Plasma carotenoids, vitamin C, tocopherols, and

AJCN. First published ahead of print January 20, 2016 as doi: 10.3945/ajcn.114.101659.

Plasma carotenoids, vitamin C, tocopherols, and retinol and the risk of breast cancer in the European Prospective Investigation into Cancer and Nutrition cohort1,2 Marije F Bakker,3* Petra HM Peeters,3,5 Veronique M Klaasen,3,6 H Bas Bueno-de-Mesquita,4,5,7,8 Eugene HJM Jansen,7 Martine M Ros,7 Noémie Travier,9 Anja Olsen,10 Anne Tjønneland,10 Kim Overvad,11 Sabina Rinaldi,12 Isabelle Romieu,12 Paul Brennan,12 Marie-Christine Boutron-Ruault,13,14,15 Florence Perquier,13,14,15 Claire Cadeau,13,14,15 Heiner Boeing,16 Krasimira Aleksandrova,16 Rudolf Kaaks,17 Tilman Ku¨hn,17 Antonia Trichopoulou,18,19 Pagona Lagiou,18,20,21 Dimitrios Trichopoulos,19,20,21 Paolo Vineis,5,22 Vittorio Krogh,23 Salvatore Panico,24 Giovanna Masala,25 Rosario Tumino,26 Elisabete Weiderpass,27,28,29,30 Guri Skeie,27 Eiliv Lund,27 J Ramón Quirós,31 Eva Ardanaz,32,33 Carmen Navarro,34,35,36 Pilar Amiano,34,37 María-José Sánchez,34,38 Genevieve Buckland,9 Ulrika Ericson,39 Emily Sonestedt,39 Matthias Johansson,12,40 Malin Sund,41 Ruth C Travis,42 Timothy J Key,42 Kay-Tee Khaw,43 Nick Wareham,44 Elio Riboli,5 and Carla H van Gils3 3

Julius Center for Health Sciences and Primary Care and 4Department of Gastroenterology and Hepatology, University Medical Center Utrecht, Utrecht, Netherlands; 5Department of Epidemiology and Biostatistics, School of Public Health, Faculty of Medicine, Imperial College London, London, United Kingdom; 6Division of Human Nutrition, Wageningen University, Wageningen, Netherlands; 7National Institute for Public Health and the Environment, Bilthoven, Netherlands; 8Department of Social & Preventive Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia; 9Unit of Nutrition, Environment and Cancer, Catalan Institute of Oncology (ICO-IDIBELL), Barcelona, Spain; 10Danish Cancer Society Research Center, Copenhagen, Denmark; 11Section for Epidemiology, Department of Public Health, Aarhus University, Aarhus, Denmark; 12International Agency for Research on Cancer, Lyon, France; 13Inserm, Centre for Research in Epidemiology and Population Health, U1018, Nutrition, Hormones and Women’s Health Team, Villejuif, France; 14University Paris-Sud, UMRS 1018, Villejuif, France; 15IGR, Villejuif, France; 16Department of Epidemiology, German Institute of Human Nutrition Potsdam-Rehbru¨cke, Nuthetal, Germany; 17German Cancer Research Center (DKFZ), Division of Cancer Epidemiology, Heidelberg, Germany; 18WHO Collaborating Center for Food and Nutrition Policies, Department of Hygiene, Epidemiology and Medical Statistics, University of Athens Medical School, Athens, Greece; 19Hellenic Health Foundation, Athens, Greece; 20Department of Epidemiology, Harvard School of Public Health, Boston, MA; 21Bureau of Epidemiologic Research, Academy of Athens, Athens, Greece; 22Human Genetic Foundation (HuGeF), Turin, Italy; 23Epidemiology and Prevention Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy; 24Dipartimento Di Medicina Clinica e Chirurgia, Federico II University, Naples, Italy; 25 Molecular and Nutritional Epidemiology Unit, Cancer Research and Prevention Institute–ISPO, Florence, Italy; 26Cancer Registry and Histopathology Unit, “Civic M.P.Arezzo” Hospital, ASP Ragusa, Italy; 27Department of Community Medicine, Faculty of Health Sciences, University of Tromsø, The Arctic University of Norway, Tromsø, Norway; 28Department of Research, Cancer Registry of Norway, Institute of Population-Based Cancer Research, Oslo, Norway; 29Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden; 30Genetic Epidemiology Group, Folkha¨lsan Research Center, Helsinki, Finland; 31Public Health Directorate, Asturias, Spain; 32Navarre Public Health Institute, Pamplona, Spain; 33CIBER de Epidemiology and Public Health (CIPERESP), Spain; 34Consortium for Biomedical Research in Epidemiology and public Health (CIBER de Epidemiología y Salud Pu´blica), Madrid, Spain; 35Department of Epidemiology, Murcia Regional Health Council, Murcia, Spain; 36Department of Health and Social Sciences, Universidad de Murcia, Murcia, Spain; 37Public Health Division of Gipuzkoa, BioDonostia Research Institute, Health Department of Basque Region, San Sebastian, Spain; 38Andalusian School of Public Health, Granada, Spain; 39Department of Clinical Sciences, Lund University, Malmo¨, Sweden; 40Department of Biobank Research and 41Department of Surgery, Umea˚ University, Umea˚, Sweden; 42Cancer Epidemiology Unit, University of Oxford, Oxford, United Kingdom; 43University of Cambridge, Cambridge, United Kingdom; and 44MRC Epidemiology Unit, University of Cambridge, Cambridge, United Kingdom

ABSTRACT Background: Carotenoids and vitamin C are thought to be associated with reduced cancer risk because of their antioxidative capacity. Objective: This study evaluated the associations of plasma carotenoid, retinol, tocopherol, and vitamin C concentrations and risk of breast cancer. Design: In a nested case-control study within the European Prospective Investigation into Cancer and Nutrition cohort, 1502 female incident breast cancer cases were included, with an oversampling of premenopausal (n = 582) and estrogen receptor–negative (ER2) cases (n = 462). Controls (n = 1502) were individually matched to cases by using incidence density sampling. Prediagnostic samples were analyzed for a-carotene, b-carotene, lycopene, lutein, zeaxanthin, b-cryptoxanthin, retinol, a-tocopherol, g-tocopherol, and

vitamin C. Breast cancer risk was computed according to hormone receptor status and age at diagnosis (proxy for menopausal status) by using conditional logistic regression and was further stratified by smoking status, alcohol consumption, and body mass index (BMI). All statistical tests were 2-sided. Results: In quintile 5 compared with quintile 1, a-carotene (OR: 0.61; 95% CI: 0.39, 0.98) and b-carotene (OR: 0.41; 95% CI: 0.26, 0.65) were inversely associated with risk of ER2 breast tumors. The other analytes were not statistically associated with ER2 breast cancer. For estrogen receptor–positive (ER+) tumors, no statistically significant associations were found. The test for heterogeneity between ER2 and ER+ tumors was statistically significant only for b-carotene (P-heterogeneity = 0.03). A higher risk of breast cancer was found for retinol in relation to ER2/progesterone receptor–negative tumors

Am J Clin Nutr doi: 10.3945/ajcn.114.101659. Printed in USA. Ó 2016 American Society for Nutrition

Copyright (C) 2016 by the American Society for Nutrition

1 of 11

2 of 11

BAKKER ET AL.

(OR: 2.37; 95% CI: 1.20, 4.67; P-heterogeneity with ER+/progesterone receptor positive = 0.06). We observed no statistically significant interaction between smoking, alcohol, or BMI and all investigated plasma analytes (based on tertile distribution). Conclusion: Our results indicate that higher concentrations of plasma b-carotene and a-carotene are associated with lower breast cancer risk of ER2 tumors. Am J Clin Nutr doi: 10.3945/ajcn.114.101659. Keywords: breast cancer, EPIC, antioxidants, carotenoids, plasma

INTRODUCTION

Vegetables and fruit contain many putatively cancer-protective substances. For breast cancer, a recent meta-analysis of 15 prospective studies showed that high intake of fruit and of fruit and vegetables combined is associated with a weak breast cancer risk reduction. For vegetables only, no reduction in risk was found (1). Because of measurement errors inherent to dietary questionnaires that were used in most of these studies (2), protective effects may have been underestimated or small effects could have been missed. Even small effects can have a large impact on public health, leading to preventive dietary recommendations on a population scale. Blood concentrations of carotenoids and vitamin C are good biomarkers of vegetable and fruit consumption and provide better estimates of the concentration actually available to cells than dietary questionnaires (3–6). Besides, carotenoids and vitamin C are thought to have cancer-protective capacities themselves (7–10). Recently, blood concentrations of 6 carotenoids were studied in relation to breast cancer risk in a meta-analysis of 15 prospective studies (11). Higher concentrations of total carotenoids, a-carotene, b-carotene, and lutein were found to be statistically significantly associated with lower breast cancer risk. Another pooled analysis of 8 prospective studies, examining circulating carotenoids, found

1 Supported by Wereld Kanker Onderzoek Fonds (WCRF NL grant number WCRF 2006/13); Europe Against Cancer Program of the European Commission; Deutsche Krebshilfe, Deutsches Krebsforschungszentrum; German Federal Ministry of Education and Research; Danish Cancer Society; Health Research Fund of the Spanish Ministry of Health (ISCIII RETICC RD06/0020), Spanish Regional Governments of Andalucia, Asturia, Basque Country, Murcia (No. 6236), and Navarra; Catalan Institute of Oncology, Red de Centros RCESP, C03/09, Spain; Cancer Research UK; Medical Research Council, United Kingdom; Stroke Association, United Kingdom; British Heart Foundation; Department of Health, United Kingdom; Food Standards Agency, United Kingdom; Wellcome Trust, United Kingdom; Helenic Health Foundation; Italian Association for Research on Cancer; Italian National Research Council, Fondazione-Istituto Banco, Napoli, Italy; Dutch Ministry of Public Health, Welfare and Sports; Dutch Prevention Funds; LK Research Funds; Dutch ZON (Zorg Onderzoek Nederland); World Cancer Research Fund; Swedish Cancer Society; Swedish Scientific Council; Regional Government of Ska˚ne, Sweden; European Research Council; French League against Cancer; National Institute for Health and Medical Research, France; Mutuelle Générale de l’Education Nationale, France; 3M Co, France; Gustave Roussy Institute, France; and General Councils of France. The funder and sponsors did not have any input into study design, study conduct, data collection, analysis, or interpretation, nor did they influence the preparation, review, or approval of the manuscript. 2 Supplemental Tables 1–3 are available from the “Online Supporting Material” link in the online posting of the article and from the same link in the online table of contents at http://ajcn.nutrition.org. *To whom correspondence should be addressed. E-mail: m.f.bakker-8@ umcutrecht.nl. Received October 22, 2014. Accepted for publication November 30, 2015. doi: 10.3945/ajcn.114.101659.

similar results and, in addition, statistically significant inverse associations for lycopene (12). Blood concentrations of vitamin C have been studied in only one nested case-control study of limited size (13), with no suggestion for a protective association. Several prospective studies suggested that fruit and vegetable consumption and dietary carotenoid intake are more strongly related to estrogen receptor–negative (ER2)45 than to estrogen receptor–positive (ER+) breast cancer (14–18) and more strongly to premenopausal than to postmenopausal breast cancer (1, 19). The pooled analysis mentioned above (12) showed that inverse associations between b-carotene concentrations and breast cancer risk were statistically significantly stronger for ER2 than for ER+ breast cancer. The aim of our study was to evaluate prediagnostic plasma concentrations of carotenoids and vitamin C in relation to subsequent risk of incident breast cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort. We also included tocopherol and retinol as was done in comparable studies (6, 20–25). The unique features of this investigation are its large sample size, allowing oversampling of premenopausal and ER2 breast cancer cases, and its long duration of follow-up, together with the fact that it includes participants from northern to southern Europe, spanning a wide range of vegetables and fruit consumption and related plasma biomarkers. METHODS

EPIC cohort EPIC is an ongoing multicenter, prospective cohort study primarily designed to investigate the relation between nutrition and cancer. The total EPIC cohort consists of 521,468 participants recruited from 23 centers in 10 European countries: Denmark, France, Germany, Greece, Italy, the Netherlands, Norway, Spain, Sweden, and the United Kingdom. Enrollment took place between 1992 and 1998. Further details have been described previously (26). Study participants Eligible cases were all first primary incident, histologically confirmed (invasive) breast cancer cases identified by follow-up based on population cancer registries in most countries. For France, Germany, and Greece, a combination of methods was used, including health insurance records, cancer and pathology registries, and active follow-up through study participants and their next of kin. Details on case and control selection have been described previously (27). In short, of all breast cancer cases (n = 5458) diagnosed before 2005, 1502 cases (invasive breast cancer with blood samples available) and 1502 controls were selected. Our study was designed to include an oversampling of premenopausal and ER2 breast cancer cases. Age at diagnosis was used as a proxy for menopausal status at the time of breast cancer diagnosis (#50 compared with .50 y of age at diagnosis for 45 Abbreviations used: EPIC, European Prospective Investigation into Cancer and Nutrition; ER2, estrogen receptor negative; ER+, estrogen receptor positive; PR2, progesterone receptor negative; PR+, progesterone receptor positive.

PLASMA ANTIOXIDANTS AND RISK OF BREAST CANCER

pre- and postmenopausal, respectively). All premenopausal cases (#50 y, n = 582) that were identified at the time of the study (irrespective of their ER2 status) were selected. From the postmenopausal cases (.50 y), all ER2 cases (n = 462) and an approximately equal sample from ER+ cases (n = 458) were selected. The ER+ cases were characterized by the same distribution of country and year of diagnosis as the ER2 cases, whereas apart from these criteria, their selection was random. Controls were selected by using an incidence density sampling design and further matched by study center, age (within 1 y), menopausal status at recruitment, use of exogenous hormones, phase of menstrual cycle, fasting status at blood collection, and time of blood collection (61 h) (27). All participants gave written informed consent. The study was approved by the local ethics committees in the participating centers and by the International Agency for Research on Cancer ethical review committee. Laboratory assays Details on laboratory analysis have been described earlier (28). In brief, each batch contained 80 plasma samples, 2 3 39 for matched case-control pairs, which were analyzed in the same batch and in random order to minimize errors from batch-tobatch variations, and 2 quality control (laboratory) samples. Laboratory technicians were blinded to the case-control status of all samples. For this study, plasma samples were used that had not been thawed previously. Intrabatch and interbatch CVs were 2.8% and 8.7% for vitamin C, 9.1% and 13.2% for a-carotene, 5.6% and 10.9% for b-carotene, 3.7% and 9.1% for b-cryptoxanthin, 6.7% and 13.9% for lutein, 9.9% and 11.3% for lycopene, 11.3% and 21.9% for zeaxanthin, 2.7% and 6.9% for retinol, 2.5% and 8.2% for a-tocopherol, and 2.5% and 6.8% for g-tocopherol, respectively. Carotenoids, retinol, and tocopherols Plasma samples (200 mL) were analyzed for a-carotene, b-carotene, b-cryptoxanthin, lutein, lycopene, zeaxanthin, retinol, a-tocopherol, and g-tocopherol. The National Institute for Public Health and the Environment (Bilthoven, Netherlands) conducted the analysis. This analysis was done by using HPLC (using an HPLC column, 250 3 4.6 mm, ChromSpher 5 mm C18; Varian Assoc.), following a method based on that of Steghens et al. (29). Vitamin C Plasma vitamin C was measured with a colorimetric assay on an LX20-Pro autoanalyzer (Beckman-Coulter). Statistical analysis Descriptives of all analytes were compared between cases and controls. Differences in means 6 SDs were tested by using paired t tests, and if variables were not normally distributed, they were log-transformed and geometric means were calculated. Spearman’s rank correlation coefficient was used to assess the correlation between all individual analytes. ORs and 95% CIs for ER+ and ER2 breast cancer in relation to plasma concentrations of all analytes were calculated with

3 of 11

conditional logistic regression models, stratified by case-control set. Analyses were based on the quintile distribution among the controls (lowest quintile as reference category). Tests for linear trend across quintiles were performed by using median values of each quintile. The final models were, in addition to conditioning on matching factors, adjusted for BMI (in kg/m2), height, age at menarche, age at first full-term pregnancy, oral contraceptive pill use, hormone therapy use, smoking status, alcohol intake, total energy, saturated fatty acids, educational level, and season of blood collection (see Table 2 for units and categories), using indicator variables for missing data. Age at menopause (for postmenopausal women), physical activity [based on Cambridge Physical Activity Index (30)], and geographic region (northern/middle/southern Europe) were evaluated but not included in the final model because they did not change the relation with breast cancer risk and the first 2 variables were missing for several centers. Analyses were also stratified by the combination of both receptors [ER+/progesterone receptor positive (PR+) compared with ER2/progesterone receptor negative (PR2)]. When further stratifying by age at diagnosis as a proxy for menopausal status, only strata with cases .50 y at diagnosis were large enough to draw conclusions on. We assessed heterogeneity between the subtypes defined by receptor status by using a log-likelihood ratio test to compare conditional logistic regression models with and without interaction terms for subtype outcome (ER2 compared with ER+ and ER2/PR2 compared with ER+/PR+). Interaction terms were created by multiplying each subtype with the linear trend over the quintile score of analyte concentrations (31). Partial Pearson correlation coefficients (adjusted for age, BMI, and season of blood collection) were used to assess the correlation between all analytes and fruit and vegetable intake. To diminish the influence of possible changes in food patterns or metabolic changes due to preclinical disease, we performed sensitivity analyses excluding cases whose blood samples were collected within 2 y before cancer diagnosis. Finally, we examined the modifying effect of smoking status (never/past compared with current), alcohol consumption (continuous, g/d), and BMI (continuous) on the relations between plasma analyte concentrations on breast cancer risk, because a protective effect of high plasma antioxidant concentrations could be hypothesized to be stronger in women with high amounts of oxidative stress, caused by smoking and alcohol consumption (19, 32–35), and lower BMI might have stronger protective effects (12). The analyses were performed with conditional logistic regression analyses by using interaction terms with all individual analytes (tertile distribution). Two-sided P values ,0.05 were considered statistically significant. All analyses were performed by using SPSS 20 (SPSS Inc.) and SAS 9.2 (SAS Institute). RESULTS

Baseline characteristics of cases and controls and their plasma concentrations of carotenoids, retinol, tocopherols, and vitamin C are shown in Tables 1 and 2, respectively. In breast cancer cases, time between blood donation and diagnosis was 4.1 y, on average. Most characteristics were equally distributed, which is partly due to the matching procedure. Because of failure of the

4 of 11

BAKKER ET AL.

laboratory method or not enough sample available, there were 5 cases and 1 control with missing values on vitamin C, 28 cases and 22 controls on a-tocopherol and g-tocopherol, and 30 cases and 22 controls on all other analytes with missing values. Spearman correlation coefficients between individual analytes are shown in Supplemental Table 1. The highest correlation was observed between a-carotene and b-carotene (r = 0.69). Spearman correlations for plasma vitamin C, b-carotene, retinol, and the sum of a- and g-tocopherol with their dietary counterparts (vitamin C, b-carotene, and vitamins A and E) were low (0.13, 0.22, 0.14, and 20.18, respectively; data not shown). The breast cancer risk associated with plasma concentrations of all analytes is shown in Table 3 for ER2 and for ER+ breast cancer. Risk of ER2 breast cancer was 39–59% lower for women in the highest quintile (quintile 5) compared with lowest

quintile (quintile 1) in the adjusted analysis for plasma concentrations of a-carotene (OR: 0.61; 95% CI: 0.39, 0.98; P-trend = 0.02) and b-carotene (OR: 0.41; 95% CI: 0.26, 0.65; P-trend = 0.002). Results for retinol (OR: 1.65; 95% CI: 0.97, 2.81) showed a higher risk of ER2 breast cancer that was borderline statistically significant (P-trend = 0.08). The other analytes were not statistically associated with ER2 breast cancer. For ER+ tumors, no statistically significant associations were found, but the test for heterogeneity between ER2 and ER+ tumors was only statistically significant for b-carotene (P-heterogeneity = 0.03). Separate analyses by ER/PR status (Figure 1) indicated a protective effect of vitamin C in relation to ER+/PR+ breast cancer (OR: 0.64; 95% CI: 0.35, 1.17; P-trend = 0.04). In relation to ER2/PR2, a similar effect was observed, but here the P-trend was not statistically significant (P-trend = 0.16,

TABLE 1 Baseline characteristics of breast cancer cases and controls at recruitment1

Age, y BMI, kg/m2 Height, cm Age at menarche, y Age at first-time pregnancy, y Alcohol consumption, g/d Saturated fatty acids, g/d Energy, kcal/d Storage time blood samples, y Time from blood draw to diagnosis, y Menopausal status at recruitment, % Postmenopausal (natural or surgical) Premenopausal Perimenopausal Use of exogenous hormones, % Oral contraceptive use Never Past Current HT use Never Past Current Smoking status, % Never Past Current Parous women, % Physically active (based on CPAI), % Secondary school or university degree, % Specific case characteristics, n #50 y at diagnosis ER+/ER2 PR+/PR2 ER+PR+/ER2PR2 .50 y at diagnosis ER+/ER2 PR+/PR2 ER+PR+/ER2PR2 1

Cases (n = 1502)

Controls (n = 1502)

49.98 6 8.58 24.86 6 4.14 162.34 6 6.49 13.03 6 1.53 25.23 6 4.37 4.21 (0.84, 13.20)3 29.99 6 11.50 1974 6 548 12.68 6 1.70 4.07 6 2.70

50.00 6 8.59 24.80 6 4.06 162.09 6 6.46 13.05 6 1.55 24.69 6 4.32 4.12 (0.66, 12.23) 29.31 6 11.47 1934 6 543 12.70 6 1.68

41.9 42.7 15.4

42.1 42.4 15.5

38.5 56.9 4.6

38.1 57.3 4.6

71.8 7.4 20.8

71.3 7.1 21.6

53.8 24.4 21.8 85.7 42.1 49.0

53.9 23.2 22.9 86.9 41.4 47.1

2

582 183/75 172/72 145/47 920 458/462 282/358 233/307

CPAI, Cambridge Physical Activity Index [incorporating occupational and nonoccupational physical activity; see Wareham et al. (30)]; ER2, estrogen receptor negative; ER+, estrogen receptor positive; HT, hormone replacement therapy; PR2, progesterone receptor negative; PR+, progesterone receptor positive. 2 Mean 6 SD (all such values). 3 Median; 25th, 75th percentiles in parentheses (all such values).

PLASMA ANTIOXIDANTS AND RISK OF BREAST CANCER TABLE 2 Concentrations of vitamin C, carotenoids, retinol, and tocopherols of breast cancer cases and controls at recruitment1 Analytes

Cases (n = 1502)

Controls (n = 1502)

Vitamin C, mmol/L a-Carotene, nmol/L b-Carotene, nmol/L b-Cryptoxanthin, nmol/L Zeaxanthin, nmol/L Lutein, nmol/L Lycopene, nmol/L Sum of carotenoids, nmol/L Retinol, mmol/L a-Tocopherol, mmol/L g-Tocopherol, mmol/L

41.3 99.5 565 312 270 52.0 326 1851 1.78 23.2 4.15

42.7 103 606 320 274 50.8 343 1937 1.76 23.3 4.05

(40.4, 42.2) (95.6, 104) (545, 586) (298, 326) (261, 279) (49.9, 54.2) (311, 342) (1679, 1906) (1.75, 1.81) (22.7, 23.4) (3.98, 4.32)

(41.8, 43.5) (99.1, 108) (585, 628) (307, 333) (265, 282) (48.7, 52.9) (328, 358) (1886, 1989) (1.73, 1.78) (22.9, 23.6) (3.88, 4.22)

1

Values are means with 95% CIs in parentheses and are based on geometric mean values. Because of failure of the laboratory method or not enough sample available, there were 5 cases and 1 control with missing values for vitamin C, 28 cases and 22 controls with missing values for a-tocopherol and g-tocopherol, and 30 cases and 22 controls with missing values for all other analytes.

P-heterogeneity = 0.64). For ER2/PR2, there was an inverse association for b-carotene (OR: 0.45; 95% CI: 0.26, 0.80; P-trend = 0.02) and a borderline statistically significant association for a-carotene (OR: 0.64; 95% CI: 0.36, 1.13; P-trend = 0.09). No statistically significant associations were observed for b- and a-carotene in relation to ER+/PR+ breast cancer. The tests for heterogeneity were not statistically significant (P-heterogeneity = 0.20 and P-heterogeneity = 0.28). Positive associations in relation to ER2/PR2 but not ER+/PR+ breast cancer were seen for zeaxanthin (OR: 2.34; 95% CI: 1.04, 5.23; P-trend = 0.06) and retinol (OR: 2.37; 95% CI: 1.20, 4.67; P-trend = 0.02). P values for heterogeneity were 0.03 and 0.06, respectively. Sensitivity analyses of receptor-specific tumor subtypes that were restricted to women diagnosed .50 y of age gave comparable results to those of Table 3 and Figure 1 (results not shown). Partial Pearson correlation coefficients between individual analyte concentrations and fruit and vegetable intakes are shown in Supplemental Table 2. a-Carotene is most strongly related with root vegetables (r = 0.25), b-carotene with total and root vegetables (both r = 0.14), vitamin C with total and root vegetables (both r = 0.07), and zeaxanthin with leafy vegetables (r = 0.24). Retinol is inversely correlated with leafy vegetables (r = 20.20). Sensitivity analyses were conducted excluding 341 cases whose blood samples were collected within 2 y before cancer diagnosis. Results were quite similar to the analysis including these cases. For ER2 breast cancer, in quintile 5 compared with quintile 1, this was a-carotene (OR: 0.73; 95% CI: 0.44, 1.22; P-trend = 0.14), b-carotene (OR: 0.48; 95% CI: 0.28, 0.81; P-trend = 0.03), and retinol (OR: 1.69; 95% CI: 0.93, 3.10; P-trend = 0.13). Interaction terms for smoking status, alcohol consumption, or BMI (Supplemental Table 3) were for all analytes (based on a tertile distribution) not statistically significant.

DISCUSSION

To our knowledge, this is the largest nested case-control study to date of plasma carotenoids as well as retinol, tocopherols,

5 of 11

and vitamin C in relation to hormone receptor–specific breast cancer risk. For ER2 breast cancers, statistically significant associations were found for plasma a-carotene and b-carotene, leading to a 39–59% reduction in breast cancer risk. Our results are largely in agreement with those of the pooled analysis of 8 prospective studies (12). Their risk estimates for ER2 breast cancer (highest compared with lowest quintile) were OR = 0.61 for a-carotene and OR = 0.52 for b-carotene; ours were OR = 0.61 and OR = 0.41, respectively. Another pooled analysis of 18 prospective studies on dietary carotenoid intake also showed inverse associations of a-carotene, b-carotene, and lutein/zeaxanthin intake being primarily present in relation to ER2 but not to ER+ breast cancer (18). This was also the case for vegetable consumption in a recently published large pooled analysis (36). Two recent nested case-control studies on carotenoids and breast cancer subtypes did not observe heterogeneity between ER2 and ER+ tumors, but the number of ER2 breast cancers was much lower (n = 56 and n = 292, respectively) than in our study (37, 38). The inverse associations found for ER2 breast cancer support the idea that the effect of antioxidants could be easier to detect in less hormone-dependent breast cancers, whereas for more hormone-dependent breast cancers, this effect might be “overshadowed” by the strong influence of hormonal factors (39). Unexpectedly, we observed higher risk ER2/PR2 breast cancer for women in the upper quintiles of zeaxanthin and retinol concentrations. These positive associations were not present in relation to ER+/PR+ breast cancer. Other nested case-control studies did not observe this relation (6, 20–25), although some describe a positive nonsignificant association with retinol (21, 22, 40). However, none was able to distinguish between receptorspecific breast cancer subtypes. The increased ER2/PR2 breast cancer risk in relation to higher concentrations of retinol could be due to the fact that retinol, in contrast to carotenoids and vitamin C, is merely derived from animal sources, although we have no reason to think this would pertain to ER2/PR2 breast cancer only. We do not have an explanation for the increased risk of ER2/PR2 breast cancer for women with high zeaxanthin concentrations. This has not been found in other studies or in relation to other cancers and could also be a chance finding. We did not observe modification of the carotenoid–breast cancer associations by smoking, alcohol consumption, or BMI, as some earlier studies did (6, 12, 22, 25). For example, some other studies showed stronger effects of high carotenoid concentrations in current smokers (12). One explanation might be that the antioxidative effect of vitamin C is not visible in smokers because there is an overload of oxidative stress caused by smoking (also depleting vitamin C concentrations). Our study has some limitations and strengths that deserve further discussion to help interpret our findings. We only have a single measurement of biomarkers available, and changes in long-term exposure and also day-to-day variation may have diluted the results. Daily variation was taken into account in part by matching on time of blood collection, fasting status, and adjusting for season. Moreover, repeatability studies on these biomarkers in blood samples collected up to 11 (vitamin C) to 15 y (retinol, a- and b-carotene) suggest that single measurements are reasonable estimators and representative to use as a predictor and suitable for assessment as risk factors (40–42). The (higher as

Vitamin C, mmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2 a-Carotene, nmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2 b-Carotene, nmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2 Lycopene, nmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2 Lutein, nmol/L ER2 Median2 Model 1 Model 2

Plasma analyte/ER status

34.21 1.00 (0.67, 1.48) 0.99 (0.65, 1.50) 34.70 0.80 (0.56, 1.13) 0.81 (0.56, 1.18)

72.31 1.11 (0.78, 1.58) 1.08 (0.74, 1.57) 71.44 0.95 (0.65, 1.40) 1.00 (0.67, 1.50)

419.88 0.47 (0.32, 0.70) 0.42 (0.28, 0.64) 413.04 0.97 (0.66, 1.43) 0.95 (0.63, 1.42)

260.73 0.74 (0.53, 1.03) 0.71 (0.50, 1.02) 259.04 0.92 (0.62, 1.37) 0.81 (0.53, 1.25)

202.25 1.04 (0.75, 1.45) 1.03 (0.73, 1.45)

19.95 1.00 (reference) 1.00 (reference)

38.33 1.00 (reference) 1.00 (reference) 36.10 1.00 (reference) 1.00 (reference)




639/638 639/638

515/514 515/514

636/632 636/632

515/514 515/514

636/632 636/632

515/514 515/514

636/632 636/632

515/514 515/514

Q2


Q1

535/535 535/535

Cases/controls, n

270.62 0.86 (0.59, 1.25) 0.82 (0.55, 1.22)

367.25 1.05 (0.70, 1.56) 0.94 (0.61, 1.44)

374.96 0.84 (0.57, 1.24) 0.83 (0.54, 1.27)

589.96 1.10 (0.76, 1.59) 1.06 (0.71, 1.57)

603.82 0.71 (0.48, 1.05) 0.66 (0.44, 1.00)

103.45 0.83 (0.57, 1.21) 0.79 (0.52, 1.19)

106.86 0.74 (0.51, 1.08) 0.73 (0.49, 1.09)

43.45 0.79 (0.55, 1.13) 0.80 (0.54, 1.18)

43.12 0.62 (0.40, 0.94) 0.63 (0.40, 0.98)

Q3

TABLE 3 Breast cancer ORs (95% CIs) according to quintiles of plasma analytes for ER2 and ER+ breast cancers1

372.56 1.27 (0.81, 2.01) 1.23 (0.76, 2.01)

510.26 1.43 (0.95, 2.15) 1.37 (0.88, 2.11)

510.26 0.60 (0.38, 0.93) 0.59 (0.36, 0.95)

847.86 1.03 (0.70, 1.50) 1.00 (0.66, 1.51)

879.73 0.56 (0.37, 0.83) 0.51 (0.33, 0.79)

146.86 1.16 (0.80, 1.69) 1.22 (0.81, 1.83)

153.49 0.82 (0.55, 1.22) 0.81 (0.53, 1.24)

51.55 0.78 (0.54, 1.13) 0.74 (0.50, 1.10)

51.37 0.89 (0.60, 1.32) 0.88 (0.58, 1.34)

Q4

616.13 1.35 (0.78, 2.33) 1.19 (0.66, 2.13)

748.86 0.93 (0.58, 1.47) 0.90 (0.55, 1.48)

742.08 1.04 (0.57, 1.89) 1.07 (0.56, 2.03)

1373.03 0.94 (0.64, 1.39) 1.02 (0.66, 1.57)

1426.18 0.46 (0.31, 0.70) 0.41 (0.26, 0.65)

266.43 0.76 (0.51, 1.14) 0.77 (0.49, 1.19)

279.56 0.67 (0.44, 1.01) 0.61 (0.39, 0.98)

64.32 0.77 (0.53, 1.14) 0.76 (0.50, 1.16)

63.58 0.71 (0.47, 1.08) 0.71 (0.45, 1.11)

Q5

0.23 0.48

0.72 0.61

0.30 0.38

0.70 0.91

0.002 0.002

0.23 0.28

0.03 0.02

0.19 0.17

0.11 0.12

P-trend

(Continued)

0.26

0.03

0.26

0.86

P-heterogeneity (ER2 compared with ER+)

6 of 11 BAKKER ET AL.

ER+ Median2 Model 1 Model 2 Zeaxanthin, nmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2 b-Cryptoxanthin, nmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2 Sum of carotenoids, nmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2 Retinol, mmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2 a-Tocopherol, mmol/L ER2 Median2 Model 1 Model 2


TABLE 3 (Continued )

203.44 0.66 (0.43, 1.01) 0.71 (0.45, 1.12)

32.62 1.27 (0.92, 1.74) 1.26 (0.90, 1.77) 33.16 1.07 (0.69, 1.65) 1.06 (0.67, 1.70)

231.17 1.00 (0.71, 1.42) 0.94 (0.65, 1.37) 237.65 0.62 (0.43, 0.88) 0.58 (0.40, 0.85)

1516.71 0.55 (0.39, 0.79) 0.51 (0.35, 0.75) 1525.42 0.90 (0.60, 1.34) 0.85 (0.55, 1.31)

1.50 1.12 (0.70, 1.80) 1.05 (0.64, 1.73) 1.48 1.09 (0.75, 1.59) 1.10 (0.74, 1.63)

19.93 1.33 (0.70, 2.52) 1.26 (0.64, 2.50)






515/514 515/514

636/632 636/632

515/514 515/514

636/632 636/632

515/514 515/514

636/632 636/632

515/514 515/514

636/632 636/632

517/516 517/516

Q2


Q1

636/632 636/632

Cases/controls, n

23.15 0.92 (0.49, 1.74) 0.78 (0.40, 1.54)

1.73 0.89 (0.61, 1.29) 0.92 (0.62, 1.37)

1.75 1.59 (0.98, 2.58) 1.67 (1.01, 2.77)

1978.02 0.97 (0.65, 1.46) 0.92 (0.59, 1.45)

1986.50 0.70 (0.47, 1.04) 0.70 (0.46, 1.06)

374.72 0.79 (0.54, 1.14) 0.74 (0.49, 1.11)

356.35 0.70 (0.48, 1.05) 0.68 (0.45, 1.03)

56.68 1.00 (0.63, 1.59) 0.95 (0.58, 1.57)

54.70 1.29 (0.89, 1.88) 1.18 (0.79, 1.75)

276.81 0.69 (0.45, 1.07) 0.67 (0.43, 1.07)

Q3

26.58 1.00 (0.54, 1.82) 0.85 (0.44, 1.63)

2.04 1.18 (0.79, 1.75) 1.15 (0.76, 1.75)

2.05 1.28 (0.79, 2.07) 1.35 (0.81, 2.25)

2545.21 1.08 (0.72, 1.63) 1.00 (0.64, 1.58)

2560.23 0.70 (0.46, 1.06) 0.65 (0.41, 1.03)

564.78 0.82 (0.58, 1.17) 0.73 (0.49, 1.09)

551.03 0.78 (0.53, 1.14) 0.70 (0.46, 1.07)

89.53 0.92 (0.56, 1.51) 0.83 (0.49, 1.41)

87.61 1.34 (0.81, 2.21) 1.21 (0.70, 2.07)

381.31 0.61 (0.39, 0.95) 0.60 (0.37, 0.97)

Q4

33.66 1.07 (0.59, 1.96) 0.88 (0.46, 1.68)

2.52 1.01 (0.65, 1.56) 1.02 (0.64, 1.63)

2.55 1.55 (0.94, 2.55) 1.65 (0.97, 2.81)

3707.60 0.86 (0.56, 1.30) 0.85 (0.53, 1.37)

3716.75 0.71 (0.44, 1.15) 0.64 (0.37, 1.09)

1005.58 0.77 (0.52, 1.13) 0.70 (0.45, 1.10)

983.98 0.91 (0.59, 1.42) 0.84 (0.51, 1.37)

146.43 0.99 (0.59, 1.64) 0.84 (0.49, 1.45)

143.01 1.34 (0.75, 2.38) 1.29 (0.69, 2.42)

633.22 0.59 (0.36, 0.97) 0.59 (0.35, 1.00)

Q5

0.99 0.51

0.89 0.92

0.13 0.08

0.55 0.66

0.27 0.23

0.79 0.68

0.44 0.29

0.88 0.42

0.23 0.45

0.14 0.15

P-trend

(Continued)

0.32

0.61

0.66

0.33

0.11



7 of 11

19.73 1.06 (0.75, 1.49) 1.04 (0.73, 1.49)

3.40 0.97 (0.68, 1.39) 1.00 (0.69, 1.45) 2.96 1.17 (0.72, 1.91) 1.18 (0.70, 1.97)


517/516 517/516

636/632 636/632

Q2


Q1

636/632 636/632

Cases/controls, n

4.15 0.99 (0.62, 1.59) 1.01 (0.61, 1.68)

4.30 1.06 (0.71, 1.57) 1.14 (0.75, 1.74)

23.12 1.06 (0.73, 1.56) 1.04 (0.70, 1.56)

Q3

6.14 1.09 (0.67, 1.78) 1.08 (0.64, 1.82)

6.42 1.15 (0.75, 1.75) 1.16 (0.74, 1.82)

26.41 0.73 (0.49, 1.10) 0.77 (0.50, 1.18)

Q4

10.19 0.93 (0.56, 1.54) 0.91 (0.53, 1.58)

10.27 1.41 (0.83, 2.38) 1.54 (0.87, 2.71)

33.45 0.86 (0.56, 1.33) 0.88 (0.56, 1.40)

Q5

0.42 0.38

0.17 0.13

0.27 0.40

P-trend

0.12

0.62


1

Ranges of all quintiles, based on controls, for all analytes are as follows: vitamin C: Q1 (2.50–28.5), Q2 (28.5–39.4), Q3 (39.4–46.4), Q4 (46.4–56.1), Q5 (56.1–145.30); a-carotene: Q1 (14.00–56.95), Q2 (56.95–88.31), Q3 (88.31–124.03), Q4 (124.03–198.07), Q5 (198.07–1520.25); b-carotene: Q1 (24.87–348.94), Q2 (348.94–497.32), Q3 (497.32–718.63), Q4 (718.63–1066.96), Q5 (1066.96–7698.56); lycopene: Q1 (14.00–192.07), Q2 (192.07–314.62), Q3 (314.62–431.27), Q4 (431.27–596.49), Q5 (465.62–1529.68); zeaxanthin: Q1 (14.00–21.00), Q2 (21.00–43.20), Q3 (43.20–69.13), Q4 (69.13–107.19), Q5 (107.19–440.69); b-cryptoxanthin: Q1 (17.00–176.09), Q2 (176.09–286.20), Q3 (286.20–444.89), Q4 (444.89–709.18), Q5 (709.18–5476.30); sum of carotenoids: Q1 (277.42–1268.75), Q2 (1268.75– 1730.40), Q3 (1730.40–2331.07), Q4 (2231.07–2982.65), Q5 (2982.65–9134.24); retinol: Q1 (0.32–1.37), Q2 (1.37–1.63), Q3 (1.63–1.90), Q4 (1.90–2.25), Q5 (2.25–6.70); a-tocopherol: Q1 (8.93–18.29), Q2 (18.29–21.54), Q3 (21.54–24.84), Q4 (24.84–29.34), Q5 (29.34–84.65); g-tocopherol: Q1 (0.07–2.33), Q2 (2.33–3.59), Q3 (3.59–5.15), Q4 (5.15–7.87), Q5 (7.87–28.95). Model 1: conditional logistic regression model adjusted for matching factors [study center, age (within 1 y), menopausal status at recruitment, use of exogenous hormones, phase of menstrual cycle, fasting status at blood collection, and time of blood collection (61 h)]. Model 2: conditional logistic regression model also adjusting for BMI (continuous), height (continuous), age at menarche (,12, 12–14, .14 y, missing), age at first full-term pregnancy (nulliparous, #20, .20 and #25, .25 and #30, .30 y, missing), oral contraceptive use (ever/never/missing, for premenopausal women), hormone therapy use (ever/never/missing, for postmenopausal women), smoking status (never, past, current, missing), alcohol consumption (g/d), educational level (none, primary school, technical/professional school, secondary school, university degree, missing), intake of saturated fatty acids (g/d), energy intake (kcal/d), and season of blood collection (winter, spring, summer, fall). ER+, estrogen receptor positive; ER2, estrogen receptor negative; Q, quintile. 2 Median value within quintile, based on controls.

ER+ Median2 Model 1 Model 2 g-Tocopherol, mmol/L ER2 Median2 Model 1 Model 2 ER+ Median2 Model 1 Model 2


TABLE 3 (Continued )

8 of 11 BAKKER ET AL.


9 of 11

FIGURE 1 ER/PR positive (n = 378) and ER/PR negative (n = 354) breast cancer ORs according to quintiles of plasma analytes. Stratified analysis based on conditional logistic regression analysis with linear trend tests across quintiles using median values of each quintile. P-heterogeneity across subtypes using a log-likelihood ratio test (with or without interaction terms). Diamonds represent ORs; lines represent 95% CIs. Lowest quintile represents reference category. ER2, estrogen receptor negative; ER+, estrogen receptor positive; PR2, progesterone receptor negative; PR+, progesterone receptor positive.

well as lower) differences in plasma concentrations of the investigated carotenoids compared with other studies investigating plasma analytes (20–25) might be related to laboratory differences, because in a study investigating another sample of the EPIC population but using different laboratory techniques, lower concentrations of plasma carotenoids were observed in general (3). Because carotenoids are fat soluble, their concentrations in plasma may be influenced by type and amount of fat consumed. We did not measure and were therefore not able to adjust for plasma lipid concentration, which may have led to underestimation of the carotenoid associations. However, it should be noted that in nested case-control studies that did have this information available, adjustment did not materially alter the results (20, 43). Although our findings support the hypothesis that high concentrations of several carotenoids and vitamin C may protect against the development of hormone receptor–specific breast cancer, this does not necessarily mean that it helps to take dietary supplements (44). The consumption of carotenoid- and vitaminrich food such as carrots as a dietary counterpart for a-carotene, carrots and leafy vegetables for b-carotene, and citrus fruits for vitamin C might lead to a possible lower (hormone receptor– specific) breast cancer risk. In general, effects of fruit and vegetable intake appear to be weak and largely confined to ER2 breast

cancer (36), but measurement errors inherent to dietary questionnaires may have diluted these effects. Also, we cannot exclude that other underlying lifestyle behaviors, genetic factors, or the availability of other bioactive compounds might be correlated with the biomarkers under study, leading to residual confounding (33, 45). Other substances present in the same foods as vitamin C and carotenoids might be responsible for the beneficial effects too. The inverse association attributed to these vitamin concentrations may also be produced by a combination of bioactive compounds working synergistically together. An important strength of this study is the inclusion of participants from 10 different countries with a large variation in consumption patterns. All plasma analytes were measured in the same laboratory and in samples collected well before breast cancer diagnosis, avoiding the potential modifying effect of cancer treatment and changes in lifestyle or dietary habits after the diagnosis of cancer. Further strengths of this study include large sample size, long follow-up time, and oversampling of ER2 and PR2 breast cancer subtypes, making it possible to evaluate these relatively infrequent but harmful breast cancer subtypes. In conclusion, women with higher plasma concentrations of b-carotene and a-carotene are at lower breast cancer risk of ER2 breast cancer.

10 of 11

BAKKER ET AL.

The authors’ responsibilities were as follows—MFB, PHMP, VMK, HBBdM, ER, and CHvG: designed the research; PHMP, HBBdM, EHJMJ, MMR, AT, KO, IR, HB, RK, AT, PV, SP, RT, JRQ, PA, M-JS, TJK, K-TK, and ER: collected data; MFB, PHMP, and CHvG: analyzed the data and wrote the manuscript; MFB, PHMP, NT, and CHvG: performed the statistical analysis; and all authors: revised the manuscript and read and approved the final manuscript. No authors declared a conflict of interest related to the study.

REFERENCES 1. Aune D, Chan DS, Vieira AR, Rosenblatt DA, Vieira R, Greenwood DC, Norat T. Fruits, vegetables and breast cancer risk: a systematic review and meta-analysis of prospective studies. Breast Cancer Res Treat 2012;134:479–93. 2. Natarajan L, Flatt SW, Sun X, Gamst AC, Major JM, Rock CL, Al-Delaimy W, Thomson CA, Newman VA, Pierce JP, et al. Validity and systematic error in measuring carotenoid consumption with dietary self-report instruments. Am J Epidemiol 2006;163:770–8. 3. Al-Delaimy WK, Ferrari P, Slimani N, Pala V, Johansson I, Nilsson S, Mattisson I, Wirfalt E, Galasso R, Palli D, et al. Plasma carotenoids as biomarkers of intake of fruits and vegetables: individual-level correlations in the European Prospective Investigation into Cancer and Nutrition (EPIC). Eur J Clin Nutr 2005;59:1387–96. 4. Block G, Norkus E, Hudes M, Mandel S, Helzlsouer K. Which plasma antioxidants are most related to fruit and vegetable consumption? Am J Epidemiol 2001;154:1113–8. 5. Fowke JH, Chung FL, Jin F, Qi D, Cai Q, Conaway C, Cheng JR, Shu XO, Gao YT, Zheng W. Urinary isothiocyanate levels, brassica, and human breast cancer. Cancer Res 2003;63:3980–6. 6. Tamimi RM, Hankinson SE, Campos H, Spiegelman D, Zhang S, Colditz GA, Willett WC, Hunter DJ. Plasma carotenoids, retinol, and tocopherols and risk of breast cancer. Am J Epidemiol 2005;161:153–60. 7. IARC Working Group on the Evaluation of Cancer Preventive Agents. IARC handbooks of cancer prevention. Vol. 2. Lyon (France): Carotenoids; 1998. 8. Kelley DS, Bendich A. Essential nutrients and immunologic functions. Am J Clin Nutr 1996;63:994S–6S. 9. Niki E. Action of ascorbic acid as a scavenger of active and stable oxygen radicals. Am J Clin Nutr 1991;54:1119S–24S. 10. Steinmetz KA, Potter JD. Vegetables, fruit, and cancer: II. Mechanisms. Cancer Causes Control 1991;2:427–42. 11. Aune D, Chan DS, Vieira AR, Navarro Rosenblatt DA, Vieira R, Greenwood DC, Norat T. Dietary compared with blood concentrations of carotenoids and breast cancer risk: a systematic review and meta-analysis of prospective studies. Am J Clin Nutr 2012;96: 356–73. 12. Eliassen AH, Hendrickson SJ, Brinton LA, Buring JE, Campos H, Dai Q, Dorgan JF, Franke AA, Gao YT, Goodman MT, et al. Circulating carotenoids and risk of breast cancer: pooled analysis of eight prospective studies. J Natl Cancer Inst 2012;104:1905–16. 13. Wu K, Helzlsouer KJ, Alberg AJ, Comstock GW, Norkus EP, Hoffman SC. A prospective study of plasma ascorbic acid concentrations and breast cancer (United States). Cancer Causes Control 2000;11:279–83. 14. Baglietto L, Krishnan K, Severi G, Hodge A, Brinkman M, English DR, McLean C, Hopper JL, Giles GG. Dietary patterns and risk of breast cancer. Br J Cancer 2011;104:524–31. 15. Boggs DA, Palmer JR, Wise LA, Spiegelman D, Stampfer MJ, AdamsCampbell LL, Rosenberg L. Fruit and vegetable intake in relation to risk of breast cancer in the Black Women’s Health Study. Am J Epidemiol 2010;172:1268–79. 16. Fung TT, Hu FB, McCullough ML, Newby PK, Willett WC, Holmes MD. Diet quality is associated with the risk of estrogen receptor– negative breast cancer in postmenopausal women. J Nutr 2006;136: 466–72. 17. Olsen A, Tjonneland A, Thomsen BL, Loft S, Stripp C, Overvad K, Moller S, Olsen JH. Fruits and vegetables intake differentially affects estrogen receptor negative and positive breast cancer incidence rates. J Nutr 2003;133:2342–7. 18. Zhang X, Spiegelman D, Baglietto L, Bernstein L, Boggs DA, van den Brandt PA, Buring JE, Gapstur SM, Giles GG, Giovannucci E, et al. Carotenoid intakes and risk of breast cancer defined by estrogen receptor and progesterone receptor status: a pooled analysis of 18 prospective cohort studies. Am J Clin Nutr 2012;95:713–25.

19. Zhang S, Hunter DJ, Forman MR, Rosner BA, Speizer FE, Colditz GA, Manson JE, Hankinson SE, Willett WC. Dietary carotenoids and vitamins A, C, and E and risk of breast cancer. J Natl Cancer Inst 1999; 91:547–56. 20. Hultén K, Van Kappel AL, Winkvist A, Kaaks R, Hallmans G, Lenner P, Riboli E. Carotenoids, alpha-tocopherols, and retinol in plasma and breast cancer risk in northern Sweden. Cancer Causes Control 2001;12: 529–37. 21. Dorjgochoo T, Gao YT, Chow WH, Shu XO, Li H, Yang G, Cai Q, Rothman N, Cai H, Franke AA, et al. Plasma carotenoids, tocopherols, retinol and breast cancer risk: results from the Shanghai Women Health Study (SWHS). Breast Cancer Res Treat 2009;117:381–9. 22. Epplein M, Shvetsov YB, Wilkens LR, Franke AA, Cooney RV, Le Marchand L, Henderson BE, Kolonel LN, Goodman MT. Plasma carotenoids, retinol, and tocopherols and postmenopausal breast cancer risk in the Multiethnic Cohort Study: a nested case-control study. Breast Cancer Res 2009;11:R49. 23. Kabat GC, Kim M, Adams-Campbell LL, Caan BJ, Chlebowski RT, Neuhouser ML, Shikany JM, Rohan TE, Investigators WHI. Longitudinal study of serum carotenoid, retinol, and tocopherol concentrations in relation to breast cancer risk among postmenopausal women. Am J Clin Nutr 2009;90:162–9. 24. Maillard V, Kuriki K, Lefebvre B, Boutron-Ruault MC, Lenoir GM, Joulin V, Clavel-Chapelon F, Chaje`s V. Serum carotenoid, tocopherol and retinol concentrations and breast cancer risk in the E3N-EPIC study. Int J Cancer 2010;127:1188–96. 25. Dorgan JF, Sowell A, Swanson CA, Potischman N, Miller R, Schussler N, Stephenson HE Jr. Relationships of serum carotenoids, retinol, alpha-tocopherol, and selenium with breast cancer risk: results from a prospective study in Columbia, Missouri (United States). Cancer Causes Control 1998;9:89–97. 26. Riboli E, Hunt KJ, Slimani N, Ferrari P, Norat T, Charrondie`re UR, Hémon B, Casagrande C, Vignat J, Overvad K, et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr 2002;5:1113–24. 27. Ku¨hn T, Kaaks R, Becker S, Eomois PP, Clavel-Chapelon F, Kvaskoff M, Dossus L, Tjonneland A, Olsen A, Overvad K, et al. Plasma 25hydroxyvitamin D and the risk of breast cancer in the European prospective investigation into cancer and nutrition: a nested case-control study. Int J Cancer 2013;133:1689–700. 28. Ros MM, Bueno-de-Mesquita HB, Kampman E, Aben KK, Bu¨chner FL, Jansen EH, van Gils CH, Egevad L, Overvad K, Tjonneland A, et al. Plasma carotenoids and vitamin C concentrations and risk of urothelial cell carcinoma in the European Prospective Investigation into Cancer and Nutrition. Am J Clin Nutr 2012;96:902–10. 29. Steghens JP, Van Kappel AL, Riboli E, Collombel C. Simultaneous measurement of seven carotenoids, retinol and alpha-tocopherol in serum by high-performance liquid chromatography. J Chromatogr B Biomed Sci Appl 1997;694:71–81. 30. Wareham NJ, Jakes RW, Rennie KL, Schuit J, Mitchell J, Hennings S, Day NE. Validity and repeatability of a simple index derived from the short physical activity questionnaire used in the European Prospective Investigation into Cancer and Nutrition (EPIC) study. Public Health Nutr 2003;6:407–13. 31. James RE, Lukanova A, Dossus L, Becker S, Rinaldi S, Tjonneland A, Olsen A, Overvad K, Mesrine S, Engel P, et al. Postmenopausal serum sex steroids and risk of hormone receptor–positive and –negative breast cancer: a nested case-control study. Cancer Prev Res (Phila) 2011;4: 1626–35. 32. Brooks PJ. DNA damage, DNA repair, and alcohol toxicity—a review. Alcohol Clin Exp Res 1997;21:1073–82. 33. Key TJ. Fruit and vegetables and cancer risk. Br J Cancer 2011;104: 6–11. 34. Lachance PA, Nakat Z, Jeong WS. Antioxidants: an integrative approach. Nutrition 2001;17:835–8. 35. Nagel G, Linseisen J, Van Gils CH, Peeters PHM, Boutron-Ruault MC, Clavel-Chapelon F, Romieu I, Tjonneland A, Olsen A, Roswall N, et al. Dietary beta-carotene, vitamin C and E intake and breast cancer risk in the European Prospective Investigation into Cancer and Nutrition (EPIC). Breast Cancer Res Treat 2010;119:753–65. 36. Jung S, Spiegelman D, Baglietto L, Bernstein L, Boggs DA, van den Brandt PA, Buring JE, Cerhan JR, Gaudet MM, Giles GG, et al. Fruit and vegetable intake and risk of breast cancer by hormone receptor status. J Natl Cancer Inst 2013;105:219–36.

PLASMA ANTIOXIDANTS AND RISK OF BREAST CANCER 37. Wang Y, Gapstur SM, Gaudet MM, Furtado JD, Campos H, McCullough ML. Plasma carotenoids and breast cancer risk in the Cancer Prevention Study II Nutrition Cohort. Cancer Causes Control 2015;26:1233–44. 38. Eliassen AH, Liao X, Rosner B, Tamimi M, Tworoger SS, Hankinson SE. Plasma carotenoids and risk of breast cancer over 20 y of followup. Am J Clin Nutr 2015;101:1197–205. 39. Buckland G, Travier N, Cottet V, González CA, Luján-Barroso L, Agudo A, Trichopoulou A, Lagiou P, Trichopoulos D, Peeters PHM, et al. Adherence to the Mediterranean diet and risk of breast cancer in the European Prospective Investigation into Cancer and Nutrition cohort study. Int J Cancer 2013;132:2918–27. 40. Toniolo P, Van Kappel AL, Akhmedkhanov A, Ferrari P, Kato I, Shore RE, Riboli E. Serum carotenoids and breast cancer. Am J Epidemiol 2001;153:1142–7. 41. Comstock GW, Alberg AJ, Helzlsouer KJ. Reported effects of long-term freezer storage on concentrations of retinol, beta-carotene, and alphatocopherol in serum or plasma summarized. Clin Chem 1993;39:1075–8.

11 of 11

42. Jenab M, Bingham S, Ferrari P, Friesen MD, Al-Delaimy WK, Luben R, Wareham N, Khaw KT, Riboli E. Long-term cryoconservation and stability of vitamin C in serum samples of the European prospective investigation into cancer and nutrition. Cancer Epidemiol Biomarkers Prev 2005;14:1837–40. 43. Jenab M, Riboli E, Ferrari P, Friesen M, Sabate J, Norat T, Slimani N, Tjonneland A, Olsen A, Overvad K, et al. Plasma and dietary carotenoid, retinol and tocopherol levels and the risk of gastric adenocarcinomas in the European Prospective Investigation into Cancer and Nutrition. Br J Cancer 2006;95:406–15. 44. Druesne-Pecollo N, Latino-Martel P, Norat T, Barrandon E, Bertrais S, Galan P, Hercberg S. Beta-carotene supplementation and cancer risk: a systematic review and metaanalysis of randomized controlled trials. Int J Cancer 2010;127:172–84. 45. Jenab M, Slimani N, Bictash M, Ferrari P, Bingham SA. Biomarkers in nutritional epidemiology: applications, needs and new horizons. Hum Genet 2009;125:507–25.