Decreasing chronic graft-versus-host disease rates in all populations

Abstract

Since 2005, there has been a steady decline in chronic graft-versus-host disease (cGVHD) at the Fred Hutchinson Cancer Center. To better understand this phenomenon, we studied the risk of cGVHD requiring systemic immunosuppression (cGVHD-IS) as a function of hematopoietic cell transplantation (HCT) date in 3066 survivors from 2005 through 2019. Cox regression models were fit to assess associations of HCT date (as a continuous linear variable) with cause-specific hazards of cGVHD using unadjusted and adjusted models. Median follow-up for study subjects was 7.0 years (range, 1.0-17.2). Two-year probabilities of cGVHD-IS declined among all survivors from 45% to 52% (2005-2007) to ∼40% (2008-2012) and then further to ∼26% by 2017. A decline was also observed when the analysis was restricted to 502 pediatric survivors, with cGVHD-IS probabilities <10% since 2013. Among 305 adult and pediatric survivors who underwent transplantation for nonmalignant diseases, cGVHD rates showed greater fluctuation but remained <20% after 2016. Each 5-year increase in HCT date was associated with a 27% decrease in the cause-specific hazard of cGVHD (unadjusted hazard ratio [HR], 0.73; 95% confidence interval [CI], 0.68-0.78; P < .0001); the HR was 0.81 (95% CI, 0.75-0.87; P < .0001) even after adjusting for various factors (age, donor/stem-cell source, race, sex, conditioning intensity, GVHD prophylaxis, among others) that could lead to cGVHD reduction. The decline in cGVHD was not fully explained by demographic shifts and greater use of HCT approaches that are generally associated with lower cGVHD rates. This observation underscores that single-cohort cGVHD prevention studies should use contemporaneous and not historical controls for comparison.

Graphical abstract

Figure 1.

Two-year point estimates of various outcomes by calendar year of HCT. Shown for all patients (A), age <18 years (B), and all patients who underwent transplantation for NMD (C). ∗Relapse in patients who underwent transplantation for malignancy; aGVHD, acute GVHD; cGVHD-All, cGVHD whether treated with systemic IS (cGVHD-IS) or not; NRM∗, NRM in all patients.

Figure 2.

HCT approaches and demographic shifts that might lower cGVHD-IS rates. Proportions shown by the 5-year eras in HCT date (A) and nonmutually exclusive GVHD prophylaxis categories (B). ATG, antithymocyte globulin; CNI, calcineurin inhibitor; CY, cyclophosphamide; MMF, mycophenolate mofetil; PTCY, posttransplant cyclophosphamide; TCD, T-cell depletion.

Figure 3.

Donor-recipient stem-cell source relationships, disease, and conditioning intensity. Proportions show 5-year eras in HCT date for donor type and stem-cell source (A); male recipients with female donors, children vs adult recipients, transplants performed for NMD, and myeloablative vs nonmyeloablative conditioning (B). BM, bone marrow; F, female; M, male; PB, peripheral blood.

Figure 4.

Association between time and cause-specific hazard of cGVHD. (A) Estimate of HR of NIH cGVHD-IS as a function of HCT date modeled as a continuous nonlinear variable (N = 3066), with HR relative to 1 January 2005; green line, unadjusted HR; red line, adjusted HR. (B) HR of relapse as a function of HCT date modeled as a continuous nonlinear variable among patients with malignant disease (n = 2756), with HR relative to 1 January 2005; green line, unadjusted HR; red line, adjusted HR.