Individualized Treatment for ImmuneThrombocytopenia: Predicting Bleeding Risk

Cindy E. Neunert

Treatment of patients with immune thrombocytopenia (ITP) is often directed at increasing the

0037-1963/& 2013 Puhttp://dx.do

Publicationoperative

Conflicts o

Assistant PCenter, G

Address coHealth S30912. E

Seminars

platelet count and preventing significant hemorrhage even when there is minimal bleeding

present. This approach, however, requires that a large number of patients receive prophylactic

treatment to prevent major bleeding events. Identification of initial risk factors for developmentof severe bleeding would allow for more directed and personalized therapy. This review

provides a summary of the current literature with the intent to explore various clinical and

laboratory risk factors for severe bleeding including mucosal bleeding, platelet count, andaspects of platelet function.

Semin Hematol 50:S55–S57. C 2013 Published by Elsevier Inc.

Recently published immune thrombocytope-

nia (ITP) guidelines recommend that chil-dren with only skin manifestations at

diagnosis can be managed with observation alone1;

however, given concerns about the risk of bleedingin a child with severe thrombocytopenia, many

physicians still opt for treatment with drug therapy.

The ultimate goal of such prophylactic treatment isto promptly raise the platelet count and reduce the

perceived risk of subsequent significant bleeding.

Yet, there is a paucity of data to support the use ofdrug therapy in this setting, so the decision to treat is

likely affected by personal experiences and attitudes

about the number needed to treat to prevent onebleeding event.2 More refined risk stratification

models would aid physicians in identifying the

patients who would most benefit from drug therapy,thus reducing the number needed to treat, and

offering more personalized treatment.

Review of cases of intracranial hemorrhage (ICH)reported in the literature may provide identification

of common traits among patients. Table 1 shows

pooled data from four case series and one case-control study of ICH.3–7 The majority of patients had

severe thrombocytopenia, over half had mucosal

$ - see front matterblished by Elsevier Inc.i.org/10.1053/j.seminhematol.2013.03.009

of this article was supported by the International Co-ITP Study Group (ICIS).

f interest: none.

rofessor, Department of Pediatrics and Member, Cancereorgia Regents University, Augusta, GA.

rrespondence to Cindy E. Neunert, MD, MSCS, Georgiaciences University, 1120 15th St, BG-2011, Augusta, GA-mail: cneunert@georgiahealth.edu

in Hematology, Vol 50, No 1, Suppl 1, January 2013, p

bleeding, and approximately one third had acute

disease. Additionally, 20% of events were associatedwith head trauma. No characteristics either at diag-

nosis or preceding ICH were shown to be inclusive

of all patients. Interpretation of these results islimited because of the variable data collected across

studies. In addition, these reports do not address

additional forms of severe hemorrhage beyond ICHand most do not comment on bleeding prior to the

development of ICH but rather only at the time of

the event. Most importantly, only one study containsa control group of patients without ICH for

comparison.4

Despite having limitations these case series pro-vide useful evaluation of possible clinical predictors

of severe hemorrhage. For example, mucosal bleed-

ing, ‘‘wet purpura’’ or oral mucosal bleeding have

been presumed to place children at greater risk for

ICH and have historically guided treatment.8 How-

ever, there are no prospective data to suggest this

and the existing retrospective data from the ICH

cases are inconclusive. In the case-control series by

Psaila et al, findings of ‘‘wet purpura’’ were not

clearly related to ICH (27.5% of cases, 19% of

controls).4 The authors did, however find hematuria

to be predictive of ICH, with nine of the 40 patients

with ICH also having gross and microscopic hema-

turia, compared to none in the control group

(P o.001). A limitation of these data is that screen-

ing for microscopic hematuria in the control group

was not known. Additionally, the investigators found

that head trauma occurred with greater frequency in

ICH cases. It is not surprising that head trauma was a

risk factor for ICH; however, it may also represent a

p S55–S57 S55

Table 1. Pooled Data on Risk Factors forIntracranial Hemorrhage3–7

Risk FactorNo. of Cases(N ¼ 153)*

Platelet count o20 x 109/L 118/127 (93%)Platelet count o10 x109/L 87/127 (68.5%)Additional mucosal bleeding 40/61 (65.6%)At diagnosis 10/17 (60%)Hematuria 20/105 (19%)Acute disease 99/147 (67.3%)Head trauma 29/136 (21.3%)*There were a total of 153 cases of ICH reported. Thedenominator represents the number of cases with the specificrisk factor reported.

C.E. NeunertS56

recall bias in questioning and trauma is not suitablefor inclusion in a risk stratification model applied at

diagnosis based on its unpredictability.

Beyond clinical findings, laboratory assessmentmay also help to explain bleeding variability. Tradi-

tionally, only the platelet count has been used to

characterize patient risk, with a cutoff of o20 � 109/L generally defining patients with severe disease.

Applying International Cooperative ITP Study Group

(ICIS) Registry II data, a platelet count o20 � 109/Lhad a sensitivity of 88% and specificity of 21% in

differentiating bleeding severity.9 This finding has

been supported by Page et al, who showed thatplatelet count was poorly correlated with bleeding

severity at a platelet count o30 � 109/L.10 While

severe thrombocytopenia appears necessary, as illus-trated in Table 1, it is not sufficient for the develop-

ment of severe hemorrhage.

It has been hypothesized that antibodies directedat sites essential for platelet activation and adhesion

may explain the heterogeneity of bleeding symp-

toms. Unfortunately, lack of evidence leaves thismatter unsettled, so the role of anti-platelet anti-

bodies in inducing abnormal platelet function in

patients with ITP is not well established.11–13 Fewinvestigations have explored the link between plate-

let function, anti-platelet antibodies, and the clinical

outcome of bleeding severity. Panzer et al found thatneither platelet function, based on platelet p-selectin

levels, nor positivity for anti-platelet antibodies was

correlated with bleeding severity (P 4.05).12 Addi-tionally platelet function assessment using the cone

and platelet analyzer was performed and surface

coverage inversely correlated with bleeding severity(odds ratio [OR] 0.45; 95% confidence interval [CI],

0.17–0.87; P ¼ .04). They concluded that the

majority of patients with ITP have highly activatedplatelets that are able to form platelet aggregates

compensating for reduced platelet numbers. An

inability for platelet activation to occur in a select

group of patients with ITP may be partially respon-

sible for development of severe hemorrhage. Micro-

particles, procoagulant vesicles released fromplatelets, may also play a role in maintaining hemo-

stasis in ITP and explain why the majority of patients

with ITP can endure very low platelet counts with-out significant hemorrhage. Results from two studies

propose that patients with higher platelet micro-

particle levels experienced less bleeding.14,15 Moreconclusive studies of laboratory risk factors may

be difficult to carry out because of the rarity of

severe bleeding, difficulty in assessing plateletfunction in patients with thrombocytopenia, and

the poor sensitivity, specificity, and variability of

such testing.The role of other components of hemostasis,

including vascular integrity and levels of coagulation

factors, on the risk for severe bleeding in ITP remainunderstudied. Recent work in a thrombocytopenic

mouse model clearly demonstrated that micedevelop a greater degree of hemorrhage at sites

where significant inflammation was induced.16 This

supports the theory that platelets are essential inmaintaining vascular integrity and that inflammation

predisposes to increased hemorrhage. This work has

been directly translated to patients showing thinningof the endothelium in patients with thrombocytope-

nia and cessation of bleeding prior to an increase in

platelet count related to closing of endothelial gapjunctions.17 Another possible source of bleeding in

patients with ITP could be a coexisting hemostatic

alteration that had otherwise been clinically silent. Amajor determinant of primary hemostasis, in addition

to the platelet count, is the amount and functional

activity of von Willebrand factor. This adhesiveglycoprotein anchors platelets to the injured vessel

wall. It is possible that a ‘‘two-hit’’ process results in

clinical bleeding with ITP in those with low vonWillebrand factor activity.

As outlined above, several factors appear to be

permissive but not sufficient for development ofsignificant hemorrhage, including severe thrombocy-

topenia, additional mucosal bleeding, and head

trauma. Therefore, ideally several factors would needto be included in further studies of risk prediction.

When assessing the value of a prediction model,

thresholds for sensitivity, specificity, and positiveand negative predictive value need to be well

established to be clinically useful and adapted by

physicians. For example, one study combining riskfactors of bleeding beyond skin, and/or head trauma,

and/or a platelet count o10 � 109/L yielded a

sensitivity of 100% but specificity of only 23%.4 Inthis model, a significant number of children would

need to be treated to prevent an episode of ICH. The

authors then refined the predictors to the morespecific markers of hematuria and/or head trauma

Individualized treatment for ITP S57

identified by comparing cases and controls. While

the specificity increased to 99% it was at the expense

of a sensitivity of 50%.4 In this case, fewer patientsreceive prophylactic therapy unnecessarily; how-

ever, a few cases at high risk of developing ICH

would not be identified early nor appropriatelytreated. Physician experience may dictate acceptable

thresholds for sensitivity and specificity regardless of

the evidence, so uniform agreement on the numberneeded to treat may be difficult to achieve.2

In conclusion, in it is unlikely that a single

predictive factor will determine patients with ITPat risk for severe bleeding. Rather, risk of major

hemorrhage is likely multifactorial and different for

each patient, eg, in some patients severe hemor-rhage is related to an unpredictable traumatic event

regardless of additional factors that may be present.

To better define a prediction model for risk factors ofsevere hemorrhage, prospective cohorts should

focus on inconclusive assessment of all risk factorsat diagnosis and provide detailed bleeding assess-

ment using rigorous bleeding scales at all visits.

While the addition of comprehensive laboratorytesting maybe useful, refinement of such testing

and its application to clinical practice makes it

limited at present. Application of uniform treatmentwith observation for patients with no or mild bleed-

ing and severe thrombocytopenia will expand our

knowledge about those rare children who subse-quently develop significant hemorrhage.

REFERENCES1. Neunert C, Lim W, Crowther M, et al. The American

Society of Hematology 2011 evidence-based prac-

tice guideline for immune thrombocytopenia. Blood.

2011;117:4190–207.

2. Lehmann HP, Dambita N, Buchanan GR, et al. Decision

modeling of disagreements: pediatric hematologists’

management of idiopathic thrombocytopenic purpura.

Med Decis Making. 2011;31:805–15.

3. Butros LJ, Bussel JB. Intracranial hemorrhage in

immune thrombocytopenic purpura: a retrospective

analysis. J Pediatr Hematol Oncol. 2003;25(8):660–4.

4. Psaila B, Petrovic A, Page LK, et al. Intracranial hemor-

rhage (ICH) in children with immune thrombocytopenia

(ITP): study of 40 cases. Blood. 2009;114(23):4777–83.

5. Elalfy M, Elbarbary N, Khaddah N, et al. Intracranial

hemorrhage in acute and chronic childhood immune

thrombocytopenic purpura over a ten-year period:

an Egyptian multicenter study. Acta Haematol.

2010;123(1):59–63.

6. Choudhary DR, Naithani R, Mahapatra M, et al. Intra-

cranial hemorrhage in childhood immune thrombocy-

topenic purpura. Pediatr Blood Cancer. 2009;52(4):

529–31.

7. Iyori H, Bessho F, Ookawa H, et al. Intracranial

hemorrhage in children with immune thrombocyto-

penic purpura. Japanese Study Group on childhood

ITP. Ann Hematol. 2000;79(12):691–5.

8. Vesely S, Buchanan GR, Cohen A, et al. Self-reported

diagnostic and management strategies in childhood

idiopathic thrombocytopenic purpura: results of a

survey of practicing pediatric hematology/oncology

specialists. J Pediatr Hematol Oncol. 2000;22(1):55–61.

9. Neunert CE, Buchanan GR, Imbach P, et al. Severe

hemorrhage in children with newly diagnosed

immune thrombocytopenic purpura. Blood. 2008;

112(10):4003–8.

10. Page LK, Psaila B, Provan D, et al. The immune

thrombocytopenic purpura (ITP) bleeding score:

assessment of bleeding in patients with ITP. Br J

Haematol. 2007;138(2):245–8.

11. Olsson A, Andersson PO, Tengborn L, et al. Serum

from patients with chronic idiopathic thrombocyto-

penic purpura frequently affect the platelet function.

Thromb Res. 2002;107(3-4):135–9.

12. Panzer S, Rieger M, Vormittag R, et al. Platelet function

to estimate the bleeding risk in autoimmune thrombo-

cytopenia. Eur J Clin Invest. 2007;37(10):814–9.

13. Chu XX, Hou M, Peng J, et al. Effects of IgG and its

F(ab’)2 fragments of some patients with idiopathic

thrombocytopenic purpura on platelet aggregation.

Eur J Haematol. 2006;76(2):153–9.

14. Tantawy AA, Matter RM, Hamed AA, et al. Platelet

microparticles in immune thrombocytopenic purpura

in pediatrics. Pediatr Hematol Oncol. 2010;27(4):

283–96.

15. Jy W, Horstman LL, Arce M, et al. Clinical significance

of platelet microparticles in autoimmune thrombocy-

topenias. J Lab Clin Med. 1992;119(4):334–45.

16. Goerge T, Ho-Tin-Noe B, Carbo C, et al. Inflammation

induces hemorrhage in thrombocytopenia. Blood.

2008;111(10):4958–64.

17. Kitchens CS, Pendergast JF. Human thrombocytopenia

is associated with structural abnormalities of the

endothelium that are ameliorated by glucocorticoste-

roid administration. Blood. 1986;67(1):203–6.