Rapid initiation of cyclosporine and eltrombopag ahead of definitive treatment decreases autoimmunity and expedites marrow recovery in a prospective phase II trial

Abstract

Abstract Introduction: Severe aplastic anemia (SAA) is characterized by pancytopenia and an empty marrow. Due to multiple diagnostic procedures and specialized treatment requirements for SAA, expert center referral is advised, but resultant delays in beginning therapy allow continued hematopoietic stem cell (HSC) destruction. Our phase 2 study (NCT04304820) evaluated the feasibility and safety of rapid initiation of oral, low dose cyclosporine (CSA) and eltrombopag (EPAG) to abrogate immune-mediated destruction and promote hematopoietic recovery followed by standard therapy, allogeneic HSC transplant (HSCT) or horse anti-thymocyte globulin (hATG)/CSA/EPAG. Methods: Since 2020, 45 patients, age ≥3 years (y), with treatment naïve SAA were enrolled. Patients with high suspicion for a constitutional marrow syndrome or evidence of a clonal myeloid disorder were excluded. Enrollment was not dependent on specialized tests, but conventional “Camitta” blood count criteria and referring physician reading of a hypocellular (<30%) marrow. Low dose CSA (2mg/kg daily) and full dose EPAG were initiated after confirming eligibility. Patients were comprehensively evaluated at the NIH, including repeat bone marrow within 8 weeks (+4 weeks if extenuating circumstances) of oral treatment initiation. The primary endpoint was safety and feasibility of rapid oral treatment initiation, measured by a composite rate of 1) treatment related serious adverse events (TRSAE), 2) misdiagnosis, and 3) nonadherence to treatment or protocol requirements. Secondary endpoints included overall hematologic response rate (ORR) at 1, 2, 3, 6, and 12 months (mo) and annually thereafter; survival; relapse; and high-risk (HR) clonal evolution (CE) to secondary myeloid neoplasm, defined as chromosome 7 abnormality acquisition or by morphology. Results: Median age was 37 (range 8-69); 49% were female. Very SAA (absolute neutrophil count [ANC] <0.2 K/µL) occurred in 7 patients; PNH clone >1% occurred in 21. Using propensity scoring for age, sex, and disease severity, we matched a historical cohort (HC) of 45 patients treated with same day initiation of hATG/CSA/EPAG (NCT01623167). The primary endpoint of feasibility and safety was met with a 4% composite rate of no TRSAE, one misdiagnosis (a telomere biology disorder), and one instance of nonadherence. All patients completed early initiation treatment. Two patients continued oral therapy for 12 weeks: one due to prolonged outside hospitalization for COVID-19 complications; another to pursue HSCT (nonadherent case). Median follow-up was 755 days and 2,415 days in the early oral and historical cohorts, respectively. ANC significantly increased from oral therapy initiation (median: 14 days, range 0-86) to immediately prior to Day 1 (D1) of hATG administration (median 0.3 vs 0.42, respectively, P=.03). ORR at 6mo was similar (79% vs HC 87%, P=.52). Among 6mo responders, 2y relapse rate appeared lower (16% vs HC 32%, P=.3). HR CE at 2y was lower at 3% (vs HC 13%, P=.07). Survival at 2y was comparable (90% vs HC 93%, P=.8). Within 6mo of hATG, 38% of patients had a grade ≥3 infectious AE and 44% of patients had an infection-related hospitalization, compared to 20% and 20% in the HC, respectively. From pre-treatment to immediately pre-D1 hATG, paired serum levels (n=25) of IFN-γ-inducible chemokines including CXCL9 (P=.02) and IP-10 (P=.005) and TNF superfamily member TRAIL (P=.006) significantly decreased and paired flow cytometry peripheral blood samples (n=26) demonstrated significant reductions in T-cell activation (CD8+CD25+, P=.03, and CD4+CD25+ T-cells, P=.004) and potent effector function (CD8+CD57+, P=.005, and CD4+CD57+ T-cells, P=.004). Conclusion:Rapid initiation of CSA/EPAG is feasible and safe for prompt SAA treatment; it significantly improves ANC, a known predictor of outcomes, and shows evidence of disease modulation, including reduction of pro-inflammatory cytokines and activated T cells, while awaiting confirmatory testing and initiation of definitive treatment. Notably, we observed a low 2y HR CE rate of 3%, which might reflect reduced stress hematopoiesis, which will be further assessed with adequate power in our extension cohort. The low toxicity profile, early neutrophil recovery, and potential reduction in long-term adverse outcomes with early initiation of lower dose CSA/EPAG followed by standard therapy provides a strong argument for revising the current standard approach in newly diagnosed SAA.