Rituximab for Children with Immune Thrombocytopenia: A Systematic Review

Background Rituximab has been widely used off-label as a second line treatment for children with immune thrombocytopenia (ITP). However, its role in the management of pediatric ITP requires clarification. To understand and interpret the available evidence, we conducted a systematic review to assess the efficacy and safety of rituximab for children with ITP. Methodology/Principal Findings We searched MEDLINE, EMBASE, Cochrane Library, CBM, CNKI, abstract databases of American Society of Hematology, American Society of Clinical Oncology and Pediatric Academic Society. Clinical studies published in full text or abstract only in any language that met predefined inclusion criteria were eligible. Efficacy analysis was restricted to studies enrolling 5 or more patients. Safety was evaluated from all studies that reported data of toxicity. 14 studies (323 patients) were included for efficacy assessment in children with primary ITP. The pooled complete response (platelet count ≥100×109/L) and response (platelet count ≥30×109/L) rate after rituximab treatment were 39% (95% CI, 30% to 49%) and 68% (95%CI, 58% to 77%), respectively, with median response duration of 12.8 month. 4 studies (29 patients) were included for efficacy assessment in children with secondary ITP. 11 (64.7%) of 17 patients associated with Evans syndrome achieved response. All 6 patients with systemic lupus erythematosus associated ITP and all 6 patients with autoimmune lymphoproliferative syndrome associated ITP achieved response. 91 patients experienced 108 adverse events associated with rituximab, among that, 91 (84.3%) were mild to moderate, and no death was reported. Conclusions/Significance Randomized controlled studies on effect of rituximab for children with ITP are urgently needed, although a series of uncontrolled studies found that rituximab resulted in a good platelet count response both in children with primary and children secondary ITP. Most adverse events associated with rituximab were mild to moderate, and no death was reported.


Introduction
Immune thrombocytopenia (ITP) is an immune-mediated disease characterized by transient or persistent decrease of the platelet count and increased risk of bleeding [1]. The incidence of ITP is 1.9,6.4 per 10 5 children/year [2] and 23.1%,47.3% of children with ITP suffer a disease course more than 6 months [3].
The goal of ITP treatment is to achieve a platelet count that is associated with adequate hemostasis, rather than a ''normal'' platelet count. The recommended first-line drug treatment for children with ITP includes corticosteroids, intravenous immunoglobulin (IVIg) and Anti-D immunoglobulin [4], but for pediatric patients refractory to first line treatment, appropriate second line treatments are needed.
Rituximab is a chimeric, monoclonal anti-CD20 antibody that targets B lymphocytes and causes Fc-mediated cell lysis [5,6]. It was initially approved for the treatment of lymphoma, as it can substantially decrease the normal and malignant B-cells [7]. It has also been approved for the treatment of rheumatoid arthritis in Europe, as it can destroy B lymphocytes in the joints to help reducing inflammation [8]. In recent years, it has been widely used off-label as the second line treatment in both adults and children with ITP refractory to first line treatment. A systematic review based on studies on adult patients with ITP found that rituximab resulted in overall response (platelet count .50610 9 /L) and complete response (platelet count .150610 9 /L) in 62.5% and 43.6% of patients, respectively [9]. Several studies on children suggested that rituximab showed a similar effect as adult ITP, but the reported response rate varied among those studies [10]. In addition, those results may be potentially biased and imprecise, as most studies are case series, and only involved a relatively small number of patients. The role of rituximab in the management of pediatric ITP still requires clarification. To understand and interpret the available evidence, we conducted a systematic review to evaluate the efficacy and safety of rituximab for children with ITP.

Searching
We searched PUBMED, EMBASE, Cochrane Central Register of Controlled Trials (CENTRAL) published in Cochran Library (2012, Issue 1) using the search strategy detailed in table S1; we searched Chinese Biomedical Literature Database (CBM) and Chinese National Knowledge Infrastructure (CNKI) for literatures published in Chinese. We also searched the electronic databases of American Society of Hematology, American Society of Clinical Oncology and Abstract Database of Pediatric Academic Society including its Late-Breaker Abstract Presentations,with the search term ritux*, child* and pediatr*. The references of all retrieved articles were scanned for additional relevant citations. We searched all databases from their earliest records to January 2012.

Eligibility Criteria
Clinical studies published in full text or abstract only were both included, and there was no restriction on study design or publication language. We included children aged less than 18 years with primary or secondary ITP, having platelet counts less than 30610 9 cell/L. Studies that included both children and adults were excluded if data of children could not be extracted separately. Patients were treated with rituximab irrespective of dosage and schedule. Outcome criteria based on guidance recently provided by international working group (IWG) consensus panel of both adult and pediatric experts [1]. Response (R) was defined as any platelet count $30610 9 /L and at least doubling of the baseline count. Complete response (CR) was defined as any platelet count $100610 9 /L. Time to response was from starting treatment to achievement of response. Duration of response was measured from the achievement to loss of response. All reported adverse events were included for safety assessment. Studies that enrolled fewer than 5 patients but failed to contribute to safety analysis were excluded.

Study Selection and Data Abstraction
Two authors (Yi Liang and Die Hu) independently screened titles and abstracts of all studies identified by the search strategy and assessed the studies for inclusion using predetermined inclusion criteria. The full texts of all potentially relevant articles were retrieved for detailed review. We resolved disagreements by discussion until consensus is achieved.
Two authors (Yi Liang and Die Hu) used a pre-designed data collection form to independently extract data from each included study. The following data were extracted: (1) characteristics of patients, including number of patients that meet the inclusion criteria, country, age, primary/secondary ITP, duration of ITP, splenectomized or not, and platelet count before rituximab treatment; (2) study design and use of controls; (3) dose and schedule of rituximab; (4) number of patients with platelet count response, complete response, and their definitions; (5) time to platelet count responses; (6) follow-up and duration of platelet count responses; (7)toxicities associated with rituximab; (8) source of funding.

Quality Assessment
Quality assessment was based on the checklist developed for assessment of case series by UK national institute for clinical excellence (NICE) [11]. We answered 8 questions included in the checklist with yes, no, or unclear. Quality assessment was only conducted in studies enrolling 5 or more patients, as those studies contributed to efficacy analysis.

Statistical Synthesis
Platelet count responses were analyzed only from those studies enrolling 5 or more patients, as smaller studies may be subject to extreme risk of reporting and selection bias. Adverse effects were considered from all studies, including those enrolling fewer than 5 patients each. We achieved exact binomial 95% confidence interval (CI) of response rate reported in each study. We estimated the between-study variance and determined pooled estimates of response rate using the software STATA 11.1 with a randomeffect model [12]. Time to response and response duration were described with medians, minimum and maximum values, and for studies that reported individual level data, we combined and summarized these data with medians and inter quartile ranges (IQR). We assessed publication bias both with Egger and Begg test in STATA 11.1, and P , = 0.05 suggested a significant publication bias [13,14].
We calculated a kappa statistic for measuring agreement between two authors making decisions on study selection. Values of kappa between 0.40 and 0.59 were considered to reflect fair agreement, between 0.60 and 0.74 to reflect good agreement and 0.75 or more to reflect excellent agreement [15]. The design and report of this review has been checked with PRISMA checklist (see table S2).  Figure 1 shows the literature selection process. Agreement between 2 reviewers for study selection was excellent (K = 0.75).
Secondary ITP. 7 patients with Evans syndrome associated ITP achieved response with time to response from 0.8 to 8 weeks, and none relapsed after 8 to 20 months follow-up [19]. 6 patients with systemic lupus erythematosus associated ITP achieved response with time to response from 1 to 12 weeks, among that, 2 patients relapsed at 12 and 17 months, but both achieved complete response again after a second course of rituximab, other patients had ongoing response from 6 to 22 months [22]. 2 of 6 patients with autoimmune lymphoproliferative syndrome associated ITP relapsed at 15 and 18 months, the other 4 patients had ongoing response from 5 to 36 months, but data of time to response was not reported [24].

Predictors of Response to Rituximab
4 studies [18,23,28,29] (175 patients) analyzed clinical characteristics associated with platelet count response to rituximab, however, their results varied. Bennett 2005 [18] (36 patients) found that attainment of response was weakly associated with diagnosis of Evans syndrome (P = 0.06), female sex (P = 0.14) and black race (P = 0.09), but the reported P value did not show significance association (P.0.05). Parodi 2009 [23] (49 patients) founded that median duration of ITP was significantly shorter in responders than in non-responders (P = 0.01). Grace 2012 [28], which was the largest (80 patients), showed that secondary ITP and response to steroid were strong predictors of response to rituximab, and was both suggested by result of univariate and multivariable analysis. Sharma 2005 [29] reported that patients showing a higher degree of response continued to be in remission for a longer period compared to ones with lesser degree of response.
Toxicities 23 studies reported 108 adverse events in 91 patients. In 11 studies (190 patients) enrolling more than 5 patients, 78 (41.1%) patients experienced adverse events. Adverse events were classified from Grade 1 to 5 according to Common Terminology Criteria for Adverse Events (CTCAE) [46], and were described in table 3. In all reported 108 adverse events, 91 (84.3%) were mild to moderate, and the most frequently described adverse events are mild allergic reactions, including pruritus, urticaria, chills and fever. 7 patients developed serum sickness after 1 or 2 doses of rituximab, presented with fever, rash, arthralgia and fatigue, and in which, 3 cases were assigned to Grade 3-4, as 1 patient has an acute reaction followed by persistent arthralgia and 2 patients discontinued rituximab for serum sickness [17,18]. 2 patients experienced immediate hypersensitivity reaction during rituximab infusion, which caused termination of the treatment [20,45]. 4 patients developed infections that may be associated with rituximab, including 2 patients with varicella [18,34], 1 patient with pneumonia [19], and 1 patient with life-threatening enteroviral meningoencephalitis presented with a progressive alteration in cognitive functions associated with aphasia and sensorimotor deafness [36]. 1 patient developed common variable immunodeficiency which caused prolonged hypogammaglobulinemia and increased susceptibility to infections [41]. 1 patient developed headache with white matter changes on brain MRI, as only described in abstract, no further information was reported [38]. No death associated with rituximab was reported.

Publication Bias
The Egger and Begg test were conducted to investigate publication bias in 14 studies reporting response in patients with primary ITP. The Egger test indicated no evidence of publication bias (P = 0.987). Begg's funnel plot with pseudo 95% (figure 4) was graphed in the logic that studies with smaller sample size have larger random error, thus are more dispersed, and the plot of effect estimates against standard errors would be skewed and asymmetrical in the presence of publication bias. In this review, the funnel plot suggests no significant asymmetry, indicating no evidence of substantial publication bias (P = 0.583).

Discussion
This is the first systematic review that summarizes the efficacy and safety of rituximab in children with ITP, and this review is very important at this time as rituximab has been widely used off-label for children with ITP. The treatment of rituximab achieved response in 68% and complete response in 39% of patients with primary ITP. According to studies reporting individual data, the median time to response was 3.0 weeks and the median response duration was 12.8 months. Studies with a small number of patients with secondary ITP reported that response rate to rituximab was from 64.7%,100%. Most adverse events were mild to moderate and no death was reported.
The response and complete response rate in children with primary ITP obtained in this review was 68% and 39%, respectively. It was similar to that reported in the previous systematic review on adults, which reported that overall response (platelet count .50610 9 /L) and complete response (platelet count .150610 9/ L) rate was 63% and 46% respectively [9]. Although response criteria differ in these two reviews, it may suggest that rituximab results in a similar response in adults and children with primary ITP. The frequently administrated dose of rituximab is 375 mg/m 2 per week for 4 weeks for both children and adult with ITP [9], although this dose was developed and approved for the treatment of lymphoma [47]. 1 study investigated the efficacy of a single dose of rituximab (375 mg/m 2 ) in children, and reported a similar response and complete response rate as 4 doses [16]. 2 studies investigated a low dose rituximab of 100 mg/dose/week for 4 weeks in 18 children with primary ITP, which suggested that the lower dose of rituximab still reached a good response [26,27]. Treatment of rituximab is still costly in the present time, and thus sometimes 4 doses of 375 mg/m 2 are not affordable, especially for patients with poor accessibility to health resources. In addition, high dose and multiple dosing may increase risk of adverse events. Low dose of rituximab may be promising for patients with ITP, and more studies are needed to investigate that.
Most of adverse events associated with rituximab in children with ITP were mild to moderate infusional reactions. More severe adverse effects included serum sickness, common variable immunodeficiency, severe virus infection including enteroviral meningoencephalitis, and white matter changes. 2 children developed viral infection after rituximab treatment, but they were both previously treated with other immunosuppressive agents like steroids, vincristine, and cyclosporine A, it is difficult to establish a direct association with rituximab [18,34]. In addition, a systematic review in cancer patients did not find any increase of infection caused by monoclonal antibodies [48]. The case of common variable immunodeficiency in patients with ITP after rituximab treatment was reported in an 8-year-old child [41], however, it was not clear whether it was caused by rituximab, as ITP might be the first clinical manifestation before development of common variable immunodeficiency. 1 patient developed headache and MRI showing white matter changes [38], but without detailed information as published only in abstract, it is not clear whether the children developed progressive multifocal leukoencephalopathy, although it is reported that there was a potential of progressive multifocal leukoencephalopathy among rituximabtreated patients [49]. No death was found in this review, but 9 cases of death (2.9%) were reported in studies on adults [9]. The mortality reported in adults treated with rituximab might be overestimated and these cases might be explained by long courses of complex treatment regimens or the selection of patients with advanced disease.
All the studies eligible for efficacy analysis in this review failed to compare effect between rituximab with non-rituximab group. The response rate obtained may be influenced by many potential factors, such as treatments before and combined with rituximab, cause and development of ITP. Studies on this topic with better methodological design like randomized controlled studies are urgently needed. Consecutive enrolling of patients can help to reduce selection bias in prospective case series studies, unfortunately, few studies provided information on that, and this needs improvement in future studies.

Supporting Information
Table S1 Search strategy for PUBMED, EMBASE and CENTRAL.