Abstract

The authors assessed risk factor profiles among 1,505 African-American and 1,809 White women in the 1993–2001 Carolina Breast Cancer Study. Multiple logistic regression models for case-control data were used to estimate odds ratios for several factors. Racial differences were observed in the prevalence of many breast cancer risk factors among both younger (aged 20–49 years) and older (aged 50–74 years) women. For older women, the magnitude and direction of associations were generally similar for African-American and White women, but important racial differences were observed among younger women. In particular, multiparity was associated with increased risk of breast cancer among younger African-American women (for three or four pregnancies: adjusted odds ratio (OR) = 1.5, 95% confidence interval (CI): 0.9, 2.6; for five or more pregnancies: OR = 1.4, 95% CI: 0.6, 3.1) but not among younger White women (for three or four pregnancies: OR = 0.7, 95% CI: 0.4, 1.2; for five or more pregnancies: OR = 0.8, 95% CI: 0.2, 3.0). The relations with age at first full-term pregnancy and nulliparity also varied by race. Case-only analyses before and after further adjustment for tumor stage and hormone receptor status revealed little effect on results. Hence, racial variations in both prevalences of and risks associated with particular factors may contribute to the higher incidence of breast cancer among younger African-American women.

Received for publication December 9, 2003; accepted for publication June 24, 2004.

Few published epidemiologic studies have examined whether racial or ethnic differences exist in the prevalence of risk factors for breast cancer or the magnitude of their associations with the disease. Of case-control studies that have analyzed breast cancer risk factors among African-American women (16), only two were population based and examined African-American women and White women in the same study (3, 6). Both of these studies examined risk factors among younger women (aged 20–54 years) and, thus, had limited ability to elucidate racial differences by age or menopausal status.

Distinguishing breast cancers by age at onset has important implications for cancer incidence and etiology. The most recent Surveillance, Epidemiology, and End Results Program statistics for 1996–2000 found the age-adjusted incidence of breast cancer to be higher among White than African-American women (142.0 per 100,000 vs. 120.8 per 100,000) (7). The same pattern was seen in a comparison of White and African-American women aged 50 years or older (398.6 vs. 322.1 per 100,000), but the trend was reversed among women aged less than 40 years. In that group, the incidence rate for each 5-year age group was higher among African-American women (7). This disparity indicates that African-American women are experiencing excess rates of breast cancer at younger ages. Evaluation of age distributions in case series has shown that African-American breast cancer patients are more likely to present at a younger age (811) and that cases of breast cancer among young women overall are more aggressive and show a poorer prognosis and response to treatment than cases among older women (8, 1115). These factors contribute to a mortality rate for breast cancer among younger African-American women that is twice that of younger White women (7).

To evaluate potential differences in risk factors for breast cancer, we compared risk factor profiles between African-American women and White women, stratified at age 50 years, in the Carolina Breast Cancer Study. To highlight differences in risk factors that may contribute to excess incidence among younger African-American women, we first present adjusted odds ratios for reproductive and lifestyle factors. Because breast tumor characteristics are known to differ between African-American and White women (16), we then analyzed the pattern of results from case-only analyses before and after further adjustment for stage and hormone receptor status.

MATERIALS AND METHODS

Subjects

The Carolina Breast Cancer Study is a population-based, case-control study of breast cancer in African-American and White women residing in a 24-county area in central and eastern North Carolina (17). Both cases and controls were sampled using a modification of randomized recruitment (18). All women between 20 and 74 years of age who were diagnosed with a first, invasive breast cancer between May 1, 1993, and December 31, 2000, were eligible as cases and were identified using the North Carolina Central Cancer Registry’s Rapid Case Ascertainment System (19). Sampling probabilities ensured approximately equal samples in the four age-race groups: younger (age 20–49 years) African-American women, older (age 50–74 years) African-American women, younger White women, and older White women. We excluded women of other races, who constituted less than 2 percent of the study population. Of the 2,397 eligible and locatable cases, 172 (7 percent) were denied contact with us by their physicians, and an additional 352 (15 percent) declined to participate. Thus, information from 1,873 cases was available for analysis, for an overall cooperation rate of 78 percent. Age- and race-specific cooperation rates were 84 percent, 80 percent, 76 percent, and 72 percent for younger White and African-American cases and older White and African-American cases, respectively.

Controls were drawn from North Carolina Division of Motor Vehicles lists for women aged 20–64 years and US Health Care Financing Administration lists for women aged 65–74 years. Sampling probabilities for controls ensured approximate frequency matching to cases by race and 5-year age groups. Of the 2,444 eligible and locatable women, 727 (30 percent) declined to participate. Thus, information from 1,717 controls was available for analysis, for an overall cooperation rate of 70 percent. Cooperation rates for the four age-race groups were 76 percent, 71 percent, 72 percent, and 63 percent for younger White and African-American controls and older White and African-American controls, respectively.

After exclusion of 276 women who agreed only to a brief telephone survey and had incomplete information on many risk factors of interest, the final data set consisted of 1,505 (45 percent) African-American women (787 cases and 718 controls) and 1,809 (55 percent) White women (991 cases and 818 controls).

Data collection

The data were obtained during in-person interviews conducted by female registered nurses. Through a questionnaire, the nurse-interviewer elicited information on demographics and potential breast cancer risk factors, including first-degree family history of breast and ovarian cancer, menstrual and reproductive history, and sociodemographic and lifestyle characteristics. The nurses drew a blood sample and measured weight, height, and waist and hip circumferences at the time of interview. The median time from diagnosis to interview for the cases was 3 months (range: 1–19 months); 80 percent were interviewed within 5 months of diagnosis. For the controls, the median time from selection to interview was 2 months (range: 0–26 months); 80 percent were interviewed within 5 months of selection.

The American Joint Committee on Cancer stage was abstracted from medical records, where available, or determined from information on tumor size, lymph node involvement, and distant metastasis (16). Estrogen receptor and progesterone receptor status was obtained from medical records for 80 percent of the cases. For the remaining cases, estrogen receptor and progesterone receptor status was determined with paraffin-embedded tumor tissues at the University of North Carolina laboratory (11 percent), or receptor status was missing (9 percent) (19, 20).

Statistical methods

Statistical analyses were performed separately by race, and they were further stratified by age or menopausal status. Women were categorized as postmenopausal if they reported natural menopause or bilateral oophorectomy or if they were more than 55 years of age and reported hysterectomy. Women who reported still having menstrual cycles or who had at least one remaining ovary and were aged less than 42 years were classified as premenopausal. Women who had a hysterectomy without bilateral oophorectomy and were aged 42–55 years were considered perimenopausal.

The frequency distributions of risk factors in the study sample were adjusted using age-specific sampling weights to estimate prevalences in the underlying population. Comparisons by race were evaluated using the chi-square statistic. Both the prevalence estimates and chi-square tests were generated using SUDAAN version 8.0.0 software (Research Triangle Institute, Research Triangle Park, North Carolina). Odds ratios with 95 percent confidence intervals were calculated using logistic regression models that examined the association between breast cancer status and risk factors after adjustment for other relevant covariates including age (continuous), age at menarche, parity, age at first full-term pregnancy, miscarriage, breastfeeding, induced abortion, oral contraceptive use, and hormone replacement therapy. Body mass index, waist/hip ratio, history of breast cancer in a first-degree relative, education, alcohol consumption, and smoking were also included as covariates in addition to a term for the sampling fraction. Analyses also were adjusted for years since last full-term pregnancy; these results are not presented because odds ratios were equivalent within plus or minus 0.1. Logistic analyses were performed using SAS PROC GENMOD software (SAS Institute, Inc., Cary, North Carolina), which permits the use of an offset term to take the sampling design into account.

Tests for interaction between each covariate and race were conducted by performing the likelihood ratio test using data with both races combined. A model containing only the main effects of race, a variable of interest, and other covariates was compared with a model containing the main effects of race, the variable of interest, the relevant interaction term, and other covariates. Tests were conducted for interaction among all women, younger women, and older women. None of the interactions by race reached statistical significance at p < 0.05; however, a few comparisons yielded likelihood ratio tests with p < 0.20 (noted in text). Because of the limited statistical power for tests of interaction, we also chose to highlight differences between risks for African-American women and White women of 40 percent or greater in magnitude of the odds ratio or odds ratios that went in opposite directions. These criteria, though admittedly arbitrary, were selected to avoid making too much of fairly small variations in odds ratios and to focus attention on variations in odds ratios by race that were potentially meaningful in our data. We thought such results warranted further examination in the discussion in relation to possible consistency with other available findings.

To adjust for racial differences in breast cancer characteristics (i.e., stage at diagnosis and hormone receptor status), we conducted a series of case-only analyses using logistic regression models. Odds ratios and 95 percent confidence intervals were derived from case-case comparisons to highlight the presence of heterogeneity between African-American and White women with breast cancer after adjustment for stage and estrogen receptor and progesterone receptor status (i.e., odds ratios deviating from 1.0 suggest racial differences). Although the case-case odds ratio can be used as a measure of heterogeneity of odds ratios (21), its magnitude reflects risk factor differences in both background prevalence and disease-associated risk, and the absence of noncase comparisons limits etiologic inferences. However, comparison of unadjusted and adjusted case-only odds ratios can provide an indication as to whether the tumor characteristics are potentially important to racial differences in risk factors.

RESULTS

The prevalences of various risk factors among controls for younger (aged 20–49 years) and older (aged 50–74 years) African-American women and White women are shown in table 1. Results stratified by menopausal status were very similar to the age-stratified results presented here. Among younger women, numerous statistically significant differences in the distribution of risk factors were seen by race, with African-American women younger at first full-term pregnancy and having more births, less likelihood of breastfeeding, less but longer use of oral contraceptives, larger body size, higher waist/hip ratio, lower levels of education, less alcohol consumption, and less smoking. No significant differences were seen in the distribution of age at menarche, menopausal status or age at menopause, first-degree family history, hormone replacement therapy usage, induced abortion, or miscarriage, although some differences were observed for prevalence of age at menarche and family history. Among younger parous women, the distributions for time since last full-term pregnancy were identical for African-American and White women: 8 percent, less than or equal to 3 years; 20 percent, 4–10 years; and 72 percent, 11 or more years. In general, the same similarities and differences were observed among older women. Exceptions to the pattern seen for younger women were that older African-American women underwent menopause at earlier ages (on average) than did White women and were more likely to have breastfed and less likely to have used hormone replacement therapy than their White peers. Among older women, all except three reported their age since last full-term pregnancy of greater than or equal to 11 years, and these three were between 4 and 10 years.

Comparison of prevalence estimates between younger and older women revealed differences in reproductive and other behaviors. Regardless of race, younger women were more educated, delayed childbearing, and had fewer children. Use of oral contraceptives, induced abortion, and alcohol use were much more common among younger women. Among African Americans, breastfeeding and miscarriage were lower in the younger group. Younger Whites were more likely to have ever smoked than were older Whites, and older African Americans were more likely to have quit smoking than were younger African Americans. In addition, older women of both races had more abdominal obesity (as measured by waist/hip ratio) and were more likely to have used hormone replacement therapy.

In table 2, associations between breast cancer and various reproductive and lifestyle characteristics are shown for younger women by race; differences in odds ratios by race were observed for most reproductive characteristics and also former smoking. Younger African-American women with average age at menarche (12–13 years) were at increased risk of breast cancer, as were their White counterparts (but not to the same degree). For younger White women, the risk increased with later age at first pregnancy, and nulliparous women were at greater risk than were parous women. In contrast, among African-American women, later age at first pregnancy did not increase risk, and nulliparous African Americans were at slightly reduced risk versus those who had a first pregnancy before age 25 years and those with three or more children. African-American women who breastfed were at reduced risk of breast cancer, but White women were not (pinteraction < 0.12). An interaction with race was also suggested for age at menopause, with odds ratios farther from 1.0 at both younger (aged ≤44 years) and later (aged 45–49 years) ages among younger White women than younger African-American women (pinteraction < 0.19). In addition, former African-American smokers were at elevated risk, but former White smokers were not (pinteraction < 0.06). Patterns and magnitudes of risk were similar by race for family history of breast cancer, body size, waist/hip ratio, education, oral contraceptive and hormone replacement therapy use, alcohol consumption, and induced and spontaneous abortion.

Associations among older women were more similar by race for reproductive factors, but there was some evidence that they varied for nulliparity, oral contraceptive use, hormone replacement therapy use, and spontaneous abortion (table 3). As was true for younger women, older African-American women contrasted with their White counterparts by not having increased risk from late age at first full-term pregnancy. Among nulliparous older women, however, both races were at higher risk than parous women, and the odds ratio for nulliparous African-American women was 50 percent higher than the odds ratio for their White counterparts. The odds ratio for extended oral contraceptive use was 90 percent higher among older African-American women than White women. In contrast, hormone replacement therapy use was significantly inversely associated for African-American women (odds ratio (OR) = 0.6, 95 percent confidence interval (CI): 0.4, 0.8) but had no association for White women (pinteraction < 0.07). For spontaneous abortion, the interaction by race was suggestive (p < 0.14), with the odds ratio for White women reporting one or more events equaling 0.7, whereas for African-American women there was essentially no association. Patterns and magnitudes of risk were similar for other risk factors, and 95 percent confidence intervals generally included 1.0, except for the twofold increase in risk with later age at menopause (≥50 years) in both races.

Among both younger and older women with breast cancer, African Americans were more likely than Whites to be diagnosed with advanced stage disease, and their tumors were more likely to be negative for hormone receptors (table 4). Case-only analyses were used to determine whether these differences in breast cancer characteristics might explain, at least in part, racial differences in associations of risk factors with disease. However, adjustment for both stage and hormone receptor status had minimal impact on case-case odds ratios, including odds ratios for all of the reproductive and lifestyle risk factors found to have associations with breast cancer that differed by race in case-control analyses (data not shown). Of the 15 risk factors assessed (comprising 26 odds ratios for various levels of exposure), only two odds ratios adjusted for estrogen receptor and progesterone receptor status among younger women and one odds ratio adjusted for stage among older women varied by 15 percent from their respective values prior to these adjustments. Adjustment for both stage and hormone receptor status made no further difference, none of these odds ratios differed by more than 20 percent from the comparison value, and all resulted in attenuated differences in risk factors between African-American and White cases.

DISCUSSION

A crossover in breast cancer incidence at 45–49 years of age has been observed for African-American women and White women, with African Americans having the higher rate at younger ages and lower incidence after age 50 years relative to White women (7, 22). In a recent review of the literature, Pathak et al. (23) concluded that this crossover should be expected rather than being considered an anomaly, given the age-specific distributions for age at first full-term pregnancy and parity in the two racial groups. The expectation of crossover is based on the theory that a transient increase in risk follows pregnancy, and a subsequent long-term protective effect follows, generally for 10 years postpregnancy, because of differentiation of the breast ducts. Because African-American women have more children at younger ages, the risk of breast cancer for younger African-American women would be expected to be higher than that of White women of similar age (23). Our results expand upon the hypothesis of Pathak et al. by confirming racial differences in the distributions of several risk factors (also reported recently by Bernstein et al. (24)) as well as uncovering possible age-specific differences between African-American and White women in the magnitude of risk conferred by certain risk factors. We propose that the combined effect of a greater prevalence of some risk factors and relatively higher odds for breast cancer from certain risk factors may be responsible for a greater number of breast cancer cases among young African-American women.

Several reproductive risk factors were among those with apparent variations in prevalence and patterns of breast cancer risk between African Americans and Whites. Older age at first full-term pregnancy was associated with a modest increased risk among White women, with nulliparous women being at highest risk, as is traditionally reported. This was evident for both younger and older women among Whites; however, the pattern did not hold for African-American women. Among younger African Americans, no associations with older age at first full-term pregnancy or nulliparity were observed. Among older African-American women, increasing age at first full-term pregnancy showed no association with breast cancer, but nulliparous women were at twice the risk of parous women. Palmer et al. (25) also found no elevation of risk due to late age at first birth in older African-American women. Brinton et al. (6) obtained similar results, reporting that nulliparous younger African-American women were at reduced risk of breast cancer relative to multiparous women. In addition, Laing et al. (4) reported no increased breast cancer risk for nulliparous African-American women compared with multiparous African-American women. Hence, it appears that neither age at first full-term pregnancy nor nulliparity helps to explain the observed racial pattern of breast cancer incidence.

Multiparity is often reported to reduce breast cancer risk in comparison with women having no children (26). We observed this relation among both younger and older White women, although there was no evidence of a dose-response trend. Again, however, relations among African-American women tended to vary by age. While a similar inverse relation was observed for older African Americans, younger African-American women with three or more children had an odds ratio 40–50 percent higher (not statistically significant, however) than the odds ratio for their nulliparous peers. They also were almost twice as likely to have families of this size as younger White women. These results replicate the findings of Palmer et al. (25), who documented that parity is associated with increased risk in younger African-American women but with decreased risk among older women. An age-dependent effect of increasing parity on breast cancer risk has been reported previously (2729). In one study, Bruzzi et al. (27) provided evidence for a positive association between breast cancer and increasing parity among women younger than age 40 years with two or more children, but (as with our data for African Americans) the trend was not significant, while they reported an inverse relation for older women. These authors also reported a transient increase in breast cancer risk following full-term pregnancy; a relative risk of 2.66 (95 percent CI: 1.31, 5.39) was obtained for women having delivered a child within 3 years versus women whose last child was born 10 or more years earlier. If younger African-American women have children at intervals of 3 years or less, the transient risk period could be prolonged. In addition, if having multiple children actually increases risk of breast cancer for young African-American women, the higher prevalence of this factor among African Americans could serve to elevate risk for breast cancer in this population, consistent with the higher incidence of breast cancer reported for younger African-American women.

If, as a consequence of higher parity, young African-American women were more likely to have a recent pregnancy (≤10 years previously) than were young White women, the relevant exposure may be time since last full-term pregnancy. These age-specific differences in odds ratios between African-American and White women might be expected if age since last full-term pregnancy is the relevant exposure and if younger African-American women were more likely to have a recent pregnancy. However, among both younger and older parous women in our study, there were no racial differences in distribution of time since last full-term pregnancy. Additionally, when the variable years since last full-term pregnancy was substituted for age at first full-term pregnancy, we saw the expected trend of an increased risk followed by a progressive reduction in odds ratios for younger White women, but no association was seen for younger African-American women (data not shown). Moreover, the odds ratios for parity were virtually unchanged after adjustment (data not shown). We had insufficient variability in time since last full-term pregnancy to conduct similar analyses for older women. Nevertheless, these results provide additional evidence for variation by race in relations between breast cancer and characteristics of reproductive history among younger women.

Previous results from the Carolina Breast Cancer Study on the effect of breastfeeding on breast cancer risk showed a 30 percent reduction for parous women who had ever breastfed relative to parous women who had not (30). Our race-specific analyses showed younger African-American women to be at greater reduced risk (OR = 0.6, 95 percent CI: 0.4, 0.8) than their White peers (OR = 1.0, 95 percent CI: 0.7, 1.4). The inverse association also was stronger among younger African-American women than their older counterparts (OR = 0.6 and OR = 0.8 for younger and older, respectively). We noted that, among older women, the prevalence of breastfeeding was higher among African-American than White women. In contrast, only 20 percent of younger African-American women, almost half the proportion of younger White women, had breastfed. Thus, many younger African-American women may not be benefiting from the potential protective effects of lactation.

Selection bias may have influenced our results, as response rates varied across age, race, and case status (31). Women who declined an in-person interview were asked to complete a brief telephone survey on basic breast cancer risk factors. Both cases and controls who responded only to the telephone survey were older, had an earlier age at first full-term pregnancy, and had less education, oral contraceptive use, and hormone replacement therapy use than women who participated fully. However, the differences were limited and in the same direction for cases and controls, minimizing the concern about selection bias (31). Moreover, comparisons by race for prevalence of risk factors when based only on women with complete data provide conservative estimates of differences. A further refinement took into account differences in disease characteristics between African Americans and Whites in supplementary case-only analyses. Adjustment for stage at disease diagnosis and hormone receptor status of the breast cancer failed to reduce racial differences in risk factor profiles among cases, suggesting that any differences in etiologic pathways are unlikely to be explained by stage at diagnosis or estrogen and progesterone receptor status alone.

Even with the relatively large numbers of cases and controls in this study, the sample sizes were modest for many comparisons when analyses were stratified by race and age; hence, confidence intervals were broad and firm conclusions not possible. Although the findings we highlighted did not represent statistically significant differences by race, we thought it was important to focus on point estimates to discern whether any patterns may emerge. This was particularly of interest because we had roughly comparable numbers of African-American and White women from the same geographic area. With the increasing availability of studies that have included reasonable numbers of African-American women (16, 25), a meta-analysis of results may provide more meaningful estimates of risk factor profiles by race.

In conclusion, published reports on the etiology of breast cancer among African-American women are sparse and often conclude that their risk factors are similar to those for White women. Our results, however, show racial differences in the prevalence of most risk factors. Furthermore, the associations between breast cancer and some of these factors appear to vary in magnitude and direction by race and age. Most of the variations in risk factors occurred among women under 50 years of age, and the observed differences cannot be explained by either stage at diagnosis or hormone receptor status. Our results therefore demonstrate the importance of assessing modification by age when discerning risk estimates, and they support the hypothesis that racial variations in risk combined with racial differences in prevalence for particular risk factors may contribute to the higher incidence of breast cancer among younger African-American women.

ACKNOWLEDGMENTS

This study was supported by the Specialized Program of Research Excellence (SPORE) in Breast Cancer (P50-CA58223) from the National Cancer Institute.

The authors thank the nurses of the Carolina Breast Cancer Study for their diligence in collecting questionnaire and medical record data and body measurements and Dr. Wen-Yi Huang and Dr. Lynn Dressler for analysis of the estrogen and progesterone receptor status on archived tumor tissue. A special acknowledgment is warranted for Jessica Tse for guidance and assistance with statistical analysis.

TABLE 1.

Prevalence* of breast cancer risk factors among controls, by race and age, Carolina Breast Cancer Study, 1993–2001

 20–49 years p value  50–74 years p value 
 African American  White   African American  White  
 No.  No.   No.  No.  
Age group (years)              
20–34 39 37  35 49         
35–44 172 42  189 34         
45–49 153 21  180 17         
50–54        80 30  92 31  
55–64        148 40  145 37  
65–74        126 30  177 32  
Age at menarche (years)              
≤11 91 25  72 16   76 22  61 16  
12–13 176 49  244 51   170 49  232 56  
≥14 97 26  87 33 0.36  105 28  117 27 0.13 
Missing        
Menopausal status†              
Pre-/perimenopausal 311 92  352 94   46 17  66 22  
Postmenopausal, ≤44 years 37  39   133 35  122 28  
Postmenopausal, 45–49 years 16  13 0.31  83 23  117 27  
Postmenopausal, ≥50 years        92 25  109 23 0.04 
Age (years) at first full-term pregnancy               
≤24 243 60  190 37   273 78  271 65  
25–29 39 11  98 20   33 10  81 20  
≥30 34  55 11   14  30  
Nulliparous 47 21  61 32 <0.01  32  32 <0.01 
Missing          
Parity              
47 21  61 32   32  32  
1–2 192 49  269 53   108 32  193 48  
3–4 102 26  68 14   92 28  155 37  
≥5 23  <0.01  122 32  34 <0.01 
Breastfed              
Never 278 79  231 63   184 55  246 62  
Ever 86 21  173 37 <0.01  170 45  168 38 0.04 
Oral contraceptive use (months)              
<3 74 20  50 10   228 61  205 42  
3–59 135 37  147 44   80 25  119 33  
60–120 97 23  133 32   34 11  54 15  
>120 56 20  71 14 <0.03  11  31 <0.01 
Missing      
Family history of breast cancer in first-degree relative              
No 319 86  354 77   293 83  344 83  
Yes 34  47 23 0.21  44 12  52 13 0.90 
Missing 11    17  18  
Body mass index (kg/m2)‡              
<25 57 14  188 53   41 11  168 41  
≥25 294 82  208 46 <0.01  303 86  241 58 <0.01 
Missing 13    10   
Waist/hip ratio§              
<0.81 166 53  292 74   112 32  245 60  
≥0.81 190 45  108 25 0.01  237 67  165 39 <0.01 
Missing      
Education (years)              
≤12 178 52  120 24   250 68  165 37  
13–15 102 24  121 41   60 18  145 36  
≥16 84 24  163 35 <0.01  43 13  104 27 <0.01 
Missing            
Hormone replacement              
Never 313 92  334 91   236 66  176 41  
Ever 51  70 0.78  118 34  238 59 <0.01 
Alcohol              
Never 116 39  68 16   160 44  152 35  
Ever 247 61  336 84 <0.01  194 56  262 65 0.02 
Missing            
Smoking              
Never 214 67  196 40   211 58  202 49  
Former 61 11  121 40   91 26  135 32  
Current 89 22  87 20 <0.01  52 16  77 19 <0.05 
Abortion              
Induced              
Never 301 82  334 81   334 94  396 95  
Ever 63 18  70 19 0.77  20  18 0.41 
Spontaneous              
Never 266 73  311 80   229 66  291 71  
Ever 98 27  93 20 0.33  125 34  123 29 0.13 
 20–49 years p value  50–74 years p value 
 African American  White   African American  White  
 No.  No.   No.  No.  
Age group (years)              
20–34 39 37  35 49         
35–44 172 42  189 34         
45–49 153 21  180 17         
50–54        80 30  92 31  
55–64        148 40  145 37  
65–74        126 30  177 32  
Age at menarche (years)              
≤11 91 25  72 16   76 22  61 16  
12–13 176 49  244 51   170 49  232 56  
≥14 97 26  87 33 0.36  105 28  117 27 0.13 
Missing        
Menopausal status†              
Pre-/perimenopausal 311 92  352 94   46 17  66 22  
Postmenopausal, ≤44 years 37  39   133 35  122 28  
Postmenopausal, 45–49 years 16  13 0.31  83 23  117 27  
Postmenopausal, ≥50 years        92 25  109 23 0.04 
Age (years) at first full-term pregnancy               
≤24 243 60  190 37   273 78  271 65  
25–29 39 11  98 20   33 10  81 20  
≥30 34  55 11   14  30  
Nulliparous 47 21  61 32 <0.01  32  32 <0.01 
Missing          
Parity              
47 21  61 32   32  32  
1–2 192 49  269 53   108 32  193 48  
3–4 102 26  68 14   92 28  155 37  
≥5 23  <0.01  122 32  34 <0.01 
Breastfed              
Never 278 79  231 63   184 55  246 62  
Ever 86 21  173 37 <0.01  170 45  168 38 0.04 
Oral contraceptive use (months)              
<3 74 20  50 10   228 61  205 42  
3–59 135 37  147 44   80 25  119 33  
60–120 97 23  133 32   34 11  54 15  
>120 56 20  71 14 <0.03  11  31 <0.01 
Missing      
Family history of breast cancer in first-degree relative              
No 319 86  354 77   293 83  344 83  
Yes 34  47 23 0.21  44 12  52 13 0.90 
Missing 11    17  18  
Body mass index (kg/m2)‡              
<25 57 14  188 53   41 11  168 41  
≥25 294 82  208 46 <0.01  303 86  241 58 <0.01 
Missing 13    10   
Waist/hip ratio§              
<0.81 166 53  292 74   112 32  245 60  
≥0.81 190 45  108 25 0.01  237 67  165 39 <0.01 
Missing      
Education (years)              
≤12 178 52  120 24   250 68  165 37  
13–15 102 24  121 41   60 18  145 36  
≥16 84 24  163 35 <0.01  43 13  104 27 <0.01 
Missing            
Hormone replacement              
Never 313 92  334 91   236 66  176 41  
Ever 51  70 0.78  118 34  238 59 <0.01 
Alcohol              
Never 116 39  68 16   160 44  152 35  
Ever 247 61  336 84 <0.01  194 56  262 65 0.02 
Missing            
Smoking              
Never 214 67  196 40   211 58  202 49  
Former 61 11  121 40   91 26  135 32  
Current 89 22  87 20 <0.01  52 16  77 19 <0.05 
Abortion              
Induced              
Never 301 82  334 81   334 94  396 95  
Ever 63 18  70 19 0.77  20  18 0.41 
Spontaneous              
Never 266 73  311 80   229 66  291 71  
Ever 98 27  93 20 0.33  125 34  123 29 0.13 

* Frequencies adjusted for sampling weights. Totals may not add to 100 because of rounding.

† Women were categorized as postmenopausal if they reported natural menopause or bilateral oophorectomy, or if they were more than 55 years of age and had a hysterectomy. Women who reported still having menstrual cycles or who had at least one remaining ovary and were less than 42 years of age were classified as premenopausal. Women who had a hysterectomy without bilateral oophorectomy and who were 42–55 years of age were considered perimenopausal. The variable also captures age at menopause of postmenopausal women.

‡ Body mass index of less than 25 was defined as normal weight by the National Heart, Lung, and Blood Institute (32).

§ Median value among controls = 0.81.

TABLE 2.

Race-specific, adjusted odds ratios for breast cancer and reproductive and lifestyle risk factors among women aged 20–49 years, Carolina Breast Cancer Study, 1993–2001

 African American  White 
 (n = 753)  (n = 972) 
 Cases  Controls  Cases  Controls 
 (n = 389)  (n = 364)  (n = 568)  (n = 404) 
 No.  No. Adjusted odds ratio* 95% confidence interval  No.  No. Adjusted odds ratio 95% confidence interval 
Age group (years)                
20–34 56 14  39 11    61 11  35   
35–44 182 47  172 47    263 46  189 47   
45–49 151 39  153 42    244 43  180 45   
Age at menarche (years)                
≥14 79 20  97 27 1.0   105 18  87 22 1.0  
12–13 210 54  176 48 1.6 0.9, 2.4  333 59  244 60 1.2 0.8, 1.7 
≤11 100 26  91 25 1.6 1.0, 2.4  129 23  72 18 1.7 1.1, 2.6 
Missing            
Menopausal status†                
Pre-/perimenopausal 343 88  311 85 1.0   511 90  352 87 1.0  
Postmenopausal, ≤44 years 29  37 10 0.7 0.4, 1.3  27  39 10 0.4 0.2, 0.8 
Postmenopausal, 45–49 years 17  16 0.8 0.4, 1.7  30  13 1.4 0.7, 2.9 
Age at first full-term pregnancy (years)‡                
≤24 263 67  243 67 1.0   240 42  190 47 1.0  
25–29 44 11  39 11 1.2 0.7, 2.0  130 23  98 24 1.1 0.7, 1.6 
≥30 28  34 0.7 0.4, 1.4  88 15  55 14 1.3 0.9, 2.1 
Nulliparous 54 14  47 13 0.8 0.5, 1.3  110 19  61 15 1.4 0.9, 2.2 
Missing              
Parity                
54 14  47 13 1.0   110 19  61 15 1.0  
1–2 192 49  192 53 1.1 0.7, 1.8  363 64  269 67 0.8 0.5, 1.2 
3–4 120 31  102 28 1.5 0.9, 2.6  88 15  68 17 0.7 0.4, 1.2 
≥5 23  23 1.4 0.6, 3.1   0.8 0.2, 3.0 
Breastfed                
Never 324 83  278 76 1.0   341 60  231 57 1.0  
Ever 65 17  86 24 0.6 0.4, 0.8  227 40  173 43 1.0 0.7, 1.4 
Oral contraceptive use (months)                
<3 71 18  74 20 1.0   68 12  50 12 1.0  
3–59 141 36  135 37 1.1 0.7, 1.8  214 38  147 36 1.2 0.8, 1.8 
60–120 111 29  97 27 1.1 0.7, 1.7  183 32  133 33 1.0 0.7. 1.7 
>120 66 17  56 15 1.3 0.8, 2.2  101 18  71 18 1.1 0.7, 1.9 
Missing          
Family history of breast cancer in first-degree relative                
No 319 82  319 88 1.0   460 81  354 88 1.0  
Yes 61 16  34 2.2 1.3, 3.5  92 16  47 12 1.7 1.1, 2.5 
Missing  11    16    
Body mass index (kg/m2)§                
<25 77 20  57 16 1.0   308 54  188 47 1.0  
≥25 305 78  294 81 0.7 0.5, 1.1  252 44  208 51 0.7 0.5, 0.9 
Missing  13       
Waist/hip ratio¶                
<0.81 156 40  166 46 1.0   393 69  292 72 1.0  
≥0.81 229 59  190 52 1.4 1.0, 1.9  166 29  108 27 1.4 1.0, 2.0 
Missing        
Education (years)                
≤12 192 49  178 49 1.0   162 28  120 30 1.0  
13–15 109 28  102 28 1.1 0.7, 1.6  164 29  121 30 0.9 0.7, 1.4 
≥16 88 23  84 23 1.1 0.7, 1.8  242 43  167 40 1.0 0.7, 1.4 
Missing                
Hormone replacement                
Never 342 88  313 86 1.0   475 84  334 83 1.0  
Ever 47 12  51 14 1.0 0.6, 1.6  16  70 17 1.1 0.7, 1.7 
Alcohol                
Never 127 33  116 32 1.0   108 19  68 17 1.0  
Ever 262 67  247 68 0.9 0.6, 1.3  460 81  336 83 0.8 0.6, 1.2 
Missing              
Smoking                
Never 218 56  214 59 1.0   305 54  196 49 1.0  
Former 90 23  61 17 1.8 1.2, 2.7  169 30  121 30 0.9 0.7, 1.3 
Current 81 21  89 24 1.0 0.6, 1.4  94 17  87 22 0.7 0.5, 1.0 
Abortion                
Induced                
Never 305 78  301 83 1.0   465 82  334 83 1.0  
Ever 84 22  63 17 1.2 0.8, 1.8  103 18  70 17 1.0 0.7, 1.4 
Spontaneous                
Never 293 75  266 73 1.0   441 78  311 77 1.0  
Ever 93 25  98 27 1.0 0.7, 1.4  127 22  93 23 1.1 0.8, 1.5 
 African American  White 
 (n = 753)  (n = 972) 
 Cases  Controls  Cases  Controls 
 (n = 389)  (n = 364)  (n = 568)  (n = 404) 
 No.  No. Adjusted odds ratio* 95% confidence interval  No.  No. Adjusted odds ratio 95% confidence interval 
Age group (years)                
20–34 56 14  39 11    61 11  35   
35–44 182 47  172 47    263 46  189 47   
45–49 151 39  153 42    244 43  180 45   
Age at menarche (years)                
≥14 79 20  97 27 1.0   105 18  87 22 1.0  
12–13 210 54  176 48 1.6 0.9, 2.4  333 59  244 60 1.2 0.8, 1.7 
≤11 100 26  91 25 1.6 1.0, 2.4  129 23  72 18 1.7 1.1, 2.6 
Missing            
Menopausal status†                
Pre-/perimenopausal 343 88  311 85 1.0   511 90  352 87 1.0  
Postmenopausal, ≤44 years 29  37 10 0.7 0.4, 1.3  27  39 10 0.4 0.2, 0.8 
Postmenopausal, 45–49 years 17  16 0.8 0.4, 1.7  30  13 1.4 0.7, 2.9 
Age at first full-term pregnancy (years)‡                
≤24 263 67  243 67 1.0   240 42  190 47 1.0  
25–29 44 11  39 11 1.2 0.7, 2.0  130 23  98 24 1.1 0.7, 1.6 
≥30 28  34 0.7 0.4, 1.4  88 15  55 14 1.3 0.9, 2.1 
Nulliparous 54 14  47 13 0.8 0.5, 1.3  110 19  61 15 1.4 0.9, 2.2 
Missing              
Parity                
54 14  47 13 1.0   110 19  61 15 1.0  
1–2 192 49  192 53 1.1 0.7, 1.8  363 64  269 67 0.8 0.5, 1.2 
3–4 120 31  102 28 1.5 0.9, 2.6  88 15  68 17 0.7 0.4, 1.2 
≥5 23  23 1.4 0.6, 3.1   0.8 0.2, 3.0 
Breastfed                
Never 324 83  278 76 1.0   341 60  231 57 1.0  
Ever 65 17  86 24 0.6 0.4, 0.8  227 40  173 43 1.0 0.7, 1.4 
Oral contraceptive use (months)                
<3 71 18  74 20 1.0   68 12  50 12 1.0  
3–59 141 36  135 37 1.1 0.7, 1.8  214 38  147 36 1.2 0.8, 1.8 
60–120 111 29  97 27 1.1 0.7, 1.7  183 32  133 33 1.0 0.7. 1.7 
>120 66 17  56 15 1.3 0.8, 2.2  101 18  71 18 1.1 0.7, 1.9 
Missing          
Family history of breast cancer in first-degree relative                
No 319 82  319 88 1.0   460 81  354 88 1.0  
Yes 61 16  34 2.2 1.3, 3.5  92 16  47 12 1.7 1.1, 2.5 
Missing  11    16    
Body mass index (kg/m2)§                
<25 77 20  57 16 1.0   308 54  188 47 1.0  
≥25 305 78  294 81 0.7 0.5, 1.1  252 44  208 51 0.7 0.5, 0.9 
Missing  13       
Waist/hip ratio¶                
<0.81 156 40  166 46 1.0   393 69  292 72 1.0  
≥0.81 229 59  190 52 1.4 1.0, 1.9  166 29  108 27 1.4 1.0, 2.0 
Missing        
Education (years)                
≤12 192 49  178 49 1.0   162 28  120 30 1.0  
13–15 109 28  102 28 1.1 0.7, 1.6  164 29  121 30 0.9 0.7, 1.4 
≥16 88 23  84 23 1.1 0.7, 1.8  242 43  167 40 1.0 0.7, 1.4 
Missing                
Hormone replacement                
Never 342 88  313 86 1.0   475 84  334 83 1.0  
Ever 47 12  51 14 1.0 0.6, 1.6  16  70 17 1.1 0.7, 1.7 
Alcohol                
Never 127 33  116 32 1.0   108 19  68 17 1.0  
Ever 262 67  247 68 0.9 0.6, 1.3  460 81  336 83 0.8 0.6, 1.2 
Missing              
Smoking                
Never 218 56  214 59 1.0   305 54  196 49 1.0  
Former 90 23  61 17 1.8 1.2, 2.7  169 30  121 30 0.9 0.7, 1.3 
Current 81 21  89 24 1.0 0.6, 1.4  94 17  87 22 0.7 0.5, 1.0 
Abortion                
Induced                
Never 305 78  301 83 1.0   465 82  334 83 1.0  
Ever 84 22  63 17 1.2 0.8, 1.8  103 18  70 17 1.0 0.7, 1.4 
Spontaneous                
Never 293 75  266 73 1.0   441 78  311 77 1.0  
Ever 93 25  98 27 1.0 0.7, 1.4  127 22  93 23 1.1 0.8, 1.5 

* These odds ratios were obtained from a logistic regression model adjusted for an offset term, age (continuous), and all other variables in the table except age at first full-term pregnancy.

† Women were categorized as postmenopausal if they reported natural menopause or bilateral oophorectomy, or if they were more than 55 years of age and had a hysterectomy. Women who reported still having menstrual cycles or who had at least one remaining ovary and were less than 42 years of age were classified as premenopausal. Women who had a hysterectomy without bilateral oophorectomy and who were 42–55 years of age were considered perimenopausal. The variable also captures age at menopause of postmenopausal women.

‡ Age at first full-term pregnancy was entered separately into a model containing all the other variables except parity.

§ Body mass index of less than 25 was defined as normal weight by the National Heart, Lung, and Blood Institute (32). The logistic model for body mass index did not include waist/hip ratio.

¶ Median value among controls = 0.81.

TABLE 3.

Race-specific, adjusted odds ratios for breast cancer and reproductive and lifestyle risk factors among women aged 50–74 years, Carolina Breast Cancer Study, 1993–2001

 African American  White 
 (n = 752)  (n = 837) 
 Cases  Controls  Cases  Controls 
 (n = 398)  (n = 354)  (n = 423)  (n = 414) 
 No.  No. Adjusted odds ratio* 95% confidence interval  No.  No. Adjusted odds ratio 95% confidence interval 
Age group (years)                
50–54 87 22  80 23    84 20  92 22   
55–64 164 41  148 42    169 40  145 35   
65–74 147 37  126 36    170 40  177 43   
Age at menarche (years)                
≥14 105 26  105 30 1.0   105 25  117 28 1.0  
12–13 193 48  170 48 1.2 0.8, 1.8  245 58  232 56 1.2 0.9, 1.7 
≤11 99 25  76 21 1.3 0.8, 2.0  73 17  61 15 1.5 0.9, 2.3 
Missing          
Menopausal status†                
Pre-/perimenopausal 38 10  46 13 1.0   51 12  66 16 1.0  
Postmenopausal, ≤44 years 129 32  133 38 1.4 0.7, 2.5  110 26  122 29 1.5 0.8, 2.6 
Postmenopausal, 45–49 years 93 23  83 23 1.6 0.8, 3.0  123 29  117 28 1.7 1.0, 2.9 
Postmenopausal, ≥50 years 138 35  92 26 2.2 1.2, 4.0  139 33  109 26 2.2 1.3, 3.8 
Age at first full-term pregnancy (years)‡                
≤24 280 70  273 77 1.0   250 59  271 65 1.0  
25–29 37  33 1.0 0.6, 1.8  85 20  81 20 1.2 0.8, 1.8 
≥30 18  14 1.0 0.5, 2.3  39  30 1.4 0.8, 2.5 
Nulliparous 57 14  32 2.0 1.1, 3.6  48 11  32 1.5 0.9, 2.6 
Missing          
Parity                
57 14  32 1.0   48 11  32 1.0  
1–2 109 27  108 31 0.5 0.3, 0.9  194 46  193 47 0.7 0.4, 1.2 
3–4 108 27  92 26 0.5 0.3, 1.0  152 36  155 37 0.7 0.4, 1.2 
≥5 124 31  122 34 0.5 0.3, 0.9  29  34 0.8 0.4, 1.6 
Breastfed                
Never 236 59  184 52 1.0   261 62  246 59 1.0  
Ever 162 41  170 48 0.8 0.6, 1.2  162 38  168 41 0.9 0.7, 1.3 
Oral contraceptive use (months)                
<3 259 65  228 64 1.0   214 51  205 50 1.0  
3–59 71 18  80 23 0.8 0.5, 1.3  101 24  119 29 0.8 0.6, 1.2 
60–120 45 11  34 10 1.4 0.8, 2.5  67 16  54 13 1.1 0.7, 1.8 
>120 23  11 2.1 0.9, 4.8  37  31 1.2 0.7, 2.2 
Missing            
Family history of breast cancer in first-degree relative                
No 325 82  293 83 1.0   334 79  344 83 1.0  
Yes 57 14  44 12 1.2 0.8, 1.9  78 18  52 13 1.5 1.0, 2.2 
Missing 16  17    11  18   
Body mass index (kg/m2)§                
<25 55 14  41 12 1.0   172 41  168 41 1.0  
≥25 323 81  303 86 0.8 0.5, 1.5  247 58  241 58 0.9 0.7, 1.3 
Missing 20  100       
Waist/hip ratio¶                
<0.81 104 26  112 32 1.0   232 55  245 59 1.0  
≥0.81 285 72  237 67 1.5 1.0, 3.0  187 44  165 40 1.3 0.9, 1.7 
Missing        
Education (years)                
≤12 283 71  250 71 1.0   183 43  165 40 1.0  
13–15 65 16  60 17 1.0 0.7, 1.6  124 29  145 35 0.8 0.6, 1.1 
≥16 50 13  43 12 0.9 0.5, 1.5  116 27  104 25 1.0 0.7, 1.4 
Missing              
Hormone replacement                
Never 290 73  236 67 1.0   185 44  176 43 1.0  
Ever 108 27  118 33 0.6 0.4, 0.8  238 56  238 57 0.9 0.7, 1.2 
Alcohol                
Never 166 42  160 45 1.0   155 37  152 37 1.0  
Ever 232 58  194 55 1.1 0.7, 1.5  267 63  262 63 1.0 0.7, 1.5 
Missing              
Smoking                
Never 218 55  211 60 1.0   187 44  202 49 1.0  
Former 126 32  91 26 1.2 0.8, 1.8  173 41  135 33 1.4 1.0, 1.9 
Current 54 14  52 15 0.9 0.5, 1.4  63 15  77 19 0.8 0.5, 1.2 
Abortion                
Induced                
Never 380 95  334 94 1.0   403 95  396 96 1.0  
Ever 18  20 1.0 0.5, 2.0  20  18 1.0 0.5, 2.1 
Spontaneous                
Never 251 63  229 65 1.0   324 77  291 70 1.0  
Ever 147 37  125 35 1.1 0.8, 1.6  99 23  123 30 0.7 0.5, 1.0 
 African American  White 
 (n = 752)  (n = 837) 
 Cases  Controls  Cases  Controls 
 (n = 398)  (n = 354)  (n = 423)  (n = 414) 
 No.  No. Adjusted odds ratio* 95% confidence interval  No.  No. Adjusted odds ratio 95% confidence interval 
Age group (years)                
50–54 87 22  80 23    84 20  92 22   
55–64 164 41  148 42    169 40  145 35   
65–74 147 37  126 36    170 40  177 43   
Age at menarche (years)                
≥14 105 26  105 30 1.0   105 25  117 28 1.0  
12–13 193 48  170 48 1.2 0.8, 1.8  245 58  232 56 1.2 0.9, 1.7 
≤11 99 25  76 21 1.3 0.8, 2.0  73 17  61 15 1.5 0.9, 2.3 
Missing          
Menopausal status†                
Pre-/perimenopausal 38 10  46 13 1.0   51 12  66 16 1.0  
Postmenopausal, ≤44 years 129 32  133 38 1.4 0.7, 2.5  110 26  122 29 1.5 0.8, 2.6 
Postmenopausal, 45–49 years 93 23  83 23 1.6 0.8, 3.0  123 29  117 28 1.7 1.0, 2.9 
Postmenopausal, ≥50 years 138 35  92 26 2.2 1.2, 4.0  139 33  109 26 2.2 1.3, 3.8 
Age at first full-term pregnancy (years)‡                
≤24 280 70  273 77 1.0   250 59  271 65 1.0  
25–29 37  33 1.0 0.6, 1.8  85 20  81 20 1.2 0.8, 1.8 
≥30 18  14 1.0 0.5, 2.3  39  30 1.4 0.8, 2.5 
Nulliparous 57 14  32 2.0 1.1, 3.6  48 11  32 1.5 0.9, 2.6 
Missing          
Parity                
57 14  32 1.0   48 11  32 1.0  
1–2 109 27  108 31 0.5 0.3, 0.9  194 46  193 47 0.7 0.4, 1.2 
3–4 108 27  92 26 0.5 0.3, 1.0  152 36  155 37 0.7 0.4, 1.2 
≥5 124 31  122 34 0.5 0.3, 0.9  29  34 0.8 0.4, 1.6 
Breastfed                
Never 236 59  184 52 1.0   261 62  246 59 1.0  
Ever 162 41  170 48 0.8 0.6, 1.2  162 38  168 41 0.9 0.7, 1.3 
Oral contraceptive use (months)                
<3 259 65  228 64 1.0   214 51  205 50 1.0  
3–59 71 18  80 23 0.8 0.5, 1.3  101 24  119 29 0.8 0.6, 1.2 
60–120 45 11  34 10 1.4 0.8, 2.5  67 16  54 13 1.1 0.7, 1.8 
>120 23  11 2.1 0.9, 4.8  37  31 1.2 0.7, 2.2 
Missing            
Family history of breast cancer in first-degree relative                
No 325 82  293 83 1.0   334 79  344 83 1.0  
Yes 57 14  44 12 1.2 0.8, 1.9  78 18  52 13 1.5 1.0, 2.2 
Missing 16  17    11  18   
Body mass index (kg/m2)§                
<25 55 14  41 12 1.0   172 41  168 41 1.0  
≥25 323 81  303 86 0.8 0.5, 1.5  247 58  241 58 0.9 0.7, 1.3 
Missing 20  100       
Waist/hip ratio¶                
<0.81 104 26  112 32 1.0   232 55  245 59 1.0  
≥0.81 285 72  237 67 1.5 1.0, 3.0  187 44  165 40 1.3 0.9, 1.7 
Missing        
Education (years)                
≤12 283 71  250 71 1.0   183 43  165 40 1.0  
13–15 65 16  60 17 1.0 0.7, 1.6  124 29  145 35 0.8 0.6, 1.1 
≥16 50 13  43 12 0.9 0.5, 1.5  116 27  104 25 1.0 0.7, 1.4 
Missing              
Hormone replacement                
Never 290 73  236 67 1.0   185 44  176 43 1.0  
Ever 108 27  118 33 0.6 0.4, 0.8  238 56  238 57 0.9 0.7, 1.2 
Alcohol                
Never 166 42  160 45 1.0   155 37  152 37 1.0  
Ever 232 58  194 55 1.1 0.7, 1.5  267 63  262 63 1.0 0.7, 1.5 
Missing              
Smoking                
Never 218 55  211 60 1.0   187 44  202 49 1.0  
Former 126 32  91 26 1.2 0.8, 1.8  173 41  135 33 1.4 1.0, 1.9 
Current 54 14  52 15 0.9 0.5, 1.4  63 15  77 19 0.8 0.5, 1.2 
Abortion                
Induced                
Never 380 95  334 94 1.0   403 95  396 96 1.0  
Ever 18  20 1.0 0.5, 2.0  20  18 1.0 0.5, 2.1 
Spontaneous                
Never 251 63  229 65 1.0   324 77  291 70 1.0  
Ever 147 37  125 35 1.1 0.8, 1.6  99 23  123 30 0.7 0.5, 1.0 

* These odds ratios were obtained from a logistic regression model adjusted for an offset term, age (continuous), and all the other variables in the table except age at first full-term pregnancy.

† Women were categorized as postmenopausal if they reported natural menopause or bilateral oophorectomy, or if they were more than 55 years of age and had a hysterectomy. Women who reported still having menstrual cycles or who had at least one remaining ovary and were less than 42 years of age were classified as premenopausal. Women who had a hysterectomy without bilateral oophorectomy and who were 42–55 years of age were considered perimenopausal. The variable also captures age at menopause of postmenopausal women.

‡ Age at first full-term pregnancy was entered separately into a model containing all the other variables except parity.

§ Body mass index of less than 25 was defined as normal weight by the National Heart, Lung, and Blood Institute (32). The logistic model for body mass index did not include waist/hip ratio.

¶ Median value among controls = 0.81.

TABLE 4.

Case-only ratios for tumor characteristics and race, stratified by age, Carolina Breast Cancer Study, 1993–2001

 20–49 years  50–74 years 
 African American  White Adjusted odds ratio* 95% confidence interval  African American  White Adjusted odds ratio 95% confidence interval 
 No.  No.    No.  No.   
Stage at diagnosis                
105 27  219 39 1.0   147 37  230 54 1.0  
II 194 50  248 44 1.7 1.2, 2.3  172 43  141 33 1.9 1.4, 2.5 
III, IV 71 18  68 12 2.4 1.6, 3.7  50 13  26 2.8 1.7, 4.8 
Missing 19  33    29  26   
ER†/PR† status                
ER+†/PR+† 107 27  255 45 1.0   162 41  225 53 1.0  
ER–†/PR–† 169 43  157 28 3.1 2.2, 4.3  114 29  69 16 2.2 1.5, 3.2 
Other‡ 56 14  76 13 1.9 1.3, 3.0  53 13  72 17 1.0 0.7. 1.5 
Missing 60 15  80 14    69 17  57 13   
 20–49 years  50–74 years 
 African American  White Adjusted odds ratio* 95% confidence interval  African American  White Adjusted odds ratio 95% confidence interval 
 No.  No.    No.  No.   
Stage at diagnosis                
105 27  219 39 1.0   147 37  230 54 1.0  
II 194 50  248 44 1.7 1.2, 2.3  172 43  141 33 1.9 1.4, 2.5 
III, IV 71 18  68 12 2.4 1.6, 3.7  50 13  26 2.8 1.7, 4.8 
Missing 19  33    29  26   
ER†/PR† status                
ER+†/PR+† 107 27  255 45 1.0   162 41  225 53 1.0  
ER–†/PR–† 169 43  157 28 3.1 2.2, 4.3  114 29  69 16 2.2 1.5, 3.2 
Other‡ 56 14  76 13 1.9 1.3, 3.0  53 13  72 17 1.0 0.7. 1.5 
Missing 60 15  80 14    69 17  57 13   

* The odds ratios were obtained from a logistic regression model adjusted for age (continuous), age at menarche, a combined variable of menopausal status and age at menopause, parity, breastfeeding, oral contraceptive use, body mass index, waist/hip ratio, education, hormone replacement, alcohol, smoking, induced and spontaneous abortion, and an offset term.

† ER, estrogen receptor; PR, progesterone receptor; ER+, estrogen receptor positive; PR+, progesterone receptor positive; ER–, estrogen receptor negative; PR–, progesterone receptor negative.

‡ “Other” refers to ER+/PR– and ER–/PR+.

Correspondence to Dr. Ingrid Hall, Epidemiology and Applied Research Branch, Division of Cancer Prevention and Control, Centers for Disease Control and Prevention, 4770 Buford Highway NE, MS K-55, Atlanta, GA 30341 (e-mail: iah9@cdc.gov).

References

1.
Austin H, Cole P, Wynder E. Breast cancer in black American women.
Int J Cancer
 
1979
;
24
:
541
–4.
2.
Schatzkin A, Palmer JR, Rosenberg L, et al. Risk factors for breast cancer in black women.
J Natl Cancer Inst
 
1987
;
78
:
213
–17.
3.
Mayberry RM, Stoddard-Wright C. Breast cancer risk factors among black women and white women: similarities and differences.
Am J Epidemiol
 
1992
;
136
:
1445
–56.
4.
Laing AE, Demenais FM, Williams R, et al. Breast cancer risk factors in African-American women: the Howard University tumor registry experience.
J Natl Med Assoc
 
1993
;
85
:
931
–9.
5.
Mayberry RM. Age-specific patterns of association between breast cancer and risk factors in black women, ages 20 to 39 and 40 to 54.
Ann Epidemiol
 
1994
;
4
:
205
–13.
6.
Brinton LA, Benichou J, Gammon MD, et al. Ethnicity and variation in breast cancer incidence.
Int J Cancer
 
1997
;
73
:
349
–55.
7.
Ries LAG, Eisner MP, Kosary CL, et al, eds. SEER cancer statistics review, 1975–2000. Bethesda, MD: National Cancer Institute, 2003. (http://seer.cancer.gov/csr/1975_2000, 2003).
8.
Aziz H, Hussain F, Sohn C, et al. Early onset of breast carcinoma in African American women with poor prognostic factors.
Am
 
J Clin Oncol
 
1999
;
22
:
436
–40.
9.
Crowe JP Jr, Gordon NH, Shenk RR, et al. Age does not predict breast cancer outcome.
Arch Surg
 
1994
;
129
:
483
–7.
10.
El-Tamer MB, Wait RB. Age at presentation of African-American and Caucasian breast cancer patients.
J Am Coll Surg
 
1999
;
188
:
237
–40.
11.
English WP, Cleveland KE, Barber WH. There is no difference in survival between African-American and white women with breast cancer.
Am Surg
 
2002
;
68
:
594
–7.
12.
Golledge J, Wiggins JE, Callam MJ. Age-related variation in the treatment and outcomes of patients with breast carcinoma.
Cancer
 
2000
;
88
:
369
–74.
13.
Bonnier P, Romain S, Charpin C, et al. Age as a prognostic factor in breast cancer: relationship to pathologic and biological factors.
Int J Cancer
 
1995
;
62
:
138
–44.
14.
Nixon AJ, Neuberg D, Hayes DF, et al. Relationship of patient age to pathologic features of the tumour and prognosis for patients with stage I or II breast cancer.
J Clin Oncol
 
1994
;
12
:
888
–94.
15.
Fowble BL, Schultz DJ, Overmoyer B, et al. The influence of young age on outcome in early stage breast cancer.
Int J Radiat Oncol Biol Phys
 
1994
;
30
:
23
–33.
16.
Furberg H, Millikan R, Dressler L, et al. Tumor characteristics in African American and white women.
Breast Cancer Res Treat
 
2001
;
68
:
33
–43.
17.
Newman B, Moorman PG, Millikan R, et al. The Carolina Breast Cancer Study: integrating population-based epidemiology and molecular biology.
Breast Cancer Res Treat
 
1995
;
35
:
51
–60.
18.
Weinberg CR, Sandler DP. Randomized recruitment in case-control studies.
Am J Epidemiol
 
1991
;
134
:
421
–32.
19.
Aldrich TE, Vann D, Moorman PG, et al. Rapid reporting of cancer incidence in a population-based study of breast cancer: one constructive use of a central cancer registry.
Breast Cancer Res Treat
 
1995
;
35
:
61
–4.
20.
Huang WY, Newman B, Millikan RC, et al. Hormone-related factors and risk of breast cancer in relation to estrogen receptor and progesterone receptor status.
Am J Epidemiol
 
2000
;
151
:
703
–14.
21.
Begg CB, Zhang ZF. Statistical analysis of molecular epidemiology studies employing case-series.
Cancer Epidemiol Biomarkers Prev
 
1994
;
3
:
173
–5.
22.
Ries LAG, Kosary CL, Handey BF, et al, eds. SEER cancer statistics review, 1973–1995. Bethesda, MD: National Cancer Institute, 1998.
23.
Pathak DR, Osuch JR, He J. Breast carcinoma etiology: current knowledge and new insights into the effects of reproductive and hormonal risk factors in black and white populations.
Cancer
 
2000
;
88(suppl)
:
120
–8.
24.
Bernstein L, Teal CR, Joslyn S, et al. Ethnicity-related variation in breast cancer risk factors.
Cancer
 
2003
;
97(suppl)
:
222
–9.
25.
Palmer JR, Wise LA, Horton NJ, et al. Dual effect of parity on breast cancer risk in African-American women.
J Natl Cancer Inst
 
2003
;
95
:
478
–83.
26.
Kelsey JL, Gammon MD, John EM. Reproductive factors and breast cancer.
Epidemiol Rev
 
1993
;
15
:
36
–47.
27.
Bruzzi P, Negri E, La Vecchia C, et al. Short term increase in risk of breast cancer after full term pregnancy.
BMJ
 
1988
;
297
:
1096
–8.
28.
Liu Q, Wuu J, Lambe M, et al. Transient increase in breast cancer risk after giving birth: postpartum period with the highest risk.
Cancer Causes Control
 
2002
;
13
:
299
–305.
29.
Lambe M, Hsieh CC, Trichopoulos D, et al. Transient increase in breast cancer risk after giving birth.
N Engl J Med
 
1994
;
331
:
5
–9.
30.
Furberg H, Newman B, Moorman P, et al. Lactation and breast cancer risk.
Int J Epidemiol
 
1999
;
28
:
396
–402.
31.
Moorman PG, Newman B, Millikan RC, et al. Participation rates in a case-control study: the impact of age, race, and race of interviewer.
Ann Epidemiol
 
1999
;
9
:
188
–95.
32.
National Heart, Lung, and Blood Institute. The practical guide: identification, evaluation, and treatment of overweight and obesity in adults. Bethesda, MD: National Heart, Lung, and Blood Institute, 2000. (NIH publication no. 00-4084) (http://www.nhlbi.nih.gov/guidelines/obesity/prctgd_c.pdf).