Page 135 - 2022_01-Haematologica-web

Predicting response to IST plus EPAG in SAA
mycophenolate mofetil and sirolimus, and by others with the addition of androgens, granulocyte – colony-stimulat- ing factor and erythropoietin to standard IST have failed to demonstrate superiority to standard hATG and cyclosporine.3 However, eltrombopag, a synthetic throm- bopoietin receptor agonist, when added to standard IST resulted in better outcomes compared to the standard reg- imen, with an overall response rate of >80% and 2-year survival rate of 97%,4 and this combination is now approved for first-line use in SAA in the USA. A large, prospective, randomized trial comparing standard IST to IST plus eltrombopag has broadly confirmed these results,5 and smaller case series in a variety of populations have also suggested improved outcomes with the addi- tion of thrombopoietin mimetics.6-11
Response to conventional ATG-based IST without eltrombopag correlates with blood counts before therapy, including absolute reticulocyte count (ARC),12-18 absolute neutrophil count (ANC)12,16,17,19,20 and absolute lymphocyte count (ALC).12,21,22 The presence of somatic mutations,23 having a paroxysmal nocturnal hemoglobinuria (PNH) clone,13,15,20,24,25 elevated thrombopoietin levels,16 normal telomere length of lymphocytes,15,26 a subpopulation of regulatory T cells27 and B cells28 have all been reported as prognostic markers of response. Studies typically have emphasized overall response rather than complete response, but a positive feature of IST combined with eltrombopag is the high complete response rate.4 Previously, the robustness of response has been correlated with better overall survival.29
The aim of this study was to elucidate whether IST response predictors before the introduction of eltrom- bopag remain associated with predictors of treatment response with IST plus eltrombopag and to identify sub- groups of patients with SAA that might particularly ben- efit from IST plus eltrombopag. We sought correlations of age, initial blood counts, thrombopoietin levels, and the presence of PNH clones with both the overall and com- plete responses in a large dataset including groups of patients with SAA treated on consecutive clinical proto- cols with IST alone and IST plus eltrombopag at the National Heart, Lung, and Blood Institute (NHLBI), National Institutes of Health (NIH).
Methods
Patients and treatment
Patients with SAA, aged ≥2 years, treated with hATG-based regimens at the NIH Clinical Center, Bethesda, between May 1999 and August 2019, were included in the analysis. Studies before 1999 were excluded as information on complete response was not available. The patients were treated on any of four clin- ical studies approved by the Institutional Review Board of the NHLBI (NCT00001964, NCT00061360, NCT00260689, NCT01623167) and were done under the principles of the Declaration of Helsinki. Informed consent was obtained on the respective protocols.
Treatment regimens have been published:4,30-32 hATG (ATGAM, Pfizer) and cyclosporine were administered to all the patients; an investigational agent, including mycophenolate mofetil, sirolimus or eltrombopag, was added to IST in some protocols (Online Supplementary Methods and Online Supplementary Figure S1). The IST plus eltrombopag group referred to four cohorts administered eltrombopag in addition to
hATG and cyclosporine; the IST alone group included other hATG-based IST.
Definitions
All definitions have been consistent across all our protocols. SAA was diagnosed when bone marrow cellularity was <30% and at least two of the following three criteria were met: ANC <0.5×109/L, ARC <60×109/L, and platelet count <20×109/L. Hematologic responses were evaluated 6 months after adminis- tration of ATG. Overall response was defined as no longer meet- ing the criteria for SAA in the absence of recent transfusions and granulocyte colony-stimulating factor administration; complete response required all ANC ≥1.0×109/L, hemoglobin level ≥10 g/dL and platelet count ≥100×109/L; and partial response was defined as an overall response not meeting the criteria for com- plete response. For the purpose of statistical analysis, patients who did not complete 6 months of initial IST due to death, HSCT or who underwent a second course of IST were counted as having had no response, consistent with our previous study.12 Blood values assessed as potential response predictors were ARC, ANC, ALC, platelet count, presence of a PNH clone and plasma thrombopoietin levels. The lowest values during the 4 weeks preceding IST were used as pretreatment blood counts, as previously described.12 Plasma thrombopoietin levels were measured by magnetic Luminex assay with a standard proto- col.33 PNH clones were defined as the presence of 1% or more clonal populations deficient in glycosylphosphatidylinositol anchored proteins on the surface of neutrophils or red blood cells; information on PNH clones was available only for neu- trophils in the IST plus eltrombopag study.
Statistics
Statistical analyses were performed using an EZR software package (version 1.41), graphical user interphase of R (version 3.6.1).34 The Fisher exact test and Mann-Whitney U test were used to compare categorical and numerical variables, respective- ly. Correlations between baseline numerical variables were test- ed with the Spearman rank correlation test. In the logistic regres- sion model, continuous variables were transformed to the natu- ral log (X+1), where X represents the variable, consistent with the previous study.12 All variables with P<0.2 in the univariate analysis were included in the multivariate analysis and a back- ward stepwise selection procedure was performed to create the final model.
Results
Patients
A total of 417 patients with SAA, aged 2 to 82 years, were initially treated with standard hATG and cyclosporine, with or without an investigational agent, between May 1999 and August 2019, on any of four prospective studies (Online Supplementary Figure S1): (i) a single arm study of standard IST plus mycophenolate mofetil performed from May 1999 to June 2003 (n=103);30 (ii) a direct comparison of standard IST (n=42) and IST plus sirolimus (n=35), from June 2003 to November 2005;31 (iii) a direct comparison of standard IST (n=60) and rabbit ATG-based IST (n=60; not included), from December 2005 to July 2010;32 and (iv) a single-arm study of IST plus eltrombopag (n=177), from July 2012 to August 2019, including cohort 1 (n=30), cohort 2 (n=31), cohort 3 (n=31) and an extension cohort after the publication (n=85).4 One patient who received IST plus eltrombopag in the exten-
haematologica | 2022; 107(1)
127