American Journal of Hematology 77:358362 (2004)

Prognostic Variables in Newly Diagnosed

Childhood Immune Thrombocytopenia
Shahid Ahmed,1* Anita K. Siddiqui,2 Rabia K. Shahid,3 Miriam Kimpo,4
Cristina P. Sison,5 and Mark A. Hoffman6
1
Saskatoon Cancer Center, University of Saskatchewan, Saskatoon, SK, Canada
2
Dartmouth General Hospital, Dalhousie University, Dartmouth, NS, Canada
3
Dow University of Health Sciences, Karachi, Pakistan
4
Schneider Children Hospital, New Hyde Park, New York
5
North Shore-Long Island Jewish Research Institute, Biostatistics Unit, Manhasset, New York
6
Long Island Jewish Medical Center, New Hyde Park, New York

Immune thrombocytopenia (ITP) has a favorable prognosis in children. Only a small

number of children go on to develop chronic ITP. However, at the time of diagnosis, it is
not possible to predict the course of the disease. In order to determine prognostic factors
that could predict the disease course at diagnosis, we retrospectively evaluated various
clinical variables in 103 pediatric patients with newly diagnosed ITP at our institution from
1995 to 2001. Sixty-eight (66%) patients had a mean platelet volume (MPV) of <8 fL on
admission. Of 72 patients who had a follow-up period of at least 6 months, 54 (75%)
achieved a durable remission within 6 months and 18 (25%) developed chronic ITP. In
univariate analysis, a low admission MPV (<8), history of viral prodrome, and a low
admission platelet count (<10 109/L) predicted for a favorable outcome. Age and sex
did not correlate with remission. In multivariate analysis, a low admission MPV and a
history of a viral prodrome were the only independent factors correlated with a durable
CR. The adjusted odds ratio for achieving a durable remission was 8.9 (95% CI: 1.5451.8)
for history of a viral prodrome and 14 (95% CI: 2.5283.3) for low admission MPV value. In
conclusion, our study showed that a majority of the children with newly diagnosed ITP
presented with a low MPV value. A history of viral illness and a low admission MPV were
found to be independent prognostic variables that predicted for the achievement of a
durable CR in childhood ITP. Am. J. Hematol. 77:358362, 2004. 2004 Wiley-Liss, Inc.

Key words: ITP; MPV; viral prodrome; prognostic variables

INTRODUCTION ITP has a strikingly different clinical course in adults

Immune thrombocytopenic purpura (ITP) is one of
the common causes of thrombocytopenia in adults and
children. It affects approximately 100 people per 1 mil-
lion subjects per year in the general population, and
about half of these cases occur in children [14]. ITP is
caused by autoantibodies that bind to platelets and result
in an increased destruction of these antibody-coated
platelets by phagocytic cells in the reticuloendothelial
system. ITP exists in two major clinical forms, an acute
self-limiting illness, and a chronic disorder characterized
by persistence of thrombocytopenia for longer than 6
months after initial presentation. Additionally, some
chronic cases show an intermittent pattern of thrombo-
cytopenia (so-called relapsing ITP).
2004 Wiley-Liss, Inc.
and children. Spontaneous remission in adults with
ITP is less frequent and majority develops chronic
ITP [1]. Children, on the other hand, have a much
Presented in part at the 43rd and 44th Annual Meetings of the
American Society of Hematology in 2001 and 2002, respectively.
*Correspondence to: Shahid Ahmed, M.D., Saskatoon Cancer
Center, University of Saskatchewan, 20 Campus Drive, University
of Saskatchewan Campus, Saskatoon, SK, Canada, S7N4H4.
E-mail: sahmed@scf.sk.ca, shahidahmed00@yahoo.com
Received for publication 4 March 2004; Accepted 27 June 2004
Published online in Wiley InterScience (www.interscience.wiley.com).
DOI: 10.1002/ajh.20205

more favorable prognosis. In more than 75%, the dis-
ease resolves within 6 months irrespective of treatment
[3]. In children, however, it is not possible to predict the
course of the disease at the time of initial diagnosis. An
insidious presentation, female sex, and age over 10
years have been reported to be prognostic factors that
predict the development of chronic ITP in children
[57]. The importance of other clinical variables such
as admission platelet count or platelet size as prognos-
tic markers in ITP in children is not known.
Mean platelet volume (MPV) is a measure of platelet
size [8]. During our investigation to validate known
variables and other clinical features as reliable predic-
tors for the disease course, we discovered a positive role
of low mean platelet volume (MPV) value at diagnosis
in childhood ITP [9]. In this study, we evaluated
whether MPV or other variables such as history of
viral illness, age, and sex can independently predict an
early complete remission in childhood acute ITP in a
larger number of children with ITP.
METHODS
The study protocol was approved by Long Island
Jewish Medical Centers Institutional Review Board.
Medical records of all pediatric patients (age below 18
years) admitted at Long Island Jewish Medical Center
with newly diagnosed immune thrombocytopenia from
January 1995 to December 2001 were reviewed. The
demographic and clinical data were obtained. All those
patients who did not have prior history of thrombo-
cytopenia were considered newly diagnosed ITP cases.
Patients who had thrombocytopenia due to other
causes, such as medication, malignancy, bone marrow
failure, and sepsis, were excluded from the analysis.
Additionally, all those patients who had ITP and
underlying autoimmune disorders, lymphoproliferative
disease, or HIV infection were excluded from the ana-
lysis. Platelet counts and MPV values were determined
by a Genesis automated cell counter (Coulter Electro-
nics, Hialeah, FL) (coefficient of variation
2.2% at
MPV values of 810 fL). In the instrument at our
institution, normal MPV values ranged from 811 fL.
Medical records of all those patients who had a follow-
up of at least 6 months or greater at our institution
were reviewed to determine the disease course. On the
basis of follow-up data, all patients were divided into
two groups: group 1, patients who achieved a durable
remission, and group 2, patients who developed
chronic ITP. Our definitions are as follows: (1)
Chronic ITP was defined as a persistent low platelets
count (<150  109/L) for more than 6 months. Those
patients who had 3 or more relapses during the
6-month follow-up period were considered relapsing
ITP and were included in the group of chronic ITP.
Prognostic Variables in Acute ITP 359
(2) A viral prodrome was defined as a documented
history of upper respiratory tract illness (URI), gastro-
intestinal (GI) symptoms, or skin exanthema within 4
weeks of diagnosis of ITP. (3) A remission was defined
as improvement in platelet count to normal values
(150  109/L) and maintenance of normal platelet
count without any therapy. All patients who did not
develop chronic ITP were considered to achieve a
durable remission. (4) A response to treatment was
defined as an increment in platelet count above 20 
109/L without platelet transfusion, if admission platelet
counts were <10  109/L or a 50% increase in platelet
count if admission platelets were 10  109/L and that
level was maintained for at least the next 24 hr.
Statistical Analysis
Data are reported as median (range). The Mann
Whitney test was carried out to compare continuous
variables between the two groups. Fishers exact test
was used to determine which of the categorical vari-
ables were associated with a complete remission rate.
The following variables were examined to determine if
any of them could predict the disease outcome: age,
male or female sex, presence or absence of a viral
prodrome, admission platelet count, admission MPV
values, treatment with single agents versus two or more
agents, time to response, treatment with steroids upon
discharge. Continuous variables were dichotomized by
specifying arbitrary cut-off points as follows: age (<10
or 10 years), admission platelet count (<10 or 10 
109/L), admission MPV values (<8 or 8), time to
response (<2 or 2 days). Those found to be signifi-
cant were included in a stepwise logistic regression to
determine which variables independently discriminated
between a self-limited course and chronic ITP. All two
tailed P values <0.05 were considered significant.
RESULTS
There were a total of 103 pediatric patients who
met the criteria of acute primary ITP during the study
period. Their clinical and laboratory data are
described in Tables I and II. Of 103 patients, 82
(77%) had bone marrow examination. All bone mar-
row biopsy/aspiration specimens were consistent with
ITP. Megakaryocytes were adequate in number. Nor-
mal maturation was observed in all three cell lineages,
and no abnormal infiltration was seen. Median age of
patients was 6 year (range 0.1017.6) and their med-
ian admission platelet count was 7  109/L (range
[169]  109/L). Sixty-eight (66%) patients on admis-
sion had MPV value <8. Relationships of admission
platelet count with patient age and admission MPV
360 Ahmed et al.

TABLE I. Characteristics of Patients With Acute ITP*

Patient characteristics All patients (N 103) No FU (N 31) Group 1a (N 54) Group 2b (N 18) P valuec

Age (years) 6 (0.1017.6) 6 (0.317.6) 5 (0.116) 7.5 (1.317) NS

Age < 10 years 72 (70%) 22 (71%) 39 (72%) 11 (61%) NS
Male:female 52:51 24:7 23:31 5:13 0.4
Viral prodrome 51/68 (75%) 17/19 (89%) 29/36 (81%) 5/13 (38%) 0.01
Bone marrow biopsy 86 (83%) 26 (83%) 42 (78%) 18 (100%) NS
Intracranial hemorrhage 0 0 0 0 NS
Treated with single agent 82 (80%) 27 (87%) 40 (74%) 15 (83%) NS
Time to response 2 days (115) 2 days (15) 2 days (115) 2 days (14) NS
Duration of stay 3 days (217) 3 days (27) 3 days (217) 3 days (27) NS
Discharge on steroids 43 (42%) 14 (45%) 21 (39%) 8 (44%) NS

*Values are given in median (range).

a
Patients who achieved a durable CR.
b
Patients who developed chronic ITP.
c
Comparison of variables between group 1 and group 2.

TABLE II. Laboratory Features on Admission of Patients With Acute ITP*

Laboratory features All patients (N 103) No FU (N 31) Group 1a (N 54) Group 2b (N 18) P valuec

Admission platelet count (109/L) 7 (169) 9 (260) 5 (169) 11.5 (244) 0.02
Admission platelet < 10 63 (61.2%) 16 (52%) 40 (74%) 7 (39%) 0.01
Admission MPV 7 (3.912.1) 7.1 (4.912.1) 6.1 (3.911.3) 8.4 (5.111.1) 0.0002
Admission MPV < 8 68 (66%) 20 (64%) 44 (82%) 4 (22%) <0.0001
Admission hemoglobin (g/dL) 12.2 (4.915.6) 12.3 (4.914.4) 11.95 (8.115.6) 11.8 (5.113.6) NS
Discharge platelet count 54 (16421) 51 (26385) 55 (16421) 60.5 (23148) NS
Discharge MPV 9.4 (6.215.5) 9.8 (6.214.9) 9.2 (6.915.5) 9.75 (7.112) NS
Discharge hemoglobin 11.3 (8.414.5) 11.6 (8.413.5) 10.95 (9.114.5) 11.4 (9.712.5) NS

*Values are given in median (range).

a
Patients who achieved a durable CR.
b
Patients who developed chronic ITP.
c
Comparison of variables between group 1 and group 2.

are shown in Figs. 1 and 2. A history of viral illness

was obtained in 68 patients (66%), and 51 (75%) of
them had documented viral illness within 4 weeks
of admission. Of 51 patients with a positive history
of viral prodrome within 4 weeks of diagnosis of ITP,
48 (94%) patients had history of URI and 3 (6%)
had GI symptoms. Five patients received antibiotic
treatment within 4 weeks of the diagnosis of ITP.
All 103 patients received treatment for ITP during
their hospital stay, and therapy was heterogeneous.
Eighty-two patients were treated with single agents
(mostly intravenous gamma globulin [IVIgG]), 19
received two agents (mostly IVIgG and prednisone),
and 2 had three agents (vincristine or intravenous
anti-D with IVIgG and prednisone). Glucocorticoids,
oral or intravenous, were given to 44 patients, 4
patients received intravenous anti-D (WinRho), and
1 received vincristine. None of the patients underwent
splenectomy at diagnosis or during the follow-up
period. Four patients had a hemoglobin level of
9 g/dL or below, and iron deficiency was the cause
of anemia in these patients.
Follow-Up
There were 72 patients who had a follow-up period
of at least 6 months. The median duration of follow-
up was 9 months (range 684 months). Of those 72
patients, 60 (83.3%) had bone marrow biopsy and/or
aspiration. All biopsy/aspiration specimens were con-
sistent with ITP. No significant difference was found
in the initial treatment of the two groups.
Eighteen patients developed chronic ITP during the
follow-up period. Of 18 patients who developed chronic
ITP, 1 patient had relapsing ITP and experienced three
relapses within 6 months whereas the remaining 17
patients had platelet counts <100  109/L for more than
6 months. On univariate analysis admission platelet
count, admission MPV, and history of viral prodrome
were associated with the disease outcome. Forty (85%)
out of 47 patients with an admission platelet count <10 
109/L went into remission compared to 14 (56%) out of
25 patients with admission platelet count 10  109/L
or above (P 0.01). Forty-four (92%) of 48 patients
with an admission MPV value <8 achieved a durable

Fig. 1. Scatter plot comparing admission platelet counts
with patient age in the two groups and in patients who did
not have more than 6 months of follow-up.
CR compared to 10 (42%) of 24 patients with an admis-
sion MPV of 8 or above (P < 0.0001). Twenty-nine
(86%) out of 34 patients with a history of viral illness
went into early remission compared to 7 (47%) out
of 15 patients who did not have history of a viral
prodrome (P 0.01). Age and sex did not correlate
with remission. Stepwise logistic regression analysis,
which initially included all three significant variables
in the model, resulted in a final predictive model with
a history of viral prodrome and admission MPV value
being independently significant. The adjusted odds
ratio for achieving a durable remission was 8.9 (95%
CI: 1.5451.8) for history of a viral prodrome and
14 (95% CI: 2.5283.3) for admission MPV value.
DISCUSSION
Our results show that a large proportion of chil-
dren with newly diagnosed ITP present with a low
MPV value. The low initial MPV in acute ITP value
gradually increased to the normal range with the
recovery of platelet count (Table II). Investigators
have earlier found an inverse nonlinear relationship
between the MPV and platelet count in normal sub-
jects [10,11]. Threatte, in a review of available data
about the clinical usefulness of MPV, has suggested
that MPV correlates with the platelet turnover and
can reflect changes in the rate of platelet production
[8]. MPV increases in disorders associated with
Prognostic Variables in Acute ITP 361
Fig. 2. Scatter plot comparing admission platelet counts
with MPV in the two groups and in patients who did not
have more than 6 months of follow-up.
accelerated platelet turnover, whereas a deficient plate-
let production usually results in a normal or low MPV.
Although a normal volumenumber relationship is
retained in ITP, the volumenumber relationship,
however, varies in the acute, chronic, and recovering
phases of ITP [12,13]. Patients with chronic ITP
usually have a high MPV; however, a low MPV has
been found in some subjects with acute ITP [8,13].
We also found that the low initial MPV value was
among the favorable prognostic factors that predict a
durable CR in childhood acute ITP. Overall, 75% of
patients in our study population achieved CR soon after
their acute episode and only 25% had a chronic disease
course. Ninety-two percent of children with an admission
MPV < 8 achieved an early complete remission com-
pared to 42% of patients with an MPV of 8 or above.
An infectious prodrome within 4 weeks of diagno-
sis of ITP was another important favorable prognos-
tic variable in our study. ITP in children may follow
an infectious prodrome or immunization. Nonspecific
viral illnesses are the most common infection preced-
ing the onset of ITP in children [2,4]. Of note, 75% of
children in our study groups had a history of an
infectious prodrome, and the majority of them had
an upper respiratory tract illness. This high incidence
was in agreement with previous observations where as
high as 84% of cases experienced such prodrome [14].
362 Ahmed et al.
Serologic studies identified many viruses associated
with ITP in children, such as HIV, rubella, rubeola,
varicella, mumps, Epstein-Barr virus, and cytomega-
lovirus among others [15]. Although the incidence of
an infectious prodrome was quite similar in the
patients with low MPV versus MPV of 8 or above,
a strong correlation was found between the history of
viral prodrome and syndrome and a complete and
durable remission rate in childhood ITP. Although
the International Childhood ITP Study Group has
noted heterogeneity in the disease outcome with
respect to age [16,17], we did not find an association
of disease outcome with age and sex.
Increased peripheral destruction of platelet remains
the main explanation of thrombocytopenia in ITP.
Although bone marrow examination in our patients
including those with low MPV was consistent with
ITP, normal or increased numbers of megakaryocytes
do not necessarily mean that platelet production is
adequate. For example, two groups of investigators
using radiolabeled platelets suggested that, in addition
to increase destruction, decreased platelet counts in
some patients with ITP (including patients younger
than 18 years) could also be on the basis of decrease
production of platelet [18,19]. Low MPV is not among
the indicators of high platelet turnover. Moreover, in
our study, 75% of children had a viral prodrome. It is
conceivable, therefore, that cytokines released during
the viral prodrome resulted in inhibiting or decreasing
platelet production in some of the children.
It is important to be aware of some limitations of our
study. There are several variables that can affect MPV
value. For example, an increase in MPV values has
been noted with time in EDTA-anticoagulated venous
blood, the most commonly used specimen for peri-
pheral blood count analysis [20,21]. Likewise, tempe-
rature, osmolarity, and pH may also affect MPV [8].
Although automated cell counters are fairly accurate in
determining platelet count and MPV, the possibility of
instrumental artifacts at low platelet counts cannot be
fully excluded. Additionally, only hospitalized children
were included in the study, and follow-up data were not
available for 30% of patients.
In summary, a majority of children with acute ITP
present with a low MPV value and commonly have an
infectious prodrome. A low admission MPV value and
history of a viral prodrome correlated with a complete
and durable response rate. Future large prospective
studies are required to confirm our findings.
ACKNOWLEDGMENT
We gratefully acknowledge the helpful suggestions
of Dr. Kanti R. Rai and his critical review of the
manuscript.
REFERENCES
1. Cines DB, Blanchette VS. Immune thrombocytopenic purpura.
N Engl J Med 2002;346:9951008.
2. Bolton-Maggs PHB. Idiopathic thrombocytopenic purpura. Arch
Dis Child 2000;83:220222.
3. George JN, Woolf SH, Raskob GE, et al. Idiopathic thrombo-
cytopenic purpura: a practice guideline developed by explicit
methods for the American Society of Hematology. Blood 1996;
88:340.
4. Souid AK, Sadowitz PD. Acute childhood immune thrombocyto-
penic purpura: diagnosis and treatment. Clin Pediatr 1995;34:
487494.
5. Ramos MEG, Newman AJ, Gross S. Chronic thrombocytopenia
in childhood. J Pediatr 1978;92:584586.
6. Robb LG, Tiedeman K. Idiopathic thrombocytopenic purpura:
predictors of chronic disease. Arch Dis Child 1990;65:502506.
7. Blanchette V, Carcao M. Approach to the investigation and man-
agement of immune thrombocytopenic purpura in children. Semin
Hematol 2000;37:29913006.
8. Threatte GA. Usefulness of the mean platelet volume. Clin Lab
Med 1993;13:937950.
9. Ahmed S, Karim A, Hoffman M. The prognostic value of mean
platelet volume (MPV) in newly diagnosed immune thrombocyto-
penic purpura (ITP). Blood 2001;98:521a.
10. Bessman JD, Williams LJ, Glimer PR. Mean platelet volume: the
inverse relation of platelet size and count in normal subjects. Am J
Clin Pathol 1981;76:289293.
11. Graham SS, Traub B, Mink IB. Automated platelet-sizing
parameters on a normal population. Am J Clin Pathol 1987;87:
365.
12. Levin J, Bessman JD. The inverse relation between platelet volume
and platelet number. Abnormalities in hematologic disease and
evidence that platelet size does not correlate with platelet age.
J Lab Clin Med 1983;101:295307.
13. Baynes RD, Lamparelli RD, Bezwoda WR, et al. Platelet param-
eter. Part II. Platelet volume-number relationships in various nor-
mal and disease states. S Afr Med J 1988;73:3943.
14. Levine SP. Thrombocytopenia caused by immunologic platelet
destruction. In: Lee GR, Foerster J, Lukens J, Paraskevas F,
Greer JP, Rodgers GM, editors. Wintrobes clinical hematology.
Baltimore, MD: Williams & Wilkins; 1999. p 15831611.
15. Yenicesu I, Yetgin S, Ozyurek E, Aslan D. Virus-associated
immune thrombocytopenic purpura in childhood. Pediatr Hema-
tol Oncol 2002;19:433437.
16. Kuhne T, Imbach P, Bolton-Maggs PH, et al. Newly diagnosed
idiopathic thrombocytopenic purpura in childhood: an observa-
tional study. Lancet 2001;358:21222125.
17. Kuhne T, Buchanan GR, Zimmerman S, et al. A prospective
comparative study of 2540 infants and children with newly
diagnosed idiopathic thrombocytopenic purpura (ITP) from
International Childhood ITP Study Group. J Pediatr 2003;143:
605608.
18. Heynes Adu P, Badenhorst PN, Lotter MG, et al. Platelet turn-
over and kinetics in immune thrombocytopenic purpura: results
with autologous 111In-labeled platelets and homologous
51
Cr-labeled platelets differ. Blood 1986;67:8692.
19. Louwes H, Lathori OAZ, Vellenga E, de Wolf JTM. Platelet
kinetic studies in patients with idiopathic thrombocytopenic pur-
pura. Am J Med 1999;106:430434.
20. Thompson CB, Diaz DD, Quinn PG, et al. The role of anti-
coagulant in the measurement of platelet volumes. Am J Clin
Pathol 1983;80:327.
21. Threatte GA, Adrados C, Ebbe S, et al. Mean platelet volume:
the need for a reference method. Am J Clin Pathol 1984;81:
769772.