Available via license: CC BY 4.0
Content may be subject to copyright.
Postmenopausal Hormone Therapy and Colorectal Cancer Risk by
Molecularly Defined Subtypes and Tumor Location
Julia D Labadie, DVM, MSPH, PhD,
1,2
Tabitha A Harrison, MPH,
1
Barbara Banbury, MS, PhD,
1
Efrat L Amtay, MPH, PhD,
3
Sonja Bernd, PharmD, PhD,
4
Hermann Brenner, MD, MPH,
3,5
Daniel D Buchanan, PhD,
6
Peter T Campbell, MS, PhD,
7
Yin Cao, MPH, ScD,
8,9,10
Andrew T Chan, MD, MPH,
11,12
Jenny Chang-Claude, PhD,
13,14
Dallas English, MSc, PhD,
15,16
Jane C Figueiredo, PhD,
17,18
Steven J Gallinger, MD, MSc,
19
Graham G Giles, MSc, PhD,
15,16,20
Marc J Gunter, PhD,
21
Michael Hoffmeister, MSc, PhD,
3
Li Hsu, PhD,
1,22
Mark A Jenkins, PhD,
16
Yi Lin, MS,
1
Roger L Milne, MSc, PhD,
15,16,20
Victor Moreno, MD, PhD,
23
Neil Murphy, MSc, PhD,
21
Shuji Ogino, MD, PhD, MS,
24
Amanda I Phipps, MPH, PhD,
1,2
Lori C Sakoda, MPH, PhD,
1,25
Martha L Slattery, MPH, PhD,
26
Melissa C Southey, PhD,
15,20,27
Wei Sun, PhD,
1
Stephen N Thibodeau,
PhD,
28
Bethany Van Guelpen, MD, PhD,
29
Syed H Zaidi, PhD,
30
Ulrike Peters, MPH, PhD,
1,2
Polly A Newcomb, MPH, PhD,
1,2,
*
1
Public Health Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA;
2
Department of Epidemiology, University of Washington, Seattle, WA,
USA;
3
Division of Clinical Epidemiology and Aging Research, German Cancer Research Center (DKFZ), Heidelberg, Germany;
4
Division of Cancer Epidemiology and
Genetics, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA;
5
Division of Preventive Oncology, German Cancer Research Center (DKFZ) and
National Center for Tumor Diseases (NCT), Heidelberg, Germany;
6
Colorectal Oncogenomics Group, Department of Clinical Pathology, The University of Melbourne,
Parkville, Victoria, Australia;
7
Behavioral and Epidemiology Research Group, American Cancer Society, Atlanta, GA, USA;
8
Division of Public Health Sciences,
Department of Surgery, Washington University School of Medicine, St Louis, MO, USA;
9
Alvin J. Siteman Cancer Center at Barnes-Jewish Hospital and Washington
University School of Medicine, St Louis, MO, USA;
10
Division of Gastroenterology, Department of Medicine, Washington University School of Medicine, St Louis, MO,
USA;
11
Division of Gastroenterology, Massachusetts General Hospital, Boston, MA, USA;
12
Clinical and Translational Epidemiology Unit, Department of Medicine,
Massachusetts General Hospital, Boston, MA, USA;
13
Division of Cancer Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany;
14
University
Medical Centre Hamburg-Eppendorf, University Cancer Centre Hamburg (UCCH), Hamburg, Germany;
15
Cancer Epidemiology Division, Cancer Council Victoria,
Melbourne, Victoria, Australia;
16
Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, The University of Melbourne,
Melbourne, Victoria, Australia;
17
Department of Medicine, Samuel Oschin Comprehensive Cancer Institute, Cedars-Sinai, Los Angeles, CA, USA;
18
Department of
Preventive Medicine, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA;
19
Lunenfeld Tanenbaum Research Institute, Mount Sinai
Hospital, University of Toronto, Toronto, Ontario, Canada;
20
Precision Medicine, School of Clinical Sciences at Monash Health, Monash University, Clayton, Victoria,
Australia;
21
Nutrition and Metabolism Section, International Agency for Research on Cancer, World Health Organization, Lyon, France;
22
Department of Biostatistics,
University of Washington, Seattle, WA, USA;
23
Oncology Data Analytics Program, Catalan Institute of Oncology-IDIBELL, L’Hospitalet de Llobregat, Barcelona, Spain;
24
Program in MPE Molecular Pathological Epidemiology, Department of Pathology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA;
25
Division of Research, Kaiser Permanente Northern California, Oakland, CA, USA;
26
Department of Internal Medicine, University of Utah, Salt Lake City, UT, USA;
27
Genetic Epidemiology Laboratory, Department of Clinical Pathology, University of Melbourne, Melbourne, Victoria, Australia;
28
Division of Laboratory Genetics,
Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, USA;
29
Department of Radiation Sciences, Oncology Unit, Umea˚ University, Umea˚,
Sweden; Wallenberg Centre for Molecular Medicine, Umea˚ University, Umea˚ , Sweden and
30
Ontario Institute for Cancer Research, Toronto, Ontario, Canada
*Correspondence to: Polly Newcomb, PhD, MPH, Fred Hutchinson Cancer Research Center, Mailstop M4-B402, PO Box 11024, Seattle, WA 98109-1024, USA (e-mail: pnew-
comb@fredhutch.org).
Abstract
Background: Postmenopausal hormone therapy (HT) is associated with a decreased colorectal cancer (CRC) risk. As CRC is a
heterogeneous disease, we evaluated whether the association of HT and CRC differs across etiologically relevant, molecularly
defined tumor subtypes and tumor location. Methods: We pooled data on tumor subtypes (microsatellite instability status,
CpG island methylator phenotype status, BRAF and KRAS mutations, pathway: adenoma-carcinoma, alternate, serrated), tu-
mor location (proximal colon, distal colon, rectum), and HT use among 8220 postmenopausal women (3898 CRC cases and
4322 controls) from 8 observational studies. We used multinomial logistic regression to estimate odds ratios (OR) and 95%
confidence intervals (CIs) for the association of ever vs never HT use with each tumor subtype compared with controls.
Received: 11 March 2020; Revised: 20 April 2020; Accepted: 12 May 2020
©The Author(s) 2020. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/),
which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
1of9
JNCI Cancer Spectrum (2020) 4(5): pkaa042
doi: 10.1093/jncics/pkaa042
First published online 19 May 2020
Article
Models were adjusted for study, age, body mass index, smoking status, and CRC family history. All statistical tests were 2-
sided. Results: Among postmenopausal women, ever HT use was associated with a 38% reduction in overall CRC risk (OR
¼0.62, 95% CI ¼0.56 to 0.69). This association was similar according to microsatellite instability, CpG island methylator
phenotype and BRAF or KRAS status. However, the association was attenuated for tumors arising through the serrated
pathway (OR ¼0.81, 95% CI ¼0.66 to 1.01) compared with the adenoma-carcinoma pathway (OR ¼0.63, 95% CI ¼0.55 to 0.73;
P
het
¼.04) and alternate pathway (OR ¼0.61, 95% CI ¼0.51 to 0.72). Additionally, proximal colon tumors had a weaker associa-
tion (OR ¼0.71, 95% CI ¼0.62 to 0.80) compared with rectal (OR ¼0.54, 95% CI ¼0.46 to 0.63) and distal colon (OR ¼0.57, 95%
CI ¼0.49 to 0.66; P
het
¼.01) tumors. Conclusions: We observed a strong inverse association between HT use and overall CRC
risk, which may predominantly reflect a benefit of HT use for tumors arising through the adenoma-carcinoma and alternate
pathways as well as distal colon and rectal tumors.
Colorectal cancer (CRC) is a heterogeneous disease that evolves
through multiple pathways defined by genetic and epigenetic
events (1,2). Four tumor markers have been commonly used to
better characterize this heterogeneity: microsatellite instability
(MSI), CpG island methylator phenotype (CIMP), somatic muta-
tions in BRAF, and somatic mutations in KRAS. Together, these
tumor markers approximate 3 distinct molecular pathways of
colorectal carcinogenesis: adenoma-carcinoma (traditional), al-
ternate, and serrated (1,3,4). These pathways are established
early in disease pathogenesis and can be identified within pre-
cancerous lesions by microscopy (3,5–8). Research has shown
that these tumor types have distinct appearances, predilections
for locations within the colon, and biologic behaviors (8–10). As
such, it is plausible that the epidemiologic factors underlying
their etiologies could also differ.
Multiple lines of evidence, including randomized controlled
trials, show that postmenopausal hormone therapy (HT) is as-
sociated with a decreased risk of CRC (11–20). The reduction in
risk, about 20%–40% in recent analyses, has been observed in
users of estrogen alone as well as combined estrogen plus pro-
gestin. Few studies have evaluated whether the association of
HT use and CRC risk differs by molecularly defined CRC sub-
types; however, such information might increase the under-
standing of the mechanisms for this beneficial effect. Current
literature suggests that HT use is associated with a lower risk of
MSI-low or microsatellite stable tumors (MSI-L/MSS) and possi-
bly with a lower risk of CIMP-negative and BRAF wild-type
tumors (21,22). HT use has only been associated with KRAS
wild-type tumors in the distal colon in 1 previous study (23).
Regarding tumor location, the association of HT use and CRC is
reportedly stronger among tumors of the distal colon compared
with the proximal colon (22,24). To our knowledge, no study
has evaluated both tumor markers and location in relation to
HT use to provide a comprehensive understanding of subtype-
specific CRC risk.
In this study, we examined HT use in relation to molecularly
defined CRC subtypes using available data from the Colon
Cancer Family Registry (CCFR) (21,25,26) and 7 studies contrib-
uting to the Genetics of Epidemiology of Colorectal Cancer
Consortium (27,28). Specifically, we evaluated each of the 4
common tumor markers (MSI, CIMP, BRAF, and KRAS) sepa-
rately, as well as 3 pathways of carcinogenesis defined by com-
binations of those markers and tumor location.
Methods
Study Populations
Data from 8 observational studies of CRC were pooled: the
Cancer Prevention Study II (CPS-II) (29), the German Darmkrebs:
Chancen der Verhutung durch Screening Study (DACHS) (30,31),
the Diet Activity and Lifestyle Study (DALS) (32), the Swedish
population of the European Prospective Investigation into
Cancer (EPIC) (33), the Melbourne Collaborative Cohort Study
(MCCS) (34), the Nurses’ Health Study (NHS) (35,36), the
Northern Sweden Health and Disease Study (NSHDS) (37), and 3
population-based sites from the Colon Cancer Family Registry
(21,25,26). Each study included women diagnosed with incident
invasive CRC and contemporaneous unaffected controls. Only
women with tumor marker data were eligible for inclusion in
this analysis. Study-specific details are described in the
Supplementary Methods (available online). All participants pro-
vided informed consent for participating in this research, and
studies were approved by their respective Institutional Review
Boards.
Data Collection and Harmonization
The harmonization procedure and ascertainment of HT use are
described in more detail in the Supplementary Methods (avail-
able online). Information on demographics and environmental
risk factors was collected via telephone or in-person interviews
and/or structured self-completed questionnaires (24,38,39). HT
use was generally ascertained as any self-reported use at base-
line survey. Additionally, ever use of formulation-specific (es-
trogen-only or estrogen plus progestin) HT use was derived
from 3 studies (CCFR, CPSII, and NHS). HT nonusers at reference
time were used as the comparison group. Postmenopausal sta-
tus was harmonized as either 1) study-derived menopausal sta-
tus, if available; 2) self-reported menopausal status, if study-
derived data were not available; or 3) age 55 years and older, if
study-derived and self-reported data were not available (40).
Tumor Characteristics and Molecular Subtyping
Tumor marker testing was conducted using DNA extracted
from formalin-fixed, paraffin-embedded tumor tissue speci-
mens. Individual study protocols varied, as outlined below and
further detailed in the Supplementary Methods (available
online).
MSI testing was primarily conducted using polymerase chain
reaction (PCR) following the National Cancer Institute Bethesda
Consensus Panel (CCFR, CPS-II, MCCS, NHS) (41). Typically, 4 or
more interpretable markers were required to classify tumors,
with some variation across studies outlined in Supplementary
Table 1 (available online). Additional methods used include im-
munohistochemistry (NSHDS, EPIC, and a subset of CCFR and
MCCS) and mononucleotide marker panels (DACHS, DALS)
(Supplementary Methods, available online). Tumors were classi-
fied as MSI-high (MSI-H) if at least 30% of the markers showed
2of9 | JNCI Cancer Spectrum, 2020, Vol. 4, No. 5
instability and MSI-L/MSS if less than 30% of the makers showed
instability. MSI status could be determined for 3639 CRC cases
(93.4%).
Most studies used Methylight (42) methylation analysis to
determine CIMP status, classified as positive or negative based
on either an 8- (CPS-II, EPIC, NHS, NSHDS) (43,44) or 5-gene
(CCFR, MCCS) (45–47) panel. The percent of methylated refer-
ence value was calculated to determine whether each gene was
positive for methylation (generally percent of methylated refer-
ence >10). DACHS used a different 5-gene panel (48,49) to de-
termine CIMP status and based methylation on the presence or
absence of the methylation-specific PCR product. DALS (50) de-
termined CIMP status using a classic panel of CpG islands (51,
52). Specific genes included in each panel, details of calling
methylation status, and number of methylated genes present to
classify a tumor as CIMP-positive are outlined in
Supplementary Table 2 (available online). CIMP status could be
determined for 3453 CRC cases (88.6%).
Studies used PCR, sequencing, and immunohistochemistry
techniques to assess BRAF and KRAS mutations, as detailed in
the Supplementary Methods (available online). The majority of
studies evaluated BRAF via V600E mutations in exon 15 and
KRAS via mutations in codons 12 and 13, although a few evalu-
ated additional loci. BRAF and KRAS status could be determined
for 3564 (91.4%) and 3435 (88.1%) CRC cases, respectively.
Tumor pathways were defined as follows, consistent with
previously suggested classifications (3,8): 1) Adenoma-
carcinoma (traditional) pathway (MSS/MSI-L, CIMP-negative,
BRAF wild-type, KRAS wild-type), 2) alternate pathway (MSS/
MSI-L, CIMP-negative, BRAF wild-type, KRAS-mutated), and 3)
serrated pathway (CIMP-positive, BRAF-mutated, KRAS wild-
type). Tumor pathway could be classified for 2401 CRC cases
(61.6%).
Tumor location was obtained from registry and pathology
reports. Location was grouped based on the International
Classification of Diseases (ICD-9) codes as follows: 1) Proximal
(153.0/Hepatic flexure, 153.1/Transverse colon, 153.4/Cecum,
153.6/Ascending colon), 2) distal (153.2/Descending colon, 152.3/
Sigmoid colon, 153.7/Splenic flexure), and 3) rectal (154.0/
Rectosigmoid junction, 154.1/Rectum). Tumor location could be
classified for 3808 CRC cases (97.7%).
Statistical Analysis
We excluded women who were pre- or perimenopausal at study
baseline (934 cases, 760 controls) and those with missing data
on HT use (208 cases, 209 controls). After exclusions, 3898 CRC
cases and 4322 controls were included in our analyses
(Figure 1).
Odds ratios (ORs) and 95% confidence intervals (CIs) from lo-
gistic regression models were used to approximate the relative
risks for the association of HT use and CRC. Separate models
were evaluated for each tumor-specific outcome using multino-
mial logistic regression with tumor marker status vs control as
the outcome (eg, BRAF-mutated or BRAF wild-type vs control).
All models included study site as well as covariates selected a
priori based on known associations with both HT and CRC.
Figure 1. Overview of participants included in analytic population. A-C ¼adenoma-carcinoma; CIMP ¼CpG island methylator phenotype; HT ¼postmenopausal hor-
mone therapy; MSI ¼microsatellite instability. *Estrogen-only and estrogen plus progestin groups are not mutually exclusive.
J. D. Labadie et al. | 3 of 9
These included age in years, body mass index (BMI; normal or
underweight [BMI <25], overweight [BMI 25–30], obese [BMI
>30], unknown), smoking status (current, former, never, un-
known), and first-degree relative with CRC (yes, no, unknown).
Secondary analyses were conducted for estrogen-only therapy
and combined estrogen plus progestin therapy. For multinomial
logistic regression models, Wald v
2
tests were used to evaluate
heterogeneity in odds ratios by tumor marker status (53).
Additionally, sensitivity analyses were conducted excluding
1) women aged 45 years and younger (n ¼131) and 2) women
with probable Lynch syndrome based on 4 tumor markers (de-
fined as MSI-H, CIMP-negative, BRAF wild-type, KRAS wild-type;
n¼89), because both populations may have unique factors al-
tering their CRC risk. We also performed a meta-analysis of the
association of any HT use and CRC risk to evaluate heterogene-
ity by study site.
All analyses were conducted using R version 3.5.2 with a 2-
sided Pless than .05 considered statistically significant.
Results
Baseline population characteristics of the 8220 postmenopausal
women in our study are shown in Table 1. Compared with
Table 1. Baseline characteristics of 8220 postmenopausal women by case-control status
Characteristics
a
Overall Case Control
(n ¼8220) (n ¼3898) (n ¼4322)
Age, mean (SD), y 65.28 (9.08) 64.79 (9.54) 65.72 (8.62)
Age group, y
<45 101 (1.2) 84 (2.2) 17 (0.4)
45–55 828 (10.1) 464 (11.9) 364 (8.4)
55–65 2780 (33.8) 1246 (32.0) 1534 (35.5)
65–75 3309 (40.3) 1561 (40.0) 1748 (40.4)
>75 1202 (14.6) 543 (13.9) 659 (15.2)
First-degree relative with CRC
Yes 1251 (15.2) 722 (18.5) 529 (12.2)
No 6633 (80.7) 2994 (76.8) 3639 (84.2)
Missing 336 (4.1) 182 (4.7) 154 (3.6)
Body mass index
Normal or underweight 3659 (44.5) 1613 ( 41.4) 2046 (47.3)
Overweight 2818 (34.3) 1322 (33.9) 1496 (34.6)
Obese 1571 (19.1) 870 (22.3) 701 (16.2)
Missing 172 ( 2.1) 93 (2.4) 79 (1.8)
Smoking
Current smoker 948 (11.5) 522 (13.4) 426 (9.9)
Former smoker 2619 (31.9) 1285 (33.0) 1334 (30.9)
Never smoker 4477 (54.5) 2012 (51.6) 2465 (57.0)
Missing 176 (2.1) 79 (2.0) 97 (2.2)
Self-reported race
White 8077 (98.3) 3780 (97.0) 4297 (99.4)
Other 113 (1.4) 98 (2.5) 15 (0.4)
Missing 30 (0.4) 20 (0.5) 10 (0.2)
Study
CCFR 1985 (24.1) 1215 ( 31.2) 770 (17.8)
CPSII 893 (10.9) 412 ( 10.6) 481 (11.1)
DACHS 2074 (25.2) 872 ( 22.4) 1202 (27.8)
DALS 891 (10.8) 427 (11.0) 464 (10.7)
EPIC Sweden 129 (1.6) 37 (0.9) 92 (2.1)
MCCS 455 (5.5) 185 (4.7) 270 (6.2)
NHS 1649 (20.1) 686 (17.6) 963 (22.3)
NSHDS 144 (1.8) 64 (1.6) 80 (1.9)
Any postmenopausal hormone therapy use
Ever 3112 (37.9) 1262 (32.4) 1850 (42.8)
Never 5108 (62.1) 2636 (67.6) 2472 (57.2)
Estrogen-only
Ever 1160 (14.1) 506 (13.0) 654 (15.1)
Never 3323 (40.4) 1778 (45.6) 1545 (35.7)
Missing 3737 (45.5) 1614 (41.4) 2123 (49.1)
Estrogen plus progestin
Ever 717 (8.7) 328 (8.4) 389 (9.0)
Never 3758 (45.7) 1961 (50.3) 1797 (41.6)
Missing 3745 (45.6) 1609 (41.3) 2136 (49.4)
a
No. (%) shown unless otherwise indicated. CCFR ¼Colon Cancer Family Registry; CPSII ¼Cancer Prevention Study II; CRC ¼colorectal cancer; DACHS ¼Darmkrebs:
Chancen der Verhutung durch Screening Study; DALS ¼Diet Activity and Lifestyle Study; EPIC ¼European Prospective Investigation into Cancer; MCCS ¼Melbourne
Collaborative Cohort Study; NHS ¼Nurses’ Health Study; NSHDS ¼Northern Sweden Health and Disease Study.
4of9 | JNCI Cancer Spectrum, 2020, Vol. 4, No. 5
controls, cases were more likely to have a family history of CRC
and be current or former smokers. Cases were less likely to be
HT users than controls (32.4% vs 42.8%). Among those with
formulation-specific data, cases were less likely than controls to
use both estrogen-only (22.2% vs 29.7%) and estrogen plus pro-
gestin formulations (14.3% vs 17.8%).
Multivariable-adjusted associations between HT use, HT for-
mulation, and overall- and tumor marker–specific CRC risk are
presented in Figure 2 and Supplementary Table 3 (available on-
line). Ever use of HT was associated with a 38% reduction in CRC
risk (OR ¼0.62, 95% CI ¼0.56 to 0.69). Both use of estrogen-only
(OR ¼0.71, 95% CI ¼0.62 to 0.83) and estrogen plus progestin
(OR ¼0.76, 95% CI ¼0.64 to 0.91) formulations were associated
with reduced CRC risk, although the effect estimates were at-
tenuated compared with any HT use in this subsample of the
study population. Few differences in baseline characteristics
were noted between women with and without formulation-
specific data (Supplementary Table 4, available online).
Among cases with respective tumor marker data, 19.8% were
MSI-H (n ¼719), 24.3% were CIMP-positive (n ¼841), 18.4% were
BRAF-mutated (n ¼654), and 32.0% were KRAS-mutated
(n ¼1098). Ever use of HT was associated with reduced risk of al-
most all tumor marker subtypes of CRC, with some variation
across subtypes (Figure 2;Supplementary Table 3, available on-
line). The association of ever HT use and CRC was attenuated
among CIMP-positive cases (OR ¼0.74, 95% CI ¼0.63 to 0.87)
compared with CIMP-negative cases (OR ¼0.62, 95% CI ¼0.55 to
0.69) (P
het
¼.04). This trend was consistent across HT formula-
tions, although the difference in odds ratios was not statistically
significant. HT use was inversely associated with both KRAS-
mutated and wild-type individuals. This association was con-
sistent for estrogen-only use; however, estrogen plus progestin
formulations were not statistically significantly associated with
KRAS-mutated individuals (OR ¼0.90, 95% CI ¼0.70 to 1.14; P
het
¼0.09). No differences were observed for MSI or BRAF mutation
status.
Of 2401 tumors (61.6%) that were able to be classified by
pathway, the majority were classified as adenoma-carcinoma
pathway tumors (48.4%; n ¼1162), with 32.9% classified as alter-
nate pathway (n ¼790) and 18.7% as the serrated pathway
(n ¼449). No major differences in baseline characteristics were
noted between women who were and were not able to be classi-
fied by pathway (Supplementary Table 4, available online). The
effect estimates for HT use in both the adenoma-carcinoma and
alternate pathways were similar to that seen for HT overall (ad-
enoma-carcinoma OR ¼0.63, 95% CI ¼0.55 to 0.73; alternate
OR ¼0.61, 95% CI ¼0.51 to 0.72). However, the effect estimate
was attenuated and no longer statistically significant for tumors
that arose via the serrated pathway (OR ¼0.81, 95% CI ¼0.66 to
1.01; P
het
vs adenoma-carcinoma ¼.04). This difference was not
consistent across HT formulation: for estrogen-only formula-
tions, ever use was statistically significantly inversely associ-
ated with all 3 pathwa ys (adenoma-carcin oma OR ¼0.71, 95%
CI ¼0.57 to 0.88; alternate OR ¼0.60, 95% CI ¼0.47 to 0.77;
Figure 2. Association between postmenopausal hormone therapy (HT) use and colorectal cancer (CRC), overall and formulation specific. A-C ¼adenoma-carcinoma;
CI ¼confidence interval; CIMP ¼CpG island methylator phenotype; CRC ¼colorectal cancer; MSI ¼microsatellite instability; OR ¼odds ratio.
*Wald P<0.05. Wald Pvalues are comparing within-group odds ratios; reference groups are BRAF wild-type, KRAS wild-type, CIMP negative, traditional, distal colon.
†
Controls are used as reference for all odds ratios. All odds ratios are adjusted for age, body mass index, smoking status, and first-degree family history of CRC.
‡Formulation-specific data were only available for a subset of women (n ¼4483 for estrogen-only; n ¼4475 for estrogen plus progestin).
J. D. Labadie et al. | 5 of 9
serrated OR ¼0.72, 95% CI ¼0.54 to 0.98), and there was no sta-
tistical difference between pathways. However, for estrogen
plus progestin formulations, ever use was only statistically sig-
nificantly associated with tumors that arose via the adenoma-
carcinoma pathway (OR ¼0.70, 95% CI ¼0.55 to 0.91).
Most tumors were located in the proximal colon (47.3%),
with tumors of the distal colon (29.4%) only slightly more com-
mon than rectal tumors (23.3%). Compared with distal colon
(OR ¼0.57, 95% CI ¼0.49 to 0.66) and rectal tumors (OR ¼0.54,
95% CI ¼0.46 to 0.63), the effect estimate for the association of
HT use and proximal colon tumors was attenuated (OR ¼0.71,
95% CI ¼0.62 to 0.80; P
het
vs distal ¼.01). This trend was consis-
tent across HT formulations, although there was no statistical
difference between proximal and distal colon tumors for
estrogen-only formulation (P
het
¼.32).
No substantial changes in results were noted after removing
either the 131 women aged 45 years and younger or the 89 wom-
enwith molecularly defined Lynch syndrome (Supplementary
Tables 5 and 6, available online). Meta-analysis results were
consistent with our pooled main analysis (summary OR ¼0.64,
95% CI ¼0.58 to 0.71; P
het
¼.10).
Discussion
In this large pooled study of postmenopausal women, HT use,
regardless of formulation type, was associated with a decreased
risk of CRC, consistent with prior research (11–20). In general,
this inverse association was observed irrespective of MSI, CIMP,
BRAF,orKRAS status. However, when considering all tumor
markers together and grouping cases by common pathways and
tumor location, the association was attenuated for tumors aris-
ing via the serrated pathway and for proximal colon tumors.
Our results do not support the hypothesis that the associa-
tion of HT and CRC differs by the individual tumor markers MSI,
BRAF, and KRAS. Strong inverse associations were observed for
HT use and CRC, regardless of BRAF and KRAS status. Prior stud-
ies found a nearly 20% reduced risk among ever HT users irre-
spective of BRAF and KRAS mutation status, although effect
estimates did not reach statistical significance (22,23). These
studies had substantially smaller samples sizes than ours, con-
tributing to reduced power to detect difference in effect. We ad-
ditionally observed strong inverse associations for HT use and
both MSI-L/MSS and MSI-H CRC. Prior research is somewhat
conflicting regarding the association of HT and MSI status, with
most studies suggesting an association only among MSI-L/MSS
patients (22,24,54). There are many possible explanations for
this discrepancy, including sample size, study design, reference
period used for ascertaining HT use, and panels used to classify
MSI status. Because prior studies have indicated high concor-
dance across MSI panels (41), we suspect the latter had the least
influence.
We found some evidence that the association of HT differs
by CIMP status, with an attenuated effect estimate observed for
CIMP-positive tumors. A previous study had similar findings,
reporting a borderline inverse association among CIMP-
negative tumors and no association for CIMP-positive tumors
(22,47). This finding should be interpreted with caution because
CIMP is not consistently defined across studies, and CIMP preva-
lence may be affected by detection method and sample quality.
Our results suggest that a comprehensive approach of con-
sidering tumor markers together as pathways may reveal other-
wise nebulous patterns. Our findings indicate that the
association of HT use and CRC was largely driven by tumors
arising via the adenoma-carcinoma and alternate pathway.
These tumors make up the majority of CRC cases, whereas ser-
rated tumors represent about 20%–30% of CRC (1,3,8,55–57).
Serrated tumors, characterized as CIMP-positive, BRAF-mutated,
and KRAS wild-type, tend to behave more aggressively, with
faster progression and poorer prognosis (3,8,10,56,58–60).
Based on the different biologic behavior, appearance, distribu-
tion of tumor markers, and genetic susceptibility of serrated
tumors, it is plausible that HT may indeed play a lesser role in
their pathogenesis.
We also observed a weaker association for tumors of the
proximal colon, consistent with prior studies (22,24). There is
evidence that serrated tumors are more likely to develop in the
proximal colon (8,61,62), so it is unclear whether these are in-
dependent associations. In our study, most serrated tumors
(n ¼391) were in the proximal colon. The association of HT use
and proximal tumors was similar (OR ¼0.68, 95% CI ¼0.60 to
0.78) after removing serrated tumors from analysis, suggesting
an independent association. However, 43.1% (n ¼776) of proxi-
mal tumors could not be classified by pathway because of in-
complete tumor marker data, so this analysis is limited. The
proximal and distal colon have different embryologic origins,
microbiomes, and microenvironments (62–66). As such, they ap-
pear to be predisposed to different tumor types. For instance,
proximal colon tumors are more likely to be MSI-H, CIMP-posi-
tive, and mucinous and occur more commonly in women and
older individuals (67–71). Further research is needed to better
elucidate whether differences in the proximal colon make it
less sensitive to the effects of estrogens (ie, fewer receptors, dif-
ferent microbiota), whether precancerous lesions in the proxi-
mal colon are estrogen insensitive based on differences in the
carcinogenic pathway, or some combination of factors.
Our results indicate that both estrogen-only and estrogen
plus progestin formulations reduce CRC risk. In general, effect
estimates were attenuated for estrogen plus progestin formula-
tions compared with estrogen-only formulations, perhaps
reflecting smaller exposure frequencies. However, overall
trends were similar. Two main exceptions were present. First,
whereas estrogen-only and any HT use were associated with
about a 40% reduction in tumors arising via the alternate path-
way, estrogen plus progestin use was not statistically signifi-
cantly associated with these tumors. This may indicate that
alternate pathway tumorigenesis is specifically modified by es-
trogen and not progestin. Likewise, there was a null association
between estrogen plus progestin use and proximal colon
tumors despite a 24%–29% reduction in risk with estrogen-only
or any HT, respectively.
To our knowledge, this is the largest study to assess whether
the association of HT use and CRC differs by individual tumor
markers and location. In addition, it is one of few investigations
that combines multiple tumor markers to evaluate tumor path-
way–specific associations. Some limitations should be consid-
ered in interpreting our results. First, all exposure and
epidemiologic covariate information included in this analysis
was based on self-report, which could lead to exposure misclas-
sification. Second, HT use was assessed only during the refer-
ence period, and detailed information on dose, frequency, and
duration of use was not routinely available. Third, we were not
able to assess endogenous hormones that may reflect age at
menarche, parity, or breast feeding, which may also influence
CRC risk. Fourth, there is some evidence that HT users may be
more likely to undergo CRC screening (72,73). It is unclear how
this may impact our results because this relationship may be
complicated by differences in sensitivity of screening detection
6of9 | JNCI Cancer Spectrum, 2020, Vol. 4, No. 5
for specific CRC subtypes. Temporal trends and regional differ-
ences in screening and HT use may also influence observed
associations. Finally, although this study includes populations
in many locales, the participants were predominantly white,
and therefore, these findings may not be generalizable to other
racial and ethnic groups.
In this large, multisite study we observed a strong inverse
association between HT use and CRC risk, regardless of individ-
ual tumor markers and HT formulation. The decreased risk may
predominantly reflect tumors of the distal colon or rectum and
those arising via the adenoma-carcinoma (traditional) pathway,
because the association was relatively weaker among proximal
colon tumors and those arising via the serrated pathway.
Further investigation into the mechanisms underlying these
differences may add to our understanding of subtype-specific
CRC risk and pathways of tumorigenesis.
Funding
Genetics and Epidemiology of Colorectal Cancer
Consortium: This work was supported by the National
Cancer Institute, National Institutes of Health, US
Department of Health and Human Services R01 CA176272,
U01 CA137088, and U01 CA164930. This research was funded
in part through the National Institutes of Health/National
Cancer Institute Cancer Center Support Grant P30 CA015704.
The Colon Cancer Family Registry (CCFR) was supported
in part by National Cancer Institute/National Institutes of
Health award number U01 CA167551 and through National
Cancer Institute/National Institutes of Health U01/U24 coop-
erative agreements with the following CCFR sites: Ontario
(OFCCR) (CA074783), Seattle (SCCFR) (CA074794 and R01
CA076366), and Australasian (ACCFR) (CA074778 and
CA097735). The content of this manuscript does not neces-
sarily reflect the views or policies of the National Cancer
Institute or any of the collaborating centers in the CCFR, nor
does mention of trade names, commercial products, or
organizations imply endorsement by the US government,
any cancer registry, or the CCFR.
CPS-II: The American Cancer Society funds the creation,
maintenance, and updating of the Cancer Prevention Study
II cohort. This study was conducted with Institutional
Review Board approval.
DACHS: This work was supported by the German
Research Council (BR 1704/6–1, BR 1704/6–3, BR 1704/6–4, CH
117/1–1, HO 5117/2–1, HE 5998/2–1, KL 2354/3–1, RO 2270/8–1,
and BR 1704/17–1), the Interdisciplinary Research Program of
the National Center for Tumor Diseases, Germany, and the
German Federal Ministry of Education and Research
(01KH0404, 01ER0814, 01ER0815, 01ER1505A, and 01ER1505B).
DALS: National Institutes of Health (R01 CA48998).
EPIC: The coordination of EPIC is financially supported by
the European Commission (DGSANCO) and the
International Agency for Research on Cancer. The national
cohorts are supported by Danish Cancer Society (Denmark),
Swedish Cancer Society, Swedish Research Council, and
County Councils of Ska˚ ne and V€
asterbotten (Sweden).
MCCS: This cohort recruitment was funded by VicHealth
and Cancer Council Victoria. The MCCS was further supported
by Australian National Health and Medical Research Council
grants 509348, 209057, 251553, and 504711 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.
Harvard cohort (NHS): NHS is supported by the National
Institutes of Health (R01 CA137178, P01 CA087969, UM1
CA186107, R01 CA151993, R35 CA197735, K07 CA190673, and
P50 CA127003).
NSHDS: Swedish Cancer Society; Cancer Research
Foundation in Northern Sweden; Swedish Research Council;
J C Kempe Memorial Fund; Faculty of Medicine, Umea˚
University, Umea˚ , Sweden; and Cutting-Edge Research
Grant from the County Council of V€
asterbotten, Sweden.
Fred Hutchinson Cancer Research Center investigators
were also supported by National Institutes of Health T32
CA094880 and National Institutes of Health K05 CA152715.
Notes
Role of the funders: The funders had no role in the design of the
study; the collection, analysis, and interpretation of the data;
the writing of the manuscript; and the decision to submit the
manuscript for publication.
Authors contributions: Conceptualization: JDL, MH, PAN; Data
curation: TAH, BB, YL, UP; Formal analysis: JDL, BB, LH, YL, WS;
Funding acquisition: HB, PTC, ATC, JC, JCF, SJG, GGG, MJG, MH,
MAJ, RLM, VM, NM, SO, LCS, MLS, SNT, BV, UP, PAN;
Investigation: all authors; Methodology: JDL, LH; Project admin-
istration: TAH, UP; Resources: HB, PTC, ATC, JC, JCF, SJG, GGG,
MJG, MH, MAJ, RLM, VM, NM, SO, LCS, MLS, SNT, BV, UP, PAN;
Supervision: UP, PAN; Visualization: JDL; Writing - original draft:
JDL, TAH, UP, PAN; Writing - review & editing: all authors.
Disclosures: The authors have no conflicts of interest to
disclose.
Acknowledgments: SCCFR: The authors would like to thank the
study participants and staff of the Hormones and Colon Cancer
and Seattle Cancer Family Registry studies (CORE studies). CPS-II:
The authors thank the CPS-II participants and Study Management
Group for their invaluable contributions to this research. The
authors would also like to acknowledge the contribution to this
study from central cancer registries supported through the
Centers for Disease Control and Prevention National Program of
Cancer Registries and cancer registries supported by the National
Cancer Institute Surveillance Epidemiology and End Results pro-
gram. DACHS: We thank all participants and cooperating clini-
cians, and Ute Handte-Daub, Utz Benscheid, Muhabbet Celik, and
Ursula Eilber for excellent technical assistance. EPIC: Where
authors are identified as personnel of the International Agency for
Research on Cancer/World Health Organization, the authors alone
are responsible for the views expressed in this article, and they do
not necessarily represent the decisions, policy, or views of the
International Agency for Research on Cancer/World Health
Organization. Harvard cohort (NHS): The study protocol was ap-
proved 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 for their valuable con-
tributions 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,
J. D. Labadie et al. | 7 of 9
TX, VA, WA, WY. The authors assume full responsibility for analy-
ses and interpretation of these data. MCCS: Melbourne
Collaborative Cohort Study cohort recruitment was funded by
VicHealth and Cancer Council Victoria. The MCCS was further
augmented 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. NSHDS: The NSHDS investigators
thank the Biobank Research Unit at Umea˚University,the
V€
asterbotten Intervention Programme, the Northern Sweden
MONICA study, and Region V€
asterbotten for providing data and
samples and acknowledge the contribution from Biobank Sweden,
supported by the Swedish Research Council (VR 2017–00650).
References
1. Jass JR. Classification of colorectal cancer based on correlation of clinical,
morphological and molecular features. Histopathology. 2007;50(1):113–130.
2. Jass JR. Molecular heterogeneity of colorectal cancer: implications for cancer
control. Surg Oncol. 2007;16(1):7–9.
3. Leggett B, Whitehall V. Role of the serrated pathway in colorectal cancer
pathogenesis. Gastroenterology. 2010;138(6):2088–2100.
4. Samadder NJ, Vierkant RA, Tillmans LS, et al Associations between colorectal
cancer molecular markers and pathways with clinicopathologic features in
older women. Gastroenterology. 2013;145(2):348–356.e1-2.
5. Kambara T, Simms LA, Whitehall VL, et al BRAF mutation is associated with
DNA methylation in serrated polyps and cancers of the colorectum. Gut.
2004;53(8):1137–1144.
6. Hawkins NJ, Ward RL. Sporadic colorectal cancers with microsatellite insta-
bility and their possible origin in hyperplastic polyps and serrated adenomas.
J Natl Cancer Inst. 2001;93(17):1307–1313.
7. M€
akinen MJ, George SMC, Jernvall P, et al Colorectal carcinoma associated
with serrated adenoma—prevalence, histological features, and prognosis. J
Pathol. 2001;193(3):286–294.
8. Kocarnik JM, Shiovitz S, Phipps AI. Molecular phenotypes of colorectal cancer
and potential clinical applications. Gastroenterol Rep (Oxf). 2015;3(4):269–276.
9. Zlobec I, Bihl MP, Foerster A, et al Stratification and prognostic relevance of
Jass’s molecular classification of colorectal cancer. Front Oncol. 2012;2:7.
10. Phipps AI, Limburg PJ, Baron JA, et al Association between molecular sub-
types of colorectal cancer and patient survival. Gastroenterology. 2015;148(1):
77–87 e2.
11. Anderson GL, Limacher M, Assaf AR, et al Effects of conjugated equine estro-
gen in postmenopausal women with hysterectomy: the Women’s Health
Initiative randomized controlled trial. JAMA. 2004;291(14):1701–1712.
12. Calle EE, Miracle-McMahill HL, Thun MJ, et al Estrogen replacement therapy
and risk of fatal colon cancer in a prospective cohort of postmenopausal
women. J Natl Cancer Inst. 1995;87(7):517–523.
13. Fernandez E, La Vecchia C, Braga C, et al Hormone replacement therapy and
risk of colon and rectal cancer. Cancer Epidemiol Biomarkers Prev. 1998;7(4):
329–333.
14. Grodstein F, Martinez ME, Platz EA, et al Postmenopausal hormone use and
risk for colorectal cancer and adenoma. Ann Intern Med. 1998;128(9):705–712.
15. Grodstein F, Newcomb PA, Stampfer MJ. Postmenopausal hormone therapy
and the risk of colorectal cancer: a review and meta-analysis. Am J Med. 1999;
106(5):574–582.
16. Campbell PT, Newcomb P, Gallinger S, et al Exogenous hormones and colo-
rectal cancer risk in Canada: associations stratified by clinically defined fa-
milial risk of cancer. Cancer Causes Control. 2007;18(7):723–733.
17. Johnson JR, Lacey JV, Lazovich D, et al Menopausal hormone therapy and risk
of colorectal cancer. Cancer Epidemiol Biomarkers Prev. 2009;18(1):196–203.
18. Rennert G, Rennert HS, Pinchev M, et al Use of hormone replacement therapy
and the risk of colorectal cancer. J Clin Oncol. 2009;27(27):4542–4547.
19. Lin KJ, Cheung WY, Lai JY, et al The effect of estrogen vs. combined estrogen-
progestogen therapy on the risk of colorectal cancer. Int J Cancer. 2012;130(2):
419–430.
20. Green J, Czanner G, Reeves G, et al Menopausal hormone therapy and risk of
gastrointestinal cancer: nested case-control study within a prospective co-
hort, and meta-analysis. Int J Cancer. 2012;130(10):2387–2396.
21. Newcomb PA, Baron J, Cotterchio M, et al Colon Cancer Family Registry: an
international resource for studies of the genetic epidemiology of colon can-
cer. Cancer Epidemiol Biomarkers Prev. 2007;16(11):2331–2343.
22. Limsui D, Vierkant RA, Tillmans LS, et al Postmenopausal hormone therapy
and colorectal cancer risk by molecularly defined subtypes among older
women. Gut. 2012;61(9):1299–1305.
23. Limburg PJ, Limsui D, Vierkant RA, et al Postmenopausal hormone therapy
and colorectal cancer risk in relation to somatic KRAS mutation status
among older women. Cancer Epidemiol Biomarkers Prev. 2012;21(4):681–684.
24. Newcomb PA, Zheng Y, Chia VM, et al Estrogen plus progestin use, microsat-
ellite instability, and the risk of colorectal cancer in women. Cancer Res. 2007;
67(15):7534–7539.
25. Cotterchio M, McKeown-Eyssen G, Sutherland H, et al Ontario familial colon
cancer registry: methods and first-year response rates. Chronic Dis Can. 2000;
21(2):81–86.
26. Winship I, Win AK. The Australasian Colorectal Cancer Family Registry. Med J
Aust. 2012;197(9):480–481.
27. Phipps AI, Newcomb PA, Garcia-Albeniz X, et al Association between colorec-
tal cancer susceptibility loci and survival time after diagnosis with colorectal
cancer. Gastroenterology. 2012;143(1):51–54.e4.
28. Peters U, Jiao S, Schumacher FR, et al Identification of genetic susceptibility
loci for colorectal tumors in a genome-wide meta-analysis. Gastroenterology.
2013;144(4):799–807.e24.
29. Calle EE, Rodriguez C, Jacobs EJ, et al The American Cancer Society Cancer
Prevention Study II Nutrition Cohort: rationale, study design, and baseline
characteristics. Cancer. 2002;94(9):2490–2501.
30. Brenner H, Chang-Claude J, Seiler CM, et al Protection from colorectal cancer
after colonoscopy: a population-based, case-control study. Ann Intern Med.
2011;154(1):22–30.
31. Amitay EL, Carr PR, Jansen L, et al Association of aspirin and nonsteroidal
anti-inflammatory drugs with colorectal cancer risk by molecular subtypes. J
Natl Cancer Inst. 2019;111(5):475–483.
32. Slattery ML, Potter J, Caan B, et al Energy balance and colon cancer—beyond
physical activity. Cancer Res. 1997;57(1):75–80.
33. Riboli E, Kaaks R. The EPIC Project: rationale and study design. European
Prospective Investigation into Cancer and Nutrition. Int J Epidemiol. 1997;
26(suppl 1):S6–14.
34. Giles GG, English DR. The Melbourne Collaborative Cohort Study. IARC Sci
Publ. 2002;156:69–70.
35. Belanger CF, Hennekens CH, Rosner B, et al The Nurses’ Health Study. Am J
Nurs. 1978;78(6):1039–1040.
36. Colditz GA, Manson JE, Hankinson SE. The Nurses’ Health Study: 20-year con-
tribution to the understanding of health among women. J Womens Health.
1997;6(1):49–62.
37. Hallmans G, Agren A, Johansson G, et al Cardiovascular disease and diabetes
in the Northern Sweden Health and Disease Study Cohort-evaluation of risk
factors and their interactions. Scand J Public Health Suppl. 2003;31(61_suppl):
18–24.
38. Colditz GA. Menopausal hormone therapy after breast cancer. Breast Cancer
Res. 2005;7(4):168–170.
39. Hoffmeister M, Raum E, Krtschil A, et al No evidence for variation in colorec-
tal cancer risk associated with different types of postmenopausal hormone
therapy. Clin Pharmacol Ther. 2009;86(4):416–424.
40. Stampfer MJ, Colditz GA, Willett WC, et al Postmenopausal estrogen therapy
and cardiovascular disease. Ten-year follow-up from the Nurses’ Health
Study. N Engl J Med. 1991;325(11):756–762.
41. Boland CR, Thibodeau SN, Hamilton SR, et al A National Cancer Institute
workshop on microsatellite instability for cancer detection and familial pre-
disposition: development of international criteria for the determination of
microsatellite instability in colorectal cancer. Cancer Res. 1998;58(22):
5248–5257.
42. Eads CA, Danenberg KD, Kawakami K, et al MethyLight: a high-throughput
assay to measure DNA methylation. Nucleic Acids Res. 2000;28(8):E32.
43. Dahlin AM, Palmqvist R, HenrikssonML, et al The role of the CpG island meth-
ylator phenotype in colorectal cancer prognosis depends on microsatellite in-
stability screening status. ClinCancer Res. 2010;16(6):1845–1855.
44. Ogino S, Kawasaki T, Kirkner GJ, et al CpG island methylator phenotype-low
(CIMP-low) in colorectal cancer: possible associations with male sex and
KRAS mutations. J Mol Diagn. 2006;8(5):582–588.
45. English DR, Young JP, Simpson JA, et al Ethnicity and risk for colorectal can-
cers showing somatic BRAF V600E mutation or CpG island methylator phe-
notype. Cancer Epidemiol Biomarkers Prev. 2008;17(7):1774–1780.
46. Weisenberger DJ, Siegmund KD, Campan M, et al CpG island methylator
phenotype underlies sporadic microsatellite instability and is tightly
associated with BRAF mutation in colorectal cancer. Nat Genet. 2006;38(7):
787–793.
47. Weisenberger DJ, Levine AJ, Long TI, et al Association of the colorectal CpG is-
land methylator phenotype with molecular features, risk factors, and family
history. Cancer Epidemiol Biomarkers Prev. 2015;24(3):512–519.
48. Carr PR, Jansen L, Bienert S, et al Associations of red and processed meat in-
take with major molecular pathological features of colorectal cancer. Eur J
Epidemiol. 2017;32(5):409–418.
49. Warth A, Kloor M, Schirmacher P, et al Genetics and epigenetics of small
bowel adenocarcinoma: the interactions of CIN, MSI, and CIMP. Mod Pathol.
2011;24(4):564–570.
50. Samowitz WS, Sweeney C, Herrick J, et al Poor survival associated with the
BRAF V600E mutation in microsatellite-stable colon cancers. Cancer Res. 2005;
65(14):6063–6069.
51. Rashid A, Shen L, Morris JS, et al CpG island methylation in colorectal adeno-
mas. Am J Pathol. 2001;159(3):1129–1135.
52. Park SJ, Rashid A, Lee JH, et al Frequent CpG island methylation in serrated
adenomas of the colorectum. Am J Pathol. 2003;162(3):815–822.
8of9 | JNCI Cancer Spectrum, 2020, Vol. 4, No. 5
53. Fox J. Applied Regression Analysis and Generalized Linear Models. 3rd ed.
Thousand Oaks, VA: SAGE Publications, Inc; 2016.
54. Slattery ML, Potter JD, Curtin K, et al Estrogens reduce and withdrawal of
estrogens increase risk of microsatellite instability-positive colon cancer.
Cancer Res. 2001;61(1):126–130.
55. Snover DC. Update on the serrated pathway to colorectal carcinoma. Hum
Pathol. 2011;42(1):1–10.
56. Bettington M, Walker N, Clouston A, et al The serrated pathway to colorectal
carcinoma: current concepts and challenges. Histopathology. 2013;62(3):367–386.
57. Vogelstein B, Fearon ER, Hamilton SR, et al Genetic alterations during
colorectal-tumor development. N Engl J Med. 1988;319(9):525–532.
58. Burnett-Hartman AN, Newcomb PA, Potter JD, et al Genomic aberrations oc-
curring in subsets of serrated colorectal lesions but not conventional adeno-
mas. Cancer Res. 2013;73(9):2863–2872.
59. Groff RJ, Nash R, Ahnen DJ. Significance of serrated polyps of the colon. Curr
Gastroenterol Rep. 2008;10(5):490–498.
60. Li SC, Burgart L. Histopathology of serrated adenoma, its variants, and differ-
entiation from conventional adenomatous and hyperplastic polyps. Arch
Pathol Lab Med. 2007;131(3):440–445.
61. Yamauchi M, Morikawa T, Kuchiba A, et al Assessment of colorectal cancer
molecular features along bowel subsites challenges the conception of distinct
dichotomy of proximal versus distal colorectum. Gut.2012;61(6):847–854.
62. Lee GH, Malietzis G, Askari A, et al Is right-sided colon cancer different to
left-sided colorectal cancer? A systematic review. Eur J Surg Oncol. 2015;41(3):
300–308.
63. Mowat AM, Agace WW. Regional specialization within the intestinal immune
system. Nat Rev Immunol. 2014;14(10):667–685.
64. Merlano MC, Granetto C, Fea E, et al Heterogeneity of colon cancer: from
bench to bedside. ESMO Open. 2017;2(3):e000218.
65. Hooper LV, Littman DR, Macpherson AJ. Interactions between the microbiota
and the immune system. Science. 2012;336(6086):1268–1273.
66. Macfarlane GT, Macfarlane LE. Acquisition, evolution and maintenance of
the normal gut microbiota. Dig Dis. 2009;27(1):90–98.
67. Iacopetta B. Are there two sides to colorectal cancer? Int J Cancer. 2002;101(5):
403–408.
68. Gervaz P, Bucher P, Morel P. Two colons-two cancers: paradigm shift and clin-
ical implications. J Surg Oncol. 2004;88(4):261–266.
69. Slattery ML, Curtin K, Wolff RK, et al A comparison of colon and rectal so-
matic DNA alterations. Dis Colon Rectum. 2009;52(7):1304–1311.
70. Minoo P, Zlobec I, Peterson M, et al Characterization of rectal, proximal and
distal colon cancers based on clinicopathological, molecular and protein pro-
files. Int J Oncol. 2010;37(3):707–718.
71. Ang PW, Loh M, Liem N, et al Comprehensive profiling of DNA methylation in
colorectal cancer reveals subgroups with distinct clinicopathological and
molecular features. BMC Cancer. 2010;10(1):227.
72. Prihartono N, Palmer JR, Louik C, et al A case-control study of use of postmen-
opausal female hormone supplements in relation to the risk of large bowel
cancer. Cancer Epidemiol Biomarkers Prev. 2000;9(4):443–447.
73. Barrett-Connor E. Postmenopausal estrogen and prevention bias. Ann Intern
Med. 1991;115(6):455–456.
J. D. Labadie et al. | 9 of 9
