Long-term outcomes in patients with severe aplastic anemia treated with immunosuppression and eltrombopag: a phase 2 study

Abstract

Patients with severe aplastic anemia (SAA) are either treated with bone marrow transplant (BMT) or immunosuppression (IST) depending on their age, comorbidities, and available donors. In 2017, our phase 2 trial reported improved hematologic responses with the addition of eltrombopag (EPAG) to standard IST for SAA when compared with a historical cohort treated with IST alone. However, the rates and characteristics of long-term complications, relapse, and clonal evolution, previously described in patients treated with IST alone, are not yet known with this new regimen, IST and EPAG. Patients were accrued from 2012 to 2020, with a total of 178 subjects included in this secondary endpoint analysis. With double the sample size and a much longer median follow-up (4 years) since the original publication in 2017, we report a cumulative relapse rate of 39% in responding patients who received cyclosporine (CSA) maintenance and clonal evolution of 15% in all treated patients at 4 years. Relapse occurred at distinct timepoints: after CSA dose reduction and EPAG discontinuation at 6 months, and after 2 years when CSA was discontinued. Most relapsed patients were retreated with therapeutic doses of CSA +/- EPAG, and two-thirds responded. Clonal evolution to a myeloid malignancy or chromosome 7 abnormality (high-risk) was noted in 5.7% of patients and conferred a poorer overall survival. Neither relapse nor high-risk evolution occurred at a higher rate than was observed in a historical comparator cohort, but the median time to both events was earlier in IST and EPAG treated patients. This trial was registered at www.clinicaltrials.gov as #NCT01623167.

Figures

Graphical abstract

Figure 1.

Enrollment flow diagram. A total of 236 patients were screened to enroll 180 in this phase 2 trial between 2012 and 2020. Most patients not enrolled in the study did not meet eligibility criteria or declined enrollment to pursue alternative therapy, primarily bone marrow transplantation. In Cohorts 1-3, one subject was found to have abnormal baseline cytogenetics, which precluded treatment on this protocol and therefore was taken off the study to pursue appropriate treatment. One patient enrolled on the extension cohort, who had ANC <200 x 109/L, was initiated on treatment prior to the results of pretreatment cytogenetics as allowed per protocol but was taken off the study to pursue appropriate treatment after chromosome 7 abnormalities were identified (marked by * in the diagram). Both patients were not included in the reported secondary analysis. A total of 178 patients were included for hematologic response at 6 months and clonal evolution analysis. A total of 145 patients who had achieved a response at 6 months were included for relapse analysis. Of note, the Clinical Center at NIH performs screening at multiple levels, making the application of a CONSORT analysis complicated.

Figure 2.

Relapse and clonal evolution in the EPAG-IST and historic IST groups. (A) Competing risk curves (death and BMT used as competing risks) depicting the time to relapse in subjects with a response at 6 months; blue line represents patients who received CSA maintenance (n = 109) and red line shows patients who did not receive CSA maintenance (n = 36). The purple dashed line indicates 6-month landmark timepoint when CSA was discontinued in the first 44 subjects and lowered in the rest and EPAG discontinued in all. The green dashed line indicates 2-year landmark timepoint when maintenance dose CSA was discontinued. Relapses were observed to occur shortly after these 2 distinct time points. (B) In the EPAG-IST group, 134 subjects were treated with 6 months of therapeutic CSA followed by 18 months of low-dose maintenance. Of these, 109 achieved a response at 6 months, and these were compared with a historic IST group who received a similar IST regimen with 24 months of CSA total. These competing risk curves (death and BMT used as competing risks) show no difference in the rates of relapse between subjects treated with IST plus EPAG versus historic IST alone (p = 0.09). (C) The cumulative incidence of clonal evolution was noted in 14.8% of all treated patients in EPAG-IST group. (D) No difference in high-risk evolution was observed in EPAG-IST and historic IST group as shown with completing risk analysis using Fine-Gray (death and BMT used as competing risks).

Figure 3.

Treatment and responses in patients with relapse in the EPAG-IST group. (A) Approximately half of the relapsed patients were treated with CSA alone, and the other half were treated with CSA and EPAG combination. Responses in these 2 groups are shown in the pie diagram. *One patient in the CSA plus EPAG treated group was non-compliant, and therefore not assessable and not included in the pie diagram. (B) Of the 43 subjects who achieved a response with reinitiation of oral therapy, about 1/3 were able to discontinue drug with ongoing response while another 1/3 remained on drug (many on low dose or being tapered) with ongoing response.

Figure 4.

The impact of initial response and late events on survival in EPAG-IST group. (A) All-cause mortality was included. There was no statistical difference in overall survival in EPAG-IST and historic IST group. (B) Patients with relapse did not have a worse outcome when compared with the ones without a relapse (this group includes patients with clonal evolution). (C) High-risk clonal evolution resulted in lower survival when compared with patients with no evolution or low-risk evolution. (D) Patients who did not achieve the initial hematologic response (non-responders) had poorer overall survival when compared with the ones with the response; however, achieving a complete response did not further improve the survival when compared with the patients with incomplete recovery.