Although acute graft-versus-host disease (aGVHD) is a major cause of non-relapse mortality (NRM) after allogeneic hematopoietic cell transplantation (allo-HCT), there have been few reports on the influence of older age on NRM after aGVHD.
We retrospectively analyzed 2954 patients with hematological malignancies who received their first allo-HCT between January 2005 and December 2009 and developed grade II to IV aGVHD, using the database of the Transplant Registry Unified Management Program (TRUMP) at the Japan Society for Hematopoietic Cell Transplantation. The median age of the patients was 48 years (range: 20–82) and the median follow-up period for survivors was 740.5 days after transplantation. The diagnosis included acute myeloid leukemia (n=1248, 42.2%), acute lymphoid leukemia (n=537, 18.2%), myelodysplastic syndrome (n=328, 11.1%), non-Hodgkin lymphoma (n=391, 13.2%), adult T cell leukemia/lymphoma (n=187, 6.3%), chronic myeloid leukemia (n=99, 3.4%), plasma cell neoplasms (n=56, 1.9%), myeloproliferative neoplasm (n=42, 1.4%), Hodgkin lymphoma (n=27, 0.9%), and other leukemia/lymphoma (n=39, 1.3%). High-risk disease was present in 1470 patients (49.8%). The stem cell sources were HLA-matched related donors in 689 patients (23.3%), mismatched related donors in 245 patients (8.3%), matched unrelated donors in 780 patients (26.4%), mismatched unrelated donors in 595 patients (20.0%), and cord blood in 645 patients (21.8%). Myeloablative conditioning was used in 1808 patients (61.2%). Cyclosporine- and tacrolimus-based GVHD prophylaxis was used in 1304 and 1600 patients (44.1% and 54.2%), and just 88 patients (3.0%) underwent in vivo T cell depletion. Five hundred and thirty-four patients (18.1%) had GVHD prophylaxis without methotrexate. At the onset of aGVHD, 2379 (80.5%), 234 (7.9%), and 958 patients (32.4%) had initial skin, liver, and gut involvement, respectively. The cumulative incidence of NRM was analyzed under the assumption that they represented competing risks, defining relapse as a competing event for NRM, and compared using Gray's test. The Fine-Gray proportional hazards model was used to test the statistical significance of several potential prognostic factors for NRM. Multivariate analysis was conducted with covariates with P-values <0.05 on univariate analyses.
On analysis with stratification into five age groups (20–29 [n=346], 30–39 [n=552], 40–49 [n=697], 50–59 [n=946], and 60 or older [n=413]), two-year NRM rates after aGVHD were 17.4%, 24.1%, 25.0%, 34.8%, and 38.4%, respectively. Of 853 NRM, the most common primary cause of death was infection (n=260, 30.5%), followed by aGVHD (n=162, 19.0%) and organ failure (n=131, 15.4%). On multivariate analysis, six factors significantly were associated with increased risks of NRM: 1) age of 50 or older when compared to 20–29 (50–59: HR: 2.1 [range:1.5–2.8], 60-: HR: 2.3 [range: 1.6–3.2]), 2) three or more points on the HCT-CI score at transplant when compared to 0 point (HR:1.4 [range:1.02–2.0]), 3) HLA mismatched related and HLA matched/mismatched unrelated donor when compared to HLA identical related donor (HLA mismatched related donor: HR: 1.9 [range: 1.4–2.5], HLA matched unrelated donor: HR: 1.7 [range: 1.3–2.2], and HLA mismatched unrelated donor: HR: 1.8 [range: 1.4–2.4]), 4) GVHD prophylaxis without methotrexate (HR:1.6 [range:1.3–2.1]), and 5) grade III at the onset of aGVHD when compared to grade I-II (HR: 1.8 [range: 1.4–2.3]), and 6) initial liver involvement at the onset of aGVHD (HR: 2.1 [range:1.6–2.6]). On sub-analysis by aGVHD grade, the impact of age on NRM was greater in patients with grade II aGVHD than in those with grades III–IV. On additional multivariate analysis limited in the 859 patients of 50 or older with reduced intensity conditioning, age of 60 or older was still a significant factor for increased NRM when compared to 50–54 (55–59: HR: 1.4 [range: 0.99–1.9], 60–64: HR: 1.6 [range:1.1–2.2], 65–70: HR: 1.7 [range: 1.1–2.6], and 70-:HR: 4.2 [range: 1.4–12.7]).
Our study showed that older age is associated with increased NRM after aGVHD. It may be important to tailor prophylaxis against infection and treatment of aGVHD and infection according to the patient's age, donor source, and the symptom of aGVHD onset in order to reduce NRM further.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.