Higher Socioeconomic Status Is Associated with Improved Overall Survival in Adults with Philadelphia Chromosome Negative Acute Lymphoblastic Leukemia: A Population Study

Background:

Socioeconomic status (SES) and other non-disease (social and demographic) characteristics are known to predict overall survival (OS) in children with acute lymphoblastic leukemia (ALL) (Petridou et al. Ann Oncol 2015). Less is known about the impact of these factors on survival of adults with ALL. We studied which non-biological risks impact OS in adults with Philadelphia chromosome negative (Ph negative) ALL, with emphasis on the impact of SES on survival.

Methods:

We assembled data from the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER-18) registry from 2000 to 2014. This registry covers 27.8% of the US population. We identified patients 18-69 years old diagnosed with Ph negative ALL between 2000-2012 using ICD-O-3 codes. We merged data with the Federal Information Processing Standards county codes and designated patients as metropolitan or non-metropolitan based on state and country 2013 rural-urban continuum codes. We assigned county-level median household income using the Census American Community Survey (www.census.gov).

Descriptive statistics were calculated for age at diagnosis, sex, race/ethnicity, marital status, insurance status,median household income, and status with respect to the 200% of the federal poverty line (FPL). We stratified the population by age: 18-39 years (adolescent and young adult, or AYA, cohort) and 40-69 years (adult cohort). We also identified 3 eras of interest: 2000 to 2003, 2004 to 2007, and 2008 to 2012. Multivariable Cox proportional hazards (PH) regression assessed predictors of OS, with separate models for AYAs and adults. Analyses were performed using SAS.

Results:

In total, 5,858 patients met criteria for analysis. Median age was 41.4 years (25-75^th%: 26.9-54.8), with 45.8%patients in the AYA cohort. 57.8% of patients were male, and half were non-Hispanic white (50.8%). Over 90% of all patients lived in metropolitan counties. One-third (35%) of patients lived below the 200% FPL. Median household income was $55,901 (25-75^th%: $51,389-67,677). 49% of patients had missing data about insurance, so we omitted this variable from analysis. ALL lineage was B cell in 57%, T-cell in 10%, and unspecified in 33% of cases. 52% were married overall with more AYAs (61.3%) not married compared to older adults (34%).

In Cox regression model for AYAs, higher median income was associated with better OS (HR=0.95 for every $10,000 above national household income, p=0.03). We also found that later era of diagnosis (2008-2012 vs. 2000-2003) was associated with better OS (HR=0.70, p<0.001), while older age (HR=1.04, p<0.0001), unspecified lineage (vs. B-cell, HR=1.20, p=0.04), and all races compared to non-Hispanic white were associated with poorer OS with the exception of Asian/Pacific Islander: American Indian (HR=2.82, p<0.001), Hispanic (HR=1.58, p<0.001), and non-Hispanic black (HR=1.50, p<0.001). Sex, marital status, and rural residence were not associated with OS in the AYA cohort.

In the adult cohort, in a Cox regression model, patients with higher income had better OS (HR=0.95 for every $10,000 above median national household income, p=0.001). We also found that later era of diagnosis (2008-2012 and 2004-2007 vs. 2000-2003), and non-metropolitan geography (HR=0.81, p<0.001) predicted better OS, while older age (HR=1.03, p<0.0001), Hispanic race (HR=1.16, p=0.03), male sex (HR=1.09, p=0.03), and non-married status conferred poorer OS (HR=1.31, p<0.0001) (Table).

Conclusion:

We identify non-biological predictors of OS in adults with ALL in a large population-based registry. Notably, higher SES portends better OS in both AYAs and older adults with ALL, consistent with findings in the pediatric literature. Marital status affects OS only in older patients, while race/ethnicity in AYAs greatly impacts OS. All ages had improved OS in more recent years, suggesting that all age cohorts are benefitting from new ALL treatment approaches. Our findings of the impact of SES on OS call for more investigation and action to improve social support and possibly adherence in the adult population.

Table.

View largeDownload slide

Table.

View largeDownload slide

Close modal