Eltrombopag in combination with immunosuppressive therapy in pediatric severe aplastic anemia: phase 2 ESCALATE trial Open Access

Severe aplastic anemia (SAA) is a rare, life-threatening disease characterized by acquired bone marrow hypoplasia and peripheral blood cytopenia due to immune dysregulation.¹ Immune activation causes T cells to attack hematopoietic stem cells (HSCs) and other progenitor cells, although the relevant antigens on HSCs are not conclusively defined.² 

For pediatric patients with SAA who have a matched sibling donor, HSC transplantation (HSCT) is the standard first-line treatment. If a suitable donor is unavailable, immunosuppressive therapy (IST) with antithymocyte globulin (ATG) and cyclosporine A (CsA) is the standard alternative treatment.^3,4 Overall response rates (ORRs; defined as a complete response [CR] or partial response [PR]) at 6 months for IST in pediatric patients range from 49% to 74%,^5-7 with overall survival rates of 74% to 97% (at 1-10 years)^5-9; however, lower CR rates (22%-35%)^5,7 and significant relapse rates remain challenges. Long-term complications, such as relapse and clonal evolution, including paroxysmal nocturnal hemoglobinuria (PNH) and myelodysplastic syndromes, result in event-free survival rates of 37% to 76%.^6,8,10 Recent studies highlight the emerging role of upfront matched unrelated donor HSCT as a promising first-line therapy for pediatric SAA.¹¹ This is under investigation in the upfront randomized TransIT trial (ClinicalTrials.gov identifier: NCT05600426).¹² 

Therapy options are limited for patients with relapsed/recurrent or refractory SAA who are unable to undergo HSCT.^10,13 Therefore, alternative efficacious and tolerable treatments are needed. Eltrombopag, a small molecule thrombopoietin receptor agonist (TPO-RA), is approved for the treatment of immune thrombocytopenia (ITP), hepatitis C virus–related thrombocytopenia, and SAA in patients with an inadequate IST response.¹⁴ In 2018, the US Food and Drug Administration expanded eltrombopag’s product label to include its use with IST as first-line treatment for SAA for patients aged ≥2 years.¹⁴ Recent trials show that adding eltrombopag to standard IST improves the hematologic response in untreated adult patients with SAA, without impacting safety.^15,16 Here we present data from ESCALATE (ClinicalTrials.gov identifier: NCT03025698), designed to evaluate the efficacy, safety, and pharmacokinetics (PK) of eltrombopag in combination with IST in pediatric patients with SAA.

ESCALATE is a phase 2, open-label, noncontrolled, intrapatient dose-escalation study evaluating the efficacy, safety, and PK profile of eltrombopag in combination with IST in pediatric patients with SAA.

Patients aged 1 to <18 years with SAA, unsuitable or ineligible for HSCT (determined by local practices or national guidelines¹⁷), were enrolled into 1 of 2 cohorts: (1) patients with relapsed/recurrent or refractory SAA who had received IST previously (relapsed/refractory [R/R] cohort); and (2) treatment-naïve patients. Eligible R/R cohort patients had to have a prior diagnosis of SAA and a diagnosis of R/R SAA or recurrent AA, and treatment-naïve patients were required to have a diagnosis of SAA at enrollment (supplemental Methods).

Eltrombopag was administered orally once daily, at an initial dose of 25 mg for patients aged <6 years and 50 mg for patients aged 6 to <18 years. Intrapatient dose escalations were permitted every 2 weeks from day 2 of week 3 until the target platelet count (50 × 10⁹/L to 200 × 10⁹/L) or the maximum dose of 150 mg (supplemental Methods) was reached. Eltrombopag starting doses were based on safety and tolerability data from pediatric ITP studies. Dose modification guidance was provided for adverse events (AEs; supplemental Methods).

Patients received eltrombopag for 26 weeks or until early discontinuation. In the R/R cohort, eltrombopag was administered in 1 of 2 regimens based on the patients’ prior drug experience and investigator’s decision: (1) horse-ATG (hATG; ATGAM, Pfizer) 40 mg/kg per day for 4 days, CsA, and eltrombopag; or (2) CsA and eltrombopag without hATG. All treatment-naïve patients received hATG 40 mg/kg per day for 4 days, CsA, and eltrombopag. All patients received CsA every 12 hours starting at day 1 to maintain whole blood trough concentrations of 200 to 400 μg/L (supplemental Methods). After eltrombopag initiation, standard supportive care (including anti-infective medications and transfusion support) was used per institutional guidelines. HSC stimulants and other TPO-RAs were not permitted.

This study included a 26-week treatment period, a 52-week follow-up period, and a 3-year long-term follow-up period (supplemental Figure 1). During the treatment period, patient visits occurred on day 1 of weeks 1 to 10 and weeks 12, 14, 18, 22, and 26. A safety follow-up visit occurred 30 days after the last dose of eltrombopag, with follow-up visits every 4 weeks to collect safety and efficacy data. After the 52-week follow-up period, patients could enroll into a 3-year long-term follow-up period, with visits every 6 months. Patients deriving clinical benefit from eltrombopag could continue treatment during the follow-up periods. Responders could continue CsA (2 mg/kg) throughout the 52-week follow-up and the initial 26 weeks of 3-year follow-up (up to week 104).

This study was designed and conducted in accordance with the guidelines for Good Clinical Practice of the International Council for Harmonisation, principles of the Declaration of Helsinki, and local laws and regulations. The study protocol and all amendments were reviewed and approved by the independent ethics committee and/or institutional review board of each study center. Before participating in the study, representatives of the patients (parents or legal guardians) provided written informed consent; additionally, assent was obtained from patients where appropriate, as defined by the institutional review board.

The objectives of this analysis were to evaluate the efficacy, safety, and PK of eltrombopag in pediatric patients with SAA. Assessment of efficacy included ORR at weeks 12, 26, 52, and 78 and was defined as the proportion of patients who achieved either CR or PR according to the North American Pediatric Aplastic Anemia Consortium criteria.⁶ Response criteria were the same for both cohorts. CR was defined as absolute neutrophil count ≥1.0 × 10⁹/L, hemoglobin ≥10 g/dL, and platelet count ≥100 × 10⁹/L in the absence of ongoing transfusions; PR was defined as absolute neutrophil count ≥0.5 × 10⁹/L, hemoglobin ≥8 g/dL, and platelet count ≥20 × 10⁹/L in the absence of ongoing transfusions. Patients whose hematologic values did not meet the criteria for CR or PR or those who received a red blood cell (RBC) transfusion within 6 weeks or a platelet transfusion or granulocyte/granulocyte-macrophage colony-stimulating factor within 2 weeks of response assessment were classified as nonresponders. Other efficacy end points were best overall response (the proportion of patients with CR or PR at any time point up to data cutoff), as well as RBC and platelet transfusion independence.

Safety was assessed by the frequency and severity of AEs. AEs were coded according to MedDRA version 26.1 and graded according to Common Terminology Criteria for Adverse Events version 4.03. Bone marrow morphology and dysplasia,¹⁸ degree of fibrosis (using European consensus grading¹⁹), cytogenetics by fluorescence in situ hybridization, and karyotyping were determined at weeks 12, 26, 52, and 78, at 2 and 3 years, and during disease progression. Clinically significant cytogenetic abnormalities were reported as AEs.

The PK parameters (area under the concentration-time curve [AUC] to the last quantifiable concentration point [AUC_last] and to the end of the dosing interval [AUC_tau], trough plasma concentration [C_trough], maximum plasma concentration [C_max], and apparent systemic or total body clearance at steady state from plasma) at the highest steady state dose (adjusted to 50 mg dose) were evaluated. Blood samples were collected from all patients who received ≥1 dose of eltrombopag, with additional sampling if the highest dose was not reached by week 12. Predose samples were collected on day 1 of weeks 3, 4, 6, 12, 14, and 26. Postdose samples were collected at 2, 4, 6, 8, and 24 hours on day 1 of weeks 3 and 12.

This analysis was based on data with a cutoff date of 4 May 2023, when all patients had reached their 78-week assessment (completion of the treatment and 52-week follow-up periods) or discontinued treatment earlier. Sample size calculation is detailed in the supplemental Methods.

The full analysis set for evaluating efficacy included all patients assigned to study treatment. The safety set included all patients who received ≥1 dose of eltrombopag. The PK analysis set included all patients who provided an evaluable PK profile, defined as having received the planned dose of eltrombopag before collecting a PK sample, followed meal restrictions per protocol, and did not have emesis within 6 hours of dosing.

Baseline characteristics were presented as frequencies and percentages for categorical data and median, minimum, and maximum for continuous data. ORRs were summarized by cohort using descriptive statistics and 95% confidence intervals. Transfusion frequency during treatment and maximum duration of transfusion independence were summarized descriptively. PK parameters were calculated by noncompartmental methods using Phoenix WinNonlin (Pharsight, Mountain View, CA) software. Descriptive statistics were used to summarize the PK parameters at the maximum dose (adjusted to a 50 mg dose), summarized by cohort group and age group.

Overall, 51 patients were enrolled: 14 in the R/R cohort and 37 in the treatment-naïve cohort. Of the 14 R/R cohort patients, 10 received regimen 1 (hATG, CsA, and eltrombopag), and 4 received regimen 2 (CsA and eltrombopag). Thirty-six patients completed the treatment phase (Figure 1; supplemental Table 1). Fifteen patients discontinued eltrombopag early: 5 (9.8%) due to an AE, 5 (9.8%) due to patient/guardian decision (2 withdrew for HSCT, 1 due to dissatisfaction with an AE, 1 for drug noncompliance, and 1 to continue a stable eltrombopag dose), 4 (7.8%) due to physician decision (2 patients due to drug noncompliance and 2 due to lack of efficacy, with 1 proceeding to HSCT), and 1 (2.0%) due to progressive disease.

Subsequently, 28 patients completed the initial follow-up period and entered the long-term follow-up period.

The median age of patients was 10 years (range, 2-17), with 54.9% male and 58.8% White (Table 1). Demographics and characteristics were well balanced between cohorts (including between the 2 R/R cohort treatment regimens; supplemental Table 2). The median time since the first SAA diagnosis was 14.7 months (range, 3.1-105.4) in the R/R cohort and 1.2 months (range, 0.5-8.5) in the treatment-naïve cohort (Table 1). Data on prior treatments received in the R/R cohort were not collected.

At baseline, >70% of patients were RBC and/or platelet transfusion dependent (Table 1). Baseline hematologic counts after patients had received transfusions and enrolled in the study are presented in Table 1.

For both cohorts combined, the ORR was 23.5% at week 12, which increased to 54.9% at week 26 and then remained constant at 52.9% and 51.0% at weeks 52 and 78, respectively. The proportion of patients with a CR increased over time (Figure 2A). In the R/R cohort, ORRs were 42.9%, 71.4%, 57.1%, and 57.1% at weeks 12, 26, 52, and 78, respectively. ORRs in the treatment-naïve cohort were 16.2%, 48.6%, 51.4%, and 48.6% at the same time points, respectively. Stratification of the R/R cohort by SAA history showed a trend toward higher ORRs in relapsed/recurrent patients vs refractory patients, although the data are limited by sample size (supplemental Figure 2).

When stratified by age, the ORRs in patients aged <12 years (n = 30) were 20.0%, 43.3%, 43.3%, and 40.0% at weeks 12, 26, 52, and 78, respectively. In patients aged ≥12 years (n = 21), the proportions were 28.6%, 71.4%, 66.7%, and 66.7% at the same time points, respectively. The trend for higher ORR in the R/R cohort vs the treatment-naïve cohort was also observed when stratified by cohort and age (supplemental Figure 3); however, data are limited by sample size.

Patients who responded tended to maintain these responses even after discontinuing eltrombopag (Figure 3). Among the 34 responders, 21 sustained their responses, including 6 (42.9%) from the R/R cohort and 15 (40.5%) from the treatment-naïve cohort. In the R/R cohort, 4 relapsed/recurrent patients and 2 refractory patients had sustained responses (supplemental Figure 4). The median time to response was 16.6 weeks in the R/R cohort (relapsed/recurrent, 14.4 weeks; refractory, 38.9 weeks) and 22.4 weeks in the treatment-naïve cohort (supplemental Figure 5). The best ORR at any time during the study among all patients was 66.7% (Figure 2B).

Across cohorts, 42 and 43 patients were dependent on RBC and platelet transfusions at baseline, respectively. At data cutoff, 28 patients (66.7%) achieved RBC transfusion independence for at least 56 days, and 33 patients (76.7%) achieved platelet transfusion independence for at least 28 days during treatment (supplemental Table 3). In the R/R cohort (n = 14), 10 patients were RBC dependent, and 10 patients were platelet transfusion dependent at baseline, of whom 7 (70.0%) and 8 (80.0%), respectively, achieved transfusion independence during treatment. At baseline, in the treatment-naïve cohort (n = 37), 32 patients were RBC transfusion dependent, and 33 were platelet transfusion dependent. Of these, 21 (65.6%) and 25 patients (75.8%), respectively, achieved transfusion independence during treatment.

At data cutoff, of the 34 patients with at least 1 RBC transfusion-free period, the median longest RBC transfusion-free period was 266 days (range, 58-1296), including 9 R/R patients (321 days [range, 185-1232]) and 25 treatment-naïve patients (262 days [range, 58-1296]; supplemental Table 4). Among the 40 patients with at least 1 platelet transfusion-free period, the median longest platelet transfusion-free period was 263 days (range, 34-1289), including 12 R/R patients (268 days [range, 36-1232]) and 28 treatment-naïve patients (249.5 days [range, 34-1289]; supplemental Table 4).

The median duration of exposure to eltrombopag was 234 days (range, 3-1611) overall, 294 days (range, 45-1239) in the R/R cohort, and 231 days (range, 3-1611) in the treatment-naïve cohort.

All patients experienced at least 1 AE, and 41 (80.4%) patients had a treatment-related AE of any grade, with the most common (occurring in ≥30% of patients) being increased blood bilirubin (43.1%), increased alanine aminotransferase (37.3%), and increased aspartate aminotransferase (33.3%); in most instances, the severity of these AEs was classified as either grade 1 or 2 (Table 2).

Serious AEs (SAEs) occurred in 29 patients (56.9%), of whom 19 (37.3%) had a grade ≥3 SAE. The most common (occurring in ≥3 patients) SAEs were pyrexia (all grade, n = 11 [21.6%]; grade ≥3, n = 4 [7.8%]), febrile neutropenia (all grade, n = 7 [13.7%]; grade ≥3, n = 6 [11.8%]), and acute kidney injury (all grade, n = 3 [5.9%]; grade ≥3, n = 2 [3.9%]).

Overall, 7 patients (13.7%) had treatment-related AEs leading to treatment discontinuation, of whom 2 (3.9%) had grade ≥3 events (Table 2). Increased alanine aminotransferase was the reason for discontinuation in 4 patients (7.8%), increased aspartate aminotransferase in 3 (5.9%), increased blood bilirubin in 2 (3.9%), and drug-induced liver injury in 1 (2.0%). AEs leading to dose adjustment or interruption occurred in 33 (64.7%), of whom 15 (29.4%) had an AE of grade ≥3. One patient had dose interruption due to platelet counts >200 000/μL; however, no patient permanently discontinued eltrombopag due to increased platelet counts.

No on-treatment deaths occurred. One patient died before initiating study treatment due to catheter-related sepsis. One patient died after study discontinuation during the 52-week follow-up period; this patient had a bone marrow transplant (day 382) and reported pneumonia on day 502 (120 days after the last dose of eltrombopag), which resulted in death.

Of 48 evaluated patients with baseline assessments, 23 (47.9%) exhibited no change in bone marrow fibrosis grade from baseline to worst postbaseline assessment. In contrast, 17 (35.4%) patients showed mild progression in bone marrow fibrosis: 16 patients progressed from grade 0 to 1, and 1 patient progressed to grade 2. In addition, 3 patients (6.3%) regressed from grade 1 to 0, and 1 patient (2.0%) regressed from grade 2 to 0 (supplemental Table 5).

Four patients exhibited moderate bone marrow dysplasia during the study. One treatment-naïve patient progressed to and died from acute myeloid leukemia (AML) ∼18 months after permanent discontinuation of eltrombopag; this patient had a normal karyotype, including at the time of AML diagnosis. Two patients had mild dysplasia at baseline; 1 in the erythroid component and 1 in both the erythroid and myeloid components. Both patients exhibited moderate dysplasia in these components during the study, which spontaneously resolved. Another patient showed moderate erythroid dysplasia and discontinued the study.

Clonal evolution to PNH was reported in 1 R/R patient during the 52-week follow-up period. This patient had permanently discontinued eltrombopag per physician decision 1 week before the diagnosis of PNH; no treatment for the event was reported.

Among the 51 patients, 37 had normal baseline karyotype results, and 10 had limited results due to insufficient metaphase cell counts (no noted abnormalities in the available metaphase cells). Four patients had unavailable central karyotype data, but local testing showed that 3 had a normal karyotype, whereas 1 patient had limited results due to insufficient metaphase counts. Postbaseline chromosomal abnormalities were identified in the bone marrow from 3 patients. One R/R patient, with a normal baseline karyotype, exhibited a pericentric inversion of chromosome 15 from week 12 onward. Despite this, the patient achieved a CR by day 175, which was maintained until the end of follow-up. The abnormality was not considered clinically significant, and thus, treatment was continued. A treatment-naïve patient experienced a transient additional chromosome Y in 2 of 20 metaphases at week 26, which was not reported as an AE. This patient’s baseline karyotype was not assessable due to insufficient metaphases but was normal at weeks 12 and 52; PR was achieved at day 302 and CR at day 484. Another treatment-naïve patient had an abnormal karyotype 46,XY,t(2,3)(p23,q12)[2]/46,XY[18] with a clonal reciprocal translocation between chromosomes 2 and 3, observed over 3 years after stopping eltrombopag, which was reported as an AE (deemed unrelated to the study treatment by the treating physician). This patient achieved CR on day 456 and had normal blood test results at the last assessment on day 1367.

At the highest steady-state dose achieved (adjusted to a 50 mg dose), eltrombopag clearance was lower, and plasma exposure (AUC_tau and C_max) was higher in patients aged 1 to <6 years than those aged 6 to <18 years. In patients aged 1 to <6 years, the geometric mean (geometric coefficient of variation, percent [geo-CV%]) of apparent clearance at steady state (CL_ss/F), AUC_tau, and C_max for the dose adjusted (to 50 mg) at the highest steady-state dose was 110 mL/h (72.0%), 460 h∗μg/mL (64.9%), and 25.6 μg/mL (42.2%), respectively (Table 3). Maximum eltrombopag plasma concentrations were observed at 2.0 hours (median T_max) after dosing. In patients aged 6 to <18 years, the corresponding geometric mean (geo-CV%) of CL_ss/F, AUC_tau, and C_max were 184 mL/h (52.6%), 285 h∗μg/mL (54.2%), and 15.2 μg/mL (49.5%), respectively, and the median T_max was 4.08 hours (Table 3).

At week 3 of the study (2 weeks after starting eltrombopag), the PK parameters (adjusted to a 50 mg dose) showed the following geometric mean (geo-CV%) values: CL_ss/F in all patients was 126 mL/h (47.5%); AUC_tau was 410 h∗μg/mL (48.3%); and C_max was 22.5 h∗μg/mL (46.8%), achieved at a median T_max of 3.3 hours. The mean clearance at the highest steady-state dose achieved for all patients was higher than that of the starting dose at week 3 (186 vs 144 mL/h).

Eltrombopag is approved as a first-line treatment for SAA in patients aged >2 years. Although data on eltrombopag in children, particularly those with R/R SAA, are scarce, observational studies indicate that eltrombopag alongside various ISTs can restore trilineage hematopoiesis.^20-22 However, clinical trials show conflicting results. This study aimed to assess the efficacy, safety, and PK of eltrombopag in pediatric patients with SAA.

The US Food and Drug Administration approved eltrombopag as a first-line SAA treatment based on a phase 1/2 study by the National Institutes of Health, in which patients aged ≥2 years with untreated SAA received eltrombopag combined with IST.¹⁴ The National Institutes of Health study revealed that combining eltrombopag with IST improved response rates compared with a historical IST-only group.¹⁵ In a subgroup analysis of 40 pediatric patients, there was no significant difference in ORR (eltrombopag plus IST, 70% vs IST, 72%; P = .78) and CR rate (eltrombopag plus IST, 30% vs IST, 23%; P = .42) at 6 months.²³ However, in adults, ORR was significantly higher with eltrombopag (eltrombopag plus IST, 82% vs IST, 58%; P < .001).²³ Further analysis indicated numerically lower response rates for patients aged <12 years receiving eltrombopag vs IST alone (ORR, 63% vs 78%, respectively [P = .29]), with significantly lower CR (6% vs 24%, respectively [P = .049]), whereas patients aged ≥12 years achieved numerically higher ORR and CR rate (ORR, 75% vs 67% [P = .48]; CR, 46% vs 21% [P = .052]).²³ A trial in Russia found no significant difference in ORR at 4 months between children with newly diagnosed SAA treated with IST with and without eltrombopag (65% vs 53%; P = .22), although the CR rate at 4 months was higher with eltrombopag (31% vs 12%; P = .027).²⁴ A retrospective analysis similarly showed higher CR rates with eltrombopag at 3 months (30% vs 8%; P = .004) and 6 months (50% vs 10%; P < .001) but no advantage in ORRs or long-term CR rates.²⁵ In ESCALATE, ORRs with eltrombopag and IST treatment were similar at 6 (48.6%), 12 (51.4%), and 18 months (48.6%) for treatment-naïve patients. Notably, the CR rates in treatment-naïve patients increased over time. A key consideration is that patients in ESCALATE had the option to continue treatment beyond 6 months, whereas, in previous studies, eltrombopag treatment was limited to 6 months. This difference in treatment duration could potentially explain the higher CR observed in ESCALATE. Furthermore, ESCALATE used more stringent response criteria than some previous studies, including a requirement of 6 weeks of RBC transfusion independence.

Eltrombopag treatment in pediatric patients with R/R SAA is not well explored. Prior studies of mostly adults show hematologic responses with eltrombopag in 40% to 50% of patients with refractory SAA at 12 to 24 weeks.^26-28 However, these studies included a limited number of pediatric patients. In the R/R cohort of the ESCALATE study, ORRs of 71.4% at week 26 and 57.1% at weeks 52 and 78 suggest that there may be a role for eltrombopag in this population. Interestingly, the ORRs at all time points were observed to be higher in patients with relapsed/recurrent SAA than in treatment-naïve patients, which may indicate that eltrombopag is more effective in these patients, although further investigation is needed. The small sample size and lack of data regarding prior IST treatment regimens in the R/R cohort preclude any meaningful conclusions. Furthermore, in both cohorts, eltrombopag plus IST treatment positively affected patients’ transfusion needs, with most patients achieving RBC and platelet transfusion independence.

Eltrombopag’s safety profile in pediatric patients aligns with that of adults with SAA.^15,16 Consistent with previous reports, the most frequently recorded AEs were hyperbilirubinemia and increased liver enzymes. Serum discoloration and interference with total bilirubin and creatinine testing have been reported in patients taking eltrombopag^29,30; consequently, high levels of total bilirubin are not reflective of liver function abnormalities. As a pediatric trial, a more conservative approach was taken, resulting in more discontinuation than in adult trials. Eltrombopag was discontinued for mildly increased liver enzymes, and all episodes of transaminitis either resolved or were resolving after eltrombopag discontinuation. No additional safety signals were identified. There were no on-treatment deaths, and neither of the 2 deaths reported was suspected to be related to study treatment.

Bone marrow fibrosis has been a concern in patients with ITP treated with TPO-RAs because of their effect on megakaryopoiesis.³¹ A single-center study evaluating bone marrow fibrosis in 66 patients with ITP found that TPO-RAs induce myelofibrosis grades 2/3 in approximately one-fifth of patients, increasingly with longer treatment (>2 years).³² Despite limited evidence of bone marrow fibrosis in patients treated with eltrombopag for up to 44 weeks, it is speculated that exposure could heighten some patients’ risk.³³ However, eltrombopag treatment did not significantly increase bone marrow reticulin or collagen formation in a 2-year longitudinal study in patients with ITP.³⁴ This was confirmed in the EXTEND study, in which patients with chronic ITP were treated with eltrombopag for ≤5.5 years.³⁵ In ESCALATE, no major impact on bone marrow fibrosis was noted.

Chromosomal abnormalities, especially monosomy 7, have been associated with an increased risk of progression to myelodysplastic syndrome or AML.² In the combined long-term data of 83 patients from 2 phase 2 nonrandomized studies in patients aged ≥2 years with refractory SAA, an evolution to an abnormal karyotype was seen in 16 (19%), most within 6 months of eltrombopag initiation. However, this cytogenetic evolution was not associated with genomic mutational status, and the mutated allele fractions of myeloid cancer genes did not increase on treatment with eltrombopag.²⁸ In our study, no chromosome 7 abnormalities arose, and cytogenetic abnormalities detected after eltrombopag initiation were not considered clinically significant. However, there was 1 case of AML. Additional studies on clonal evolution in pediatric patients with SAA treated with eltrombopag are warranted.

Eltrombopag starting doses were selected based on a population PK model from pediatric ITP data, adjusted for higher exposure in SAA. After starting treatment with the age-specific dose, patients had their dose assessed and modified every 2 weeks based on platelet counts until the maximum dose was achieved. The final dose after the completion of dose escalation aimed to provide an exposure equivalent to that in adult patients with SAA. Overall, eltrombopag clearance was lower, and plasma exposure was higher in patients aged 1 to <6 years than those aged 6 to <18 years. Specifically, younger patients tended to show higher exposure, whereas older pediatric patients exhibited similar exposure levels to adults. Given the small sample size and high variability, further investigation is warranted. Furthermore, there was no clear difference in PK parameters between the 2 cohorts.

The single-arm phase 2 study design of ESCALATE with no control arm is inherently limited. The small sample size also limits the statistical analysis of treatment efficacy. The R/R cohort included only 14 patients for whom prior treatment was unknown, and a subset of these patients did not receive hATG; hence, these data should be interpreted with caution. The absence of data on prior treatments for the R/R cohort makes it difficult to account for the potential influence of prior therapies on the outcome of eltrombopag. The high patient attrition rate posed a significant limitation. Comparing response rates in ESCALATE with previously published response rates is challenging due to variations in study designs, efficacy assessment criteria, and sample sizes.

In conclusion, this study demonstrated that hematologic parameters improved over time, and over half of the patients achieved transfusion independence. Eltrombopag plus IST showed a trend toward a favorable ORR in patients with R/R disease and revealed no new safety concerns. However, given the small sample size, lack of data regarding prior treatments in the R/R cohort, and the short follow-up period, further studies are needed to confirm these findings and optimize dosing strategies. Although the data reported are encouraging, individual patient factors and existing guidelines should be considered when deciding on this treatment regimen. Further studies in pediatric patients with SAA are needed.

The authors thank the patients and their families for participating in ESCALATE and all medical staff involved in the trial. The authors also thank Xun Li of Navigate for conducting the bone marrow assessments for this study. Medical writing support was provided by Helen Findlow of Novartis Pharmaceuticals UK Ltd (London, United Kingdom), in accordance with the Good Publication Practice 2022 guidelines (https://www.ismpp.org/gpp-2022).

This study was funded by Novartis Pharma AG, Basel, Switzerland.

Contribution: All authors provided substantial contribution to the study design and/or collection, analysis, and/or interpretation of data; were involved in the drafting and/or critical reviewing of the manuscript; provided final approval of the version to be published; and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Conflict-of-interest disclosure: C.B. reports grants from Novartis and Sobi. S.S. reports personal fees from Pfizer. J.E.F. reports clinical trial support from Novartis. J.A.R. reports funding from Novartis for the conduct of the study. T.A.N. has been a consultant for Novartis. T.F.W., Q.W., and P.U. are employees of Novartis. B.S. serves on a Novartis steering committee and has received personal fees from Pfizer. A.S., W.W., and A.V. serve on a Novartis steering committee. D.A.W. serves on Novartis steering committee; serves on scientific advisory boards with Beam Therapeutics and Skyline Therapeutics (formerly Geneception); is a consultant for Tessera Therapeutics, Verve Therapeutics, and Vertex Pharmaceuticals; and reports research funding from ExCellThera. The remaining authors declare no competing financial interests.

The current affiliation for A.V. is Johns Hopkins All Children's Hospital, Cancer and Blood Disorders Institute, Johns Hopkins University, St. Petersburg, FL.

Correspondence: David A. Williams, Boston Children’s Hospital, Dana-Farber/Boston Children's Cancer and Blood Disorders Center, Harvard Medical School, 300 Longwood Ave, Karp 08125.3, Boston, MA 02115; email: dawilliams@childrens.harvard.edu.