Background: Eltrombopag (EPAG) with or without immunosuppressant (IST) has been applied in acquired aplastic anemia (AA), yet data of EPAG+IST in relapsed/refractory AA was limited, and no study has compared efficacy and safety between EPAG+IST and EPAG monotherapy in relapsed/refractory AA patients.Aims: To evaluate and compare the efficacy and safety between EPAG+IST and EPAG monotherapy in relapsed/refractory AA patients in a real-world setting.Methods: Data from patients diagnosed as acquired AA in our center were retrospectively collected. All the enrolled patients were refractory/relapsed to the standard IST for at least 6 months before EPAG. All patients had been treated with EPAG, which was started at 25 mg/day and increased every 2 weeks to a maximum of 150 mg/day until a best response was achieved. Meanwhile, some patients were treated with cyclosporin A (CsA) or tacrolimus (FK506) at the same time. EPAG had to be prescribed for at least 6 months before evaluation. Complete response (CR), overall response (OR) and relapse rate, as well as adverse events and factors which could affect efficacy were analyzed.Results: Totally 99 patients (83 non-severe AA (NSAA) and 16 SAA) were included in the study. The median age at EPAG initiation was 46 (13-88) years old, the median time of EPAG treatment was 11 (6-41) months and the median time of follow-up was 18 (6-41) months. 72 patients were treated with EPAG+IST, including 41 (56.9%) treated with EPAG+FK506 and 31 (43.1%) treated with EPAG+CsA. 27 patients were treated with EPAG alone. No significant difference was found between EPAG+IST group and EPAG group in patient baseline characteristics like age, male proportion, NSAA proportion, presence of PNH clone, proportion of previous ATG+CsA / CsA treatment, previous IST duration and dosage. With compatible follow-up time, EPAG exposure duration and dosage, there was no significant difference in OR/CR rate at 3 rd/6 th/12 th month between patients who was treated with EPAG+FK506 and EPAG+CsA. Under similar compatible baseline conditions, the OR rate was 33.3% vs 22.2% (P=0.284) at 3 rd month, 61.1% vs 37.0% (P=0.032) at 6 th month, and 67.2% vs 42.1% (P=0.051) at 12 th month for patients treated with EPAG+IST and EPAG alone, respectively, but no significant difference was found in time to response (3 (1-12) vs 3 (1-7) months, P=0.679) or CR rate at 3 rd/6 th/12 th month (6.9%/12.5%/20.7% vs 3.7%/7.4%/5.3%, P>0.05) between the two groups. Relapse occurred at 6 th to 12 th month of EPAG treatment, and the relapse rate at 12 th month was 9.8% and 27.3% (P=0.154) for patients treated with EPAG+IST and EPAG alone, respectively. For patients treated with EPAG+IST, responders had a significantly higher baseline reticulocyte count (60.25 (11.5-230.5)×10 9/L vs 16.7 (6.6-56.6)×10 9/L, P=0.040) compared with non-responders. No predictive factors for the overall response were found for patients treated with EPAG alone. Adverse events which led to dosage regulation were gastrointestinal disorders (2.8% vs 3.7%, P=1.000), elevated creatinine (2.8% vs 0, P=0.599), elevated ALT (1.4% vs 0, P=1.000) and arthralgia (0 vs 3.7%, P=0.280) for patients with EPAG+IST and EPAG, respectively. No deaths were found in either group, while the clone evolution rate was 2.8% and 3.7% (P=1.000) in EPAG+IST and EPAG monotherapy group, respectively.Conclusion: EPAG+IST had higher OR rate than EPAG monotherapy with similar side effects for patients with relapsed/refractory acquired AA. Those with higher baseline reticulocyte count were more likely to respond to EPAG+IST.Key words: relapsed/refractory, aplastic anemia, eltrombopag, immunosuppressant, efficacy DisclosuresNo relevant conflicts of interest to declare. OffLabel Disclosure:In the presented study, eltrombopag was prescribed in relapsed/refractory aplastic anemia patients.