Patients and data collection
Data were collected and reviewed from patients newly diagnosed with SAA in the period from February 2013 to March 2022 at Peking Union Medical College Hospital (PUMCH) or Beijing Children's Hospital affiliated with Capital Medical University who had been treated with standard IST alone or IST + EPAG. Those patients did not undergo allogeneic HSCT due to the absence of HLA-matched sibling donors, age or personal preference. Patients who met the following criteria were included in the final analysis: (1) newly diagnosed with acquired SAA; (2) having been treated with rabbit ATG with CsA or rabbit ATG with CsA and EPAG; and (3) had a last follow-up date of September 30th, 2023, with all surviving patients followed for at least 12 months before enrolment. Patients who were diagnosed with hereditary bone marrow failure syndrome, myelodysplastic syndrome (MDS) or flaer-negative granulocytes larger than 50% with uncontrolled infection or severe liver or kidney impairment before treatment were excluded. Paroxysmal nocturnal haemoglobinuria (PNH) clones were detected by flow cytometry, and a PNH clone was considered to be present when the percentage of flaer-negative granulocytes exceeded 1%.
This study was approved by the Ethics Committee of Peking Union Medical College Hospital (Ethics No. K5283) and Beijing Children's Hospital of Capital Medical University (Ethics No. 2019-k-276). Informed consent was waived for retrospective studies.
Clinical data, including sex, age, baseline assessments (such as blood cell count, reticulocyte count, liver and kidney function, serum ferritin (SF), PNH clone, bone marrow smear and trephine biopsy, chromosomal karyotype and genetic mutations), time from diagnosis to ATG treatment, and treatment regimens, including the ATG dosage, CsA dose and trough concentration, EPAG dosage and duration, were recorded.
Criteria for diagnosis and response
The diagnosis of SAA and haematological response assessment were performed according to the Guidelines for the Diagnosis and Treatment of Aplastic Anaemia (2016 edition) [17]. For patients younger than 50 years, chromosome breakage tests, comet tests, telomere length tests, cytogenetic analyses and myeloid gene sequencing were conducted to exclude congenital bone marrow failure disorders. SAA was defined as a blood count that met at least two of the following three criteria: absolute neutrophil count (ANC) < 0.5×109/L, platelet count (PLT) < 20×109/L, and reticulocyte count (RET) < 20×109/L. Cases with ANC < 0.2×109/L were defined as very severe aplastic anaemia (VSAA).
The haematologic response was evaluated, including the overall response rate (ORR) and complete response rate (CRR), at 3, 6 and 12 months after standard IST or IST + EPAG treatments and at the end of follow-up. The definition of response was as follows: (1) complete response (CR) was defined as a haemoglobin (Hb) level reaching the normal range for that sex and age group, ANC > 1.0×109/L, and PLT > 100×109/L; (2) partial response (PR) was defined as transfusion independence and no longer meeting the SAA diagnostic criteria; (3) no response (NR) was defined as still meeting the diagnostic criteria for SAA; and (4) overall response (OR) was defined as achieving at least a PR (PR + CR) [17].
Possible predictive factors for the ORR and CRR at 6 months, such as sex, age, time from diagnosis to treatment (months), baseline neutrophil count, lymphocyte count, haemoglobin, platelet count, reticulocyte count, SAA or VSAA, the presence of PNH clones, the presence of chromosomal karyotype abnormalities, the presence of myeloid mutations and the use of EPAG, were analysed.
Clonal evolution, overall survival (OS) and event-free survival (EFS) were also documented.
Overall survival (OS): OS was calculated from the first day of IST treatment until the date of death or the last known survival.
Event-free survival (EFS) was defined as the time interval from the start of IST ± EPAG treatment to the occurrence of an event (death, relapse, clonal evolution, haematopoietic stem cell transplantation, or loss to follow-up).
Relapse was defined as the recurrence of lower peripheral blood cell counts after effective IST therapy which required second-line therapy, such as adding EPAG to those with IST alone or HSCT for those with IST + EPAG.
PNH clonal evolution was defined as the absence of PNH clones with subsequent presentation of PNH clones or as the expansion of PNH clones during follow-up.
Clonal evolution, which is the progression to myeloid cancer, was defined as the progression to myelodysplastic syndrome (MDS) or acute myeloid leukaemia (AML) [18].
Possible influencing factors for OR, CR, OS and EFS, such as sex, age, time from diagnosis to treatment (months), baseline neutrophil count, lymphocyte count, haemoglobin, platelet count, reticulocyte count, SAA or VSAA, the presence of PNH clones, the presence of chromosomal karyotype abnormalities, the presence of myeloid mutations and the use of EPAG, were analysed.
Treatment regimen
IST ATG: rabbit ATG (rATG, Sanofi, France) 3–5 mg/(kg·d) for 5 days. Methylprednisolone was given to prevent serum sickness. CsA was started at 3–5 mg/kg twice a day, and the plasma trough concentration was maintained at 100–200 ng/ml. After the optimal efficacy was reached, the drug was continued for at least 6 months before the dose was tapered. The total duration of treatment was 2–3 years (Fig. 1).
TPO-RAs: Adult group: EPAG at a dosage of 100–150 mg qd. Children: EPAG at a dosage of 2.5 mg/kg qd for those aged 2–5 years, 50–75 mg qd for those aged 6–12 years, and 100 mg qd for those > 12 years (Fig. 1).
Supportive therapy: Patients were given anti-infective treatment, transfusion, and granulocyte colony-stimulating factor if needed. No other concomitant medications were given.
Statistical analysis
SPSS 24.0 statistical software (IBM SPSS, Somers, NY, USA) was used. Nonnormally distributed data are presented as medians (interquartile ranges, IQRs), whereas count data are expressed as cases (%). Differences between groups were evaluated via chi-square tests and Kruskal‒Wallis rank‒sum tests for categorical variables. Chi-square tests, Kruskal‒Wallis rank-sum tests and logistic regression were used to identify the factors that influenced treatment efficacy, and factors with P < 0.2 in the univariate analysis were included in the multivariate regression analysis. The Kaplan‒Meier method was used to estimate OS and EFS, and the OS and EFS of subgroups were compared via the log-rank test. Multivariable Cox proportional hazard regression models were used to analyse the factors influencing both OS and EFS. All the statistical tests were two-sided, and P values of < 0.05 were considered significant.