Recent trends in breast cancer among younger women in the United States

Abstract

Increases in the incidence of postmenopausal breast cancers have been linked to screening and menopausal hormone use, but younger women have received less attention. Thus, we analyzed trends in breast cancer incidence (N = 387 231) using the National Cancer Institute's Surveillance, Epidemiology, and End Results Program 13-Registry database (1992-2004). Whites had higher incidence rates than blacks after age 40 years, but the reverse was true among younger women (black-white crossover). Among younger women, the rate per 100,000 woman-years was 16.8 for black vs 15.1 for white women; the highest black-white incidence rate ratio (IRR) was seen among women younger than 30 years (IRR = 1.52, 95% confidence interval = 1.34 to 1.73). This risk pattern was not observed in other ethnic groups. The black-white crossover among younger women was largely restricted to breast cancers with favorable tumor characteristics. The annual percentage change in the incidence of invasive breast cancers decreased modestly among older women but increased among younger (<40 years) white women. Continued surveillance of trends is needed, particularly for molecular subtypes that preferentially occur among young women.

Figure 1

Annual percentage changes (APCs) for age-adjusted incidence rate trends and absolute numbers in the National Cancer Institute's Surveillance, Epidemiology, End Results Program 13-Registry database (1992–2004) for all breast cancers combined (in situ + invasive), in situ cancers only, and invasive cancers only. A) APCs and absolute numbers by age (<40, 40–49, ≥50 years). B) APCs and absolute numbers by age (<40, 40–49, ≥50 years) and race (white, black). APCs are recorded with point estimates; 95% confidence intervals are in parentheses. Under the null hypothesis, non–statistically significant APCs indicate that the trend line for a given age group was no different than a horizontal or flat trend line with a slope of zero. Using Poisson regression models we also assessed whether trends varied by age group (A) or by race within age group, see text for details.