Long-Term Follow-Up of Eltrombopag Treatment for Patients with Cyclosporin A Refractory/Relapsed Transfusion-Dependent Non-Severe Aplastic Anemia: A Report from a Single Center in China

Aplastic anemia (AA) is a bone marrow hematopoietic disorder characterized by immune-mediated peripheral blood cytopenia and bone marrow hypoproliferation. It is caused by abnormal activation of cytotoxic T cells that produce type 1 cytokines through the induction of apoptosis via Fas and Fas ligands [1‒4]. AA is classified as severe AA (SAA) and non-severe AA (NSAA) according to the Camitta and Bacigalupo criteria [5, 6]. In SAA, patients with an absolute neutrophil count (ANC) of <0.2 × 10⁹/L are considered to have very severe AA (VSAA). Allogeneic hematopoietic stem cell transplantation (allo-HSCT) from human leukocyte antigen-identical donors is the preferred treatment for patients with SAA/VSAA aged <40 years. According to the results of the randomized phase 3 RACE trial, eltrombopag (ELT) combined with immunosuppressive therapy (IST) is a new standard treatment for patients with newly diagnosed SAA who are not eligible for allo-HSCT [7]. For transfusion-dependent non-severe aplastic anemia (TD-NSAA), similar treatment regimens as those for SAA are recommended as first-line treatment according to the guidelines published by the UK Hematology Standards Committee [8].

ELT is an orally bioavailable small-molecule non-peptide TPO receptor agonist (TPO-RA) that binds to the transmembrane region outside the c-MPL ligand-binding pocket to stimulate hematopoiesis by non-competitive activation of c-MPL through JAK-STAT-activated signaling. This leads to the proliferation and differentiation of bone marrow progenitor cells into megakaryocytes and promotes platelet production [9]. Olnes et al. [10] demonstrated for the first time that ELT was effective in patients with refractory AA, with an efficiency of 44%. In a prospective study by Townsley et al. [11], 92 patients with primary SAA were divided into three cohorts and administered IST in combination with ELT. The overall response rates (ORRs) at 6 months were 80%, 87%, and 94% in the three cohorts, respectively, and 66% in the overall cohort. At the median follow-up of 2 years, the survival rate was 97%. In a phase I-II clinical trial that enrolled 197 patients with SAA, the ORR at 6 months was 41% and 68% in the IST and IST + ELT groups, respectively. The addition of ELT to standard IST increases the rate, speed, and intensity of hematologic responses in previously untreated patients with SAA without additional toxic effects [12]. Therefore, ELT was approved in combination with IST as first-line treatment in patients with SAA or as salvage therapy for patients with AA who are refractory to IST.

Although ELT is widely used in refractory/relapsed or newly diagnosed AA, there is no consensus on the duration of ELT treatment. In addition, there are several reports on the relapse and clone evolution rates in patients with newly diagnosed or relapsed SAA; however, there are limited reports on the rates in patients with NSAA, especially in those with relapsed/refractory TD-NSAA. Therefore, we aimed to report the long-term effectiveness, recurrence, and clonal evolution rates of ELT for the treatment of patients with refractory/relapsed TD-NSAA.

Data from patients with CsA refractory/relapsed TD-NSAA who had been treated with ELT for at least 3 months at Peking Union Medical College Hospital (PUMCH) from January 2018 to September 2021 were retrospectively collected. The diagnosis of AA was based on the 1975 Camitta criteria [5], and the severity classification was based on the Camitta [5] and Bacigalupo [6] criteria; SAA: marrow cellularity <25% (or 25–50% with <30% residual hematopoietic cells), plus at least 2 of (i) neutrophils <0.5 × 10⁹/L, (ii) platelets <20 × 10⁹/L, (iii) reticulocyte count <20 × 10⁹/L; VSAA: same as for SAA but neutrophils <0.2 × 10⁹/L; NSAA: AA not fulfilling the criteria for SAA or VSAA. Referring to the criteria of TD-NSAA proposed by the German AA Study Group and the European Working Group on Blood and Marrow Transplantation [13], the diagnostic criteria of TD-NSAA in this study were meeting the NSAA criteria of Camitta and Bacigalupo, and the peripheral blood cell counts met at least one of the following: PLT <20 × 10⁹/L, HGB <80 g/L, or ANC <0.5 × 10⁹/L. Baseline characteristics including age, sex, disease duration, treatment prior to ELT use, disease status at the time of ELT treatment, baseline laboratory parameters, and follow-up data after ELT treatment were recorded. Refractory patients were defined as those requiring transfusion despite no hematologic response after at least 6 months of standard CsA therapy. Relapsed patients were defined as those who responded to CsA but relapsed and did not respond again. Patients who met the criteria of TD-NSAA, those who did not receive HSCT or were not HSCT candidates, those who had adequate initial liver and kidney function before ELT treatment, those who had been treated with ELT for at least 3 months if not responding, and those who had been followed up for at least 12 months after ELT treatment were enrolled in the final analysis. Patients included in the analysis were taking only ELT without other drugs. In China, most patients with TD-NSAA often refuse ATG for financial reasons, so all the patients included in the analysis had prior IST with cyclosporine or cyclosporine combined with stanozolol, and none of these patients had ELT as the first-line therapy. Patients who were diagnosed with congenital AA (e.g., Fanconi anemia and congenital dyskeratosis) by chromosomal and molecular studies, those who had paroxysmal nocturnal hemoglobinuria (PNH) granulocyte clone size ≥50% determined by flow cytometry, those who had evidence of a clonal hematologic bone marrow disorders on cytogenetics, and those who had any concomitant malignancies before ELT treatment were excluded.

This study protocol was reviewed and approved by the Ethics Committee of PUMCH, Approval No. I-23PJ2051. Written informed consent was obtained from patients before data collection. For enrolled patients, a complete workup including regular blood tests and bone marrow smears, biopsies, chromosomal examinations, myeloid-related gene examinations, tests for inherited bone marrow failure syndromes in some patients, and PNH clone detection were conducted to exclude other causes for cytopenia prior to administering ELT. Bone marrow smears, biopsies, and PNH clone detection were routinely performed every 6 months during follow-up.

ELT was initiated at 75 mg qd for 2 weeks and then increased by 25 mg every 2 weeks until a maximum of 150 mg (if tolerated); after the optimal response was achieved in patients who responded, the dose was tapered gradually. The ELT dose in patients who relapsed after ELT tapering was increased to the maximum tolerated dose; such patients were considered to have relapsed if they did not respond again. No concomitant treatments, except transfusions, were administered to the selected patients. The ELT dose was increased to the previous dose in patients who relapsed after ELT tapering; if they did not respond, the dose was increased to a maximum of 150 mg qd (if tolerated).

Baseline information, including the clinical and laboratory data, was recorded. Data on symptoms (including adverse effects), laboratory assessments (including complete blood count with differential and serum chemistry profile and electrocardiography) at 3, 6, and 12 months after ELT treatment and at the end of follow-up were collected from patient files and occasionally from local hospitals. Red blood cell or PLT transfusions were performed according to the guidelines, and the frequency of transfusions was recorded. NSAA efficacy evaluation criteria were based on the guidelines of the British Standards Institute [8], including complete response (CR), partial response (PR) and no response (NR). CR was defined as: hemoglobin concentration normal for age and gender, neutrophil count >1.5 × 10⁹/L, platelet count >150 × 10⁹/L. PR was defined as: Transfusion independence (if previously dependent) or doubling or normalization of at least 1 cell line or increase of baseline, hemoglobin concentration of >30 g/L (if initially <60), neutrophils of >0.5 × 10⁹/L (if initially <0.5), platelets of >20 × 10⁹/L (if initially <20). NR was defined as a decreased blood count or failure to meet the CR and PR criteria above. The overall response was defined as CR + PR. Death within 3 months of treatment initiation was defined as early death and considered NR. Transfusion independence was defined as not being on PLT or red blood cell transfusions for 8 consecutive weeks. Relapse was defined as a substantial or progressive decline in blood counts or a substantial decline requiring transfusion or reinitiation of AA treatment [12, 14]. Adverse events (AEs) were recorded and classified based on the Common Terminology Criteria for Adverse Events version 5.0.

SPSS (version 26.0) was used for data description and statistical analyses. Data conforming to a normal distribution are expressed as mean ± standard deviation. The t-test or analysis of variance was used for comparison between groups. Data that did not conform to a normal distribution are expressed as median (range Min-Max). Non-parametric tests were used for comparisons between groups. Count data are expressed as percentages. Univariate and multivariate logistic regression analyses were performed to explore the influencing factors of response and relapse. Variables with p < 0.05 in the univariate analysis were included in the multivariate regression analysis (stepwise, bidirectional). Univariate and multivariate analyses of curative effect predictors were based on the effective number of responders with the best curative effect during the follow-up period. Statistical significance was set at p < 0.05. The factors considered in the analysis of predictors for response and relapse encompassed age at diagnosis, age at initiation of ELT, duration from diagnosis to initiation of ELT, gender, refractoriness/relapse status, cumulative ELT dose, average daily dose of ELT, baseline Hb level, white blood cell count, ANC, lymphocyte count, platelet count, reticulocyte count, presence of PNH clone, and serum ferritin level.

Out of the 55 patients included, 24 (43.6%) had relapsed TD-NSAA, while 31 (56.4%) had refractory TD-NSAA. There were 20 (36.4%) men and 35 (63.6%) women. The median age at diagnosis was 46 (19–80) years. The median duration from diagnosis to ELT treatment was 48 (6–384) months. Before ELT treatment, 26 (47.3%) and 29 (52.8%) patients were treated with CsA alone and CsA + stanozolol, respectively. Eight (14.6%) patients had a median of 3 (1–23)% FLAER-negative neutrophils. Chromosomal abnormalities were found in four (7.3%) patients, including 2 patients with +8, 1 patient with 13q-, and 1 patient with −7. Genetic mutations were found in nine (16.4%) patients, including 5, 2, and 2 patients with mutations in PIGA, DNMT3A, and BCOR, respectively. Baseline patient characteristics are listed in Table 1.

At the median follow-up of 28 (12–48) months, the ORR was 38.2%, 60.0%, and 52.7%, at 3, 6, and 12 months after ELT treatment, respectively, and 53.8% at the end of the follow-up. The CR rate (CRR) was 9.1%, 14.6%, and 9.1% at 3, 6, and 12 months after ELT treatment, respectively, and 9.6% at the end of the follow-up. Figure 1 showed a bar graph of the number of lineage responses in 3, 2, and 1 Lineages Involving relapsed and refractory TD-NSAA during follow-up. At the end of the follow-up, the effective number of patients was 28 (53.8%), including 17 (32.7%) patients with refractory TD-NSAA and 11 (21.2%) patients with relapsed TD-NSAA. Among them, 14 patients (26.9%) had one-lineage response, 9 patients (17.3%) had two-lineage response, and 5 patients (9.6%) had three-lineage response. The ORR seemed higher in refractory TD-NSAA than in relapsed TD-NSAA 6 months after ELT, but the difference was not statistically significant (p = 0.513, Fig. 2a). During ELT treatment, the CRR for refractory TD-NSAA seemed higher than that of relapsed TD-NSAA, but the difference was not statistically significant (p = 0.181) (Fig. 2b).

The median time to treatment response was 3 (1–7) months. The median time to CRR was 4 (2–12) months. At the end of the follow-up, the median duration of ELT treatment was 12 (4–48) months, and the median ELT exposure was 24.0 (4.5–131.5) g.

AEs occurred in 21 of 55 (38.2%) patients, including five (9.1%) patients with gastrointestinal disorders, four (7.3%) patients with joint pain, three (5.5%) patients with hyperkalemia, three (5.5%) patients with pruritus, two (3.6%) patients with elevated creatinine levels, two (3.6%) patients with elevated liver enzyme levels, and two (3.6%) patients with elevated bilirubin levels. Most AEs were mild (grades I and II) and were resolved with symptomatic treatment. However, 1 patient discontinued ELT treatment due to severe elevation in the creatinine level. Another patient required a reduction in the ELT dose owing to severe gastrointestinal distress.

Analysis of factors that might influence ORR in responders and non-responders (Table 2) showed a statistically significant difference in ANC at ELT initiation (p = 0.029). Univariate logistic regression found that ANC at ELT initiation (odds ratio, 2.6; 95% confidence interval [CI], 1.2–7.7; p = 0.041) might be associated with ORR (Table 3), but no association was found between ANC at ELT initiation (odds ratio, 2.6; 95% CI, 1.1–8.6; p = 0.083) and ORR in multivariate logistic regression (Table 4). Therefore, no factors were found to be related to the ORR.

At the median follow-up of 28 (12–48) months, 21.2% (7/33) patients relapsed (Figure 3), and the median time from ELT treatment to relapse was 14 (6–45) months. Six (10.9%) clone evolutions were observed at the end of the follow-up. Three (5.5%) patients showed increased PNH clones (2 patients who did not have PNH clones at baseline presented with PNH clones of 1% and 4%, respectively, and 1 patient had increased PNH clones from 3 to 7.5%). One patient had normal karyotype 46,XX at baseline, and developed chromosomal abnormalities 46,XX, t (10; 11)(p12; q21). One patient with a normal karyotype of 46,XX at baseline and failed to respond to ELT developed low blast myelodysplastic syndrome (MDS-LB) at 12 months after ELT with a karyotype of 46,XX,-7.

At the median of 16 (13–22) months after ELT treatment, three (5.5%) patients who did not respond to ELT died from respiratory failure, heart failure, and pulmonary infection. Three patients who did not respond to or relapsed after ELT treatment underwent HSCT and achieved CR. Eleven patients switched to avatrombopag, and the remaining patients continued with blood transfusions.

When analyzing the characteristics of patients with and without relapse (Table 5), there were statistically significant differences in age at diagnosis (p = 0.017) and age at ELT initiation (p = 0.033). In univariate logistic regression, there might be a statistically significant difference between age at diagnosis (odds ratio, 1.1; 95% confidence interval [CI], 1.0–1.1; p = 0.038) and age at ELT initiation (odds ratio, 1.1; 95% CI, 1.0–1.2; p = 0.043) and relapse (Table 6). Further multivariate logistic regression analysis with age at diagnosis and age at ELT initiation did not identify any factors that could influence relapse (Table 7).

Several studies have demonstrated the efficacy of ELT in patients with newly diagnosed or refractory/relapsed AA. However, most studies have focused on patients with SAA, and the long-term outcomes of refractory/relapsed patients are rare. In a retrospective study of patients with TD-NSS+NSAA, 16% progressed to SAA within a median of 31 (4–162) months [15]. The mutational burden in TD-NSS+NSAA patients was less pronounced than in SAA, whereas cumulative mortality was similar between TD-NSS+NSAA patients and SAA [15]. Although TD-NSAA differs from SAA, it has a higher tendency to transform into SAA [16].

This study is not only an extension of our previous reports on refractory/relapsed AA [17] but also a study focusing on TD-NSAA with a relatively large population and long-term follow-up. Most patients with TD-NSAA in China are treated with CsA alone rather than with CsA + ATG/ALG. This is probably due to the relatively high price of ATG/ALG, potential injection-related AEs, and the unavailability of hospitalization. Because limited patients were treated with CsA + ATG/ALG, to decrease possible variations, we only selected those who were treated with CsA or CsA + stanozolol. Here, we reported 55 patients treated with ELT alone for at least 3 months and were followed up for at least 12 months. The ORR reported in our present study is similar to that reported in our previous study [17] and the study by Olnes et al. [10]. At the end of the follow-up period, 14 (25.5%) patients had one-lineage response, nine (16.4%) patients had two-lineage responses, and five (9.1%) patients had a three-lineage response, with a median time to response of 3 (1–7) months. All these results are consistent with those reported in our previous study and other studies [18‒20]. Meanwhile, tolerance to ELT was good, and no new AEs were found in our study compared to those reported in other studies [21, 22]. Zaimoku et al. [23] reported that for patients with SAA initially treated with IST + ELT, blood counts remained the best predictor, especially those with an absolute reticulocyte count between 10–30 × 10⁹/L. However, PLT count or the presence of PNH clones did not influence the response to ELT. The addition of ELT changed the natural course of SAA. However, our patient had refractory/relapsed TD-NSAA, which differed from the above patients. We did not find any factors that could predict the response to ELT. This finding is consistent with the results of our previous study [17].

Few studies have reported long-term follow-up data of refractory or newly diagnosed SAA. Desmond et al. [24] reported that for refractory SAA, with an ORR of 40%, most patients (14/17) remained on ELT for a median time of 12 months (6–37 months) and continued to show improvement; only 17.6% (3/17) patients relapsed. Notably, in this study, 5 patients with near normalization of blood counts discontinued the drug at a median of 28.5 months after initiation, and they all maintained stable blood counts at a median of 13 months after ELT treatment. For newly diagnosed SAA, at a median follow-up of 4 years, Patel et al. [14] reported a cumulative relapse rate of 39% in responsive patients and a clonal evolution rate of 15% in all treated patients. Relapse occurred at the time of ELT discontinuation at 6 months and 2 years after CsA discontinuation. Most patients with relapse were retreated with therapeutic doses of CSA +/− ELT, and two-thirds of them responded. These results showed that the inclusion of ELT improved the response rate and shortened the time to response, but ELT and IST did not prevent recurrence and provided an outline for SAA, either newly diagnosed or refractory. The study Fan et al. [25] included 31 patients with TD-NSAA + NSAA treated or untreated previously with IST, and showed a 50% response in at least one lineage, with good tolerance. At a median follow-up of 16 months for all patients and 27 months for responsive patients, most patients required reinitiation of ELT for declining blood counts, and all responded to ELT treatment.

Almost no studies have focused on the factors predicting relapse after ELT treatment. Our findings showed that age, sex and baseline cell count did not influence the relapse rate and no predictors of relapse were found. Patel et al. [14] observed a significantly higher risk of relapse in older patients with SAA and reported that no biomarkers other than age-predicted relapse. Our results differ from theirs because their study involved patients with newly diagnosed SAA.

Clonal evolution was observed in six (10.9%) patients at the end of the follow-up. Three (5.5%) patients exhibited an increased number of PNH clones. One patient developed a chromosomal abnormality at t (10,11) 26 months after ELT. Another patient progressed to MDS-LB >1 year after ELT treatment. At the end of the follow-up period, 3 patients who did not respond to ELT died from respiratory failure, heart failure, and pulmonary infection. Patel et al. [15] reported that in an MAA cohort that had not been exposed to TPO-RA, at a median follow-up period of 31 (4–162) months, 11% (9/85) patients progressed to clinically significant PNH requiring the initiation of eculizumab therapy and 1% (1/85) patients progressed to AML after IST. The adverse clone evolution rate was higher in our cohort, even with a shorter follow-up time, which might raise concerns regarding the impact of ELT. However, our cohort included refractory/relapsed patients who might have had worse outcomes and fewer PNH clones than the general patient population. Before the use of TPO-RA, the incidence of late clones in patients with acquired AA receiving IST was 6–10% [26‒28], consistent with our data.

This study has several limitations. First, this was a single-center retrospective study, and some patients treated with ELT were either lost to follow-up or had incomplete medical data and were, therefore, excluded from the study, which might have caused bias. Patients were treated with CsA or CsA + stanozolol before ELT; therefore, they were not fully immunosuppressed. The number of patients was limited and the follow-up duration was relatively short. Despite these limitations, our data provide long-term effectiveness, relapse, and clone evolution data for patients with TD-NSAA and may serve as a reference for future prospective studies.

The study complied with the ethical requirements of the World Medical Association Declaration of Helsinki (https://www.wma.net/policies-post/wma-declaration-of-helsinki-ethical-principles-for-medical-research-involving-human-subjects/). This study protocol was reviewed and approved by the Ethics Committee of Peking Union Medical College Hospital, Approval No. I-23PJ2051. Given the retrospective nature of the data collection, the study was exempt from the requirement for written informed consent from the Peking Union Medical College Hospital Ethics Committee.

The authors have no conflicts of interest to declare.

This study was supported by the National High Level Hospital Clinical Research Funding (2022-PUMCH-C-026, 2022-PUMCH-D-002, and 2022-PUMCH-B-046), the CAMS Innovation Fund for Medical Sciences (CIFMS 2021-I2M-1-003), and the Beijing Natural Science Foundation (7232109).

Qinglin Hu contributed to the collection of data, analysis, and writing. Yang Yuan was involved in data collection. Yang Chen and Chen Miao were involved in data collection and data interpretation. Han Bing participated in the design, data collection, and data interpretation.

The data that support the findings of this study are not publicly available due to privacy or ethical restrictions but are available from the corresponding author B. Han (E-mail: hanbingtg123@163.com) upon reasonable request.