Incidence Rates of the Leukemias in the United States Are Significantly Associated with Birth Cohort,

Leukemia (all types) is a common cancer in the United States (US) with 44,600 new cases expected in 2011. There are currently only a few established risk factors for any major type. If leukemia risks are substantially modulated by known or unknown environmental and lifestyle exposures, then incidence rates in the population should vary significantly by birth cohort. However, prior studies have not examined birth cohort effects using contemporary data and methods.

We used nationally representative data from the National Cancer Institute's Surveillance, Epidemiology and End Results Program for 1992 – 2008 (68,481 leukemias and 6.6×10⁸ person-years of follow-up). For each major type in male and female pediatric (ages 0 – 17) and adult (ages 18 – 85) populations, we estimated the average annual percentage change in incidence attributable to calendar period and/or birth cohort (net drift), and the significance of unique (non-linear) birth cohort effects, using age-period-cohort statistical models and Poisson regression.

In adults, birth cohort effects for AML were significant in men (P = 0.002) and borderline significant in women (P = 0.053). Compared to men born during 1947 – 1951, men born in the 1920s–1930s were 1.4-fold more likely to develop AML; AML rates were stable in men born after 1951. Birth cohort patterns for AML were qualitatively similar in women compared to men. For CLL, birth cohort effects were highly significant in men (P=5.7×10⁻⁶), peaking among men born circa 1939 and falling by 38% among men born circa 1963. In contrast, birth cohort effects for CLL were almost completely absent in women. For CML, incidence declined steadily by 1.3%/year among men (P = 0.0003) and by 1.5%/year among women (P = 0.0008). For ALL, incidence increased steadily by 2.0%/year among women (P = 0.001); birth cohort patterns were similar in men but not statistically significant (P = 0.36). In the pediatric population, AML rates were stable over time. In contrast, ALL rates increased by 1.4%/year among males (P = 0.001) and by 1.1%/year among females (P=0.04).

In the US, leukemia risks increase or decrease substantially by birth cohort for each major leukemia type except adult female CLL and pediatric AML, on the order of 1–2%/year or 20–40% per generation. These results are consistent with the hypothesis that leukemia risks are substantially affected by known (i.e. smoking, certain chemicals) or suspected (i.e. obesity) environmental and lifestyle exposures, a number of which are potentially modifiable.

No relevant conflicts of interest to declare.