Childhood ITP: knowing when to worry?

In this issue of Blood, Psaila and colleagues describe a large cohort of patients with ITP and intracranial hemorrhage.

Immune thrombocytopenia (ITP) is a fairly rare, generally benign illness in the pediatric population, affecting between 5 to 13 children per 100 000. Intracranial hemorrhage (ICH) is considered a rare complication occurring in a small subset of patients (< 1% in multiple studies). Because ICH is a rare but significant complication of ITP in childhood, when to consider treatment remains an issue. In this issue of Blood, Psaila et al describe a cohort of patients with ITP and ICH, examining risk factors for hemorrhage and attempting to identify an algorithm to stratify patients according to risk and allow for rational treatment decisions.¹ 

In this study, the authors examined medical information obtained through a survey that combined a retrospective and a prospective multi-institutional analyses. The calculated incidence of ICH was 0.19% to 0.78% of pediatric patients with ITP, depending on the estimate of the percentage of all bleeding episodes in this population that were captured by the survey. Controls were the next 2 patients presenting with ITP of similar duration and severity from the reporting institution and were supplemented with cases from Cornell/Weill if they were not locally available. As reported in other studies,^2,3  the majority of children had platelet counts less than 20 × 10⁹/mcL at the time of ICH, but 10% had counts higher than 20 × 10⁹/mcL. The authors also report that almost half (45%) of the episodes of ICH occurred in patients within the first week of diagnosis, whereas 30% occurred in patients with ITP of more than 6 months' duration, supporting older studies that the frequency of ICH is highest in the first week after diagnosis.⁴ 

Outcomes for these patients are also reported and whereas 8 of 10 patients whose ICH was a presenting feature of their ITP survived without sequelae, 50% of the patients who presented within the first week of diagnosis died and 38% had neurologic sequelae. Whether this good outcome for patients with ICH on presentation is a peculiarity of this study or has some biological meaning will require further studies with more detailed clinical information on treatment in relation to ICH and greater detail on the ICH.

Next, the authors examined risk factors for ICH to identify a population of patients who might be at higher risk of bleeding and therefore might benefit from more aggressive therapy. The patients who presented with ICH as the first manifestation of the ITP were excluded from treatment/cost analysis. As in previous studies,^2,5,6  head trauma and hemorrhagic manifestations (beyond mild cutaneous) were significant risk factors for ICH and, in the case of more bleeding manifestations, increased mortality. In contrast to other studies, the authors found no relationship between concurrent nonsteroidal anti-inflammatory drug use, wet purpura, and gastrointestinal or vaginal bleeding, but this may be due in part to the limitations of retrospective reporting in a survey study. The patients with ICH were also noted to have a lower incidence of petechiae.

The most interesting part of this paper was the presentation of algorithms stratifying patients with ITP according to risk for ICH. The algorithms were used to analyze the cost effectiveness of therapy (excluding in this analysis those patients who presented with ICH as the initial manifestation of ITP). According to this analysis, if all patients with any bleeding other than petechiae and ecchymosis and/or head trauma and/or platelet count less than 10 × 10⁹/mcL are treated, 327 children will be treated to prevent 1 ICH, whereas treating those with hematuria and/or head trauma (regardless of platelet count) resulted in treating 8.5 children to prevent 1 ICH. Using this last approach, in this cohort, none of the subjects with ICH would have been missed. If this finding is confirmed, and the benefit of treatment is confirmed, it may be possible to treat a small subgroup of children with ITP who are at highest risk for ICH, thereby minimizing risk to patients from overtreatment while reducing cost.

Although this paper has limitations because of the nature of the study (survey study with retrospective and prospective components), important information about ICH and ITP is presented. Besides presenting one of the largest cohorts of patients with ICH, this study shows that the incidence of ICH is low, but morbidity and mortality from ICH remain high. This study suggests that those patients who present with ICH during the first week of diagnosis do poorly, and further studies to understand why are needed. Perhaps more aggressive therapy for these patients would help, but perhaps there is something about the pathophysiology of ICH in these patients that makes them more likely to suffer neurologic sequelae (location of bleeding, duration prior to presentation, size of blood collection, etc). Could therapeutic intervention worsen the severity of these bleeds? Does intravenous immunoglobulin or Rho(D) immune globulin predispose a person to either ICH or worsen a subclinical ICH? Unfortunately, information regarding treatment and response in relation to the ICH was not collected in this study. However, studies have suggested an increased risk of thrombosis and stroke with intravenous immunoglobulin therapy for ITP and other indications.^7,8  Further studies about the pathophysiology of the ICH resulting in neurologic sequelae versus the ICH with complete recovery, and the temporal relationship of ICH and previous treatment are needed.

Patients with head trauma and hematuria (certainly) or other hemorrhage besides mild mucocutaneous bleeding (according to other studies and supported by the current study) are at increased risk for developing ICH (even if the platelet count is > 20 × 10⁹/mcL) and should be considered for treatment. Psaila et al have taken steps toward understanding ICH in ITP but further multi-institutional studies, not only about whether or not ICH has occurred but also about those hemorrhages and the response/relationship to therapy, are needed to answer these important questions.