Fax: (313) 966-7322
Ethnicity related differences in the survival of young breast carcinoma patients
Article first published online: 28 JUN 2002
Copyright © 2002 American Cancer Society
Volume 95, Issue 1, pages 21–27, 1 July 2002
How to Cite
Newman, L. A., Bunner, S., Carolin, K., Bouwman, D., Kosir, M. A., White, M. and Schwartz, A. (2002), Ethnicity related differences in the survival of young breast carcinoma patients. Cancer, 95: 21–27. doi: 10.1002/cncr.10639
- Issue published online: 28 JUN 2002
- Article first published online: 28 JUN 2002
- Manuscript Accepted: 21 DEC 2001
- Manuscript Revised: 20 DEC 2001
- Manuscript Received: 13 NOV 2001
- breast carcinoma survival;
- early-onset breast carcinoma;
African-American women face an increased risk of early-onset breast carcinoma compared to white American women, and breast carcinoma has been reported to be particularly aggressive in premenopausal women.
Surveillance, Epidemiology, and End Results Program data were analyzed for 507 African-American and 1378 white patients from Detroit diagnosed with breast carcinoma under the age of 40 between 1990 and 1999.
The proportion of in situ disease detected in African-American patients between 1995 and 1999 nearly doubled compared to the 1990-1994 interval (11.3% compared to 6.4%) but was consistently lower than the proportion of in situ disease seen in white patients for the same intervals (15.7% and 16.4% respectively). Evaluation of patients with invasive disease revealed that African-American patients had larger mean tumor size (3.4 cm versus 2.6 cm; P < 0.001), lower rates of localized disease (42.4% versus 52.1%; P < 0.001), higher rates of estrogen receptor negativity (61.9% versus 44.4%; P < 0.001), and higher proportions of medullary tumors (5.8% versus 3.3%; P = 0.021). Cox proportional hazards survival analysis adjusted for age, tumor size, nodal status, hormone receptor status, and histology showed higher mortality rates for African-American patients at all disease stages. Relative risk of death for African-American patients was 1.94 in patients with localized disease (95% confidence interval [CI], 1.23–3.05), 1.58 for regional disease (95% CI = 1.18–2.11), and 2.32 for distant disease (95% CI = 1.15–4.69).
These findings show that young African-American breast carcinoma patients face an increased mortality risk. Additional studies evaluating risk and treatment response in this subset of patients are warranted. Cancer 2002;95:21–7. © 2002 American Cancer Society.
Breast carcinoma survival is influenced by many factors.1–6 Volume of the invasive tumor component and regional nodal involvement are associated with increased risk of harboring micrometastases. Expression of particular molecular markers (estrogen receptor, progesterone receptor, HER2/neu) correlates with responsiveness to systemic therapies. Other clinical factors, such as age at diagnosis and ethnicity, are prominent features in the patient profile and are known to influence survival, but the biology underlying their prognostic impact remains poorly understood.
Young age at diagnosis and nonwhite ethnicity are generally found to exert a negative effect on outcome of American breast carcinoma patients.7, 8 The survival disadvantage conferred by these two features may be cumulative for African-American women. African-American women overall have a lower incidence of breast carcinoma compared to white American women, yet their mortality is substantially higher, and they have a younger age distribution for breast carcinoma. For women under the age of 45 years, breast carcinoma incidence is higher for African-American than for white American women.9
A review of the literature shows that evaluation of the age-related epidemiology of breast carcinoma in African-American women is complicated by inconsistent definitions of young,7, 10–14 with age generally functioning as a surrogate for menstrual status; and by the fact that single-institution studies are unlikely to have data on large sample sizes of young breast carcinoma patients with minority ethnicity backgrounds.
The goal of the current study was to analyze the outcome of African-American breast carcinoma patients compared to white breast carcinoma patients from the city of Detroit, Michigan, after stratifying for stage at presentation and other standard prognostic features. The current study is based on data from the Surveillance Epidemiology and End Results (SEER) Program. Because of its large African-American community, Detroit has been a major contributor of information regarding the epidemiology of breast carcinoma in African-American women since the SEER Program's inception.15
PATIENTS AND METHODS
Surveillance Epidemiology and End Results Program records were analyzed for African-American and white patients diagnosed with breast carcinoma in metropolitan Detroit between 1990 and 1999. The study was restricted to women diagnosed with breast carcinoma under the age of 40 years. Surveillance Epidemiology and End Results data is collected from clinical records; there is no central pathology review.
Stage of disease at presentation was categorized as local (disease confined to breast), regional (involvement of regional lymph nodes), or distant (clinically apparent spread to distant organ sites). In addition, primary tumor size, estrogen receptor status, progesterone receptor status, and tumor histology were evaluated.
Patterns of stage distribution (including in situ versus invasive/infiltrating cancers) and survival in the two patient subsets were analyzed for the separate time intervals of 1990–1994 and 1995–1999 and for the combined 1990–1999 decade.
All statistical analyses were done with SAS (SAS Institute, Cary, NC), version 8.1. Continuous variables were compared by unpaired t-test, and categorical variables were compared by chi-square test or Fisher exact test. The Cox proportional hazards model16 was used to estimate relative risks of death by stage of disease among African-Americans and whites, with adjustments for age, tumor size, nodal status, hormone receptor status, and histology.
A total of 1885 patients were identified who met the study criteria of being diagnosed with breast carcinoma under the age of 40 years; 507 (27%) were African-American and 1378 (73%) were white.
Table 1 shows trends in the detection of in situ cancers for the African-American and white patients during the chronologic intervals 1990-1994 and 1995-1999. A ductal pattern made up the majority of the in situ cases for all patient subsets, 75% of all in situ cases contained a ductal carcinoma in situ component, and 25% were pure lobular carcinoma in situ. The proportion of in situ cancers remained stable (16%) for the white patients over the entire decade. The proportion of in situ tumors increased for the African-American patients over the same time period; however, this fraction remained lower than that seen for the white patients (6.4% in situ tumors for the African-American patients 1990–1994, 11.3% for 1995–1999).
|Interval||Proportion of in situ cases||P value|
For the 462 African-American and 1157 white American patients with invasive cancer 1990-1999, the mean tumor size was 2.8 cm; 49.3% presented with disease localized to the breast; and 49.0% of tumors were estrogen receptor negative.
Clinical features for the invasive tumors, stratified by ethnicity, are shown in Table 2. The mean age for both the white and African-American patients was approximately 35 years. The African-American patients were less likely to present with localized breast carcinomas compared to white patients (42.4% vs. 52.1%, P < 0.001), and African-American patients had larger primary tumor sizes (34 mm vs. 26 mm, P < 0.001). The predominant tumor histopathology was invasive ductal cancer for both ethnic groups (78%); however, medullary cancers were more common among the African-American patients (5.8% vs. 3.3%, P = 0.021).
|Not otherwise specified||18||3.9%||29||2.5%|
|Estrogen receptor statusa||Negative||190||61.9%||388||44.4%||< 0.001|
|Progesterone receptor statusb||Negative||185||59.7%||412||48.8%||0.001|
The African-American patients were also more likely than white patients to have estrogen receptor and progesterone receptor negative tumors (61.9% vs. 44.4%, P < 0.001 and 59.7% vs. 48.8%, P < 0.001 respectively). Among the African-American patients, 51.5% had tumors that were negative for both estrogen and progesterone receptors, compared to 40.3% of white patients. Estrogen and progesterone receptor positive tumors were found in 30.6% of African-American patients compared to 46.3% of white patients.
The African-American patients had similar rates of axillary node positive disease (41.8% vs. 39.8%; P = 0.56) and the mean number of metastatic lymph nodes was also similar (4.4 vs. 4.7, P = 0.58). There was a larger fraction of African-American patients with unknown axillary nodal status (24.9% versus 14.5%).
Survival analyses were restricted to the study population with invasive cancers. Mean overall survival for the African-American patients was 45.2 months (range, 0–129 months); this was significantly worse than the mean overall survival of 49.9 months (range, 0–127 months) seen for white patients (P = 0.01). For the subset of patients diagnosed during the earlier time interval 1990–1994, the magnitude of this disparity was greater; mean survival for the African-American patients was 59.8 months, compared to 72.3 months for the white patients (P < 0.001). Mean overall survival, stratified by stage of disease at presentation, is shown in Figure 1.
Results of the Cox proportional hazards analysis of survival (adjusted for age, tumor size, nodal status, hormone receptor status, and histology), as shown in Table 3, showed that the African-American patients faced statistically significant greater risks of death at each stage of disease. Survival curves for the African-American and white patients with localized and regional, stages of disease are shown in Figures 2 and 3.
|Stage||Risk ratio||95% confidence interval|
The complex interactions of issues related to age at diagnosis, stage distribution, and access to care cumulatively exert a negative impact on breast carcinoma outcome in African-American patients. Five year survival for African-American breast carcinoma patients is 71%, compared to 86% for white patients.9 Early onset breast carcinoma has been associated with a greater incidence of unfavorable prognostic features, such as hormone receptor negativity, advanced stage distribution, and worse outcome.7, 10–12 African-American women have a younger age distribution for breast carcinoma, and these prognostic patterns probably contribute to the ethnicity-related survival disparities that are seen in the United States.
Age as a factor in assessing breast carcinoma prognosis is a surrogate marker for menstrual status and the influence of endogenous circulating hormones on tumor growth rates and response to treatment. However, there is no universally-accepted age limit for the designation of young. Studies of early onset breast carcinoma are therefore characterized by variable methodology, with different age limits used to define the young patient population. Fifty is frequently the age selected to demarcate pre-versus postmenopausal status, often without documentation of actual menopausal history.
In the current study, 40 years was selected as the cutoff for early onset breast carcinoma because this represents the age interval correlating with the crossover in ethnicity-related variation in breast carcinoma age distribution. Breast carcinoma is relatively rare prior to the fourth decade of life; however, the public health implications of early onset breast carcinoma are quite severe because this younger age range represents prime years of professional and familial productivity. For women of at least 50 years, the breast carcinoma incidence rate is 373 per 100,000 for whites and 317 per 100,000 for African-Americans. For women under 40, the incidence rate is 57 per 100,000 for whites and 64 per 100,000 for African-Americans. The age distributions equalize and crossover in the 40-50 year age interval, with higher incidence rates among white women over age 45 years.9 Despite the overall relatively low incidence of breast carcinoma in very young women, the disparate age distributions for African-American and white women seen in clinical practice can be quite striking, as approximately 30–35% of African-American breast carcinoma patients are under the age of 50 years, compared to only 20% of white breast carcinoma patients.13, 17–19
Similarities in patterns of disease among African-American women and patients with BRCA mutation-associated cancers have motivated speculation that genetic factors may be more significant among African-American women with breast carcinoma than previously expected.20 For example, early-onset breast carcinoma is a well-recognized hallmark for genetically-predisposed breast carcinoma, and, as discussed, this younger age distribution is also seen for African-American breast carcinoma patients. In addition, African-American breast carcinoma patients have an increased prevalence of estrogen receptor-negative, high-grade medullary tumors,19, 21–26 features that are similarly reported in hereditary breast carcinoma patients.27–32
Unfortunately, little research has been completed regarding the incidence of BRCA germline mutations among African-American families. However, the epidemiology of breast carcinoma in African-American women parallels the breast carcinoma profile seen among native African women, and this suggests the presence of a founder mutation-effect. Breast carcinoma is an uncommon disease in Africa, but afflicted patients are substantially younger compared to American breast carcinoma patients. Research from the University of Chicago33 has already characterized a few germline breast carcinoma mutations that appear unique to African-American pedigrees. Additional research is needed in this area.
An alternative explanation for the younger age distribution seen among African-American compared to white breast carcinoma patients has been proposed by Pathak et al.34 These investigators have incorporated the increased prevalence of young age at first live birth that is seen in African-American women and the increased breast carcinoma risk that occurs in the short-term postpartum period into a model that may at least partially account for the increased incidence of breast carcinoma in African-American women under the age of 40 years.
The survival disadvantage seen for African-American breast carcinoma patients is likely to be multifactorial in etiology. The increased rates of poverty and higher proportions of medically-uninsured individuals within the African-American community contribute to this disparity by creating barriers to health care access. These access issues result in delayed diagnoses and advanced stage distribution. The National Cancer Institute explored the impact of socioeconomic factors on breast carcinoma outcome in the Black White Cancer Survival Study, a case control study of African-American and white breast carcinoma patients. That project reported that 75% of the ethnicity-related disparities in breast carcinoma outcome were explained by sociodemographic variables.35 Similarly, Heimann et al.36 compared prognostic features, treatment, and outcome for white and African-American breast carcinoma patients treated at the University of Chicago between 1946 and 1987 and found that controlling for stage and treatment eradicated the overall survival disadvantage seen for the African-American patients.
However, findings such as in this study, indicating poorer breast carcinoma outcome within discrete stage groups for African-American compared to white patients,7, 8, 18, 19 motivate questions regarding differences in tumor biology. Furthermore, studies of African-American and white women seen in equal-access health care systems have yielded provocative results. Wojcik et al.37 evaluated over 6,000 breast carcinoma patients (11% African-American) through the Department of Defense (DOD) Tumor Registry. All subjects were beneficiaries of the DOD health program, entitling them to minimal or no-cost definitive care at military medical facilities. These investigators reported that adjustment for stage and demographic variables reduced the increased mortality risk in African-American compared to white women from 1.45 (95% confidence interval [CI] = 1.20–1.76) to 1.41 (95% CI = 1.16–1.70). The investigators reported that the overall five year breast carcinoma mortality rate (24.8%) appeared somewhat improved for African-American patients receiving care in this equal-access system compared to national population-based mortality rates for African-American breast carcinoma patients, yet the ethnicity-related survival disparity persisted.
Yood et al.38 evaluated breast carcinoma outcome in a managed care environment; this study evaluated nearly 900 breast carcinoma patients (30% African-American) seen through a large health maintenance organization-based medical group in Detroit, Michigan. Logistic regression modeling of the outcome data indicated that a more advanced stage distribution and sociodemographic factors exerted significant and possibly independent effects on the survival rates of African-American patients. Baseline hazard ratio of breast carcinoma mortality for African-Americans compared to whites was 1.6 (95% CI = 1.1–2.2). Adjusting for sociodemographic factors alone reduced this risk to 1.2 (95% CI = 0.8–1.9) and adjusting for stage alone reduced risk to 1.3 (95% CI = 0.9–1.9); adjustment for both resulted in the strongest effect on outcome, with risk reduced to 1.0 (95% CI = 0.7–1.5).
The unresolved issue remains: is African-American ethnicity merely a surrogate risk factor for limited socioeconomic resources and delayed diagnosis when looking at breast carcinoma outcome, or conversely, do the socioeconomic disadvantages that are more prevalent among African-Americans magnify and simultaneously camouflage some underlying difference in biologic tumor aggressiveness that results in younger age and more advanced stage at presentation? Differences in breast carcinoma survival may also be related to gynecologic/menstrual factors, lifestyle issues, nutrition/body weight, and environmental exposures. Poverty rates and ethnicity can interact with each of these, and sorting out their independent contributions to breast carcinoma incidence and survival is challenging. While it is imperative that we aim to eliminate poverty and its associated barriers to health care, research regarding tumor biology in breast carcinoma patients representing all sociodemographic and ethnic strata is essential.
The current study focused on breast carcinoma in African-American and white breast carcinoma patients under the age of 40 years, and showed a clear disparity in outcome, with the African-American patients facing a significantly greater risk of death at every stage of disease. An element of more aggressive tumor biology may be exerting an effect; however, future research should be directed at separating the genetic, environmental, and hormonal components of this effect. A research agenda directed at the challenging question of why the youngest African-American women face the highest breast carcinoma mortality risk will not only enable us to more effectively address and eradicate this particular outcome disparity, but it will also contribute to an improved understanding of the impact of breast carcinoma on young women of all ethnic backgrounds.
In conclusion, young African-American women are clearly facing increased risks for breast carcinoma mortality, and this public health issue warrants further investigation and research. Breast health awareness programs should be emphasized among this population subset, and efforts should be strengthened to include young African-American women in studies of breast carcinoma genetics as well as risk assessment.
- 7Survival patterns among younger women with breast cancer: the effects of age, race, stage, and treatment. J Natl Cancer Inst Monogr. 1994; 16: 69–77., .
- 9RiesLA, EisnerMP, KosaryCL, et al., editors. SEER Cancer Statistics Review, 1973-1997. Bethesda, MD: National Cancer Institute, 2000.
- 12Breast cancer outcome and predictors of outcome: are there age differentials? J Natl Cancer Inst Monogr. 1994; 16: 35–42., , .
- 16Regression models and life-tables. JR Statistical Society. 1972; B4: 187–220..
- 17SteeleGD OsteenRT, WinchesterDP, et al, editors. National Cancer Database Annual Review of Patient Care. Atlanta, GA: American Cancer Society 1994.
- 20Breast cancer genetics: toward molecular characterization of individuals at increased risk for breast cancer. Part I. In: DeVitaT, HellmanS, RosenbergSA, editors. Cancer Principles and Practice of Oncology Updates. 1998; 12: 1–12., .