Romiplostim Significantly Reduces Chemo Dose Delays, Modifications Due to Thrombocytopenia in GI Cancers

Weekly romiplostim (Nplate) significantly reduced chemotherapy dose modifications attributable to chemotherapy-induced thrombocytopenia (CIT) compared with placebo in patients with gastrointestinal (GI) cancers receiving oxaliplatin-based multiagent cytotoxic treatment, according to results from the phase 3 RECITE trial (NCT03362177) published in the New England Journal of Medicine.¹

In the primary end point analysis, no CIT-induced modifications of the chemotherapy dose occurred in the second or third chemotherapy cycles in 84% of patients receiving romiplostim (n = 109) vs 36% of those receiving placebo (n = 56), yielding an OR of 10.16 (95% CI, 4.44-23.72; P < .001) and a risk ratio of 2.77 (95% CI, 1.78-4.30; P < .001).¹

“Romiplostim was effective in treating persistent CIT and preventing its recurrence in patients with GI cancers,” lead study author Hanny Al-Samkari, MD, and colleagues wrote in the paper.¹ “With objective, drug label–based thresholds for chemotherapy dose reductions or delays, one-sixth of patients receiving romiplostim had chemotherapy dose modifications due to CIT, as compared with nearly two-thirds of those receiving placebo.”

Al-Samkari is the Peggy S. Blitz Endowed Chair in Hematology/Oncology at Massachusetts General Hospital and an associate professor of medicine at Harvard Medical School in Boston.

“We very clearly proved that romiplostim is effective. Of course, we already knew that because it's been published in smaller studies, but it's very good to have a large single study like this [to confirm its effectiveness],” Gerald Soff, MD, a professor of clinical medicine in the Division of Hematology and chief of the Classical Hematology Section at the University of Miami Miller School of Medicine in Florida, added in an exclusive interview with OncLive®.

What was the rationale for evaluating romiplostim for CIT management in GI cancers?

CIT is a common and challenging complication associated with myelosuppressive chemotherapy that carries an increased risk of bleeding, reduces chemotherapy relative dose intensity (RDI), and may compromise treatment outcomes in patients receiving adjuvant or systemic therapy for advanced disease. The prevalence of CIT following chemotherapy initiation is 62% overall among patients with colorectal cancer (CRC), and 32% among those on platinum-based regimens; the 3-month incidence is 13.5% and 12.9%, respectively. Despite the clinical burden of CIT, no widely available agents are approved for its treatment, and platelet transfusions are limited in availability and offer only a transient benefit while increasing infection risk.

Romiplostim is a subcutaneously administered thrombopoietin-receptor agonist peptibody that binds and activates c-MPL, the thrombopoietin receptor; it had previously demonstrated clinical benefit in CIT management in phase 2 trial (NCT02052882) and observational settings.

How was the RECITE trial designed?

This phase 3, international, double-blind, randomized, placebo-controlled RECITE trial enrolled patients receiving oxaliplatin-based multiagent cytotoxic chemotherapy for CRC, gastroesophageal cancer, or pancreatic cancer at any stage or treatment line, including adjuvant and advanced settings. Patients must have had persistent CIT, defined as a platelet count of 85 × 10⁹ per liter or lower on trial day 1, despite at least 14 days of recovery time from the prior cycle nadir (or at least 21 days for capecitabine and oxaliplatin [CAPOX]–based regimens); and at least 3 additional planned chemotherapy cycles remaining.

Eligible patients were randomly assigned 2:1 to receive subcutaneous romiplostim once per week or placebo, with dose adjustment targeting platelet counts of 100 to 250 × 10⁹ per liter, from trial day 1 through chemotherapy reinitiation and for up to 3 additional chemotherapy cycles.¹ Chemotherapy regimens included FOLFOX (leucovorin calcium [folinic acid], fluorouracil, and oxaliplatin), CAPOX, FOLFIRINOX (leucovorin calcium, fluorouracil, irinotecan hydrochloride, and oxaliplatin), or FOLFOXIRI (folinic acid, 5-fluorouracil, oxaliplatin, and irinotecan).

The primary end point was the absence of CIT-induced chemotherapy dose modifications in both the second and third planned chemotherapy cycles, as assessed by blinded independent central review. The trial was designed to have 93% power under assumed rates of 52.8% (romiplostim) and 23.7% (placebo) for no dose modification, with an overall 2-sided type I error rate of 5% and a hierarchical testing strategy to control for type I error.

The enrollment period ran from September 30, 2019, to October 24, 2023, with a primary analysis date of January 25, 2024, and a final trial visit on January 9, 2025. The trial was conducted at 55 sites across 14 countries.

What were the baseline characteristics of the study population?

In the overall population (romiplostim, n = 109; placebo, n = 56), baseline characteristics were as follows:

• Median age: 63 years
• Race: 90% White, 24% Hispanic, and 4% Black participants
• Median baseline platelet count: 69 × 10⁹ per liter
• Cancer type: 75% CRC, 13% gastroesophageal, 12% pancreatic
• Stage IV disease: 72% in the romiplostim group vs 61% in the placebo group
• ECOG performance status greater than 0: 53% for romiplostim vs 41% for placebo
• Fewer than 2 prior chemotherapy lines: 73% of all patients
• More than 2 prior chemotherapy cycles: 84% of all patients
• Prior bevacizumab (Avastin) use: 35% with romiplostim vs 20% with placebo
• Concurrent bevacizumab use: 25% in the romiplostim group vs 14% in the placebo group

Although 75% of patients overall completed the randomly assigned regimen, completion was more likely with romiplostim than placebo (82% vs 61%). The most common reasons for discontinuation in the placebo group vs the romiplostim group were patient request (12% vs 1%) and nonresponse (11% vs 0%).

Patients were treated with romiplostim for a median of 7.1 weeks (range, 0.1-22.0), at a median average weekly dose of 2.4 μg/kg, and 88% of patients had a maximum dose of no more than 5 μg/kg. The placebo was received for a median of 8.9 weeks (range, 0.1-16.0). The median duration of trial participation, including long-term follow-up, was 16.0 months (range, 1.9-61.4) in the romiplostim group and 19.0 months (range, 0.3-59.2) in the placebo group.

What secondary efficacy end point data were reported?

Median platelet nadirs during the treatment period were higher with romiplostim than with placebo (87 × 10⁹ per liter [IQR, 62-112] vs 58 × 10⁹ per liter [IQR, 48-64]; estimated mean difference, 26 × 10⁹ per liter [95% CI, 8-44]; P = .005).

Time to first platelet response, defined as a platelet count of 100 × 10⁹ per liter or higher with no platelet transfusions in the prior 7 days, was significantly faster with romiplostim than with placebo at a median of 1.1 weeks (95% CI, not estimable) vs 2.1 weeks (95% CI, 1.1-3.0; P < .001), respectively (HR, 2.67; 95% CI, 1.81-3.95; P < .001). Overall, platelet responses were achieved in 97% of romiplostim patients and 77% of placebo patients. The duration-adjusted rate of bleeding events of grade 2 or higher per 100 patient-years was 4.0 in the romiplostim group and 7.6 in the placebo group (HR, 0.53; 95% CI, 0.04-6.77; P = .63).

What did data from the post hoc analysis reveal about the effect of romiplostim on RDI?

In a post hoc analysis, the percentage of patients with a chemotherapy RDI greater than 85% was 75% with romiplostim vs 38% with placebo. The median RDI across the 3 planned cycles was 98% (IQR, 87-100) with romiplostim and 77% (IQR, 56-98) with placebo. Notably, 95% of patients in the romiplostim group vs 73% of those in the placebo group received all 3 planned chemotherapy cycles during the trial. Additionally, by week 4, platelet responses were observed in 96% vs 66% of patients in these respective groups.

What was the agent’s safety profile?

Through long-term follow-up, death occurred in 53% of romiplostim patients (n = 58) and 45% of placebo patients (n = 25), with progressive disease accounting for 71% and 72% of deaths in each respective group. No deaths were attributed to adverse effects (AEs) considered by the investigator to be related to romiplostim or placebo.

In total, 87% and 62% of patients in the romiplostim group and placebo group, respectively, experienced any-grade AEs. Grade 3 or higher AEs occurred in 37% and 22% of patients, respectively. AEs considered related to romiplostim or placebo by the investigator were reported in 12% and 7% of patients, respectively, with the most frequent being nausea (2% in each group) and headache (2% with romiplostim; 0% with placebo). No serious AEs were considered to be related to romiplostim or placebo, and no AEs related to either study agent led to death or discontinuation of romiplostim, placebo, or chemotherapy. No thrombocytosis was reported.

The incidence of serious AEs (SAEs) was 21% with romiplostim and 5% with placebo. The most common SAEs were anemia (3% vs 0%, respectively) and COVID-19 pneumonia (3% vs 0%). Neutropenia of grade 3 or higher occurred in 8% of patients receiving romiplostim and 7% of patients receiving placebo; grade 3 or higher febrile neutropenia was observed in 1 patient receiving romiplostim. Granulocyte colony-stimulating factor or pegylated granulocyte colony-stimulating factor was received by 20% and 13% of patients in the romiplostim vs placebo groups, respectively.

Thromboembolic AEs occurred in 2% vs 0% of patients receiving romiplostim vs placebo. Newly diagnosed myelodysplastic syndrome (MDS) and secondary cancers, respectively, were observed at respective rates of 1% and 2% in the romiplostim group and 0% and 5% in the placebo group; the MDS case was considered unrelated to romiplostim. Two fatal AEs of COVID-19 pneumonia and multiple organ dysfunction syndrome occurred in the romiplostim group within 30 days after the last dose; neither was considered related to romiplostim. Discontinuation of study treatment due to AEs occurred in 3 romiplostim patients and 1 placebo patient.

What were the study's limitations?

The investigators acknowledged that the 3-cycle intervention period and lack of standardized posttrial follow-up inherently limited the interpretation of cancer outcomes. Baseline imbalances between the treatment groups, along with patient population heterogeneity across tumor types, stages, chemotherapy regimens, and lines of therapy, further confounded the interpretation of oncologic end points. Although the trial was adequately powered for the primary end point, a low number of grade 2 or higher bleeding events limited the ability to formally assess subsequent secondary efficacy end points, restricting meaningful interpretation to platelet support and chemotherapy RDI outcomes. Additionally, the trial was not designed to detect late AEs during extended follow-up.

What’s next for romiplostim?

The investigators noted that ongoing real-world studies, which can enroll larger and broader patient populations and accommodate longer treatment durations, may provide further insight into the effect of romiplostim-supported CIT treatment on long-term outcomes.¹ The findings also provide a foundation for investigation in broader populations, including an ongoing phase 3 trial (NCT03937154) evaluating romiplostim vs placebo for CIT in patients with breast, ovarian, or non–small cell lung cancers.^1,2

References

1. Al-Samkari H, Muñoz C, Geredeli Ç, et al. Romiplostim versus placebo for chemotherapy-induced thrombocytopenia. N Engl J Med. 2026;394(11):1061-1073. doi:10.1056/NEJMoa2511882
2. Study of romiplostim for chemotherapy-induced thrombocytopenia in adult subjects with non-small cell lung cancer (NSCLC), ovarian cancer, or breast cancer. ClinicalTrials.gov. Updated April 7, 2026. Accessed June 5, 2026. https://clinicaltrials.gov/study/NCT03937154