Effectiveness and Safety of Romiplostim Among Patients with Newly Diagnosed, Persistent and Chronic Immune Thrombocytopenia in European Clinical Practice

This article is published with digital features, including a summary slide, to facilitate understanding of the article. To view digital features for this article, go to https://​doi.​org/​10.​6084/​m9.​figshare.​14258513.

Primary immune thrombocytopenia (ITP) is an acquired immune-mediated disorder characterized by platelet counts less than 100 × 10⁹/L and increased risk of bleeding [1]. Primary ITP is relatively rare, with an annual incidence of 3.3/100,000 and prevalence of 9.5/100,000 adults [2]. Romiplostim was approved by the European Medicines Agency (EMA) on 4 February 2009 for the treatment of thrombocytopenia in adult patients with chronic ITP (currently defined as ITP lasting a duration of greater than 12 months) with insufficient response to prior therapy [3]. Romiplostim was approved by the United States Food and Drug Administration in 2008, and, in 2019, the label was revised to include all patients who have had an insufficient response to previous ITP treatments, regardless of ITP duration when the treatment is initiated [4].

Evidence from randomized clinical trials suggests romiplostim is effective at raising platelet levels among patients with chronic ITP as well as those in earlier phases (ITP ≤ 12 months) [5]. For example, in an integrated analysis of nine studies involving 911 ITP patients treated with romiplostim, 53% of those with newly diagnosed (ITP < 3 months) or persistent (ITP 3–12 months), and 49% of those with chronic ITP achieved durable platelet response within 24 weeks of initiating romiplostim. In real-world clinical practice, estimates of romiplostim treatment effect have been examined overall [6]; however, such estimates of romiplostim treatment effect among patients by duration of ITP have not previously been described. The objective of this study was to describe patient characteristics, treatment patterns, and ITP-related health and safety outcomes in romiplostim users, both overall and stratified by duration of ITP at initiation of romiplostim.

Overall, 340 patients were enrolled in the study, representing adults with primary ITP who received at least one dose of romiplostim [6]. Four patients died during the study and one patient withdrew consent during the 24-week follow-up. Sixty-four (19%) patients had newly diagnosed ITP, 50 (15%) had persistent ITP and 226 (66%) had chronic ITP at romiplostim initiation (Table 1). Patients with newly diagnosed ITP were older than patients with persistent or chronic ITP (median age = 67, 59, 60 years, respectively). Median baseline platelet counts were similar across strata, ranging from 17 to 21 × 10⁹/L. A higher proportion of patients with newly diagnosed ITP at romiplostim initiation had a history of bleeding in the 6 months prior to initiation (75% compared to 50% and 39% for persistent and chronic ITP, respectively). These patients also more frequently had active bleeding at romiplostim initiation [25% compared to 8% (persistent) and 12% (chronic)].

Regarding the nature and extent of ITP treatments prior to romiplostim initiation, most patients (more than 60%) had previous exposure to oral steroids or IVIg, regardless of the duration of ITP at romiplostim initiation (Table 1). Forty-five percent of newly diagnosed patients (n = 29), 60% of persistent (n = 30), and 76% of chronic ITP patients (n = 172) were considered refractory at romiplostim initiation (Table S5), and all analyses were repeated using only refractory patients (Tables S6–S9). Prior IV steroid administration and receipt of a platelet transfusion at romiplostim initiation was highest in patients with newly diagnosed ITP (48% and 36%, respectively), and was lowest in those with chronic ITP (27% and 12%, respectively). Conversely, previous receipt of splenectomy was most frequent in patients with chronic ITP: 47% of chronic ITP patients were splenectomized versus 8% of newly diagnosed or persistent ITP patients initiating romiplostim. Prior rituximab use was also relatively common, particularly in patients with persistent (48%) or chronic (44%) ITP at romiplostim initiation. Overall, the number of ITP treatments received prior to romiplostim initiation increased with increasing duration of ITP. However, concomitant treatments at romiplostim initiation was highest in those with newly diagnosed ITP at romiplostim initiation, namely with respect to oral steroids, IVIg, IV steroids, and platelet transfusions (Table 2).

In this cohort of adults with primary ITP initiating romiplostim in European clinical practice, one-third initiated romiplostim treatment within 12 months following their ITP diagnosis. This is the first study to evaluate real-world patient characteristics, romiplostim dosing, and outcomes stratified by duration of ITP at romiplostim initiation (newly diagnosed, persistent, chronic). Generally, duration of ITP did not appear to affect dosing or response to romiplostim, but newly diagnosed patients in this study had markers of more severe disease at romiplostim initiation, including a higher prevalence of prior and active bleeding and lower platelet counts compared with those with persistent or chronic ITP at romiplostim initiation. In these patients, it is possible that romiplostim was used as an emergency therapy to control bleeding. While the majority of this group had a starting dose of 1 µ/kg, a higher dose of ≥ 3 µ/kg was not uncommon, and was more frequent than seen in persistent or chronic ITP patients. However, newly diagnosed ITP patients had the lowest cumulative dose over the 24 weeks of follow-up, reflecting a higher incidence of discontinuation in this group. Despite these differences in severity of ITP at romiplostim initiation and romiplostim discontinuation rates during follow-up, platelet responses were similar across strata of ITP duration, and there were few ADRs.

The safety and effectiveness of romiplostim that we observed in routine clinical practice in Europe is generally consistent with other real-world data studies. Specifically, Lozano et al. [14], found that adverse events were rare in 54 patients using romiplostim in Spain, and Mihaylov et al. [15] found that median platelet count increased in response to romiplostim in 100 patients in Central and Eastern Europe within the first 6 months of treatment. At 24 weeks post-initiation of romiplostim in our study, the weighted median platelet count was similar across ITP duration categories (88–131 × 10⁹/L). These values are consistent with the median platelet count observed by ITP duration in an integrated analysis of nine clinical trials comparing romiplostim to standard of care or placebo [5]. Among the trial participants, patients who initiated romiplostim with newly diagnosed or persistent primary ITP had median platelet counts ranging from 75 to 120 × 10⁹/L during weeks 2–24 after initiation, while patients with chronic ITP at initiation had counts from 75 to 90 × 10⁹/L [5].

After romiplostim initiation, a durable platelet response was achieved in 32%, 53%, and 49% of patients with newly diagnosed, persistent, and chronic ITP, respectively. The response in chronic ITP patients appears comparable to that measured in chronic patients in the aforementioned integrated clinical trial data (49%), which employed a similar definition of durable platelet response (platelet response for 6 weeks or more during weeks 17–24 after romiplostim initiation) [5]. Among newly diagnosed and persistent ITP patients combined, the integrated analysis demonstrated a durable platelet response of 53%. While our estimate of durable platelet response in persistent patients aligns with this, our estimate in newly diagnosed patients is lower. Given the combined nature of the estimate from the clinical trial data, it is difficult to contextualize our stratified estimates and comment on possible reasons for any differences. One explanation for the lower durable response rate for patients newly diagnosed with ITP could be that newly diagnosed patients in our study had a lower weekly romiplostim dose administered (median weekly dose: 2.4 μg/kg compared with 3.0 μg/kg in the trial data). Additionally, 25% of newly diagnosed patients in our study discontinued romiplostim by 24 weeks, and the majority of these patients discontinued within the first 6 weeks (Figure S2). In contrast, the median duration of exposure among patients in the integrated analysis of trial data was 52 weeks and there were important differences in how the endpoint was operationalized. In the integrated analysis of trial data, missing platelet counts were imputed if the most immediate prior and following measurements were ≥ 50 × 10⁹/L, and patients who received rescue therapy could contribute to the endpoint after application of an 8-week wash-out period. In our study, patients were censored for missing platelets during specific weeks of follow-up, and we considered rescue therapy as an endpoint failure. These reasons may help explain any differences in the proportions of patients who achieved a durable platelet response in the real-world versus clinical trial setting.

A significant proportion (45%) of patients with newly diagnosed ITP at romiplostim initiation received rescue therapy during follow-up, and most use occurred within the first month following romiplostim treatment initiation. This could reflect the more severe nature of ITP in these patients at romiplostim initiation (e.g., more active bleeding) compared with those with persistent or chronic ITP. While the integrated analysis of clinical trial data also reported a high proportion of rescue therapy among patients with ITP for 1 year or less (44%), our estimates for use among patients with chronic ITP was lower (20.4% vs. 50% in the trials). It should be noted, however, that we were unable to capture any changes in dose or frequency of a specific treatment in our rescue therapy definition, which could explain why some of the estimates in the clinical trial setting were higher.

In our study population, splenectomy occurred relatively infrequently following romiplostim initiation and patients with persistent ITP at initiation were the most likely to undergo this procedure during the 24 weeks of follow-up (13%). Use of splenectomy in those with either newly diagnosed or chronic ITP at romiplostim initiation was just 2%. Splenectomy has historically been considered the second-line therapy of choice; however, rates have declined in recent years partially due to increased availability of medical options, including thrombopoietin-receptor agonists, romiplostim, and eltrombopag [2, 16, 17]. Notably, 47% of chronic ITP patients initiating romiplostim were already splenectomized. Given the high proportion of splenectomized patients in the chronic ITP group at romiplostim initiation, and that splenectomy is typically reserved for patients who have had ITP for a certain duration (to allow for potential spontaneous remission) and/or have already received medical alternatives [7, 18], these statistics are not unexpected. Infrequent use of splenectomy following romiplostim initiation has also been observed in clinical practice in the United States [19].

Nearly 75% of patients with newly diagnosed ITP at romiplostim discontinued all non-romiplostim ITP treatments during weeks 14–24 post-initiation, and over half of patients with persistent or chronic ITP at romiplostim initiation also did so. Given the toxicities associated with many first-line ITP treatment, this is an important finding. For example, in our cohort, over one-third of patients started romiplostim with concomitant oral steroids, regardless of ITP duration. Clinical decision to administer romiplostim early in the course of ITP may have been due to the severity of the presenting symptoms and the desire to limit steroids due to side effects [18]. Known side effects of corticosteroid use includes hypertension, sleep and mood disturbances, gastric irritation or ulcer formation, glaucoma, myopathy, and osteoporosis [20, 21], while previous and current clinical guidelines recommend against prolonged use [18, 22].

The incidence of ADRs during follow-up did not vary appreciably by ITP duration at romiplostim initiation: ADRs of any type occurred in 11–17% of patients, thrombotic events were rare, and there were no instances of bone marrow fibrosis. In the clinical trial setting, Kuter et al. (2019) found that serious adverse event rates were similar between newly diagnosed and chronic patients [5]. They considered any event that required or prolonged hospitalization as an adverse event, such as bleeding, dehydration, and pyrexia, in addition to any life-threatening condition.

This study has several important limitations. First, this was a descriptive study: no statistical comparisons of outcomes across categories of ITP duration were conducted. It is clear that several baseline characteristics differed by ITP duration, some of which were likely also associated with some of the outcomes. This apparent confounding was not addressed in this descriptive analysis. Second, the data were collected soon after romiplostim was approved in Europe for patients with chronic ITP and before the indication was expanded in Europe to include non-splenectomized patients in 2015 [3]. Therefore, usage patterns and patient characteristics may have changed since the time of data collection. Despite the risk of channeling bias in our data [23], i.e., that patients with worse prognoses may be more likely to be treated with romiplostim, we observed improved effectiveness outcomes. Further, our findings should be generalizable to patients managed in routine care in the European countries participating in this study, but patterns of use of romiplostim may be different in other countries with different prescribing and reimbursement practices. Third, as medical chart review was the primary source of data for this study, some clinical data had limited availability. For example, patients with chronic ITP may have had less complete histories of covariates if their ITP diagnosis predated their available medical record. To limit differential measurement error, we limited the lookback period for important variables (e.g., hospitalized bleeding 6 months before index, liver event occurring in the 5 years before romiplostim initiation). Additionally, the bleeding events reported here likely represent bleeding episodes severe enough to warrant medical care. Chart reviews likely do not capture less severe bleeding, which may not have required medical attention but are important from the patient perspective [24]. Finally, patients with ITP for more than 12 months at romiplostim initiation may represent those with less severe ITP and may be more likely to respond to therapy [25], which could impact future comparative studies. We examined this directly by stratifying by duration of ITP and examining distributions of health and medication variables.

This study builds upon prior real-world evidence of romiplostim use in Europe [6, 14, 15] and prior clinical trial evidence [5] to support the effectiveness and safety in newly diagnosed and persistent ITP patients. Regardless of duration of ITP at romiplostim initiation, in our cohort of adult primary ITP patients being managed in routine European clinical practice, romiplostim effectively raised platelet counts and, importantly, allowed for the discontinuation of other types of ITP treatments not intended for prolonged use. Additionally, the safety profile of romiplostim was consistent across categories of ITP duration and with that demonstrated in the clinical trial setting.