Original article
Addition of etoposide to CHOP chemotherapy in untreated patients
with high-grade non-Hodgkin's lymphoma
F. Celsing,
1S.Widell,
1K. Merk,
2tP. Bernell,
3G. Grimfors,
1A. Hedlund,
3J. Liliemark,
4E. Svedmyr,
4E. Osby
1& M. Bjorkholm
1Department of Hematology and Infectious Diseases, Karolinska Hospital; 2 Department of Oncology, South Hospital, Stockholm;
3
Section ofHematology, Department of Medicine, Danderyd Hospital, Danderyd; 4Department ofGeneral Oncology, Karolinska Hospital, Stockholm, Sweden; ^deceased
* Seepage 1217 for list of participating centers
Summary
Background: Second- and third-generation chemotherapy
pro-tocols for the treatment of aggressive non-Hodgkin's
lympho-mas (NHL) have considerable, and age-related, toxic effects.
In addition, they do not seem to prolong overall survival in
comparison to standard CHOP chemotherapy. In this phase II
study we investigated the feasibility and efficacy of the addition
of etoposide to the conventional CHOP regimen.
Patients and methods: Toxicity and clinical efficacy were
determined in 132 patients with previously untreated
high-grade NHL. There were 51 patients in clinical stage I and II
and 81 patients in stage III and IV, with a median age of 54
years (range 17-85). Patients received standard-dose CHOP
plus etoposide 100 mg/m
2i.v. on day 1 and 200 mg/m
2p.o. on
days 2-3.
Results: The overall response rate was 84%, with 70%
com-plete and 14% partial responses. The predicted three- and
five-year survivals for the group as a whole were 60% and 53%,
respectively, and the corresponding disease-free survivals for
patients achieving complete remissions were 65% and 56%,
respectively. Outcome was not different from that of
CHOP-treated patients in a recently completed Nordic study
per-formed during the same time period. Myelosuppression (WHO
grade 3—4), observed in 87% of patients and infectious
compli-cations (WHO grade 3-4) in 33%, dominated the toxicity profile
of this regimen. Fifty-seven of 92 complete responders (62%)
received 6-8 CHOP-E cycles with no reductions in planned dose
intensity. LDH level higher than normal, extranodal sites = 2,
stage III-IVat diagnosis were all indicators of a poor survival.
Conclusions: We conclude that CHOP-E treatment is
effec-tive in high-grade NHL. However, mainly due to severe
mye-losuppression frequent schedule modifications were required
and the results are not obviously superior to those of
conven-tional CHOP.
Key words: chemotherapy, CHOP, etoposide, high-grade
non-Hodgkin's lymphoma, risk factors
Introduction
The introduction of multiagent chemotherapy has
radi-cally improved the prognosis of patients with high-grade
malignant non-Hodgkin's lymphoma (NHL). More than
two decades ago McKelvey et al. [1] reported on the
successful use of the first-generation regimen, CHOP,
combining cyclophosphamide, doxorubicin, vincristine,
and prednisone. Durable complete remissions were seen
in 30% of patients. Since then the results of studies
employing a variety of second- and third-generation
chemotherapeutic regimens have indicated that
long-term complete responses can be achieved in 45% to 65%
of patients [2-6]. Most of these prospective studies have
been performed in single institutions and so far the results
have contrasted with those of prospective randomised
trials comparing first- with second- and third-generation
chemotherapy programs [7-10]. Furthermore, the
tox-icity of these newer drug combinations is considerable
and patients frequently require institutional care. It is
also well recognised that most patients > 60 years of age
have a poor tolerance for this type of intensive
chemo-therapy, primarily because of the attendant severe
mye-losuppression and high-dose anthracycline and vinca
alkaloid-related toxicity [4, 5, 11]. Thus, despite their
poor prognoses, elderly patients with or without
con-current disorders are frequently excluded from trials
using intensive regimens.
Among the drugs which are effective in the treatment
of aggressive NHL, etoposide, together with
doxorubi-cin and cyclophosphamide, has become the most firmly
established. Its efficacy as a single agent has been
docu-mented [12], also in refractory NHL [13]. However, most
of the patients in studies adding etoposide to
conven-tional CHOP chemotherapy have had refractory or
re-lapsed disease. Furthermore, in studies where CHOP/
etoposide has been given up-front, modifications of the
CHOP component have been introduced [14-18]. Thus,
it seemed logical to explore the addition of etoposide
(given over three days) to standard CHOP chemotherapy
Table 1 Patient characteristics. Number of patients Total 132 Age (years) Median 54 Range 17-85 <60 82 5=60 50 Sex Male 74 Female 58 Histological subtype Centroblastic 72 Immunoblastic 7 Lymphoblastic 19 Large cell anaplastic 11 Pleomorphic T cell 5 Unclassified B cell 10 Unclassified T cell 4 Unclassified T & B 4 Clinical stage I 23 II 28 III 36 IV 45 A 74 B 58 Bulky mass > 10 cm 29 Bulky mass not evaluated 2 LDH above normal limit ( > 8.0 ukat/1) 77 LDH not done 10 Extra nodal sites
None 42 One site 47 Ss 2 sites 43
with its well-established efficacy and limited toxicity in
an effort to increase its anti-neoplastic effect in
aggres-sive NHL. This regimen, intended for outpatient use,
looked promising in relapsing patients [19]. We report
the therapeutic efficacy, toxicity and risk factors in
pre-viously untreated patients with aggressive NHL receiving
CHOP-E.
the WHO criteria. Patient data did not permit a clear distinction between performance statuses 2 and 3, which is why this variable was not included in the multivariate prognostic factor analysis.
The predominant histologic subtype was centroblastic NHL (55%). Sixty-one percent of patients had stage III—IV disease, 44% had B symptoms and 22% presented with bulky disease.
Treatment
The CHOP-E regimen consisted of cyclophosphamide 750 mg/m2 i.v., doxorubicin 50 mg/m2 i.v., vincristine 1.4 mg/m2 i.v. (maximum 2.0 mg), and etoposide 100 mg/m2 i.v., all administered on day 1. Oral etoposide 200 mg/m2 was given on days 2 and 3 and oral prednisone 50 mg/m2 on days 1-5. The doxorubicin and etoposide doses were reduced by 25% if infectious complications and/or severe neutropenia (^0.5 x 109/l > 6 days) were recorded after the preceding course of chemotherapy. Treatment was repeated every three weeks. Patients were restaged after four courses or earlier if clinically indicated and after completion of the therapy. Patients with a response less than partial remission (PR) after four courses were excluded from further CHOP-E chemotherapy and were treated according to the discretion of the investigator. However, most non-responding patients were given the MIME regimen [22]. No antiviral, antibacterial or antifungal prophylaxis was used. Involved field radiation was given for bulky disease at presentation.
The use of hematopoietic growth factors was not restricted and 29% of patients received G- or GM-CSF after one or several chemo-therapy courses. Growth factors were predominantly given to patients with a prior chemotherapy-related infectious complication, often in combination with a CHOP-E dose attenuation.
Eleven patients received high-dose chemotherapy with or without total body irradiation and blood stem cell or bone marrow trans-plantation (nine autologous and two allogeneic; one patient in first CR and 10 patients in second CR).
Response evaluation
At restaging, complete remission (CR) was defined as the complete disappearance of all clinical disease manifestations and the reversal of all previously abnormal investigations. All patients achieving CR re-ceived two to four additional CHOP-E cycles. Partial remission (PR) was denned as a reduction of > 50% of all measurable tumours with-out new lesions. Response less than PR, but not fulfilling the criteria for progressive disease (PD) was defined as stable disease (SD). PD was defined as an increase in size ( > 25%) of one or more measurable lesions or the appearance of new lesions.
Patients and methods
Patients
A total of 132 patients ( > 16 years) with previously untreated high-grade NHL were included. The clinical characteristics of the patients receiving CHOP-E are listed in Table 1. The median age was 54 years (range 17-85) with 50 patients (38%) being >60 years. The median follow-up of surviving patients was 53 months (range 12+-95+).
The diagnosis was based on morphological and immunohisto-chemical examination of tumour biopsy material. For clinical reasons the diagnostic material was limited in a few patients to several needle aspiration biopsies. Diagnostic specimens were classified according to the updated Kiel classification [20]. The pre-treatment evaluation consisted of physical examination, hematological and chemical sur-veys, chest X-ray, computerised tomography of the abdomen and pelvis and when clinically indicated of the chest, and bone marrow biopsy. None of the patients underwent staging laparotomy. Bulky disease was defined as a tumour mass > 10 cm. The Ann Arbor staging classifica-tion was used [21]. Toxicity and performance status were assessed by
Statistics
Overall survival (OS) and disease free-survival (DFS) were calculated according to the method of Kaplan and Meier [23]. OS was calculated from the start of therapy and DFS from the onset of CR to date of relapse; two patients who died without evidence of lymphoma (myo-cardial infarction and septicemia not related to treatment) were cen-sored at the time of death. Two emigrating patients were cencen-sored at the time of their last follow-up. The log-rank test was used to assess the significance of differences in OS and DFS [24]. A non-parametric chi-square test was used for 2 x 2 tables.
Results
Response to therapy
Ninety-two patients (70%) achieved CR and 14% PR,
for an overall response rate of 84%, while 16% failed to
Table 2. Response to treatment.
Total number of patients (n = 132) B-cell lymphoma (n = 105) T-cell Lymphoma (n =21) Unclassified (n = 6) Age (years) < 60 (n = 82) > 60 (n = 50) Stage I-II(n = 51) III-IV(n =81) Extra nodal sites
0-1 (n = 89) >2(n=43) LDH level Normal (n = 45) Elevated ( > 8.0 ukat/1; n = 11) Not evaluated (n = 10) CR (%) 92 (70) 76(72) 11(52) 5(83) 64 (79)' 28 (56) 43(84)' 49(60) 70(79)* 22(51) 36 (80)' 48 (62) 8(80) PR (%) 18(14) 16(15) 2(10) -7(8) 11 (22) 4(8) 14(17) 10(11) 8(19) 4(9) 13(17) 1(10) SD (%) 14(10) 10(10) 4(19) -7(8) 7(14) 3(6) 11(14) 7(8) 7(16) 4(9) 9(12) 1(10) PD (%) 8(6) 3(3) 4(19) 1(17) 4(5) 4(8) 1(2) 7(9) 2(2) 6(14) 1(2) 7(9)
-Predictive five years overall survival (%) 58 45 67* 44 63" 31 76" 39
Abbreviations: CR - complete remission; PR - partial remission; SD - stable disease; PD - progressive disease.
n
P < 0.05.
respond or had progressive disease (Table 2). Fifty-seven
of 92 CR patients (62%) received six to eight cycles of
CHOP-E chemotherapy with no reduction in relative
dose intensity. The CR rate was significantly lower in
patients > 60 years (56% vs. 78% for younger patients;
P < 0.05) and in patients with stage III—IV disease
(60% vs. 84% for stage I—II patients; P < 0.05; Table 2).
Survival
The predicted three- and five-year OS for the entire
CHOP-E group were 60% and 53%, respectively
(Figure 1). Patient survival according to the
interna-tional prognostic factor index [25] is shown in Figure 2.
Stage III—IV patients had a significantly worse survival
than patients with limited disease (I—II; P < 0.05;
Table 2). The OS curve of stage I—II patients reached a
plateau phase of about 65% at four years, in contrast
to the slowly declining survival curve of patients with
advanced disease. Performance status was excluded in
this analysis (see under 'patients and methods'). Patients
with elevated LDH (>8.0 ukat/1) had significantly
shorter survivals than the remainder (P < 0.05) and
pa-tients >60 years tended to have worse survivals than
younger patients (P - 0.06). The predicted three- and
five-year DFS for CR patients were 65% and 56%,
respectively. DFS at five years were almost identical for
patients < 60 years (56%, n = 64), patients ^ 60- < 70
years (63%, n = 19), and patients ^ 70 years (64%, n - 9).
Toxicity
The toxic effects of chemotherapy are shown in Table 3.
In general, neutropenia with infection was the major
side effect of this regimen while other toxicities were
modest. There were four therapy-related deaths (3%),
> m
Time (years)
Figure I Overall survival of all patients calculated according to the ICaplan-Meier method [23].
Time (years)
Figure 2. Comparison of overall survival between patients with 0-1 risk factors (n = 46), two risk factors (n = 28), three risk factors (n = 38) and four risk factors (n = 10), according to Shipp et al. [25].
Table 3. CHOP-E treatment toxicity (% of patients). Diarrhoea Stomatitis Infection Neurotoxicity Alkaline Phosphatase sALAT WBC count Haemoglobin Platelet count n 129 129 130 128 131 132 128 127 126 WHO grade 0 (%) 76 78 39 61 nd nd 4 24 59 WHO grade 1-2 (%) 20 20 28 37 nd nd 9 57 24 WHO grade 3-4 (%) 4 2 33 2 1 1 87 19 17
three due to septicemia and one due to haemorrhage.
Eighty-seven percent of patients developed leukopenia
WHO grade 3-4 and one-third of the patients received
antibiotic treatment for severe infection or neutropenic
fever of unknown origin. Other severe toxicities, i.e.
WHO grade 3-4, were infrequent; stomatitis (2%),
diar-rhoea (4%) neurotoxicity (2%) and liver toxicity (1%).
No patient with a chemotherapy-related leukemia or
other hematological malignancy was recorded during
the study period. The majority (>80%) of
chemother-apy courses were administered on an outpatient basis.
Discussion
New chemotherapeutic regimens for the treatment of
aggressive NHL have frequently been introduced in
recent years. The majority of them have included at least
six different cytostatic drugs and the dose schedule of
etoposide has been employed with wide variations. In
view of the lack of cross-resistance between vincristine
and etoposide [26, 27] and the efficacy of etoposide as a
single drug in NHL [12, 13], the addition of etoposide to
the conventional CHOP regimen seemed to be a
poten-tial way to improve efficacy while maintaining toxicity at
a tolerable level. Such attempts have been made in a
restricted number of studies with variants of the
CHOP-E regimen used in this study, reporting CR rates of
73%-87% in patients with aggressive NHL [14, 16, 18].
However, there are no available results from a strict
comparison between CHOP and CHOP-E in this patient
population.
The standard CHOP regimen with the addition of i.v.
(day 1) and oral (day 2-3) etoposide used in the present
study yielded a CR rate of 70% and predicted three- and
five-year OS of 60% and 53%, respectively. If stage I
patients (mainly with extranodal and/or bulky disease
and/or advanced age) were excluded the corresponding
survival figures were 57% and 49%. These survival rates
may be compared with those of a similar study
per-formed in the Nordic countries during the same time
period in which patients with aggressive NHL were
randomised to receive CHOP or MACOP-B. In the
CHOP arm of this study including only stage II-IV
patients (n - 166; median age 51 years) the predicted
three- and five-year OS were 60% and 56%, respectively
(M. Jerkeman, personal communication).
Although this was not a randomized study it seems
unlikely that CHOP-E chemotherapy as used here is
superior to CHOP with regard to complete remission
rate, disease-free or overall survival. One possible
ex-planation for this suggested lack of improvement could
be the fact that 42% of courses given to patients
achiev-ing CR were reduced by 25% or more and 14% of the
patients received fewer than the initially planned six
courses. The main reason for these dose and schedule
modifications was myelosuppression (87% of courses
WHO grade 3—4) with 33% of patients developing an
infectious complication requiring hospital care. These
results are in contrast to those recently presented by
Bergman et al. [18] using a CHOP modification plus
etoposide in aggressive NHL. In their patient population
with the same median age (54 years) as the one in the
present study only 7% of the chemotherapy courses were
associated with severe infections despite the fact that
both the doxorubicin and etoposide doses were
consid-erably higher. Forty-five percent of their patients received
hemopoietic growth factor support (only GM-CSF)
compared to 29% of the patients in the present study
(predominantly G-CSF) which could only partly explain
the drastic difference in severe infections/septicemias.
The distribution of risk factors [25] and their impact on
treatment outcome were similar in the two series.
There were three deaths due to septicemia in our
patient population which compares well with the results
of other studies using CHOP in high-grade NHL [7, 10].
The addition of G- or GM-CSF to patients receiving
chemotherapy for malignant NHL generally reduces the
infection rate and duration of neutropenia. However, the
use of growth factors does not produce any significant
improvement with regard to dose intensity, CR rate or
survival [28-30]. It is well established that patients > 60
years with high-grade malignant NHL present a high
mortality rate from different causes, one of them being a
greater risk of increased organ toxicity. Interestingly, the
median age of the 43 patients (33%) who developed
severe infectious complications requiring i.v. antibiotics
was 55 years, i.e., not different from that of the series as
a whole. However, there was a clear tendency to more
frequent dose reductions (and reduced survival) in the
age group above 65 years. The challenge to optimize
treatment for the elderly patient with aggressive NHL
remains and one important step would be the
identifica-tion of the relatively few patients in this age group who
can tolerate the toxicity of curative combination
chemo-therapy. It is hoped that the results of a large, recently
closed, Nordic trial randomising patients (> 60 years) to
CHOP or CNOP (N=mitoxantrone) chemotherapy with
or without G-CSF support will add to our present
knowledge in this respect.
In conclusion, CHOP-E, with the addition of
etopo-side for three days to conventional CHOP, does not
seem to be more efficacious than CHOP alone.
Myelo-suppression was the most limiting toxicity. Although
eagerly awaited, the results of a controlled randomised
trial comparing CHOP with CHOP plus etoposide are
not yet available. However, soon to be closed is one such
large German study in which the impact of G-CSF and
shorter time interval between courses will also be
eval-uated (M. Pfreundschuh, personal communication).
Acknowledgements
The trial was supported by the Swedish Cancer Society
and the Karolinska Institute foundations. The Nordic
Lymphoma Study Group is gratefully acknowledged for
supplying information on CHOP treated patients.
* The following centres participated in the study
Department of Hematology and Infectious Diseases, Karolinska Hos-pital, Stockholm: M. Bjorkholm, F. Celsing, G. Grimfors, S. Widell, E. Osby; Department of Oncology, South Hospital, Stockholm: K. Merk, O. Tullgren; Section of Hematology, Department of Medi-cine, Danderyd Hospital, Danderyd: P. Bernell, R. Hast, A. Hedlund, E. Kimby, L. Stenke; Department of General Oncology, Karolinska Hospital, Stockholm: J. Liliemark, C. Lindemalm, E. Svedmyr, C. Wedelin, A. Osterborg.
References
1. McKelvey EM, Gottlieb JA, Wilson HE et al. Hydroxyldaunomy-cin (adriamyHydroxyldaunomy-cin) combination chemotheraphy in malignant lym-phoma. Cancer 1976; 38: 1484-93.
2. Shipp MA, Young B, Harrington DP et al. The m-BACOD combination chemotherapy regimen in large-cell lymphoma: Analysis of the completed trial and comparison with the m-BACOD regimen. J Clin Oncol 1990; 8. 84-93.
3. Boyd DB, Coleman M, Papish SW et al. COPBLAM III infu-sional combination chemotherapy for diffuse large-cell lympho-ma. J Clin Oncol 1988; 6: 425-33.
4. Longo D, DeVita VT, Duffey P et al. Superiority of ProMACE-CytaBOM over ProMACE-MOPP in the treatment of advanced diffuse aggressive lymphoma: Results of a prospective random-ized trial. J Clin Oncol 1991; 9: 25-38.
5. Connors SM, Klimo P. MACOP-B chemotherapy for malignant lymphomas and related conditions: 1987 update and additional observations. Semin Hematol 1988; 25: 41-6.
6. Sweetenham JW, Mead GM, Whitehouse JMA. Intensive weekly combination chemotherapy for patients with intermediate-grade and high-grade non-Hodgkin's lymphoma. J Clin Oncol 1991; 9: 2202-9.
7. Fisher RI, Gayner ER, Dahlberg M et al. Comparison of CHOP vs. m-BACOD vs. ProMACE-CytaBOM vs. MACOP-B in pa-tients with intermediate or high-grade non-Hodgkin's lympho-mas. N Engl J Med 1993; 328: 1002-6.
8. Takagi T, Sampi K, Sawada U et al. A comparative study of CHOP versus MEVP (mitoxantrone, etoposide, vindesine, pre-dnisolone) therapy for intermediate-grade and high-grade non-Hodgkin's lymphoma: A prospective randomized study. Int J Hematol 1993; 57: 67-71.
9. Cooper IA. Wolf MM, Fox RM et al for the New Zealand Lymphoma Group. Randomized comparison of MACOP-B with CHOP in patients with intermediate-grade non-Hodgkin's lym-phoma. J Clin Oncol 1994; 12: 769-78.
10. Linch DC.Vaughan Hudson B, Hancock BWet al. A randomised comparison of a third-generation regimen (PACEBOM) with a standard regimen (CHOP) in patients with histologically aggres-sive non-Hodgkin's lymphoma: A British National Lymphoma Investigation report. Br J Cancer 1996; 74: 318-22.
11. Fiorentino MV. Lymphoma in the elderly. Leukemia 1991; 5: 79-85. 12. Jacobs P, King HS, Dent DM, Hayes M.VP-16-213 in the treat-ment of stage III and IV diffuse large-cell lymphoma. Cancer 1985; 56: 1006-13.
13. Hainsworth JD. Chronic administration of etoposide in the treat-ment of non-Hodgkin's lymphoma. Leuk Lymph 1993; 10 (Suppl)1 65-72.
14. Hansen MM, Blomfield CD, Jorgensen J et al. VP-16-213 in combination with cyclophosphamide, doxorubicin, vincristine and prednisone in the treatment of non-Hodgkin's lymphomas. Cancer Treat Rep 1980; 64: 1135-7.
15. Wilson WH, Bryant G, Bates et al. EPOCH chemotherapy: Toxicity and efficacy in relapsed and refractory non-Hodgkin's lymphoma. J Clin Oncol 1993; 11: 1573-82.
16. Koppler H, Pfluger KH, Eschenbach I et al. Randomised com-parison of CHOEP versus alternating hCHOP/IVEP for high-gTade non-Hodgkin's lymphomas: Treatment results and prog-nostic factor analysis in a multi-centre trial. Ann Oncol 1994; 5: 49-55.
17. Carrion JR, Garcia Arroyo FR, Salinas P. Infusional chemo-therapy (EPOCH) in patients with refractory or relapsed lympho-ma. Am J Clin Oncol 1995; 18: 44-6.
18. Bergmann L, Karakas T, Lautenschlager G et al. Vincristine, doxorubicin, cyclophosphamide, prednisone and etoposide (VACPE) in high-grade non-Hodgkin's lymphoma-a multicenter phase II study. Ann Oncol 1995, 6: 1019-24.
19. Merk K. An epidemiological approach to the etiology of Hodg-kin's disease. Academic thesis, Stockholm 1992.
20. Stansfeld A, Diebold J, Kapanci Yet al. Updated Kiel Classifica-tion for lymphoma. Lancet 1988; i: 292-3.
21 Carbone PP, Kaplan HS, Musshoff K et al. Report of the com-mittee on Hodgkin's disease staging classification. Cancer Res
1971; 31: 1860-1.
22. Cabanillas F, Hagemeister FB, McLaughlin P et al. Results of MIME salvage regimen for recurrent or refractory lymphoma. J Clin Oncol 1987; 5: 407-12.
23. Kaplan EM, Meier P. Nonparametnc estimation from incom-plete observations. J Am Stat Assoc 1958, 53: 457-81.
24. Peto R, Pike MC, Armitage P et al. Design and analysis of randomised clinical trials requiring prolonged observations of each patient. Br J Cancer 1971; 38: 1484-93.
25. Shipp MA. Prognostic factors in aggressive non-Hodgkin's lym-phoma who has high-risk disease. Blood 1994; 83: 1165-73. 26. O'Reilly SE, Klimo P, Connors JM. The evolving role of
etopo-side in the management of lymphomas and Hodgkin's disease. Cancer 1991; 67: 271-80.
27. Australian and New Zealand lymphoma group. Investigation of the addition potential of tenoposide and vincristine in non-Hodg-kin lymphomas. Cancer Treat Rep 1986; 70: 985-90.
28. Zinzani PL, Pavone E, Storti S et al. Randomized trial with or without granulocyte colony-stimulating factor as adjunct to in-duction VNCOP-B treatment of elderly high-grade non-Hodg-kin's lymphoma. Blood 1997; 89: 3974-9.
29. Pettengell R, Gurney H, Radford JA et al. Granulocyte colony-stimulating factor to prevent dose-limiting neutropenia in non-Hodgkin's lymphoma: A randomized controlled trial. Blood 1992; 80. 1430-6.
30. Gerhartz HH, Engelhard M, Meusers P et al. Randomized, double-blind, placebo-controlled, phase III study of recombinant human granulocyte-macrophage colony-stimulating factor as ad-junct to induction treatment of high-grade malignant
non-Hodg-kin's lymphomas. Blood 1993; 82: 2329-39. Received 26 June 1998; accepted 16 September 1998.
Correspondence to:
M. Bjdrkholm, MD
Department of Hematology and Infectious Diseases Karolinska Hospital
SE-17176 Stockholm Sweden