In this study, we have demonstrated the efficacy of TPA-RAs without ATG for patients with AA in a real-world setting. The median age of the patients in this cohort was 76.5 years old, which is higher than that in previous prospective studies (32-53 years) [2–4, 11, 17–19], although in the one retrospective study we reviewed , the median age of patients naïve for ATG treatment was 74 years. The cohort in this study may therefore be more representative of the patients who would typically be naïve for ATG treatment in daily clinical practice. Of the 45 patients included in this study, 32 were ineligible for ATG treatment because of their age (≥ 70 years) and/or comorbidities, and the other 13 had only mild symptoms and were not indicated for ATG treatment according to the Japanese guideline . Hematologic response in at least one lineage was observed in 62% of patients at six months from the initiation of eltrombopag treatment, and an optimal response was observed in 84% of patients regardless of time. A trilineage response was observed in 25% of eligible patients at any point during the follow-up period. The median time to achieve hematological responses in at least one lineage and a trilineage response were 84 and 397 days, respectively.
Several prospective and retrospective studies have published data on the efficacy and safety of TPO-RA for newly diagnosed and refractory AA [2–4, 9, 11, 17–19]. In these studies, the response rates for at least unilineage and trilineage responses were approximately 40-80% and 20-40%, respectively.
In these studies, most patients had previously been treated with ATG, and one of these studies demonstrated that patients previously treated with ATG exhibited faster hematopoietic cell recovery (transfusion independence rate at three months: 44% versus 0%) and a slightly better hematological response rate than patients not treated with ATG (response rate in at least one lineage: 74% vs 64%; trilineage response rate: 34% vs 27%) . Although the numbers of patients are small and there are obvious caveats concerning attempts to compare independent studies, the hematologic response rates and speed of hematopoietic recovery in our study do not seem to be inferior to the results of previous studies, including those involving younger patients and those previously treated with ATG.
This study cohort included 10 patients with SAA. Of these, six were over 75 years old, and the response rate for at least unilineage and trilineage responses in this group was 66% and 33%, respectively. Immunosuppressive therapy is effective regardless of the patient’s age; however, the mortality rate is likely to be higher in older patients than younger patients [20, 21]. Therefore, clinicians should take extra care when administering ATG to elderly patients. However, previous case reports have demonstrated the efficacy of eltrombopag without ATG for treating SAA [22, 23], and findings in our study were consistent with these reports even in elderly patients who are ineligible for ATG treatment. These findings suggest that TPO-RA should be administered even to elderly SAA patients who are ineligible for ATG.
The safety profile of the eltrombopag treatment in this study was consistent with previous studies [2, 4, 11]. The most frequent adverse event was anorexia and nausea. These events were resolved within two to four weeks of discontinuation or reduction of eltrombopag administration. Although the adverse events did not recur in most of the patients when the eltrombopag was reinitiated, they did in three female patients, who were then switched to romiplostim. After the switch, their symptoms were completely resolved. Most of the adverse events were not serious and appear not to have been related to age. Instead, anorexia and nausea were more likely to be observed in female patients. Because the number of patients in this study was small, however, further data collection will be necessary to confirm these tendencies.
By the time of data analysis, five patients had been switched from eltrombopag to romiplostim. The reason for this in the other two patients (who were male) was lack of efficacy. After switching to romiplostim, the response was improved in one of these patients, but not in the other, due to the onset of MDS. Previous studies have shown that switching from one TPO-RA to another is effective in clinical practice in patients with not only immune thrombocytopenia [24, 25] but also AA [10, 26]. Our observational findings were consistent with their results. In this study, the eltrombopag dosage tended to affect the hematological response. An adequate TPO-RA dosage appears to be necessary for hematological recovery. If the first TPO-RA fails because an adequate dosage cannot be provided, switching to another TPO-RA should be considered to achieve treatment success.
Our study cohort included two patients who developed hematological malignancy. These patients eventually died as a result of AML and MDS, respectively. In previous studies on TPO-RA treatment for AA, clonal evolution and AML have been observed at an incidence of 3-19% and 0-4%, respectively [2, 4, 17, 19, 27]. However, it is also reported that secondary MDS or AML occurs in 15–20% of patients with AA regardless of the use of TPO-RAs . TPO-RAs can stimulate the expansion of hematopoietic stem cells via the c-MPL signaling pathway  and this could potentially affect the emergence of abnormal clones. Clonal evolution is of concern, and the precise mechanism underlying the development of hematologic malignancy remains controversial. Therefore, long-term follow-up and close surveillance of patients treated with TPO-RAs are essential.
We did not find any factors that significantly predicted hematological response to eltrombopag treatment. A shorter period from AA diagnosis to treatment initiation was likely to predict a hematopoietic response, but it was not significant. Previous studies have shown that higher reticulocyte counts and a shorter period from first immunosuppressive therapy to initiation of eltrombopag administration were likely to predict a hematological response to eltrombopag [2, 27]. While there are sufficient hematopoietic stem cells remaining in the bone marrow (i.e., before the hematopoietic stem cells are depleted), the introduction of a TPO-RA may be able to effectively stimulate a hematologic response.
This study has some limitations. First, it was a retrospective observational study with a small number of patients being treated at a single institution. The decisions to switch from pretreatment to TPO-RA when the TPO-RA was introduced and to switch the TPO-RA from eltrombopag to romiplostim were therefore left to the discretion of the treating physicians. To confirm our findings, we need to conduct a prospective study with a larger number of patients. Second, the reasons for not using ATG involved both cases in which the symptoms were mild and the use of ATG was unnecessary (i.e., not recommended according to the guidelines), and cases in which the symptoms were severe but ATG could not be used because of patient factors such as age and complications. There is thus a need for studies involving sufficient numbers of cases that the patients can be divided into separate cohorts according to these two sets of characteristics. Third, clonal evolution was not evaluated systematically, and it might occur more frequently than was documented here.
In conclusion, we investigated the efficacy of TPA-RAs in ATG- naïve patients with AA. Our findings reflect the nature of real-world TOP-RA treatment in patients with AA and suggest that TPO-RAs should be attempted in refractory patients with AA, even if they are elderly, have severe AA, and are ATG-ineligible. Further studies may identify factors that predict efficacy. Better understanding of TPO-RA action in AA and the proper use of TPO-RA in AA patients could improve AA treatment strategies and encourage successful long-term treatment of patients with ATG- naïve AA.