Letters to the Editor
Confirmation of the utility of the International
Table 1. Patients' characteristics (n=1112).
Staging System and identification of a unique
pattern of disease in Brazilian patients with
Characteristic
Information available (N)
N (%)
multiple myeloma
Age, median (range)
1112
60.5 years (28 94)
Multiple myeloma (MM) is one of the most fre-
Sex
1112
Female
553 (49.7)
quent hematologic malignancies, and its incidence
Male
559 (50.3)
varies worldwide. Except for occasional case series or
Ethnicity
1112
correlative biological studies, little is known about the
Caucasian/mixed
926 (83.3)
incidence and clinical features of MM in Latin
African
176 (15.8)
America. In Brazil, national estimates for the inci-
Other
10 (0.9)
Creatinine
1004
dence of MM are currently unavailable. Sixteen
> 2 mg/dL
231 (23)
Brazilian institutions provided information on
2 mg/dL
773 (77)
patients diagnosed with MM between 1998 and
Immunoglobulin (Ig) isotype
926
2004. The investigators included all patients whose
IgG
570 (61.5)
IgA
181 (19.5)
charts were available. All patients were undergoing
IgM
3 (0.3)
care at these institutions. The diagnosis was based on
Light-chain only
112 (12.1)
the new criteria,1 and/or standard clinical, laboratory
Non-secretory
40 (4.3)
and radiographical features, as well as on bone mar-
Hemoglobin
1007
<10 g/dL
585 (58)
row findings compatible with MM. Since the majori-
Bone lesions
1043
ty of patients had Durie-Salmon stage (DSS) III, the
Absent
155 (14.9)
diagnosis was not in doubt. Patients' data were
Present
888 (85.1)
obtained from institutional charts, and were entered
Durie-Salmon stage
1066
on a web-based system specifically designed for the
I
68 (6.4)
II
182 (17.1)
study under the auspices of the International
III
816 (76.5)
Myeloma Foundation. We collected information
International Staging System stage
756
regarding the demographical features of the patient,
I
152 (20.1)
date of diagnosis, stage according to the DSS2 and
II
368 (48.7)
III
236 (31.2)
ISS,3 type of monoclonal component, results of perti-
Hypercalcemia
1003
nent laboratory tests, type of treatment administered,
Absent
764 (76.2)
and date of last follow-up or death. From 1998 to
Present
239 (23.8)
2004, the patients who were not candidates for high-
High-dose chemotherapy
1057
dose chemotherapy (i.e., those who presented with
Yes
270 (25.5)
No
787 (74.5)
poor performance status or for whom high-dose
chemotherapy was not available) were treated with
melphalan and prednisone.
survival was 52.1 months. After 5 years, the estimat-
The database was analyzed in August 2006. A total
ed proportions of survival for patients in DSS stages I,
number of 1,112 patients were included. 2 tests were
II and III were 73.3%, 54.7% and 44.3% respectively.
used for the comparisons between proportions.
The survival curves for patients in ISS stages I, II
Survival, defined from the date of diagnosis until
and III are clearly distinct, with no overlap until 100
death, was analyzed by the Kaplan-Meier method4
months of follow-up (Figure 1, Panel B). In addition,
and survival curves were compared using the log-rank
20.1%, 48.7% and 31.2% of patients were in ISS I, II
test.5 Patients who were alive on the last visit or who
and III respectively. Corresponding proportions in the
were lost to follow-up were censored for survival
ISS database were 28.9%, 37.5% and 33.6% respec-
analyses. Significant baseline univariate predictors of
tively.3 The median overall survival for patients in ISS
survival in the Kaplan-Meier analyses were consid-
stage I had not been reached at the time of the analy-
ered for inclusion in a Cox regression model.6
sis; for patients in stages II and III, median overall sur-
Backward selection of all factors was used, and a vari-
vival was 65.5 and 26.0 months respectively.
able was retained in the model if the associated two-
Corresponding figures in the original ISS database
sided p value was 0.05. All statistical tests were two-
were 62, 44 and 29 months respectively. After 5
sided, and final p values 0.05 were considered signif-
years, the estimated proportions of survival for
icant.
patients in ISS stages I, II and III were 68.2%, 52.7%
Main demographical and clinical features of the
and 30.4% respectively.
patients are shown in Table 1. The median follow-up
The differences between the survival of our
was 20.5 months, and the estimated median overall
patients in stages I and II and the corresponding
survival was 57.7 months. At the time of the analysis,
patients from the original ISS database may reflect
392 patients (35.3%) were already deceased. Overall
patient selection, treatment modalities, or be due to
survival was analyzed according to DSS and ISS
chance. Our database contains a lower proportion of
(Figure 1, Panels A and B). There was a statistically
patients in stage I, and a higher proportion in stage II,
significant difference between the overall survival of
but these differences are unlikely to have contributed
patients with MM in stages I, II and III according to
to the longer overall survival observed in the current
both staging systems (p<0.001). With DSS, however,
study. In the ISS database, 26.1% of patients received
there was a considerable overlap between the sur-
high-dose chemotherapy within nine months of diag-
vival curves for patients in stage I and II (Figure 1,
nosis, figures very similar to those found in the pres-
Panel A). The median overall survival for patients in
ent study (25.5%). We, therefore, cannot explain the
DSS I and II had not been reached at the time of the
survival difference observed for patients in this study
analysis; for patients in stage III, the median overall
and
patients
included
in
the
ISS
database.
haematologica | 2008; 93(5) | 791 |
Letters to the Editor
The multivariate prognostic model included
Table 2. Multivariate prognostic model for mortality.
patients' baseline variables that were associated with
mortality in the Kaplan-Meier univariate analyses.
Variable
Category
Hazard ratio 95% confidence
p
Only hypercalcemia, ISS stage III and age were inde-
interval
value
pendent predictors of mortality (Table 2). According
to the model, DSS had no independent prognostic
role when hypercalcemia, ISS and age were consid-
Hypercalcemia
Absent
1.00
Present
2.47
1.863.28
<0.001
ered simultaneously.
ISS stage
I
1.00
The current study, which is the largest case series of
II
1.22
0.791.88
0.364
MM in Brazil, identifies the unique aspect of the pat-
III
2.40
1.553.72
<0.001
tern of disease for Brazilian myeloma patients, with
Age
60 years
1.00
> 60 years
1.54
1.182.01
0.002
hypercalcemia as the main risk. It recognizes the fea-
Durie-Salmon stage
I
1.00
sibility of large, collaborative, observational studies
II
0.97
0.442.12
0.939
between various tertiary-care hematology centers,
III
1.28
0.632.60
0.501
and confirms that ISS is indeed a more useful prog-
nostic index than the DSS system in Brazilian
ISS: International Staging System.
patients.
Vania T.M. Hungria,1 Angelo Maiolino,2 Gracia Martinez,3
1.0
Gisele W.B. Colleoni,4 Érika O.D.M. Coelho,5 Laís Rocha,6
A
DSS
Renata Nunes,7 Rosane Bittencourt,8 Luciana C.O. Oliveira,9
III
Rosa Malena O. Faria,10 Ricardo Pasquini,11 Sílvia M.M.
0.8
III
Magalhães,12 Cármino A. Souza,13 Jorge V. Pinto Neto,14
Luciana Barreto,15 Elizabeth Andrade,16 Maria do Socorro O.
0.6
Portella,17 Vanessa Bolejack,18 Brian G.M. Durie19 on behalf of
the International Myeloma Working Group Latin America
0.4
1Santa Casa de São Paulo, Brazil; 2Universidade Federal do
Rio de Janeiro, Brazil; 3Universidade de São Paulo, Brazil;
0.2
4Universidade Federal de São Paulo, Brazil; 5HEMOPE,
p<0.001
Recife, Brazil; 6Universidade Federal da Bahia, Brazil;
0.0
7
0
20
40
60
80
100
Hospital Brigadeiro de São Paulo, Brazil; 8Universidade
Time (months)
Federal do Rio Grande do Sul, Brazil; 9Medical School of
1.0
Ribeirão Preto, Brazil; 10Universidade Federal de Minas Gerais,
B
ISS
Brazil; 11Universidade Federal do Paraná, Brazil;
I
12
0.8
II
Universidade Federal do Ceará, Brazil; 13Universidade
III
Estadual de Campinas, Brazil; 14Hospital de Base de Brasília,
Brazil; 15Instituto Nacional de Câncer do Rio de Janeiro, Brazil;
0.6
16Fundação de Oncologia de Manaus, Brazil 17Novartis
Oncology, Brazil; 18Cancer Research and Biostatistics, Seattle,
0.4
WA, USA; 19Aptium Oncology, Inc. Cedars-Sinai Outpatient
Cancer Center, Los Angeles, CA, USA
0.2
Correspondence: Vania T. M. Hungria, MD, Rua Tucumã,
p<0.001
113 apto 41, São Paulo, Brazil, CEP, 01455-010.
0.0 0
20
40
60
80
100
Phone: international +55.11. 38419377. Fax: international
+55.11.38468790 E-mail: hungria@dialdata.com.br
Time (months)
Figure 1. Overall survival according to the Durie-Salmon Staging
References
(DSS) system (Panel A, N=1,066) and to the International Staging
system (ISS; Panel B, N=756).
1. Criteria for the classification of monoclonal gam-
mopathies, multiple myeloma and related disorders: a
report of the International Myeloma Working Group. Br J
Haematol 2003;121:749-57.
Interestingly, no patients from Brazil or any other
2. Durie BG, Salmon SE. A clinical staging system for multi-
Latin American countries were included in the origi-
ple myeloma. Correlation of measured myeloma cell mass
nal ISS database.
with presenting clinical features, response to treatment,
When survival was analyzed according to other
and survival. Cancer 1975;36:842-54.
potential prognostic factors, age >60 years (52.1 vs.
3. Greipp PR, San Miguel J, Durie BG, Crowley JJ, Barlogie
65.5 months for those with age 60 years), presence
B, Bladé J, et al. International staging system for multiple
myeloma. J Clin Oncol 2005;23:3412-20.
of hypercalcemia (21.3 vs. 63.9 months for those with
4. Kaplan E, Meier P. Nonparametric estimation from incom-
normal serum calcium) and receipt of high-dose
plete observations. J Am Stat Assoc 1958;53:457-81.
chemotherapy (82.2 vs. 43.5 months for patients not
5. Mantel N. Evaluation of survival data and two new rank
treated with high-dose chemotherapy) were found to
order statistics arising in its consideration. Cancer
be significantly associated with increased mortality
Chemother Rep 1966;50:163-70.
(p<0.001 in all cases). Age and receipt of high-dose
6. Cox DR. Regression models and life tables (with discus-
sion). J R Statist Soc 1972;34 (series B):187-220.
chemotherapy were correlated: 83.0% of patients
undergoing high-dose chemotherapy were aged 60
Citation: Hungria VTM, Maiolino A, Martinez G, Colleoni GWB, Coelho EOD, Rocha L,
years, whereas 62.9% of patients not treated with
Nunes R, Bittencourt R, Oliveira LCO, Faria RMO, Pasquini R, Magalhães SMM, Souza CA,
Pinto Neto JV, Barreto L, Andrade E, Portella MdSO, Bolejack V, Durie BGM. Confirmation
high-dose chemotherapy were aged > 60 years
of the utility of the International staging system and identification of a unique pattern of
(p<0.001).
disease in Brazilian patients with multiple myeloma. Haematologica 2008 May;
93(5):791-792. doi: 10.3324/haematol.11637
| 792 | haematologica | 2008; 93(5)