ITP is an autoimmune bleeding disorder in which the immune system destroys platelets and suppresses their production in the bone marrow. In children, the condition often resolves on its own, but a substantial minority develop persistent or chronic disease requiring second-line therapy. TPO-RAs, which stimulate platelet production by activating the thrombopoietin receptor on megakaryocytes, are recommended second-line treatments. The class includes oral agents such as eltrombopag and avatrombopag, and the injectable peptide-mimetic romiplostim. When one agent fails, guidelines have suggested that switching within the class, particularly from an oral drug to romiplostim, may still succeed, because the agents differ in structure, metabolism, and pharmacology.

That assumption, however, had never been rigorously tested in the specific scenario that matters most: children who have failed not just one but several oral TPO-RAs. The Chinese research team, led by Shuyue Dong and Yumeng Jia with senior authors Xiaoling Cheng and Runhui Wu, was able to examine this question thanks to a unique real-world circumstance. Romiplostim reached the market late in China, which meant that children there often received it only after prolonged exposure to, and failure of, multiple oral TPO-RAs. This sequencing created a natural experiment that would be difficult to construct deliberately.

The study was a single-center, retrospective, observational analysis of patients aged 1 to 18 years with persistent or chronic ITP who received romiplostim for at least four weeks. The researchers assessed clinical characteristics, prior therapies, platelet response, bleeding outcomes, the need for rescue therapy, and adverse drug reactions. The work was approved by the Medical Ethics Committee of Beijing Children’s Hospital, and all patients provided written informed consent. The study was registered as ChiCTR2400086099 and funded by the National Key R&D Program of China and the National Natural Science Foundation of China.

Sixteen patients were analyzed in total. Thirteen of them had refractory ITP and had previously used two or more TPO-RAs before starting romiplostim. The results in this refractory group were strikingly poor. Eleven of the thirteen showed no response to romiplostim at all, and the remaining two exhibited only a transient partial response. In other words, not a single child in the refractory cohort achieved a complete or durable response once multiple oral TPO-RAs had already failed. The clinical burden in this group was substantial: twelve of the thirteen patients experienced bleeding episodes, eight required rescue therapy, and all thirteen ultimately switched to other drugs after romiplostim also proved ineffective.

The remaining three patients told a very different story, and the contrast is central to the study’s message. These children chose romiplostim not because other TPO-RAs had failed to work, but because of practical concerns such as potential drug interactions or liver dysfunction, which are recognized limitations of oral agents like eltrombopag. Only one of the three had previously used eltrombopag. In this selected group, two of the three patients achieved a complete response, while one showed no response. The findings suggest that romiplostim retains real value when it is used as a genuine alternative rather than as a last resort within an exhausted drug class.

On the safety front, the news was unambiguously positive. No adverse drug reactions were observed in any of the sixteen patients during the study period. This is consistent with the generally favorable tolerability profile that romiplostim has shown in pediatric populations, and it confirms that the drug itself is not the problem. The issue, the authors argue, is not whether romiplostim is safe or even whether it can work in children with ITP, but whether it can work in children whose disease has already demonstrated resistance to multiple members of its own class.

That resistance pattern carries important biological implications. TPO-RAs all converge on the same molecular target, the thrombopoietin receptor, also known as c-Mpl. If a child’s platelet production pathway remains unresponsive to two or more agents that activate this receptor through different chemical scaffolds, the likelihood that a third agent hitting the same target will succeed appears low. The study’s data give that reasoning empirical weight: within-class switching after multiple failures produced essentially no durable responses. This challenges the common clinical habit of cycling through TPO-RA options sequentially and suggests that the time and exposure spent on further within-class trials may come at the cost of delaying more appropriate therapies.

The authors conclude that romiplostim remains a safe and effective option for carefully selected pediatric ITP patients, but that children who have failed multiple oral TPO-RAs derive minimal benefit from switching to it. Instead, they recommend prioritizing alternative therapies that target immune dysregulation, the underlying mechanism of the disease, rather than continuing to push on the same thrombopoietin receptor pathway. Such alternatives could include immunomodulatory or immunosuppressive approaches that address the autoimmune attack driving platelet destruction, although the study itself does not evaluate specific alternative regimens.

As with any retrospective, single-center study, the findings come with limitations. The cohort was small, with only sixteen patients, and the refractory subgroup of thirteen, while internally consistent, limits statistical precision. Retrospective designs are also vulnerable to selection bias, since the decision to try romiplostim after multiple failures reflected real-world clinical judgment rather than a randomized protocol. Nevertheless, the near-universal failure of romiplostim in the multiply pretreated group is difficult to dismiss as chance, and the real-world sequencing that Chinese late-market conditions created gives the observation unusual relevance for clinicians everywhere who face the same decision. For families navigating refractory pediatric ITP, the study offers a clear if sobering message: after multiple oral TPO-RA failures, the next step should probably not be another drug from the same shelf.

Subject of Research: Romiplostim switching after multiple oral thrombopoietin receptor agonist failures in pediatric refractory immune thrombocytopenia

Article Title: Is switching to romiplostim still worthwhile after failure of multiple oral TPO-RAs in pediatric refractory ITP?

Article References: Is switching to romiplostim still worthwhile after failure of multiple oral TPO-RAs in pediatric refractory ITP?. (n.d.). https://doi.org/10.1007/s00277-026-07174-y

Image Credits: AI Generated

DOI: 10.1007/s00277-026-07174-y

Keywords: immune thrombocytopenia, romiplostim, eltrombopag, avatrombopag, TPO receptor agonists, pediatric hematology, refractory ITP, switching strategy, platelet response, drug safety, retrospective study, Beijing Children's Hospital

The research, conducted by a team led by Xiaohui Xiao, Guiping Chen, and Cong Xu across multiple Chinese hematology centers, compared two first-line strategies in 123 newly diagnosed patients with severe aplastic anemia aged between 40 and 65, treated between 2016 and 2024. Sixty-seven patients underwent haploidentical hematopoietic stem cell transplantation, a procedure in which the donor is a family member, typically a parent or child, who shares only one matched human leukocyte antigen haplotype with the recipient. The remaining 56 patients received the preferred conventional regimen: immunosuppressive therapy combined with eltrombopag, a combination that has become the default first-line option for older adults precisely because it avoids the risks of transplantation in a population less able to tolerate intensive procedures.

The logic of the comparison is rooted in a long-standing clinical dilemma. Immunosuppressive therapy, usually built around horse antithymocyte globulin and cyclosporine, works by damping down the aberrant immune destruction of hematopoietic stem and progenitor cells that drives the disease. Eltrombopag was added to this regimen after trials showed it could boost response rates and improve blood counts. But the combination is far from curative. A substantial fraction of patients never respond, others relapse, and some evolve to clonal complications such as myelodysplastic syndrome. Transplantation, by contrast, replaces the failed marrow entirely with healthy donor stem cells, offering the possibility of a durable cure, but at the cost of transplant-related mortality, graft-versus-host disease, and infections during the period of immune reconstitution. For younger patients with a fully matched sibling donor, transplantation has long been the favored first option. The open question has been whether the same logic applies to middle-aged and older patients when only a half-matched family donor is available.

The study’s findings on response rates were striking. At three months after treatment, 17.9 percent of the transplant group had achieved a complete response, meaning fully normalized blood counts, compared with just 5.4 percent of those on immunosuppressive therapy plus eltrombopag. By six months, the gap had widened: 44.8 percent versus 17.9 percent. At twelve months, 64.2 percent of transplant recipients had achieved complete response compared with 33.9 percent of the drug-treated group, differences that were all statistically significant. In a disease where the depth and durability of blood count recovery correlate closely with quality of life and survival, these numbers represent a substantive clinical signal, not a marginal one.

Survival outcomes told a more nuanced story. Three-year overall survival was remarkably similar between the two groups, at 78.0 percent for the transplant arm and 80.3 percent for the immunosuppressive therapy plus eltrombopag arm, a difference that did not reach statistical significance. The researchers also performed a landmark analysis restricted to patients who had survived their first six months of treatment, a statistical technique designed to correct for the tendency of early transplant mortality to bias survival comparisons. Conditioned on surviving to that checkpoint, three-year overall survival was again comparable: 92.2 percent for transplant recipients versus 88.9 percent for the drug group. In other words, although transplantation produced deeper and faster hematologic recovery, it did not translate into a survival advantage across the entire 40-to-65 age range during the study’s follow-up window.

The most clinically provocative result emerged when the investigators stratified patients by age. Among those aged 40 to 49, three-year failure-free survival, a measure that captures both treatment failure and relapse, was significantly higher after transplantation: 82.9 percent versus 60.1 percent with drugs, a statistically significant difference. For patients aged 50 to 65, however, there was essentially no difference between the strategies, with the comparison yielding a P value of 0.984. This age-dependent divergence suggests that the durability advantage of a donor-derived immune system may only outweigh the upfront hazards of transplantation in patients whose physiological reserve and life expectancy are sufficient for those late benefits to materialize before competing risks intervene.

The technical significance of this study lies partly in its scale and design. Retrospective comparisons are inherently vulnerable to selection bias, since clinicians and patients choose treatments based on factors such as donor availability, comorbidities, and disease severity that can independently influence outcomes. By pooling patients across multiple centers and analyzing a relatively homogeneous age window, the researchers aimed to reduce some of that confounding, though the authors and independent commentators alike would acknowledge that only a prospective randomized trial can definitively settle the question. The study also benefited from independent verification of its survival analyses, an unusual methodological safeguard that strengthens confidence in the reported statistics.

For the field of haploidentical transplantation, the findings arrive at a moment of rapid evolution. Half-matched transplants were once considered a last resort because of high rates of graft rejection and graft-versus-host disease. Improvements in graft engineering, post-transplant cyclophosphamide prophylaxis, and supportive care have progressively narrowed the gap between haploidentical and fully matched transplantation. The current study, spanning cases from 2016 to 2024, reflects that modern era, and its results give weight to the argument that a readily available family donor should not be discounted simply because the patient has passed the conventional age threshold for upfront transplantation.

Clinically, the study does not demand an immediate rewrite of treatment guidelines, but it does complicate the conversation that hematologists have with newly diagnosed middle-aged patients. The current standard of care, immunosuppressive therapy with eltrombopag, remains a reasonable first choice that spares many patients the risks of a transplant, and overall survival in this cohort was excellent with both strategies. But the data indicate that patients in their forties who have a suitable half-matched family donor may achieve more durable, treatment-free blood count recovery by proceeding to transplantation upfront, rather than waiting for the drug regimen to fail and then transplanting in a compromised state. For patients in their fifties and early sixties, the calculus appears more balanced, and the lower-intensity medical route may be preferable.

The study was funded in part by China’s National Key Research and Development Program and the National Natural Science Foundation of China, and the authors declared no conflicts of interest. Its publication as an open-access article means that clinicians worldwide, including in regions where fully matched donors are scarce and haploidentical transplantation carries the greatest practical weight, can examine the full dataset. As follow-up extends and prospective trials are designed, the central lesson of this work is likely to endure: age alone is an incomplete guide to treatment selection in severe aplastic anemia, and the decade between forty and fifty may mark a genuine biological tipping point in the trade-off between the slow, uncertain path of immune re-education and the swift, definitive, but riskier replacement of the marrow itself.

Subject of Research: First-line treatment of severe aplastic anemia in patients aged 40–65 comparing haploidentical stem cell transplantation with immunosuppressive therapy plus eltrombopag

Article Title: Haploidentical hematopoietic stem cell transplantation compared to immunosuppressive therapy plus eltrombopag as first-line treatment for severe aplastic anemia patients aged 40–65: a multicenter retrospective study

Article References: Xiao, X., Chen, G., Xu, C., Wu, L., Liu, X., Zhou, M., Chen, X., Zhou, R., Mo, W., Li, Y., Xu, S., Zhang, Y., Lin, D., & Wang, S. (2026). Haploidentical hematopoietic stem cell transplantation compared to immunosuppressive therapy plus eltrombopag as first-line treatment for severe aplastic anemia patients aged 40–65: a multicenter retrospective study. Annals of Hematology. https://doi.org/10.1007/s00277-026-07289-2

Image Credits: AI Generated

DOI: 10.1007/s00277-026-07289-2

Keywords: severe aplastic anemia, haploidentical transplantation, hematopoietic stem cells, immunosuppressive therapy, eltrombopag, bone marrow, failure-free survival, overall survival, hematology, first-line treatment, graft-versus-host disease, retrospective study