Progressive Decline in Late Mortality after Hematopoietic Cell Transplantation (HCT) over 40 Years - a Report from BMTSS

Background: The high intensity of therapeutic exposures (for autologous and allogeneic HCT) and chronic GvHD and its management (in allogeneic HCT) increase the risk for late mortality. Significant changes in transplantation strategies have been instituted over the past 4 decades, with the overarching goal of improving outcomes. The impact of these changes on late mortality remains unknown.

Methods: We evaluated late mortality in 5,566 patients who had survived at least 2y after HCT performed between 1974 and 2010. Vital status information was ascertained as of May 2016, using medical records, National Death Index and Accurint databases. Separate analyses were conducted for allogeneic and autologous HCT. For all-cause mortality, relapse-related (RRM) and non-relapse-related (NRM) mortality, we examined trends over 4 time periods: <1990 (n=656); 1990-1999 (n=1650); 2000-2004 (n=1292) and 2005-2010 (n=1968). Multivariable Cox regression analysis was used to identify predictors of all-cause mortality. Proportional subdistribution hazards model (Fine-Gray) for competing risks was used for RRM and NRM. Multivariable analysis included demographics, primary disease, conditioning regimens, disease status at HCT, stem cell source (for allogeneic and autologous HCT recipients); for allogeneic HCT recipients, it also included transplant intensity, GvH prophylaxis, chronic GvHD.

Results: Allogeneic HCT:The cohort included 2,999 2y survivors followed for a median of 12.4y (range, 2.0-40.7) from HCT. Significant differences in HCT strategies (p<0.0001) observed over time in the cohort included increases in age at HCT (median: 21.7y to 40.4y); use of peripheral blood stem cells (PBSCs: 0% to 62%) and cord blood stem cells (0% to 21%); reduced intensity HCTs (0% to 54%); GvH prophylaxis with tacrolimus (0% to 57%), sirolimus (0% to 49%), MMF (0% to 27%); prevalence of chronic GvHD (43% to 60%); and decreases in GvH prophylaxis with methotrexate (71% to 32%), steroids (63% to 0.3%). Multivariable analysis revealed a 44% reduction over the 4 decades in risk of all-cause late mortality (Figure 1, <1990: HR=1.0; 1990-1999: HR=0.77, p=0.01; 2000-2004: HR=0.53, p=0.0003; 2005-2010: HR=0.56, p=0.005). The reduction in risk was more marked in NRM (80% reduction over 4 decades: <1990: HR=1.0; 1990-1999: HR=0.52, p<0.001; 2000-2004: HR=0.41, p=0.0005; 2005-2010: HR=0.2, p<0.0001) than RRM (51% decline over 4 decades: <1990: HR=1.0; 1990-1999: HR=0.74, p=0.02; 2000-2004: HR=1.06, p=0.7; 2005-2010: HR=0.49, p<0.0001) (Figure 1). Autologous HCT: The cohort included 2,567 2y survivors followed for a median of 9.3y (2-31) from HCT. Significant (p<0.0001) changes in transplantation strategies observed over time included increases in: median age at HCT (29.3y to 53.1y); use of PBSCs (39% to 100%); and decreases in: HCT for AML/MDS (18% to 5%), conditioning with TBI (43% to 12%). Multivariable analysis revealed a 75% reduction in risk of all-cause late mortality over 4 decades (Figure 2, <1990: HR=1.0; 1990-1999: HR=0.55, p<0.0001; 2000-2004: HR=0.35, p<0.0001; 2005-2010: HR=0.25, p<0.0001). The decline in risk of NRM was pronounced (96% decline over 4 decades; Figure 2, <1990: 1.0; 1990-1999: HR=0.41, p<0.0001; 2000-2004: HR=0.22, p<0.0001; 2005-2010: HR=0.04, p<0.0001). The decline in risk of RRM was 83% over 4 decades, and was statistically significant only after 2000 (Figure 2, <1990: HR=1.0; 1990-1999: HR=0.77, p=0.3; 2000-2004: HR=0.46, p<0.0001; 2005-2010: HR=0.17, p<0.0001). Analyses restricted to autologous and allogeneic HCT recipients transplanted between 1974 and 2004 with follow-up for all for the first 10y after HCT (to ensure comparable follow-up across all time periods) showed similar findings (data not shown).

Conclusions: Changes in transplantation strategies have contributed to a progressive decline in late mortality for both allogeneic and autologous HCT. The all-cause mortality has declined 44% for allogeneic HCT recipients and 75% for autologous HCT recipients over the 4 decades examined in this study. The decline in risk, while evident for relapse-related mortality, is more prominent for non-relapse mortality.

View largeDownload slide

View largeDownload slide

Close modal