While vast majority of patients achieve robust engraftment within a few weeks after allogeneic hematopoietic stem cell transplantation (HSCT), some patients may experience persistent cytopenias, mostly thrombocytopenia, which are usually referred to as poor graft function (PGF). Unfortunately, treatment options for PGF remain limited, because except the possible treatment of the cause (which is often missing, and not always treatable) the only therapy is a boost of HSCT from the same donor. Eltrombopag is a thrombopoietin-mimetic agent which is approved for the treatment of idiopathic thrombocytopenia; more recently, it has been shown effective for the treatment of aplastic anemia. Between January 2011 and April 2018, 66 patients suffering from different malignant or non-malignant hematological diseases underwent an HSCT at the Federico II University of Naples. We report on a retrospective analysis of 13 transplanted patients who have consistently received eltrombopag for the treatment of post-HSCT persistent cytopenia. Patients were selected according the following criteria: lack of neutrophil engraftment at day +28 (primary graft failure), or persistent unilineage or multilineage cytopenia at anytime after HSCT (PGF; demonstration of impaired hematopoiesis by bone marrow analysis). All but one patients (who had severe aplastic anemia) were transplanted for a malignant disease, after a fully myelo-ablative conditioning regimen. There was a female prevalence (9 out 13); the median Sorror comorbidity index was 0 (range 0-2) while the median EBMT score was 4 (range 2-7). The most common HSC source was bone marrow (BM) (10 out 13 patients, 83,3%), and the median dose of CD34+ was 2.24 x 106/kg (range 0.68-7.2), with 4 out of the 13 patients (33%) who have received <2 x 106 CD34+/kg (but only 1 <1.5 x 106 CD34+/kg). Major ABO-incompatibility was present in 4/13 cases (25%); these patients underwent 2-3 sessions of plasmapheresis before receiving an unmanipulated BM. The vast majority of patients (9 out 13, 75%) were transplanted from a haploidentical donor, while 2 and 2 patient were grafted from unrelated or sibling HLA-matched donor. Considering patients evaluable for engraftment (alive at day +35, n=52) the incidence of PGF was not increased in haplo-identical HSCT (Fisher two-tailed exact test, p=0.32). All but one patients had ANC engraftment (median 18 days, range 14-34), with donor engraftment confirmed by full donor chimerism on bulk BM specimen. The patient suffering from primary graft failure underwent a second HSCT from the same donor, eventually dying from acute GvHD. Indeed, 11/12 patients were classified as having PGF based on low platelet counts; 6 of them were considered primary PGF because they have never reached platelets >20000/µL after HSCT. In most patients, PGF was not limited to the megakaryocytic lineage, since anemia and neutropenia affected respectively 6 and 2 patients. At the time of PGF, 6/12 patients had grade II-IV acute GvHD. Among the 12 patients with PGF, 5 patients achieved a complete response and 2 patient a partial response; among the 5 non-responders, 2 suffered from early deaths due to infectious complications and were not evaluable for response. Among the 6 responders, 3 of them have discontinued eltrombopag, while retaining their hematological response. The long-term outcome of this cohort was good, since only 3 deaths were observed (the 2 infections and one grade IV aGvHD). No patient experienced relapse of his underlying disease, eventually suggesting that eltrombopag should not preferentially stimulate possible residual malignant cells. In conclusion, our study suggest that eltrombopag is safe and can be hypothesized as etiologic treatment of PGF; indeed, eltrombopag may be considered in post-HSCT prolonged cytopenia due to PGF when other etiologic treatments are elusive or ineffective. Not all PGF share the same cause and pathogenic mechanism; in our preliminary experience eltrombopag seems more effective when additional specific causes can be identified (e.g., immune mechanism associated with GvHD, or infectious complications and their treatments). Future studies are needed to better understand the pathogenic events underlying PGF in individual patients, and to identify the setting of PGF patients who may best benefit from eltrombopag treatment. DisclosuresPane:BMS: Speakers Bureau; AMGEN: Speakers Bureau; Novartis: Research Funding, Speakers Bureau. Risitano:Amyndas Pharmaceuticals: Consultancy; Pfizer Inc.: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Alexion Pharmaceuticals, Inc.: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; Jazz Pharmaceuticals: Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Alnylam Pharmaceuticals, Inc.: Consultancy, Honoraria, Research Funding; Novartis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding, Speakers Bureau; Ra Pharmaceuticals, Inc.: Consultancy, Honoraria, Research Funding.