Articles

Eltrombopag for children with chronic immune

thrombocytopenia (PETIT2): a randomised, multicentre,
placebo-controlled trial
John D Grainger, Franco Locatelli, Thirachit Chotsampancharoen, Elena Donyush, Bunchoo Pongtanakul, Patcharee Komvilaisak, Darintr Sosothikul,
Guillermo Drelichman, Nongnuch Sirachainan, Susanne Holzhauer, Vladimir Lebedev, Richard Lemons, Dagmar Pospisilova, Ugo Ramenghi,
James B Bussel, Kalpana K Bakshi, Malini Iyengar, Geoffrey W Chan, Karen D Chagin, Dickens Theodore, Lisa M Marcello, Christine K Bailey

Summary
Background The thrombopoietin receptor agonist eltrombopag has been shown to be safe, tolerable, and effective for Published Online
adults with chronic immune thrombocytopenia. We aimed to investigate the safety and efficacy of eltrombopag for July 29, 2015
[Link]
children with chronic immune thrombocytopenia. S0140-6736(15)61107-2
See Online/Comment
Methods PETIT2 was a two part, randomised, multicentre, placebo-controlled study done at 38 centres in [Link]
12 countries (Argentina, Czech Republic, Germany, Hong Kong, Israel, Italy, Russia, Spain, Taiwan, Thailand, UK, S0140-6736(15)61223-5
and USA). Paediatric patients aged 1–17 years who had chronic immune thrombocytopenia and platelet counts less See Online/Articles Lancet Haem
than 30 × 10⁹ per L were randomly assigned (2:1) to receive eltrombopag or placebo. We stratified patients by age 2015; published online July 29.
[Link]
into three cohorts (patients aged 12–17 years, 6–11 years, and 1–5 years) before randomly entering them into a
S2352-3026(15)00114-3
13 week, double-blind period. Randomisation was done by the GlaxoSmithKline Registration and Medication
Department of Haematology,
Ordering System and both patients and study personnel were masked to treatment assignments. Patients who were Royal Manchester Children’s
allocated eltrombopag received tablets (except for those aged 1–5 years who received an oral suspension formulation) Hospital, University of
once per day for 13 weeks. Starting doses for patients aged 6–17 were based on bodyweight, and ethnic origin and Manchester, Manchester, UK
ranged between 50 mg/day and 25 mg/day (starting dose for patients aged 1–5 years was 1·2 mg/kg/day or (J D Grainger MD); IRCCS
Ospedale Pediatrico Bambino
0·8 mg/kg/day for east Asian patients). Patients who completed the double-blind period entered a 24 week open- Gesù, University of Pavia,
label treatment period in which all patients received eltrombopag at either the starting dose (if they were formerly Rome, Italy (F Locatelli MD);
on placebo) or their established dose. The primary outcome was the proportion of patients achieving platelet counts Prince of Songkla University,
Songklanagarind Hospital,
of at least 50 × 10⁹ per L in the absence of rescue therapy for 6 or more weeks from weeks 5 to 12 of the double-blind
Bangkok, Thailand
period. The intention-to-treat population included in the efficacy assessment consisted of all patients who were (T Chotsampancharoen MD);
randomly assigned to one of the treatment groups, and the safety population included all patients who received at Izmaylovskaya Children’s City
least one dose of study drug. This trial is registered with [Link], number NCT01520909. Clinical Hospital, Moscow
Board of Health, Moscow,
Russia (E Donyush MD); Siriraj
Findings Beginning in March 15, 2012, 92 patients were enrolled, and the trial was completed on Jan 2, 2014. Hospital, Bangkok, Thailand
63 patients were assigned to receive eltrombopag and 29 were assigned to receive placebo. In the double-blind period, (B Pongtanakul MD);
three patients discontinued treatment because of adverse events: two patients in the eltrombopag group withdrew Srinagarind Hospital, Khon
Kaen University, Khon Kaen,
because of increased liver aminotransferases and one in the placebo group withdrew because of abdominal
Thailand (P Komvilaisak MD);
haemorrhage. 25 (40%) patients who received eltrombopag compared with one (3%) patient who received placebo Chulalongkorn University,
achieved the primary outcome of platelet counts of at least 50 × 10⁹ per L for 6 of the last 8 weeks of the double-blind Bangkok, Thailand
period (odds ratio 18·0, 95% CI, 2·3–140·9; p=0·0004). Responses were similar in all cohorts (eltrombopag vs placebo: (D Sosothikul MD); Hospital de
Niños Ricardo Gutierrez,
39% vs 10% for patients aged 12–17 years, 42% vs 0% for patients aged 6–11 years, and 36% vs 0% for patients aged
Buenos Aires, Argentina
1–5 years). Proportionately fewer patients who received eltrombopag (23 [37%] of 63 patients) had WHO grades 1–4 (G Drelichman MD);
bleeding at the end of the double-blind period than did those who received placebo (16 [55%] of 29 patients); grades Ramathibodi Hospital,
2–4 bleeding were similar (three [5%] patients who received eltrombopag vs two [7%] patients who received placebo). Bangkok, Thailand
(N Sirachainan MD); Charité,
During the 24-week open-label treatment period, 70 [80%] of 87 patients achieved platelet counts of 50 × 10⁹ per L or University Medicine, Berlin,
more at least once. Adverse events that occurred more frequently with eltrombopag than with placebo included Germany (S Holzhauer MD);
nasopharyngitis (11 [17%] patients), rhinitis (10 [16%] patients), upper respiratory tract infection (7 [11%] patients), GUZ Regional Children’s Clinical
and cough (7 [11%] patients). Serious adverse events occurred in five (8%) patients who received eltrombopag and Hospital, Krasnodar, Russia
(V Lebedev MD); Primary
four (14%) who received placebo. Safety was consistent between the open-label and double-blind periods. No deaths, Children’s Medical Center, Salt
malignancies, or thromboses occurred during the trial. Lake City, UT, USA
(R Lemons MD); Faculty
Interpretation Eltrombopag, which produced a sustained platelet response in 40% of patients with chronic immune Hospital of Palacky University,
Olomouc, Czech Republic
thrombocytopenia, is a suitable therapeutic option for children with chronic symptomatic immune thrombocytopenia. (D Pospisilova MD); Regina
We identified no new safety concerns and few patients discontinued treatment because of adverse events. Margherita Children’s Hospital,
Turin, Italy (U Ramenghi MD);
Funding GlaxoSmithKline. Weill Cornell Medical College,

[Link] Published online July 29, 2015 [Link] 1

 Articles

New York, NY, USA Introduction Jak/STAT and MAPK, which results in proliferation and
(J B Bussel MD); Immune thrombocytopenia is an autoimmune disease in differentiation of megakaryocytes and increased platelet
GlaxoSmithKline, Research
Triangle Park, NC, USA
which antibodies develop against platelets,1,2 resulting in production. Eltrombopag does not compete with
(D Theodore MD); and premature destruction of platelets and inhibition of platelet endogenous thrombopoietin binding at the extracellular
GlaxoSmithKline, Collegeville, production by megakaryocytes in the bone marrow.2 thrombopoietin receptor domain and might therefore
PA, USA (K K Bakshi MS, Immune thrombocytopenia has a prevalence of roughly have an additive effect with thrombopoietin.18 Eltrombopag
M Iyengar PhD, G W Chan MD,
L M Marcello MS,
five in 100 000 children.3–5 For most affected children, is approved for the treatment of adults with chronic
C K Bailey PhD, K D Chagin MD) immune thrombocytopenia is a self-limiting disease that immune thrombocytopenia, severe aplastic anaemia, and
Correspondence to: resolves with or without intervention.3 Chronic immune chronic hepatitis C-associated thrombocytopenia. In the
Dr John D Grainger, Royal thrombocytopenia, defined as ongoing disease more phase 3 PETIT2 study, we aimed to investigate the safety
Manchester Children’s Hospital, than 12 months after diagnosis,6 occurs in 13–36% of and efficacy of eltrombopag in children using doses
Manchester M13 9WL, UK
[Link]@[Link]
children with immune thrombocytopenia.7 Some of informed by the dose-finding phase of the PETIT study
these children will have persistently low platelet counts ([Link] identifier: NCT00908037).19
with an ongoing risk of significant bleeding and
limitations in quality of life.7–9 Methods
The American Society of Hematology (ASH) and Study design
International Working Group (IWG) guidelines PETIT2 was a two part, randomised, multicentre,
acknowledge the scarcity of evidence about the treatment placebo-controlled study done at 38 centres in
of chronic immune thrombocytopenia in children.9,10 The 12 countries (Argentina, Czech Republic, Germany,
most common treatments include immunosuppression Hong Kong, Israel, Italy, Russia, Spain, Taiwan, Thailand,
with corticosteroids or other drugs, rituximab, and UK, and USA). The first part of the trial was a 13 week,
splenectomy.9 However, the side-effects and potential placebo-controlled, double-blind period, in which we
short-term and long-term risks of these interventions can randomly assigned patients 2:1 to receive either
be problematic,11–14 and splenectomy is unacceptable to eltrombopag or matching placebo. In the second part,
many families and physicians.15–17 patients who completed the double-blind period entered
Eltrombopag is an oral non-peptide thrombopoietin a 24 week, open-label treatment period. The open-label
receptor agonist that binds to the transmembrane domain period of the study allowed us to assess long-term safety
of the thrombopoietin receptor, leading to signal and efficacy without burdening paediatric patients with a
transduction through various pathways, including longer than necessary placebo-controlled period.

Research in context
Evidence before this study study of the phase 1/2 trial; one retrospective single-institution
The thrombopoietin receptor agonist eltrombopag is approved study; and one prospective, single-arm, single-institution
in the USA for treatment of chronic immune thrombocytopenia study. One article was a long-term retrospective study of
in adults who have had an insufficient response to rituximab, and two articles were single-arm, single-institution
corticosteroids, immunoglobulins, or splenectomy, and in studies of new formulations of anti-immunoglobulin D
Europe for patients who have had splenectomy and do not (intramuscular and subcutaneous). The disease setting was
respond to treatment with corticosteroids or immunoglobulins. described as chronic, refractory, or severe.
Most treatment guidelines for children with chronic immune
Added value of this study
thrombocytopenia are based on evidence from adult studies
PETIT2 is the first international, randomised, phase 3 trial of
with few randomised trials assessing treatments in children.
eltrombopag for the treatment of chronic immune
The phase 2 PETIT trial established the dosing and safety of
thrombocytopenia in paediatric patients. Our results provide
eltrombopag in paediatric patients. We searched PubMed for
further evidence to guide treatment choices for paediatric
clinical trials of existing and emerging treatments for chronic
patients with chronic immune thrombocytopenia.
immune thrombocytopenia in the paediatric setting using the
search terms “immune thrombocytopenia” and “children” with Implications of all the available evidence
the advanced settings to limit these terms to the title of the Our results for PETIT2, in addition to those of a phase 1/2 trial
article, combined with “and treatment”. The search identified of romiplostim and PETIT, provide evidence for the use of
52 articles. Of these, we excluded case reports (n=3) and those thrombopoietin receptor agonists in paediatric patients with
not assessing the efficacy of a treatment for chronic immune persistent or chronic immune thrombocytopenia who have
thrombocytopenia in children (n=42), leaving seven articles for troublesome bleeding. Many physicians and families might
further review. We excluded another article because the setting now judge thrombopoietin receptor agonists to be preferable
was primary immune thrombocytopenia. Four of the remaining to splenectomy or rituximab. Future research is needed to focus
articles were about romiplostim and included one randomised, on the long-term safety and remission rates of children treated
phase 1/2, placebo-controlled trial; one long-term follow-up with these drugs.

2 [Link] Published online July 29, 2015 [Link]

 Articles

The study was reviewed and approved by an aged 1–5 years started treatment at 1·2 mg/kg/day
independent ethics committee or institutional review (0·8 mg/kg/day for east Asian patients). Patients aged
board in accordance with applicable country-specific 6–17 years received eltrombopag tablets and patients
requirements, the International Conference on aged 1–5 years received an oral suspension formulation.
Harmonisation Good Clinical Practice guidelines, and We adjusted the dose to a maximum of 75 mg/day on
the ethical principles outlined in the Declaration of the basis of individual platelet counts. The dose was
Helsinki 2008. The trial protocol is available online. decreased if platelet counts reached more than For the trial protocol see http://
200 × 10⁹ per L. If platelet counts increased to more [Link]-clinicalstudyregister.
com/study/115450#ps
Patients than 400 × 10⁹ per L, treatment was interrupted until
Patients were eligible if they were aged 1–17 years with a platelet counts decreased to less than 150 × 10⁹ per L.
confirmed diagnosis of chronic immune Patients who received eltrombopag during the double-
thrombocytopenia (duration of >12 months) in blind period continued at the same dose when they
accordance with the IWG guidelines;6 had a platelet entered the open-label period. Patients assigned to
count less than 30 × 10⁹ per L; had relapsed or refractory receive placebo during the double-blind period received
disease after one or more previous treatments for the prespecified eltrombopag starting dose when they
immune thrombocytopenia; or were unable to continue entered the open-label period.
other treatment for immune thrombocytopenia. We
allowed concomitant drugs for immune
thrombocytopenia if given at a stable dose for 4 weeks or 118 patients assessed for eligibility 26 ineligible
more before the study start. We obtained written 11 platelet count ≥30 × 109 per L
9 laboratory values outside
informed consent from each patient’s legal guardian. normal range
Patients older than 6 years were asked to sign an assent 3 withdrew consent before
form in accordance with local laws. 92 enrolled randomisation
1 exceeded screening period
2 unable to comply with protocol
Randomisation and masking 92 randomised
Patients were randomly assigned (2:1) to eltrombopag or
matching placebo. Randomisation was stratified by age
cohort (12–17, 6–11, and 1–5 years), and a block size of six
was used within each stratum. Randomisation of patients 63 assigned eltrombopag 29 assigned placebo
and allocation of masked study drugs was done with the
GlaxoSmithKline Registration and Medication Ordering
2 discontinued treatment 1 discontinued treatment
System (RAMOS), a telephone-based interactive voice 2 increased liver 1 abdominal haemorrhage
response system. The randomisation schedule was aminotransferases
computer generated by the statistician in the
GlaxoSmithKline Biostatistical Department, using 61 completed double-blind 28 completed double-blind
RandAll, a GlaxoSmithKline-validated randomisation treatment treatment
system. Patients and study personnel (including those
employed by the funder) were masked to treatment
63 included in intention-to-treat 29 included in intention-to-treat
assignment during the double-blind period. Because analysis analysis
RAMOS allocated the masked medication, the
investigator did not know what treatment (eltrombopag
or placebo) was being given to the patient. Once the week 2 withdrew
2 insufficient efficacy
12 data were locked, the investigator was unmasked to
the treatment assignment for each patient by an
independent statistician to work out the appropriate
starting dose for the open-label period. 87 assigned open-label eltrombopag 7 withdrew
4 adverse events
3 hepatobiliary laboratory
Procedures abnormalities
We based starting doses on results from the open-label, 1 thrombocytopenia
80 completed treatment 2 insufficient efficacy
dose-finding period of the PETIT study.19 Patients of east 1 withdrawn consent from
Asian ancestry received a lower starting dose because parent or guardian
they have increased plasma eltrombopag exposures. 77 completed 24 week ocular follow-up
Patients in the cohorts aged 6–11 years and 12–17 years
who weighed 27 kg or more started treatment at
50 mg/day (25 mg/day for east Asian patients), and those 87 included in intention-to-treat analysis
weighing less than 27 kg started treatment at 37·5 mg/day
(25 mg/day for east Asian patients). Patients in the cohort Figure 1: Trial profile

[Link] Published online July 29, 2015 [Link] 3

 Articles

All patients were allowed to receive standard-of-care doses of concomitant drugs for immune thrombocytopenia,
treatment for chronic immune thrombocytopenia. Any platelet transfusions, and splenectomies during the study
new drugs for immune thrombocytopenia, increased were deemed rescue medications. Complete blood counts,
including platelets, were assessed at weekly visits for the
first 16 weeks, then every 4 weeks for patients on a stable
Eltrombopag (n=63) Placebo (n=29) dose of eltrombopag. A patient was classified as a non-
Age (years) 9·4 [8·2–10·5] 9·8 [8·3–11·3] responder after the initiation of rescue treatment and for
Males 33 (52%) [39–65] 15 (52%) [33–71] the duration of its use until platelet counts decreased to
Weight (kg) 41·0 (35·5–46·4) 42·7 (33·2–52·3) less than 50 × 10⁹ per L. Patients with subsequent
Time since diagnosis of 41 (34·1) 53 (40·3) assessments of platelet counts of 50 × 10⁹ per L or more
immune thrombocytopenia were classified as responders.
(months)
Patients completed 24–28 weeks of follow-up visits,
Baseline immune 13 (21%) [11–33] 1 (3·4%) [0–18]
thrombocytopenia drug use
including an ophthalmic examination 24 weeks after the
Previous immune
end of the open-label period. Patients who completed
thrombocytopenia both the double-blind period and open-label period were
medication deemed to have completed the study. Participation from
Any 60 (95%) [87–99] 28 (97%) [82–100] screening to the end of the 24 week follow-up period
≥2 46 (73%) 26 (90%) lasted about 16 months.
Rituximab 9 (14%) 6 (21%)
Anti-D 13 (21%) 3 (10%) Outcomes
Splenectomy 4 (6%) [2–15] 0 [0–12] The primary outcome was the proportion of patients
Platelets ≤15 × 10⁹ per L 38 (60%) [47–72] 19 (66%) [46–82] achieving platelet counts of at least 50 × 10⁹ per L in the
Ethnic origin absence of rescue therapy for 6 or more weeks from
East Asian* 20 (32%) [21–45] 10 (34%) [18–54] weeks 5–12 of the double-blind period. Secondary
Other† 43 (68%) [55–79] 19 (66%) [46–82] endpoints in both the double-blind period and open-label
period were safety, need for rescue treatment, incidence
Data are mean [95% CI], mean (range), mean (SD), n (%) [95% CI], or n (%). *East
of WHO-classified bleeding, and the maximum period of
Asian=Japanese, east Asian heritage, or southeast Asian heritage. †Includes white
(Arabic, north African, and European heritage), African American (n=1), and continuous platelet counts of 50 × 10⁹ per L or more.
mixed race (n=1). Secondary endpoints with respect to platelet response
assessments in the double-blind period were weighted
Table 1: Baseline characteristics
mean platelet change up to week 12, the proportion of

12–17 years 6–11 years 1–5 years Total Odds ratio p value
(95% CI)
Eltrombopag Placebo Eltrombopag Placebo Eltrombopag Placebo Eltrombopag Placebo
(n=23) (n=10) (n=26) (n=13) (n=14) (n=6) (n=63) (n=29)
Primary endpoint
Responders 9 (39%) 1 (10%) 11 (42%) 0 5 (36%) 0 25 (40%) 1 (3%) 18·0 (2·3–140·9) 0·0004
95% CI 0·20–0·61 0·00–0·45 0·23–0·63 NA 0·13–0·65 NA ·· ··
East Asian responders, n/N ·· ·· ·· ·· ·· ·· 7/20 (35%) 0/10 NC
(%)
Secondary endpoints
Responders during weeks ·· ·· ·· ·· ·· ·· ·· ·· 25·3 (8·2–78·7) <0·0001
1–12
Responders any time ·· ·· ·· ·· ·· ·· 47 (75%) 6 (21%) 11·7 (4·0–34·5) <0·0001
weeks 1–12
Responders any time 15 (65%) 2 (20%) 17 (65%) 3 (23%) 7 (50%) 0 39 (62%) 5 (17%) 8·3 (2·7–25·1) 0·00018
weeks 1–6
Maximum continuous <0.0001
response (first 12 weeks)
Mean (SD), weeks 3·6 (3·2) 1·0 (2·5) 3·4 (3·0) 0·2 (0·4) 2·8 (3·5) 0·0 (0·0) 3·3 (3·1)* 0·4 (1·5) ·· ··
Median (range), weeks 2·0 (0–10) 0·0 (0–8) 3·0 (0–11) 0·0 (0–1) 1·5 (0–12) 0·0 (0–0) 3·0 (0–12) 0·0 (0–8) ·· ··
Rescue therapy needed 3 (13%) 2 (20%) 6 (23%) 3 (23%) 3 (21%) 2 (33%) 12 (19%) 7 (24%) 0·44 (0·2–0·9) 0·032

Data are n (%) unless otherwise indicated. Data for east Asian responders were not analysed by cohort. Data for the secondary endpoints responders at any time and responders at weeks 1–12 were only done for
the treatment groups overall and not by cohort. NA=not available. NC=cannot be calculated. *Van Elteren stratified rank test.

Table 2: Efficacy during the double-blind period

4 [Link] Published online July 29, 2015 [Link]

 Articles

patients achieving platelet counts of 50 × 10⁹ per L or A

more at any time during the first 6 weeks and 12 weeks in 100 Eltrombopag
the absence of rescue therapy, and odds ratio (OR) of a 90
Placebo
patient achieving a platelet count of 50 × 10⁹ per L or
80
more on at least one occasion from weeks 1 to 12 in the

Patients with a response (%)

absence of rescue therapy. Secondary endpoints with 70
respect to platelet response assessments in the open- 60
label treatment period included the proportion of patients 50
who achieved platelet counts of 50 × 10⁹ per L or more at
40
any time and from weeks 4 to 24, and reduction or
discontinuation of concomitant drugs for immune 30

thrombocytopenia. We collected samples for 20

pharmacokinetic studies (data not shown). The appendix 10
shows the liver chemistry stopping criteria. 0
We coded and grouped adverse events by system organ 1 2 3 4 5 6 7 8 9 10 11 12
class and preferred term in accordance with the Medical Number of
Study week
Dictionary for Regulatory Activities and graded them evaluable
patients
using the National Cancer Institute (NCI) Common Eltrombopag 63 63 63 62 63 63 63 63 63 63 63 63
Terminology Criteria for Adverse Events (CTCAE) v4.0, Placebo 29 29 28 29 29 29 29 29 29 29 29 29
except for grade 4 thrombocytopenia, which was judged
B
as treatment failure. Patients underwent eye
100
examinations at screening, every 12 weeks while on study
treatment, and at the 24 week follow-up visit. The ocular 90

data were assessed by an independent clinical events 80

Patients with a response (%)

committee that was blinded to the treatment assignment 70

of the patients. 60

50
Statistical analysis
Assuming response rates of 50% for eltrombopag and 40

10% for placebo based on response rates in the adult 30

clinical trial (RAISE),20 we calculated that 66 evaluable 20
patients were needed to provide 90% or more power at
10
the 5% significance level (two-sided), calculated using
0
the Fisher-exact test option in Power Analysis and Sample 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37
Size System 2005. To compensate for 10% missing data, Study week
Number of
we planned to enrol at least 75 patients across all groups. evaluable
Hypothesis testing and 95% CIs were all two-sided. patients
The intention-to-treat population included in the efficacy Eltrombopag 87 86 86 68 62 55 52 53 45 42 40 63 30 38 43 47 36 34 44 39 41 34 36 79

assessment consisted of all patients who were randomly Figure 2: Response by week in the double-blind period (A) and open-label period (B)
assigned to one of the treatment groups. The safety
population included all patients who received at least one
dose of study drug. We did the primary efficacy analysis generalised linear mixed model with a logit canonical See Online for appendix
with stratified Cochran-Mantel-Haenszel χ² test statistics link function for repeated binary data, allowing for age For information about the
that adjusted for age cohorts. To assess the treatment by cohort (1–5 years, 6–11 years, and 12–17 years) and Medical Dictionary for
Regulatory Activities see
cohort interaction, we used the Breslow-Day test for treatment as fixed effects. Additionally, “patient” was
[Link]
homogeneity of treatment effect. If data were missing treated as a random effect and assumed to follow a
during weeks 5 to 12 of the double-blind period normal distribution (~N[0,σ2]). This model relies on the
(irrespective of the reason), we treated this absence as a assumption that missing data are missing at random.
negative response (at the corresponding visit) in the We used logistic regression to calculate the proportion of
computation of the primary efficacy endpoint. patients who achieved a response at any time period. We
During the secondary analyses of the double-blind analysed weighted mean platelet counts by covariance.
period, we analysed the OR of a patient achieving a To calculate the maximum duration of continuous
platelet response with a generalised linear mixed-effects response, we used the van Elteren stratified rank test.
model for repeated binary data, including supportive Descriptive summaries are reported in the results for all
analysis that used the generalised estimating equation. other analyses.
We included platelet data collected every week for An external Data Safety Monitoring Board was provided
12 weeks in the repeated measures analysis. We used a with unmasked safety and efficacy data after 35 patients

[Link] Published online July 29, 2015 [Link] 5

 Articles

assistance for manuscript preparation. All authors had

A
150 Eltrombopag full access to all the data in the study and had final
140 Placebo responsibility for the decision to submit for publication.
130
120 Results
Median platelet count (× 109 per L)

110
The trial was started on March 15, 2012, and completed
100
90 on Jan 2, 2014. We screened 118 patients for eligibility
80 and excluded 26 of them (figure 1), leaving 92 patients.
70 These 92 patients were randomly assigned to treatment
60 in the double-blind period, with 63 patients assigned to
50 receive eltrombopag and 29 patients assigned to receive
40
placebo (figure 1). 33 patients were aged 12–17 years,
30
20 39 were aged 6–11 years, and 20 were aged 1–5 years. In
10 the double-blind period, three patients discontinued
0 treatment because of adverse events: two patients in the
Baseline 1 2 3 4 5 6 7 8 9 10 11 12 13
eltrombopag group withdrew because of increased liver
Study week
Number of aminotransferases and one in the placebo group
evaluable
patients withdrew because of abdominal haemorrhage. Two
Eltrombopag 62 63 63 63 62 62 61 63 63 61 62 61 61 60 patients who received eltrombopag withdrew after
Placebo 29 29 29 28 29 27 28 28 28 27 28 28 28 28
completing the double-blind period because of
B insufficient efficacy. 59 (94%) patients receiving
200 eltrombopag and 28 patients (97%) receiving placebo
190 completed the double-blind period and entered the open-
180 label treatment period. 55 (87%) patients who originally
170
received eltrombopag and 25 (86%) who originally
160
150
received placebo completed the study.
At baseline, 72 (78%) of 92 patients had received at least
Median platelet count (× 109 per L)

140
130 two previous treatments for immune thrombocytopenia,
120 and only four patients had undergone splenectomy, all of
110 whom were assigned to receive eltrombopag (table 1).
100
57 (62%) patients had baseline platelet counts of 15 × 10⁹
90
80
per L or less (table 1). More patients assigned who
70 received eltrombopag (13 [21%] of 63 patients) were
60 receiving drugs for immune thrombocytopenia at
50 baseline than were patients who received placebo (1 [3%]
40 of 29 patients).
30
More patients who received eltrombopag (25 [40%] of
20
10
63 patients) achieved the primary endpoint of a platelet
0 count at least 50 × 10⁹ per L without rescue therapy for
14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 6 weeks or more from weeks 5 to 12 than did those who
Study week
Number of received placebo (1 [3%] of 29 patients; OR 18·0, 95% CI
evaluable
patients
2·3–140·9, p=0·0004; table 2). The response to
Eltrombopag 87 86 86 68 62 55 52 53 45 42 40 63 30 38 43 47 36 34 44 39 41 34 36 79 eltrombopag was consistent in all age cohorts. In the
double-blind period, a significantly greater proportion of
Figure 3: Median platelet count during the double-blind period (A) and open-label period (B)
patients in the eltrombopag group had a platelet response
during weeks 1–6 (OR 8·3, 95% CI 2·7–25·1, p=0·00018)
completed the double-blind period. Additionally, a and weeks 1–12 (11·7, 4·0–34·5, p<0·0001) than did
clinical events committee reviewed the data from the patients in the placebo group (table 2; figure 2A). The
ophthalmic assessments after all patients had completed magnitude of the platelet count response during the
the study. This trial is registered with [Link], double-blind period, as shown by the weighted mean
number NCT01520909. platelet change, was higher for patients treated with
eltrombopag than for those treated with placebo (mean
Role of the funding source area under the curve 63·9 for eltrombopag vs 23·7 for
Authors employed by the funder of the study contributed placebo, p<0·0001). A repeated-measures analysis of
to the study concept and design, data interpretation and platelet response during 12 weeks in the double-blind
analysis, and final approval to submit for publication. period showed that patients treated with eltrombopag
The study funder provided funding for editorial had higher likelihood of maintaining a response than did

6 [Link] Published online July 29, 2015 [Link]

 Articles

Eltrombopag Placebo 12–17 years 6–11 years 1–5 years Total

(n=31) (n=37) (n=19) (n=87)
Double-blind period (baseline)
Grades 1–4 (any bleeding) 45/63 (71%) 20/29 (69%) Responder any time 24 (77%) 28 (76%) 18 (95%) 70 (80%)

Grades 2–4 (clinically significant) 16/63 (25%) 6/29 (21%) Cumulative weeks of response, weeks 4–24

Double-blind period (week 12) Mean (SD), weeks 10·0 (8·1) 9·9 (7·8) 10·1 (7·1) 10·0 (7·7)

Grades 1–4 (any bleeding) 23/63 (37%) 16/29 (55%) Median (range), weeks 11 (0–21) 11 (0–21) 10 (0–21) 11 (0–21)

Grades 2–4 (clinically significant) 3/63 (5%) 2/29 (7%) Maximum continuous response, weeks 1–24

Open-label period (baseline) Mean (SD), weeks 8·5 (7·7) 8·6 (8·0) 8·6 (8·1) 8·6 (7·8)

Grades 1–4 (any bleeding) 55/87 (63%) ·· Median (range), weeks 8·0 (0–24) 8·0 (0–24) 6·0 (0–24) 6·0 (0–24)

Grades 2–4 (clinically significant) 17/87 (20%) ·· Rescue therapy use 3 (10%) 6 (16%) 2 (10·5%) 11 (13%)

Open-label period (week 24) Concomitant drugs for immune

thrombocytopenia
Grades 1–4 (any bleeding) 19/79 (24%) ··
Any 5 (16%) 9 (24%) 6 (32%) 20 (23%)
Grades 2–4 (clinically significant) 5/79 (6%) ··
Corticosteroids 5 (16%) 8 (22%) 5 (26%) 18 (21%)
Data are n/N (%). Intravenous immunoglobulin 2 (6%) 4 (11%) 0 1 (1%)
Ciclosporin 0 1 (3%) 0 1 (1%)
Table 3: Bleeding by WHO bleeding scale
Mycophenolate 1 (3%) 0 0 1 (1%)
Dapsone 0 0 1 (5%) 1 (1%)

patients treated with placebo (OR 25·3, 95% CI 8·2–78·7, Data are n (%) unless indicated otherwise.
p<0·0001). Figure 3 shows median changes in platelet
Table 4: Efficacy during the open-label period
counts during the double-blind period. The mean
maximum continuous response during the double-blind
period was also significantly longer with eltrombopag
than with placebo (table 2). Finally, the proportion of 12–17 years 6–11 years 1–5 years
patients who received rescue treatment in the double- Eltrombopag Placebo Eltrombopag Placebo Eltrombopag Placebo
blind period was smaller for eltrombopag (12 [19%] of
Double-blind period
63 patients) than for placebo (7 [24%] of 29 patients;
Number of patients 23 10 26 13 14 6
p=0·032; table 2).
Median average dose per 69·0 70·1 50·7 63·7 26·7 35·2
Table 3 shows bleeding events for the double-blind day (mg)
period. Of patients with WHO grades 2–4 bleeding, three
Average dose per day (mg/kg) 1·0 1·0 1·6 2·0 1·7 1·8
patients treated with placebo (one in each cohort)
Median maximum dose 1·2 1·2 1·9 2·4 2·1 2·4
reported WHO grade 3 bleeding during the double-blind (mg/kg)
period. No patients treated with eltrombopag reported Maximum average dose per 71 73 71 71 46 45
grade 3 bleeding during the same period. No patients in day (mg)
either group had WHO grade 4 bleeding. Number of east Asian 6 3 10 5 4 2
Figure 2 shows the proportion of patients receiving patients*
eltrombopag who achieved a platelet response each Median average dose per 56·8 63·3 40·2 63·4 27·2 35·8
day (mg)
week during the open-label period. Median platelet
Median average dose per day 1·3 1·1 1·2 1·9 2·2 2·0
counts were at least 50 × 10⁹ per L at most assessments
(mg/kg)
(figure 3), and 70 (80%) of 87 patients achieved platelet
Median maximum dose 1·5 1·3 1·9 2·5 4·7 4·1
counts of 50 × 10⁹ per L or more at least once during the (mg/kg)
24 weeks (table 4). During the open-label period, Open-label period
duration of response was consistent in responders from Number of patients 31 ·· 37 ·· 19 ··
all cohorts (table 4). The proportion of patients needing Median average dose per day 67·7 ·· 56·9 ·· 42·8 ··
rescue therapy decreased during the open-label (mg)
treatment period (11 [13%] of 87 patients; table 4). Average dose per day (mg/kg) 1·0 ·· 1·6 ·· 1·9 ··
During the open-label treatment period, the proportion Maximum average dose per 75 ·· 75 ·· 72 ··
of patients reporting any WHO-classified bleeding day (mg)
decreased (table 3). No grade 3 or 4 bleeding was Number of east Asian patients 9 ·· 14 ·· 6 ··
reported. Median average dose per 58·6 ·· 37·2 ·· 48·9 ··
Of the 87 patients in the open-label treatment period, day (mg)

15 were taking concomitant drugs for immune Average dose per day 1·1 ·· 1·3 ·· 3·4 ··
(mg/kg)
thrombocytopenia at baseline of the open-label period.
Nine patients attempted a reduction or discontinuation *Starting dose was 25 mg for east Asian patients aged 6–12 years and 0·8 mg/kg for east Asian patients aged 1–5 years.
of concomitant drugs. Seven patients permanently
Table 5: Dose by cohort
discontinued all baseline concomitant drugs and did not

[Link] Published online July 29, 2015 [Link] 7

 Articles

Eltrombopag Placebo Eltrombopag in Placebo in

(n=63) (n=29) double-blind double-blind
period (n=63) period (n=29)
Any adverse event* 51 (81%) 21 (72%)
Nasopharyngitis 11 (17%) 2 (7%) Any serious adverse event 5 (8%) 4 (14%)

Rhinitis 10 (16%) 2 (7%) Gingivitis 1 (2%) 0

Epistaxis 8 (13%) 6 (21%) Influenza 1 (2%) 0

Upper respiratory tract infection 7 (11%) 1 (3%) Aseptic meningitis 1 (2%) 0

Cough 7 (11%) 0 Pneumonia 1 (2%) 0

Headache 6 (10%) 3 (10%) Fungal pneumonia 1 (2%) 0

Abdominal pain 6 (10%) 0 Abnormal ALT 1 (2%)* 0

Pyrexia 4 (6%) 1 (3%) Abnormal AST 1 (2%)* 0

AST increased 4 (6%) 0 Epistaxis 0 1 (3%)

Upper abdominal pain 3 (5%) 4 (14%) Petechiae 0 1 (3%)

Vomiting 2 (3%) 3 (10%) Haemorrhage 0 1 (3%)

Hypertensive crisis 0 1 (3%)
Data are n (%). Adverse events that occurred in ≥5% of patients in the
double-blind period are reported. Patients could experience more than one event. Data are n (%). Frequencies are based on number of patients experiencing the event.
AST=aspartate aminotransferase. *Adverse events occurring at any frequency. Patients could experience more than one event. ALT=alanine aminotransferase.
AST=aspartate aminotransferase. *Judged by the investigator to be drug-related.
Table 6: Adverse events in the double-blind period
Table 7: Serious adverse events in the double-blind period

need rescue therapy, and an eighth patient had a Similar trends were seen for the frequency of adverse
sustained reduction of baseline corticosteroids without events during the open-label treatment period, with
needing rescue. The ninth patient attempted to reduce or 69 (79%) of 87 patients reporting an adverse event. The
stop a concomitant drug but subsequently received most common adverse events were similar in type and
rescue medication. Table 5 shows the median average incidence to those in patients receiving eltrombopag
daily dose for the whole study population and for east during the double-blind period. Nine (10%) patients had
Asian patients during the double-blind period and open- serious adverse events and eight (9%) patients had
label treatment period. grade 3 or 4 adverse events. One patient had grade 4
During the double-blind period, the most common neutropenia, which was judged to be unrelated to
adverse events that occurred in a higher proportion of eltrombopag by the local investigator. No
patients receiving eltrombopag than in those receiving thromboembolic adverse events, malignancies, or deaths
placebo included nasopharyngitis, rhinitis, upper occurred during the study. Of the 92 patients in the
respiratory tract infection, and cough (table 6). Adverse PETIT2 study, two patients treated with eltrombopag
events were similar in east Asian patients (appendix). were judged by an external ocular clinical events
The most common adverse events that occurred in a committee to have a new cataract (n=1) or progression of
higher proportion of patients given placebo than in those a pre-existing cataract (n=1); both patients were receiving
given eltrombopag were epistaxis, upper abdominal corticosteroids.
pain, headache, and vomiting. Hepatobiliary laboratory findings during both the
Serious adverse events were reported in a greater double-blind period and open-label period were mostly
proportion of patients given placebo than in those given mild, reversible, and not accompanied by clinically
eltrombopag (table 7). Serious adverse events occurred in significant symptoms that would suggest impaired liver
five (8%) of 63 patients treated with eltrombopag function. During the double-blind period, five (8%) of
(infections and transaminitis events; appendix) and four 63 patients receiving eltrombopag (three white patients
(14%) of 29 patients who were given placebo (bleeding and two east Asian patients) had alanine aminotransferase
events assessed by NCI CTCAE). The most common (ALT) levels three or more times the upper limit of
drug-related adverse events were aminotransferase normal (ULN), two of whom met the liver chemistry
abnormalities, which we detected in four patients treated stopping criteria (appendix). Hepatobiliary laboratory
with eltrombopag but no patients given placebo. parameters resolved in two of the three remaining
Most adverse events were CTCAE grades 1 and 2 for patients with ALT concentrations three or more times the
both patients treated with eltrombopag and those given ULN while they were still on treatment and in one patient
placebo. We identified grade 3 or 4 adverse events in after study treatment discontinuation in the open-label
3 (5%) of 63 patients who received eltrombopag and treatment period. During the open-label treatment
8 (28%) of 29 patients who received placebo. No patients period, six (7%) of 87 patients receiving eltrombopag (one
treated with eltrombopag had a grade 4 adverse event, white patient and five east Asian patients) had ALT
but one patient who received placebo had a grade 4 concentrations three or more times the ULN, three of
haemorrhage (not assessed by WHO criteria). whom met the protocol-defined stopping criteria.

8 [Link] Published online July 29, 2015 [Link]

 Articles

Discussion frequent in east Asian patients; the abnormal liver test

Eltrombopag is licensed by the US Food and Drug results resolved after patients stopped taking eltrombopag.
Administration and European Medicines Agency for the Two patients showed early signs of ocular lens changes
treatment of chronic immune thrombocytopenia in while receiving simultaneous corticosteroids, which might
adults. The ASH and IWG guidelines concluded that be a confounding factor. We identified no thromboses or
evidence was insufficient to support the use of malignancies.
thrombopoietin receptor agonists in children.9 However, The side-effect profile of eltrombopag in PETIT2 is
the IWG added that, if the safety and efficacy data similar to those reported in the PETIT and adult
reported for adults were reproduced in children, “they studies.18,25,28,29 33% of patients in PETIT2 were east Asian;
could be used not only for children with chronic the doses needed and response rates detected for east
refractory immune thrombocytopenia, but also in those Asian patients were similar to those of the overall
with persistent but highly symptomatic disease resistant population, and adverse event rates were similar, with
to usual first-line treatments”.10 the exception of an increased number of liver adverse
Many children with persistent or chronic immune events.
thrombocytopenia have minimal bleeding and retain The safety and efficacy data in PETIT2 fulfil the IWG
good health-related quality of life.20,21 However, treatment criteria for use in chronic immune thrombocytopenia, as
options are inadequate and might not improve health- also seen in PETIT. These studies suggest that the use of
related quality of life22 for children with problematic eltrombopag might be a reasonable approach to support
disease, and the most frequently used interventions are platelet counts while waiting for spontaneous
immunosuppressive drugs, rituximab, or splenectomy.9 improvement in children. Liver function tests and eye
Immunosuppressive drugs have been available for a examinations might be needed, especially when
long time but are associated with adverse events;9 therefore, eltrombopag is used in combination with steroids. In
new treatments are needed. Splenectomy is an alternative PETIT and PETIT2, treatment was restricted to
approach for chronic immune thrombocytopenia, but 6–9 months; longer follow-up of paediatric patients
risks of infection, thrombotic complications, and cancer receiving eltrombopag is needed to fully assess its role in
have been reported in meta-analyses and in long-term the treatment of chronic immune thrombocytopenia.
follow-up.23–25 Since chronic immune thrombocytopenia Although immune thrombocytopenia-specific bleeding
can spontaneously resolve in children,8 the long-term risks scales are available, we chose to use the WHO bleeding
for children undergoing splenectomy are unacceptable to scale and the NCI CTCAE to assess bleeding. Use of the
many physicians and families. This unacceptability was WHO bleeding scale allowed us to compare our results
shown by a splenectomy rate at recruitment of only 7% in to those of eltrombopag trials in adults, but we accept
PETIT and 4% in PETIT2.19 that not using a more appropriate bleeding scale for
Rituximab is not licensed for immune thrombocytopenia. patients with chronic immune thrombocytopenia is a
A randomised controlled trial showed no benefit of limitation of this trial. Although very few data were
rituximab monotherapy compared with placebo in adults.26 missing during the double-blind period, we assessed the
Pooled data from a meta-analysis of paediatric studies robustness of the analysis to missingness. This analysis
reported a 39% complete response rate with a median did not suggest departure from the primary inferences.
response duration of only 12·8 months.27 In one study, Treatment with eltrombopag led to stable platelet
Patel and colleagues14 estimated that only 26% of children responses, reduction in grades 1–4 bleeding from baseline
would have a persisting response to rituximab for 5 years.27 to the end of the study, and allowed discontinuation of
A sustained response rate was achieved by 40% of baseline immune thrombocytopenia drugs in children
patients receiving eltrombopag in PETIT2, as shown by with chronic immune thrombocytopenia. We identified
platelet counts of at least 50 × 10⁹ per L without rescue for no new safety concerns, and the only discontinuations
6 or more weeks during weeks 5–12 of the double-blind because of adverse events were related to hepatobiliary
period. In the open-label period of PETIT2, the efficacy laboratory abnormalities, which resolved after drug
of eltrombopag was shown by responses in 80% of discontinuation. The responses seen in this study are
patients as defined by platelet counts of 50 × 10⁹ per L or consistent with those seen in PETIT19 and the adult
more at least once. Proportions of patients with responses studies of eltrombopag for treatment of chronic immune
were similar between age cohorts. A clinical response thrombocytopenia.25,28–30 In addition to the PETIT study19
was shown by a reduction in bleeding severity (WHO and phase 1/2 studies of romiplostim,31,32 PETIT2 provides
grades 1–4) from 71% at baseline to 24% at week 24. Most evidence for the use of thrombopoietin receptor agonists
patients taking concomitant drugs for immune in paediatric patients with chronic immune
thrombocytopenia discontinued these drugs. thrombocytopenia.
Eltrombopag was well tolerated. The doses needed for Contributors
children were broadly similar to those needed for adults. All authors contributed to the writing of the manuscript through critical
Increased liver enzyme concentrations led to drug review, comments, and approval of early drafts of the manuscript, and all
authors approved the final draft for publication. TC contributed to the
discontinuation for five of 92 patients and were most

[Link] Published online July 29, 2015 [Link] 9

 Articles

conception and design, acquisition of data, and the data analysis and 11 Blanchette V, Imbach P, Andrew M, et al. Randomised trial of
interpretation of the study. KKB, MI, GWC, LMM, and CKB contributed intravenous immunoglobulin G, intravenous anti-D, and oral
to the conception and design and the data analysis and interpretation of prednisone in childhood acute immune thrombocytopenic purpura.
the study. DT and KDC contributed to the data analysis and interpretation Lancet 1994; 344: 703–07.
of the study. FL, ED, BP, PK, DS, GD, NS, VL, DP, UR, and JBB 12 Bennett CM, Rogers ZR, Kinnamon DD, et al. Prospective phase
contributed to the acquisition of data for the study. JDG, SH, and RL 1/2 study of rituximab in childhood and adolescent chronic
contributed to the acquisition of data and the data analysis and immune thrombocytopenic purpura. Blood 2006; 107: 2639–42.
interpretation of the study. 13 Cooper N, Bussel JB. The long-term impact of rituximab for childhood
immune thrombocytopenia. Curr Rheumatol Rep 2010; 12: 94–100.
Declaration of interests 14 Patel VL, Mahevas M, Lee SY, et al. Outcomes 5 years after response
JDG has received research funding from Baxter and honoraria from to rituximab therapy in children and adults with immune
Amgen, Baxter, and GlaxoSmithKline. BP and PK have received research thrombocytopenia. Blood 2012; 119: 5989–95.
funding from GlaxoSmithKline. DS has received research funding from 15 Neunert CE, Bright BC, Buchanan GR. Severe chronic refractory
CSL Behring and GlaxoSmithKline. JBB has been a consultant for immune thrombocytopenic purpura during childhood: a survey of
Portola; received research funding from Amgen, Cangene, Eisai, physician management. Pediatr Blood Cancer 2008; 51: 513–16.
Genzyme, GlaxoSmithKline, IgG of America, Immunomedics, Ligand, 16 Vianelli N, Palandri F, Polverelli N, et al. Splenectomy as a curative
Shionogi, and Sysmex; participated in advisory boards for Amgen, Eisai, treatment for immune thrombocytopenia: a retrospective analysis
GlaxoSmithKline, Ligand, Shionogi, and Symphogen, and has received of 233 patients with a minimum follow up of 10 years.
honoraria from Amgen, Bristol-Myers Squibb, Cangene, Haematologica 2013; 98: 875–80.
GlaxoSmithKline, Novartis, and Rigel. KKB, MI, GWC, KDC, DT, LMM, 17 Wang KKW, Charles C, Heddle NM, et al. Understanding why
and CKB were employees at GlaxoSmithKline during the time of study patients with immune thrombocytopenia are deeply divided on
conduct and initial publication development. KKB, DT, and CKB hold splenectomy. Health Expect 2014; 17: 809–817.
equity ownership in GlaxoSmithKline. DP received a fee for being 18 Erickson-Miller CL, Delorme E, Tian SS, et al. Preclinical activity of
principal investigator at a local site in the Czech Republic from eltrombopag (SB-497115), an oral, nonpeptide thrombopoietin
receptor agonist. Stem Cells 2009; 27: 424–30.
GlaxoSmithKline. LMM, MI, GWC, DT, CKB, and KKB are now
employees of Novartis Pharmaceuticals Corporation. FL, TC, ED, GD, 19 Bussel J, Garcia de Miguel P, Despotovic J, et al. Eltrombopag
treatment of childhood persistent and chronic immune
NS, SH, VL, RL, and UR declare no competing interests.
thrombocytopenia: final results of the PETIT study (TRA108062):
Acknowledgments a phase 2, placebo-controlled clinical trial. Haematologica 2014;
The study was funded by GlaxoSmithKline; eltrombopag is an asset of 99: S733 (abstr).
Novartis AG as of March 2, 2015. We thank Andres Brainsky for his 20 Cheng G, Saleh MN, Marcher C, et al. Eltrombopag for
contributions to the study design, and Rosanna Tedesco for editorial management of chronic immune thrombocytopenia (RAISE):
assistance with the preparation of the manuscript. We also thank the a 6-month, randomised, phase 3 study. Lancet 2011; 377: 393–402.
patients and the principal investigators and their institutions for their 21 Strullu M, Rakotonjanahary J, Tarral E, et al. Evaluation of health
contributions to the study (appendix). Editorial assistance was provided related quality of life in children with immune thrombocytopenia
by Vanessa A Marrero and Nancy E Price of AOI Communications and with the PedsQL 4.0 Generic Core Scales: a study on behalf of the
pays de Loire pediatric hematology network.
was funded by GlaxoSmithKline and Novartis Pharmaceuticals
Health Qual Life Outcomes 2013; 11: 193.
Corporation.
22 Grainger JD, Young NL, Blanchette VS, Klaassen RJ. Quality of life
References in immune thrombocytopenia following treatment.
1 Ballem PJ, Segal GM, Stratton JR, Gernsheimer T, Adamson JW, Arch Dis Child 2013; 98: 895–97.
Slichter SJ. Mechanisms of thrombocytopenia in chronic 23 Bisharat N, Omari H, Lavi I, Raz R. Risk of infection and death
autoimmune thrombocytopenic purpura. Evidence of both impaired among post-splenectomy patients. J Infect 2001; 43: 182–86.
platelet production and increased platelet clearance. 24 Boyle S, White RH, Brunson A, Wun T. Splenectomy and the
J Clin Invest 1987; 80: 33–40. incidence of venous thromboembolism and sepsis in patients with
2 McMillan R, Wang L, Tomer A, Nichol J, Pistillo J. Suppression of immune thrombocytopenia. Blood 2013; 121: 4782–90.
in vitro megakaryocyte production by antiplatelet autoantibodies 25 Kristinsson SY, Gridley G, Hoover RN, Check D, Landgren O.
from adult patients with chronic ITP. Blood 2004; 103: 1364–69. Long-term risks after splenectomy among 8,149 cancer-free
3 Reid MM. Chronic idiopathic thrombocytopenic purpura: American veterans: a cohort study with up to 27 years follow-up.
incidence, treatment, and outcome. Arch Dis Child 1995; Haematologica 2014; 99: 392–98.
72: 125–28. 26 Arnold DM, Heddle NM, Carruthers J, et al. A pilot randomized
4 Hedman A, Henter JI, Hedlund I, Elinder G. Prevalence and trial of adjuvant rituximab or placebo for nonsplenectomized
treatment of chronic idiopathic thrombocytopenic purpura of patients with immune thrombocytopenia. Blood 2012; 119: 1356–62.
childhood in Sweden. Acta Paediatr 1997; 86: 226–27. 27 Liang Y, Zhang L, Gao J, Hu D, Ai Y. Rituximab for children with
5 Zeller B, Helgestad J, Hellebostad M, et al. Immune immune thrombocytopenia: a systematic review. PLoS One 2012;
thrombocytopenic purpura in childhood in Norway: a prospective, 7: e36698.
population-based registration. Pediatr Hematol Oncol 2000; 28 Bussel JB, Cheng G, Saleh MN, et al. Eltrombopag for the treatment
17: 551–58. of chronic idiopathic thrombocytopenic purpura. N Engl J Med 2007;
6 Rodeghiero F, Stasi R, Gernsheimer T, et al. Standardization of 357: 2237–47.
terminology, definitions and outcome criteria in immune 29 Saleh MN, Bussel JB, Cheng G, et al. Safety and efficacy of
thrombocytopenic purpura of adults and children: report from an eltrombopag for treatment of chronic immune thrombocytopenia:
international working group. Blood 2009; 113: 2386–93. results of the long-term, open-label EXTEND study. Blood 2013;
7 George JN, Woolf SH, Raskob GE, et al. Idiopathic thrombocytopenic 121: 537–45.
purpura: a practice guideline developed by explicit methods for the 30 Bussel JB, Saleh MN, Wong RSM, et al. Update on the safety and
American Society of Hematology. Blood 1996; 88: 3–40. efficacy of EXTENDed treatment with eltrombopag (EPAG) in
8 Neunert CE, Buchanan GR, Imbach P, et al. Bleeding adults with chronic immune thrombocytopenia (ITP). Blood 2013;
manifestations and management of children with persistent and 122: 2315 (abstr).
chronic immune thrombocytopenia: data from the Intercontinental 31 Bussel JB, Buchanan GR, Nugent DJ, et al. A randomized,
Cooperative ITP Study Group (ICIS). Blood 2013; 121: 4457–62. double-blind study of romiplostim to determine its safety and efficacy
9 Neunert C, Lim W, Crowther M, et al. The American Society of in children with immune thrombocytopenia. Blood 2011; 118: 28–36.
Hematology 2011 evidence-based practice guideline for immune 32 Bussel JB, Hsieh L, Buchanan GR, et al. Long-term use of the
thrombocytopenia. Blood 2011; 117: 4190–207. thrombopoietin-mimetic romiplostim in children with severe
10 Provan D, Stasi R, Newland AC, et al. International consensus chronic immune thrombocytopenia (ITP).
report on the investigation and management of primary immune Pediatr Blood Cancer 2015; 62: 208–13.
thrombocytopenia. Blood 2010; 115: 168–86.

10 [Link] Published online July 29, 2015 [Link]