Real World Romiplostim Dosing, Response, and Home Administration in Pediatric Immune Thrombocytopenia

Romiplostim, a weekly subcutaneously administered thrombopoietin receptor agonist (TPO-RA), is approved by the US Food and Drug Administration (FDA) for treatment of immune thrombocytopenia (ITP) in children ≥1 year of age. TPO-RAs increase platelet count by promoting growth and development of megakaryocytes and their precursors.¹ Guidelines recommend treatment with TPO-RAs in children with ITP with mucosal bleeding and/or impaired quality of life.² Recent studies demonstrate efficacy of TPO-RAs early in the disease course.^{3, 4}

The FDA approved prescribing information for romiplostim recommends an initial dose of 1 mcg/kg, with weekly increases of 1 mcg/kg if platelet counts remain <50 x10⁹ cells/uL. Many patients are initiated on higher doses of romiplostim with a goal of minimizing bleeding and time to platelet response.¹ It has been speculated that higher starting doses may be associated with increased risk of thrombotic complications, although randomized studies of romiplostim have not shown a higher risk of thrombosis in treated patients compared with placebo or a correlation between thrombosis and platelet count.^5–9 Studies of romiplostim in adults have reported increased efficacy with starting doses >1 mcg/kg.^10–13 Limited data are available regarding real-world romiplostim dosing and administration for ITP in children.^{13, 14}

This study aimed to describe starting doses, administration, early platelet response rates, and incidence of thrombosis in children with ITP treated with romiplostim in both clinical and home administration settings.

This observational cohort study included patients age 0-22 with primary or secondary ITP who received romiplostim between 2008 and 2023. Collected data included clinical characteristics, romiplostim prescribing practices, bleeding scores, and complications. Secondary ITP was defined as ITP occurring in Evans syndrome, inborn error of immunity, or post-transplant. Single platelet response was defined as platelet count > 50 x10⁹/L at least once in the absence of rescue therapy. Early platelet response was defined as platelet count >50 x10⁹/L within one week following initial romiplostim dose. Romiplostim dosing and discontinuation were determined by the treating clinician. Outcomes were summarized using descriptive statistics. Denominators reflect available data for each data point. Fisher’s exact tests were performed for categorical data and Mann-Whitney tests were performed for continuous variables. Statistical analysis was performed using GraphPad Prism version 10.0.0.

There were 58 children with ITP treated with romiplostim over the 15-year period, including patients with newly diagnosed (<3 months, 31%) persistent (3-12 months, 38%) and chronic (>12 months, 29%) ITP at time of romiplostim initiation. Patients were followed for a median of 5.8 years (range 0-17.3). There were 17 (29.3%) patients with secondary ITP who received romiplostim, including 14 (29.3%) with Evans syndrome. The median patient age at romiplostim initiation was 12.3 years (range 0.8-21.9) and 67% (39/58) were female. Patients received a median of 3 (range 1-6) ITP-directed therapies prior to romiplostim. Highest bleeding score prior to romiplostim was grade 1 (1.7%, n=1), grade 2 (6.9%, n=4), grade 3 low risk (13.8%, n=8), grade 3 high risk (55.2%, n=32), grade 4 (13.8%, n=8), grade 5 (3.4%, n=2) (Supplementary Table 1).

Median initial romiplostim dose was 3.6 mcg/kg (range 1-10) and median maximum romiplostim dose was 8 mcg/kg (range 1-10) (Figure 1A, Table 1). Most patients (84.5%, n=49/57) received an initial romiplostim dose >1 mcg/kg (Figure 1A). Initial romiplostim dose was ≥ 5 mcg/kg in 21% (12/57). The single platelet response rate was 81% (44/54). Early platelet response was achieved in 37% (19/52) (Table 1). Early response was seen at all disease phases (newly diagnosed (n=10), persistent (n=4), chronic (n=5)). Patients with an early platelet response were started on significantly higher initial doses than those without an early response (median 4.1 mcg/kg versus 3.0 mcg/kg, p=0.002) (Figure 1B). Patients with an initial romiplostim dose ≥3 mcg/kg had a higher likelihood of romiplostim response on their initial dose than those with initial dose <3 mcg/kg (initial response rate 52% versus 15%, p=0.0171) (Figure 1C). Among those with an initial romiplostim dose <3 mcg/kg, 95% (20/21) required a subsequent dose increase. The initial romiplostim dose was not significantly different between overall responders and non-responders.

Figure 1. Initial romiplostim dose and initial platelet response.

A. Distribution of initial romiplostim doses. B. Initial romiplostim dose in patients who did and did not achieve an early platelet response. C. Response to initial romiplostim dose ≥3 mcg/kg as compared to initial dose <3 mcg/kg.

Table 1.

Romiplostim dosing and response

	Median (range) or n (%)
Initial romiplostim dose (mcg/kg)	3.6 (1-10)
Maximum romiplostim dose (mcg/kg)	8 (1-10)
Single platelet response	44 (81%)
Initial platelet response on first romiplostim dose	19 (37%)
Romiplostim dose increase required	45 (80%)
Time to platelet response (days)	16 (5-91)
Length of romiplostim treatment (days)	105 (1-2150)

There was no difference in single platelet response rate by ITP phase (Supplementary Figure 1A), highest bleeding score (Supplementary Figure 1B), or primary versus secondary ITP (78% vs 94%, respectively, p=0.24). Among those whose ITP had no response to eltrombopag, 77% (10/13) had a single platelet response to romiplostim (Supplementary Table 1).

Romiplostim was held due to thrombocytosis in 22% (12/54), attributable to combined romiplostim and corticosteroids in 50% (6/12). Median platelet count at dose hold was 621 x10⁹ cells/uL (range 315-1199 x10⁹ cells/uL). One week following dose hold, 42% (5/12) had platelet counts <30 x10⁹ cells/uL, including 3 patients with platelet counts <10 x10⁹ cells/uL.

Romiplostim was administered at home in 33.3% (19/57) with median time of home administration of 8 months (range 1-56). Median romiplostim dose at time of first home administration was 4.75 mcg/kg (range 2-7 mcg/kg). In all cases, at least one dose was initially administered in the infusion center or clinic, and doses were adjusted during home administration. The single platelet response rate to romiplostim was the same with home and hospital administration (68.4% (13/19) vs 88.2% (30/34), p>0.05), with no safety issues identified in the home administration group.

The median duration of romiplostim treatment was 3.5 months. Romiplostim was tapered prior to discontinuation in 44% (21/48). Tapering regimens varied. Reasons for discontinuation included platelet count normalization (35.4%, 17/48), no response (18.8%, 9/48), intermittent response or waning effect (12.5%, 6/48), transitioned to other therapies (27.1%, 13/28), short pre-procedure course (4.2%, 2/48), and single dose on presentation (2.1%, 1/48).

Thrombosis occurred in 3.4% (n=2) and was attributable to causes other than romiplostim. One patient with an immunodeficiency was treated with romiplostim prior to stem cell transplant. Four months after discontinuing romiplostim, the patient developed graft versus host disease with a portal vein thrombus. The second patient underwent emergent splenectomy and developed portal vein thrombus on post-operative day 3. Platelet counts at time of thrombus were 68 and 259 x10⁹ cells/uL, respectively.

In this real-world cohort of pediatric patients with ITP with significant bleeding, most patients received initial romiplostim doses >1 mcg/kg. Patients who received initial romiplostim doses of ≥3 mcg/kg had higher likelihood of early platelet response. This finding differs from a similar study of adults with acute ITP, likely due to differences in clinical characteristics including phase of ITP and ages of treated patients.⁵ The majority of patients in this pediatric cohort had persistent or chronic ITP and had received multiple ITP-directed treatments prior to romiplostim. Published findings of a similar cohort in adults demonstrated that higher initial romiplostim dosing led to fewer bleeding events and improved rates of platelet response.¹⁵ In clinical practice, initial romiplostim doses of ≥3 mcg/kg should be considered to decrease duration of thrombocytopenia, bleeding risk, and ongoing bleeding.^{S16, S17} Thrombosis after romiplostim was rare (3.4%) and attributable to secondary causes.

Overall rate of romiplostim response in this cohort was high (81%). Notably, there was a high rate of romiplostim response (77%) even among children who did not have platelet response to eltrombopag, suggesting clinical utility of TPO-RA switching in pediatric patients. Romiplostim response occurred regardless of disease phase and in patients with primary and secondary ITP, demonstrating broad efficacy in children with ITP.

Romiplostim dosing was variable, with inconsistent approaches to dose tapers and holds. Prescribing recommendations are to hold romiplostim for platelet counts >400 x10⁹ cells/uL. Dose holds for high platelet count were not uncommon (22.2%) and associated with rebound thrombocytopenia and increased bleeding risk. These data support consideration of a dosing decrease rather than reflexive dose holds in response to thrombocytosis.

Romiplostim is not approved for in-home administration in the US. One-third of patients who initially received romiplostim in the clinical setting were able to receive romiplostim at home, including titration of doses during home administration. The efficacy and safety data in this cohort supports the feasibility of home administration in children. Administration at home alleviates the burden of travel and time at an infusion center, making treatment easier for patients and families, and eliminates the higher cost for infusion-related charges.

This study is limited by retrospective design and small sample size. Due to the retrospective nature of this study, the impact of romiplostim on quality of life and fatigue could not be evaluated, and information about adverse events or missed or inaccurate dosing with home administration is limited to data available in the medical record. Since some patients in the cohort were treated with short courses of romiplostim, durable platelet response could not be assessed. Future prospective study in larger cohorts is required to comprehensively address these questions.

These findings underscore the potential benefit of higher initial romiplostim dosing than is currently recommended in the FDA-approved prescribing information, particularly in patients with severe bleeding, and highlight the risk of rebound thrombocytopenia associated with holding doses. Furthermore, this study suggests that in-home romiplostim administration is feasible and could enhance treatment convenience with reduced costs. There is an ongoing need to standardize real-world dosing strategies to optimize the efficacy and safety of romiplostim for the treatment of ITP.

Supplementary Material

Supplementary materials

Supplementary Figure 1. Single platelet response to romiplostim by A. disease phase and B. bleeding score.

Data availability:

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.