BACKGROUND AND OBJECTIVES
The usefulness of the triple combination of horse anti‐human thymocyte immunoglobulin (hATG) + ciclosporin (CsA) + eltrombopag (EPAG) in patients with severe aplastic anaemia (SAA) was reported in the phase III RACE study.1 The NIH (National Institutes of Health) group also reported that approximately 90% of patients with SAA responded when EPAG was used in combination with hATG + CsA therapy.2 The risk of chromosomal abnormalities did not differ between the EPAG‐combination group and the historical controls. Based on the results of these clinical studies, triple therapy with anti‐human thymocyte immunoglobulin (ATG) + CsA + EPAG is recommended as the first‐line treatment for transfusion‐dependent patients with moderate to SAA who are ineligible for haematopoietic stem cell transplantation (HSCT) in Japan.3
Rabbit anti‐human thymocyte immunoglobulin (rATG) was the only ATG approved in Japan until recently. It is unclear whether rATG + CsA + EPAG therapy would have the same efficacy as hATG because rATG is less effective against aplastic anaemia (AA) than hATG.4, 5 Additionally, because rATG may exert a stronger immunosuppressive effect than hATG, the incidence of infectious complications and chromosomal abnormalities with rATG + CsA + EPAG may differ from that observed with hATG.
Therefore, to address these clinical concerns in Japan, we prospectively evaluated the efficacy and safety of rATG + CsA + EPAG in adult patients with transfusion‐dependent non‐SAA or SAA.
PATIENTS AND METHODS
Patients
Adult patients with non‐SAA who required transfusions or SAA6 and had no prior treatment with CsA, EPAG, anabolic steroids or allogeneic HSCT, who were registered in the W‐JHS AA02 study (trial registration number: jRCTs071190032, 21 October 2019) and who were managed by 29 institutions in Japan, including the West Japan Hematology Study Group (W‐JHS) from October 2019 to September 2022 were included in this study.
Other eligibility criteria were 18–80 years of age and satisfying at least two of the following: for non‐SAA patients requiring transfusions—neutrophil count <1 × 109/L, platelet count <50 × 109/L and reticulocyte count <60 × 109/L; and for SAA patients—neutrophil count <0.5 × 109/L, platelet count <20 × 109/L and reticulocyte count <40 × 109/L. In addition, the percentage of cellular components should be <25% as judged by a bone marrow biopsy, without findings of fibrosis or haematopoietic malignancy, and the Eastern Cooperative Oncology Group performance status should be 0–2. Patients with morphologically diagnosed myelodysplastic syndromes (MDS) or those with chromosomal abnormalities common to MDS, as defined in the WHO 2017 (revised fourth edition) diagnostic criteria, were excluded.
However, cases with isolated del(13q) were included in this study.
Blood samples for a biomarker analysis were collected in the morning on the day that therapy was initiated. Paroxysmal nocturnal haemoglobinuria (PNH)‐type cells were detected using high‐sensitivity flow cytometry (SRL, Inc., Tokyo, Japan)7, 8 and plasma thrombopoietin levels were measured using an enzyme‐linked immunosorbent assay (SRL, Inc.).9 Thrombopoietin ≥320 pg/mL was considered high based on the previous study.10 HLA class I allele‐lacking cells were analysed by flow cytometry, as previously described.11
Treatment
rATG was infused at a dosage of 2.5 mg/kg/day for 5 days, followed by CsA (5 mg/kg/day) divided into twice‐daily (oral) doses, and EPAG (75 mg/day) was administered from day 6. rATG dosage was determined based on previous randomized study.12 CsA concentration was measured using the chemiluminescence immunoassay both before administration (trough) and 2 h (peak) after administration. The dosage of CsA was adjusted to maintain a concentration of 600–900 ng/mL, 2 h after oral administration to maximize efficacy,13 while the trough level was kept below 250 ng/mL to minimize the risk of renal dysfunction. EPAG dosing and adjustment are available in Supporting Information—Methods.
If the blood trough level exceeded 250 ng/mL or the serum creatinine level was ≥150% of the baseline level, the CsA dose was reduced by 25%. Methylprednisolone was administered at 2 mg/kg/day on days 1–5; 1 mg/kg/day on day 6; and 0.5 mg/kg/day on days 8, 10, 12, 14, 16, 18 and 20.
For patients who responded at week 26, the EPAG and CsA were continued for at least 52 weeks, with doses adjusted at the institution's discretion. In contrast, for patients who did not respond, treatment after 27 weeks was not specified.
Efficacy assessment
The primary end‐point was the haematological response rate (HRR) 12 weeks after the start of the study treatment (hereafter referred to as week 12), based on the Camitta criteria14 (Table S1). Secondary end‐points included: (1) HRR at week 26 and 52; (2) duration of response and recurrence rate in patients with a haematological response (HR); and (3) relationships between the HR and PNH clones, HLA class I allele‐lacking cells or plasma thrombopoietin levels.
Safety evaluation
All adverse events (AEs) after the start of treatment were recorded using the National Cancer Institute Common Terminology Criteria for Adverse Events (v5.0; NCI‐CTCAE; Japanese translation Japan Clinical Oncology Group/Japan Society of Clinical Oncology (JCOG/JSCO) version). AEs and progression to MDS/acute myeloid leukaemia (AML) were also evaluated. AEs were monitored every 2 weeks until week 12 and thereafter every 4 weeks until week 52.
Chromosome and somatic mutations
Bone marrow examination and cytogenetic analysis were performed at baseline and week 26 and 52. Targeted capture sequencing of peripheral blood samples was conducted using a panel of common driver genes implicated in MDS15 at the beginning of treatment and at week 12, 26 and 52. The presence and development of the mutant clones were evaluated.
Statistical analysis
The target number of patients was set to be 20% higher than the response rate at week 26 in the NIH study group, based on the treatment outcomes of rATG + CsA therapy, which showed a response rate of 37% at week 12.16 To establish thresholds and expectations for the primary end‐point, we estimated the response rates to be 37% and 57% respectively. Assuming a one‐sided significance level of 2.5% and statistical power of ≥80%, the required sample size, calculated based on the exact binomial distribution, was at least 54 patients. Considering a 10% ineligibility rate, 60 patients were enrolled in the final study.
Efficacy was evaluated using the full analysis set (FAS) of all eligible patients. The primary end‐point, the HRR at week 12, was analysed using an exact method based on binomial distribution, with a null hypothesis of a 37% threshold. The significance level was set at 2.5% on one side.
For secondary end‐points, 95% confidence intervals (CIs) were estimated as needed. FAS was used to analyse chromosomal abnormalities and somatic gene mutation clones. All statistical analyses were performed using SAS (ver.
9.4, SAS Institute, Cary, NC, USA).
Research logistics and ethical regulations
Central monitoring was conducted based on the data collected from the case report forms (Viedoc 4). All investigators adhered to the principles of the Declaration of Helsinki, Ethical Guidelines for Human Genome and Gene Analysis Research (28 February 2017), Clinical Research Act No. 16 (2017) and Enforcement Regulations of the Clinical Research Act.
This study was conducted in accordance with the latest version of the Japanese Ministry of Health, Labour and Welfare Ordinance No. 17 of 2018 and related notices.
The principal investigator obtained approval from the medical institutions' administrators for conducting the study, following a research protocol approved by the Clinical Research Network Fukuoka Certified Review Board (CRB7180004) in accordance with each institution's regulations. Prior to enrolment, all patients provided their written informed consent.
RESULTS
Patients
Sixty patients were enrolled in this study. One patient's diagnosis was changed to MDS after registration, resulting in an FAS population of 59, while the safety analysis population remained at 60. Table 1 shows the patient characteristics.
The median age was 61.5 (range, 18–77) years. Forty‐eight patients (80.0%) had SAA. None of the patients <40 years old had a matched sibling donor.
Twenty‐seven patients (45.0%) discontinued treatment. The reasons for discontinuation included HSCT or the addition of another therapeutic agent owing to insufficient efficacy (n = 11), failure to achieve a response at week 26 (n = 11), the attending physician's judgement that discontinuation was appropriate (n = 3) and other reasons (n = 2). The median time to discontinuation was 24.4 (range 0.4–45.1) weeks.
Haematological response rate (HRR)
The primary end‐point, the HRR at week 12, was 52.6% (95% CI, 39.0–66.0%) (p = 0.012) (Table 2), which was lower than the expected overall response rate of 57%. Therefore, this study did not meet its primary end‐point. Two patients were excluded from the analysis due to early mortality and the development of MDS.
Three (5.3%) patients achieved a complete response (CR) and 27 (47.4%) achieved a partial response (PR). At week 26, the HRR increased to 67.9% (95% CI, 54.0–79.7). The HRR in non‐SAA patients was 63.6% at week 12 and 72.7% at week 26, whereas in SAA patients, it was 50.0% and 66.7% respectively.
Platelet and reticulocyte counts began to increase after week 6, followed by an increase in neutrophil count after week 8 (Figure S1).
There was no significant association between the HR at week 12 or 26 and factors such as age, sex, disease severity (non‐severe versus severe) or baseline platelet, reticulocyte or neutrophil counts in the univariate analysis (Table S2). In addition, the HRRs at week 12 were 51.1% in patients with PNH clones at baseline and 60.0% in those without (p = 0.733), whereas at week 26, they were 68.2% and 60.0% (p = 0.715) respectively. No significant association was observed between HRs at week 12 and 26 and HLA class I allele‐lacking blood cells or elevated thrombopoietin levels (Table S3).
Long‐term outcomes
Of the 38 patients who achieved a HR at week 26, 37 continued treatment until week 52. Among 37 patients who continued treatment, the HRR at week 52 was 91.9%, with 10 patients achieving a CR and 24 patients achieving a PR. Six patients with PR at week 26 achieved CR by week 52 following continuation of EPAG.
On the other hand, three patients lost their response at week 52 despite continuous EPAG and CsA administration. In addition, the percentage of patients who achieved a response duration of 1 year, calculated using the Kaplan–Meier method, was 81.7% (95% CI, 62.0–91.8) (Figure 1). The dosage of EPAG and CsA after week 26 and treatment after discontinuation of study were available in Supporting Information—Results.
Safety
All 59 patients received rATG (2.5 mg/kg/day) for 5 days, except for one patient who discontinued treatment due to grade 3 alanine aminotransferase elevation. In total, 259 AEs were reported in 51 patients (85.0%) (Table 3). Thirty‐six patients (60.0%) experienced grade 3/4 AEs.
Grade 3/4 AEs occurring in >5% of the patients included febrile neutropenia (20.0%), alanine aminotransferase increased (13.3%) and hypertension (8.3%). Three patients developed grade ≥3 toxicity related to EPAG: liver function test (LFT, n = 2) and atrial fibrillation (n = 1); EPAG was temporarily discontinued in the patients with LFT abnormalities and reinitiated at the same dose after LFT values improved; none permanently discontinued EPAG. Fifty‐six patients achieved the target CsA concentration.
An increase in creatinine of all grades was observed in 22 patients, including one patient with a grade 3 elevation. Eight serious AEs were reported in six patients. One patient died of sepsis.
The patient was a 76‐year‐old man with almost no neutrophils in his peripheral blood. After immunosuppressive therapy (IST), his immunodeficiency worsened and he died on day 8. Grade 4 serious AEs included CPK increased, decreased neutrophil count, AML and MDS.
One grade 3 serious AE (anaemia) and one grade 2 AE (blood bilirubin and creatinine increased) were reported.
The incidence of MDS/AML was 3.3% (2/60). One case that was registered as having AA at the initial enrolment was later reclassified as MDS before progressing to AML. Neither of the patients showed chromosomal abnormalities at baseline; however, somatic mutations were detected.
The AML patient had TP53 biallelic mutations, whereas the MDS patient had U2AF1 mutations along with 1p loss, 2q loss and 15qUPD. Due to disease progression, the study was discontinued in these patients at week 12 and 30.
Chromosomal abnormalities and somatic mutations
A chromosomal analysis was performed in 59 patients. Five patients (8.5%) had chromosomal abnormalities at baseline, including del(13q) (n = 2), −Y (n = 2) and del(20q) (n = 1) (Table S4). Two had additional abnormalities at week 26: +15 (n = 1) and +Y (n = 1), and one patient exhibited expansion of the del(13q) clone at week 26.
Among patients without chromosomal abnormalities at baseline, chromosomal abnormalities were detected in five patients at week 26 and three Ypatients at week 52: del(13q) (n = 6), inv(3) (p21q27) (n = 1), and −Y, +Y (n = 1). Notably, two patients with del(13q) no longer exhibited del(13q) at week 52.
Somatic mutations were observed at baseline in 23 patients (39.0%) (Figure S2). Common somatic mutations included 6pUPD (n = 5); mutations in U2AF1 (n = 4); ASXL1, DNMT3A and JARID2 (n = 3, each); and BCORL1, SH2B3 and PIGA (n = 2, each). Nine had two or more somatic mutations.
Mutated clones present at baseline expanded in 12 patients (57.1%), whereas new mutated clones emerged after treatment in 37 patients (33.9%). There was no association between the presence of somatic mutations at baseline and HRs at week 12 and 26. In particular, there was no correlation between the presence of PIGA, BCOR or BCORL1 mutations and HRs at week 26 (p = 0.654).
DISCUSSION
In this prospective study, we assessed the efficacy of rATG combined with CsA and EPAG in newly diagnosed transfusion‐dependent patients with AA. The overall HRR at week 12 was significantly higher than that observed in a retrospective study of rATG and CsA alone.16 A sustained response at week 52 was observed in 91.9% of the responders at week 26. This study confirmed that the addition of EPAG enhanced the response rate of IST to rATG.17, 18
According to previous studies, patients treated with rATG demonstrated a slower response than those treated with hATG, with the highest response typically observed 12 months posttreatment.19, 20 In this study, the overall response rate increased over time, reaching 50% at week 12 and 66.7% at week 26 in the patients with SAA. Notably, the response rate at week 26 was comparable to that reported in the RACE trial.1 However, direct comparisons are difficult because the EPAG was initiated on day 6 in this study, whereas it was started on day 14 in the RACE trial. The concurrent addition of EPAG to IST may further improve the response rate in patients treated with rATG, similar to the benefits observed with hATG.2
Clonal evolution after EPAG treatment remains a significant concern. Patel et al. reported no significant difference in the incidence of high‐risk clonal evolution, including abnormalities of chromosome 7, between newly diagnosed patients treated with hATG plus CsA with EPAG and a historical cohort treated without EPAG (5.7% vs. 10.3%).21 However, the median time for clonal evolution was notably shorter in patients treated with EPAG than in those not treated with EPAG (186 vs.
777 days). In contrast, in the RACE trial, monosomy 7 was observed at 6 months in one patient treated without EPAG, whereas no cases emerged in patients treated with EPAG.1 In the present study, no abnormalities of chromosome 7 were observed within 52 weeks. However, two patients (3.3%) developed MDS or AML within 35 weeks of treatment.
These patients had no chromosomal abnormalities before treatment but were later found to carry biallelic TP53 or U2AF1 mutations. Given the early development of MDS (week 30) and AML (week 11) after the initiation of treatment, it is possible that these patients had undiagnosed myeloid malignancies rather than AA at the time of treatment initiation. Indeed, the patient who developed AML had mild fibrosis and aggregation of megakaryocytes in the bone marrow and showed a low plasma thrombopoietin level and an elevated immature platelet fraction at the time of the diagnosis.
The final diagnosis was later revised to MDS. Nevertheless, the addition of EPAG may have accelerated the progression to MDS or AML. Therefore, precise evaluation of the underlying disease is essential when treating patients with an EPAG‐inclusive regimen.
The deletion of 13q was the most common cytogenetic abnormality that emerged following treatment in this study. A similar finding was observed in a phase 2 EPAG study conducted by NIH.21 PNH‐type cells were detected in all patients with del(13q) at the time of the diagnosis (data not shown), consistent with a previous report.22 Since the del(13q) abnormality is observed in wild‐type cells, but not in PNH‐type cells,22 the expansion of the del(13q) clone may result from both persistent immune attack, which allows PNH clones to expand after IST, and stimulation of haematopoietic stem cells with del(13q) by EPAG.23
The presence of PNH‐type cells or HLA class I allele‐lacking leucocytes is generally associated with a favourable response to IST, as these markers reflect the immune mechanisms underlying AA, allowing haematopoietic stem cells to evade immune attack.7, 11, 24, 25, 26 However, these immune markers did not predict a good response. Similarly, a reticulocyte count >10 × 109/L was not associated with an improved response rate.27 The diminished predictive value in the present study was primarily attributed to the increased response rate in patients who tested negative for these immune markers. This finding is consistent with previous observations that EPAG alone can stimulate haematopoiesis in approximately 40%–50% of patients with SAAs refractory to IST.23 The increased response rate in patients whose bone marrow failure was not clearly immune‐mediated may explain the overall improved response rates in the broader AA patient population.
This study was associated with some limitations. As this was a single‐arm study, the efficacy of rATG plus CsA in combination with EPAG had to be compared with that of a previous retrospective study using rATG plus CsA alone. A randomized trial is required to confirm the improved efficacy of triple combination therapy.
In conclusion, this prospective study demonstrated that the combination of rATG and CsA with EPAG achieved faster and higher HRRs than rATG and CsA alone, without inducing high‐risk chromosomal abnormalities, regardless of the presence of known immune markers.
AUTHOR CONTRIBUTIONS
KI, YO, KU, HK, KT and SN designed the study; KI and SN collected clinical data; FN, KI, SO, KM TT, KH, YZ and SN analysed and interpreted the data; FN, KI, and SN wrote the manuscript; and all authors reviewed and approved the final manuscript.
FUNDING INFORMATION
The present study was conducted with funding from Novartis International AG. to the Kyushu Clinical Research Support Center (CReS Kyushu), based on a Clinical Research Agreement. The funders had no role in the writing of the manuscript or in the decision to submit the manuscript for publication.
CONFLICT OF INTEREST STATEMENT
FN: Nippon Shinyaku, Sanofi, Genmab, Chugai: Honoraria. KI: Bristol Myers Squibb: Honoraria; Alexion: Consultancy. RSu: Chugai, Kyowa Kirin, AbbVie, Bristol Myers Squibb, Eisai, Otsuka, MSD, Janssen, Takeda, Meiji Seika Pharma, Novartis, AstraZeneca: Honoraria; Chugai, Kyowa Kirin, Shionogi, Taiho, Eisai, Otsuka: Research funding.
YO: Kissei, Sanofi, IQVIA, Nippon Shinyaku, Symbio, Daiichi Sankyo, Astellas, Amgen, MSD, Kyowa Kirin, Chugai, Asahi Kasei Pharma, Janssen: Honoraria; Novartis, Pfizer, Bristol Myers Squibb: Honoraria, Research funding; Incyte, Meiji Seika, Shionogi, Sumitomo Pharma, AbbVie, JCR Pharma: Research Funding. KU: Agios, Alexion, Aperis, Asahi Kasei Pharma, Astellas, Bristol Myers Squibb, Chugai, Daiichi Sankyo, Eisai, Gilead, Incyte, Kyowa Kirin, Nippon Shinyaku, Novartis, Ohara, Ono, Otsuka, Pharma Essentia, Sanofi, Takeda: Research funding; AbbVie, Amgen, Asahi Kasei Pharma, AstraZeneca, Bristol Myers Squibb, Chugai, Eisai, Genmab, Incyte, Kyowa Kirin, Nippon Shinyaku, Novartis, Ono, Pfizer, Sandoz, Takeda: an entity's board of directors, speaker's bureau, and its advisory committees. HK: Novartis, Honoraria.
RSa: Nippon Shinyaku: Research Funding. SO: Eisai, Chordia Therapeutics, Montage Bio.: Consultancy; Asahi Genomics: Ownership interest; Eisai, Chordia Therapeutics: Research Funding. KM: AbbVie, Bristol Myers Squibb, Chugai: Honoraria, Research Funding; Sumitomo, Taiho, Kyowa Kirin: Research Funding; Otsuka: Consultancy, Honoraria, Research Funding; Asaki Kasei Pharma, Astellas, AstraZeneca, Eisai, Janssen, Novartis, Pfizer, Sanofi: Honoraria.
KH: GenoDive: Consultancy, Research Funding. IM: Chugai, AbbVie, Otsuka, Ono, Novartis: Honoraria, Research Funding; Kyowa Kirin, Sumitomo Pharma, Asahi Kasei pharma, Eisai, Taiho, Bristol Myers Squibb: Honoraria; Genmab, Incyte: Research Funding; Pfizer, Astellas, Janssen, Takeda, Ono, SymBio, Sanofi, AstraZeneca: Honoraria. KA: Asahi Kasei Pharma, AbbVie, Kyowa Kirin, Chugai, Bristol Myers Squibb, Janssen: Honoraria; AbbVie, Kyowa Kirin: Research Funding.
SN: Alexion, SymBio, Sobi, Asahi Kasei Pharma, Pfizer, Sanofi, Novartis, Kyowa Kirin: Honoraria. The remaining author declares no competing financial interests.
PATIENT CONSENT STATEMENT
All patients signed written informed consent forms.
Supporting information