

recorded hematological toxicity only in patients treated with 800 mg/day:

anemia 2% and neutropenia 3%.

Several reports have discussed on the efficacy of imatinib dose escalation

in CML resistant patients, which is considered one of the possible therapeutic

approaches. Kantarjian et al. in 2003 [11], investigated whether increasing

the dose of imatinib mesylate may overcome drug resistance in relapsed or

refractory Ph1CML patients. Fifty-four patients entered the study and were

treated with 800 mg/daily (47 patients) or 600 mg/daily (seven patients):

among 20 patients treated for hematological resistance or relapse, 65%

improved their response but only two had a CyR; among 34 patients treated

for cytogenetic resistance or relapse, 56% obtained an improvement of CyR.

Following this report, two additional papers [8,9] described the lack of efficacy

of imatinib dose escalation. Marin et al. [8] reported on the outcome of 36

consecutive patients with CML in chronic phase in CHR in whom imatinib dos-

age was increased when they failed to achieve CCyR on an initial dose of 400

mg daily; 14 patients (39%) improved their CyR’s and seven (19%) achieved

CCyR. The responses were short lasting, and six patients (43%) lost their

best response and, at latest follow-up, only nine patients (25%) had sustained

improvement with five of them (14%) remaining in CCyR.

Zonder et al. [9] reported on 20 patients who had their dose increased (14

to 800 mg/day and 2 to 600 mg/day) because of progressive disease

(mostly clonal evolution) or inadequate CyR after at least 1 year of therapy.

Six patients had MCyR (three complete and three partial) following dose

increase and two others had MinCR. Two patients with clonal evolution tran-

siently lost the additional clonal aberrations. Almost all of the responses

occurred within 6 months and typically lasted 3–6 months. However, three

patients had persisting MCyR of >18 months duration. Also these results

suggested that the majority of patients, who receive dose escalation, do not

have a sustained meaningful response.

Kantarjian et al. [7] reported on the responses and survival of 106 patients

with newly diagnosed CP-CML who were enrolled in the International

Randomized Study of Interferon and STI571 (IRIS) trial, who began treat-

ment with imatinib at a dose of 400 mg daily, and who subsequently under-

went dose escalation to either 600 mg or 800 mg daily. Reasons for dose

escalation were evaluated retrospectively based on two sets of criteria: the

IRIS protocol-defined criteria and the European LeukemiaNet (ELN) recom-

mendations. The rates of freedom from progression to accelerated phase or

blast phase and overall survival were 89% and 84%, respectively, at 3 years

after dose increase. A CyR was obtained in 42% of patients who had their

dose escalated based on protocol criteria and in 38% of patients whom had

their dose escalated according to the ELN recommendations. The results of

TABLE I. Clinical Features of Patients before Dose Escalation

Features No. patients 74

Age (years)median 50range 19–85

SexM 52F 22

WBC (3 109/L) 4.5 (3.8–6.2)Plts (3 109/L) 220 (180–350)Hb (gr/dl) 13 (11.8–15)SokalLow 41Intermediate 24High 9

Type of bcr transcriptb2a2 36b3a2 38

Treatment historyIFN (late CP) 22Only imatinib (early CP) 52

Cause of imatinib dose escalationPrimary resistance 34-Hematologic 10-Cytogenetic 24

Secondary resistance 36-Hematologic 3-Cytogenetic 33

Suboptimal response 4Imatinib dose escalation from400 mg to 600 mg 54400 mg to 800 mg 20

Figure 1. Kaplan Meier plot of overall survival.

Figure 2. Kaplan Meier plot of progression free survival.

letters

376 American Journal of Hematology

this study provided reasonable evidence for the benefit of imatinib dose

escalation in patients with CP-CML who did experience a suboptimal CyR or

a cytogenetic relapse. Furthermore, Jabbour et al. [10] in 2009 reported on

the long-term efficacy of imatinib dose escalation in 84 patients with CP-

CML who met the criteria of failure. Twenty-one patients with hematologic

failure and 63 with cytogenetic failure had their imatinib dose escalated from

400 to 800 mg daily (n 5 72) or from 300 to 600 mg daily (n 5 12). CCyR

were achieved in 40% of patients, including 52% of those with cytogenetic

failure and 5% of those with hematologic failure. The estimated 2- and 3-

year EFS and OS rates were 57% and 47%, and 84% and 76%, respec-

tively. Responses were long lasting with 88% of patients with MCyR main-

taining their response beyond 2 years. Our experience confirms that imati-

nib dose escalation was effective in patients with cytogenetic failure, partic-

ularly in those with acquired resistance as compared with patients with

primary cytogenetic or hematological resistance. Toxicity was limited and

high doses of imatinib were generally well tolerated. We used dose escala-

tion in only one patient with suboptimal molecular response at 18 months

of imatinib, and we did not observe molecular improvement: Rea et al. [12]

reported 38% of molecular improvement in patients who escalated the

dose for molecular suboptimal response and concluded that, in this subset

of patients, dose escalation seems to be less effective than in cytogenetic

suboptimal response and the potential of second generation inhibitors

should be evaluated. In conclusion, imatinib dose escalation can induce

sustained responses in a subset of patients with cytogenetic failure or

acquired resistance after a previous CyR to standard-dose imatinib,

whereas it appears less effective in hematological failure patients or in

molecular suboptimal responders. The availability of second generation

TKI should thus be taken into account as early intervention in these latter

categories of patients.

Methods

Patient features. Between 1999 and 2006, 433 CP-CML patients were

treated with imatinib at standard dose, after interferon failure (20) or as first

line therapy (54), in two different Institutions.

Responses to imatinib standard dose were defined according to European

LeukemiaNet recommendations (1): failure was considered the lack of CHR

at 3 months, and of CyR at 6 months, the attainment of less than partial

CyR at 12 months, of less than complete CyR at 18 months, or the loss of

CHR, CCyR or acquisition of BCR-ABL mutations at any time. Suboptimal

response was defined by incomplete hematologic response at 3 months,

less than partial CyR at 6 months, less than complete CyR at 12 months

and less than MMR at 18 months, or acquisition of cytogenetic abnormalities

in Ph1 cells, mutations of BCR-ABL or loss of MMR at any time points.

Although these definitions were created for patients who received imatinib

as first line therapy, we applied these criteria also for patients who received

imatinib after IFN-alpha failure.

Procedures. All patients had cytogenetic analysis and RT- and RQ-PCR

studies before dose escalation and thereafter at 3 and 6 months of therapy

and then every 6 months. Conventional cytogenetic analysis was performed

on bone marrow cells by G-banding technique and at least 20 metaphases

were analyzed on direct and short-term cultures after 24 h. For molecular

analysis, mononuclear cells were isolated from 20-mL peripheral blood after

separation on a Ficoll-Hypaque gradient. RNA extraction, RT-PCR, and RQ-

PCR were performed as already described [2]. For mutational screening,

the methodology used for mutation detection was the following: after RNA

extraction and reverse transcription, overlapping fragments covering the

entire kinase domain were generated by nested PCR and screened by

DHPLC. In positive cases, a direct sequencing was performed.

Response criteria were the following: complete hematologic remission

(CHR) was defined as a white blood cell count less than 10 3 109/L with no

immature cells in the peripheral blood, a platelet count less than 450 3 109/L,

and disappearance of all signs and symptoms related to leukemia. Partial

hematologic response was defined as the persistence of peripheral imma-

ture cells or persistence with improvement of more than 50% of splenome-

galy and degree of thrombocytosis. CyR, based on results of conventional

analysis, was categorized as: complete (CCyR), 0% Ph-positive meta-

phases; partial (PCyR), 1% to 35% Ph-positive metaphases; major (MCyR),

0% to 35% Ph-positive metaphases; minor, 36% to 65% Ph-positive meta-

phases; minimal (mCyR), 66% to 95% Ph-positive metaphases; no

response, more than 95% Ph-positive metaphases. We defined MMR as a

BCR-ABL/ABL ratio less than 0.1% and a CMR as a ratio less than 0.001.

To confirm a CMR value for a single assay, our definition required that the

analysis gave an ABL copy number corresponding to a 4.5-log sensitivity of

the assay and that the samples negative by RQ-PCR were analyzed by

nested PCR: CMR was defined as nested PCR negativity.

Statistical analysis. Statistical analysis was carried out using the SPSS

software package. In particular, responses and frequency of toxicity were

compared using w2 or Fisher’s exact test. Survival was defined as the time

from diagnosis to death or date of last follow-up. Progression free survival

was defined from the time of start of imatinib to progression to advanced

phase of disease. Actuarial probability of survival and progression free sur-

vival was calculated using the Kaplan–Meier method and compared by the

log-rank test. P values were considered significant when were <0.05.

1Department of Cellular Biotechnology and Hematology, Sapienza University,Rome, Italy

2Hematology Section, Department of Biomedical Sciences, University of Catania,Catania, Italy

3General Pathology Section, Department of Biomedical SciencesUniversity of Catania, Catania, Italy

*Correspondence to: Massimo Breccia, Department of Human Biotechnologiesand Hematology, Via Benevento 6, Rome 00161, Italy

E-mail: breccia@bce.uniroma1.itReceived for publication 22 January 2010; Accepted 3 February 2010

Conflict of interest: Nothing to report.Published online 8 February 2010 in Wiley InterScience

(www.interscience.wiley.com).DOI: 10.1002/ajh.21680

References

1. Baccarani M, Saglio G, Goldman J, et al. European leukemia net. Evolving

concepts in the management of chronic myeloid leukemia: Recommendations

from an expert panel on behalf of the European Leukemia Net. Blood 2006;

108:1809–1820.2. Baccarani M, Pane F, Saglio G. Monitoring treatment of chronic myeloid leu-

kemia. Haematologica 2008;93:161–169.3. Deininger M, O’Brien SG, Guilhot F, et al. International randomized study of

interferon Vs STI571 (IRIS) 8-year follow up: Sustained survival and low riskfor progression or events in patients with newly diagnosed chronic myeloidleukemia in chronic phase (CML-CP) treated with imatinib. Blood 2009;114:abst 1126

4. Apperley JF. Part 1: Mechanisms of resistance to imatinib in chronic myeloidleukaemia. Lancet Oncol 2007;8:1018–1029.

5. Melo JV, Chuah C. Resistance to imatinib mesylate in chronic myeloid leukae-mia. Cancer Lett 2007;249:121–132.

6. Hochhaus A, Erben P, Ernst T, Mueller MC. Resistance to targeted therapy inchronic myelogenous leukemia. Semin Hematol 2007;44:S15–S24.

7. Kantarjian HM, Larson RA, Guilhot F, et al. International randomized study ofinterferon vs STI571 (IRIS) investigators efficacy of imatinib dose escalationin patients with chronic myeloid leukemia in chronic phase. Cancer 2009;115:551–560.

8. Marin D, Goldman JM, Olavarria E, Apperley JF. Transient benefit onlyfrom increasing the imatinib dose in CML patients who do not achievecomplete cytogenetic remissions on conventional doses. Blood 2003;102:2702–2703.

9. Zonder JA, Pemberton P, Brandt H, et al. The effect of dose increase of imati-nib mesylate in patients with chronic or accelerated phase chronic myeloge-nous leukemia with inadequate hematologic or cytogenetic response to initialtreatment. Clin Cancer Res 2003;9:2092–2097.

10. Jabbour E, Kantarjian H, Jones D, et al. Imatinib mesylate dose escalation is

associated with durable responses in patients with chronic myeloid leukemia

after cytogenetic failure on standard-dose imatinib therapy. Blood 2009;113:

2154–2160.11. Kantarjian HM, Talpaz M, O’Brien S, et al. Dose escalation of imatinib mesy-

late can overcome resistance to standard-dose therapy in patients with

chronic myelogenous leukemia. Blood 2003;101:473–475.12. Rea D, Etienne G, Corm S, et al. Imatinib dose escalation for chronic phase-

chronic myelogenous leukaemia patients in primary suboptimal response to

imatinib 400 mg daily standard therapy. Leukemia 2009;23:1193–1196.

letters

American Journal of Hematology 377

Acute leukemia association with psoriasis: A report on 100patients from a single center in ChinaYing Wang, Yingchang Mi, Dapeng Li, Yanping Xue, Shougeng Bian, and Jianxiang Wang*

The association between psoriasis and leukemia is a controversial topic.

One hundred acute leukemia (AL) patients with psoriasis were divided

into three groups according to treatment of psoriasis: (i) a bimolane

treatment group (n 5 35); (ii) a nonbimolane treatment group (n 5 45);

and (iii) an untreated group (n 5 20). The most common type of leukemia

associated with psoriasis was acute promyelocytic leukemia (APL). The

proportion of nonrecurrent clonal chromosome structural aberration in

the bimolane treatment group (42.9%) was significantly higher than that

of the nonbimolane treatment group (17.1%) and the untreated group

(21.4%; P < 0.05). In patients with non-APL, the complete remission rate

of the first course in the nonbimolane treatment group (84.2%) was sig-

nificantly higher than that of the bimolane treatment group (8.3%) and

the untreated group (37.5%; P < 0.05). The 5-year overall survival rate

(31.2%) and 5-year relapse-free survival rate (27.2%) in the bimolane

group were lower than those of the nonbimolane treatment group

(60.3%, 58.9%) and the untreated group (54.6%, 51.6%; P < 0.01). Nonre-

current chromosome aberration (CA) was prone to occur in bimolane

treated patients, who have lower remission rates and poor prognoses.

Psoriasis is a chronic inflammatory skin disease characterized by a benign

proliferation of keratinocytes. Because of characteristically rapid proliferation

of the epidermal cells, some anticancer drugs, which prevent DNA synthe-

sis, may be effective. Imine C, Imine B, and bimolane are such drugs that

can be used to treat psoriasis. Hundreds of psoriasis patients who were

treated with bimolane or analogs and suffered from acute leukemia (AL)

were reported [1–3]. Further experimental studies [4–6] confirmed that bimo-

lane and Imine B, which belong to topoisomerase II inhibitors with a strong

mutagenic ability, promoted chromosome aberration and fracture. Clinical

and laboratory evidence showed that leukemia in some of the patients with

psoriasis was likely to be secondary to the treatment of bimolane and Imine

B. However, some psoriatic patients who never took bimolane or other drugs

suffered from leukemia [7]. In addition, it was reported that the most com-

mon subtype of bimolane treatment-related leukemia is acute promyelocytic

leukemia (APL) [2–3]. This is significantly different from the constituent ratio

of AL in general population, in which APL accounts for about 20%.

We studied one-hundred (70 males, 30 females) de novo AL patients with

psoriasis who were registered in our center from November 1983 to January

2008. Median follow-up time was 13.5 months (range, 2–206 months).

Eighty of 100 had been treated for psoriasis. Of them, 35 patients had taken

bimolane or analogs (referred to as a bimolane treatment group) and 45

patients had taken medicine other than bimolane or analogs (referred to as

a nonbimolane treatment group). The remaining 20 patients did not receive

any treatment for psoriasis (referred to as the untreated group).

APL patients received all trans-retinoic acid (ATRA) or arsenic trioxi-

de(As2O3) as induction therapy. Acute myeloid leukemia (AML) patients

received standard HA [Homoharringtonine (HHT), cytosine arabinoside (Ara-

C)], DA [Daunorubicin (DNR), Ara-C], or HAD (HHT, Ara-C, and DNR) regi-

men as induction therapy. Acute lymphocytic leukemia (ALL) patients

received standard VDCP [Vincristine-VCR, DNR, Cyclophosphamide (CTX),

Prednisone (PDN)] or VDCLP (VCR, DNR, CTX, L-asparaginase, PDN) regi-

men as induction therapy. All patients who achieved complete remission (CR)

received postremission therapy.

The median ages were 45 years (range, 19–67 years), 38 years

(range,11–62 years), and 42.5 years (range,19–66 years) in the bimolane

treatment group, the nonbimolane treatment group, and the untreated group,

respectively. There were more patients above 60 years of age in the bimo-

lane treatment group than the other two groups (P < 0.05). Median white

blood cell (WBC) count was 4.45 3 109 per liter (range, 0.9–200.2 3 109

per liter), 4.20 3 109 per liter (range, 0.36–217.6 3 109 per liter), 2.72 3

109 per liter (range, 0.58–171.4 3 109 per liter) and median hemoglobin

(Hb) count were 94.5 g/l (range, 40–165 g/L), 80.5 g/L (range, 42–149 g/L),

62.5 g/L (range, 37–120 g/L) in the bimolane treatment group, the nonbimo-

lane treatment group, and the untreated group at diagnosis, respectively.

Anemia in the untreated group was more frequent and worse than in the

other two groups (P < 0.05).

There were more patients with more than 10 years course of psoriasis in

the bimolane treatment group than in the other two groups (P < 0.05).

The most common subtype of leukemia is APL, followed by AML with t(8;

21), which accounts for about 75%. The constituent ratio of leukemia sub-

types had no significant statistical difference (P > 0.05) among the three

groups. The details are shown in Table I.

Chromosome aberration (CA) can be divided into two types: one is recur-

rent CA, including t(15;17) (q22; q12), t (8;21) (q22;q22), and inv (16)

(p13q22), 11q23 and the other is nonrecurrent CA. The incidence of CA in

70 patients with available karyotype analysis was 72.9% (51/70), nonrecur-

rent CA occurred in 31/70 (45.7%) patients (including 20 patients with both

nonrecurrent CA and recurrent CA simultaneously). The proportion of nonre-

current clonal structural aberration in the bimolane treatment group (42.9%)

was significantly higher than that of the nonbimolane treatment group

(17.1%) and the untreated group (21.4%; P < 0.05).

Eighty-seven patients received induction therapy, in which 48 cases were

APL patients and 39 cases were non-APL patients. Of 48 (89.6%) APL

patients, 43 achieved CR, and there was no significant difference among the

three groups in the CR rate (P > 0.05). In non-APL patients, CR rate after

the first course of treatment in the nonbimolane treatment group (84.2%)

was significantly higher than that in the bimolane treatment group (8.3%)

and the untreated group (37.5%; P < 0.05). Further analysis showed that

non-APL patients with psoriasis less than 10-years-course, treated with

medicine other than bimolane or analogs, were more likely to achieve CR

after the first course of induction treatment.

The rates of 5-year overall survival (OS) and 5-year relapse-free survival (RFS)

in the bimolane treatment group were 31.2% and 27.2%, respectively, which

were lower than those of the nonbimolane treatment group (60.3%, 58.9%) and

the untreated group (54.6%, 51.6%; P < 0.01) (shown in Fig. 1). WBC greater

than 10 3 109 per liter when diagnosed was independent of poor prognostic fac-

tors for OS by COX regression analysis (P < 0.05, relative risk (RR) 5 5.60).

Nonrecurrent clonal chromosomal structural aberration when diagnosed was the

independent poor prognostic factor for RFS (P < 0.05, RR5 4.09).

Previous studies have shown that patients with psoriasis have a higher

risk of suffering from hematological malignancies [8–10]. An antigenic stimu-

lation hypothesis was proposed, in which immune stimulation may lead to

an increased risk of malignancy. Soderberg et al. [10] conducted a nation-

wide case-control study in support of this hypothesis. Their results revealed

increased risks of hematological malignancies in patients with psoriasis or

Sjogren’s syndrome that were Th1-mediated chronic inflammatory proc-

esses, whereas they did not find elevated risks of hematological malignan-

cies in patients with pernicious anemia, which was Th1-mediated in the

absence of chronic inflammation or multiple sclerosis where inflammation is

intermittent.

Lu et al. [11] analyzed chromosome fragile sites of peripheral blood lym-

phocytes in 24 patients with psoriasis vulgaris and found that the incidence

of CA in the psoriasis group was higher than in the normal control group.

There were 47 common fragile sites in the psoriasis group, and seven of

these were in accordance with leukemia-related fragile sites: 3p14, 2p13,

6q25, 6q26, 3q27, 7q22, xq22. These results suggested that there was cyto-

genetic instability and defect of DNA repair in some psoriatic patients.

Why is the incidence rate of some leukemia subtypes in psoriatic patients

different from the general population and why is APL the most common

occurrence? The promyelocytic leukemia (PML) gene plays a role in the

pathogenesis of psoriasis. Liu et al. [12] showed a significantly higher

expression of the PML protein in progressive plaque lesions than in unin-

volved skin. Low-level expression of the retinoic acid receptor (RAR)a in the

epidermis of psoriatic patients results in uninhibited proliferation of epidermal

letters

378 American Journal of Hematology

cells [13]. The efficacy of retinoic acid in the treatment of psoriasis relies on

its ability to enhance the function of AP-1 by binding with RAR, thus inhibit-

ing the abnormal proliferation and apoptosis of epidermal cells and inducing

normal differentiation [14]. The role of the PML/RARa fusion gene in the

pathogenesis of APL has been very clear, and ATRA was used successfully

to treat APL. The involvement of PML and RAR genes in the pathogenesis

of both psoriasis and APL may provide an explanation for the susceptibility

of psoriatic patients to APL.

Although the constituent ratio of leukemia subtypes in the three groups

had no significant difference, the treatment outcome and prognosis were

obviously different. The prognosis of the bimolane treatment group was sig-

nificantly worse than that of the other two groups. Prognosis of the untreated

group was worse than that of the nonbimolane treatment group. More than

10 years course of psoriasis significantly affected CR rate, especially the

CR rate after the first course of treatment; an initial WBC count greater than

10 3 109 per liter was a poor independent prognostic factor of OS; nonre-

current clonal chromosomal structural aberration was a poor independent

prognostic factor of RFS. There were more psoriatic patients with more than

10 years history, higher incidence of nonrecurrent clonal chromosomal struc-

tural aberration, and more elderly patients (over age 60) in the bimolane

treatment group than the other two groups, which may be the reasons why

the bimolane treatment group had a worse outcome.

In conclusion, APL and AML with t(8; 21) were the most common types of

ALs in psoriatic patients. Psoriatic patients who were treated with biomolane

and analogs were subject to nonrecurrent CA and had lower remission rates

and poor long-term survival.

Department of Clinical Hematology, Institute of Hematology and Blood DiseasesHospital, Chinese Academy of Medical Sciences,

Peking Union Medical College, Tianjin, ChinaContract grant sponsor: National Science Foundation for Distinguished

Young Scholars; Grant number: 30025019. Contract grant sponsor: The MajorState Basic Research Development Program; Contract grant number:

2006CB910406; Contract grant sponsor: 863 Project; Contract grant number:2006AA02A405

*Correspondence to: Jianxiang Wang, MD, Department of Clinical Hematology,State Key Laboratory of Experimental Hematology, Institute of Hematology andBlood Diseases Hospital, Chinese Academy of Medical Sciences, Peking Union

Medical College, 288 Nanjing Road, Tianjin 300020, ChinaE-mail: wangjx@ihcams.ac.cn

Conflict of interest: Nothing to report.Published online 16 February 2010 in Wiley InterScience

(www.interscience.wiley.com).DOI: 10.1002/ajh.21683

References1. Wang SH, Ma C. A case report of psoriasis treated with Imine B which

induced acute promyelocytic leukemia. Clin J Dermatol 1984;13:48.2. Zhang MH, Wang XY, Gao LS, et al. 140 cases of acute leukemia caused by

bimolane. Chin J Inter Med 1993;32:668–671.3. Sheng Q, Chen BA, Jiang ZY. Clinical analysis on 146 cases of psoriasis

treated with tumoricidal drugs which induced leukemia. Dermatol Venerol2003;25:5–7.

4. Christopher E, Heather Smith H. Bimolane: In vitro inhibitor of human topois-merse II. Cancer Lett 1997;120:135–140.

5. Wang YZ, Li JX, Yan MR, et al. Study on pathogenesis of bimolane inducingleukemia. Chin J Haematol 1989;10:19–21.

6. Xue KX, Wang S, Ma JG, et al. Bimalone induced chromosome aberration ofhuman lymphocyte and micronucleus emerging. Chin J Hematol 1991;12:321–322.

7. Paslin D. Psoriasis without neutrphils. Int J Dermatol 1990;29:37–40.8. Hannuksela-Svahn A, Pukkala E, Laara E, et al. Psoriasis, its treatment,

and cancer in a cohort of Finnish patients. J Invest Dermatol 2000;114:587–590.

9. Gelfand JM, Berlin J, Van Voorhees A, et al. Lymphoma rates are low butincreased in patients with psoriasis: Results from a population-based cohortstudy in the United Kingdom. Arch Dermatol 2003;139:1425–1429.

10. Soderberg KC, Jonsson F, Winqvist O, et al. Autoimmune diseases, asthmaand risk of haematological malignancies: A nationwide case-control study inSweden. Eur J Cancer 2006;42:3028–3033.

11. Lu HB, Li WD, Wen HQ, et al. Chromosome aberration rate of lymphocytes inperipheral blood and fragile site of psoriasis patients. Chin J Dermatol 1998;11:112.

12. Liu ER, Peng ZH, Tan SS. Role of promyelocytic leukemia protein expressionin the pathogenesis of psoriasis. J Xi’an Jiaotong University (Medical Scien-ces) 2005;26:186–187.

13. Feng S, Lin L, Wu Q, et al. Study on the expression of RXRalpha in patientswith psoriasis vulgaris. Eur J Dermatol 2006;16:33–38.

14. Kerkhof PC. Update on retinoid therapy of psoriasis in: An update on the useof retinoids in dermatology. Dermatol Ther 2006;19:252–263.

Figure 1. Overall survival (OS) and relapse-free survival (RFS) of three groups.

TABLE I. The Constituent Ratio of Leukemia Types in the Three Groups

Diagnosis

No. (%)

Bimolane treatment group Nonbimolane treatment group Untreated group Total

AML-M1 0/35 (0) 1/45 (2.3) 0/20 (0) 1/100 (1.0)AML-M2 5/35 (14.3) 12/45 (27.3) 4/20 (20.0) 21/100 (21.0)t(8;21)AML 4/35 (11.4) 12/45 (27.3) 4/20 (20.0) 20/100 (20.0)APL 21/35 (60.0) 22/45 (50.0) 10/20 (50.0) 53/100 (53.0)AML-M4 3/35 (8.6) 1/45 (2.3) 2/20 (10.0) 6/100 (6.0)AML-M4eo 0/35 (0) 3/45 (6.7) 2/20 (10.0) 5/100 (5.0)AML-M5 3/35 (8.6) 5/45 (11.4) 1/20 (5.0) 9/100 (9.0)AML-M6 1/35 (2.9) 0/45 (0) 0/20 (0) 1/100 (1.0)ALL 2/35 (5.7) 1/45 (2.3) 1/20 (5.0) 4/100 (4.0)

letters

American Journal of Hematology 379

Procalcitonin is a reliable marker of severe systemic infectionin neutropenic haematological patients with mucositisLoredana Sarmati,1* Angela Beltrame,1 Luca Dori,1 Gaetano Maffongelli,1 Laura Cudillo,2

Gottardo De Angelis,2 Alessandra Picardi,2 Licia Ottaviani,3 Maria Giovanna Cefalo,3

Adriano Venditti,3 Sergio Amadori,3William Arcese,2 and Massimo Andreoni1

Patients with neutropenia are exposed to a high risk for infections in

which fever is often the unique symptom [1]. Systemic infections remain

the main cause of mortality in these patients therefore, the policy for

infection management is to promptly administer empirical antibiotic ther-

apy in order to avoid the increased risk of mortality related to the treat-

ment delay [2]. However, microbiological diagnostic tests are not suffi-

ciently rapid, sensitive or specific to indentify the microbial causes of

fever, and a considerable number of patients suffer febrile episodes

over a prolonged period without a definite microbiological etiology.

Procalcitonin (PCT) [3] has become increasingly popular as a novel

marker of infection. Several studies have underscored its value in clinical

conditions by identifying infectious processes [4–6]. The use of this marker

in hematological patients has provided controversial results and no agree-

ment exists about the capacity of PTC to differentiate fever by other inflam-

matory processes such as mucositis and graft versus host diseases (GVHD)

[7–11].

To evaluate the usefulness of PCT in diagnostic and therapeutic

approaches to fever in hematological patients, we evaluated the values of

PCT, C-reactive protein (CRP), and erythrocyte sedimentation rate (ESR) in

the course of febrile neutropenia by assessing the influence of mucositis

and GVHD on these three parameters.

The demographics and the clinical characteristics of the study population

are shown in Table I. One hundred and sixteen infectious events were

observed, 50% of them occurring in patients with mucositis. Nine subjects

(30%) out of 30 transplanted patients developed GVHD. FUO, pneumonia,

sepsis, and infection of central venous catheter (CVC) were the most com-

mon clinical features associated with fever.

To assess the power of the investigated inflammatory parameters in discrim-

inating the clinical severity of infection episodes, the median values of PCT,

CRP, and ESR as measured during the major events were compared with

those recorded during the minor events. During major infectious events, the

PCT levels were significantly higher at any considered time point (PCT1: 0.55

vs. 0.13 ng/ml, P < 0.001; PCT2: 1.04 vs. 0.21 ng/ml, P < 0.001; PCT3: 0.54

vs. 0.16 ng/ml, P < 0.001), CRP was significantly higher only at times one and

two (CRP1: 112 vs. 56 mg/dl, P 5 0.001; CRP2: 150 vs. 104 mg/dl, P 5 0.02)

and the ESR values were not significantly different at any time.

Figure 1 shows the results of the PCT and CRP determinations during

febrile major and minor events in patients with and without mucositis. The

analysis showed that the PCT values were significantly higher at all three

time points during major events compared with minor events, regardless of

the presence of mucositis. On the contrary, the values of CRP were signifi-

cantly higher at the time point one in patients with major events and mucosi-

tis, and at time-point one and two in those with major events without mucosi-

tis, compared with patients with minor events.

The sensitivity and specificity of PCT and CRP to identify major infectious

events in patients with mucositis was evaluated through ROC curves (see

Fig. 2). The area under the curve accounts for the statistical significance for

PCT1 (68%; 95% CI: 54–82; P 5 0.016), PCT2 (68%; 95% CI: 53.9–81.7; P

5 0.02), and PCT3 (65%; 95% CI: 51–80; P 5 0.035). Otherwise, the area

under the ROC curve was statistical significant for CRP1 (68%, 95% CI: 54–

81; P 5 0.019), but not for CRP2 (54%, 95% CI: 38.8–68.9; P 5 0.6) and

CRP3 (48%, 95% CI: 33–63; P 5 0.8). Moreover, in patients with mucositis,

PCT2 values of �0.5 ng/ml and �0.2 ng/ml discriminates major infectious

events with sensitivities of 48% and 79% and specificities of 70% and 52%,

respectively.

A bacterial isolate was obtained in 50/59 (85%) major infectious events

and in 22/57 (38%) minor events. Gram-negative isolates were significantly

more frequent during major events compared with minor events (44/59 vs.

4/57, P < 0.001).

The PCT values, but not the CRP values, were significantly higher at all

three time points in patients with a gram-negative isolate relative to patients

with an isolate other than a gram negative (PCT1: 1.9 vs. 0.2 ng/ml, P 5

0.001; PCT2: 3.5 vs. 0.4 ng/ml, P < 0.001; PCT3: 0.7 vs. 0.4 ng/ml, P 5

0.005). Moreover, a PCT2 value of �0.98 ng/ml was found to discriminate

between the presence or absence of a gram-negative bacterial isolation

during fever with a sensitivity of 80% and a specificity of 78%.

The study of PCT and CRP was conducted separately on the group of

transplant patients with the occurrence of GVHD. To this end, only patients

at risk of GVHD (allogeneic, haploidentical and umbilical cord blood trans-

plants) have been considered. GVHD occurred in 12 out of 30 patients, and

it was skin localized in 11/12 cases and bowel localized in one case. Major

infectious events occurred in 9/12 cases and in all of these patients high lev-

els of PCT (>1.3 ng/ml) were detected at all three time points. During minor

infectious events, the values of PCT1, 2, and 3 were �0.2 ng/ml regardless

of GVHD.

The present study shows that PCT is a reliable marker of infection in

patients with hematological neoplasms and febrile neutropenia. A fair number

of publications [7–16] have evaluated the utility of PCT as a marker of infection

in neutropenic patients with solid-organ cancer or hematological malignancy.

Overall, these studies showed a correlation between an elevation of PCT and

the presence of infection. However, in a recent review [9] that explored studies

published between January 1990 and October 2006, the authors were unable

to reach definitive conclusions about the usefulness of PCT in febrile neutro-

penia because of the heterogeneity of the populations evaluated and the varia-

bility of the results obtained in the different articles.

In our study, the significant increase of PCT during infectious events was

not affected by inflammatory conditions such as mucositis and GVHD. To

date, only two articles have addressed the role of PCT in hematological

patients with neutropenic fever and mucositis and/or GVHD, and they came

to conflicting conclusions. Both these studies used the LUMI Test

B.R.A.H.M.S. to assess the levels of PCT. Blijlevens et al. [7] demonstrated

the inability of PCT to discriminate between infection and other inflammatory

conditions in 12 hematological transplanted patients. In particular, the

authors found out that GVHD rather than infection accounted for the

increase of PCT. Fleischhack et al. [11], in a study including pediatric

patients with hematological or solid tumors without infections, concluded that

chemotherapy-induced tissue damage (such as severe mucositis) did not

cause considerable increases of the PCT plasma levels. Some considera-

tions can be made to justify these different results. In the first study, the lim-

ited number of patients impacted the statistical evaluation of the results; in

the second study, none of the patients was transplanted and half of the

febrile events occurred in children with solid tumors.

In our study, the levels of PCT at all three time points detected during the

course of major infectious events in patients with and without mucositis were

significantly higher than in patients with minor infections. Regarding the influ-

ence of GVHD on the PCT increase in our study population, although the

sample size does not allow statistical evaluation, it is clear that the PCT was

not affected by GVHD and values greater than 1 ng/ml are present only in

patients with severe infections regardless of the simultaneous occurrence of

GVHD.

In some studies [12,17,18], it was reported that PCT elevation could be

significantly higher in patients with gram-negative bacteraemia than in those

with gram-positive bacteraemia. Kocazeybek et al. [18], in a study on

patients with infectious endocarditis, found that the median serum PCT level

of gram-positive-related endocarditis was 2.92 ng/ml, whereas this level was

8.62 ng/ml, in cases of endocarditis due to gram-negative bacteria. Charles

et al. [17], in a review on bacteraemia in critically ill patients, observed that

a high-PCT value was found to be independently associated with gram-neg-

letters

380 American Journal of Hematology

ative bacteraemia and that a PCT level of 16.0 ng/ml yielded an 83% posi-

tive predictive value and a 74% negative predictive value. In our study, val-

ues of PCT of �0.98 ng/ml were able to discriminate between the presence

and the absence of gram-negative bacteraemia with a sensitivity of 80% and

a specificity of 78%. This is of great importance in relation to the clinical

severity of gram-negative bacteraemia in patients with neutropenia and sug-

gests the use of PCT in guiding the choice of empirical antibiotic therapy

during febrile neutropenia.

In conclusion, our experience indicates that the determination of PCT at

the given time points might represent a valuable tool to predict the occur-

rence of major infectious episodes in patients with febrile neutropenia, even

in presence of confounding associated conditions such as mucositis and/or

GVHD. This would allow an early and appropriate antibiotic therapy to be

administered in a very high-risk population such as the one including

patients with hematological malignancies and chemotherapy-related febrile

neutropenia.

Methods

Patients. From January 2008 to December 2008, 116 febrile episodes

occurring in 88 neutropenic patients were investigated. All the patients were

affected with hematological neoplasms and 30 of them underwent stem cell

transplantation (SCT).

Fever was defined as a single measurement of a temperature of 38.58C

or two or more measurements of 388C. Neutropenia was defined as an

absolute neutrophils count <1,000/mm3. Infection events included all epi-

sodes of fever associated with evidence of microbiological isolations. Sepsis,

septic shock, bacteraemia (BSI), and pneumonia were defined as major

infectious events; a fever of unknown origin (FUO), infections of the site of

insertion of the CVC, sinusitis, Herpes simplex virus manifestations, and fol-

liculitis were defined as minor infectious events.

BSI was defined as a microbial growth in one blood culture bottle; how-

ever, at least two positive blood cultures bottles collected from different sites

were required for microorganisms usually considered as potential contami-

nants (i.e., coagulase-negative Staphylococcus species). Hemoculture sam-

ples were collected upon the first appearance of fever and then on every

occasion in which fever was �388C. Sepsis was defined by the presence of

fever or hypothermia, tachypnea, and tachycardia. Instances of sepsis asso-

ciated with hypotension not corrected by proper fluid therapy were scored as

septic shock.

Pneumonia was assumed in cases in which typical clinical and radiological

findings were seen on the chest radiograph obtained upon fever onset. CVC

infection was suspected in presence of skin inflammation or vein thrombosis

TABLE 1. Patients and Febrile Episode Characteristics

Characteristic Na (%)

Total no. of patients 88No. of nontransplanted patients 58 (66)LMAa 40 (69)LLAb 9 (16)LMCc 1 (2)LNHd 3 (5)LTe 1 (2)LBf 2 (3

No. of transplanted patients 30 (34)Allogeneic 7 (23)Autologous 14 (47)Haploidentical 3 (10)Umbilical cord Blood 6 (20)

Age (years)Median (range) 47 (16–70)

GenderMale 52 (59)Female 36 (41)

Total no. of febrile episodes 116Sepsis 21 (18)Septic shock 9 (8)Pneumonia 22 (19)Bacteraemia 9 (8)CVCg 20 (17)FUOh 29 (25)Otheri 6 (5)

Febrile episodes associated with mucositis (in total) 58 (50)Febrile episodes with GVHD (only in transplanted) 9 (30)

aLMA Acute myeloid leukaemia; bAcute lymphatic leukaemia; cChronic myeloidleukaemia; dLymphoma non Hodgkin; eT-cell lymphoma; fBurkitt lymphoma; gCen-tral venous catheter infection; hFever of unknown origin; iFolliculitis, Herpes sim-plex, sinusitis.

Figure 1. PCT (a and c) and CRP (b and d) determinations during febrile major and minor events in patients with and without mucositis.

letters

American Journal of Hematology 381

at the CVC insertion site. Episodes of fever not associated with microbiologi-

cal or clinical evidence of infection were regarded as FUO.

Oral and gastrointestinal mucositis was defined and graded according to

the Common Terminology Criteria for Adverse Events (CTCAE) version

three [19]. Acute GVHD was evaluated according to the standard

criteria [20].

PCT, ESR, and CPR evaluations. PCT was determined by a commer-

cially available immunoluminetric assay, LUMI Test PCT kit (B.R.A.H.M.S.

Diagnostica, Berlin Germany). Values greater than or equal to 0.5 ng/ml

were considered to be pathological in accordance with the manufacturer’s

instructions (Al-Nawas & Shah, 1996). The precision of the test was esti-

mated in accordance with CLSI guidelines EP5-A (evaluation of precision

performance of clinical chemistry devices). The interassay coefficient var-

iation at PCT value of 0.1 ng/ml was 15%, 0.2 ng/ml was 10%, and >0.3

ng/ml was <6%.

ESR was determined with the VES Matic Cube 200 (Diesse, FL), an auto-

matic system using 1 ml of undiluted blood sample with K2-EDTA as an anti-

coagulant. Blood collected in the EDTA tube and accurately mixed was left

to settle. By analogical sensors, the instrument automatically reads the

erythrocyte sedimentation levels. The results obtained were compared with

the Westergren method and then printed.

The CRP measurement was determined with the use of a Tina-quant

immunoturbimetric kit for CRP (latex) highly sensitive assay (Roche). The

measuring range is from 0.01 to 2 mg/dl. The turbidity resulting from

the immunoprecipitation of CRP in the serum with a specific antibody, in

the presence of polyethylene glycol, is measured as an increase in absorb-

ance at 340 nm and is proportional to the concentration of CRP in the

sample. The samples were read with the Hitachi 704/705 instruments

(Roche).

Blood samples to evaluate PCT, ESR, and CRP were collected upon the

first appearance of fever and then every day after until its resolution. For the

statistical analysis, the values of three specific time points were considered:

time 1 (PCT1, ESR1, and CRP1) on the first day of fever, time 2 (PCT2,

ESR2, and CRP2) on the day with the highest values during the febrile epi-

sode, and time 3 (PCT3, ESR3, and CR 3) on the last day of fever.

Statistical analysis. The statistical analysis was carried out using the

Mann–Whitney U test for the medians of nonparametric data and Fisch-

er’s Exact Test for the comparison of proportions. The sensitivity, specific-

ity and area under the receiving operator characteristic (ROC) curve were

calculated for both PCT and CRP. For the statistical analysis, the data

were stored by Excel 2003 and later processed using the software SPSS

version 16.

1Clinical Infectious Diseases, Tor Vergata University, Rome, Italy2Haematology Stem Cell Transplant Unit, Tor Vergata University, Rome, Italy

3Oncohaematologic Unit, Tor Vergata University, Rome, Italy*Correspondence to: Loredana Sarmati, Clinical Infectious Diseases,

Tor Vergata University, V. Montpellier 1, 00133 Rome, Italy.E-mail: sarmati@med.uniroma2.it

Conflict of interest: Nothing to report.Published online 16 February 2010 in Wiley InterScience

(www.interscience.wiley.com).DOI: 10.1002/ajh.21685

References1. Link H, Bohme A, Cornely OA, et al. Antimicrobial therapy of unexplained fever

in neutropenic patients guidelines of the infectious diseases working party(AGIHO) of the German society of hematology and oncology (DGHO), studygroup interventional therapy of unexplained fever, arbeitsgemeinschaft suppor-tivmassnahmen in der onkologie (ASO) of the Deutsche krebsgesellschaft(DKG-German cancer society). Ann Hematol 2003;82 (Suppl. 2):105–111.

2. Bodey GP, Buckley M, Sathe YS, et al. Quantitative relationships betweencirculating leukocytes and infection in patients with acute leukemia. Ann InternMed 1966;64:328–340.

3. Maruna P, Nedelnikova R, Gurlich R. Physiology and genetics of procalcito-nin. Physiol Res 2000;49:S57–S61.

4. Assicot M, Gendrel D, Carsin H, et al. High serum procalcitonin concentra-tions in patients with sepsis and infection. Lancet 1993;341:515–518.

5. Simon P, Milbrandt EB, Emlet L. Procalcitonin-guided antibiotics in severesepsis. Critical Care 2008;12:309.

6. Charles PE, Ladoire S, Snauwaert A, et al. Impact of previous sepsis on theaccuracy of procalcitonin for the early diagnosis of blood stream infection incritically ill patients. BMC Infect Dis 2008;8:163.

7. Blijlevens NMA, Donnelly JP, Meis JFG, et al. Procalcitonin does not discrimi-nate infection from inflammation after allogeneic bone marrow transplantation.Clin Diagn Lab Immunol 2000;7:889–892.

8. Schuttrumpf S, Binder L, Hagemann T, et al. Procalcitonin: A useful discrimi-nator between febrile conditions of different origin in hemato-oncologicalpatients? Ann Hematol 2003;82:98–103.

9. Sakr Y, Sponholz C, Tuche F, et al. The role of procalcitonin in febrile neutro-penic patients: Review of the literature. Infection 2008;36:396–407.

10. Svaldi M, Hirber J, Lanthaler A, et al. Procalcitonin-reduced sensitivity andspecificity in heavily leucopenic and immunosuppressed patients. Br J Hae-matol 2001;115:53–57.

11. Fleischhack G, Kambeck I, Cipic D, et al. Procalcitonin in paediatric cancerpatients: Its diagnostic relevance is superior to that of C-reactive protein,interleukin 6, interleukin 8, soluble interleukin 2 receptor and soluble tumournecrosis factor receptor II. Br J Haematol 2000;111:1093–1102.

12. Svaldi M, Hirber J, Lanthaler AI, et al. Procalcitonin-reduced sensitivity andspecificity in heavily leucopenic and immunosuppressed patients. Br J Hae-matol 2001;115:53–57.

13. Carnino L, Betteto S, Loiacono M, et al. Procalcitonin as a predictive markerof infections in chemoinduced neutropenia. J Cancer Res Clin Oncol, 2010;136:611–615.

14. Semeraro M, Thomee C, Rolland E, et al. A predictor of unfavourable out-come in neutropenic paediatric patients presenting with fever of unknown ori-gin. Pediatr Blood Cancer, 2010;54:284–290.

Figure 2. Comparison between the PCT and CRP ROC curves in febrile eventsin the presence of mucositis according to major infectious events. (a) PTC1,68% (95% CI: 54–82; P 5 0.016); CRP1, 68% (95% CI: 54–81; P 5 0.019). (b)PCT2, 68% (95% CI: 53.9–81.7; P 5 0.02); CRP2, 54% (95% CI: 38.8–68.9;P 5 0.6). (c) PCT3, 65% (95% CI: 51–80; P 5 0.035); CRP3, 48% (95% CI:33–63; P 5 0.8).

letters

382 American Journal of Hematology

15. Hatzistilianou M, Rekliti A, Athanassiadou F, et al. Procalcitonin as an earlymarker of bacterial infection in neutropenic febrile children with acute lympho-blastic leukemia. Inflamm Res 2009, in press.

16. Martinez-Albarran M, Perez-Molina Jde J, Gallegos-Castorena S, et al. Procalcito-nin and C-reactive protein serum levels as markers of infection in a pediatric popula-tion with febrile neutropenia and cancer. Pediatr Hematol Oncol 2009;26:414–425.

17. Charles PE, Ladoire S, Aho S, et al. Serum procalcitonin elevation in criticallyill patients at the onset of bacteraemia caused by either gram negative orgram positive bacteria. BMC Infect Dis 2008;8:381–8.

18. Kocazeybek B, Kucukoglu S, Oner YA. Procalcitonin and C-reactive protein ininfective endocarditis: Correlation with etiology and prognosis. Chemotherapy2003;49:76–84.

19. Cancer Therapy Evaluation Program. Common Terminology Criteria forAdverse Events, Version 3.0, DCTD, NCI, Bethesda, MD, USA. NIH, DHHS,March 31, 2003. Available at: http://ctep.cancer.gov, Publish on August 9,2006.

20. Przepiorka D, Weisdorf D, Martin P, et al. Consensus conference on acuteGVHD grading. Bone Marrow Transplant 1995;15:825–828.

JAK2V617F mutation persists in blasts and mature cells oftransformed JAK2V617F-positive-myeloproliferative neoplasia:A European Leukemia Net (ENL) study.Ciro R. Rinaldi,1* Paola Rinaldi,1 Marica Gemei,1 Francesco Grimaldi,1 Giorgia Battipaglia,1

Luigi Del Vecchio,1 Bruno Martino,2 Giorgina Specchia,3 Anna Candoni,4 Luigi Gugliotta,5

Alessandro M. Vannucchi,6 Tiziano Barbui,7 and Fabrizio Pane1

Transformation to acute myeloid leukemia (AML) is a known complica-

tion of myeloproliferative neoplasia (MPN). Recent studies reported the

high incidence (53%) of JAK2 negative blasts from transformed

JAK2V617F-MPN. We collected, by cell sorting, blast cells, and mature

cells (GRA) from total bone marrow (BM) of 34 patients newly diag-

nosed of secondary AML. At MPN diagnosis (PMF n 5 18; PV n 5 9;

ET n 5 7), JAK2 was mutated in 22 of 34 patients. Twenty of 22

JAK2V617F-MPN (91%) maintained the mutation in blasts and GRA

after leukemic switch, while in 2 of 22 patients the selected compart-

ments lost the mutation. Surprisingly, we also found the first case of

JAK2V617F-AML from a wild type (WT)-MPN. In contrast to the pre-

vious study, we conclude that JAK2V617F-MPN yields rarely (9%) a

JAK2WT-AML and any JAK2-status modification/persistence involves

always the entire BM during leukemic transformation.

MPN are a heterogeneous group of diseases characterized by the hyper-

proliferation of myeloid precursors in the bone marrow (BM) and/or in extra-

medullary sites leading to elevated number of nonlymphoid cells and/or pla-

telets in the peripheral blood [1]. The discovery of an acquired somatic

mutation in the JAK2 gene resulting in a valine-to-phenylalanine substitution

at position 617 (JAK2V617F) has substantially modified the molecular knowl-

edge’s behind these syndromes [2–5]. By a sensitive allele-specific polymer-

ase chain reaction (ASO-PCR) assay [6], it has been well established that

around 97% of patients affected by polycythemia vera (PV), 60%–70% of

patients with essential thrombocythemia (ET), and 60% of primary myelofib-

rosis (PMF) are positive for the JAK2V617F mutation. Transformation to

acute myeloid leukemia (AML) is a known complication of MPN but the role

of JAK2V617F mutation is still undefined [7–14]. In 2006, Campbell ana-

lyzed four patients with secondary AML describing a possible model for the

development of a JAK2WT-AML in a patient with JAK2V617F-MPN [10] and

more recently, Theocharides et al.[15] reported that in up to 53% of the

patients who developed secondary AML from a JAK2-mutated MPN the

mutation was no longer detectable; however, the results, in a smaller cohort

of patients (n 5 27), obtained with DNA extracted from cells scraped or

laser-capture-microdissected from bone marrow (BM) or peripheral blood

(PB) smears, were confirmed in fresh samples only in some cases with no-

uniform cell sources (PB granulocytes, BM-blasts, PB-blasts). In this study,

we collected, by cell sorting, blast cells, and mature myeloid cells (granulo-

cytes, GRA) from whole BM aspirates of 34 newly diagnosed patients with

AML secondary to MPN (18 derived from PMF; 9 from PV and 7 from ET)

and analyzed the JAK2 status before and after leukemic transformation in

selected cell compartments. In contrast to the previous studies, we conclude

that a JAK2V617F-positive-MPN yields rarely (9%) a JAK2V617F-negative-

AML. At the time of MPN diagnosis, JAK2V617F was detectable in 22 of 34

patients (65%) (10 of 18 PMF; 9 of 9 PV; and 3 of 7 ET) (Table I). Informa-

tion on treatment during the chronic phase was available in 25 of 34

patients. Twenty patients had received cytoreductive treatment with hydrox-

yurea (HuOH), two patients with HuOH associated to busulphan, one patient

with IFN-alpha, one patient with busulphan plus anagrelide, one patient

HuOH plus clorambucil (Table I). Four PV-untreated-patients had a sponta-

neous transformation to AML. No cytogenetic abnormalities were detected

during the chronic phase of the disease in 30/34 evaluated patients. No

MPL or JAK2-exone 12 mutations were present at this stage (data not

shown). Median time to AML progression (TTP) was 5.09 years (yrs) (range

0.38–27.81). A significant difference (P 5 0.02) in TTP was found grouping

patients according to JAK2 status during the MPN phase [JAK2WT-MPN n

5 12, TTP median 15.10 yrs (0.38–16.32); JAK2 mutated-MPN n 5 22,

TTP median 4.07 yrs (0.67–27.81)]. Cytogenetic data were available in 23 of

TABLE I. Summary of Patient Characteristics: MPN Diagnosis and Previous

Treatment

Patient Gender Age Diagnosis JAK2Cytoreductivetreatment

TTP(yrs)

1 M 72 PMF WT HuOH 16.422 M 67 ET WT BU1HuOH 15.433 F 67 PMF WT BU1HuOH 20.34 M 35 PMF WT – 6.015 F 32 PMF WT HuOH 5.096 M 68 PMF MUT HuOH 2.017 F 70 ET MUT HuOH 4.718 F 88 PMF MUT HuOH 27.819 M 66 PV MUT HuOH 4.99

10 M 45 ET MUT IFN 2.6411 F 57 PV MUT HuOH 4.7112 F 59 PMF MUT – 1.0013 M 55 PV MUT None 5.9814 F 74 PV MUT None 7.5115 F 77 PMF WT Clor1HuOH 16.3216 M 74 PMF WT HuOH 4.1317 F 81 ET MUT – 1.0018 M 58 PMF WT HuOH 0.3819 F 68 PMF MUT HuOH 13.1620 F 65 PV MUT None 1.0221 F 60 PMF MUT HuOH 5.9622 M 49 PMF MUT HuOH 1.0123 F 71 PMF MUT HuOH 4.0724 F 66 ET WT HuOH 15.1125 F 70 PMF MUT – 11.0926 F 77 ET WT HuOH 9.0327 F 71 PV MUT HuOH 10.1628 F 75 PV MUT HuOH 1.0829 M 81 PMF MUT HuOH 2.2930 M 83 PV MUT None 5.3031 F 76 PV MUT HuOH 11.3232 M 65 PMF MUT HuOH 0.6733 M 64 PMF WT – 8.1334 F 67 ET WT BU1ANA 1.09

M, male; F, female; PMF, primary myelofibrosis, ET, essential thrombocytemia; PV,polycythemia vera; HuOH, hydroxiurea; BU, busulphan; IFN, interferon; ANA, ana-grelide; Clor, clorambucil; (2), treatment information’s not available; TTP, time toprogression; yrs, years

letters

American Journal of Hematology 383

TABLEII.Summary

ofPatientCharacteristics:MolecularAnalysis

inAMLSecondary

toMPN

MPN

AML

Patient

Diagnosis

JAK2MNCs(AB

%±SD)

Source

%BLAST

Cytogenetic

JAK2MNCs(AB%

±SD)

JAK2BLAST(AB%

±SD)

JAK2CD151

Cells

(AB%

±SD)

JAK2CD31

Cells

1PMF

WT

BM

80

ND

WT

WT

WT

WT

2ET

WT

BM

90

COMPLEX

WT

WT

WT

WT

3PMF

WT

BM

95

DEL(5)1

TRI(8)1

TRI(1)

WT

WT

WT

WT

4PMF

WT

BM

70

DEL(7)1

DEL(9)1

TRI(14)

WT

WT

WT

WT

5PMF

WT

BM

70

ND

MUT(42±2.85)

MUT(41±1.75)

MUT(42±2)

WT

6PMF

MUT(58±1.75)

BM

76

46,XY

MUT(57±2.5)

MUT(58±2)

MUT(57±2.25)

WT

7ET

MUT(55±1.51)

BM

40

46,XX

MUT(54±1.5)

MUT(55±2)

MUT(55±1.25)

WT

8PMF

MUT(57±2.00)

BM

83

46,XX

MUT(56±1.75)

MUT(58±2.5)

MUT(57±0.75)

WT

9PV

MUT(62±2.13)

BM

71

46,XY

WT

WT

WT

WT

10

ET

MUT(32±0.95)

BM

72

46,XY

MUT(30±2.05)

MUT(32±1.5)

MUT(35±1.5)

WT

11

PV

MUT(60±1)

BM

40

46,XX

MUT(61±1.05)

MUT(60±1)

MUT(59±1.5)

WT

12

PMF

MUT(56±1.2)

PB

59

MONO(7),TRI(1)

MUT(57±2)

MUT(56±2.5)

MUT(56±2)

WT

13

PV

MUT(59±2)

BM

78

ND

MUT(55±1.75)

MUT(60±1.5)

MUT(59±1.75)

WT

14

PV

MUT(81±3)

BM

55

ND

MUT(76.2)

MUT(71±2.75)

MUT(69±2.2)

WT

15

PMF

WT

BM

90

ND

WT

WT

WT

WT

16

PMF

WT

BM

62

ND

WT

WT

WT

WT

17

ET

MUT(26±1.95)

BM

27

ND

MUT(21±3.5)

MUT(30±2.76)

MUT(23±2.7)

WT

18

PMF

WT

BM

69

46,XY

WT

WT

WT

WT

19

PMF

MUT(57±1.75)

BM

60

COMPLEX

MUT(57±2.5)

MUT(58±2)

MUT(58±2)

WT

20

PV

MUT(56±1.15)

BM

37

46,XX

MUT(55±2.5)

MUT(5612)

MUT(55±2)

WT

21

PMF

MUT(57±1.6)

BM

85

DEL(20)

MUT(55±1.67)

MUT(55±1.25)

MUT(56±1.5)

WT

22

PMF

MUT(56±2.02)

BM

60

ND

MUT(56±0.75)

MUT(55±1)

MUT(56±1.2)

WT

23

PMF

MUT(58±3.05)

BM

100

46,XX

MUT(58±0.5)

MUT(57±1)

MUT(57±1.97)

WT

24

ET

WT

BM

100

46,XX

WT

WT

WT

WT

25

PMF

MUT(52±3.5)

PB

47

46,XX

MUT(50±1.5)

MUT(50±2)

MUT(51±2.3)

WT

26

ET

WT

BM

24

TRI(8)

WT

WT

WT

WT

27

PV

MUT(71±2)

BM

78

46,XX

MUT(69±2)

MUT(65±1.5)

MUT(68±2.3)

WT

28

PV

MUT(59±1.5)

BM

90

46,XX

MUT(58±1.75)

MUT(58±1.5)

MUT(59±1.5)

WT

29

PMF

MUT(61±2)

BM

53

ND

MUT(60±1.15)

MUT(60±1.5)

MUT(59±1.75)

WT

30

PV

MUT(49±3)

BM

72

ND

WT

WT

WT

WT

31

PV

MUT(58±3)

BM

22

ND

MUT(58±2)

MUT(59±2)

MUT(59±0.95)

WT

32

PMF

MUT(56±2)

BM

66

46,XY

MUT(57±2.65)

MUT(55±2)

MUT(55±2.13)

WT

33

PMF

WT

BM

60

46,XY

WT

WT

WT

WT

34

ET

WT

BM

77

t(9:12);TRI(8);ISO(17)

WT

WT

WT

WT

MPN,mye

loproliferative

neoplasia;BM,bonemarrow;PB,peripheralblood;WT,wild

type;MUT,V617F

mutated;AB,allele

burden;SD,standard

deviation;MNCs,

mononucleatedcells;ND,notdone;DEL(x),

deletion;TRI(x),

trisomy;

ISO(x),isochromosome.

letters

384 American Journal of Hematology

34 patients during AML transformation: Eight patients showed additional

abnormalities involving chromosomes 1, 5, 7, 8, 9, 12, 14, 17, and 20 (Table II)

while normal karyotype was confirmed in 14 patients. No other AML-associ-

ated mutations (FLT3, NPM, CEBPA, RUNX1) were detectable at this stage.

To evaluate the modification in the JAK2 status before and during leukemic

transformation we performed ASO-PCR and (QRT)-PCR assay on total BM of

the MPN phase and sorted cell populations from AML phase. No JAK2V617F

mutation was detectable in sorted CD31 cells (Table II). In our cohort of

patients, we found that JAK2V617F mutation was still present at the blast

transformation in both compartments: CD341 cells (blasts) and CD151 cells

(GRA) in 20 of 22 JAK2 mutated MPN (91%) (Table II). Two of 22 patients

(9%) developed JAK2V617F negative AML starting from a mutated PV with a

mean TTP of 5.14 yrs. Interestingly, the WT status was confirmed in blast cells

but also in GRA (Table II). Surprisingly, we found a case of JAK2V617F mutated

AML transforming from a WT-PMF. The negativity for the mutation at MPN stage

was confirmed in two different laboratories using three different PCR assays.

Also in this case the JAK2V617F positivity in the AML phase occurred in both

GRA and blast compartments with an allele burden ratio of 42 and 41%, respec-

tively. The remaining 11 JAK2WT-MPN maintained the same JAK2 status during

blast crisis. To determine whether leukemic transformation could modify the

JAK2-mutational burden we performed (QRT)-PCR assay. For patients with per-

sisting JAK2-mutated disease (n 5 20), no differences (P 5 0.3) in the allele

burden were found comparing MNCs from chronic phase (allelic ratio, median ±

SD: 57% ± 1.75) with MNCs of leukemic transformations (allelic ratio, median ±

SD: 56% ± 1.75). No differences in the allele burden (P 5 0.4) were also found

comparing GRA (allelic ratio, median ± SD: 56 ± 2.5%) with blasts (allelic ratio,

median ± SD: 56 ± 1.5%) in AML phase. These results contrast with previous

studies, in which the JAK2V617F mutation levels decreased during leukemic

evolution [16,17] (data summarized in Table II). At the time of the study, six of 34

patients were eligible for intensive chemotherapy: Four of six patients achieved

complete remission after induction treatment: 3 JAK2mutated-AML from JAK2-

mutated-MPN (2 ET and 1 PV); 1 JAK2WT-AML from JAK2mutated-PMF; 2

JAK2WT-AML from JAK2WT-MPNs obtained partial remission. Twenty-three

patients underwent palliative treatment whereas five deceased before starting

cytotoxic treatment. One patient, JAK2mutated-AML from JAK2mutated-ET

(patient 9), underwent allogeneic stem cell transplantation from sibling donor

and achieved a stable remission. Remarkably, the JAK2V617F mutation was no

longer detectable by quantitative PCR.

In conclusion, these results contrast with the previous study in which the

JAK2 mutation was lost in 53% of blasts during leukemia transformation. In

our work, the loss of JAK2V617F mutation during AML progression is a rare

event (9%). One possible explanation may relate to the different approach

used by Theocharides et al. [15] who first isolated DNA from cells collected

using laser-dissection or brushing from stored BM or PB smears and then

confirmed the results in half of patients in fresh isolated cells. It can be pos-

sible, in theory, that a poor quality-DNA could not be efficiently amplified by

using mutation-specific, single-base mismatched primers while amplification

of wild-type sequence could have been favored. Furthermore, in several

cases laser dissection could have collected a nonrepresentative part of the

leukemic clone considering the nonhomogenous distribution in BM or PB

smears implying an underestimation of the JAK2 mutated blasts, which

could have been diluted in a majority of WT cells and not detected by con-

ventional PCR. For those cases in which the results were confirmed in fresh

cells the source of the cells was often represented by PB-CD341 cells

collected using noncomparable techniques (FACS or MACS), sometimes

mixing PB and BM blasts but no additional information about number and

quality of the cells were provided. When the type of MPN studied is taken

into account one aspect seems of interest: no mutated-PMF patients (n 5

10), who represent the majority of our mutated cases, lost the JAK2V617F

mutation during leukemia. PMF patients are much more likely to transform

than ET and PV, which represent a more benign diseases also more fre-

quently JAK2-mutated, but surprisingly in this view the only two patients

changing the JAK2 status after transformation were PV. This finding is in

accordance with recently published data [18] that demonstrated how JAK2-

mutant AML, but not JAK2-WT AML, usually arises from a preceding myelo-

fibrotic phase. This allow to think that probably the mechanisms of leukemic

transformation are different in PMF comparing with ET and PV and that the

JAK2V617F status of AML secondary to PMF may play a different role

respect to the JAK2V617F status of AML secondary to ET and PV. One

PMF patient developed a secondary mutated-AML from a WT disease. Both

ASO-PCR and real time assays excluded the presence of JAK2V617 muta-

tion during the chronic phase of the MPN in both BM cells and in peripheral

GRA, reinforcing strongly the hypothesis that the mutation could have been

acquired during leukemia switch as it has been reported to occur in 3–4% of

de novo AML [19]. However, the possibility that a very low burden of JAK2

mutated allele during the chronic phase may have gone undetected cannot

be ruled out. One additional finding resulting from this study is that presence

of the JAK2V617F mutation during the MPN phase seems to be associated

with an acceleration to leukemic transformation compared to WT disease,

although the incidence of AML in the two groups was comparable. Additional

studies in larger patient series and multivariate analysis are needed before a

prognostic role of JAK2V617F mutation regarding time to leukemia transfor-

mation can be definitely assessed. However, the persistence of the

JAK2V617F mutation in the actually largest patient series provides a new

intriguing scenery that opens again the discussion.

Patients and Methods

Patients. We performed the molecular analysis on 34 BM aspirates

obtained by patients newly diagnosed with secondary AML (18 derived

from PMF; 9 from PV and 7 from ET). Informed consent for the study

from all patients was obtained in accordance with the Declaration of Hel-

sinki. The study was conducted according to the guidelines of Italian

local ethics committee. The diagnosis of MPN was made according to

WHO [20]. MPN-secondary-AML was diagnosed according to WHO cri-

teria using 20% blasts in BM as cut off. Patient’s characteristics are

given in Table I.

Cell sorting and DNA preparation. Mononucleated cells (MNCs) were

isolated by Ficoll centrifugation. After erythrocyte lyses, 10 million BM (or

PB) cells were labeled by incubation with the antibodies CD15-APC, CD3-

FITC, CD45-PerCP, CD34-Pe (BD Biosciences, San Jose, CA), analyzed

and sorted with a fluorescence-activated cell sorter FACSAria (BD Bioscien-

ces) using FACS Diva software (BD Biosciences), according to manufac-

turer’s recommendations. Leukemic blasts were collected using side scatter

according to CD45 and CD34 expression. Mature cells (GRA) and T-cell

were sorted by FACS using CD15 and CD3 antibodies, respectively. Leuke-

mic blasts and GRA were �98% pure by morphological criteria and FACS

reanalysis. Genomic DNA was isolated using QIAmp DNA Blood Kit

(Qiagen).

ASO-PCR and QRT-PCR. The JAK2V617F mutation status was deter-

mined by ASO-PCR as previously described [2] with a sensitivity of 1%.

Measurement of JAK2V617F allele burden was performed by a quantitative

real time (QRT)-PCR assay, using 80 ng DNA. PCR amplification and detec-

tion were performed on an ABI Prism 7300 analyzer (Applied Biosystem)

using the following cycling conditions: 10 min at 958C followed by 50 cycles

of 15 s at 958C and 60 s at 608C. QRT-PCR assay was performed with

specific forward primers [JAK2-wild type (WT): 50-GCGCGGTTTTAAATTATG

GAGTATGTG-30; JAK2V617F: 50-GCGCGGTTTTAAATTATGGAGTATGTT-30),common reverse primers (50-GCGGTGATCCTGAAACTGAATTTTC-30). FAM-

labeled MGB probe sequence was: 50-TGGAGACGAGAGTAAGTAAAACTA

CAGGCT-30. All samples were analyzed in triplicate and the amount of

JAK2V617F allele was calculated by comparison with cloned JAK2 plasmids.

The mean of triplicate DCT determinations (CTJAK2V617F-CT

JAK2WT) was used

to calculate the percentage of mutated allele. Positive and negative controls

were included in each assay; inter- and intra-assay variation was 3 and 5%,

respectively [5]. To compare continuous variables among the groups, we used

the Mann-Whitney U test and Student’s t-Test.

Acknowledgment

This work was conducted on behalf of the European Leukemia Net (ENL).

Authorship

Contribution: C.R.R. performed research, analyzed data, and wrote the

paper; P.R., F.G. and G.B. performed research and analyzed data; B.M.,

G.S., A.C., L.G. provided patients; M.G., L.D.V. performed the FACS sorting;

A.V. performed real time PCR; T.B. and F.P. designed the research and

wrote the paper.

letters

American Journal of Hematology 385

1Hematology Division, Department of Biochemistry and Medical BiotechnologyUniversity Federico II, Naples, Italy

2Division of HematologyReggio Calabria Hospital, Reggio Calabria, Italy

3Division of Hematology,University of Bari, Bari, Italy4Division of Hematology, Bone Marrow Transplantation Unit, University of Udine,

Udine, Italy5Department of Hematology, Arcispedale Santa Maria Nuova, Reggio Emilia, Italy

6Department of Hematology, University of Florence, Florence, Italy7Division of Hematology, Ospedali Riuniti di Bergamo

Bergamo, Italy*Correspondence to: Dr. Ciro Roberto Rinaldi, Hematology Division - Federico II

University, Via Pansini 5, 80131 Naples – ItalyE-mail: rinaldic@dbbm.unina.it

Conflict of interest: Nothing to report.Received for publication 10 Febuary 2010; Accepted 11 February 2010

Published online 16 February 2010 in Wiley InterScience(www.interscience.wiley.com).

DOI: 10.1002/ajh.21687

References1. Vardiman JW, Harris NL, Brunning RD. The World Health Organization

(WHO) classification of the myeloid neoplasms. Blood 2002;100:2292–2302.2. Baxter EJ, Scott LM, Campbell PJ, et al. Acquired mutation of the tyrosine kin-

ase JAK2 in human myeloproliferative disorders. Lancet 2005;365:1054–1061.3. Kralovics R, Passamonti F, Buser AS, et al. A gain-of-function mutation of

JAK2 in myeloproliferative disorders. N Engl J Med 2005;352:1779–1790.4. Levine RL, Wadleigh M, Cools J, et al. Activating mutation in the tyrosine kin-

ase JAK2 in polycythemia vera, essential thrombocythemia, and myeloidmetaplasia with myelofibrosis. Cancer Cell 2005;7:387–397.

5. Lippert E, Boissinot M, Kralovics R, et al. The JAK2-V617F mutation is fre-quently present at diagnosis in patients with essential thrombocythemia andpolycythemia vera. Blood 2006;108:1865–1867.

6. Veneri D, Capuzzo E, de Matteis G, et al. Comparison of JAK2V617F muta-tion assessment employing different molecular diagnostic techniques. BloodTransfus 2009;7:204–209.

7. Schnittger S, Bacher U, Kern W, et al. JAK2V617F as progression marker inCMPD and as cooperative mutation in AML with trisomy 8 and t(8;21): A

comparative study on 1103 CMPD and 269 AML cases. Leukemia 2007;21:1843–1845.

8. Rossi D, Deambrogi C, Capello D, et al. JAK2 V617F mutation in leukaemictransformation of philadelphia-negative chronic myeloproliferative disorders.Br J Haematol 2006;135:267–268.

9. Panovska-Stavridis I, Cevreska L, Ivanovski M, et al. JAK2V617F mutationsin myeloid malignancies: Single center experience. Prilozi 2009;29:257–268.

10. Campbell PJ, Baxter EJ, Beer PA, et al. Mutation of JAK2 in the myeloprolifer-ative disorders: Timing, clonality studies, cytogenetic associations, and role inleukemic transformation. Blood 2006;108:3548–3555.

11. Frohling S, Lipka DB, Kayser S, et al. Rare occurrence of the JAK2V617F mutation in AML subtypes M5, M6, and M7. Blood 2006;107:1242–1243.

12. Hsiao HH, Yang WC, Liu YC, et al. Disappearance of JAK2 V617F mutationin a rapid leukemic transformed essential thrombocythemia patient. Leuk Res2008;32:1323–1324.

13. Iwanaga E, Nanri T, Matsuno N, et al. A JAK2-V617F activating mutation inaddition to KIT and FLT3 mutations is associated with clinical outcome inpatients with t(8;21)(q22;q22) acute myeloid leukemia. Haematologica 2009;94:433–435.

14. Lee JW, Kim YG, Soung YH, et al. The JAK2 V617F mutation in de novoacute myelogenous leukemias. Oncogene 2006;25:1434–1436.

15. Theocharides A, Boissinot M, Girodon F, et al. Leukemic blasts in trans-formed JAK2-V617F-positive myeloproliferative disorders are frequently nega-tive for the JAK2-V617F mutation. Blood 2007;110:375–379.

16. Zhang SJ, Li JY, Zhang JF, et al. The decrease of JAK2 V617F allele burdenin leukemia transformation of an elderly patient with myelofibrosis. Leuk Res.2009;33:e116–e118.

17. Hussein K, Bock O, Theophile K, et al. JAK2 (V617F) allele burden discrimi-nates essential thrombocythemia from a subset of prefibrotic-stage primarymyelofibrosis. Exp Hematol 2009;37:1186–1193.

18. Beer PA, Delhommeau F, Lecouedic JP, et al. Two routes to leukemic trans-formation following a JAK2 mutation-positive myeloproliferative neoplasm.Blood 2009; [Epub ahead of print].

19. Lee JW, Kim YG, Soung YH, et al. The JAK2 V617F mutation in de novoacute myelogenous leukemias. Oncogene 2006;25:1434–1436.

20. Tefferi A, Thiele J, Vardiman JW. The 2008 World Health Organization classi-fication system for myeloproliferative neoplasms: Order out of chaos. Cancer2009;115:3842–3847.

Persistent splenomegaly during imatinib therapy and thedefinition of complete hematological response in chronicmyelogenous leukemia

Zdenek Racil,1* Hana Klamova,2 Jaroslava Voglova,3 Edgar Faber,4 Filip Razga,1 Daniela Zackova,1

Lucie Buresova,5 Petr Cetkovsky,2 and Jiri Mayer1

Splenomegaly belongs among typical findings on physical examina-

tion in patients with newly diagnosed chronic myelogenous leukemia

(CML) [1]. Its disappearance is a part of achieving complete hemato-

logical response (CHR), that is nowadays (when second generation of

tyrosine kinase inhibitors are available) of particular interest during

imatinib treatment. However, the kinetics of the disappearance of sple-

nomegaly in patients with CML has still never been studied. We have

analyzed 20 out of 245 patients with newly diagnosed chronic phase

CML that had a still palpable spleen at the 3rd month of imatinib ther-

apy in terms of treatment response at 18 months from the start of ther-

apy. Our analysis have showed that eight (40%) of these 20 patients

had achieved a treatment response at these time points. Moreover 11

patients had still a palpable splenomegaly at the 6th month after the

start of imatinib therapy and six (54%) of them had a therapeutic

response at the 18th month, suggesting that slower spleen shrinkage

in patients with newly diagnosed chronic phase CML does not neces-

sarily mean the failure of the therapy in the future.

Very recently Baccarani et al. [2] postulated the new European Leukemia-

Net (ELN) criteria for the treatment response in patients with newly diagnosed

CML in the early chronic phase. Compared to previous recommendations [3]

from 2006 prepared by the same group, there is (among others) one important

change: the nonachievement of CHR at the 3rd month of treatment is now

defined as a ‘failure of therapy’. In the 2006 criteria [3], the absence of CHR at

the 3rd month was defined as a suboptimal response and failure of therapy

was considered only when the CHR was not present at the 6th month. In

some patients, even this 2006 criterion could be very strict to define failure.

CHR definition namely includes the complete disappearance of initial spleno-

megaly. However, to our best knowledge there are no clear data about the

kinetics of the shrinkage of an enlarged spleen in patients with CML during dif-

ferent types of therapy, including tyrosine kinase inhibitors; namely imatinib.

We would therefore like to point out that while the achievement of CHR within

the meaning of blood count parameters is undoubted, splenomegaly may per-

sist for several months even after the normalization of blood count and may

not be a clear criterion for failure of this therapy in the future.

We have analyzed data from 245 patients with newly diagnosed chronic

phase CML treated between 2005 and 2008. We have focused on the pres-

ence of palpable splenomegaly during physical examination at the 3rd month

of imatinib treatment and its further disappearance as a prognostic marker

for imatinib treatment response.

One hundred and one (41%) out of 245 patients had a palpable spleen at

the time of diagnosis and the start of imatinib treatment.

At the 3rd month after the start of imatinib therapy—the time point

when CHR should have been achieved according to 2009 ELN criteria

[2]—20 patients (8% from total number of our patients and 20% from the

group of patients with initial splenomegaly) still had a palpable spleen.

These patients were further analyzed in terms of treatment response at

letters

386 American Journal of Hematology

18 months from the start of therapy, as well as at the last available fol-

low-up. Baseline clinical characteristics of these 20 patients are given in

Table I. The median length of imatinib treatment at the time of evaluation

or at the moment of treatment change in this group of patients was 19.5

months (range 3–36 months).

At 18 months from the start of imatinib treatment, 10 (50%) out of these

20 patients with persistent splenomegaly at the 3rd month had been classi-

fied (according to ELN criteria [2,3]) as ‘imatinib failure’ (including nine

patients with change of treatment within the 3rd–18th month). However, 8

(40%) had achieved a treatment response (5 optimal, 3 suboptimal) accord-

ing to these criteria and two were not evaluable (cytogenetic data are miss-

ing). At the last follow-up (median 30 months; range 16–42 months) of these

20 patients with splenomegaly at the 3rd month, 10 had still been on stand-

ard dose of imatinib and eight (40%) had achieved a treatment response

(five optimal, two suboptimal, and one suboptimal with very recently

detected increase in BCR-ABL transcript levels).

There weren’t any significant differences among spleen sizes below the

left costal margin at the time of diagnosis in patients with persisting spleno-

megaly at the 3rd month that had failed (median 13 cm; range 3–19 cm) or

hadn’t failed (median 12 cm; range 1–18 cm) the imatinib treatment at 18

months (P 5 0.360; Mann-Whitney test). A high Sokal score index at the

time of diagnosis was more frequent among patients with splenomegaly at 3

months that were classified as failure at 18 months (90%) compared to

patients that did not fail (40%) at this point of time (P 5 0.043; Fisher exact

test).

Between the 3rd and the 6th months splenomegaly had disappeared in

nine of the 20 presented patients, but was still palpable in 11 of them and

thus these patients fulfilled the criterion of ‘‘failure’’ according to the 2006

ELN criteria [3]. However, when these patients were followed up at 18

months, six (54%) had a therapeutic response to imatinib according to the

ELN criteria [2,3], three (27%) had failed, and two were not evaluable (cyto-

genetic data are missing). Finally, one patient out of 20 with splenomegaly

at the 3rd month had a palpable spleen at 12 months. This patient had an

optimal response at 18 months (Table II).

During the natural course of CML, the spleen is infiltrated by Philadelphia-

chromosome (Ph) positive cells, causing its enlargement. Interestingly, the

spleen may be more important than the bone marrow for futher karyotypic

evaluation of Ph positive cells that announce the transformation to advanced

phases of the disease [4,5].

We know from thework of Kamada et al.[6,7] and other researchers, that spleen

enlargement correlates with the grade of leukocytosis as a marker of disease bur-

den, as well as with the progression of CML to advanced stages and thus with the

prognosis of patients with this malignancy. However, data showing that slower

treatment response in splenomegaly correspondswith imatinib failure aremissing.

Our data, from patients with chronic phase CML that were treated with

front-line imatinib with persistent splenomegaly at the 3rd month, denotes

that even though the frequency of imatinib failure at the 18th month seems

to be higher than in the general population of patients with chronic phase

CML, solely its presence at the 3rd month does not necessarily mean failure

of imatinib therapy in the future in each individual patient (Table III).

In the view of published recommendations of Baccarani et al. [2] all 20

presented patients that do not reach the CHR due to persistent splenome-

galy at the 3rd month would have been classified as ‘failure of imatinib ther-

apy’ and should have been switched to the second generation of tyrosine

kinase inhibitors or allogeneic stem cell transplantation should be offered to

them. However, as showed, 40% of them did not need this change in ther-

apy during all the observed period and might not need it in the future.

Thus, in our opinion and based on our data mentioned above, previous

recommendations from ELN [3] where not achieving CHR at the 3rd month

of imatinib treatment was only the ‘suboptimal response’ were probably

more appropriate for routine clinical practice.

Acknowledgments

We would like to thank Ms. Shira Timilsina for the English correction of

the manuscript. This work was supported by CELL - The CzEch Leukemia

Study Group for Life.

1Department of Internal Medicine Hemato-Oncology, Masaryk University andUniversity Hospital Brno, Brno, Czech Republic; 2Institute of Hematology andBlood Transfusion, Prague, Czech Republic; 3Department of Internal Medicine

Hematology; University Hospital Hradec Kralove, Hradec Kralove, Czech Republic4Department of Hemato-Oncology,UniversityHospital Olomouc,Olomouc,CzechRepublic5Institute of Biostatistics and Analyses at the Faculty of Medicine, and the Faculty

of Science of the Masaryk University, Brno, Czech Republic*Correspondence to: Zdenek Racil, Department of Internal Medicine, Hemato-Oncology, University Hospital Brno, Jihlavska 20, 625 00 Brno, Czech Republic

E-mail: zracil@fnbrno.czConflict of interest: Nothing to report.

Received for publication 4 February 2010; Accepted 11 February 2010Published online 16 February 2010 in Wiley InterScience

(www.interscience.wiley.com).DOI: 10.1002/ajh.21689

TABLE I. Patients Characteristics

Characteristics No. (%) or median [range]

Total number of patients 245No. of pts. with splenomegaly at diagnosis 101 (41%)No. of patients with splenomegaly at the 3rd month 20 (8% of total No. of pts. and 20% of pts.

with splenomegaly at diagnosis)Male/female 8/12Sokal index low/intermediate/high 2 (10%)/5 (25%)/13 (65%)Additional cytogenetic abnormalities in Ph1 clone 2 (10%)a

Median hemoglobin level at the time of diagnosis (g/l) 93 [60–158]Median leukocyte count at the time of diagnosis (10.9/l) 110 [42–612]Median thrombocyte count at the time of diagnosis (10.9/l) 394 [95–1695]Median spleen size at the time of diagnosis (cm) 13 [1–19]Median spleen size at the 3rd month of imatinib treatment (cm) 4.5 [1–15]

No. of patients with splenomegaly at the 6th month 11 (4.5% of total No. of pts. and 11% of pts.with splenomegaly at diagnosis)

Median spleen size at the 6th month of imatinib treatment (cm) 1 [1–6]No. of patients with splenomegaly at the 12th month 1 (0.4% of total No. of pts. and 1% of pts.

with splenomegaly at diagnosis)

aonly 19 patients were evaluable for this analysis

TABLE II. Response of Patients with Persistent Splenomegaly at the 3rd, 6th, and 12th Months Assessed at 18 Months After the Start of Imati-

nib Treatment According to European LeukemiaNet Criteria [2]

Response at 18 months Optimal Suboptimal Failure Not evaluable

Presence of splenomegaly at the 3rd month (n 5 20) 5 (25%) 3 (15%) 10 (50%) 2 (10%)Presence of splenomegaly at the 6th month (n 5 11) 4 (36%) 2 (18%) 3 (27%) 2 (18%)Presence of splenomegaly at the 12th month (n 5 1) 1 (100%) – – –

letters

American Journal of Hematology 387

TABLE

III.

Patients

with

Splenomegaly

atthe3rd

Month

-BaselineCharacteristics,WhiteBlood

CellCount,

WhiteBlood

CellDifferentialCount,

CytogeneticResponse,MolecularResponse,and

Treatm

entofChronic

MyelogenousLeukemia

atthe3rd,6th,12th,18th

MonthsandattheLastFollow-up

Grayfields-notperform

ed.Black

fields-notapplicable.WBC,whitebloodcells;WBC

diff.,whitebloodcelldifferentialcount;eos.,eosinophiles;baso.,basophiles;mye

loc.,mye

locytes;minim

.CgR,minim

alcytogeneticresponse;PCgR,

partialcytogeneticresponse;CCgR,complete

cytogeneticresponse;MMoR,majormolecularresponse;CMoR,complete

molecularresponse.*S

pleensizein

centimeters

below

costalmargin.aPatientrefusedbonemarrow

examination

afterthe3rd

month.

letters

388 American Journal of Hematology

References1. O’Brien SG, Guilhot F, Larson RA, et al. Imatinib compared with interferon

and low-dose cytarabine for newly diagnosed chronic-phase chronic myeloidleukemia. N Engl J Med 2003;348:994–1004.

2. Baccarani M, Cortes J, Pane F, et al. Chronic myeloid leukemia: An update ofconcepts and management recommendations of European LeukemiaNet. JClin Oncol 2009;27:6041–6051.

3. Baccarani M, Saglio G, Goldman J, et al. Evolving concepts in the man-agement of chronic myeloid leukemia: Recommendations from an expertpanel on behalf of the European LeukemiaNet. Blood 2006;108:1809–1820.

4. Stoll C, Oberling F, Flori E. Chromosome analysis of spleen and/or lymph

nodes of patients with chronic myeloid leukemia (CML). Blood 1978;52:828–838.

5. Skorski T, Nieborowska-Skorska M, Calabretta B. A model of Ph’ positivechronic myeloid leukemia-blast crisis cell line growth in immunodeficient SCIDmice. Folia Histochem Cytobiol 1992;30:91–96.

6. Kamada N, Uchino H. Chronologic sequence in appearance of clinical and

laboratory findings characteristic of chronic myelocytic leukemia. Blood 1978;

51:843–850.

7. Sokal J. Prognosis in chronic myeloid leukaemia: Biology of the disease vs.

treatment. Baillieres Clin Haematol 1987;1:907–929.

Successful unmanipulated stem cell transplantation fromHLA-haploidentical 3-loci-mismatched parents in two childrenwith severe aplastic anemia not responding toimmunosuppressive therapyZhidong Wang, Hongmin Yan, Ling Zhu, and Hengxiang Wang*

Hematopoietic stem cell transplantation (HSCT) and immunosuppres-

sive therapy (IST) are two therapeutic modalities for severe aplastic

anemia (SAA). Several reports indicate that 2- to 5-year overall sur-

vival for young patients following IST who lack HLA-matched sibling

donors is between 60 and 90% [1,2]. The optimal treatment for the

patients who have no response to IST has not been established. Hap-

loidentical stem cell transplantation (haplo-HSCT) might be a feasible

option as it has been successfully used in the management of malig-

nant hematological diseases, and the transplantation protocols have

been greatly improved since it was first introduced into clinical use

by Aversa et al. [3]. Here, we report successful haplo-HSCT in two

children with SAA who were refractory to IST and lacked HLA-identi-

cal donors.

The conditioning regimen consisted of busulhan (Bu, 4 mg/kg/day) from

Day 210 to Day 29, fludarabine at 30 mg/m2/day from Day 210 to Day

27, cyclophosphamide (CY) at 50 mg/kg/day from Day 26 to Day 23,

rabbit ATG (Fresenius, Munich, Germany) at 5 mg/kg/day or swine ALG

(Wuhan, China) at 20 mg/kg/day from Day 24 to Day 21. Both patients

received cyclosporine (CsA), methotrexate (MTX), mycophemolate mofeil

(MMF), and basiliximab for graft-versus-host disease (GVHD) prophylaxis.

Basiliximab was given at 20 mg/day on Days 1 and 4. Additionally, mes-

enchymal stem cells (MSCs) derived from the donor’s bone marrow were

cultured as previously described, and a total dose of 1.25 3 106/kg was

infused on Day 1 in Case 2 to prevent graft rejection and reduce GVHD

[4]. Donors received 5 mg/kg/day granulocyte-colony stimulating factor for

5 consecutive days. Bone marrow was collected on Day 1, whereas

peripheral blood stem cell was harvested on Day 2. Engraftment was

defined as a sustained (3-day) neutrophil count �0.5 3 109/L and an

unsupported platelet count (7-day) �20 3 109/L.

Characteristics of the patients are shown in Table I. Case 1: A 5-year-

old boy was diagnosed as idiopathic SAA in September 2007. He received

IST consisting of ATG and CsA with granulocyte-colony stimulating factor

in October 2007, and no hematological response was observed. From Jan-

uary 2008, he suffered from recurrent fever around 398C, while no infection

was found. In April 2008, he transferred to our hospital, and multiple pyo-

genic abscesses in the liver were identified. He responded effectively to the

antibiotics treatment including ampicillin/sulbactam and imipenem/cilastatin

after 2 months. However, his white blood cell count was 0.9 3 109/L with

1.9% neutrophils and transfusion dependent. The number of packed red

cells and platelet transfusions received before HSCT was 27 and 48, respec-

tively. HLA-identical sibling and unrelated donors were unavailable. He

received haplo-HSCT from his 3-loci-mismatched mother in July 2008. Graft

included 5.78 3 108/kg bone marrow total nucleated cells (TNCs) and 11.60

3 108/kg peripheral blood TNCs. The percentage of CD34 cell was 0.12

and 0.30%, respectively. Total infused CD31 T cells were 3.53 3 108/kg.

Case 2: A 3-year-old girl was diagnosed as idiopathic SAA in August

20 08. After CsA and androgens treatment for 3 months, her reticulocyte

counts once increased to 2.22%, whereas it decreased to 0.07% for

recurrent respiratory tract infection. Her white blood cell count was 1.8

3 109/L with 2.8% neutrophils and transfusion dependent. The number

of packed red cells and platelet transfusions received before HSCT was

21 and 24, respectively. The HLA-identical unrelated donor was unavail-

able. She underwent haplo-HSCT from his 3-loci-mismatched father in

April 2009. Graft included 1.98 3 108/kg bone marrow TNCs and 11.78

3 108/kg peripheral blood TNCs. The percentage of CD34 cell was

0.03 and 0.78%, respectively. Total infused CD31 T cells were 2.51 3

108/kg.

Regimen-related toxicity was mild including vomiting and fever for ATG/

ALG. Both patients experienced febrile neutropenia for several days. The

time for neutrophil and platelet recovery was 12 and 17 days in Case 1 and

that was 11 and 14 days in Case 2. Chromosomal analysis indicated com-

plete donor-type engraftment. Grade II skin rash acute GVHD developed in

Case 1 and was controlled completely. The liver abscess disappeared after

antibiotics treatment for further 5 months. Hepatic chronic GVHD presented

with elevated transaminases, and cholestasis was observed at Month 4 and

resolved after 1 month later. No acute or chronic GVHD occurred in Case 2.

Immunosuppression was continued for 9 and 5 months posttransplant in

Case 1 and Case 2, respectively. Now, they are alive in good performance

with normal blood cell counts in good performance status for 20 and

8 months, respectively.

Haplo-HSCT is an alternative transplant strategy characterized by severe

GVHD, high graft rejection rate, and impaired immune reconstitution [3]. It

was mainly performed in patients with advanced hematological malignancies

lacking HLA-identical donors. Recently, it has also been emerged as a novel

strategy for SAA patients who had failed intensive IST and lacked unrelated

donors [5,6]. However, there have not yet been large successful reports of

haplo-HSCT for patients with SAA [7]. It was limited by high risk of graft

rejection and severe GVHD, especially in heavily transfused patients with

refractory SAA. In this preliminary experience with haplo-HSCT for SAA,

both patients achieved rapid myeloid/platelet engraftment and survived in

good performance.

Recently, a novel study indicated that the addition of fludarabine and

ATG to CY in conditioning regimen may contribute to low graft failure with

less toxicity in children undergoing allogeneic SCT for SAA [8]. Fludarabine,

a purine analog, is endowed with potent immunosuppressive activity, inhibits

or eradicates recipient’s T cells, and enhances engraftment. Fludarabine-

based conditioning regimen might reduce the graft failure incidence from

11 to 0% comparing with standard regimen in HLA-identical SCT for SAA

[9]. The overall survival was superior also. In our study, both children condi-

tioned with Bu/Fludarabine/CY/ATG protocol achieved sustained engraftment

that the time of hematological recovery was comparable to that in HLA-

letters

American Journal of Hematology 389

identical SCT. Additionally, only mild regimen-related toxicities were found in

these two patients, indicating that this regimen is an acceptable for haplo-

HSCT in refractory SAA.

For aGVHD prophylaxis, basiliximab or with MSCs was used in the two

patients. Basiliximab, which binds specially to IL-2 receptor-a of activated

T cells with high affinity, is a potent inhibitor of IL-2-mediated T-cell activation

and proliferation. Our previous report demonstrated a significant reduction in

the Grades II–IV aGVHD incidence from 33 to 11% when the combination of

basiliximab and CsA 1 MTX 1 MMF was administered following haploidenti-

cal SCT [10]. In this study, there was no severe aGVHD occurring both in

Case 1 and Case 2, suggesting that basiliximab may exert GVHD prevention

in SAA transplantation although high-dose haplotype TNCs and T cells were

infused.

MSCs administration has been increasingly reported to prevent GVHD in

clinical trials [11]. MSCs exert their immunosuppressive effect by secreting

soluble factors such as hepatocyte growth factor and transforming growth

factor-a. In this study, MSCs of donor origin were infused in Case 2. No

GVHD occurring in Case 2 indicated MSCs play roles in prophylaxis of

GVHD.

In conclusion, although this preliminary experience involves only two

patients, several important findings should be highlighted. First, a condition-

ing regimen consisting of fludarabine, CY, ATG, and low-dose busulphan is

powerfully immunosuppressive and enable for stable engraftment. Second,

based on these first two patients, the use of basiliximab and MSCs will con-

tinued to be evaluated as GVHD prevention. Further study is to establish

that haplo-HSCT is a rational choice for patients with refractory SAA who

lack HLA-identical sibling or unrelated donors.

Department of Hematology, The General Hospital of Air Force P.L.A.,Beijing, China

*Correspondence to: Hengxiang Wang, Department of Hematology, The GeneralHospital of Air Force P.L.A., Beijing 100142, China

E-mail: wanghengxiang123@yahoo.com.cnPublished online 22 February 2010 in Wiley InterScience

(www.interscience.wiley.com).DOI: 10.1002/ajh.21693

Conflict of interest: Nothing to report.

References1. Gurion R, Gafter-Gvili A, Paul M, et al. Hematopoietic growth factors in aplas-

tic anemia patients treated with immunosuppressive therapy-systematicreview and meta-analysis. Haematologica 2009;94:712–719.

2. Osugi Y, Yagasaki H, Sako M, et al. Antithymocyte globulin and cyclosporinefor treatment of 44 children with hepatitis associated aplastic anemia. Haema-tologica 2007;92:1687–1690.

3. Aversa F, Tabilio A, Velardi A, et al. Treatment of high-risk acute leukemiawith T-cell-depleted stem cells from related donors with one fully mismatchedHLA haplotype. N Engl J Med 1998;339:1186–1193.

4. Jin JD, Wang HX, Xiao FJ, et al. A novel rich source of human mesenchymalstem cells from the debris of bone marrow samples. Biochem Biophys ResCommun 2008;376:191–195.

5. Dey BR, Spitzer TR. Current status of haploidentical stem cell transplantation.Br J Haematol 2006;135:423–437.

6. Tsutsumi Y, Tanaka J, Miura T, et al. Successful non-T-cell-depleted nonmye-loablative hematopoietic stem cell transplantation (NST) from an HLA-haploi-dentical 2-loci-mismatched sibling in a heavily transfused patient with severeaplastic anemia based on the fetomaternal microchimerism. Bone MarrowTransplant 2004;34:267–269.

7. Lacerda JF, Martins C, Carmo JA, et al. Haploidentical stem cell transplanta-tion with purified CD341 cells after a chemotherapy-alone conditioning regi-men in heavily transfused severe aplastic anemia. Biol Blood Marrow Trans-plant 2005;11:399–400.

8. George B, Mathews V, Viswabandya A, et al. Fludarabine based reducedintensity conditioning regimens in children undergoing allogeneic stem celltransplantation for severe aplastic anemia. Pediatr Transplant 2008;12:14–19.

9. Maury S, Bacigalupo A, Anderlini P, et al. Improved outcome of patients olderthan 30 years receiving HLA-identical sibling hematopoietic stem cell trans-plantation for severe acquired aplastic anemia using fludarabine-based condi-tioning: A comparison with conventional conditioning regimen. Haematologica2009;94:1312–1315.

10. Ji SQ, Chen HR, Yan HM, et al. Anti-CD25 monoclonal antibody (basiliximab)for prevention of graft-versus-host disease after haploidentical bone marrowtransplantation for hematological malignancies. Bone Marrow Transplant2005;36:349–354.

11. Tisato V, Naresh K, Girdlestone J, et al. Mesenchymal stem cells of cordblood origin are effective at preventing but not treating graft-versus-host dis-ease. Leukemia 2007;21:1992–1999.

TABLE I. Characteristics and Transplant Outcome of Two Patients

Case 1 Case 2

Age(years)/Sex 5/Male 3/FemaleDisease status Very severe Very severeTreatment protocol before HSCT ATG 1 CSA CSA 1 androgensRecipient-HLA A 33/11 A 02/11

B 61/52 B 40/52DRB1 9/15 DRB1 08/09

Donor age/Sex 26/Female 27/MaleDonor-HLA A 33/2 A 33/02

B 61/58 B 40/58DRB1 17/9 DRB1 03/08

ABO blood type (donor/recipient) O/B B/OGraft Bone marrow: Bone marrow:

CD341 0.69 3 106/kg CD341 0.06 3 106/kgTNC 5.78 3 108/kg TNC 1.98 3 108/kgPeripheral blood: Peripheral blood:CD3413.48 3 106/kg CD341 9.12 3 106/kgTNC 11.60 3 108/kg TNC 11.78 3 108/kgCD313.53 3 108/kg CD312.51 3 108/kg

Neutrophil engraftment Day 12 Day 11Platelet engraftment Day 17 Day 14Acute GVHD Grade II NoneChronic GVHD Limited NoneFollow-up (months) 20 8

letters

390 American Journal of Hematology