original articles
Annals of Oncology Annals of Oncology 25: 1422–1428, 2014 doi:10.1093/annonc/mdu150 Published online 8 April 2014
Circulating prolactin and breast cancer risk among pre- and postmenopausal women in the EPIC cohort
1 Division of Cancer Epidemiology, German Cancer Research Center (DKFZ) Heidelberg, Heidelberg, Germany; 2Section of Nutrition and Metabolism, International Agency for Research on Cancer (IARC), Lyon, France; 3Danish Cancer Society Research Center, Copenhagen; 4Section for Epidemiology, Department of Public Health, Aarhus University, Aarhus, Denmark; 5INSERM, Centre for Research in Epidemiology and Population Health [CESP], Nutrition, Hormones and Women’s Health Team, Villejuif; 6 University of Paris Sud, UMRS, Villejuif; 7IGR, Villejuif, France; 8Cancer Epidemiology Centre, Cancer Council Victoria, Melbourne; 9Centre for Molecular, Environmental, Genetic, and Analytic Epidemiology, The University of Melbourne, Melbourne, Australia; 10Department of Epidemiology, German Institute of Human Nutrition (DIfE) Potsdam-Rehbrücke, Nuthetal, Germany; 11Department of Hygiene, Epidemiology and Medical Statistics, University of Athens Medical School, Athens; 12Hellenic Health Foundation, Athens, Greece; 13Department of Epidemiology, Harvard School of Public Health, Boston, USA; 14Bureau of Epidemiologic Research, Academy of Athens, Athens, Greece; 15Molecular and Nutritional Epidemiology Unit, Cancer Research and Prevention Institute—ISPO, Florence, Italy; 16Epidemiology and Prevention Unit, National Tumor Institute (IRCCS), Milano; 17Cancer Registry and Histopathology Unit, ‘Civic-M. P. Arezzo’ Hospital ASP, Ragusa; 18Piedmont Cancer Registry, Centre for Epidemiology and Prevention in Oncology in Piedmont, Torino; 19Department of Clinical and Experimental Medicine, Federico II University, Naples, Italy; 20Unit of Nutrition, Environment and Cancer, Catalan Institute of Oncology-ICO, IDIBELL, Barcelona; 21Public Health Directorate, Asturias; 22Granada Cancer Registry, Andalusian School of Public Health, Granada; 23Consortium for Biomedical Research in Epidemiology and Public Health (CIBER), Madrid; 24Public Health Division of Gipuzkoa, Basque Regional Health Department, San Sebastian; 25Department of Epidemiology, Murcia Regional Health Authority, Murcia; 26Navarre Public Health Institute, Pamplona, Spain; 27 National Institute for Public Health and the Environment (RIVM), Bilthoven; 28Department of Gastroenterology and Hepatology, University Medical Centre, Utrecht, The Netherlands; 29School of Public Health, Imperial College, London, UK; 30Department of Epidemiology, Julius Center for Health Sciences and Primary Care, University Medical Center, Utrecht, The Netherlands; 31Department of Radiation Sciences, University of Umeå, Umeå; 32Department of Surgical and Perioperative Sciences, Umeå University, Umeå, Sweden; 33Department of Community Medicine, Faculty of Health Sciences, University of Tromsø, Tromsø; 34Department of Research, Cancer Registry of Norway, Oslo, Norway; 35Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden; 36Society of Public Health, Helsinki, Finland; 37 School of Clinical Medicine, University of Cambridge, Cambridge; 38Cancer Epidemiology Unit, University of Oxford, Oxford, UK
Received 24 January 2014; revised 26 February 2014; accepted 26 March 2014
Background: Experimental and epidemiological evidence suggests that prolactin might play a role in the etiology of breast cancer. We analyzed the relationship of prediagnostic circulating prolactin levels with the risk of breast cancer by menopausal status, use of postmenopausal hormone replacement therapy (HRT) at blood donation, and by estrogen and progesterone receptor status of the breast tumors. Patients and methods: Conditional logistic regression was used to analyze the data from a case–control study nested within the prospective European EPIC cohort, including 2250 invasive breast cancer and their matched control subjects. Results: Statistically significant heterogeneity in the association of prolactin levels with breast cancer risk between women who were either pre- or postmenopausal at the time of blood donation was observed (Phet = 0.04). Higher serum levels of prolactin were associated with significant increase in the risk of breast cancer among postmenopausal women [odds ratio (OR)Q4–Q1 = 1.29 (95% confidence interval, CI, 1.05–1.58), Ptrend = 0.09]; however, this increase in risk seemed to be confined to women who used postmenopausal HRT at blood donation [ORQ4–Q1 = 1.45 (95% CI 1.08– 1.95), Ptrend = 0.01], whereas no statistically significant association was found for the non-users of HRT [ORQ4–Q1 = 1.11 (95%CI 0.83–1.49), Ptrend = 0.80] (Phet = 0.08). Among premenopausal women, a statistically non-significant inverse
*Correspondence to: Prof. Rudolf Kaaks, Department of Cancer Epidemiology, German Cancer Research Center (DKFZ) Heidelberg, Im Neuenheimer Feld 581, D-69120 Heidelberg, Germany. Tel: +49-6221-42-22-19; Fax: +49-6221-42-22-03; E-mail: [email protected]
© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
K. Tikk1, D. Sookthai1, T. Johnson1, S. Rinaldi2, I. Romieu2, A. Tjønneland3, A. Olsen3, K. Overvad4, F. Clavel-Chapelon5,6,7, L. Baglietto8,9, H. Boeing10, A. Trichopoulou11,12, P. Lagiou11,13,14, D. Trichopoulos12,13,14, D. Palli15, V. Pala16, R. Tumino17, S. Rosso18, S. Panico19, A. Agudo20, V. Menéndez21, M.-J. Sánchez22,23, P. Amiano23,24, J. M. Huerta Castaño23,25, E. Ardanaz23,26, H. B. Bueno-de-Mesquita27,28,29, E. Monninkhof30, C. Onland-Moret30, A. Andersson31, M. Sund32, E. Weiderpass33,34,35,36, K.-T. Khaw37, T. J. Key38, R. C. Travis38, M. J. Gunter29, E. Riboli29, L. Dossus5,6,7 & R. Kaaks1*
Annals of Oncology
original articles
association was observed [ORQ4–Q1 = 0.70 (95% CI 0.48–1.03), Ptrend = 0.16]. There was no heterogeneity in the prolactin–breast cancer association by hormone receptor status of the tumor. Conclusion: Our study indicates that higher circulating levels of prolactin among the postmenopausal HRT users at baseline may be associated with increased breast cancer risk. Key words: breast cancer, prolactin levels, hormone replacement therapy, estrogen receptor, progesterone receptor, prospective cohort
introduction
study population and methods the EPIC cohort The EPIC cohort is based on 366 521 women and 153 457 men recruited between the years 1992 and 2000 in 10 European countries: Denmark, France, Germany, Greece, Italy, Norway, Spain, Sweden, The Netherlands, and the UK. Details on the subject recruitment, baseline data, and blood collection protocols have been reported previously [11]. In brief, anthropometric measurements and questionnaire data on diet, health status,
Volume 25 | No. 7 | July 2014
design of the nested case–control study General criteria for the inclusion of cases and matched controls into the study were that women could be clearly classified as being either pre- or postmenopausal at the time of blood donation, and did not have any previous diagnosis of cancer (except non-melanoma skin cancer). Women who were ≤42 years of age or reported having had at least nine menstrual periods in the last 12 months were classified as premenopausal. Women were classified as postmenopausal when they reported not having any menses over the past 12 months, when older than 55 years of age, or when reporting bilateral ovariectomy. Women aged 42 years or older whose questionnaire data on menopausal status were incomplete were not included to the present case– control study. For each case-subject, one control subject closest to the case (based on matching criteria) with an available blood sample was chosen among appropriate risk sets consisting of all cohort members alive and free of cancer at the time of diagnosis of the index case. Matching criteria were the study recruitment center, menopausal status at blood donation, age at blood donation (±6 months), time at blood donation (±2 h), fasting status, phase of menstrual cycle [matching categories for premenopausal women: early follicular (days 0–7 of the cycle), late follicular (days 8–11), mid-cycle (days 12–16), early luteal (days 17–19), mid-luteal (days 20–24), late luteal (days 25+)], and current use of HRT at the time of blood donation (for postmenopausal women). In total, 2271 invasive breast cancer case subjects and an equal number of control subjects were selected for the study. After exclusion of case or control subjects with missing prolactin
doi:10.1093/annonc/mdu150 |
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
The polypeptide hormone prolactin promotes the proliferation of breast epithelial cells, and in combination with progesterone promotes the differentiation of specialized structures—alveoli—that synthesize and secrete milk during lactation [1]. Associations of several breast cancer risk factors, such as nulli-parity and high mammographic breast density with increased serum prolactin concentrations, [2, 3] in combination with animal experimental data led to the hypothesis that elevated prolactin might promote breast tumor development [4, 5]. One large and four very small prospective studies have previously addressed possible associations between circulating prolactin levels and breast cancer risk [4, 6–9]. In the Nurses’ Health Study (NHS), the only large-scale prospective study to date, a series of nested case–control studies indicated a modest positive risk association between serum prolactin concentrations and breast cancer risk among postmenopausal women and especially with the risk of estrogen receptor-positive (ER+) disease, but showed no significant associations among premenopausal women or with ER-negative breast cancer [4, 10]. The other four studies, which each included only small numbers of incident breast cancer cases (n = 20–173), showed no significant relationship between serum prolactin and breast cancer risk among either pre- or postmenopausal women [6–9] and did not provide any information with regard to hormone receptor status of the tumors. In this large case–control study nested within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort, we examined the association of prediagnostic serum levels of prolactin among pre- and postmenopausal women with subsequent risk of breast cancer, overall, by hormone receptor status of the tumor and by the use of postmenopausal hormone replacement therapy (HRT) at the time of blood donation. With a total of 2250 incident breast cancer cases, including 817 postmenopausal case subjects reporting HRT use at baseline, this is the second major prospective study to date on the relationships between prediagnostic prolactin levels and subsequent breast cancer risk.
reproductive and menstrual history, and medication use was collected at baseline. All EPIC study participants provided written consent for the use of questionnaire information and blood samples in research studies. Ethical approval for the nested case–control study reported here was obtained from the ethical review boards of International Agency for Research on Cancer (IARC) (Lyon, France) and from the local ethics committees in participating countries/study centers. Details on the identification of vital status, breast cancer incidence, and information/quantification descriptions for the hormone receptor status of the tumors across the study centers were reported previously [11, 12]. For the present study, the closure date for follow-up was the last date of complete followup for both cancer incidence and vital status, which ranged from 2003 to 2006, depending on the study center. Sweden was not included in the analysis because independent studies were being completed on breast cancer risk and endogenous hormones [9].
original articles
Annals of Oncology
measurements (n = 17) and with the prolactin value above detection range of the assay (more than 133 ng/ml, n = 4), 2250 incident cases and an equal number of controls remained for the analyses.
laboratory assays The prolactin analyses were performed in the laboratory of the Division of Cancer Epidemiology at the German Cancer Research Centre and determined by immunoradiometric assay [IRMA (CT), IBL International GMBH, Germany] from blood taken at baseline. The detection range of the assay was 0.35–133 ng/ml. The same quality controls were included in each analytical batch. The mean inter- and intra-assay coefficients of variation were 4.62% and 2.17%, respectively.
Conditional logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for breast cancer occurrence, calculating ORs over the quartile levels and on a continuous log2 scale of circulating prolactin. As circulating prolactin levels show variation over the menopausal status, the quartile cutoff points were calculated separately based on the control population of pre- and postmenopausal women. We a priori decided to adjust all the models for number of full-term pregnancies (0, 1, 2, 3, ≥4, missing), smoking status (current, never, previous, missing), and body mass index (continuous scale) due to modest variation in prolactin levels over these factors in healthy women. A more detailed description of the statistical analysis used in this study is described in supplementary Material, available at Annals of Oncology online. All P-values presented are two-sided and P < 0.05 was considered statistically significant. Statistical analyses were conducted using SAS software, version 9.2 (SAS Institute, Cary, NC).
results This nested case–control study consisted of 2250 case subjects with incident, invasive breast cancer matched to an equal number of control subjects. A total of 1738 case–control pairs were sampled among women who could be unambiguously classified
Table 1. Baseline characteristics of case and control subjects [data presented as median (min, max) or n (%)]
Age at blood donation (years) Age at menopause (years) Age at diagnosis (years) Lag time till diagnosis (years) Body mass index (kg/m2) Ever had a full-term pregnancy Baseline smoking Baseline HRTb use a
Premenopausal womena Cases (512) 46.2 (26.7–56.9) — 50.9 (35.2–64.2) 5.2 (0.04–11.7) 23.6 (16.5–49.1) 438 (86%) 109 (21%) –
Controls (512)
Postmenopausal womena Cases (1738)
Controls (1738)
46.1 (26.7–57.1) — — — 23.9 (16.8–46.3) 442 (88%) 111 (22%) –
58.3 (41.1–75.5) 50 (15–63) 63.0 (45.3–83.6) 4.6 (0.01–12.0) 24.7 (13.8–47.7) 1453 (85%) 325 (19%) 817 (47%)
58.3 (41.2–76.8) 50 (21–63) — — 24.5 (16.0–45.3) 1497 (88%) 324 (19%) 817 (47%)
Menopausal status at the time of blood donation. Hormone replacement therapy.
b
| Tikk et al.
Volume 25 | No. 7 | July 2014
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
statistical analyses
as being postmenopausal at blood donation, and 512 pairs were sampled among premenopausal women. For the premenopausal case–control pairs, the median age at blood donation (baseline) was 46.2 (range: 26.7–57.1) years, and the median age at tumor diagnosis for the case subjects was 50.9 (range: 35.2–64.2) years; for the postmenopausal women, the ages at blood donation and at breast cancer diagnosis were 58.3 (range: 41.1–76.8) and 63.0 (range: 45.3–83.6) years, respectively (Table 1). A total of 817 postmenopausal case–control pairs were using HRT at the time they provided their blood sample. In linear regression analyses adjusted for matching factors, the number of full-term pregnancies, smoking status, and body mass index, our data showed significantly higher prolactin levels among premenopausal women when compared with postmenopausal women [adjusted geometric mean levels 7.87 (95% CI 7.4–8.3) versus 6.54 (95% CI 6.3–6.8) ng/ml], and among postmenopausal HRT users (especially among combined estrogenplus-progestin users) compared with non-users at baseline [adjusted geometric mean levels 6.72 (95% CI 6.4–7.1) versus 5.94 (95% CI 5.7–6.2) ng/ml]. In case–control comparisons, prolactin levels differed on average ∼3% between case and control subjects, with the greatest difference among postmenopausal women who used HRT at baseline (5.9%) (data not shown). With regard to our risk models (logistic regression), as statistical tests indicated significant heterogeneity in the strength of the association between breast cancer and circulating prolactin levels by menopausal status at baseline (Phet = 0.04), all statistical analyses were conducted separately for pre- and postmenopausal women. Among the premenopausal women, conditional logistic regression analyses adjusting for smoking status, number of full-term pregnancies, and body mass index suggest a negative association of serum prolactin levels with overall breast cancer risk, which was not statistically significant [ORQ4–Q1 = 0.70 (95% CI 0.48–1.03); Ptrend = 0.16] (Table 2). Among the postmenopausal women, there was a positive association of prolactin levels with breast cancer risk overall, with a statistically significant OR of 1.29 (95% CI 1.05–1.58) in the highest quartile of prolactin, but no statistically significant trend based on prolactin as a continuous log2 variable (Ptrend = 0.09). When the analyses were stratified further by baseline HRT use, statistically
original articles
Annals of Oncology
Table 2. Adjusted odds ratiosa [OR (95% CI)] for overall breast cancer by quartile levels and on a continuous log2 scale of circulating prolactin Quartiles 1
2
Premenopausal women ca/co 145/128 128/128 OR Ref. 0.83 (0.57, 1.22) Postmenopausal women (all) ca/co 383/435 418/437 OR Ref. 1.09 (0.90, 1.33) Postmenopausal non-HRT users ca/co 250/272 250/245 OR Ref. 1.09 (0.84, 1.41) Postmenopausal HRT users ca/co 133/163 168/192 OR Ref. 1.09 (0.80, 1.50)
Log2 OR (95% CI)
Ptrend b
3
4
128/128 0.83 (0.57, 1.22)
111/128 0.70 (0.48, 1.03)
0.87 (0.71, 1.06)
0.16
455/432 1.22 (0.99, 1.49)
482/434 1.29 (1.05, 1.58)
1.10 (0.98, 1.22)
0.09
239/229 1.12 (0.85, 1.48)
182/175 1.11 (0.83, 1.49)
0.98 (0.84, 1.15)
0.80
216/203 1.34 (0.98, 1.84)
300/259 1.45 (1.08, 1.95)
1.21 (1.04, 1.41)
0.01
Phet c 0.04
0.08
Adjusted for full-term pregnancies, smoking status, and body mass index. Linear trends for OR estimates over a continuous scale of prolactin levels. c Statistical tests for heterogeneity were based on likelihood-ratio test, comparing the model fit for logistic regression models with and without corresponding interaction term. b
significant ORs were observed among the HRT users [ORQ4– Q1 = 1.45 (95% CI 1.08–1.95), on a continuous log2 scale ORlog2 = 1.21 (95% 1.04–1.41), Ptrend = 0.01], but not among the non-users [ORQ4–Q1 = 1.11 (95% CI 0.83–1.49), Ptrend = 0.80], although the test for heterogeneity between HRT users and nonusers was only of borderline statistical significance (based on continuous log2 scale, Phet = 0.08). Stratifying the analyses by hormone receptor status of the tumor, there was no evidence of heterogeneity in the prolactin–breast cancer relationship by tumor subtype, overall or in strata of users or non-users of HRT (Table 3). Among postmenopausal women using combined estrogen-plus-progestin HRT, the relative risk appeared to be more pronounced for ER + tumors [ORlog2 = 1.27 (95% CI 1.03–1.57), Ptrend = 0.03, on a continuous log2 scale, n = 338/425] than among users of estrogens only therapy [ORlog2 = 1.06 (95% CI 0.72–1.58), Ptrend = 0.77, n = 103/206], although the test for statistical heterogeneity was not statistically significant (Phet = 0.44) (data not shown). In subgroup analyses, statistical tests showed no evidence of heterogeneity in the strength of the association with overall breast cancer risk by lag time until breast cancer diagnosis (supplementary Table S1, available at Annals of Oncology online), and overall, there was also no significant heterogeneity by age at tumor diagnosis in either pre- or postmenopausal women (data not shown). Among the postmenopausal women using HRT at baseline, however, circulating levels of prolactin were significantly associated with breast cancer diagnosed above the median age (61 years) within this subgroup [ORlog2 = 1.46 (95% CI 1.18–1.80), Ptrend = 0.0004, n = 370 sets], but not with breast cancer diagnosed at a younger age [ORlog2 = 0.95 (95% CI 0.76–1.19), Ptrend = 0.65, n = 447 sets] (Phet = 0.01) (data not shown). Complementary statistical analyses showed no evidence for major confounding or interaction effects by other reproductive
Volume 25 | No. 7 | July 2014
and lifestyle factors, or by levels of the other endogenous hormones (data not shown). Furthermore, sensitivity analyses in which the first 2 years of follow-up were excluded did not show any major difference in risk estimates compared with the full dataset.
discussion In this large prospective study, we observed a statistically significant heterogeneity in the association of prolactin levels with breast cancer risk between women who were either pre- or postmenopausal at the time of blood donation. Among postmenopausal women, there was a significant positive association of prolactin levels with overall breast cancer risk; however, this increase in risk seemed to be confined to women who used HRT at the time of blood donation. Among premenopausal women, there was a statistically non-significant inverse association of prolactin with breast cancer risk. Our data showed no evidence for heterogeneity in the relationship of prolactin levels with risk of hormone receptor positive or negative breast tumors. Our findings show partial agreement with those from the Nurses’ Health Study cohorts I and II, in which a series of nested case–control studies were conducted to address the relationship of serum prolactin levels with breast tumor development [4, 10] that included up to 3421 breast cancer cases [4]. In the Nurses’ studies, using data based on only the first 10 years of prospective follow-up (comparable to the follow-up time our present study) with 2468 breast cancer (of whom 1023 among pre- and 1445 among postmenopausal women) and 4021 control subjects, there was a positive association between serum prolactin and overall breast cancer risk among postmenopausal women, and no association among women who were premenopausal at the time of blood draw (Phet = 0.01). Our observations are similar to these findings. However, contrary to our doi:10.1093/annonc/mdu150 |
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
a
original articles
Annals of Oncology
Table 3. Adjusted odds ratiosa [OR (95% CI)] for breast cancer by receptor status of the tumor by quartile levels and on a continuous log2 scale of circulating prolactin 1
2
3
Log2 OR
Ptrend b
88/101 0.70 (0.44, 1.09)
0.85 (0.68, 1.07)
0.18
23/25 0.69 (0.30, 1.61)
0.96 (0.65, 1.42)
0.84
56/73 0.56 (0.32, 0.99)
0.80 (0.61, 1.05)
0.11
18/14 1.02 (0.31, 3.33)
1.31 (0.78, 2.19)
0.30
384/358 1.29 (1.02, 1.63)
1.10 (0.97, 1.24)
0.13
1.09 (0.85, 1.40)
0.48
1.23 (0.88, 1.70) 71/55
1.07 (0.90, 1.27)
0.42
1.14 (0.66, 1.95)
1.00 (0.75, 1.32)
0.98
137/139 1.12 (0.80, 1.57)
0.97 (0.81, 1.17)
0.76
0.99 (0.71, 1.39)
0.96
1.11 (0.69, 1.78) 35/26
0.97 (0.75, 1.24)
0.78
0.94 (0.43, 2.05)
0.97 (0.66, 1.41)
0.86
247/219 1.42 (1.01, 1.99)
1.21 (1.03, 1.43)
0.02
1.18 (0.81, 1.73)
0.38
1.22 (0.75, 2.00) 36/29
1.15 (0.90, 1.47)
0.27
1.52 (0.65, 3.54)
1.10 (0.67, 1.78)
0.71
96/74 1.36 (0.87, 2.14) 210/196
44/34 1.14 (0.59, 2.21) 73/76
52/40 1.67 (0.86, 3.24) 137/120
Phet c
0.77
0.24
0.96
0.76
0.83
0.92
0.96
0.58
a
Adjusted for full-term pregnancies, smoking status, and body mass index. Linear trends for OR estimates over a continuous scale of prolactin levels. c Statistical tests for heterogeneity were based on likelihood-ratio test, comparing the model fit for logistic regression models with and without corresponding interaction term. b
| Tikk et al.
Volume 25 | No. 7 | July 2014
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
Premenopausal women ER+ ca/co 114/102 91/87 92/95 OR Ref. 0.91 (0.57, 1.43) 0.82 (0.53, 1.27) ER− ca/co 30/26 34/40 36/32 OR Ref. 0.57 (0.25, 1.27) 0.80 (0.34, 1.88) ER+PR+ ca/co 88/76 62/62 68/63 OR Ref. 0.78 (0.45, 1.35) 0.83 (0.49, 1.42) ER−PR− ca/co 19/16 21/25 19/22 OR Ref. 0.69 (0.24, 1.94) 0.71 (0.22, 2.23) All postmenopausal women combined ER+ ca/co 302/349 348/344 373/356 OR Ref. 1.20 (0.96, 1.50) 1.25 (0.99, 1.57) ER− ca/co 80/85 68/92 78/71 OR Ref. 0.77 (0.5, 1.19) 1.13 (0.72, 1.79) ca/co 155/171 191/174 191/206 ER+PR+ OR Ref. 1.22 (0.89, 1.69) 1.03 (0.74, 1.44) ca/co 50/45 49/66 47/51 ER−PR− OR Ref. 0.65 (0.37, 1.14) 0.77 (0.43, 1.39) Postmenopausal women, not using HRT at the time of blood donation ER+ ca/co 198/223 206/190 193/182 OR Ref. 1.24 (0.92, 1.66) 1.21 (0.88, 1.66) ER− ca/co 51/48 42/54 42/43 OR Ref. 0.74 (0.42, 1.29) 0.84 (0.45, 1.57) ca/co 102/114 117/99 97/100 ER+PR+ OR Ref. 1.32 (0.87, 2.01) 1.09 (0.70, 1.70) ca/co 31/24 28/40 27/31 ER−PR− OR Ref. 0.51 (0.25, 1.07) 0.54 (0.24, 1.25) Postmenopausal women, using HRT at the time of blood donation ER+ ca/co 104/126 142/154 180/174 OR Ref. 1.14 (0.80, 1.63) 1.29 (0.91, 1.83) ER− ca/co 29/37 26/38 36/28 OR Ref. 0.87 (0.41, 1.88) 1.69 (0.82, 3.48) ca/co 53/57 74/75 94/106 ER+PR+ OR Ref. 1.10 (0.65, 1.84) 0.93 (0.56, 1.55) ca/co 19/21 21/26 20/20 ER−PR− OR Ref. 0.82 (0.29, 2.36) 1.23 (0.49, 3.09)
4
original articles
Annals of Oncology
Volume 25 | No. 7 | July 2014
by receptor status has been shown to be relatively robust and to have only moderate dependence on the assay used, especially for medium- and high-expressing tumors [17], staining for receptor status was performed in different laboratories both within and across EPIC centers, and this may have led to some misclassification by receptor status. Finally, in spite of the large study size (a total of 2250 matched case–control pairs), our study still had relatively limited numbers of case subjects for hormone receptor negative tumors within strata of menopausal status and postmenopausal HRT use, and thus the statistical power for detecting associations of prolactin with risk in these subgroups was limited.
conclusions Our study shows that higher circulating levels of prolactin among the postmenopausal women who used HRT at baseline are associated with increased breast cancer risk. In contrast, statistically non-significant inverse association of premenopausal prolactin levels and subsequent breast cancer risk was seen. Our data did not show any heterogeneity in the prolactin-breast cancer association by hormone receptor status of the tumor among either pre- or postmenopausal women.
acknowledgements We thank all the EPIC cohort participants. Furthermore, we thank Britta Lederer and Sigrid Henke for their work in conducting the immunoassays; Sabine Rohrmann, Jutta Schmitt, and Jutta Kneisel for their assistance during the collection of hormone receptor status data. The coordination of EPIC is financially supported by the European Commission (DG-SANCO) and the International Agency for Research on Cancer. The national cohorts are supported by Danish Cancer Society (Denmark); Ligue contre le Cancer, Mutuelle Générale de l’Éducation Nationale, Institut National de la Santé et de la Recherche Médicale (France); Deutsche Krebshilfe, Deutsches Krebsforschungszentrum and Federal Ministry of Education and Research (Germany); the Hellenic Health Foundation (Greece); Italian Association for Research on Cancer (AIRC) and National Research Council (Italy); Dutch Ministry of Public Health, Welfare and Sports (VWS), Netherlands Cancer Registry (NKR), LK Research Funds, Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF), Statistics Netherlands (The Netherlands); Norwegian Research Council, Norwegian Cancer Society, ERC-2009-AdG 232997 and Nordforsk, Nordic Centre of Excellence programme on Food, Nutrition and Health (Norway); Health Research Fund (FIS), The Spanish Ministry of Health (ISCIII RETICC RD06/0020/0091) and the Catalan Institute of Oncology, Regional Governments of Andalucía, Asturias, Basque Country, Murcia (no 6236) and Navarra, ISCIII RETIC (RD06/ 0020; Spain); Swedish Cancer Society, Swedish Scientific Council and Regional Government of Skåne and Västerbotten (Sweden); Cancer Research UK, Medical Research Council (UK).
funding This research did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.
doi:10.1093/annonc/mdu150 |
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
observations, in the Nurses’ cohorts, the direct association among postmenopausal women appeared to be clearly stronger for ER/PR-positive tumors [RRQ4–Q1 = 1.53 (95% CI 1.17–1.98), Ptrend = 0.0002] and the absence of an association in ER/PRnegative tumors was reported [RRQ4–Q1 = 0.88 (95% CI 0.52– 1.47), Ptrend = 0.94] (Phet = 0.04) [4]. In our data, there was no evidence for heterogeneity of the prolactin–breast cancer association by receptor status. In several other, much smaller prospective studies, the heterogeneity of relationships of prolactin with breast cancer risk by receptor status was not addressed [6–9]. The observed associations between circulating prolactin and subsequent breast cancer risk, both in our study and in the Nurses’ cohorts, were of a modest magnitude compared with the associations reported for other endogenous hormones, such as testosterone and estradiol, with relative risk estimates up to 2.5 for top versus bottom tertile levels [12]. Nonetheless, a vast body of evidence from animal and in vitro studies supports the involvement of prolactin in breast cancer development. Especially in combination with the ovarian steroid hormones estradiol and progesterone, prolactin has effects on normal epithelial cell expansion, ductal side branching of the breast during puberty, and formation of lobuloalveolar structures during pregnancy [1]. Furthermore, prolactin and progesterone have synergic roles in inducing cell growth and proliferative activity in the terminal duct-lobular units of the breast—the site of origin of most breast cancers—and evidence is accumulating for crosstalk between progesterone, prolactin, and their receptor signaling pathways in normal and malignant breast cells [13–15]. We observed a more pronounced association between circulating prolactin and breast cancer among postmenopausal women who were using HRT at the time of blood donation, compared with the association among HRT non-users. Given the well-documented biological synergy between prolactin and progesterone, it is striking that our data indicated a significant association of prolactin with the risk of receptor-positive tumors among women using combined estrogen-plus-progestin HRT, but not among the users of estrogen-only. However, only a reduced number of women in our study had detailed information about the types of HRT used, and the statistical test for heterogeneity between estrogen-plus-progestin users and estrogen-only users was not statistically significant. Furthermore, in the Nurses’ cohorts associations of prolactin with breast cancer risk were reported to be relatively independent of exogenous hormone use at the time of blood donation (Phet = 0.65) [4, 10]. At the time the NHS was conducted, most women were on estrogen only HRT and relatively few were on combined estrogenplus-progestin HRT ( personal communication), which may explain why NHS did not see a difference by HRT use. Major strengths of our study are its prospective design and the large number of incident cases with the information on hormone receptor status of the tumor, which allowed subgroup analyses describing risk associations by menopausal status, postmenopausal HRT use, and receptor status of the tumor. A limitation of our study is that for each woman we had prolactin measurements for only one blood sample, which may limit interpretations about women’s average prolactin levels over time periods longer than 5–10 years [4, 16]. Furthermore, although classification of breast cancer outcomes
original articles disclosure The authors have declared no conflicts of interest.
references
9. Manjer J, Johansson R, Berglund G et al. Postmenopausal breast cancer risk in relation to sex steroid hormones, prolactin and SHBG (Sweden). Cancer Causes Control 2003; 14: 599–607. 10. Tworoger SS, Eliassen AH, Rosner B et al. Plasma prolactin concentrations and risk of postmenopausal breast cancer. Cancer Res 2004; 64: 6814–6819. 11. Riboli E, Hunt KJ, Slimani N et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr 2002; 5: 1113–1124. 12. James RE, Lukanova A, Dossus L et al. Postmenopausal serum sex steroids and risk of hormone receptor-positive and -negative breast cancer: a nested case− control study. Cancer Prev Res 2011; 4: 1626–1635. 13. Ormandy CJ, Hall RE, Manning DL et al. Coexpression and cross-regulation of the prolactin receptor and sex steroid hormone receptors in breast cancer. J Clin Endocrinol Metab 1997; 82: 3692–3699. 14. Lee HJ, Ormandy CJ. Interplay between progesterone and prolactin in mammary development and implications for breast cancer. Mol Cell Endocrinol 2012; 357: 101–107. 15. Perks CM, Keith AJ, Goodhew KL et al. Prolactin acts as a potent survival factor for human breast cancer cell lines. Br J Cancer 2004; 91: 305–311. 16. Arslan AA, Gu Y, Zeleniuch-Jacquotte A et al. Reproducibility of serum pituitary hormones in women. Cancer Epidemiol Biomarkers Prev 2008; 17: 1880–1883. 17. von Wasielewski R, Mengel M, Wiese B et al. Tissue array technology for testing interlaboratory and interobserver reproducibility of immunohistochemical estrogen receptor analysis in a large multicenter trial. Am J Clin Pathol 2002; 118: 675–682.
Annals of Oncology 25: 1428–1436, 2014 doi:10.1093/annonc/mdu154 Published online 25 April 2014
Patient perception of the benefit of a BRAF inhibitor in metastatic melanoma: quality-of-life analyses of the BREAK-3 study comparing dabrafenib with dacarbazine J.-J. Grob1*, M. M. Amonkar2, S. Martin-Algarra3, L. V. Demidov4, V. Goodman2, K. Grotzinger2, P. Haney2, E. Kämpgen5, B. Karaszewska6, C. Mauch7, W. H. Miller, Jr8, M. Millward 9, B. Mirakhur2, P. Rutkowski10, V. Chiarion-Sileni11, S. Swann2 & A. Hauschild12 1
Aix-Marseille University, APHM, Hôpital Timone, Marseille, France; 2GlaxoSmithKline, Collegeville, USA; 3Department of Medical Oncology, Clínica Universidad de Navarra, Pamplona, Spain; 4Department of Tumor Biotherapy, N.N. Blokhin Russian Cancer Research Center, Moscow, Russian Federation; 5Department of Dermatology, Skin Cancer Center, University Hospital Erlangen, Erlangen, Germany; 6Przychodnia Lekarska KOMED, Konin, Poland; 7Department for Dermatology and Venereology and CIO KölnBonn, University Hospital Cologne, Cologne, Germany; 8Departments of Oncology and Medicine, Lady Davis Institute and Segal Cancer Centre, Jewish General Hospital, McGill University, Montreal, Canada; 9Department of Medical Oncology, Sir Charles Gairdner Hospital and School of Medicine and Physiology, University of Western Australia, Perth, Australia; 10Department of Soft Tissue/Bone Sarcoma and Melanoma, Maria Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Warsaw, Poland; 11Melanoma Cancer Unit, Veneto Oncology Institute-IRCCS, Padova, Italy; 12Department of Dermatology, University Hospital Schleswig-Holstein, Kiel, Germany
Received 5 December 2013; revised 26 February 2014; accepted 6 April 2014
Background: In a randomized phase III study (BREAK-3), dabrafenib showed prolonged progression-free survival (PFS) (median 5.1 versus 2.7 months; hazard ratio = 0.30; 95% confidence interval 0.18–0.53; P < 0.0001) compared with dacarbazine (DTIC) in patients with BRAF V600E metastatic melanoma. Assessing how these results are transformed into a real health benefit for patients is crucial.
*Correspondence to: Prof. Jean-Jacques Grob, Aix-Marseille University, APHM, Hôpital Timone, Bouches-du-Rhône, 264 Rue St Pierre, Marseille 13005, France. Tel: +33-49138-85-91; Fax: +33-491-38-79-89; Email: [email protected]
© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
1. Macias H, Hinck L. Mammary gland development. Wiley Interdiscip Rev Dev Biol 2012; 1: 533–557. 2. Eliassen AH, Tworoger SS, Hankinson SE. Reproductive factors and family history of breast cancer in relation to plasma prolactin levels in premenopausal and postmenopausal women. Int J Cancer 2007; 120: 1536–1541. 3. Greendale G, Huang MH, Ursin G et al. Serum prolactin levels are positively associated with mammographic density in postmenopausal women. Breast Cancer Res Treat 2007; 105: 337–346. 4. Tworoger SS, Eliassen AH, Zhang X et al. A 20-year prospective study of plasma prolactin as a risk marker of breast cancer development. Cancer Res 2013; 73: 4810–4819. 5. Bernichtein S, Touraine P, Goffin V. New concepts in prolactin biology. J Endocrinol 2010; 206: 1–11. 6. Wang DY, De Stavola BL, Bulbrook RD et al. Relationship of blood prolactin levels and the risk of subsequent breast cancer. Int J Epidemiol 1992; 21: 214–221. 7. Helzlsouer KJ, Alberg AJ, Bush TL et al. A prospective study of endogenous hormones and breast cancer. Cancer Detect Prev 1994; 18: 79–85. 8. Kabuto M, Akiba S, Stevens RG et al. A prospective study of estradiol and breast cancer in Japanese women. Cancer Epidemiol Biomarkers Prev 2000; 9: 575–579.
Annals of Oncology
Annals of Oncology Annals of Oncology 25: 1422–1428, 2014 doi:10.1093/annonc/mdu150 Published online 8 April 2014
Circulating prolactin and breast cancer risk among pre- and postmenopausal women in the EPIC cohort
1 Division of Cancer Epidemiology, German Cancer Research Center (DKFZ) Heidelberg, Heidelberg, Germany; 2Section of Nutrition and Metabolism, International Agency for Research on Cancer (IARC), Lyon, France; 3Danish Cancer Society Research Center, Copenhagen; 4Section for Epidemiology, Department of Public Health, Aarhus University, Aarhus, Denmark; 5INSERM, Centre for Research in Epidemiology and Population Health [CESP], Nutrition, Hormones and Women’s Health Team, Villejuif; 6 University of Paris Sud, UMRS, Villejuif; 7IGR, Villejuif, France; 8Cancer Epidemiology Centre, Cancer Council Victoria, Melbourne; 9Centre for Molecular, Environmental, Genetic, and Analytic Epidemiology, The University of Melbourne, Melbourne, Australia; 10Department of Epidemiology, German Institute of Human Nutrition (DIfE) Potsdam-Rehbrücke, Nuthetal, Germany; 11Department of Hygiene, Epidemiology and Medical Statistics, University of Athens Medical School, Athens; 12Hellenic Health Foundation, Athens, Greece; 13Department of Epidemiology, Harvard School of Public Health, Boston, USA; 14Bureau of Epidemiologic Research, Academy of Athens, Athens, Greece; 15Molecular and Nutritional Epidemiology Unit, Cancer Research and Prevention Institute—ISPO, Florence, Italy; 16Epidemiology and Prevention Unit, National Tumor Institute (IRCCS), Milano; 17Cancer Registry and Histopathology Unit, ‘Civic-M. P. Arezzo’ Hospital ASP, Ragusa; 18Piedmont Cancer Registry, Centre for Epidemiology and Prevention in Oncology in Piedmont, Torino; 19Department of Clinical and Experimental Medicine, Federico II University, Naples, Italy; 20Unit of Nutrition, Environment and Cancer, Catalan Institute of Oncology-ICO, IDIBELL, Barcelona; 21Public Health Directorate, Asturias; 22Granada Cancer Registry, Andalusian School of Public Health, Granada; 23Consortium for Biomedical Research in Epidemiology and Public Health (CIBER), Madrid; 24Public Health Division of Gipuzkoa, Basque Regional Health Department, San Sebastian; 25Department of Epidemiology, Murcia Regional Health Authority, Murcia; 26Navarre Public Health Institute, Pamplona, Spain; 27 National Institute for Public Health and the Environment (RIVM), Bilthoven; 28Department of Gastroenterology and Hepatology, University Medical Centre, Utrecht, The Netherlands; 29School of Public Health, Imperial College, London, UK; 30Department of Epidemiology, Julius Center for Health Sciences and Primary Care, University Medical Center, Utrecht, The Netherlands; 31Department of Radiation Sciences, University of Umeå, Umeå; 32Department of Surgical and Perioperative Sciences, Umeå University, Umeå, Sweden; 33Department of Community Medicine, Faculty of Health Sciences, University of Tromsø, Tromsø; 34Department of Research, Cancer Registry of Norway, Oslo, Norway; 35Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden; 36Society of Public Health, Helsinki, Finland; 37 School of Clinical Medicine, University of Cambridge, Cambridge; 38Cancer Epidemiology Unit, University of Oxford, Oxford, UK
Received 24 January 2014; revised 26 February 2014; accepted 26 March 2014
Background: Experimental and epidemiological evidence suggests that prolactin might play a role in the etiology of breast cancer. We analyzed the relationship of prediagnostic circulating prolactin levels with the risk of breast cancer by menopausal status, use of postmenopausal hormone replacement therapy (HRT) at blood donation, and by estrogen and progesterone receptor status of the breast tumors. Patients and methods: Conditional logistic regression was used to analyze the data from a case–control study nested within the prospective European EPIC cohort, including 2250 invasive breast cancer and their matched control subjects. Results: Statistically significant heterogeneity in the association of prolactin levels with breast cancer risk between women who were either pre- or postmenopausal at the time of blood donation was observed (Phet = 0.04). Higher serum levels of prolactin were associated with significant increase in the risk of breast cancer among postmenopausal women [odds ratio (OR)Q4–Q1 = 1.29 (95% confidence interval, CI, 1.05–1.58), Ptrend = 0.09]; however, this increase in risk seemed to be confined to women who used postmenopausal HRT at blood donation [ORQ4–Q1 = 1.45 (95% CI 1.08– 1.95), Ptrend = 0.01], whereas no statistically significant association was found for the non-users of HRT [ORQ4–Q1 = 1.11 (95%CI 0.83–1.49), Ptrend = 0.80] (Phet = 0.08). Among premenopausal women, a statistically non-significant inverse
*Correspondence to: Prof. Rudolf Kaaks, Department of Cancer Epidemiology, German Cancer Research Center (DKFZ) Heidelberg, Im Neuenheimer Feld 581, D-69120 Heidelberg, Germany. Tel: +49-6221-42-22-19; Fax: +49-6221-42-22-03; E-mail: [email protected]
© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
K. Tikk1, D. Sookthai1, T. Johnson1, S. Rinaldi2, I. Romieu2, A. Tjønneland3, A. Olsen3, K. Overvad4, F. Clavel-Chapelon5,6,7, L. Baglietto8,9, H. Boeing10, A. Trichopoulou11,12, P. Lagiou11,13,14, D. Trichopoulos12,13,14, D. Palli15, V. Pala16, R. Tumino17, S. Rosso18, S. Panico19, A. Agudo20, V. Menéndez21, M.-J. Sánchez22,23, P. Amiano23,24, J. M. Huerta Castaño23,25, E. Ardanaz23,26, H. B. Bueno-de-Mesquita27,28,29, E. Monninkhof30, C. Onland-Moret30, A. Andersson31, M. Sund32, E. Weiderpass33,34,35,36, K.-T. Khaw37, T. J. Key38, R. C. Travis38, M. J. Gunter29, E. Riboli29, L. Dossus5,6,7 & R. Kaaks1*
Annals of Oncology
original articles
association was observed [ORQ4–Q1 = 0.70 (95% CI 0.48–1.03), Ptrend = 0.16]. There was no heterogeneity in the prolactin–breast cancer association by hormone receptor status of the tumor. Conclusion: Our study indicates that higher circulating levels of prolactin among the postmenopausal HRT users at baseline may be associated with increased breast cancer risk. Key words: breast cancer, prolactin levels, hormone replacement therapy, estrogen receptor, progesterone receptor, prospective cohort
introduction
study population and methods the EPIC cohort The EPIC cohort is based on 366 521 women and 153 457 men recruited between the years 1992 and 2000 in 10 European countries: Denmark, France, Germany, Greece, Italy, Norway, Spain, Sweden, The Netherlands, and the UK. Details on the subject recruitment, baseline data, and blood collection protocols have been reported previously [11]. In brief, anthropometric measurements and questionnaire data on diet, health status,
Volume 25 | No. 7 | July 2014
design of the nested case–control study General criteria for the inclusion of cases and matched controls into the study were that women could be clearly classified as being either pre- or postmenopausal at the time of blood donation, and did not have any previous diagnosis of cancer (except non-melanoma skin cancer). Women who were ≤42 years of age or reported having had at least nine menstrual periods in the last 12 months were classified as premenopausal. Women were classified as postmenopausal when they reported not having any menses over the past 12 months, when older than 55 years of age, or when reporting bilateral ovariectomy. Women aged 42 years or older whose questionnaire data on menopausal status were incomplete were not included to the present case– control study. For each case-subject, one control subject closest to the case (based on matching criteria) with an available blood sample was chosen among appropriate risk sets consisting of all cohort members alive and free of cancer at the time of diagnosis of the index case. Matching criteria were the study recruitment center, menopausal status at blood donation, age at blood donation (±6 months), time at blood donation (±2 h), fasting status, phase of menstrual cycle [matching categories for premenopausal women: early follicular (days 0–7 of the cycle), late follicular (days 8–11), mid-cycle (days 12–16), early luteal (days 17–19), mid-luteal (days 20–24), late luteal (days 25+)], and current use of HRT at the time of blood donation (for postmenopausal women). In total, 2271 invasive breast cancer case subjects and an equal number of control subjects were selected for the study. After exclusion of case or control subjects with missing prolactin
doi:10.1093/annonc/mdu150 |
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
The polypeptide hormone prolactin promotes the proliferation of breast epithelial cells, and in combination with progesterone promotes the differentiation of specialized structures—alveoli—that synthesize and secrete milk during lactation [1]. Associations of several breast cancer risk factors, such as nulli-parity and high mammographic breast density with increased serum prolactin concentrations, [2, 3] in combination with animal experimental data led to the hypothesis that elevated prolactin might promote breast tumor development [4, 5]. One large and four very small prospective studies have previously addressed possible associations between circulating prolactin levels and breast cancer risk [4, 6–9]. In the Nurses’ Health Study (NHS), the only large-scale prospective study to date, a series of nested case–control studies indicated a modest positive risk association between serum prolactin concentrations and breast cancer risk among postmenopausal women and especially with the risk of estrogen receptor-positive (ER+) disease, but showed no significant associations among premenopausal women or with ER-negative breast cancer [4, 10]. The other four studies, which each included only small numbers of incident breast cancer cases (n = 20–173), showed no significant relationship between serum prolactin and breast cancer risk among either pre- or postmenopausal women [6–9] and did not provide any information with regard to hormone receptor status of the tumors. In this large case–control study nested within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort, we examined the association of prediagnostic serum levels of prolactin among pre- and postmenopausal women with subsequent risk of breast cancer, overall, by hormone receptor status of the tumor and by the use of postmenopausal hormone replacement therapy (HRT) at the time of blood donation. With a total of 2250 incident breast cancer cases, including 817 postmenopausal case subjects reporting HRT use at baseline, this is the second major prospective study to date on the relationships between prediagnostic prolactin levels and subsequent breast cancer risk.
reproductive and menstrual history, and medication use was collected at baseline. All EPIC study participants provided written consent for the use of questionnaire information and blood samples in research studies. Ethical approval for the nested case–control study reported here was obtained from the ethical review boards of International Agency for Research on Cancer (IARC) (Lyon, France) and from the local ethics committees in participating countries/study centers. Details on the identification of vital status, breast cancer incidence, and information/quantification descriptions for the hormone receptor status of the tumors across the study centers were reported previously [11, 12]. For the present study, the closure date for follow-up was the last date of complete followup for both cancer incidence and vital status, which ranged from 2003 to 2006, depending on the study center. Sweden was not included in the analysis because independent studies were being completed on breast cancer risk and endogenous hormones [9].
original articles
Annals of Oncology
measurements (n = 17) and with the prolactin value above detection range of the assay (more than 133 ng/ml, n = 4), 2250 incident cases and an equal number of controls remained for the analyses.
laboratory assays The prolactin analyses were performed in the laboratory of the Division of Cancer Epidemiology at the German Cancer Research Centre and determined by immunoradiometric assay [IRMA (CT), IBL International GMBH, Germany] from blood taken at baseline. The detection range of the assay was 0.35–133 ng/ml. The same quality controls were included in each analytical batch. The mean inter- and intra-assay coefficients of variation were 4.62% and 2.17%, respectively.
Conditional logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) for breast cancer occurrence, calculating ORs over the quartile levels and on a continuous log2 scale of circulating prolactin. As circulating prolactin levels show variation over the menopausal status, the quartile cutoff points were calculated separately based on the control population of pre- and postmenopausal women. We a priori decided to adjust all the models for number of full-term pregnancies (0, 1, 2, 3, ≥4, missing), smoking status (current, never, previous, missing), and body mass index (continuous scale) due to modest variation in prolactin levels over these factors in healthy women. A more detailed description of the statistical analysis used in this study is described in supplementary Material, available at Annals of Oncology online. All P-values presented are two-sided and P < 0.05 was considered statistically significant. Statistical analyses were conducted using SAS software, version 9.2 (SAS Institute, Cary, NC).
results This nested case–control study consisted of 2250 case subjects with incident, invasive breast cancer matched to an equal number of control subjects. A total of 1738 case–control pairs were sampled among women who could be unambiguously classified
Table 1. Baseline characteristics of case and control subjects [data presented as median (min, max) or n (%)]
Age at blood donation (years) Age at menopause (years) Age at diagnosis (years) Lag time till diagnosis (years) Body mass index (kg/m2) Ever had a full-term pregnancy Baseline smoking Baseline HRTb use a
Premenopausal womena Cases (512) 46.2 (26.7–56.9) — 50.9 (35.2–64.2) 5.2 (0.04–11.7) 23.6 (16.5–49.1) 438 (86%) 109 (21%) –
Controls (512)
Postmenopausal womena Cases (1738)
Controls (1738)
46.1 (26.7–57.1) — — — 23.9 (16.8–46.3) 442 (88%) 111 (22%) –
58.3 (41.1–75.5) 50 (15–63) 63.0 (45.3–83.6) 4.6 (0.01–12.0) 24.7 (13.8–47.7) 1453 (85%) 325 (19%) 817 (47%)
58.3 (41.2–76.8) 50 (21–63) — — 24.5 (16.0–45.3) 1497 (88%) 324 (19%) 817 (47%)
Menopausal status at the time of blood donation. Hormone replacement therapy.
b
| Tikk et al.
Volume 25 | No. 7 | July 2014
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
statistical analyses
as being postmenopausal at blood donation, and 512 pairs were sampled among premenopausal women. For the premenopausal case–control pairs, the median age at blood donation (baseline) was 46.2 (range: 26.7–57.1) years, and the median age at tumor diagnosis for the case subjects was 50.9 (range: 35.2–64.2) years; for the postmenopausal women, the ages at blood donation and at breast cancer diagnosis were 58.3 (range: 41.1–76.8) and 63.0 (range: 45.3–83.6) years, respectively (Table 1). A total of 817 postmenopausal case–control pairs were using HRT at the time they provided their blood sample. In linear regression analyses adjusted for matching factors, the number of full-term pregnancies, smoking status, and body mass index, our data showed significantly higher prolactin levels among premenopausal women when compared with postmenopausal women [adjusted geometric mean levels 7.87 (95% CI 7.4–8.3) versus 6.54 (95% CI 6.3–6.8) ng/ml], and among postmenopausal HRT users (especially among combined estrogenplus-progestin users) compared with non-users at baseline [adjusted geometric mean levels 6.72 (95% CI 6.4–7.1) versus 5.94 (95% CI 5.7–6.2) ng/ml]. In case–control comparisons, prolactin levels differed on average ∼3% between case and control subjects, with the greatest difference among postmenopausal women who used HRT at baseline (5.9%) (data not shown). With regard to our risk models (logistic regression), as statistical tests indicated significant heterogeneity in the strength of the association between breast cancer and circulating prolactin levels by menopausal status at baseline (Phet = 0.04), all statistical analyses were conducted separately for pre- and postmenopausal women. Among the premenopausal women, conditional logistic regression analyses adjusting for smoking status, number of full-term pregnancies, and body mass index suggest a negative association of serum prolactin levels with overall breast cancer risk, which was not statistically significant [ORQ4–Q1 = 0.70 (95% CI 0.48–1.03); Ptrend = 0.16] (Table 2). Among the postmenopausal women, there was a positive association of prolactin levels with breast cancer risk overall, with a statistically significant OR of 1.29 (95% CI 1.05–1.58) in the highest quartile of prolactin, but no statistically significant trend based on prolactin as a continuous log2 variable (Ptrend = 0.09). When the analyses were stratified further by baseline HRT use, statistically
original articles
Annals of Oncology
Table 2. Adjusted odds ratiosa [OR (95% CI)] for overall breast cancer by quartile levels and on a continuous log2 scale of circulating prolactin Quartiles 1
2
Premenopausal women ca/co 145/128 128/128 OR Ref. 0.83 (0.57, 1.22) Postmenopausal women (all) ca/co 383/435 418/437 OR Ref. 1.09 (0.90, 1.33) Postmenopausal non-HRT users ca/co 250/272 250/245 OR Ref. 1.09 (0.84, 1.41) Postmenopausal HRT users ca/co 133/163 168/192 OR Ref. 1.09 (0.80, 1.50)
Log2 OR (95% CI)
Ptrend b
3
4
128/128 0.83 (0.57, 1.22)
111/128 0.70 (0.48, 1.03)
0.87 (0.71, 1.06)
0.16
455/432 1.22 (0.99, 1.49)
482/434 1.29 (1.05, 1.58)
1.10 (0.98, 1.22)
0.09
239/229 1.12 (0.85, 1.48)
182/175 1.11 (0.83, 1.49)
0.98 (0.84, 1.15)
0.80
216/203 1.34 (0.98, 1.84)
300/259 1.45 (1.08, 1.95)
1.21 (1.04, 1.41)
0.01
Phet c 0.04
0.08
Adjusted for full-term pregnancies, smoking status, and body mass index. Linear trends for OR estimates over a continuous scale of prolactin levels. c Statistical tests for heterogeneity were based on likelihood-ratio test, comparing the model fit for logistic regression models with and without corresponding interaction term. b
significant ORs were observed among the HRT users [ORQ4– Q1 = 1.45 (95% CI 1.08–1.95), on a continuous log2 scale ORlog2 = 1.21 (95% 1.04–1.41), Ptrend = 0.01], but not among the non-users [ORQ4–Q1 = 1.11 (95% CI 0.83–1.49), Ptrend = 0.80], although the test for heterogeneity between HRT users and nonusers was only of borderline statistical significance (based on continuous log2 scale, Phet = 0.08). Stratifying the analyses by hormone receptor status of the tumor, there was no evidence of heterogeneity in the prolactin–breast cancer relationship by tumor subtype, overall or in strata of users or non-users of HRT (Table 3). Among postmenopausal women using combined estrogen-plus-progestin HRT, the relative risk appeared to be more pronounced for ER + tumors [ORlog2 = 1.27 (95% CI 1.03–1.57), Ptrend = 0.03, on a continuous log2 scale, n = 338/425] than among users of estrogens only therapy [ORlog2 = 1.06 (95% CI 0.72–1.58), Ptrend = 0.77, n = 103/206], although the test for statistical heterogeneity was not statistically significant (Phet = 0.44) (data not shown). In subgroup analyses, statistical tests showed no evidence of heterogeneity in the strength of the association with overall breast cancer risk by lag time until breast cancer diagnosis (supplementary Table S1, available at Annals of Oncology online), and overall, there was also no significant heterogeneity by age at tumor diagnosis in either pre- or postmenopausal women (data not shown). Among the postmenopausal women using HRT at baseline, however, circulating levels of prolactin were significantly associated with breast cancer diagnosed above the median age (61 years) within this subgroup [ORlog2 = 1.46 (95% CI 1.18–1.80), Ptrend = 0.0004, n = 370 sets], but not with breast cancer diagnosed at a younger age [ORlog2 = 0.95 (95% CI 0.76–1.19), Ptrend = 0.65, n = 447 sets] (Phet = 0.01) (data not shown). Complementary statistical analyses showed no evidence for major confounding or interaction effects by other reproductive
Volume 25 | No. 7 | July 2014
and lifestyle factors, or by levels of the other endogenous hormones (data not shown). Furthermore, sensitivity analyses in which the first 2 years of follow-up were excluded did not show any major difference in risk estimates compared with the full dataset.
discussion In this large prospective study, we observed a statistically significant heterogeneity in the association of prolactin levels with breast cancer risk between women who were either pre- or postmenopausal at the time of blood donation. Among postmenopausal women, there was a significant positive association of prolactin levels with overall breast cancer risk; however, this increase in risk seemed to be confined to women who used HRT at the time of blood donation. Among premenopausal women, there was a statistically non-significant inverse association of prolactin with breast cancer risk. Our data showed no evidence for heterogeneity in the relationship of prolactin levels with risk of hormone receptor positive or negative breast tumors. Our findings show partial agreement with those from the Nurses’ Health Study cohorts I and II, in which a series of nested case–control studies were conducted to address the relationship of serum prolactin levels with breast tumor development [4, 10] that included up to 3421 breast cancer cases [4]. In the Nurses’ studies, using data based on only the first 10 years of prospective follow-up (comparable to the follow-up time our present study) with 2468 breast cancer (of whom 1023 among pre- and 1445 among postmenopausal women) and 4021 control subjects, there was a positive association between serum prolactin and overall breast cancer risk among postmenopausal women, and no association among women who were premenopausal at the time of blood draw (Phet = 0.01). Our observations are similar to these findings. However, contrary to our doi:10.1093/annonc/mdu150 |
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
a
original articles
Annals of Oncology
Table 3. Adjusted odds ratiosa [OR (95% CI)] for breast cancer by receptor status of the tumor by quartile levels and on a continuous log2 scale of circulating prolactin 1
2
3
Log2 OR
Ptrend b
88/101 0.70 (0.44, 1.09)
0.85 (0.68, 1.07)
0.18
23/25 0.69 (0.30, 1.61)
0.96 (0.65, 1.42)
0.84
56/73 0.56 (0.32, 0.99)
0.80 (0.61, 1.05)
0.11
18/14 1.02 (0.31, 3.33)
1.31 (0.78, 2.19)
0.30
384/358 1.29 (1.02, 1.63)
1.10 (0.97, 1.24)
0.13
1.09 (0.85, 1.40)
0.48
1.23 (0.88, 1.70) 71/55
1.07 (0.90, 1.27)
0.42
1.14 (0.66, 1.95)
1.00 (0.75, 1.32)
0.98
137/139 1.12 (0.80, 1.57)
0.97 (0.81, 1.17)
0.76
0.99 (0.71, 1.39)
0.96
1.11 (0.69, 1.78) 35/26
0.97 (0.75, 1.24)
0.78
0.94 (0.43, 2.05)
0.97 (0.66, 1.41)
0.86
247/219 1.42 (1.01, 1.99)
1.21 (1.03, 1.43)
0.02
1.18 (0.81, 1.73)
0.38
1.22 (0.75, 2.00) 36/29
1.15 (0.90, 1.47)
0.27
1.52 (0.65, 3.54)
1.10 (0.67, 1.78)
0.71
96/74 1.36 (0.87, 2.14) 210/196
44/34 1.14 (0.59, 2.21) 73/76
52/40 1.67 (0.86, 3.24) 137/120
Phet c
0.77
0.24
0.96
0.76
0.83
0.92
0.96
0.58
a
Adjusted for full-term pregnancies, smoking status, and body mass index. Linear trends for OR estimates over a continuous scale of prolactin levels. c Statistical tests for heterogeneity were based on likelihood-ratio test, comparing the model fit for logistic regression models with and without corresponding interaction term. b
| Tikk et al.
Volume 25 | No. 7 | July 2014
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
Premenopausal women ER+ ca/co 114/102 91/87 92/95 OR Ref. 0.91 (0.57, 1.43) 0.82 (0.53, 1.27) ER− ca/co 30/26 34/40 36/32 OR Ref. 0.57 (0.25, 1.27) 0.80 (0.34, 1.88) ER+PR+ ca/co 88/76 62/62 68/63 OR Ref. 0.78 (0.45, 1.35) 0.83 (0.49, 1.42) ER−PR− ca/co 19/16 21/25 19/22 OR Ref. 0.69 (0.24, 1.94) 0.71 (0.22, 2.23) All postmenopausal women combined ER+ ca/co 302/349 348/344 373/356 OR Ref. 1.20 (0.96, 1.50) 1.25 (0.99, 1.57) ER− ca/co 80/85 68/92 78/71 OR Ref. 0.77 (0.5, 1.19) 1.13 (0.72, 1.79) ca/co 155/171 191/174 191/206 ER+PR+ OR Ref. 1.22 (0.89, 1.69) 1.03 (0.74, 1.44) ca/co 50/45 49/66 47/51 ER−PR− OR Ref. 0.65 (0.37, 1.14) 0.77 (0.43, 1.39) Postmenopausal women, not using HRT at the time of blood donation ER+ ca/co 198/223 206/190 193/182 OR Ref. 1.24 (0.92, 1.66) 1.21 (0.88, 1.66) ER− ca/co 51/48 42/54 42/43 OR Ref. 0.74 (0.42, 1.29) 0.84 (0.45, 1.57) ca/co 102/114 117/99 97/100 ER+PR+ OR Ref. 1.32 (0.87, 2.01) 1.09 (0.70, 1.70) ca/co 31/24 28/40 27/31 ER−PR− OR Ref. 0.51 (0.25, 1.07) 0.54 (0.24, 1.25) Postmenopausal women, using HRT at the time of blood donation ER+ ca/co 104/126 142/154 180/174 OR Ref. 1.14 (0.80, 1.63) 1.29 (0.91, 1.83) ER− ca/co 29/37 26/38 36/28 OR Ref. 0.87 (0.41, 1.88) 1.69 (0.82, 3.48) ca/co 53/57 74/75 94/106 ER+PR+ OR Ref. 1.10 (0.65, 1.84) 0.93 (0.56, 1.55) ca/co 19/21 21/26 20/20 ER−PR− OR Ref. 0.82 (0.29, 2.36) 1.23 (0.49, 3.09)
4
original articles
Annals of Oncology
Volume 25 | No. 7 | July 2014
by receptor status has been shown to be relatively robust and to have only moderate dependence on the assay used, especially for medium- and high-expressing tumors [17], staining for receptor status was performed in different laboratories both within and across EPIC centers, and this may have led to some misclassification by receptor status. Finally, in spite of the large study size (a total of 2250 matched case–control pairs), our study still had relatively limited numbers of case subjects for hormone receptor negative tumors within strata of menopausal status and postmenopausal HRT use, and thus the statistical power for detecting associations of prolactin with risk in these subgroups was limited.
conclusions Our study shows that higher circulating levels of prolactin among the postmenopausal women who used HRT at baseline are associated with increased breast cancer risk. In contrast, statistically non-significant inverse association of premenopausal prolactin levels and subsequent breast cancer risk was seen. Our data did not show any heterogeneity in the prolactin-breast cancer association by hormone receptor status of the tumor among either pre- or postmenopausal women.
acknowledgements We thank all the EPIC cohort participants. Furthermore, we thank Britta Lederer and Sigrid Henke for their work in conducting the immunoassays; Sabine Rohrmann, Jutta Schmitt, and Jutta Kneisel for their assistance during the collection of hormone receptor status data. The coordination of EPIC is financially supported by the European Commission (DG-SANCO) and the International Agency for Research on Cancer. The national cohorts are supported by Danish Cancer Society (Denmark); Ligue contre le Cancer, Mutuelle Générale de l’Éducation Nationale, Institut National de la Santé et de la Recherche Médicale (France); Deutsche Krebshilfe, Deutsches Krebsforschungszentrum and Federal Ministry of Education and Research (Germany); the Hellenic Health Foundation (Greece); Italian Association for Research on Cancer (AIRC) and National Research Council (Italy); Dutch Ministry of Public Health, Welfare and Sports (VWS), Netherlands Cancer Registry (NKR), LK Research Funds, Dutch Prevention Funds, Dutch ZON (Zorg Onderzoek Nederland), World Cancer Research Fund (WCRF), Statistics Netherlands (The Netherlands); Norwegian Research Council, Norwegian Cancer Society, ERC-2009-AdG 232997 and Nordforsk, Nordic Centre of Excellence programme on Food, Nutrition and Health (Norway); Health Research Fund (FIS), The Spanish Ministry of Health (ISCIII RETICC RD06/0020/0091) and the Catalan Institute of Oncology, Regional Governments of Andalucía, Asturias, Basque Country, Murcia (no 6236) and Navarra, ISCIII RETIC (RD06/ 0020; Spain); Swedish Cancer Society, Swedish Scientific Council and Regional Government of Skåne and Västerbotten (Sweden); Cancer Research UK, Medical Research Council (UK).
funding This research did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.
doi:10.1093/annonc/mdu150 |
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
observations, in the Nurses’ cohorts, the direct association among postmenopausal women appeared to be clearly stronger for ER/PR-positive tumors [RRQ4–Q1 = 1.53 (95% CI 1.17–1.98), Ptrend = 0.0002] and the absence of an association in ER/PRnegative tumors was reported [RRQ4–Q1 = 0.88 (95% CI 0.52– 1.47), Ptrend = 0.94] (Phet = 0.04) [4]. In our data, there was no evidence for heterogeneity of the prolactin–breast cancer association by receptor status. In several other, much smaller prospective studies, the heterogeneity of relationships of prolactin with breast cancer risk by receptor status was not addressed [6–9]. The observed associations between circulating prolactin and subsequent breast cancer risk, both in our study and in the Nurses’ cohorts, were of a modest magnitude compared with the associations reported for other endogenous hormones, such as testosterone and estradiol, with relative risk estimates up to 2.5 for top versus bottom tertile levels [12]. Nonetheless, a vast body of evidence from animal and in vitro studies supports the involvement of prolactin in breast cancer development. Especially in combination with the ovarian steroid hormones estradiol and progesterone, prolactin has effects on normal epithelial cell expansion, ductal side branching of the breast during puberty, and formation of lobuloalveolar structures during pregnancy [1]. Furthermore, prolactin and progesterone have synergic roles in inducing cell growth and proliferative activity in the terminal duct-lobular units of the breast—the site of origin of most breast cancers—and evidence is accumulating for crosstalk between progesterone, prolactin, and their receptor signaling pathways in normal and malignant breast cells [13–15]. We observed a more pronounced association between circulating prolactin and breast cancer among postmenopausal women who were using HRT at the time of blood donation, compared with the association among HRT non-users. Given the well-documented biological synergy between prolactin and progesterone, it is striking that our data indicated a significant association of prolactin with the risk of receptor-positive tumors among women using combined estrogen-plus-progestin HRT, but not among the users of estrogen-only. However, only a reduced number of women in our study had detailed information about the types of HRT used, and the statistical test for heterogeneity between estrogen-plus-progestin users and estrogen-only users was not statistically significant. Furthermore, in the Nurses’ cohorts associations of prolactin with breast cancer risk were reported to be relatively independent of exogenous hormone use at the time of blood donation (Phet = 0.65) [4, 10]. At the time the NHS was conducted, most women were on estrogen only HRT and relatively few were on combined estrogenplus-progestin HRT ( personal communication), which may explain why NHS did not see a difference by HRT use. Major strengths of our study are its prospective design and the large number of incident cases with the information on hormone receptor status of the tumor, which allowed subgroup analyses describing risk associations by menopausal status, postmenopausal HRT use, and receptor status of the tumor. A limitation of our study is that for each woman we had prolactin measurements for only one blood sample, which may limit interpretations about women’s average prolactin levels over time periods longer than 5–10 years [4, 16]. Furthermore, although classification of breast cancer outcomes
original articles disclosure The authors have declared no conflicts of interest.
references
9. Manjer J, Johansson R, Berglund G et al. Postmenopausal breast cancer risk in relation to sex steroid hormones, prolactin and SHBG (Sweden). Cancer Causes Control 2003; 14: 599–607. 10. Tworoger SS, Eliassen AH, Rosner B et al. Plasma prolactin concentrations and risk of postmenopausal breast cancer. Cancer Res 2004; 64: 6814–6819. 11. Riboli E, Hunt KJ, Slimani N et al. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr 2002; 5: 1113–1124. 12. James RE, Lukanova A, Dossus L et al. Postmenopausal serum sex steroids and risk of hormone receptor-positive and -negative breast cancer: a nested case− control study. Cancer Prev Res 2011; 4: 1626–1635. 13. Ormandy CJ, Hall RE, Manning DL et al. Coexpression and cross-regulation of the prolactin receptor and sex steroid hormone receptors in breast cancer. J Clin Endocrinol Metab 1997; 82: 3692–3699. 14. Lee HJ, Ormandy CJ. Interplay between progesterone and prolactin in mammary development and implications for breast cancer. Mol Cell Endocrinol 2012; 357: 101–107. 15. Perks CM, Keith AJ, Goodhew KL et al. Prolactin acts as a potent survival factor for human breast cancer cell lines. Br J Cancer 2004; 91: 305–311. 16. Arslan AA, Gu Y, Zeleniuch-Jacquotte A et al. Reproducibility of serum pituitary hormones in women. Cancer Epidemiol Biomarkers Prev 2008; 17: 1880–1883. 17. von Wasielewski R, Mengel M, Wiese B et al. Tissue array technology for testing interlaboratory and interobserver reproducibility of immunohistochemical estrogen receptor analysis in a large multicenter trial. Am J Clin Pathol 2002; 118: 675–682.
Annals of Oncology 25: 1428–1436, 2014 doi:10.1093/annonc/mdu154 Published online 25 April 2014
Patient perception of the benefit of a BRAF inhibitor in metastatic melanoma: quality-of-life analyses of the BREAK-3 study comparing dabrafenib with dacarbazine J.-J. Grob1*, M. M. Amonkar2, S. Martin-Algarra3, L. V. Demidov4, V. Goodman2, K. Grotzinger2, P. Haney2, E. Kämpgen5, B. Karaszewska6, C. Mauch7, W. H. Miller, Jr8, M. Millward 9, B. Mirakhur2, P. Rutkowski10, V. Chiarion-Sileni11, S. Swann2 & A. Hauschild12 1
Aix-Marseille University, APHM, Hôpital Timone, Marseille, France; 2GlaxoSmithKline, Collegeville, USA; 3Department of Medical Oncology, Clínica Universidad de Navarra, Pamplona, Spain; 4Department of Tumor Biotherapy, N.N. Blokhin Russian Cancer Research Center, Moscow, Russian Federation; 5Department of Dermatology, Skin Cancer Center, University Hospital Erlangen, Erlangen, Germany; 6Przychodnia Lekarska KOMED, Konin, Poland; 7Department for Dermatology and Venereology and CIO KölnBonn, University Hospital Cologne, Cologne, Germany; 8Departments of Oncology and Medicine, Lady Davis Institute and Segal Cancer Centre, Jewish General Hospital, McGill University, Montreal, Canada; 9Department of Medical Oncology, Sir Charles Gairdner Hospital and School of Medicine and Physiology, University of Western Australia, Perth, Australia; 10Department of Soft Tissue/Bone Sarcoma and Melanoma, Maria Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Warsaw, Poland; 11Melanoma Cancer Unit, Veneto Oncology Institute-IRCCS, Padova, Italy; 12Department of Dermatology, University Hospital Schleswig-Holstein, Kiel, Germany
Received 5 December 2013; revised 26 February 2014; accepted 6 April 2014
Background: In a randomized phase III study (BREAK-3), dabrafenib showed prolonged progression-free survival (PFS) (median 5.1 versus 2.7 months; hazard ratio = 0.30; 95% confidence interval 0.18–0.53; P < 0.0001) compared with dacarbazine (DTIC) in patients with BRAF V600E metastatic melanoma. Assessing how these results are transformed into a real health benefit for patients is crucial.
*Correspondence to: Prof. Jean-Jacques Grob, Aix-Marseille University, APHM, Hôpital Timone, Bouches-du-Rhône, 264 Rue St Pierre, Marseille 13005, France. Tel: +33-49138-85-91; Fax: +33-491-38-79-89; Email: [email protected]
© The Author 2014. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Brown University on October 27, 2014
1. Macias H, Hinck L. Mammary gland development. Wiley Interdiscip Rev Dev Biol 2012; 1: 533–557. 2. Eliassen AH, Tworoger SS, Hankinson SE. Reproductive factors and family history of breast cancer in relation to plasma prolactin levels in premenopausal and postmenopausal women. Int J Cancer 2007; 120: 1536–1541. 3. Greendale G, Huang MH, Ursin G et al. Serum prolactin levels are positively associated with mammographic density in postmenopausal women. Breast Cancer Res Treat 2007; 105: 337–346. 4. Tworoger SS, Eliassen AH, Zhang X et al. A 20-year prospective study of plasma prolactin as a risk marker of breast cancer development. Cancer Res 2013; 73: 4810–4819. 5. Bernichtein S, Touraine P, Goffin V. New concepts in prolactin biology. J Endocrinol 2010; 206: 1–11. 6. Wang DY, De Stavola BL, Bulbrook RD et al. Relationship of blood prolactin levels and the risk of subsequent breast cancer. Int J Epidemiol 1992; 21: 214–221. 7. Helzlsouer KJ, Alberg AJ, Bush TL et al. A prospective study of endogenous hormones and breast cancer. Cancer Detect Prev 1994; 18: 79–85. 8. Kabuto M, Akiba S, Stevens RG et al. A prospective study of estradiol and breast cancer in Japanese women. Cancer Epidemiol Biomarkers Prev 2000; 9: 575–579.
Annals of Oncology
