The Role of Structural Violence in Acute Myeloid Leukemia Outcomes

Background:

Non-Hispanic Black (NHB) and Hispanic patients with Acute Myeloid Leukemia (AML) have higher mortality rates than non-Hispanic white (NHW) patients despite lower incidence, more favorable genetics, and a younger age at presentation (Darbinyan, Blood Adv. 2017). We performed a multilevel analysis of disparities in AML patients to investigate the contribution of structural violence, specifically neighborhood SES, on racial/ethnic differences in leukemia-specific survival.

Methods:

Adult AML (non-APL) patients diagnosed between 2012 and 2018 at six academic cancer centers in the Chicago area were included. Census tract data was collected using the FFIEC Geocoding/Mapping System and computed tract disadvantage and tract affluence scores were categorized into distribution tertiles (low, moderate, high). Time to relapse and death from leukemia were examined, adjusting for age, gender and race/ethnicity (baseline models), and for potential mediators of racial disparities including distal (Charlson Comorbidity Index (CCI), obesity, concentrated disadvantage and affluence, health insurance status), and proximal mediators (somatic mutations, and European Leukemia Network (ELN) prognostic score categories).

Results

Patient characteristics are shown in Table 1 (n = 822). Significant heterogeneity in age and comorbidities at diagnosis was observed, with Hispanic patients being the youngest and with the lowest CCI. Morbid obesity was more prevalent in NHB and Hispanic (23% and 20%, respectively) compared with NHW (11%) patients. Payer source also differed significantly; private insurance was twice as frequent among NHW than NHB (51% vs. 25%) patients, while the largest uninsured population was Hispanic.

ELN adverse risk disease was most prevalent in NHW subjects, NPM1 mutations were least prevalent in Hispanic patients, and p53 mutations more prevalent in NHB (26%) compared to NHW (12%) and Hispanics (9%) although due to low numbers this did not reach significance (p=0.10). NHB and Hispanic patients tended to reside in more disadvantaged and less affluent areas. Treatment data was available for 764 patients (Table 2); 75% received intensive induction therapy and choice of first-line treatment did not differ by race or tract disadvantage. Allogeneic transplant rates however differed by race, age, insurance status, tract disadvantage, and ELN score.

Treatment complications of induction chemotherapy, as reflected by ICU admissions during induction, were significantly lower in NHW (25%) compared to NHB (39%) and Hispanic (42%) patients. ICU admission rates were significantly higher in patients with morbid obesity and low tract affluence.

Minority (vs. NHW) ethnicity was associated with a 42% increased hazard of death from leukemia (HR=1.42, 95% CI: 1.09, 1.85), and a 36% increased hazard of death from all causes (HR=1.36, 95% CI: 1.07, 1.72), each after controlling for age, gender and study site. Adjustment for continuous tract disadvantage and affluence and their interaction lowered both the hazard of leukemia and all cause death to 1.18 (95% CI: 0.88, 1.60) and 1.14 (95% CI: 0.88, 1.49), respectively. In formal mediation analysis, neighborhood SES accounted for 37% (p=0.09) and 50% (p=0.02) of the racial disparity in death from leukemia and all causes, respectively.

Discussion: This study is the first to integrate data at the individual patient level with neighborhood characteristics, using census tract level variables to examine their contribution to AML patient outcomes. To date, formal mediation methods had not been employed to disentangle race/ethnic disparities in adult AML survival. Notably, our mediation analysis shows that census tract level SES explains a substantial proportion of the disparity in hazard of leukemia death. In addition, the observed disparities in treatment complications of induction chemotherapy, as reflected by ICU admissions, and the continued disparity in allogeneic transplant utilization all warrant further study. These results draw attention to the need for deeper investigation into the social and economic barriers to successful treatment outcomes for leukemia patients and represent an important first step toward designing strategies to mitigate these persistent health inequities.