IJC International Journal of Cancer
Anthropometric factors and risk of molecular breast cancer subtypes among postmenopausal Norwegian women Julie Horn1,2, Mirjam D.K. Alsaker1,3, Signe Opdahl1, Monica J. Engstrïm4, Steinar Tretli1,5, Olav A. Haugen4, Anna M. Bofin4, Lars J. Vatten1 and Bjïrn Olav A˚svold1,6 1
Department of Public Health, Norwegian University of Science and Technology, Trondheim, Norway Department of Gynecology and Obstetrics, Levanger Hospital, Nord-Trïndelag Hospital Trust, Norway 3 Department of Oncology, St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway 4 Department of Laboratory Medicine, Children’s and Women’s Health, Norwegian University of Science and Technology, Trondheim, Norway 5 Cancer Registry of Norway, Institute of Population-Based Cancer Research, Oslo, Norway 6 Department of Endocrinology, St. Olavs Hospital, Trondheim University Hospital, Trondheim, Norway 2
Epidemiology
Adult height and body weight are positively associated with breast cancer risk after menopause, but few studies have investigated these factors according to molecular breast cancer subtype. A total of 18,562 postmenopausal Norwegian women who were born between 1886 and 1928 were followed up for breast cancer incidence from the time (between 1963 and 1975) height and weight were measured until 2008. Immunohistochemical and in situ hybridization techniques were used to subtype 734 incident breast cancer cases into Luminal A, Luminal B [human epidermal growth factor receptor 2 (HER22)], Luminal B (HER21), HER2 subtype, basal-like phenotype (BP) and five-negative phenotype (5NP). We used Cox regression analysis to assess adult height and body mass index (BMI) in relation to risk of these subtypes. We found a positive association of height with risk of Luminal A breast cancer (ptrend, 0.004), but there was no clear association of height with any other subtype. BMI was positively associated with risk of all luminal breast cancer subtypes, including Luminal A (ptrend, 0.002), Luminal B (HER22) (ptrend, 0.02), Luminal B (HER21) (ptrend, 0.06), and also for the HER2 subtype (ptrend, 0.04), but BMI was not associated with risk of the BP or 5NP subtypes. Nonetheless, statistical tests for heterogeneity did not provide evidence that associations of height and BMI differed across breast cancer subtypes. This study of breast cancer risk among postmenopausal women suggests that height is positively associated with risk of Luminal A breast cancer. BMI is positively associated with risk of all luminal subtypes and for the HER2 subtype.
It is well established that adult height and body weight are positively associated with risk of breast cancer in postmenopausal women.1–4 A few prospective studies have examined associations of adult height and body weight by hormone receptor status, and the results suggest that body height and weight are positively associated with risk of estrogen receptor (ER) or progesterone receptor (PR) positive tumors but Key words: height, BMI, breast cancer, molecular subtype Abbreviations: BMI: body mass index; BP: basal-like phenotype; CI: confidence interval; CISH: chromogenic in situ hybridization; CK5: cytokeratin 5; EGFR: epidermal growth factor receptor; ER: estrogen receptor; HER2: human epidermal growth factor receptor 2; HR: hazard ratio; IQR: interquartile range; 5NP: five negative phenotype; PR: progesterone receptor; TMA: tissue microarray Grant sponsor: Norwegian Cancer Society DOI: 10.1002/ijc.28912 History: Received 28 Jan 2014; Accepted 3 Apr 2014; Online 18 Apr 2014 Correspondence to: Julie Horn, Department of Public Health, Faculty of Medicine, NTNU, Post Box 8905, MTFS, 7491 Trondheim, Norway, Tel.: 47-7359-7529, Fax: 47-7359-7577, E-mail:
[email protected] C 2014 UICC Int. J. Cancer: 135, 2678–2686 (2014) V
associations with hormone receptor negative breast cancer are less clear.5–8 Recently, breast cancer has been categorized into molecularly defined subtypes with different clinical characteristics. The prognosis seems to differ according to subtype,9–11 and it has been suggested that the underlying etiology may also differ.12,13 Thus, tumors may be divided into three luminal (ER and/or PR positive) and three nonluminal (ER and PR negative) subtypes, based on the expression of ER, PR, human epidermal growth factor receptor 2 (HER2), cytokeratin 5 (CK5), epithelial growth factor receptor 1 (EGFR1) and the proliferation marker Ki67. Studies of risk of different subtypes have shown that weight or weight gain may be positively associated with the luminal subtypes,13–16 but no clear associations have been found for nonluminal subtypes. In some studies, body weight or central adiposity appear to be positively associated with risk of triple negative breast cancer (tumors negative for ER, PR and HER2).14,15 Another study reported a positive association between waist-hip-ratio (a marker for central adiposity) and risk of basal-like breast cancer (triple negative tumors that are positive for CK5 and/or EGFR).12 However, other investigators found no association between weight and
Horn et al.
2679
nonluminal subtypes.13,17 Height may be positively associated with risk of luminal tumors, but only one study has assessed the associations according to molecular subtype.16 We have prospectively assessed the associations of adult height and body mass index (BMI) with the risk of six breast cancer subtypes. The analyses were done within the context of a historic cohort of postmenopausal Norwegian women who were largely unaffected by menopausal hormone therapy and mammography screening, both of which may modify the association of body mass with breast cancer risk.2,18–21 In this cohort, incident breast cancers were subtyped using immunohistochemical and in situ hybridization techniques under standardized conditions in tissue microarrays (TMA) constructed from archival breast cancer tissue.
Material and Methods Study population and follow-up
Between 1956 and 1959, the Norwegian Cancer Society organized a survey to evaluate the effect of clinical breast examination as a screening tool for breast cancer. All women living in Nord-Trïndelag County, Norway, who were born between 1886 and 1928, were invited to participate. Each participant was interviewed according to a standardized questionnaire and had a clinical breast examination carried out by a physician (not including a mammography). Later (between 1963 and 1975), these women were also invited to a mandatory national tuberculosis screening that included standardized measurements of height and weight. A total of 19,373 women who were invited to the breast cancer screening in 1956–1959 attended the mandatory tuberculosis screening in 1963–1975. Both the screenings have been described in detail previously.22,23 At the tuberculosis screening, height was measured to the nearest centimeter and weight was measured to the nearest kilogram on regularly calibrated scales. Inaccuracy of measurements, for example, due to pregnancy, kyphosis or participants wearing shoes, was recorded. BMI was calculated as weight in kilograms divided by the squared value of height in meters. Using the unique 11-digit identification number of every Norwegian citizen, we linked individual information on study participants to information about breast cancer incidence recorded at the Cancer Registry of Norway, and to information on vital status and emigration provided by the National Population Register at Statistics Norway. Reporting of C 2014 UICC Int. J. Cancer: 135, 2678–2686 (2014) V
incident cancer to the Cancer Registry is mandatory by law and the registration is considered to be nearly complete.24 As the association of body weight with breast cancer risk differs by menopausal status,1,2 we restricted the study population to postmenopausal women using attained age of 55 years as a surrogate criterion for menopause. We excluded 141 women who died or emigrated before the age of 55 from the analysis. We also excluded 453 women with inaccurate height or weight measurements, as well as 217 women who were diagnosed with breast cancer before the height and weight measurements were conducted or before 55 years of age. Thus, 811 of the 19,373 women were excluded from analysis, leaving 18,562 women for follow-up. For all women, information on socioeconomic factors (place of residence, marital status and occupation (own or husband’s)) was provided by the National Population Register. Information on reproductive history (age at menarche, parity and age at first birth) was collected at the breast cancer screening in 1956– 1959, and was available for 16,419 women (88.5%). The participants were followed up for breast cancer incidence from the year they completed the weight and height measurement, or after reaching the age of 55 years, whichever event occurred last. End of follow-up was set to the date of a first breast cancer diagnosis, the date of death, the date of emigration or to the end of follow-up on December 31, 2008, whichever occurred first. Tissue microarray construction and immunohistochemical analysis
TMA construction, assay methods and immunohistochemical classification of the breast cancer cases in this cohort have been described in detail previously.25 Briefly, for women with a breast cancer diagnosis during follow-up, we used archival tumor tissue that was available at the Department of Pathology, St. Olavs Hospital in Trondheim, which is the primary pathology center serving Nord-Trïndelag County. The original diagnoses were confirmed as invasive breast carcinomas by two independent pathologists and classified according to histological type and grade.26,27 Three 1 mm tissue cores were obtained from representative areas of each tumor sample and placed into TMA blocks. Immunohistochemical staining was done for ER, PR, HER2, CK5, EGFR and Ki67. HER2 status was determined by chromogenic in situ hybridization (CISH) or, if CISH was unsuccessful, the immunohistochemical result was used.
Epidemiology
What’s new? Breast cancer comes in several different molecular subtypes, and different treatments may be called for depending on the subtype. Height and weight are known to affect risk of breast cancer after menopause, but for which subtypes does risk increase? In this study, the authors obtained height and weight data from more than 18,000 Norwegian women and compared these against breast cancer subtype. They report that height associates with risk of the Luminal A subtype, and BMI associates with all luminal subtypes and the HER2 subtype.
89.2
Rural
47.4 28.9
Manual
Domestic, others
42.4 25.5
25–29
30
32.8 16.7 5.0
160–164
165–169
170
6.3 7.6 1.3
III
IV
Missing
70.0 13.8 2.4 13.8
Invasive ductal
Invasive lobular
Medullary
Others
Histology (%)
34,1
II
1
50.7
74 (67–81)
6.2
17.6
36.6
39.5
30.0
41.2
28.8
26.1
44.6
29.3
87.7
12.3
54 (47–61)
All breast cancers, n 5 969
I
Stage at diagnosis (%)
Age at diagnosis (years), median (IQR)
45.6