Introduction: Persistent thrombocytopenia (PT) is a common and severe complication after haploidentical allogeneic hematopoietic stem cell transplantation (haplo-HSCT), and there are currently no effective treatment strategies. Avatrombopag has shown promising results in the treatment of several thrombocytopenia disorders. However, no studies have examined the efficacy and safety of avatrombopag in the treatment of PT after haplo-HSCT, and predicting clinical response to avatrombopag before initiation is not possible. Thrombopoietin (TPO) is the chief cytokine that regulates megakaryocyte production and maturation, and endogenous TPO may be a biomarker of avatrombopag response. Methods: This study evaluated outcomes in patients who received the treatment of avatrombopag for persistent severe thrombocytopenia after haplo-HSCT between October 2019 and March 2022. On starting avatrombopag treatment, all included patients were with platelet counts ≤20×109/L and/or with platelet transfusion-dependence. Efficacy was assessed based on the platelet counts. Response was defined as platelet recovery to independence from platelet transfusion during or within 7 days after avatrombopag treatment, and complete response (CR) was defined as platelet recovery to ≥ 50×109/L for 7 consecutive days without platelet transfusion support. Endogenous TPO levels were measured by ELISA. Results: We report a cohort of 69 patients with PT after haplo-HSCT who were treated with avatrombopag. The 69 patients included 43 (62.3%) males and 26 (37.7%) females. The median age was 37 (11-60) years. Before starting avatrombopag therapy, the median platelet count was 15×109/L (range, 2-20×109/L). Forty patients (58.0%) had delayed platelet engraftment, and 29 patients (42.0%) had secondary failure of platelet recovery. Sixteen (23.2%) of the 69 patients had a history of using other TPO-RA (rhTPO or eltrombopag) for at least one round. Avatrombopag was initiated on a median of 59 (range 16-824) days after haplo-HSCT. Both the median starting and maximum doses were 40 mg daily. The rate of platelet recovery to transfusion independence was 72.5%, and the rate of platelet recovery to ≥ 50 ×109/L without transfusion support was 56.5%. Forty-seven (94.0%) of the 50 responders were able to taper off avatrombopag. The median duration of treatment was 40 (range 15-195) days. In addition, we present a novel clinical investigation of the use of TPO levels to predict the response to avatrombopag in patients with PT after haplo-HSCT. Logistic regression models demonstrated a significant predictive relation between TPO levels and probability of response. The TPO levels were inversely correlated with response (r = −0.998, P =0.011).(Figure) ROC analysis identified TPO levels ≤1714 pg/mL as the optimal threshold for discriminating responders and nonresponders. Based on univariate analysis, the presence of megakaryocytes in bone marrow before initiation (P=0.016), lower endogenous TPO levels (P =0.024) and no prior history of TPO-RA treatment (P =0.021) were independent factors that influenced the efficacy of avatrombopag. In the multivariate analysis, lower endogenous TPO levels (P =0.021) and no prior history of TPO-RA treatment (P=0.018) were independent factors that influenced the efficacy of avatrombopag. Avatrombopag was well tolerated in all patients, and patients responding to avatrombopag achieved improved overall survival (77.3% vs. 96.6%, P=0.008). Conclusions: Avatrombopag is a safe and effective therapy for severe PT after haplo-HSCT, and endogenous TPO levels prior to treatment could predict the response to avatrombopag. Key words: Avatrombopag, Persistent thrombocytopenia, Allogeneic hematopoietic stem cell transplantation, Haploidentical Figure legends:Multiple logistic regression models for the probability of response to avatrombopag based on endogenous TPO levels. The shaded area represents the 95% CI. Figure 1View largeDownload PPTFigure 1View largeDownload PPT Close modal