Early Immune Recovery Predicts Overall and Disease-Free Survival After Allogeneic Hematopoietic Stem Cell Transplantation.

Abstract

Abstract 2222Poster Board II-199 Background:Immune recovery is an important determinant of outcome after allogeneic hematopoietic stem cell transplantation (alloHSCT). Previous studies have demonstrated that early recovery of absolute lymphocyte count (ALC) predicts overall (OS) and disease free survival (DFS) in conventional and partially T-cell depleted (TCD) alloHSCT. Recipients of conventional and especially TCD alloHSCT can have prolonged recovery of their immune systems, which may extend beyond one year. We have previously shown that the rate of CD4+ T cell recovery after TCD alloHSCT correlates with the risk of opportunistic infections. However, the impact of lymphocyte subset recovery on relapse and long-term survival following alloHSCT remains an important unknown. Methods:We conducted a retrospective study on 353 consecutive patients (median age 39, range 2-68) who received a fully TCD alloHSCT at MSKCC between January 1997 and December 2005. We excluded patients transplanted after 2005, because they received KGF for which pre-clinical data suggest an effect on immune recovery. Diseases included ALL (85), AML (146), CML (43), MDS (31), NHL (47), and undifferentiated leukemia (1). We excluded patients who died or did not have stable engraftment by 30 days. Immune recovery data, including ALC through day 90 and lymphocyte subsets through one year were grouped by quartile of cohort (table). The primary endpoints were OS and DFS. The score statistic from the Cox proportional hazards model defined the association of these endpoints with the values of ALC, CD4+ and CD8+ T cells, and NK cells. Subsequent to individual score tests, the Cox model determined the values associated with survival endpoints, in addition to prognostic clinical factors. Because markers were measured post-HSCT, a landmark analysis explored their prognostic significance. The landmark times were 1, 2, 3, 6, 9, and 12 months post-HSCT. Additional factors in the Cox model were: patient sex, age, number of remissions, stem cell source, and donor-recipient HLA match. The Kruskal-Wallis test evaluated if the diagnoses were associated with lymphocyte recovery.Variable1MinLower QuartileMedianUpper QuartileMaxALC 3000.30.50.811.6ALC 6000.40.81.26.7ALC 9000.50.71.217.9NK 3001683485252393NK 6002203034662229NK 90201722704271569CD4 61751362091771CD4 1201462484091574CD8 60521322873643CD8 12212625149436921ALC in K/mcL; NK/CD4/CD8 in cell/mcL Results:In univariate analysis, high ALC at day 30 (ALC 30, p < 0.001) and day 60 (ALC 60, p = 0.015) were associated with improved OS. High ALC 30 was also associated with improved DFS (p = 0.005). Multivariate analysis demonstrated that high ALC 30 was independently associated with improved OS (p = 0.011). NK cell count at day 60 was associated with improved OS and DFS by univariate (p = 0.014, 0.010) and multivariate (p = 0.007, 0.002) analysis. Similarly, CD4+ T-cell count at 6 and 12 months predicted survival. By univariate analysis, CD4+ T-cell count at 6 and 12 months predicted OS (p = 0.007, p < 0.001) and DFS (p= 0.032, 0.002). This relationship remained with multivariate analysis. CD4+ T-cell count at 6 and 12 months independently predicted OS (p = 0.013, 0.007) and DFS (p = 0.059, 0.043). There was no association of CD8+ T cell count with OS or DFS. Other predictors of improved survival included younger age, higher degree of HLA match, peripheral blood as stem-cell source, and transplant in 1st remission. Diagnosis did not differentiate clinical outcomes except for CML patients, who had decreased DFS (but not OS) secondary to increased relapse with TCD. Finally, diagnosis was not associated with post transplant immune recovery. Discussion:We demonstrate that several measures of immune recovery, including time points as early as day 30, are important predictors of OS and DFS after fully TCD alloHSCT. Our finding that ALC 30 predicts OS in the TCD setting is consistent with prior studies in the conventional and partially TCD settings. We further demonstrate that reconstitution of NK cells and, later, CD4+ T cells predicts survival and relapse in the TCD setting. Finally, we show there is no relationship between diagnosis and immune recovery. It is likely that these findings will also apply to conventional alloHSCT given the concordance of our ALC 30 outcomes. These results may help identify patients most likely to benefit from interventions to enhance post-alloHSCT immune reconstitution. Disclosures:Castro-Malaspina:Alexion: Consultancy, Honoraria.