Age-Related Differences in Utilization of Allogeneic HCT for Acute Myeloid Leukemia in California: Results of a Population-Based, Novel Linked Dataset

Background: Acute myeloid leukemia (AML) predominantly affects older adults (median age at diagnosis is 67yrs), and often requires allogeneic hematopoietic cell transplantation (alloHCT) for cure. According to the Center for International Blood and Marrow Transplant Research (CIBMTR), beginning in 2001, the largest growth in alloHCT for malignant diseases occurred in patients ≥65yrs. However, whether this is due to expanded utilization of alloHCT for newly diagnosed AML is unclear. Using a novel linked dataset, we examined patients with newly diagnosed AML in California to assess population-based alloHCT utilization over time and factors associated with receipt of alloHCT across age groups.

Methods: Data obtained from a linkage between the CIBMTR database, California Cancer Registry (CCR) and Patient Discharge Data (PDD) included adolescents and young adults (AYA, 15-39y), adults (40-64y) and older adults (65-79y) with AML diagnosed between 2001-2016. Patients who received no induction chemotherapy, those who received autologous HCT, or had unspecified HCT type were excluded. Multivariable Fine-Gray regression analyses accounting for the competing risk of death were modeled separately across age cohorts.

Results: 7,925 patients with newly diagnosed AML were included. AYAs comprised 15% of the cohort compared with adults (41%), and older adults (43%). Within the adult and older adult cohorts, the majority were non-Hispanic White, while 41% of the AYAs were Hispanic. Most AYA and adult patients had private health insurance; most older adults received Medicare. More AYAs (25%) had the lowest neighborhood SES relative to adults (15%) and older adults (14%). Approximately 30-40% of adults and older adults had 3+ comorbidities (using Elixhauser Comorbidity Index). Within each age cohort, approximately 25% of patients lived >50 miles from a transplant center. Examination of alloHCT utilization over time demonstrated that the use of alloHCT has increased within each age cohort (Figure 1). Although utilization among older adult patients increased from 2% to 13% at 2 years post AML diagnosis, the absolute magnitude of change was lowest in this age group and remains far lower than AYA and adult patients. Multivariable models (Figure 2) demonstrate that patient-level variables associated with alloHCT utilization also vary by age group. For AYAs, alloHCT utilization was lower for low and middle neighborhood SES (vs. high), uninsured and those covered by County/Public Health Service (vs. private insurance); and utilization was higher for males and more recent year of diagnosis. For adults, alloHCT utilization was lower for patients with increased age, male sex, non-Hispanic Black or Asian race/ethnicity (vs. non-Hispanic White), unmarried status (vs. married), low and middle SES, uninsured, Medicaid, Medicare, and County/Public Health Service, and 1-2 or ≥3 comorbidities (vs none); utilization was higher for more recent diagnosis year and patients living <50 miles distance from a transplant center (vs. 100+ miles). For older adults, utilization decreased with increasing age and was lower among Asian and unmarried patients, whereas more recent diagnosis year was associated with higher utilization.

Conclusions: Using a novel population-based linked database encompassing >99% of newly diagnosed AML patients in California, we found that alloHCT utilization increased in all age cohorts over time; however, fewer than 14% of older adults who received leukemia induction ultimately received alloHCT even in the most recent era studied. As potential patient-level barriers to alloHCT differ across age groups, studies should explore further modifiable age-related factors and develop age-specific interventions to mitigate barriers to alloHCT utilization.