Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Breast Feeding, Parity and Breast Cancer Subtypes in a Spanish Cohort

  • Carmen M. Redondo ,

    Contributed equally to this work with: Carmen M. Redondo, Manuela Gago-Domínguez

    carmen.redondo.marey@sergas.es

    Affiliation Oncology and Genetics Unit, Genomic Medicine Group, Complejo Hospitalario Universitario de Vigo, Vigo, Spain

  • Manuela Gago-Domínguez ,

    Contributed equally to this work with: Carmen M. Redondo, Manuela Gago-Domínguez

    Affiliation Genomic Medicine Group, Galician Foundation of Genomic Medicine, Complejo Hospitalario Universitario de Santiago, IDIS, Santiago de Compostela, Spain

  • Sara Miranda Ponte,

    Affiliation Oncology and Genetics Unit, Genomic Medicine Group, Complejo Hospitalario Universitario de Vigo, Vigo, Spain

  • Manuel Enguix Castelo,

    Affiliation Radiotherapy Department, Complejo Hospitalario Universitario de Vigo, Vigo, Spain

  • Xuejuan Jiang,

    Affiliation Ophtalmology and Preventive Medicine Department, University of Southern California, Los Angeles, California, United States of America

  • Ana Alonso García,

    Affiliation Radiotherapy Department, Hospital Universitario Central de Asturias, Oviedo, Spain

  • Maite Peña Fernández,

    Affiliation Ginecology Department, Complejo Hospitalario Universitario de Santiago, Santiago de Compostela, Spain

  • María Ausencia Tomé,

    Affiliation Endocrinology Department, Complejo Hospitalario Universitario de Santiago, Santiago de Compostela, Spain

  • Máximo Fraga,

    Affiliation Pathology Department, Complejo Hospitalario Universitario de Santiago, Santiago de Compostela, Spain

  • Francisco Gude,

    Affiliation Clinical Epidemiology Unit, Complejo Hospitalario Universitario de Santiago, Santiago de Compostela, Spain

  • María Elena Martínez,

    Affiliation University of California San Diego, Moores Cancer Center, San Diego, California, United Sates of America

  • Víctor Muñoz Garzón,

    Affiliation Radiotherapy Department, Complejo Hospitalario Universitario de Vigo, Vigo, Spain

  • Ángel Carracedo,

    Affiliation Genomic Medicine Group, Galician Foundation of Genomic Medicine, Complejo Hospitalario Universitario de Santiago, IDIS, Santiago de Compostela, Spain

  • J. Esteban Castelao

    Affiliation Oncology and Genetics Unit, Genomic Medicine Group, Complejo Hospitalario Universitario de Vigo, Vigo, Spain

Abstract

Background

Differences in the incidence and outcome of breast cancer among Hispanic women compared with white women are well documented and are likely explained by ethnic differences in genetic composition, lifestyle, or environmental exposures.

Methodolgy/Principal Findings

A population-based study was conducted in Galicia, Spain. A total of 510 women diagnosed with operable invasive breast cancer between 1997 and 2010 participated in the study. Data on demographics, breast cancer risk factors, and clinico-pathological characteristics were collected. The different breast cancer tumor subtypes were compared on their clinico-pathological characteristics and risk factor profiles, particularly reproductive variables and breastfeeding. Among the 501 breast cancer patients (with known ER and PR receptors), 85% were ER+/PR+ and 15% were ER-&PR-. Among the 405 breast cancer with known ER, PR and HER2 status, 71% were ER+/PR+/HER2- (luminal A), 14% were ER+/PR+/HER2+ (luminal B), 10% were ER−/PR−/HER2- (triple negative breast cancer, TNBC), and 5% were ER−/PR−/HER2+ (non-luminal). A lifetime breastfeeding period equal to or longer than 7 months was less frequent in case patients with TNBC (OR = 0.25, 95% CI = 0.08–0.68) compared to luminal A breast cancers. Both a low (2 or fewer pregnancies) and a high (3–4 pregnancies) number of pregnancies combined with a long breastfeeding period were associated with reduced odds of TNBC compared with luminal A breast cancer, although the association seemed to be slightly more pronounced among women with a low number of pregnancies (OR = 0.09, 95% CI = 0.005–0.54).

Conclusions/Significance

In case-case analyses with the luminal A cases as the reference group, we observed a lower proportion of TNBC among women who breastfed 7 or more months. The combination of longer breastfeeding duration and lower parity seemed to further reduce the odds of having a TNBC compared to a luminal A breast cancer.

Introduction

In the US, breast cancer impacts each racial group differently [1][4]. Compared with non-Hispanic White (NHW) women, Hispanic women have a lower incidence rate of breast cancer, however, once diagnosed with this disease they are more likely to die from it [5]. Studies [6], [7] have found that despite equal access to health care services, differences persist in the presentation of Hispanic women with breast cancer compared with NHW women, suggesting a biologic basis for the racial/ethnic differences. The potential biological differences among breast cancers may result from racial/ethnic differences in genetic composition, lifestyles or environmental exposures [7].

It has been reported that women diagnosed with estrogen receptor-positive (ER+)/progesterone receptor-positive (PR+) tumors are more responsive to hormonal treatment and have a better prognosis than those diagnosed with estrogen receptor-negative (ER-)/progesterone receptor negative (PR-) tumors, suggesting etiologic heterogeneity of hormone-receptor defined subtypes of breast cancer [8], [9]. Consistently, disparate risk factor profiles for breast cancer according to ER and PR status have been reported [1], [10]. In general, Hispanic patients with breast cancer tend to have ER- tumors more disproportionately than NHW women although the difference was not as great as that seen between black and NHW women [11], [12].

In this study, we describe the characteristics of breast cancer subtypes defined by ER, PR and HER2 receptor status and assessed the associations between reproductive factors and breastfeeding and tumor subtypes in a case series of female breast cancer patients from Galicia, a region located in the northwest part of Spain, whose history has been defined by mass emigration to Latin America [13]. Because this region has been the European state with one of the highest emigration to Latin America in the 1800s and 1900s, its population could be, at least partially, a contributor of the European ancestry to Hispanics in the US. In addition, the Galician population could provide a contrast group to Hispanics from regions in the U.S. such as the San Luis Valley, Colorado in the US, many of whom self-identify as being of “Spanish origin” [14].

To our knowledge, this is in one of the first studies to explore these relationships in this population.

Materials and Methods

Study Population

A population-based study, which is part of the Breast Oncology Galician Network (BREOGAN) Study, was conducted in the city of Vigo, Spain within a geographically defined health region that covers aproximately 437,000 inhabitants. The study involved 510 women with operable invasive breast cancer diagnosed and treated between 1997 and 2010 at the Clinical University Hospital of Vigo (Vigo, Spain). Ethics approval for this study was obtained from the Galician Ethics and Research Committee (CEIC, Comité Ético de Investigación Clínica de Galicia) associated with the Complejo Hospitalario Universitario de Vigo from where all participants were recruited. Written informed consent was obtained for this study, which was conducted according to the Spanish law including adherence to the Helsinki Principles of 1975, as revised in 1983.

Data Collection

Risk factor information was collected through a risk factor questionnaire adapted from the Ella Binational Breast Cancer Study [15]. Clinical and histopathological information was abstracted from computerized medical records by trained physicians. The following variables were recorded: lifetime breastfeeding (categorized as no breastfeeding, < mean lifetime breastfeeding duration (7 months), ≥ mean lifetime breastfeeding duration (7 months)), age at menarche (categorized as ≤13 years, 14 years, ≥15 years), age at first full-term pregnancy (categorized as ≤22 years, 23–27 years, ≥28 years), parity (categorized as never vs. ever pregnant), age at diagnosis (categorized when being the main variable of the analysis as <50 years, ≥50 years, otherwise treated as a continous variable), age at menopause (categorized as <50 years, ≥50 years), menopausal status at diagnosis (categorized as pre, peri and postmenopausal), number of pregnancies (categorized as none, 1–2, ≥3), family history (categorized as none vs. one or more first degree relatives with breast and/or ovarian cancer), ER, PR and HER2 status (categorized as positive and negative), grade (categorized as I – well differenciated -, II – moderately differenciated- and III – poorly differenciated or undiferenciated), histology type (categorized as invasive ductal carcinoma, invasive lobular carcinoma and medular carcinoma), and tumor size (categorized as ≤1 cm, >1 - <2 cm, ≥2 cm). Of the 510 women who participated in the study, 1 had unknown ER status, 8 had unknown PR status, 9 had unknown ER and PR status and 105 had unknown joint ER, PR and HER2 status. Thirty eight women had unknown grade, one had unknown histological type and 21 had unknown tumor size. Two women had unknown age at menarche and two (out of 423 parous women) had unknown lifetime breastfeeding.

Clinico-pathological data.

Histopathological information was abstracted from computerized medical records by trained physicians. Immunohistochemistry (IHC) analyses on paraffin-embedded material have been previously performed following standard procedures in Galician hospitals to determine the status of ER and PR. In every tumor, 4-µm histological sections were cut and stained with hematoxylin and eosin for histopathological examination according to the criteria of the World Health Organization [16]. Histological grading was evaluated using the Nottingham modification of the Bloom-Richardson system [17]. IHC analysis on paraffin-embedded material was performed using a universal second antibody kit that used a peroxidase-conjugated labeleddextran polymer (EnVision®, Peroxidase/DAB; Dako, Glostrup, Denmark), with antibodies for ER (clone 6F11, dilution 1∶50, water bath; Novocastra, Newcastle-upon- Tyne, UK), PR (clone PgR 636, dilution 1∶50, water bath; Dako, Glostrup, Denmark). Negative and positive controls were concurrently run for all antibodies with satisfactory results. Cells were considered immunopositive when diffuse or dot-like nuclear staining was observed regardless of the intensity of the staining; only nuclear immunoreactivity was considered specific. The number of positive cells was counted by two different observers independently. Whenever necessary, a consensus was reached using a double-headed microscope. ER and PR were considered positive when the percent of immunostained nuclei was ≥10%.

Immunohistochemistry (IHC) analyses were performed to determine HER2 status (Dako). No immunostaining (0) or weak membrane immunostaining (1+) was considered low HER2 expression (HER2-). Strong membrane immunostaining (3+) was considered HER2 overexpression (HER2+). Moderate membrane staining (2+) samples were further analyzed using fluorescence in situ hybridization techniques; they were considered to be HER2+ if the ratio of cerb-B2/centromere 17 copy number was >2.0.

Statistical Analyses

We classified breast tumors according to their expression of ER (n = 509), PR (n = 502) and joint status of both ER and PR expression (n = 501). For patients with data available on ER, PR, and HER2 (n = 405 case patients), we defined four tumor subtypes (ER+/HER2- or PR+/HER2- [luminal A], ER+/HER2+ or PR+/HER2+ [luminal B], ER−/PR−/HER2+ [non-luminal], and ER−/PR−/HER2-[triple negative]), shown in Table 1). Case-case analysis was conducted. Multivariate logistic regression was used to estimate odds ratios (ORs), 95% confidence intervals (CIs), and P values for associations between the different risk factors and breast cancer subtypes while simultaneously controlling for age at diagnosis, age at menarche, age at first full-term pregnancy (only in analyses restricted to parous women), menopausal status at diagnosis and family history of first degree relatives with breast and/or ovarian cancer. Outcome (dependent) variables were breast cancer subtypes defined by ER, PR, and HER2 status, and explanatory variable was the risk factor being studied at the time. P values were calculated using likelihood ratio tests.

All statistical analyses were performed using the R statistical software (R_2.13.0). All reported test significance levels (P values) were two-sided.

thumbnail
Table 1. Characteristics of the breast cancer patients included in the study (N = 510).

https://doi.org/10.1371/journal.pone.0040543.t001

Results

A total of 501 breast cancer patients with known ER and PR and 405 with known ER, PR, and HER2 were identified. Among those with known ER and PR status, 85% were ER+/PR+ and 15% were ER-&PR- and among those with data for all three markers, 71% were ER+/PR+/HER2- (luminal A), 14% were ER+/PR+/HER2+ (luminal B), 10% were ER−/PR−/HER2- (TNBC), and 5% were ER−/PR−/HER2+ (non-luminal). The age of these patients ranged from 28 to 84 years, with a mean of 54.7±12.7 years. Detailed characteristics of the study population are presented in Table 1.

Table 2 shows the associations between breast cancer risk factors and TNBC in comparison to luminal A breast cancer. TNBC phenotype was significantly associated with shorter duration of breastfeeding after adjustment for other breast cancer risk factors. A lifetime breastfeeding period equal or longer than 7 months was inversely associated with the odds of having a triple-negative tumor (versus a luminal A tumor) (OR = 0.25, 95% CI = 0.08– 0.68). Parity was more frequent in case patients with TNBC compared to luminal A breast cancers (OR = 1.81, 95% CI = 0.67–6.32). No meaningful associations were found between other reproductive or menstrual factors and TNBC.

thumbnail
Table 2. Associations between breast cancer risk factors/tumor characteristics and triple-negative breast cancer.

https://doi.org/10.1371/journal.pone.0040543.t002

We further examined the joint effect of breastfeeding and parity and tumor subtype. Table 3 shows the association between the average lifetime duration of breastfeeding and odds of having TNBC versus luminal A breast cancer within case groups defined by parity. Among women with 2 or fewer full-term pregnancies, breastfeeding for 7 months or longer was inversely associated with the odds of having a triple-negative tumor versus a luminal A tumor (OR = 0.09, 95% CI = 0.005–0.54) after adjustment for age at diagnosis, age at menarche, age at first full-term pregnancy, menopausal status and family history; however, this finding is based on only 1 TNBC with number of full-term pregnancies 2 o lower who breastfed for 7 months or longer. Among women with 3 or more pregnancies, breastfeeding duration of 7 months or longer was also inversely associated with the odds of having a triple-negative tumor (versus luminal A tumors) although the association lacked precison (OR = 0.37, 95% CI = 0.08–1.65). No statistically significant interaction between breastfeeding and parity and odds of TNBC vs. luminal A subtype was shown (P = 0.41).

thumbnail
Table 3. Associations between parity and lifetime breastfeeding and luminal A and triple-negative breast cancer.

https://doi.org/10.1371/journal.pone.0040543.t003

No associations were shown between other breast cancer risk factors (age at menarche, age at first pregnancy, age at menopause, and family history) and the different tumor subtypes (Table S1).

Regarding tumor characteristics, high grade (OR = 47.38, 95% CI = 6.14–365.30, P<0.001) and medullar type breast cancer (OR = 35.30, 95% CI = 6.84–182.10, P<0.001), were more frequent in case patients with triple-negative tumors compared to luminal A tumors (Table S2).

Discussion

In a population-based study of breast cancer patients from Spain, we observed that the proportion of cases with TNBC versus luminal A breast cancer was lower in women who breastfed for 7 or more months than in those who did not breastfeed. Also, compared with luminal A breast cancers, TNBCs were more common in parous vs. non-parous women. Both a low (2 or fewer pregnancies) and a high (3–4 pregnancies) number of pregnancies combined with a long breastfeeding period were associated with reduced odds of TNBC compared with luminal A breast cancer; although the association seemed to be slightly more pronounced among women with a low number of pregnancies we lacked the statistical power to detect any difference. No other associations were detected between tumor subtypes and other reproductive/lifestyle breast cancer risk factors despite the accumulating evidence favoring distinct reproductive profiles among the differing tumor subtypes [18], [19].

Several case-control and cohort studies have examined the association between parity and breastfeeding and the risk of TNBC. Although an inverse association of parity with risk of ER+ breast cancer has been found by many studies, studies of ER– breast cancer indicate a positive, risk-enhancing association with parity [18][21]. Studies including “intrinsic” breast cancer subtypes based on additional molecular markers such as the basal-like subtype also found high parity to be associated with an increased risk [18], [19]. In some of these studies, this association was present only among women who had never breastfed [19], [21]. Several studies have reported a lower risk of TNBC in parous women who have ever breastfed a child [22], or who breastfed for a cumulative duration of at least 4 [19], 6 [23], [24], or 12 months [25]. Only two case–case analyses have compared the TNBC subtype with the ER+/PR+ subtype and both found a strong positive association of parity with TNBC [26] or with ER−/PR−/HER2+ tumors [27]. Both studies also observed a reduced odds of TNBC (compared to non-TNBC and to luminal A breast cancer) associated with breastfeeding [26], [27], and, in one, the positive association of parity with TNBC was present only among women who had never breastfed [27]. We found some evidence in support of these findings in our study of Spanish women from Galicia. We observed reduced odds of TNBC compared to luminal A breast cancer associated with breastfeeding, and some indication that the combination of a longer breastfeeding duration and decreased parity may reduce the odds of having a TNBC compared to a luminal A breast cancer, although this finding was based on very small numbers (only one TNBC with number of full-term pregnancies 2 o lower who breastfed for 7 months or longer).

It is unclear why the inverse association between breastfeeding and TNBC tends to be more pronounced among Hispanics than NHWs [26]. In general, Hispanic patients with breast cancer tend to have ER-negative tumors more frequently than NHW women [11], [12], [28], [29]. In one study using data from 13 Surveillance, Epidemiology, and End Results cancer registries over a period from 1992 through 2004 [30], researchers analyzed age-adjusted incidence rates and trends from women older than 50 years and showed an increase in incidence rates of ER and PR-negative tumors among Hispanic women compared with NHW women (14.2% in white women compared with 17.3% in Hispanic women). They reported a relative increase in rates of TNBC in Hispanic women (from 2001 through 2004) of 26.8%, with an absolute increase from 34.6 to 43.9 [30]. Our findings that TNBC was more likely to be of a high grade, and high tumor size at diagnosis support the hypothesis that the presence or absence of ER and PR represents distinct biological entities rather than different stages in the natural history of the disease.

It is possible that genetic susceptibility to breast cancer may differ among Hispanics and other ethnicities. Major differences in gene expression between Hispanics and NHW have been detected [5]. In one such study, researchers [5], [31] hypothesized that during pregnancy, progenitor cells that coexpress GABRP (the gamma-aminobutyric acid A receptor, GABRP, a type A receptor that is expressed in reproductive tissues), proliferate within the breast lobules and then are lost during breastfeeding. These investigators evaluated 203 newly diagnosed invasive breast cancers and showed that higher GABRP gene expression was more common in younger women with a limited history of lactation after pregnancy study [31]. The same study showed that GABRP gene expression was higher in Hispanic women compared with white controls. This investigation suggested that GABRP gene expression may be associated with high-grade breast cancer in Hispanic women; although this needs additional study.

The present study was conducted in Galicia, a region located in the northwest part of Spain, whose history has been defined by mass emigration to Latin America [13]. In the United States, Hispanics are a diverse and growing community that represents 12% of the US population [32]. Hispanic ethnicity represents a culturally and genetically heterogeneous group [33]. Hispanics are basically tri-hybrid, i.e., their ancestral populations being European, African, and Native American with the European contribution usually being the highest, although this varies to a degree [34]. Because Galicia has been the European state with one of the highest emigration rates to Latin America, its population could be, at least partially, a contributor of the European ancestry to Hispanics in the US. Thus, in a BRCA1 screening study in familial breast cancer carried out in different centers in Spain, France and the United Kingdom, the missense mutation 330 A>G was independently identified in six families, all of them with Spanish/Galician ancestors [35]. This mutation has been observed in families in diverse geographical locations (Spain, Caribbean, France, United Kingdom) who would appear to be of Spanish origin [35]. These studies are consistent with the possibility that the families studied shared a common ancestry with BRCA1 330 A>G being a founder mutation of Spanish/Galician origin. Similarly, the BRCA1 185delAG mutation has been identified among several apparently non-Jewish families in Spain [36][40], one Chilean family [41], and in several Mexican and Hispanic American cohorts [42][44]. The 185delAG families identified among the Spanish cohorts [36][40] are likely descendants of the Jewish that remained in Spain, whereas the carriers reported in the other cohorts are likely descendants of Spanish Jews who emigrated to the Americas in the late 15th century and assimilated into the larger Hispanic society [44].

Although our results are mostly in agreement with those of other studies, limitations of our study should be discussed, notably its small sample size particularly in subset analyses stratified by hormonal receptor status ER, PR or HER2. Sample sizes for the less common subtypes were limited. In addition, the breastfeeding data were based on self reported information collected years later. In general, breastfeeding history has been shown to be accurate and reliable [45], [46]. However, other authors have shown missclassification [47], although it has been found to be non-differential, i.e., to attenuate the true strength of the association between breastfeeding and the health event under study [47]. Thus, even if there was misclassification in the present study, the true association between breastfeeding and TNBC would have possibly been stronger than the observed. Another limitation relates to the fact that since breast cancer receptor status had been determined from medical records in the present study, there could be the possibility of potential heterogeneity in reading stains and scoring of immunohistochemistry; however this limitation would be expected to bias the study results towards the null. Finally, the case-case study design has obvious limitations [9]. A study that does not include a disease-free population does not provide a traditional risk ratio and may not provide a valid estimate of the association between a risk factor and disease; thus, the case-case OR estimated in this study cannot be interpreted as a measure of risk for the specific subtype. Furthermore, the magnitude of the association is not the magnitude of risk, but rather an indicator of the general direction of the correlation between risk factor and subtype. Thus, results from case-case analyses like ours must be validated in traditional case-control and cohort studies to asses risk and estimate the magnitude of the effect.

Our case-case design study can be particularly useful in assessing the relative correlation of established risk factors and the different tumor subtypes. We have included in our study detailed tumor marker information, such as ER, PR and Her-2 receptor status, which is needed to identify etiologic heterogeneity for established breast cancer risk factors and disease subtypes [9].

In conclusion, this analysis shows that associations between breast cancer and reproductive factors or breastfeeding vary by breast cancer tumor subtypes defined by ER, PR, and HER2 status, particularly luminal A and TNBC. These results are in concordance with emerging evidence that relationships for genetic susceptibility loci also vary by expression levels of markers in tumors [48]. Our results support the view that there may be more than one type of breast cancer from an etiological perspective, and specifically support the hypothesis that hormone receptor negative tumors may have a different etiology than hormone receptor positive tumors. Given the proposed disease heterogeneity observed in breast cancer, future large epidemiological studies will be helpful in identifying etiologic heterogeneity for the established risk factors by disease subtype. Breastfeeding, for example, may be a potential modifiable factor that may be related to the development of a specific breast cancer tumor subtype, and not to all tumor subtypes. Knowledge gained from these studies is likely to produce important information on specific risk factors by tumor subtype which would help in risk prediction models and risk reduction strategies.

Supporting Information

Table S1.

Associations between other reproductive and lyfestyle breast cancer risk factors and different tumor subtypes.

https://doi.org/10.1371/journal.pone.0040543.s001

(DOC)

Table S2.

Associations between tumor characteristics and different tumor subtypes.

https://doi.org/10.1371/journal.pone.0040543.s002

(DOC)

Author Contributions

Conceived and designed the experiments: MGD JEC. Performed the experiments: CMR SMP. Analyzed the data: CMR MGD JEC XJ. Contributed reagents/materials/analysis tools: MEC VMG AAG MPF MAT AC MF FG MEM. Wrote the paper: CMR MGD JEC.

References

  1. 1. Li CI, Malone KE, Daling JR (2003) Differences in breast cancer stage, treatment, and survival by race and ethnicity. Arch Intern Med 163: 49–56.
  2. 2. Shavers VL, Harlan LC, Stevens JL (2003) Racial/ethnic variation in clinical presentation, treatment, and survival among breast cancer patients under age 35. Cancer 97: 134–147.
  3. 3. Chu KC, Anderson WF (2002) Rates for breast cancer characteristics by estrogen and progesterone receptor status in the major racial/ethnic groups. Breast Cancer Res Treat 74: 199–211.
  4. 4. Chu KC, Anderson WF, Fritz A, Ries LA, Brawley OW (2001) Frequency distributions of breast cancer characteristics classified by estrogen receptor and progesterone receptor status for eight racial/ethnic groups. Cancer 92: 37–45.
  5. 5. Patel TA, Colon-Otero G, Bueno Hume C, Copland JA, Perez EA (2010) Breast Cancer in Latinas: Gene Expression, Differential Response to Treatments, and Differential Toxicities in Latinas Compared with Other Population Groups. The Oncologist 15: 466–475.
  6. 6. Wojcik BE, Spinks MK, Stein CR (2003) Effects of Screening Mammography on the Comparative Survival Rates of African American, White, and Hispanic Beneficiaries of a Comprehensive Health Care System. The Breast Journal 9: 175–183.
  7. 7. Watlington AT, Byers T, Mouchawar J, Sauaia A, Ellis J (2007) Does having insurance affect differences in clinical presentation between Hispanic and non-Hispanic white women with breast cancer? Cancer 109: 2093–2099.
  8. 8. Setiawan VW, Monroe KR, Wilkens LR, Kolonel LN, Pike MC, et al. (2009) Breast Cancer Risk Factors Defined by Estrogen and Progesterone Receptor Status. American Journal of Epidemiology 169: 1251–1259.
  9. 9. Martinez ME, Cruz GI, Brewster AM, Bondy ML, Thompson PA (2010) What can we learn about disease etiology from case-case analyses? Lessons from breast cancer. Cancer Epidemiol Biomarkers Prev 19: 2710–2714.
  10. 10. Colditz GA, Rosner BA, Chen WY, Holmes MD, Hankinson SE (2004) Risk factors for breast cancer according to estrogen and progesterone receptor status. J Natl Cancer Inst 96: 218–228.
  11. 11. Vona-Davis L, Rose DP (2009) The influence of socioeconomic disparities on breast cancer tumor biology and prognosis: a review. J Womens Health (Larchmt) 18: 883–893.
  12. 12. Elledge RM, Clark GM, Chamness GC, Osborne CK (1994) Tumor biologic factors and breast cancer prognosis among white, Hispanic, and black women in the United States. J Natl Cancer Inst 86: 705–712.
  13. 13. Galicia Xd (2001) Galicia in the World. The exterior action of the Xunta of Galicia. Xunta de Galicia 2001: 456–465.
  14. 14. Shetterly SM, Baxter J, Morgenstern NE, Grigsby J, Hamman RF (1998) Higher instrumental activities of daily living disability in Hispanics compared with non-Hispanic whites in rural Colorado. The San Luis Valley Health and Aging Study. Am J Epidemiol 147: 1019–1027.
  15. 15. Martínez ME, Gutiérrez-Millan LE, Bondy M, Daneri-Navarro A, Meza-Montenegro MM, et al. (2010) Comparative study of breast cancer in Mexican and Mexican-American women. Health 2: 1040–1048.
  16. 16. Ellis IO, Schnitt SJ, Sastre-Garau X, Bussolati G, Tavassoli FA, et al. (2003) Invasive breast carcinomas.. In: Tavassoli FA, Devilee P, editors. pp. 9–110. Lyon: IARC Press.
  17. 17. Frierson HF Jr, Wolber RA, Berean KW, Franquemont DW, Gaffey MJ, et al. (1995) Interobserver reproducibility of the Nottingham modification of the Bloom and Richardson histologic grading scheme for infiltrating ductal carcinoma. Am J Clin Pathol 103: 195–198.
  18. 18. Yang XR, Sherman ME, Rimm DL, Lissowska J, Brinton LA, et al. (2007) Differences in risk factors for breast cancer molecular subtypes in a population-based study. Cancer Epidemiol Biomarkers Prev 16: 439–443.
  19. 19. Millikan RC, Newman B, Tse CK, Moorman PG, Conway K, et al. (2008) Epidemiology of basal-like breast cancer. Breast Cancer Res Treat 109: 123–139.
  20. 20. Phipps AI, Chlebowski RT, Prentice R, McTiernan A, Wactawski-Wende J, et al. (2011) Reproductive history and oral contraceptive use in relation to risk of triple-negative breast cancer. J Natl Cancer Inst 103: 470–477.
  21. 21. Palmer JR, Boggs DA, Wise LA, Ambrosone CB, Adams-Campbell LL, et al. (2011) Parity and lactation in relation to estrogen receptor negative breast cancer in African American women. Cancer Epidemiol Biomarkers Prev 20: 1883–1891.
  22. 22. Xing P, Li J, Jin F (2010) A case-control study of reproductive factors associated with subtypes of breast cancer in Northeast China. Med Oncol 27: 926–931.
  23. 23. Phipps AI, Malone KE, Porter PL, Daling JR, Li CI (2008) Reproductive and hormonal risk factors for postmenopausal luminal, HER-2-overexpressing, and triple-negative breast cancer. Cancer 113: 1521–1526.
  24. 24. Ma H, Wang Y, Sullivan-Halley J, Weiss L, Marchbanks PA, et al. (2010) Use of four biomarkers to evaluate the risk of breast cancer subtypes in the women's contraceptive and reproductive experiences study. Cancer Res 70: 575–587.
  25. 25. Trivers KF, Lund MJ, Porter PL, Liff JM, Flagg EW, et al. (2009) The epidemiology of triple-negative breast cancer, including race. Cancer Causes Control 20: 1071–1082.
  26. 26. Shinde SS, Forman MR, Kuerer HM, Yan K, Peintinger F, et al. (2010) Higher parity and shorter breastfeeding duration: association with triple-negative phenotype of breast cancer. Cancer 116: 4933–4943.
  27. 27. Kwan ML, Kushi LH, Weltzien E, Maring B, Kutner SE, et al. (2009) Epidemiology of breast cancer subtypes in two prospective cohort studies of breast cancer survivors. Breast Cancer Res 11: R31.
  28. 28. Dunnwald LK, Rossing MA, Li CI (2007) Hormone receptor status, tumor characteristics, and prognosis: a prospective cohort of breast cancer patients. Breast Cancer Res 9: R6.
  29. 29. Hines LM, Risendal B, Slattery ML, Baumgartner KB, Giuliano AR, et al. (2008) Differences in estrogen receptor subtype according to family history of breast cancer among Hispanic, but not non-Hispanic White women. Cancer Epidemiol Biomarkers Prev 17: 2700–2706.
  30. 30. Hausauer AK, Keegan TH, Chang ET, Clarke CA (2007) Recent breast cancer trends among Asian/Pacific Islander, Hispanic, and African-American women in the US: changes by tumor subtype. Breast Cancer Res 9: R90.
  31. 31. Symmans WF, Fiterman DJ, Anderson SK, Ayers M, Rouzier R, et al. (2005) A single-gene biomarker identifies breast cancers associated with immature cell type and short duration of prior breastfeeding. Endocrine-Related Cancer 12: 1059–1069.
  32. 32. Lapham SJ (1993) 1990 Ethnic Profiles for States. Washington, DC: Ethnic and Hispanic Branch, Population Division, Bureau of the Census, CPH L-136.
  33. 33. Chakraborty BM, Fernandez-Esquer ME, Chakraborty R (1999) Is being Hispanic a risk factor for non-insulin dependent diabetes mellitus (NIDDM)? Ethn Dis 9: 278–283.
  34. 34. Sans M (2000) Admixture studies in Latin America: from the 20th to the 21st century. Hum Biol 72: 155–177.
  35. 35. Vega A, Torres M, Martinez JI, Ruiz-Ponte C, Barros F, et al. (2002) Analysis of BRCA1 and BRCA2 in breast and breast/ovarian cancer families shows population substructure in the Iberian peninsula. Annals of Human Genetics 66: 29–36.
  36. 36. Diez O, Domenech M, Alonso MC, Brunet J, Sanz J, et al. (1998) Identification of the 185delAG BRCA1 mutation in a Spanish Gypsy population. Hum Genet 103: 707–708.
  37. 37. Blesa JR, Garcia JA, Ochoa E (2000) Frequency of germ-line BRCA1 mutations among Spanish families from a Mediterranean area. Hum Mutat 15: 381–382.
  38. 38. Osorio A, Barroso A, Martinez B, Cebrian A, San Roman JM, et al. (2000) Molecular analysis of the BRCA1 and BRCA2 genes in 32 breast and/or ovarian cancer Spanish families. British Journal of Cancer 82: 1266–1270.
  39. 39. Diez O, Osorio A, Robledo M, Barroso A, Domenech M, et al. (1999) Prevalence of BRCA1 and BRCA2 Jewish mutations in Spanish breast cancer patients. Br J Cancer 79: 1302–1303.
  40. 40. Osorio A, Robledo M, Albertos J, Diez O, Alonso C, et al. (1998) Molecular analysis of the six most recurrent mutations in the BRCA1 gene in 87 Spanish breast/ovarian cancer families. Cancer Lett 123: 153–158.
  41. 41. Ah Mew N, Hamel N, Galvez M, Al-Saffar M, Foulkes WD (2002) Haplotype analysis of a BRCA1: 185delAG mutation in a Chilean family supports its Ashkenazi origins. Clin Genet 62: 151–156.
  42. 42. Alvarez-Franco M LG, Gerson R, Bale AE (2001) BRCA1 and BRCA2 founder mutations in Spanish and Mexican populations. Am J Hum Genet 69: A. 494 p.
  43. 43. Mullineaux LG, Castellano TM, Shaw J, Axell L, Wood ME, et al. (2003) Identification of germline 185delAG BRCA1 mutations in non-Jewish Americans of Spanish ancestry from the San Luis Valley, Colorado. Cancer 98: 597–602.
  44. 44. Weitzel JN, Lagos V, Blazer KR, Nelson R, Ricker C, et al. (2005) Prevalence of BRCA mutations and founder effect in high-risk Hispanic families. Cancer Epidemiol Biomarkers Prev 14: 1666–1671.
  45. 45. Li R, Scanlon KS, Serdula MK (2005) The validity and reliability of maternal recall of breastfeeding practice. Nutr Rev 63: 103–110.
  46. 46. Launer LJ, Forman MR, Hundt GL, Sarov B, Chang D, et al. (1992) Maternal recall of infant feeding events is accurate. J Epidemiol Community Health 46: 203–206.
  47. 47. Promislow JH, Gladen BC, Sandler DP (2005) Maternal recall of breastfeeding duration by elderly women. Am J Epidemiol 161: 289–296.
  48. 48. Garcia-Closas M, Hall P, Nevanlinna H, Pooley K, Morrison J, et al. (2008) Heterogeneity of breast cancer associations with five susceptibility loci by clinical and pathological characteristics. PLoS Genet 4: e1000054.