Discussion
To our knowledge, this is the first case series to report on the real-world use of TPO-RAs to treat thrombocytopenia in pediatric and AYA patients with hematologic malignancies. Our study demonstrated that treatment with TPO-RAs effectively increased platelet counts and reduced transfusion requirements in 93% (14/15) of patients. Platelet counts significantly increased within two weeks of initiating TPO-RA therapy, with sustained responses and an associated significant reduction in platelet transfusion requirements. These findings align with previous studies in adult populations and other pediatric contexts focused predominantly on their use in solid malignancies, suggesting that TPO-RAs can be valuable in managing thrombocytopenia in pediatric and AYA patients with hematologic malignancies.5,6,9,10,25,26
Refractory thrombocytopenia was the most common reason for TPO-RA administration. Nine patients (60%) had either an existing bleeding complication or a comorbidity that placed them at high risk of severe bleeding events and insufficient response to platelet transfusions. No patients experienced adverse events related to TPO-RA therapy. Three patients experienced self-resolving thrombocytosis without thromboembolic events, a known but manageable risk associated with TPO-RA use.
While more data are required to identify appropriate indications for TPO-RA therapy, our data suggest that both romiplostim and eltrombopag are safe across a range of doses, durations of therapy, diagnoses, and stages of treatment. Three representative patient courses are detailed in vignettes to highlight the broad range of potential contexts in which TPO-RAs may be useful (Supplemental Data). Despite the broad potential utility of TPO-RAs in managing thrombocytopenia, their use was limited in this cohort due to the high cost of these agents. The cost of 125mcg romiplostim is approximately $1390, and a two-week supply of 50 mg eltrombopag is approximately $6500. As such, in our cohort, use of TPO-RA therapy was limited to those patients with thrombocytopenia refractory to platelet transfusions and/or those with comorbidities that predisposed to high bleeding risk.
Notably, our study included a high proportion of patients (67%) with multiple life-threatening comorbidities and patients with relapsed disease, nearly half of whom had received prior HSCT. As such, more than half of the patients died of complications unrelated to TPO-RA use. In addition, about one third of patients in this cohort with high-risk hematologic malignancies experienced relapse after initiation of TPO-RA therapy.
Historically, concerns have been raised about the potential for TPO-RAs to stimulate the proliferation c-MPL-expressing myeloid blasts.27 Subsequent preclinical studies28 and clinical trials25,26,29 of TPO-RAs in the context of AML and MDS have not supported these concerns. In our study, two of four patients with myeloid malignancies relapsed after receiving TPO-RA therapy: one patient with TCF3-ZN384-postive B/myeloid MPAL (patient 3) and one with GATA2-driven MDS transformed to AML (patient 4). Patient 7 with t(8;21) AML had a transiently rising RUNX1-RUNX1T1 fusion transcript by reverse transcription polymerase chain reaction but did not experience overt hematologic relapse. Patient 14 with inv(16) AML has been relapse-free for 48 months. While this study was neither designed nor powered to assess a causal relationship between TPO-RA use and leukemia relapse, the relapse rates both in the overall cohort and specifically among patients with myeloid malignancies are concordant with those reported in the literature for similar populations30,31.
Although our cohort represents a highly pretreated group with multiple comorbidities, including severe infections, bleeding complications, and hyperinflammation, 14/15 (93%) of patients achieved a platelet count of 50 x 109/L or greater and had decreased platelet transfusion requirements within 8 weeks (range 1-8 weeks, median 3 weeks). Seven patients became transfusion independent during this time. We hypothesize that the magnitude of benefit may be even greater in populations with fewer comorbidities and prior cancer-directed therapies.
This study has several limitations that may restrict the external generalizability of our findings, including the retrospective design, single institution nature, and small sample size comprising a highly heterogeneous population enriched for patients with poor clinical outcomes. Further prospective studies with larger, more homogenous cohorts are needed to confirm these findings and establish standardized guidelines for the use of TPO-RAs in pediatric and AYA patients with hematologic malignancies. Moreover, our study did not include a matched control group of patients who did not receive TPO-RA therapy, raising the possibility that some patients could have experienced spontaneous platelet recovery. However, the close temporal association between TPO-RA administration and the significant increase in platelet counts in nearly all patients, most of whom had severe, transfusion-refractory thrombocytopenia, make it likely that platelet recovery was attributable to TPO-RA treatment. In addition, the rationale for selecting a specific TPO-RA among the four available agents was not available from our chart review. While we postulate that practical considerations, such as patient preferences around route of administration, dosing frequency, cost, and insurance coverage influenced the choice of therapy, we are unable to draw conclusions about the comparative efficacy and safety of these agents in this patient population. More detailed prospective data collection, including clinician rationale for therapy selection, would help to better understand these decision-making factors and optimize treatment strategies.
Despite these limitations, our findings suggest that romiplostim and eltrombopag are safe and effective agents that may be incorporated into supportive care regimens to manage thrombocytopenia in pediatric and AYA patients with hematologic malignancies and thrombocytopenia across a broad range of clinical contexts. However, neither eltrombopag nor romiplostim takes effect immediately, requiring a median of three weeks to improve platelet counts; therefore, these agents should not be a substitute for platelet transfusions during this period. Given the dearth of data on TPO-RA use in pediatric and young adult patients with hematologic malignancies, prospective clinical trials are warranted to validate our findings and explore the long-term outcomes of TPO-RA therapy. Additionally, future research should focus on identifying patient-specific factors that predict response to TPO-RAs, optimizing dosing strategies, and evaluating the cost-effectiveness of TPO-RA therapy in this population.