Impact of Graft Composition on Graft-Versus-Host Disease in Peripheral Blood HLA-Identical Sibling Transplants: Protective Role of CD8 Cell Dose

Introduction and Objective: Acute graft-versus-host disease (aGVHD) is a leading cause of transplant related mortality following allogeneic hematopoietic stem cell transplantation (HSCT). Data on the impact of graft composition, namely CD34 and T-cell subsets, and aGVHD is conflicting. We hypothesized an upper cell dose limit of CD34 would exist predictive of aGVHD without affecting engraftment and sought to explore T-cell subsets and their impact on aGVHD outcomes amongst peripheral blood progenitor cell (PBPC) HLA-identical sibling transplants (Allo-Sib). Methods: In an IRB approved protocol, consecutive Allo-Sib patients underwent HSCT for hematologic malignancy at our center from 1/2004 to 12/2012. Demographic and clinical data were prospectively collected and retrospectively confirmed. Patients received the entire D1 CD34 PBPC collection from their donors and from a second day of collection if necessary to achieve a dose of 4 x 106 CD34/kg cells. aGVHD prophylaxis was uniformly tacrolimus and mini-methotrexate. Tacrolimus was monitored to achieve a trough of 10-15 ng/ml. aGVHD was graded as per the International Bone Marrow Transplant Registry System. Exact binary logistic regression models were used to estimate the odds of aGVHD as a function of patient demographic and clinical data. For the association between cell counts and aGVHD, binary cut scores were determined by finding the point along the receiver operating characteristic curve that maximized sensitivity and specificity. Non-parametric Spearman correlation coefficients were used to estimate the association between CD34, CD3, CD4, and CD8 cell counts. Results: We analyzed 160 consecutive patients. In this mostly male (63%) population, median age was 51 and most patients underwent a myeloablative transplant (93%). Disease risk as per Armand et al 2012 was low 45%, intermediate 13.2%, and high 41.7%. Co-Morbidity Index by Sorror et al (HCT-CI) was 30.5% for 0, 33.8% for 1-2, and 35.8% for > 2. Sixteen patients (10%) developed grade II-IV aGVHD while 9 (6%) developed grade III-IV aGVHD. Median CD34 cell dose infused was 6.27 X 106/kg. Median CD3, CD4, and CD8 cell doses infused were 2.51 X 108/kg, 1.80 X 108/kg, and 0.66 X 108/kg, respectively. In univariate analysis, patients (n=67) receiving > 9.94 x 106 CD34/kg were 12 times more likely to develop grade III-IV aGVHD (odds ratio =12.308; p=0.01)(figure 1). Conversely, patients receiving > 0.50 x 108 CD8/kg were less likely to develop grade II-IV aGVHD (odds ratio = 0.30; p=0.049)(Figure 2), but not Grade III-IV aGVHD. The only other factor associated with aGVHD was a diagnosis of AML, which was protective (p=0.01). CD3 and CD4 cell counts did not correlate with aGVHD in our patient cohort. There was no correlation between engraftment and CD34, CD3, CD4, or CD8 cell dose. Finally, we found no direct correlation between CD34 cells counts and CD3, CD4, or CD8 cell counts. Conclusion: This study represents the largest evaluation examining the correlation between both infused CD34 cells and T-cell composition with aGVHD among patients undergoing strictly matched related donor HSCT. Importantly, the data demonstrates patients receiving > 10 x 106 CD34/kg have increased risk of developing clinically significant aGVHD, a dose that far exceeds adequate engraftment doses and serves now as a cap on CD34 cell dose in our program. Interestingly, our data suggests infusion of > 0.50 x 108 CD8/kg may protect from severe aGVHD whereas prior reports have demonstrated no association between CD8 cells and aGVHD. Further investigation is needed to better characterize the relationship between CD8 cells and aGVHD and better define the correlation between CD8 and CD34 cell doses in the development of aGVHD. Disclosures No relevant conflicts of interest to declare.