Context The role of race/ethnicity in survival of children with acute lymphoblastic
leukemia (ALL) is unclear, with some studies reporting poorer survival among
minority children and others reporting equivalent survival across race/ethnicity
in the modern, risk-stratified treatment era.
Objective To investigate the relation between race/ethnicity and survival in a
large, population-based analysis of incident ALL cases in the United States.
Design, Population, and Setting This study included 4952 individuals diagnosed with ALL between 1973
and 1999 at age 19 years or younger. ALL cases were identified from 9 population-based
registries of the National Cancer Institute's Surveillance, Epidemiology,
and End Results program.
Main Outcome Measures Survival probabilities were compared among white, black, Hispanic, Asian/Pacific
Islander, and American Indian/Alaskan Native children. Kaplan-Meier curves
and proportional hazard ratios from Cox regression analysis were calculated,
accounting for treatment era (1973-1982, 1983-1989, and 1990-1999), age at
diagnosis (<1, 1-9, and 10-19 years), and sex.
Results Although overall 5-year survival probabilities improved with each successive
treatment era, differences according to race/ethnicity persisted. For 1990-1999,
5-year survival was 84% for white children, 81% for Asian/Pacific Islander
children, 75% for black children, and 72% for both American Indian/Alaskan
Native children and Hispanic children. The largest difference by race/ethnicity
was observed among children diagnosed between ages 1 and 9 years. Compared
with white children, after adjusting for treatment era, age at diagnosis,
and sex, children of black, Hispanic, and American Indian/Alaskan Native descent
had hazard ratios of 1.50 (95% CI, 1.0-2.2; P = .03),
1.83 (95% CI, 1.4-2.4; P<.001), and 1.90 (95%
CI, 0.8-4.6; P = .16).
Conclusions Black, Hispanic, and American Indian/Alaskan Native children with ALL
have worse survival than white and Asian/Pacific Islander children, even in
the contemporary treatment era. Future work must delineate the social and
biological factors, including any differences in pharmacokinetics associated
with chemotherapeutic agents, that account for disparities in outcome.