Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2023 Jan 25.
Published in final edited form as: Br J Sports Med. 2022 Sep 6;56(20):1157–1170. doi: 10.1136/bjsports-2021-105132

Physical activity, sedentary time and breast cancer risk: A Mendelian randomization study

Suzanne C Dixon-Suen 1, Sarah J Lewis 2,3, Richard M Martin 3,4,5, Dallas R English 1,6, Terry Boyle 7,8, Graham G Giles 1,6,9, Kyriaki Michailidou 10,11,12, Manjeet K Bolla 12, Qin Wang 12, Joe Dennis 12, Michael Lush 12; ABCTB Investigators13, Thomas U Ahearn 14, Christine B Ambrosone 15, Irene L Andrulis 16,17, Hoda Anton-Culver 18, Volker Arndt 19, Kristan J Aronson 20, Annelie Augustinsson 21, Päivi Auvinen 22,23, Laura E Beane Freeman 14, Heiko Becher 24, Matthias W Beckmann 25, Sabine Behrens 26, Marina Bermisheva 27, Carl Blomqvist 28,29, Natalia V Bogdanova 30,31,32, Stig E Bojesen 33,34,35, Bernardo Bonanni 36, Hermann Brenner 19,37,38, Thomas Brüning 39, Saundra S Buys 40, Nicola J Camp 40, Daniele Campa 26,41, Federico Canzian 42, Jose E Castelao 43, Melissa H Cessna 44,45, Jenny Chang-Claude 26,46, Stephen J Chanock 14, Christine L Clarke 47, Don M Conroy 48, Fergus J Couch 49, Angela Cox 50, Simon S Cross 51, Kamila Czene 52, Mary B Daly 53, Peter Devilee 54,55, Thilo Dörk 31, Miriam Dwek 56, Diana M Eccles 57, A Heather Eliassen 58,59, Christoph Engel 60,61, Mikael Eriksson 52, D Gareth Evans 62,63, Peter A Fasching 25,64, Olivia Fletcher 65, Henrik Flyger 66, Lin Fritschi 67, Marike Gabrielson 52, Manuela Gago-Dominguez 68,69, Montserrat García-Closas 14, José A García-Sáenz 70, Mark S Goldberg 71,72, Pascal Guénel 73, Melanie Gündert 74,75, Eric Hahnen 76,77, Christopher A Haiman 78, Lothar Häberle 25, Niclas Håkansson 79, Per Hall 52,80, Ute Hamann 81, Steven N Hart 82, Michelle Harvie 83, Peter Hillemanns 31, Antoinette Hollestelle 84, Maartje J Hooning 84, Reiner Hoppe 85,86, John L Hopper 6, Anthony Howell 87, David J Hunter 59,88, Anna Jakubowska 89,90, Wolfgang Janni 91, Esther M John 92,93, Audrey Jung 26, Rudolf Kaaks 26, Renske Keeman 94, Cari M Kitahara 95, Stella Koutros 14, Peter Kraft 59,96, Vessela N Kristensen 97,98, Katerina Kubelka-Sabit 99, Allison W Kurian 92,93, James V Lacey 100,101, Diether Lambrechts 102,103, Loic Le Marchand 104, Annika Lindblom 105,106, Sibylle Loibl 107, Jan Lubiński 89, Arto Mannermaa 108,109,110, Mehdi Manoochehri 81, Sara Margolin 80,111, Maria Elena Martinez 69,112, Dimitrios Mavroudis 113, Usha Menon 114, Anna Marie Mulligan 115,116, Rachel A Murphy 117,118; NBCS Collaborators97,98,119,120,121,122,123,124,125,126,127,128, Heli Nevanlinna 129, Ines Nevelsteen 130, William G Newman 62,63, Kenneth Offit 131,132, Andrew F Olshan 133, Håkan Olsson 21,§, Nick Orr 134, Alpa V Patel 135, Julian Peto 136, Dijana Plaseska-Karanfilska 137, Nadege Presneau 56, Brigitte Rack 91, Paolo Radice 138, Erika Rees-Punia 135, Gad Rennert 139, Hedy S Rennert 139, Atocha Romero 140, Emmanouil Saloustros 141, Dale P Sandler 142, Marjanka K Schmidt 94,143, Rita K Schmutzler 76,77,144, Lukas Schwentner 91, Christopher Scott 82, Mitul Shah 48, Xiao-Ou Shu 145, Jacques Simard 146, Melissa C Southey 1,9,147, Jennifer Stone 6,148, Harald Surowy 74,75, Anthony J Swerdlow 149,150, Rulla M Tamimi 59,151, William J Tapper 57, Jack A Taylor 142,152, Mary Beth Terry 153, Rob AEM Tollenaar 154, Melissa A Troester 133, Thérèse Truong 73, Michael Untch 155, Celine M Vachon 156, Vijai Joseph 131, Barbara Wappenschmidt 76,77, Clarice R Weinberg 157, Alicja Wolk 79,158, Drakoulis Yannoukakos 159, Wei Zheng 145, Argyrios Ziogas 18, Alison M Dunning 48, Paul DP Pharoah 12,48, Douglas F Easton 12,48, Roger L Milne 1,6,9,†, Brigid M Lynch 1,6,160,†,*, on behalf of the Breast Cancer Association Consortium
PMCID: PMC9876601  NIHMSID: NIHMS1822689  PMID: 36328784

Abstract

Objectives:

Physical inactivity and sedentary behaviour are associated with higher breast cancer risk in observational studies, but ascribing causality is difficult. Mendelian randomization (MR) assesses causality by simulating randomized trial groups using genotype. We assessed whether lifelong physical activity or sedentary time, assessed using genotype, may be causally associated with breast cancer risk overall, pre/post-menopause, and by case-groups defined by tumour characteristics.

Methods:

We performed two-sample inverse-variance-weighted MR using individual-level Breast Cancer Association Consortium case-control data from 130,957 European-ancestry women (69,838 invasive cases), and published UK Biobank data (n=91,105–377,234). Genetic instruments were single nucleotide polymorphisms (SNPs) associated in UK Biobank with wrist-worn accelerometer-measured overall physical activity (nsnps=5) or sedentary time (nsnps=6), or accelerometer-measured (nsnps=1) or self-reported (nsnps=5) vigorous physical activity.

Results:

Greater genetically-predicted overall activity was associated with lower breast cancer risk, overall (OR=0.59; 95%CI 0.42–0.83 per-standard deviation [SD; ~8 milligravities acceleration]) and for most case-groups. Genetically-predicted vigorous activity was associated with lower risk of pre/perimenopausal breast cancer (OR=0.62; 95%CI 0.45–0.87, ≥3 vs. 0 self-reported days/week), with consistent estimates for most case-groups. Greater genetically-predicted sedentary time was associated with higher hormone-receptor-negative tumour risk (OR=1.77; 95%CI 1.07–2.92 per-SD [~7% time spent sedentary]), with elevated estimates for most case-groups. Results were robust to sensitivity analyses examining pleiotropy (including weighted-median-MR, MR-Egger).

Conclusion:

Our study provides strong evidence that greater overall physical activity, greater vigorous activity, and lower sedentary time are likely to reduce breast cancer risk. More widespread adoption of active lifestyles may reduce the burden from the most common cancer in women.

Keywords: Breast cancer, Physical activity, Sedentary time, Mendelian randomization, Causal inference

Introduction

Greater physical activity and less sedentary time are associated with lower breast cancer risk in observational studies. International and national cancer agencies have concluded that physical activity may reduce breast cancer risk, particularly postmenopausal disease, with associations strongest for vigorous activity.(1–3) Sedentary (sitting/reclining) time, a distinct exposure affecting ‘active’ and ‘inactive’ people, has been less well-studied, with conflicting findings.(4, 5) Physical inactivity or excess sitting may plausibly influence breast cancer initiation and/or growth. However, whether observed associations are causal or produced by biases (e.g. confounding, selection bias, reverse causation) is unclear. Mendelian randomization (MR) can simulate randomized controlled trials using observational data by substituting genotypes, which are randomly assigned at meiosis (before conception), as instruments (proxies) for exposures of interest.(6) Subject to meeting specific assumptions of instrumental variable analysis,(7) some of which can be investigated using sensitivity analyses (see Methods), MR can minimise confounding and reverse causation, potentially providing stronger evidence for causal inference.

A recent MR study assessed physical activity and breast cancer risk overall and by oestrogen-receptor (ER) status,(8) but did not examine other breast tumour types, vigorous activity, or sedentary time. We aimed to appraise the causal nature of associations between overall activity, vigorous activity, and sedentary time, and breast cancer risk, overall and by menopausal status, stage, grade, morphology, and molecular subtypes defined by hormone-receptor (ER, progesterone [PR]) and human epidermal growth factor receptor-2 (HER2) status.

Methods

Data sources

We performed two-sample MR using individual-level data from 130,957 European-ancestry women (69,838 with invasive breast cancers; 6,667 with in situ breast cancers; 54,452 controls) from 76 Breast Cancer Association Consortium (BCAC) studies (Tables 1, S1)(outcome dataset), and genetic estimates for movement-related exposures from published genome-wide association studies (GWAS) using UK Biobank data (exposure datasets; n=91,105–377,234).(9–11) Instruments were single-nucleotide polymorphisms (SNPs) associated in the UK Biobank GWAS with overall physical activity (all movement), vigorous physical activity, or sedentary time (Table S2).

Table 1.

Characteristics of 76 Breast Cancer Association Consortium studies, and 130,957 study participants, included in the individual-level analysis

Study acronym a Country Diagnosis years Invasive cases(N) In situ cases(N) Controls (N)
ABCFS Australia 1963–2013 1,117 - 187
ABCTB Australia 2004–2013 920 6 375
BCEES Australia 2009–2011 783 - 834
MCCS Australia 1981–2012 870 180 978
HMBCS Belarus 1994–2007 212 - 249
LMBC Belgium 1994–2011 784 21 1,268
CBCS Canada 2005–2009 568 108 817
MTLGEBCS Canada 2007–2011 341 - 170
OFBCR Canada 1967–2015 1,721 2 643
CGPS Denmark 1981–2012 1,408 3 716
EPIC Europe (Multiple countries) n.r. 3,435 412 3,597
HEBCS Finland 1997–2012 281 - 177
KBCP Finland 1990–2012 522 34 245
CECILE France 2005–2007 280 26 159
BBCC Germany 1988–2013 403 8 253
BSUCH Germany 1990–2013 252 1 168
ESTHER Germany 2001–2004 291 3 187
GC-HBOC Germany 1947–2014 3,378 256 1,593
GENICA Germany 2000–2004 459 1 284
GEPARSIXTO Germany n.r. 386 - -
GESBC Germany 1992–1995 312 39 181
HABCS Germany 1984–2010 909 19 863
MARIE Germany 2001–2005 506 6 289
PREFACE Germany 2001–2011 2,923 - -
SKKDKFZS Germany 1993–2005 1,086 9 -
SUCCESSB Germany 2008–2011 440 - -
SUCCESSC Germany 2001–2011 2,836 - -
CCGP Greece 1983–2013 667 5 322
BCINIS Israel 1999–2012 1,337 100 724
MBCSG Italy 1977–2012 549 72 366
ABCS Netherlands 2003–2011 347 - 189
ORIGO Netherlands 1991–2005 921 113 -
RBCS Netherlands 1975–2009 444 23 -
NBCS Norway 1973–2011 1,163 38 -
PBCS Poland 1998–2003 1,740 111 2,045
SZBCS Poland 2010–2012 352 9 174
MABCS Republic of North Macedonia 1993–2013 89 1 90
HUBCS Russia 1977–2009 211 - 116
BREOGAN Spain 1991–2019 1,535 129 910
HCSC Spain 1975–2013 423 3 -
KARBAC Sweden 1966–2013 499 3 -
KARMA Sweden 1969–2017 2,839 339 6,983
MISS Sweden 1983–2013 633 68 1,529
pKARMA Sweden 1980–2015 748 86 48
SMC Sweden 1987–2013 1,509 - 661
BBCS UK 1985–2009 122 - 440
DIETCOMPLYF UK 2004–2007 708 3 -
FHRISK UK 1987–2015 146 31 644
POSH UK 2000–2007 1,088 - -
PROCAS UK 1988–2018 380 93 1,648
SBCS UK 2012–2015 126 2 -
SEARCH UK 2003–2012 4,057 - 2,653
UKBGS UK 1985–2014 1,048 584 705
UKOPS UK n.a. - - 974
2SISTER USA n.r. 919 151 -
AHS USA 1994–2013 513 1 1,137
BCFR-NY USA 1949–2011 401 53 27
BCFR-PA USA 1969–2011 67 6 -
BCFR-UTAH USA 1952–2009 100 1 -
CPSII USA 1992–2009 2,393 598 3,028
CTS USA 1998–2010 1,156 - 610
MCBCS USA 1998–2014 749 167 212
MEC USA 1972–2012 668 5 724
MMHS USA 2003–2013 275 99 1,635
MSKCC USA 1982–2012 136 2 -
NBHS USA 2001–2009 483 112 652
NC-BCFR USA 1967–2012 759 15 150
NCBCS USA 1993–2012 2,074 315 1,006
NHS USA 1976–2012 1,103 333 1,804
NHS2 USA 1989–2011 1,112 409 1,905
PLCO USA 1994–2013 1,822 483 2,595
SISTER USA 2003–2008 1,504 498 1,556
TNBCC USA 2003–2013 113 - -
UBCS USA 1960–2015 606 60 -
UCIBCS USA 1994–2003 427 74 258
USRT USA 1945–2005 1,354 338 1,699
Total 1945–2019 69,838 6,667 54,452
Open in a new tab

n.a., not applicable; n.r., not recorded

a

See Supplementary Table S1 (Online Resource) for study names and references.

Exposures

Overall physical activity

As our primary physical activity instrument we used five SNPs associated with overall activity (p<5×10−8) in a prior GWAS of accelerometer-assessed movement in the UK Biobank (n=91,105) (9), which explain 0.10% of the variance in activity. Doherty and colleagues assessed overall activity as average vector magnitude (milligravities) per 30-second period,(9, 12) with mean (standard deviation, SD) 29.0 (8.0) milligravities among women in UK Biobank.(13) One SD (8 milligravities) corresponds to ~50 minutes of moderate (e.g. brisk walking) activity per week.(8)

For comparability with the previous MR study on this topic,(8) we used an expanded set of ten SNPs as a secondary instrument for overall activity. These SNPs were associated at relaxed significance (p<5×10−7) with the accelerometer-assessed overall activity phenotype in a separate UK Biobank GWAS of physical activity by Klimentidis and colleagues.(10, 11)

Vigorous physical activity

Klimentidis and colleagues identified one SNP associated (p<5×10−9) with high-intensity movement, assessed as the fraction of 30-second intervals containing accelerations over 425 milligravities.(10) This threshold approximates expenditure output for vigorous activity (>6 metabolic equivalents of task [METs]).(14) This SNP explains approximately 0.02% of variance in high-intensity movement. They identified five SNPs associated (p<5×10−9) with self-reported engagement in vigorous activity for at least ten minutes ≥3 vs. 0 days/week (n=377,234), (10) which explain approximately 0.06% of variance in this exposure. We examined both instruments as complementary measures for vigorous activity, each likely subject to different error (weak instrument or reporting bias).

Sedentary time

Doherty and colleagues applied machine-learning models, trained using body-camera and diary data, to UK Biobank accelerometry data to identify sedentary periods (sitting/reclining; MET-value typically ≤1.5).(9, 13) They identified six SNPs associated (p<5×10−8) with the probability of engaging in sedentary behaviours, defined as the ratio of sedentary-to-total 30-second periods.(9) On average UK Biobank women spent 34.6% (SD=7.2%) of their time sedentary.(13) We used these six variants, explaining 0.12% of variance in sedentariness, as our sedentary time instrument.

Outcomes

We estimated breast cancer risk overall, by menopausal status, and by case-groups defined by molecular/morphological subtype, stage, or grade at diagnosis, using BCAC clinical data to assign case-groups according to hypotheses arising from the literature. We defined separate case/control groups for invasive pre/peri-menopausal (n=23,999 cases; 17,686 controls) and postmenopausal (n=45,839 cases; 36,766 controls) breast cancers, using age at diagnosis/interview (</≥50 years) to assign missing menopausal status (27%). We examined subtypes separately by hormone-receptor (HR) status (ER+/− n=46,528/11,246; PR+/− n=34,891/16,432) and HER2 status (+/− n=6,945/33,214), and jointly including HER2-enriched (ER−/PR−/HER2+; n=1,974) and triple-negative (ER−/PR−/HER2−; n=4,964) cancers. We examined invasive ductal/lobular cancers (n=42,223/8,795), ductal carcinoma in situ (n=3,510), and risk by stage (stage I, n=17,583; stage II, n=15,992; stages III/IV, n=4,553) and grade (well/moderately differentiated, n=34,647; poorly/undifferentiated, n=16,432).

SNP-exposure (UK Biobank) and SNP-outcome (BCAC) associations

We extracted or derived estimates of association (beta coefficients, standard errors [SEs]) between SNPs and exposures from the UK Biobank GWAS publications,(9, 10) standardised to refer to the trait-increasing allele. Where required,(10) we converted estimates to per-SD changes in activity/sedentary time using UK Biobank activity data.

Genotypes in BCAC were determined using the OncoArray, an Illumina custom array, and imputed using IMPUTE2.(15) We harmonised UK Biobank and BCAC data so SNP-exposure and SNP-outcome estimates related to the same allele, using allele frequency information to resolve strand-ambiguous SNPs where possible (i.e., unless allele frequencies were 45%−55%). For each SNP, we derived trait-specific effect-allele dosages (range 0–2) by summing alleles predicting more activity (activity instruments) or sedentary time (sitting instrument). We assessed the association between each SNP and each outcome from individual-level BCAC data by fitting logistic regression models, adjusted for age at diagnosis (cases) or interview (controls), country, and ten principal components of genetic population structure (accounting for genetic substructure within Europeans), obtaining beta coefficients and SEs for use in the MR analysis. Table 1 summarises the BCAC studies and participants.

Statistical analysis

We used SNP-exposure and SNP-outcome beta coefficients and SEs to estimate odds ratios (OR) and 95% confidence intervals (CI) of the effect of each trait on each outcome. For single SNPs, we divided the SNP-outcome association by the SNP-exposure association to obtain the causal estimate (Wald ratio). For multi-SNP instruments, we used inverse-variance weighted (IVW)-MR, which averages Wald ratios across SNPs, weighted by SNP-exposure beta coefficients.(16–18) IVW-MR assumes all instruments are valid or that pleiotropy is balanced,(17) and we assumed linearity in the associations between the SNPs and exposure, and between SNPs and outcome. We performed case-only analyses to test for differences between subtypes.

Core assumptions of MR, which can be investigated using sensitivity analyses, are that the instrument: predicts exposure; is not associated with confounders of the exposure/outcome association; and influences the outcome only via the exposure (no horizontal pleiotropy) (6, 7, 19), summarised in Figure S1. We undertook sensitivity analyses to assess the robustness of our findings and the potential for violations of assumptions, most critically horizontal pleiotropy. We calculated Cochran’s Q-statistic for between-SNP heterogeneity of effects. We applied complementary methods relaxing different MR assumptions, weighted-median MR (allows invalid instruments)(20) and MR-Egger (allows horizontal pleiotropy, although prone to imprecision)(19, 21). We inspected per-SNP causal estimates (scatter, forest plots) and leave-one-out analyses to identify SNPs distorting results. We performed MR Pleiotropy Residual Sum and Outlier (MR-PRESSO) to identify outlying SNPs with evidence of horizontal pleiotropy (global-pleiotropy and SNP-outlier tests p<0.05).(22) We examined the effect of excluding two SNPs with imputation quality <0.9. We checked whether SNPs are associated with other relevant traits (possible confounders, adiposity, cancer risk) or gene expression using the NHGRI-EBI GWAS Catalog(23) and PhenoScanner.(24, 25)

Data preparation and analyses were performed using R software (R Foundation for Statistical Computing, Vienna), including the ‘MendelianRandomization’(18) and ‘MR-PRESSO’ packages.(22) Statistical power was calculated using the mRnd Mendelian randomization power calculation online tool.(26) Further details are in Supplementary Methods (Online Resource).

Results

Overall physical activity

Greater genetically-predicted physical activity was associated with lower risk of invasive breast cancer (OR=0.48;95%CI 0.30–0.78 per-SD [~8 milligravities] in overall activity), with no clearly differential effects by menopausal status, molecular subtype, morphology, stage, or grade (Table 2). We observed ORs less than 1 for all outcomes, including ER+ (OR=0.45;95%CI 0.25–0.83), PR+ (OR=0.43;95%CI 0.22–0.85), HER2+ (OR=0.48;95%CI 0.26–0.89), and HR+/HER2+ (OR=0.42;95%CI 0.20–0.88) disease. Weighted-median MR and MR-Egger results were broadly consistent (Table S3).

Table 2.

Association between the primary instrumental genetic variables for overall physical activity (per standard deviation) and risk of breast cancer

Full instrument (five SNPs) Excluding one pleiotropic SNP for outcomes with detected pleiotropya
Type of breast cancer N cases (vs. 54,452 controls) Odds ratios (95% CI)b P for heterogeneityc Odds ratios (95% CI) b P for heterogeneity c
Invasive cancers
All invasive 69,838 0.48 (0.30–0.78) 0.016 0.59 (0.42–0.83) 0.312
Pre/perimenopausal d 23,999 0.51 (0.31–0.83) 0.419 --
Postmenopausal e 45,839 0.48 (0.28–0.80) 0.054 --
By receptor status
ER+ 46,528 0.45 (0.25–0.83) 0.004 0.60 (0.43–0.85) 0.459
ER− 11,246 0.79 (0.37–1.66) 0.069 --
PR+ 34,891 0.43 (0.22–0.85) 0.003 0.58 (0.37–0.91) 0.223
PR− 16,432 0.65 (0.38–1.13) 0.186 --
HER2+ 6,945 0.48 (0.26–0.89) 0.479 --
HER2− 33,214 0.58 (0.35–0.98) 0.060 --
Combined hormone receptor- and/or HER2-defined subtypes
ER+ or PR+; HER2+ 4,816 0.42 (0.20–0.88) 0.478 --
ER+ or PR+; HER2− 27,874 0.57 (0.28–1.18) 0.004 0.79 (0.49–1.26) 0.254
ER−; PR−; HER2+ 1,974 0.53 (0.18–1.57) 0.700 --
ER−; PR−; HER2− 4,964 0.60 (0.17–2.12) 0.015 0.95 (0.37–2.44) 0.224
ER− and PR− (all) 9,215 0.65 (0.27–1.56) 0.036 0.46 (0.22–0.96) 0.226
By morphology
Ductal 42,223 0.52 (0.32–0.84) 0.053 --
Lobular 8,795 0.32 (0.18–0.58) 0.500 --
By stage at diagnosis
Stage I 17,583 0.51 (0.32–0.82) 0.333 --
Stage II 15,992 0.36 (0.22–0.58) 0.576 --
Stage III/IV 4,553 0.37 (0.17–0.81) 0.499 --
By tumour grade
Grade 1/2 34,647 0.43 (0.23–0.81) 0.011 0.58 (0.39–0.85) 0.514
Grade 3 16,432 0.46 (0.30–0.72) 0.552 --
In situ cancers
All in situ 6,667 0.63 (0.34–1.18) 0.390 --
Ductal carcinoma in situ 3,510 f 0.92 (0.25–3.43) 0.039 --
Open in a new tab

Abbreviations: CI, confidence interval; ER+/−, oestrogen receptor positive/negative; GWAS, genome wide association study; HER2+/−, human epidermal growth factor receptor 2 positive/negative; PR+/−, progesterone receptor positive/negative; SNP, single nucleotide polymorphism.

a

Outlying SNP rs564819152 was excluded from analyses of all invasive, ER+, PR+, HR+/HER2−, and well/moderately differentiated cancers (outlier identified by MR-PRESSO, global-pleiotropy test p<0.05), and HR− cancers (outlier suggested by scatter plots and leave-one-out analyses; MR-PRESSO global-pleiotropy test p=0.053). Outlying SNP rs6775319 was excluded from analyses of triple negative cancers (ER−/PR−/HER2−), and was identified by MR-PRESSO.

b

Causal odds ratios were estimated by inverse-variance weighted Mendelian randomization, using SNPs identified in a GWAS of accelerometer-measured movement traits by Doherty et al (9)

c

p-value associated with the heterogeneity test statistic (Cochran’s Q statistic) measuring heterogeneity of causal effects between SNPs

d

vs pre/perimenopausal controls (n=17,686), assigned using age (<50 years) if menopause status was unknown

e

vs postmenopausal controls (n=36,766), assigned using age (≥50 years) if menopause status was unknown

f

For analyses of ductal carcinoma in situ, likely pleiotropy was indicated by the Cochran’s Q statistic (phet=0.04) and the MR-Egger intercept test for horizontal pleiotropy (pintercept=0.01). However, a clear outlying SNP could not be identified, although leave-one-out analyses suggested substantial variation in results by instrument composition.

-- No outlying SNPs were identified.

Heterogeneity of causal effects between SNPs was evident for some outcomes (Cochran’s-Q phet<0.05)(Table 2); this was resolved after removing outliers rs564819152 (associated previously with ovarian cancer; outlying for six outcomes) or rs6775319 (one outcome), detected by MR-PRESSO, per-SNP, and leave-one-out analyses (Figures S2–S3; Table S4). Evidence of protective associations remained strong after excluding rs564819152 (Table 2). Outlier-corrected results (OR [95%CI]) were 0.59 (0.42–0.83) for all invasive breast cancer, 0.60 (0.43–0.85) for ER+, and 0.58 (0.37–0.91) for PR+ disease (HER2+ and HR+/HER2+ analyses had no outlying SNPs).

The protective effects were consistent across leave-one-out analyses (Table S4). SNPs were not associated in prior GWAS with confounders of the exposure/outcome relationship, but two had been identified in an ovarian cancer GWAS (Table S5). Excluding these made little difference to results (Table S4). Two SNPs have been reported to be associated (p<5×10−8) with adiposity in UK Biobank,(24, 25, 27) consistent with reduced adiposity being a downstream effect of increased activity (Table S5).

Results were similar although slightly attenuated using the expanded ten-SNP instrument(10)(Table S6–S7). Estimates generally remained protective upon removing outlying SNPs detected by pleiotropy investigations (IVW heterogeneity tests [Table S6], MR-PRESSO, per-SNP effects [Figures S4–S5], leave-one-out analysis [Table S8]). Most estimates were similar (Table S8) upon excluding one SNP with imputation quality <0.9 (Table S8). Four of the ten SNPs were associated in prior GWAS with confounders (including height, alcohol intake, education) or cancer risk. Furthermore, rs55657917 is associated with gene expression in breast tissue, including in two genes associated with breast cancer risk (Table S5).(23–25) However, results excluding potentially confounded SNPs were relatively unchanged (Table S8). For four SNPs, the activity-increasing allele is associated with reduced adiposity in UK Biobank.(27)

Vigorous physical activity

There was little evidence that genetically-predicted acceleration over 425 milligravities (one SNP) was associated with risk of breast cancer, with wide confidence intervals crossing one, although most estimates were in the protective direction (Table 3). The activity-increasing allele has been associated in GWAS(24, 25, 27) with greater height and decreased adiposity (Table S5).

Table 3.

Association between instrumental genetic variables for vigorous physical activity, assessed in two ways, and risk of breast cancer

Accelerometer-measured activity over 425 milligravities, per fraction of time, using one SNP a Self-reported vigorous physical activity (≥ 3 vs. 0 days/week)
Full instrument (five SNPs) Excluding one pleiotropic SNP for outcome with detected pleiotropy b
Type of breast cancer N cases (vs. 54,452 controls) Odds ratios (95% CI) c Odds ratios (95% CI) c P for heterogeneity d Odds ratios (95% CI) c P for heterogeneity d
Invasive cancers
All invasive 69,838 0.63 (0.32–1.22) 0.83 (0.69–1.01) 0.650 --
Pre/perimenopausal e 23,999 0.80 (0.25–2.58) 0.62 (0.45–0.87) 0.788 --
Postmenopausal f 45,839 0.53 (0.24–1.21) 0.95 (0.75–1.19) 0.630 --
By receptor status
ER+ 46,528 0.74 (0.35–1.55) 0.86 (0.70–1.07) 0.917 --
ER− 11,246 0.58 (0.17–1.94) 0.86 (0.61–1.21) 0.418 --
PR+ 34,891 0.68 (0.30–1.54) 0.77 (0.61–0.98) 0.544 --
PR− 16,432 0.56 (0.19–1.59) 0.95 (0.70–1.28) 0.948 --
HER2+ 6,945 0.31 (0.07–1.39) 0.83 (0.53–1.31) 0.327 --
HER2- 33,214 1.01 (0.44–2.31) 0.86 (0.68–1.10) 0.550 --
Combined hormone receptor- and/or HER2-defined subtypes
ER+ or PR+; HER2+ 4,816 0.41 (0.07–2.35) 1.00 (0.58–1.70) 0.321 --
ER+ or PR+; HER2− 27,874 0.84 (0.35–2.02) 0.82 (0.64–1.06) 0.560 --
ER−; PR−; HER2+ 1,974 0.21 (0.02–2.87) 0.57 (0.27–1.20) 0.727 --
ER−; PR−; HER2− 4,964 2.16 (0.39–12.1) 1.30 (0.79–2.12) 0.593 --
ER− and PR− (all) 9,215 0.78 (0.21–2.91) 0.95 (0.66–1.39) 0.559 --
By morphology
Ductal 42,223 0.62 (0.29–1.32) 0.81 (0.65–1.00) 0.932 --
Lobular 8,795 0.60 (0.15–2.45) 0.78 (0.53–1.17) 0.809 --
By stage at diagnosis
Stage I 17,583 0.47 (0.16–1.36) 0.88 (0.65–1.19) 0.598 --
Stage II 15,992 0.66 (0.21–2.07) 0.82 (0.59–1.14) 0.788 --
Stage III/IV 4,553 0.41 (0.06–2.63) 0.75 (0.44–1.27) 0.910 --
By tumour grade
Grade 1/2 34,647 0.54 (0.24–1.22) 0.84 (0.66–1.06) 0.640 --
Grade 3 16,432 0.51 (0.18–1.46) 0.99 (0.73–1.33) 0.557 --
In situ cancers
All in situ 6,667 0.47 (0.11–2.09) 0.94 (0.43–2.08) 0.007 1.30 (0.72–2.34) 0.189
Ductal carcinoma in situ 3,510 0.65 (0.09–4.72) 0.85 (0.42–1.69) 0.204 --
Open in a new tab

Abbreviations: CI, confidence interval; ER+/−, oestrogen receptor positive/negative; GWAS, genome wide association study; HER2+/−, human epidermal growth factor receptor 2 positive/negative; PR+/−, progesterone receptor positive/negative; SNP, single nucleotide polymorphism.

a

This SNP is a missense mutation in the gene PML, which plays a role in tumour suppression and is associated with height. PML is not expressed in breast tissue, but highly expressed in adipose tissue, suggesting that inverse (protective) associations observed do not derive from direct oncosuppression.

b

Excluding one outlying SNP identified by MR-PRESSO: rs2764261 (for analyses modelling the association with in situ tumours)

c

Causal odds ratios were estimated by inverse-variance weighted Mendelian randomization, using SNPs identified in a GWAS of physical activity by Klimentidis et al (10)

d

p-value associated with the heterogeneity test statistic (Cochran’s Q statistic) measuring heterogeneity of causal effects between SNPs

e

vs pre/perimenopausal controls (n=17,686), assigned using age (<50 years) if menopause status was unknown

f

vs postmenopausal controls (n=36,766), assigned using age (≥50 years) if menopause status was unknown

-- No outlying SNPs were identified by MR-PRESSO.

There was weak evidence that genetically-predicted self-reported vigorous activity was associated with decreased breast cancer risk overall (OR=0.83;95%CI 0.69–1.01, ≥3 days/week vs. none), and ORs for most case-groups were less than 1 (Table 3). A protective association was seen for pre/perimenopausal breast cancer (OR=0.62;95%CI 0.45–0.87), with little evidence for an association with postmenopausal breast cancer risk (OR=0.95;95%CI 0.75–1.19) (p=0.82 for the difference in pre/peri- vs. post-menopausal estimates). A protective relationship was seen for PR+ disease (OR=0.77;95%CI 0.61–0.98). There was little evidence of pleiotropic effects (Table 3, S9–S10) except one outlier in modelling in situ cancers (Figures S6–S7), a SNP previously associated with height, age at menarche, and adiposity (Table S5).(24, 25, 27) After excluding this SNP, the in situ OR was elevated (from OR=0.94;95%CI 0.43–2.08 to OR=1.30;0.72–2.34)(Table 3); other estimates remained similar (Table S10). Excluding one SNP associated in UK Biobank GWAS with past smoking and childhood height (Table S5)(24, 25, 27) attenuated estimates slightly (Table S10). The association with pre/perimenopausal cancers remained substantially inverse (protective), with confidence intervals that did not cross the null, in all sensitivity analyses (Table S10).

Sedentary time

The estimates for genetically-predicted sedentary time were elevated (in the direction of increased risk) for almost every case-group, although CIs were wide (Table 4). Greater sedentary time was associated with higher risk of hormone-receptor-negative (HR−) tumours (OR=1.77;95%CI 1.07–2.92 per-SD [~7% time spent sedentary]), including triple-negative (ER−/PR−/HER2−) cancers (OR=2.04;95%CI 1.06–3.93) (p=0.11 for the difference in ORs by HR-status). ORs were substantially elevated for in situ cancers (OR=1.75;95%CI 1.00–3.07), specifically ductal carcinoma in situ (OR=2.11;95%CI 0.99–4.49). The point estimate was elevated for stage I tumours (OR=1.62;95%CI 0.99–2.65), with little evidence of association with stage III/IV (OR=0.91;95%CI 0.45–1.84) (p=0.25 for the difference in estimates for risk of stage I vs stage III/IV tumours).

Table 4.

Association between instrumental genetic variables for sedentary time (per standard deviation in percent time spent sedentary) and risk of breast cancer

Type of breast cancer N cases (vs. 54,452 controls) Odds ratios (95% CI) a P for heterogeneity b
Invasive cancers
All invasive 69,838 1.20 (0.93–1.55) 0.962
Pre/perimenopausal c 23,999 1.22 (0.78–1.90) 0.589
Postmenopausal d45,839 1.21 (0.89–1.65) 0.983
By receptor status
ER+ 46,528 1.19 (0.90–1.57) 0.992
ER− 11,246 1.43 (0.90–2.26) 0.926
PR+ 34,891 1.19 (0.87–1.63) 0.386
PR− 16,432 1.40 (0.94–2.09) 0.435
HER2+ 6,945 1.17 (0.67–2.06) 0.718
HER2− 33,214 1.27 (0.93–1.74) 0.955
Combined hormone receptor- and/or HER2-defined subtypes
ER+ or PR+; HER2+ 4,816 0.86 (0.44–1.67) 0.585
ER+ or PR+; HER2− 27,874 1.12 (0.80–1.56) 0.801
ER−; PR−; HER2+ 1,974 1.94 (0.71–5.25) 0.646
ER−; PR−; HER2− 4,964 2.04 (1.06–3.93) 0.500
ER− and PR− (all) 9,215 1.77 (1.07–2.92) 0.819
By morphology
Ductal 42,223 1.21 (0.91–1.62) 0.992
Lobular 8,795 1.12 (0.66–1.91) 0.695
By stage at diagnosis
Stage I 17,583 1.62 (0.99–2.65) 0.187
Stage II 15,992 1.23 (0.79–1.90) 0.820
Stage III/IV 4,553 0.91 (0.45–1.84) 0.640
By tumour grade
Grade 1/2 34,647 1.15 (0.84–1.57) 0.901
Grade 3 16,432 1.32 (0.88–1.97) 0.967
In situ cancers
All in situ 6,667 1.75 (1.00–3.07) 0.933
Ductal carcinoma in situ 3,510 2.11 (0.99–4.49) 0.487
Open in a new tab

Abbreviations: CI, confidence interval; ER+/−, oestrogen receptor positive/negative; GWAS, genome wide association study; HER2+/−, human epidermal growth factor receptor 2 positive/negative; PR+/−, progesterone receptor positive/negative; SNP, single nucleotide polymorphism.

a

Causal odds ratios were estimated by inverse-variance weighted Mendelian randomization, using six SNPs identified in a GWAS of accelerometer-measured movement traits by Doherty et al (9)

b

p-value associated with the heterogeneity test statistic (Cochran’s Q statistic) measuring heterogeneity of causal effects between SNPs

c

vs pre/perimenopausal controls (n=17,686), assigned using age (<50 years) if menopause status was unknown

d

vs postmenopausal controls (n=36,766), assigned using age (≥50 years) if menopause status was unknown

Heterogeneity between SNPs was not detected (all phet>0.2)(Table 4), all MR methods produced broadly consistent results (Table S11), and MR-PRESSO did not identify outlying SNPs. Estimates were consistently elevated across leave-one-out analyses, including after omitting: one SNP correlated with a physical activity variant; one SNP predicting greater education and adiposity in prior GWAS(24, 25, 27, 28); or one strand-ambiguous SNP with minor allele frequency ~50%, for which effect-allele harmonisation was not definitive (Table S12). After excluding a SNP with imputation quality <0.9, which may have been an outlier for PR+ analyses (Figures S8–S9; MR-Egger ppleiotropy=0.046 for PR+), point estimates for PR+ and most other outcomes including HR−, triple-negative, and in situ cancers, moved further from null (Table S12). Estimates for HR− and in situ cancers remained substantially elevated in all sensitivity analyses (Table S12).

Discussion

Main findings

We conducted a Mendelian randomization study using individual-level data on 130,957 women. We found that women with genetic variants predisposing them to be more active had lower breast cancer risk overall and for most case-groups defined by tumour subtypes, stage, or grade. Effect estimates for vigorous physical activity were in the protective direction for most types of breast cancer; reporting more frequent vigorous activity was associated with reduced risk of pre/perimenopausal breast cancer. Women with genetic variants predisposing them to more sedentary time had higher risk of HR− breast cancer, but there was no strong evidence of differences in association by subtypes and weak evidence of an increased risk overall.

Strengths and limitations

A strength of our study is the use of individual-level BCAC data, which permitted examination of more outcomes than previously possible. Large sample sizes are another strength. BCAC is the largest collaboration of breast cancer studies, and we employed the most powerful available genetic instruments identified by the largest GWAS for movement-related behaviours, likely improving precision of our estimates. While statistical power was limited by the limited proportion of variation in exposure explained by the genetic instruments available (we had 52% power to detect expected effects for overall activity and overall breast cancer risk, and less power for other exposure/outcome combinations; Table S13), there were no larger datasets available to increase power. The UK Biobank studies are the only GWAS of accelerometer-assessed movement, which substantially decreases measurement error compared to self-report. Measurement error in assessing genotype is typically very low (often estimated as less than 1% (29, 30)).

The UK Biobank GWAS which identified our instruments used wrist-worn accelerometers, which may not capture ambulation as well as hip-worn accelerometers;(31) while this may have slightly affected precision, no superior data are available. Gene-exposure associations were estimated from a population (UK Biobank) including men, but no strong evidence of sexual dimorphism was reported in UK Biobank,(9) so we assume that SNP-exposure estimates adequately reflect associations in women. While our instruments predict only a small fraction of variance in exposure, any weak-instrument bias would have biased estimates towards the null and cannot explain our findings.(19) Some contributing studies within BCAC did not provide sufficient data on cancer diagnosis to classify cases into case groups (for example tumour subtype or stage), and therefore numbers (32)included in these analyses were much lower. Women without these tumour-specific outcome data may have differed from those included in analyses. Our analyses took a conventional approach of assuming linearity in SNP-exposure and SNP-outcome relationships. Satisfying this assumption is not required for valid causal inference, so even in the presence of nonlinearity our results would still provide information on probable causality, approximating an population-average causal effect of intervening on the exposure.(32–34)

Due to the nature of the data and study design, we estimated odds ratios as the measure of effect, which in some circumstances can be prone to non-collapsibility and sparse-data bias.(35, 36) These issues are most severe when many covariates are included in models (which was not the case for the current analysis), and when outcomes are neither rare nor very common (many of the outcomes we investigated are rare, limiting the extent of noncollapsibility). Overall activity and sedentary time results for pre/peri- and postmenopausal breast cancer (the only sub-outcome where all participants could be classified), demonstrate a slight pattern of noncollapsibility, where the odds ratio for all invasive breast cancers does not lie between the odds ratios for each group separately. This is not a bias but a mathematical property of odds ratios.(35)

Implications

This analysis extends findings from a recent MR study of overall physical activity and breast cancer risk overall and by ER-status, using BCAC summary data.(8) Our study, using individual-level data, confirmed those findings, and showed that the risk reduction holds across multiple subtypes. Our study also examined vigorous activity and sedentary time, not previously studied in relation to breast cancer risk using MR. We assessed associations with multiple outcomes (overall and by case-group) and our results may be subject to false positives. There was no strong evidence of differences in association by case-group.

While MR may provide estimates which more closely reflect underlying causal relationships, core assumptions must be satisfied before causal conclusions can be drawn. We satisfied the first (instrument predicts exposure) by selecting genome-wide significant SNPs identified by the largest GWAS of our traits of interest. We maximised the possibility of meeting the second (no confounding) by checking whether the SNPs were reported in prior GWAS of possible confounders (known breast cancer risk factors), and confirming that results remained consistent after excluding any SNPs that were (e.g., smoking [vigorous activity analyses], education [sedentary behaviour analyses]). We interrogated the third assumption (instrument influences outcome only through exposure) using several pleiotropy-detection approaches, acting on detected violations, and confirming consistency of results from methods relaxing this assumption. Our conclusions remained unchanged following exclusion of potentially-pleiotropic SNPs.

Several SNPs in the analyses were associated with adiposity in previous GWAS. While we cannot rule out horizontal pleiotropy (SNPs influencing adiposity independently of physical activity/sedentary time), vertical pleiotropy (same causal pathway) is more plausible; reduced adiposity is a downstream effect of increased physical activity. Vertical pleiotropy does not violate MR assumptions and excluding vertically-pleiotropic variants may distort causal estimates.(19) Nevertheless, previous MR analysis has shown evidence of a bi-directional relationship between overall activity and adiposity.(9)

Although it is possible that our findings arose by chance, our results for physical activity are consistent with observational studies, which have suggested a 20–25% breast cancer risk reduction for the most vs. least active women, with evidence of dose-response.(3, 37) Our findings support this and furthermore suggest that these relationships are likely to be causal. The observational evidence for risk reduction, particularly for premenopausal breast cancer, is strongest for vigorous physical activity, suggesting that vigorous activity may be particularly important in preventing carcinogenesis.(3, 38) Short bouts of intense activity may be more protective than equivalent energy expenditure accumulated from light activity. We found that self-reported vigorous activity was associated with lower pre/perimenopausal breast cancer risk and found weak evidence for a protective effect of vigorous activity overall. Future studies should continue to explore this with more powerful instruments.

For sedentary time, the observational evidence is sparse and inconsistent. Our results, which minimise likelihood of confounding (e.g. by unhealthy diet), are suggestive of a causal association with elevated risk of breast cancer, particularly for HR− and in situ cancer. While there is debate about the independence of physical activity and sedentary behaviour, they have different determinants and correlates and are often treated as separate traits. In our study the genetic instruments for sedentary behaviour and physical activity were mostly distinct; removing one SNP which predicted both traits did not change our findings, suggesting that both behaviours independently influence breast cancer risk.

Robust causal inference should triangulate findings across methods.(39) Our findings must be considered in light of biological plausibility. A reasonable body of mechanistic evidence supports numerous causal pathways between physical activity and breast cancer risk. Pathways involving adiposity, metabolic dysfunction, sex hormones, and inflammation have been most thoroughly described.(40–42) Mechanisms linking sedentary time and cancer are likely to at least partially overlap with those underpinning the physical activity relationship.(43, 44) Our findings cannot shed light on drivers of carcinogenesis. We saw suggestive differences by HR-status, but this may be a chance finding. Known adiposity-related SNPs did not seem to unduly influence our results, perhaps indicating that multiple pathways are important.

Conclusion

Increasing physical activity and reducing sedentary time are already recommended for cancer prevention. Our study adds further evidence that such behavioural changes are likely to lower future breast cancer incidence. A stronger cancer-control focus on physical activity and sedentary time as modifiable cancer risk factors is warranted, given the heavy burden of disease attributed to the most common cancer in women.

Supplementary Material

Supp1

KEY MESSAGES.

What is already known on this topic:

  • Observational studies have reported that active lifestyles are associated with lower breast cancer risk, but whether activity is the protective (causative) factor cannot be conclusively determined from observational evidence.

What this study adds:

  • This study, using individual-level data from the Breast Cancer Association Consortium, provides strong evidence that greater levels of physical activity and less sedentary time are likely to reduce breast cancer risk, with results generally consistent across breast cancer subtypes.

  • A systematic Mendelian randomization approach enhanced the ability to draw causal conclusions by minimising the effect of biases such as confounding, which are likely to have affected previous studies.

How this study might affect research, practice or policy:

  • Upon triangulating multiple evidence types, there is now robust evidence that insufficiently active lifestyles are a modifiable cause of breast cancer risk, and a stronger focus on promoting active lifestyles is likely to reduce the high burden from breast cancer.

  • It would be of public health benefit for physical activity researchers to establish whether Mendelian randomization supports the observational findings regarding active lifestyles and cancer risk for other cancer types.

Acknowledgements

BCAC: We thank all the individuals who took part in these studies and all the researchers, clinicians, technicians and administrative staff who have enabled this work to be carried out. ABCFS thank Maggie Angelakos, Judi Maskiell, Gillian Dite. ABCS thanks the Blood bank Sanquin, The Netherlands. ABCTB Investigators: Christine Clarke, Deborah Marsh, Rodney Scott, Robert Baxter, Desmond Yip, Jane Carpenter, Alison Davis, Nirmala Pathmanathan, Peter Simpson, J. Dinny Graham, Mythily Sachchithananthan. Samples are made available to researchers on a non-exclusive basis. BBCS thanks Eileen Williams, Elaine Ryder-Mills, Kara Sargus. BCEES thanks Allyson Thomson, Christobel Saunders, Terry Slevin, BreastScreen Western Australia, Elizabeth Wylie, Rachel Lloyd. The BCINIS study would not have been possible without the contributions of Dr. K. Landsman, Dr. N. Gronich, Dr. A. Flugelman, Dr. W. Saliba, Dr. F. Lejbkowicz, Dr. E. Liani, Dr. I. Cohen, Dr. S. Kalet, Dr. V. Friedman, Dr. O. Barnet of the NICCC in Haifa, and all the contributing family medicine, surgery, pathology and oncology teams in all medical institutes in Northern Israel. The BREOGAN study would not have been possible without the contributions of the following: Manuela Gago-Dominguez, Jose Esteban Castelao, Angel Carracedo, Victor Muñoz Garzón, Alejandro Novo Domínguez, Maria Elena Martinez, Sara Miranda Ponte, Carmen Redondo Marey, Maite Peña Fernández, Manuel Enguix Castelo, Maria Torres, Manuel Calaza (BREOGAN), José Antúnez, Máximo Fraga and the staff of the Department of Pathology and Biobank of the University Hospital Complex of Santiago-CHUS, Instituto de Investigación Sanitaria de Santiago, IDIS, Xerencia de Xestion Integrada de Santiago-SERGAS; Joaquín González-Carreró and the staff of the Department of Pathology and Biobank of University Hospital Complex of Vigo, Instituto de Investigacion Biomedica Galicia Sur, SERGAS, Vigo, Spain. The BSUCH study acknowledges the Principal Investigator, Barbara Burwinkel, and thanks Peter Bugert, Medical Faculty Mannheim. CBCS thanks study participants, co-investigators, collaborators and staff of the Canadian Breast Cancer Study, and project coordinators Agnes Lai and Celine Morissette. CCGP thanks Styliani Apostolaki, Anna Margiolaki, Georgios Nintos, Maria Perraki, Georgia Saloustrou, Georgia Sevastaki, Konstantinos Pompodakis. CGPS thanks staff and participants of the Copenhagen General Population Study. For the excellent technical assistance: Dorthe Uldall Andersen, Maria Birna Arnadottir, Anne Bank, Dorthe Kjeldgård Hansen. The Danish Cancer Biobank is acknowledged for providing infrastructure for the collection of blood samples for the cases. Investigators from the CPSII cohort thank the participants and Study Management Group for their invaluable contributions to this research. They also acknowledge the contribution to this study from central cancer registries supported through the Centers for Disease Control and Prevention National Program of Cancer Registries, as well as cancer registries supported by the National Cancer Institute Surveillance Epidemiology and End Results program. The authors would like to thank the California Teachers Study Steering Committee that is responsible for the formation and maintenance of the Study within which this research was conducted. A full list of California Teachers Study (CTS) team members is available at https://www.calteachersstudy.org/team. DIETCOMPLYF thanks the patients, nurses and clinical staff involved in the study. The DietCompLyf study was funded by the charity Against Breast Cancer (Registered Charity Number 1121258) and the NCRN. We thank the participants and the investigators of EPIC (European Prospective Investigation into Cancer and Nutrition). ESTHER thanks Hartwig Ziegler, Sonja Wolf, Volker Hermann, Christa Stegmaier, Katja Butterbach. FHRISK and PROCAS thank NIHR for funding. GC-HBOC thanks Stefanie Engert, Heide Hellebrand, Sandra Kröber and LIFE - Leipzig Research Centre for Civilization Diseases (Markus Loeffler, Joachim Thiery, Matthias Nüchter, Ronny Baber). The GENICA Network: Dr. Margarete Fischer-Bosch-Institute of Clinical Pharmacology, Stuttgart, and University of Tübingen, Germany [Hiltrud Brauch, Wing-Yee Lo], German Cancer Consortium (DKTK) and German Cancer Research Center (DKFZ), Partner Site Tübingen, 72074 Tübingen, Germany [Hiltrud Brauch], gefördert durch die Deutsche Forschungsgemeinschaft (DFG) im Rahmen der Exzellenzstrategie des Bundes und der Länder - EXC 2180 – 390900677 [Hiltrud Brauch], Department of Internal Medicine, Johanniter GmbH Bonn, Johanniter Krankenhaus, Bonn, Germany [Yon-Dschun Ko, Christian Baisch], Institute of Pathology, University of Bonn, Germany [Hans-Peter Fischer], Molecular Genetics of Breast Cancer, Deutsches Krebsforschungszentrum (DKFZ), Heidelberg, Germany [Ute Hamann], Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr University Bochum (IPA), Bochum, Germany [Thomas Brüning, Beate Pesch, Sylvia Rabstein, Anne Lotz]; and Institute of Occupational Medicine and Maritime Medicine, University Medical Center Hamburg-Eppendorf, Germany [Volker Harth]. HEBCS thanks Johanna Kiiski, Taru A. Muranen, Kristiina Aittomäki, Kirsimari Aaltonen, Karl von Smitten, Irja Erkkilä. HMBCS thanks Peter Hillemanns, Hans Christiansen and Johann H. Karstens. HUBCS thanks Darya Prokofyeva and Shamil Gantsev. KARMA and SASBAC thank the Swedish Medical Research Council. KBCP thanks Eija Myöhänen. LMBC thanks Gilian Peuteman, Thomas Van Brussel, EvyVanderheyden and Kathleen Corthouts. MABCS thanks Milena Jakimovska (RCGEB “Georgi D. Efremov”), Snezhana Smichkoska, Emilija Lazarova, Marina Iljoska (University Clinic of Radiotherapy and Oncology), Dzengis Jasar, Mitko Karadjozov (Adzibadem-Sistina Hospital), Andrej Arsovski and Liljana Stojanovska (Re-Medika Hospital) for their contributions and commitment to this study. MARIE thanks Petra Seibold, Nadia Obi, Sabine Behrens, Ursula Eilber and Muhabbet Celik. MBCSG (Milan Breast Cancer Study Group): Paolo Peterlongo, Siranoush Manoukian, Bernard Peissel, Jacopo Azzollini, Erica Rosina, Daniela Zaffaroni, Irene Feroce, Mariarosaria Calvello, Aliana Guerrieri Gonzaga, Monica Marabelli, Davide Bondavalli and the personnel of the Cogentech Cancer Genetic Test Laboratory. The MCCS was made possible by the contribution of many people, including the original investigators, the teams that recruited the participants and continue working on follow-up, and the many thousands of Melbourne residents who continue to participate in the study. We thank the coordinators, the research staff and especially the MMHS participants for their continued collaboration on research studies in breast cancer. MSKCC thanks Marina Corines, Lauren Jacobs. MTLGEBCS would like to thank Martine Tranchant (CHU de Québec – Université Laval Research Center), Marie-France Valois, Annie Turgeon and Lea Heguy (McGill University Health Center, Royal Victoria Hospital; McGill University) for DNA extraction, sample management and skilful technical assistance. J.S. is Chair holder of the Canada Research Chair in Oncogenetics. The following are NBCS Collaborators: Kristine K. Sahlberg (PhD), Anne-Lise Børresen-Dale (Prof. Em.), Lars Ottestad (MD), Rolf Kåresen (Prof. Em.), Dr. Ellen Schlichting (MD), Marit Muri Holmen (MD), Toril Sauer (MD), Vilde Haakensen (MD), Olav Engebråten (MD), Bjørn Naume (MD), Alexander Fosså (MD), Cecile E. Kiserud (MD), Kristin V. Reinertsen (MD), Åslaug Helland (MD), Margit Riis (MD), Jürgen Geisler (MD), OSBREAC and Grethe I. Grenaker Alnæs (MSc). NBHS and SBCGS thank study participants and research staff for their contributions and commitment to the studies. For NHS and NHS2 the study protocol was approved by the institutional review boards of the Brigham and Women’s Hospital and Harvard T.H. Chan School of Public Health, and those of participating registries as required. We would like to thank the participants and staff of the NHS and NHS2 for their valuable contributions as well as the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA, WY. The authors assume full responsibility for analyses and interpretation of these data. The OFBCR thanks Teresa Selander, Nayana Weerasooriya and Steve Gallinger. ORIGO thanks E. Krol-Warmerdam, and J. Blom for patient accrual, administering questionnaires, and managing clinical information. The LUMC survival data were retrieved from the Leiden hospital-based cancer registry system (ONCDOC) with the help of Dr. J. Molenaar. PBCS thanks Louise Brinton, Mark Sherman, Neonila Szeszenia-Dabrowska, Beata Peplonska, Witold Zatonski, Pei Chao, Michael Stagner. The ethical approval for the POSH study is MREC /00/6/69, UKCRN ID: 1137. We thank staff in the Experimental Cancer Medicine Centre (ECMC) supported Faculty of Medicine Tissue Bank and the Faculty of Medicine DNA Banking resource. The authors wish to acknowledge the roles of the Breast Cancer Now Tissue Bank in collecting and making available the samples and/or data, and the patients who have generously donated their tissues and shared their data to be used in the generation of this publication. PREFACE thanks Sonja Oeser and Silke Landrith. The RBCS thanks Jannet Blom, Saskia Pelders, Wendy J.C. Prager – van der Smissen, and the Erasmus MC Family Cancer Clinic. SBCS thanks Sue Higham, Helen Cramp, Dan Connley, Ian Brock, Sabapathy Balasubramanian and Malcolm W.R. Reed. We thank the SEARCH and EPIC teams. SKKDKFZS thanks all study participants, clinicians, family doctors, researchers and technicians for their contributions and commitment to this study. We thank the SUCCESS Study teams in Munich, Duessldorf, Erlangen and Ulm. SZBCS thanks Ewa Putresza. UBCS thanks all study participants, the ascertainment, laboratory and research informatics teams at Huntsman Cancer Institute and Intermountain Healthcare, and Justin Williams, Brandt Jones, Myke Madsen, Stacey Knight and Kerry Rowe for their important contributions to this study. UCIBCS thanks Irene Masunaka. UKBGS thanks Breast Cancer Now and the Institute of Cancer Research for support and funding of the Generations Study, and the study participants, study staff, and the doctors, nurses and other health care providers and health information sources who have contributed to the study. We acknowledge NHS funding to the Royal Marsden/ICR NIHR Biomedical Research Centre.

Funding

This work was supported by the following agencies. Funders had no role in study design, data collection, analysis, interpretation, writing of the report, or the decision to submit the paper for publication.

BCAC is funded by the European Union’s Horizon 2020 Research and Innovation Programme (grant numbers 634935 and 633784 for BRIDGES and B-CAST respectively), and the PERSPECTIVE I&I project, funded by the Government of Canada through Genome Canada and the Canadian Institutes of Health Research, the Ministère de l’Économie et de l’Innovation du Québec through Genome Québec, the Quebec Breast Cancer Foundation. The EU Horizon 2020 Research and Innovation Programme funding source had no role in study design, data collection, data analysis, data interpretation or writing of the report. Additional funding for BCAC is provided via the Confluence project which is funded with intramural funds from the National Cancer Institute Intramural Research Program, National Institutes of Health.

Genotyping of the OncoArray was funded by the NIH Grant U19 CA148065, and Cancer Research UK Grant C1287/A16563 and the PERSPECTIVE project supported by the Government of Canada through Genome Canada and the Canadian Institutes of Health Research (grant GPH-129344) and, the Ministère de l’Économie, Science et Innovation du Québec through Genome Québec and the PSRSIIRI-701 grant, and the Quebec Breast Cancer Foundation. Funding for iCOGS came from: the European Community’s Seventh Framework Programme under grant agreement n° 223175 (HEALTH-F2-2009-223175) (COGS), Cancer Research UK (C1287/A10118, C1287/A10710, C12292/A11174, C1281/A12014, C5047/A8384, C5047/A15007, C5047/A10692, C8197/A16565), the National Institutes of Health (CA128978) and Post-Cancer GWAS initiative (1U19 CA148537, 1U19 CA148065 and 1U19 CA148112 - the GAME-ON initiative), the Department of Defence (W81XWH-10-1-0341), the Canadian Institutes of Health Research (CIHR) for the CIHR Team in Familial Risks of Breast Cancer, and Komen Foundation for the Cure, the Breast Cancer Research Foundation, and the Ovarian Cancer Research Fund.

The BRIDGES panel sequencing was supported by the European Union Horizon 2020 research and innovation program BRIDGES (grant number, 634935) and the Wellcome Trust (v203477/Z/16/Z).

The Australian Breast Cancer Family Study (ABCFS) was supported by grant UM1 CA164920 from the National Cancer Institute (USA). The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the Breast Cancer Family Registry (BCFR), nor does mention of trade names, commercial products, or organizations imply endorsement by the USA Government or the BCFR. The ABCFS was also supported by the National Health and Medical Research Council of Australia, the New South Wales Cancer Council, the Victorian Health Promotion Foundation (Australia) and the Victorian Breast Cancer Research Consortium. J.L.H. is a National Health and Medical Research Council (NHMRC) Senior Principal Research Fellow. M.C.S. is a NHMRC Senior Research Fellow. The ABCS study was supported by the Dutch Cancer Society [grants NKI 2007–3839; 2009 4363]. The Australian Breast Cancer Tissue Bank (ABCTB) was supported by the National Health and Medical Research Council of Australia, The Cancer Institute NSW and the National Breast Cancer Foundation. The AHS study is supported by the intramural research program of the National Institutes of Health, the National Cancer Institute (grant number Z01-CP010119), and the National Institute of Environmental Health Sciences (grant number Z01-ES049030). The work of the BBCC was partly funded by ELAN-Fond of the University Hospital of Erlangen. The BBCS is funded by Cancer Research UK and Breast Cancer Now and acknowledges NHS funding to the NIHR Biomedical Research Centre, and the National Cancer Research Network (NCRN). The BCEES was funded by the National Health and Medical Research Council, Australia and the Cancer Council Western Australia and acknowledges funding from the National Breast Cancer Foundation (JS). For the BCFR-NY, BCFR-PA, BCFR-UT this work was supported by grant UM1 CA164920 from the National Cancer Institute. The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the Breast Cancer Family Registry (BCFR), nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government or the BCFR. The BCINIS study is supported in part by the Breast Cancer Research Foundation (BCRF). The BREast Oncology GAlician Network (BREOGAN) is funded by Acción Estratégica de Salud del Instituto de Salud Carlos III FIS PI12/02125/Cofinanciado and FEDER PI17/00918/Cofinanciado FEDER; Acción Estratégica de Salud del Instituto de Salud Carlos III FIS Intrasalud (PI13/01136); Programa Grupos Emergentes, Cancer Genetics Unit, Instituto de Investigacion Biomedica Galicia Sur. Xerencia de Xestion Integrada de Vigo-SERGAS, Instituto de Salud Carlos III, Spain; Grant 10CSA012E, Consellería de Industria Programa Sectorial de Investigación Aplicada, PEME I + D e I + D Suma del Plan Gallego de Investigación, Desarrollo e Innovación Tecnológica de la Consellería de Industria de la Xunta de Galicia, Spain; Grant EC11-192. Fomento de la Investigación Clínica Independiente, Ministerio de Sanidad, Servicios Sociales e Igualdad, Spain; and Grant FEDER-Innterconecta. Ministerio de Economia y Competitividad, Xunta de Galicia, Spain. The BSUCH study was supported by the Dietmar-Hopp Foundation, the Helmholtz Society and the German Cancer Research Center (DKFZ). CBCS is funded by the Canadian Cancer Society (grant # 313404) and the Canadian Institutes of Health Research. CCGP is supported by funding from the University of Crete. The CECILE study was supported by Fondation de France, Institut National du Cancer (INCa), Ligue Nationale contre le Cancer, Agence Nationale de Sécurité Sanitaire, de l’Alimentation, de l’Environnement et du Travail (ANSES), Agence Nationale de la Recherche (ANR). The CGPS was supported by the Chief Physician Johan Boserup and Lise Boserup Fund, the Danish Medical Research Council, and Herlev and Gentofte Hospital. The American Cancer Society funds the creation, maintenance, and updating of the CPS-II cohort. The California Teachers Study (CTS) and the research reported in this publication were supported by the National Cancer Institute of the National Institutes of Health under award number U01-CA199277; P30-CA033572; P30-CA023100; UM1-CA164917; and R01-CA077398. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health. The collection of cancer incidence data used in the California Teachers Study was supported by the California Department of Public Health pursuant to California Health and Safety Code Section 103885; Centers for Disease Control and Prevention’s National Program of Cancer Registries, under cooperative agreement 5NU58DP006344; the National Cancer Institute’s Surveillance, Epidemiology and End Results Program under contract HHSN261201800032I awarded to the University of California, San Francisco, contract HHSN261201800015I awarded to the University of Southern California, and contract HHSN261201800009I awarded to the Public Health Institute. The opinions, findings, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the official views of the State of California, Department of Public Health, the National Cancer Institute, the National Institutes of Health, the Centers for Disease Control and Prevention or their Contractors and Subcontractors, or the Regents of the University of California, or any of its programs. The University of Westminster curates the DietCompLyf database funded by Against Breast Cancer Registered Charity No. 1121258 and the NCRN. 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: Ligue Contre le Cancer, Institut Gustave Roussy, Mutuelle Générale de l’Education Nationale, Institut National de la Santé et de la Recherche Médicale (INSERM) (France); German Cancer Aid, German Cancer Research Center (DKFZ), Federal Ministry of Education and Research (BMBF) (Germany); the Hellenic Health Foundation, the Stavros Niarchos Foundation (Greece); Associazione Italiana per la Ricerca sul Cancro-AIRC-Italy 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); Health Research Fund (FIS), PI13/00061 to Granada, PI13/01162 to EPIC-Murcia, Regional Governments of Andalucía, Asturias, Basque Country, Murcia and Navarra, ISCIII RETIC (RD06/0020) (Spain); Cancer Research UK (14136 to EPIC-Norfolk; C570/A16491 and C8221/A19170 to EPIC-Oxford), Medical Research Council (1000143 to EPIC-Norfolk, MR/M012190/1 to EPIC-Oxford) (United Kingdom). The ESTHER study was supported by a grant from the Baden Württemberg Ministry of Science, Research and Arts. Additional cases were recruited in the context of the VERDI study, which was supported by a grant from the German Cancer Aid (Deutsche Krebshilfe). FHRISK and PROCAS are funded from NIHR grant PGfAR 0707–10031. DGE, AH and WGN are supported by the NIHR Manchester Biomedical Research Centre (IS-BRC-1215-20007). The GC-HBOC (German Consortium of Hereditary Breast and Ovarian Cancer) is supported by the German Cancer Aid (grant no 110837 and 70114178, coordinator: Rita K. Schmutzler, Cologne) and the Federal Ministry of Education and Research, Germany (grant no 01GY1901). This work was also funded by the European Regional Development Fund and Free State of Saxony, Germany (LIFE - Leipzig Research Centre for Civilization Diseases, project numbers 713–241202, 713–241202, 14505/2470, 14575/2470). The GENICA was funded by the Federal Ministry of Education and Research (BMBF) Germany grants 01KW9975/5, 01KW9976/8, 01KW9977/0 and 01KW0114, the Robert Bosch Foundation, Stuttgart, Deutsches Krebsforschungszentrum (DKFZ), Heidelberg, the Institute for Prevention and Occupational Medicine of the German Social Accident Insurance, Institute of the Ruhr University Bochum (IPA), Bochum, as well as the Department of Internal Medicine, Johanniter GmbH Bonn, Johanniter Krankenhaus, Bonn, Germany. The GEPARSIXTO study was conducted by the German Breast Group GmbH. The GESBC was supported by the Deutsche Krebshilfe e. V. [70492] and the German Cancer Research Center (DKFZ). The HABCS study was supported by the Claudia von Schilling Foundation for Breast Cancer Research, by the Lower Saxonian Cancer Society, and by the Rudolf Bartling Foundation. The HEBCS was financially supported by the Helsinki University Hospital Research Fund, the Sigrid Juselius Foundation and the Cancer Foundation Finland. The HMBCS was supported by a grant from the Friends of Hannover Medical School and by the Rudolf Bartling Foundation. The HUBCS was supported by a grant from the German Federal Ministry of Research and Education (RUS08/017), B.M. was supported by grant 17-44-020498, 17-29-06014 of the Russian Foundation for Basic Research, and the study was performed as part of the assignment of the Ministry of Science and Higher Education of the Russian Federation (№AAAA-A16-116020350032-1). Financial support for KARBAC was provided through the regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institutet, the Swedish Cancer Society, The Gustav V Jubilee foundation and Bert von Kantzows foundation. The KARMA study was supported by Märit and Hans Rausings Initiative Against Breast Cancer. The KBCP was financially supported by the special Government Funding (VTR) of Kuopio University Hospital grants, Cancer Fund of North Savo, the Finnish Cancer Organizations, and by the strategic funding of the University of Eastern Finland. LMBC is supported by the ‘Stichting tegen Kanker’. DL is supported by the FWO. The MABCS study is funded by the Research Centre for Genetic Engineering and Biotechnology “Georgi D. Efremov”, MASA. The MARIE study was supported by the Deutsche Krebshilfe e.V. [70-2892-BR I, 106332, 108253, 108419, 110826, 110828], the Hamburg Cancer Society, the German Cancer Research Center (DKFZ) and the Federal Ministry of Education and Research (BMBF) Germany [01KH0402]. MBCSG is supported by grants from the Italian Association for Cancer Research (AIRC). The MCBCS was supported by the NIH grants R35CA253187, R01CA192393, R01CA116167, R01CA176785 a NIH Specialized Program of Research Excellence (SPORE) in Breast Cancer [P50CA116201], and the Breast Cancer Research Foundation. The Melbourne Collaborative Cohort Study (MCCS) cohort recruitment was funded by VicHealth and Cancer Council Victoria. The MCCS was further augmented by Australian National Health and Medical Research Council grants 209057, 396414 and 1074383 and by infrastructure provided by Cancer Council Victoria. Cases and their vital status were ascertained through the Victorian Cancer Registry and the Australian Institute of Health and Welfare, including the National Death Index and the Australian Cancer Database. The MEC was supported by NIH grants CA63464, CA54281, CA098758, CA132839 and CA164973. The MISS study is supported by funding from ERC-2011-294576 Advanced grant, Swedish Cancer Society, Swedish Research Council, Local hospital funds, Berta Kamprad Foundation, Gunnar Nilsson. The MMHS study was supported by NIH grants CA97396, CA128931, CA116201, CA140286 and CA177150. MSKCC is supported by grants from the Breast Cancer Research Foundation and Robert and Kate Niehaus Clinical Cancer Genetics Initiative. The work of MTLGEBCS was supported by the Quebec Breast Cancer Foundation, the Canadian Institutes of Health Research for the “CIHR Team in Familial Risks of Breast Cancer” program – grant # CRN-87521 and the Ministry of Economic Development, Innovation and Export Trade – grant # PSR-SIIRI-701. The NBCS has received funding from the K.G. Jebsen Centre for Breast Cancer Research; the Research Council of Norway grant 193387/V50 (to A-L Børresen-Dale and V.N. Kristensen) and grant 193387/H10 (to A-L Børresen-Dale and V.N. Kristensen), South Eastern Norway Health Authority (grant 39346 to A-L Børresen-Dale) and the Norwegian Cancer Society (to A-L Børresen-Dale and V.N. Kristensen). The NBHS was supported by NIH grant R01CA100374. Biological sample preparation was conducted the Survey and Biospecimen Shared Resource, which is supported by P30 CA68485. The Northern California Breast Cancer Family Registry (NC-BCFR) and Ontario Familial Breast Cancer Registry (OFBCR) were supported by grant U01CA164920 from the USA National Cancer Institute of the National Institutes of Health. The content of this manuscript does not necessarily reflect the views or policies of the National Cancer Institute or any of the collaborating centers in the Breast Cancer Family Registry (BCFR), nor does mention of trade names, commercial products, or organizations imply endorsement by the USA Government or the BCFR. The Carolina Breast Cancer Study (NCBCS) was funded by Komen Foundation, the National Cancer Institute (P50 CA058223, U54 CA156733, U01 CA179715), and the North Carolina University Cancer Research Fund. The NHS was supported by NIH grants P01 CA87969, UM1 CA186107, and U19 CA148065. The NHS2 was supported by NIH grants UM1 CA176726 and U19 CA148065. The ORIGO study was supported by the Dutch Cancer Society (RUL 1997–1505) and the Biobanking and Biomolecular Resources Research Infrastructure (BBMRI-NL CP16). The PBCS was funded by Intramural Research Funds of the National Cancer Institute, Department of Health and Human Services, USA. Genotyping for PLCO was supported by the Intramural Research Program of the National Institutes of Health, NCI, Division of Cancer Epidemiology and Genetics. The PLCO is supported by the Intramural Research Program of the Division of Cancer Epidemiology and Genetics and supported by contracts from the Division of Cancer Prevention, National Cancer Institute, National Institutes of Health. The POSH study is funded by Cancer Research UK (grants C1275/A11699, C1275/C22524, C1275/A19187, C1275/A15956 and Breast Cancer Campaign 2010PR62, 2013PR044. The RBCS was funded by the Dutch Cancer Society (DDHK 2004–3124, DDHK 2009–4318). The SBCS was supported by Sheffield Experimental Cancer Medicine Centre and Breast Cancer Now Tissue Bank. SEARCH is funded by Cancer Research UK [C490/A10124, C490/A16561] and supported by the UK National Institute for Health Research Biomedical Research Centre at the University of Cambridge. The University of Cambridge has received salary support for PDPP from the NHS in the East of England through the Clinical Academic Reserve. The Sister Study (SISTER) is supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (Z01-ES044005 and Z01-ES049033). The Two Sister Study (2SISTER) was supported by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (Z01-ES044005 and Z01-ES102245), and, also by a grant from Susan G. Komen for the Cure, grant FAS0703856. SKKDKFZS is supported by the DKFZ. The SMC is funded by the Swedish Cancer Foundation and the Swedish Research Council (VR 2017–00644) grant for the Swedish Infrastructure for Medical Population-based Life-course Environmental Research (SIMPLER). The SZBCS was supported by Grant PBZ_KBN_122/P05/2004 and the program of the Minister of Science and Higher Education under the name “Regional Initiative of Excellence” in 2019-2022 project number 002/RID/2018/19 amount of financing 12 000 000 PLN. The TNBCC was supported by: a Specialized Program of Research Excellence (SPORE) in Breast Cancer (CA116201), a grant from the Breast Cancer Research Foundation, a generous gift from the David F. and Margaret T. Grohne Family Foundation. UBCS was supported by funding from National Cancer Institute (NCI) grant R01 CA163353 (to N.J. Camp) and the Women’s Cancer Center at the Huntsman Cancer Institute (HCI). Data collection for UBCS was supported by the Utah Population Database (UPDB) and Utah Cancer Registry (UCR). The UPDB is supported by HCI (including the Huntsman Cancer Foundation), University of Utah program in Personalized Health and Center for Clinical and Translational Science, and NCI grant P30 CA42014. The UCR is funded by the NCI’s SEER Program, Contract No. HHSN261201800016I, the US Center for Disease Control and Prevention’s National Program of Cancer Registries, Cooperative Agreement No. NU58DP0063200, the University of Utah and Huntsman Cancer Foundation. The UCIBCS component of this research was supported by the NIH [CA58860, CA92044] and the Lon V Smith Foundation [LVS39420]. The UKBGS is funded by Breast Cancer Now and the Institute of Cancer Research (ICR), London. ICR acknowledges NHS funding to the NIHR Biomedical Research Centre. The UKOPS study was funded by The Eve Appeal (The Oak Foundation) and supported by the National Institute for Health Research University College London Hospitals Biomedical Research Centre. The USRT Study was funded by Intramural Research Funds of the National Cancer Institute, Department of Health and Human Services, USA.

RMM is a National Institute for Health Research Senior Investigator (NIHR202411). RMM is supported by a Cancer Research UK (C18281/A19169) programme grant (the Integrative Cancer Epidemiology Programme). RMM is also supported by the NIHR Bristol Biomedical Research Centre which is funded by the NIHR and is a partnership between University Hospitals Bristol and Weston NHS Foundation Trust and the University of Bristol. RMM is affiliated with the Medical Research Council Integrative Epidemiology Unit at the University of Bristol which is supported by the Medical Research Council (MC_UU_00011/1, MC_UU_00011/3, MC_UU_00011/6, and MC_UU_00011/4) and the University of Bristol. Department of Health and Social Care disclaimer: The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care. DGE, AH and WGN are supported by the NIHR Manchester Biomedical Research Centre (IS-BRC-1215-20007). BML is funded by the Victorian Cancer Agency (MCRF-18005).

Footnotes

Competing interests

Matthias W. Beckmann conducts research funded by Amgen, Novartis and Pfizer. Peter A. Fasching conducts research funded by Amgen, Novartis and Pfizer. He received honoraria from Roche, Novartis and Pfizer. Allison W. Kurian declares research funding to her institution from Myriad Genetics for an unrelated project (funding dates 2017–2019). Sibylle Loibl declares grants and honoraria paid to her institution from Amgen, Novartis, Pfizer, Roche, and, outside the submitted work, grants and/or honoraria paid to her institution from AbbVie, Celgene, Seattle Genetics, PrIME/Medscape, Daiichi-Sankyo, Lilly, Samsung, BMS, Puma, Immunomedics, AstraZeneca, Pierre Fabre, Merck, GlaxoSmithKlein, EirGenix, and Bayer, and personal fees from Chugai; Dr. Loibl also has a patent EP14153692.0 pending. Usha Menon declares stock ownership in Abcodia Ltd. Rachel A. Murphy has been a consultant for Pharmavite. No other authors have conflicts to declare.

Ethics approval

This analysis and each contributing study received approval from the appropriate institutional review board or committee.

Patient involvement

Patient co-production was not adopted for this large multi-study analysis. We thank all participants for providing their data to the contributing BCAC studies.

Data sharing statement

The data used in this study are de-identified patient data from 76 studies participating in the Breast Cancer Association Consortium (BCAC). Enquiries about accessing BCAC data can be directed to the BCAC coordinators at the University of Cambridge: https://bcac.ccge.medschl.cam.ac.uk/

References

  • 1.International Agency for Research on Cancer. Weight Control and Physical Activity. Lyon; 2002. [Google Scholar]
  • 2.Physical Activity Guidelines Advisory Committee. Physical Activity Guidelines Advisory Committee Report, 2008. Washington, DC: U.S. Department of Health and Human Services; 2008. [DOI] [PubMed] [Google Scholar]
  • 3.World Cancer Research Fund International / American Institute for Cancer Research. Continuous Update Project Report: Diet, Nutrition, Physical Activity and Breast Cancer. 2017.
  • 4.Lynch BM, Mahmood S, T B. Chapter 10: Sedentary Behaviour and Cancer. In: Leitzmann M, Jochem C, Schmid D, editors. Sedentary Behaviour Epidemiology. Springer Series on Epidemiology and Public Health. Cham, Switzerland: Springer International Publishing; 2018. p. 245–98. [Google Scholar]
  • 5.Chong F, Wang Y, Song M, Sun Q, Xie W, Song C. Sedentary behavior and risk of breast cancer: a dose-response meta-analysis from prospective studies. Breast cancer (Tokyo, Japan). 2020. [DOI] [PubMed] [Google Scholar]
  • 6.Davey Smith G, Hemani G. Mendelian randomization: Genetic anchors for causal inference in epidemiological studies. Human molecular genetics. 2014;23(R1):R89–R98. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Greenland S. An introduction to instrumental variables for epidemiologists. Int J Epidemiol. 2000;29(4):722–9. [DOI] [PubMed] [Google Scholar]
  • 8.Papadimitriou N, Dimou N, Tsilidis KK, Banbury B, Martin RM, Lewis SJ, et al. Physical activity and risks of breast and colorectal cancer: A Mendelian randomisation analysis. Nature communications. 2020;11(1):597. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Doherty A, Smith-Byrne K, Ferreira T, Holmes MV, Holmes C, Pulit SL, et al. GWAS identifies 14 loci for device-measured physical activity and sleep duration. Nature communications. 2018;9(1):5257. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Klimentidis YC, Raichlen DA, Bea J, Garcia DO, Wineinger NE, Mandarino LJ, et al. Genome-wide association study of habitual physical activity in over 377,000 UK Biobank participants identifies multiple variants including CADM2 and APOE. Int J Obes. 2018;42(6):1161–76. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Choi KW, Chen CY, Stein MB, Klimentidis YC, Wang MJ, Koenen KC, et al. Assessment of bidirectional relationships between physical activity and depression among adults: A 2-sample Mendelian randomization study. JAMA psychiatry. 2019;76(4):399–408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Doherty A, Jackson D, Hammerla N, Plötz T, Olivier P, Granat MH, et al. Large scale population assessment of physical activity using wrist worn accelerometers: The UK Biobank study. PLoS One. 2017;12(2):e0169649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Willetts M, Hollowell S, Aslett L, Holmes C, Doherty A. Statistical machine learning of sleep and physical activity phenotypes from sensor data in 96,220 UK Biobank participants. Scientific reports. 2018;8(1):7961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Hildebrand M, VANH VT, Hansen BH, Ekelund U. Age group comparability of raw accelerometer output from wrist- and hip-worn monitors. Medicine and science in sports and exercise. 2014;46(9):1816–24. [DOI] [PubMed] [Google Scholar]
  • 15.Amos CI, Dennis J, Wang Z, Byun J, Schumacher FR, Gayther SA, et al. The OncoArray consortium: A network for understanding the genetic architecture of common cancers. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 2017;26(1):126–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Ehret GB, Munroe PB, Rice KM, Bochud M, Johnson AD, Chasman DI, et al. Genetic variants in novel pathways influence blood pressure and cardiovascular disease risk. Nature. 2011;478(7367):103–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Burgess S, Butterworth A, Thompson SG. Mendelian randomization analysis with multiple genetic variants using summarized data. Genetic epidemiology. 2013;37(7):658–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Yavorska OO, Burgess S. MendelianRandomization: an R package for performing Mendelian randomization analyses using summarized data. Int J Epidemiol. 2017:1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Burgess S, Davey Smith G, Davies N, Dudbridge F, Gill D, Glymour MM, et al. Guidelines for performing Mendelian randomization investigations [version 2]. Wellcome Open Res 2020;4:186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator. Genetic epidemiology. 2016;40(4):304–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Bowden J, Davey Smith G, Burgess S. Mendelian randomization with invalid instruments: Effect estimation and bias detection through Egger regression. Int J Epidemiol. 2015;44(2):512–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Verbanck M, Chen CY, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nature genetics. 2018;50(5):693–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, et al. The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic acids research. 2019;47(D1):D1005–D12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Staley JR, Blackshaw J, Kamat MA, Ellis S, Surendran P, Sun BB, et al. PhenoScanner: A database of human genotype-phenotype associations. Bioinformatics (Oxford, England). 2016;32(20):3207–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Kamat MA, Blackshaw JA, Young R, Surendran P, Burgess S, Danesh J, et al. PhenoScanner V2: An expanded tool for searching human genotype-phenotype associations. Bioinformatics (Oxford, England). 2019;35(22):4851–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Brion M-JA, Shakhbazov K, Visscher PM. Calculating statistical power in Mendelian randomization studies. International journal of epidemiology. 2012;42(5):1497–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Neale B. UK Biobank GWAS round 2 2018. [Available from: http://www.nealelab.is/uk-biobank/.
  • 28.Okbay A, Beauchamp JP, Fontana MA, Lee JJ, Pers TH, Rietveld CA, et al. Genome-wide association study identifies 74 loci associated with educational attainment. Nature. 2016;533(7604):539–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Wall JD, Tang LF, Zerbe B, Kvale MN, Kwok PY, Schaefer C, et al. Estimating genotype error rates from high-coverage next-generation sequence data. Genome research. 2014;24(11):1734–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Ross MG, Russ C, Costello M, Hollinger A, Lennon NJ, Hegarty R, et al. Characterizing and measuring bias in sequence data. Genome biology. 2013;14(5):R51–R. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Matthews CE, Keadle SK, Berrigan D, Staudenmayer J, P FS-M, Troiano RP, et al. Influence of accelerometer calibration approach on moderate-vigorous physical activity estimates for adults. Medicine and science in sports and exercise. 2018;50(11):2285–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Burgess S, Small DS, Thompson SG. A review of instrumental variable estimators for Mendelian randomization. Stat Methods Med Res. 2017;26(5):2333–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Burgess S, Thompson SG. Interpreting findings from Mendelian randomization using the MR-Egger method. European journal of epidemiology. 2017;32(5):377–89. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Hernán MA, Robins JM. Instruments for causal inference: an epidemiologist’s dream? Epidemiology. 2006;17(4):360–72. [DOI] [PubMed] [Google Scholar]
  • 35.Greenland S. Noncollapsibility, confounding, and sparse-data bias. Part 1: The oddities of odds. Journal of clinical epidemiology. 2021;138:178–81. [DOI] [PubMed] [Google Scholar]
  • 36.Greenland S. Noncollapsibility, confounding, and sparse-data bias. Part 2: What should researchers make of persistent controversies about the odds ratio? Journal of clinical epidemiology. 2021;139:264–8. [DOI] [PubMed] [Google Scholar]
  • 37.Neilson HK, Farris MS, Stone CR, Vaska MM, Brenner DR, Friedenreich CM. Moderate-vigorous recreational physical activity and breast cancer risk, stratified by menopause status: A systematic review and meta-analysis. Menopause (New York, NY). 2017;24(3):322–44. [DOI] [PubMed] [Google Scholar]
  • 38.McTiernan A, Friedenreich CM, Katzmarzyk PT, Powell KE, Macko R, Buchner D, et al. Physical activity in cancer prevention and survival: A systematic review. Medicine and science in sports and exercise. 2019;51(6):1252–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Munafò MR, Davey Smith G. Robust research needs many lines of evidence. Nature. 2018;553(7689):399–401. [DOI] [PubMed] [Google Scholar]
  • 40.Lynch BM, Neilson HK, Friedenreich CM. Physical activity and breast cancer prevention. In: Courneya KS, Friedenreich CM, editors. Recent Results in Cancer Research. Physical Activity and Cancer. Berlin: Springer-Verlag; 2011. [DOI] [PubMed] [Google Scholar]
  • 41.Lynch B, Leitzmann M. An evaluation of the evidence relating to physical inactivity, sedentary behavior, and cancer incidence and mortality. Curr Epidemiol Rep. 2017;4(3):221–31. [Google Scholar]
  • 42.Neilson HK, Conroy SM, Friedenreich CM. The influence of energetic factors on biomarkers of postmenopausal breast cancer risk. Current nutrition reports. 2014;3:22–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Dempsey PC, Matthews CE, Dashti SG, Doherty AR, Bergouignan A, van Roekel EH, et al. Sedentary behavior and chronic disease: Mechanisms and future directions. Journal of physical activity & health. 2020;17(1):52–61. [DOI] [PubMed] [Google Scholar]
  • 44.Lynch BM. Sedentary behavior and cancer: A systematic review of the literature and proposed biological mechanisms. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 2010;19(11):2691–709. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp1
NIHMS1822689-supplement-Supp1.pdf (1.4MB, pdf)

Data Availability Statement

The data used in this study are de-identified patient data from 76 studies participating in the Breast Cancer Association Consortium (BCAC). Enquiries about accessing BCAC data can be directed to the BCAC coordinators at the University of Cambridge: https://bcac.ccge.medschl.cam.ac.uk/

RESOURCES