Abstract
Immunosuppressive therapy (IST) offers effective treatment for 40% - 70% of patients with aplastic anemia (AA). However, 15% - 40% of AA patients given IST will develop a clonal disorder of hematopoiesis clinically presenting as myelodysplastic syndrome (post-AA MDS) many of whom will proceed to allogeneic hematopoietic cell transplantation (HCT). Whether the success of HCT is comparable to or different from that in patients with de novo MDS is not known.
We identified 123 patients who underwent HCT for post-AA MDS (N = 89 with refractory anemia; N = 34 with more advanced MDS). Transplants occurred from 1977 through 2011 (75% since 2000). Patients with post-AA MDS were 2-68 (median 28) years old; 64% were male, 36% female. The interval from diagnosis of post-AA MDS to HCT was <1 – 152 (median 6) months; 25% had good risk, 25% intermediate risk, and 50% poor risk cytogenetics by the International Prognostic Scoring System criteria. Eighty percent of HCTs were from unrelated donors, and among these, 75% were HLA-matched. In a matched-pair analysis the 123 post-AA MDS patients (cases) were matched to patients with de novo MDS (N = 393; controls) for age, disease status, donor type and HLA-match, intensity of conditioning regimen and transplant period. Seventy-four cases were matched to 4 controls, 13 to 3 controls, 22 to 2 controls, and the remaining 14 cases to 1 control. Controls were more likely to have a longer interval from diagnosis of MDS to HCT (median 9 vs. 6 months; p<0.001) and less likely to have poor risk cytogenetics (39% vs. 53%; p=0.008) or to have received ATG (38% vs. 53%, p=0.003). The source of stem cells for controls and cases was bone marrow (42% vs. 53%), blood progenitor cells (44% vs. 35%) or cord blood (13% vs. 12%). Endpoints in the analysis were graft failure, GVHD, relapse, non-relapse mortality (NRM), relapse-free and overall survival. Relapse and NRM were treated as competing risks.
Time to neutrophil engraftment (18 vs. 17 days) and engraftment rate (87% vs. 89%; p=0.44) were similar for cases and controls; there were 11 primary and 3 secondary graft failures among cases and 21 primary and 9 secondary graft failures among controls (p=0.17). The day-100 incidence of grades II-IV acute GVHD was 52% vs. 47% (p=0.30), and the 1-year incidence of chronic GVHD 49% vs. 41% (p=0.13) for cases and controls, respectively. With a median follow-up of 4 years for cases and controls, the 5-year probabilities of NRM, relapse, overall and relapse-free survival were 37%, 19%, 49% and 45% for cases, and 27%, 30%, 49% and 43%, respectively, for controls. In multivariate analysis the risk of overall mortality was similar for cases and controls (HR 1.03, p=0.83). The NRM risk was marginally higher for cases (HR 1.44, p=0.06) while relapse risk was lower (HR 0.60, p=0.02) than for controls, and there was no difference in relapse-free survival (HR 0.94, p=0.62). Cytogenetic risk was independently associated with mortality. Among cases, compared to good cytogenetic risk, intermediate risk (HR 2.13, p=0.03) but not poor risk (HR 0.79, p=0.41) was associated with higher mortality. Among controls, both intermediate risk (HR 1.55, p=0.02) and poor risk cytogenetics (HR 1.44, p=0.04) were associated with higher mortality.
Transplant success with related and unrelated donors in patients with post-AA MDS was similar to that in patients with de novo MDS, with a 5-year relapse-free survival of 41%. The slightly higher NRM risk in post-AA MDS patients, which deserves further investigation, was negated by lower relapse risk. The apparent higher risk of mortality with intermediate risk cytogenetics among patients with post-AA MDS was unexpected, but may be related to small numbers or is the expression of undetermined biologic features.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
