University of Southern California Dissertations and Theses

The role of steroid hormones in the etiology of urologic diseases

(USC Thesis Other)

The role of steroid hormones in the etiology of urologic diseases

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THE ROLE OF STEROID HORMONES IN THE ETIOLOGY OF UROLOGIC DISEASES

by
Carol Ann Davis-Dao

______________________________________________________________

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements of the Degree
DOCTOR OF PHILOSOPHY
(EPIDEMIOLOGY)

May 2011

Copyright 2011 Carol Ann Davis-Dao
ii

Acknowledgements
I would like to thank the many people who helped and encouraged me in the process of
completing my dissertation. First of all, I would to express gratitude to my committee chair and
advisor, Dr. Victoria Cortessis. I would like to thank Dr. Rebecca Sokol and Ellenie Tuazon for
their advice and assistance on my infertility meta-analysis. Thank you to Dr. Leslie Bernstein
and the group from the California Teachers Study, and Dr. Malcolm Pike and the group from the
Los Angeles-Shanghai Bladder Cancer Study, for your assistance and collaboration. I am
grateful to my dissertation committee members, Dr. Stanley Azen, Dr. Roberta McKean-Cowdin,
Dr. Kimberly Siegmund and Dr. Gerry Coetzee, for their guidance, questions and advice.
Most of all, I would like to thank my good friend Jennifer Thomas for helping me along
the way, and my husband, Vinh Dao, I could never have made it this far without your unwavering
support.
iii

Table of Contents
Acknowledgements ii
List of Tables iv
List of Figures v
Abbreviations vi
Abstract vii

Chapter 1: Male infertility and variation in CAG repeat length in the androgen
receptor gene: A meta-analysis 1
1.1 Abstract 1
1.2 Introduction 2
1.3 Materials and Methods 3
1.4 Results 7
1.5 Discussion 21
Chapter 1 References 25

Chapter 2: Length of the polymorphic CAG tract in the androgen receptor gene
varies by histologic type in testicular germ cell tumor cases from California 29
2.1 Abstract 29
2.2 Introduction 30
2.3 Materials and Methods 32
2.4 Results 36
2.5 Discussion 50
Chapter 2 References 56

Chapter 3: Lower risk in parous women suggests hormonal factors important in
bladder cancer etiology 59
3.1 Abstract 59
3.2 Introduction 60
3.3 Materials and Methods 61
3.4 Results 67
3.5 Discussion 92
Chapter 3 References 98

Comprehensive References 102

iv

List of Tables

Table 1.1: Published studies on associations between male infertility and 12
length of the CAG and GGC (GGN) trinucleotide tracts in the
androgen receptor gene

Table 1.2: Results of overall and stratified meta-analyses 14

Table 1.3: Results of multivariate meta-regression analyses 17

Table 2.1: Characteristics of all testicular germ cell tumor cases 40

Table 2.2 Androgen receptor CAG repeat lengths comparing the allele 45
transmitted from mothers to testicular germ cell tumor (TGCT)
cases (“cases”) to the allele not transmitted (“mothers”)

Table 2.3: Case-case analyses displaying odds ratios and 95% confidence 48
intervals for the association of androgen receptor CAG repeat
length with testicular germ cell tumor histology comparing
seminoma cases with: nonseminoma and mixed germ cell tumor
cases, mixed germ cell tumor cases only and pure nonseminoma
cases only

Table 3.1: Selected characteristics of women from the Los Angeles-Shanghai 69
Bladder Cancer Study

Table 3.2: Adjusted odds ratios and 95% confidence intervals for 72
associations between hormonal and reproductive factors and risk
of bladder cancer among women in the Los Angeles-Shanghai
Bladder Cancer Study

Table 3.3: Selected characteristics of the California Teachers Study cohort, 75
by smoking status

Table 3.4: Adjusted relative risks and 95% confidence intervals for 78
associations between selected menstrual, hormonal and
reproductive factors and risk of bladder cancer in the California
Teachers Study

Table 3.5: Published reports on hormonal and reproductive exposures and 83
risk of bladder cancer in women

Table 3.6: Contributing data and summary relative risks and 95% confidence 85
intervals from the meta-analysis of hormonal and reproductive
exposures and risk of bladder cancer in women

v

List of Figures

Figure 1.1A: Forrest plot of the full set of 33 studies 18

Figure 1.1B: Begg’s Funnel plot showing standardized mean difference (SMD) 19
estimates for each of the 33 studies plotted against the standard
error of each SMD

Figure 1.2A: Forrest plot of the sub-set of 13 studies that used stringent case and 20
control criteria

Figure 1.2B: Begg’s Funnel plot showing standardized mean difference (SMD) 20
estimates for each of the 13 studies from the sub-set that used
more stringent case-control criteria plotted against the standard
error of each SMD

Figure 2.1: Age-specific incidence rates of testicular germ cell tumors by 49
histologic type, 1992-2007, SEER 13

Figure 3.1A: Forrest plots displaying contributing data and results of 87
meta-analyses relating parity to risk of bladder cancer, comparing
all parous and nulliparous women

Figure 3.1B: Forrest plots displaying contributing data and results of 88
meta-analyses relating parity to risk of bladder cancer, comparing
parous and nulliparous women within strata of smoking history

Figure 3.2A: Forrest plots displaying contributing data and results of 89
meta-analyses relating history of exogenous hormone use to risk
of bladder cancer, comparing never and ever users of
oral contraceptives

Figure 3.2B: Forrest plots displaying contributing data and results of 90
meta-analyses relating history of exogenous hormone use to risk
of bladder cancer, comparing never and ever users of any
menopausal hormone therapy (HT)

Figure 3.2C: Forrest plots displaying contributing data and results of 90
meta-analyses relating history of exogenous hormone use to risk
of bladder cancer, comparing never and ever users of estrogen
alone for HT

Figure 3.2D: Forrest plots displaying contributing data and results of 91
meta-analyses relating history of exogenous hormone use to risk
of bladder cancer, comparing never and ever users of estrogen plus
progestin for HT

vi

Abbreviations

AR Androgen receptor
AR Androgen receptor gene
SMD Standardized mean difference
CI Confidence interval
TGCT Testicular germ cell tumor
CIS Carcinoma in situ
OR Odds ratio
UC Urothelial carcinoma
ER Estrogen receptor
PR Progesterone receptor
CTS California Teachers Study
RR Relative risk
HT Menopausal hormone therapy
OC Oral contraceptives
vii

Abstract

One of the main goals of my doctoral training has been to gain experience working with a
variety of types of epidemiologic data. The three projects that compose this dissertation have
provided me with the opportunity to work with several different data structures: (1) published
summary data, (2) population-based family data, (3) prospective cohort data, and (4) population-
based case-control data.
For my dissertation, I have employed each of these data structures to investigate the role
of steroid hormones in the etiology of several urologic diseases. In the last several decades of
epidemiologic research, it has become clear that steroid hormones are involved in many disease
processes. For my research, I have studied the relationships between steroid hormone exposure
and molecules of steroid hormone response and female bladder cancer, as well as two male
urologic diseases: infertility and testicular germ cell tumors.
In my first paper, I investigated the association between male infertility and a
polymorphic CAG repeat tract in the androgen receptor gene (AR) through the meta-analysis of a
large body of existing literature on the topic. I quantitatively summarized 33 studies, and found
that infertile male cases had statistically significantly longer mean CAG repeat length than
controls (SMD: 0.19, 95% CI: 0.09-0.29). Meta-analysis of a select subset of 13 studies that used
more stringent case and control selection criteria revealed a larger difference between cases and
controls (SMD: 0.31, 95% CI: 0.14-0.47). This meta-analysis provides support for an association
between increased androgen receptor CAG length and idiopathic male infertility, and brings
together for the first time results from many published reports on this topic.
In my second paper, I expanded upon the results of the meta-analysis to conduct an
original data analysis studying the association between androgen receptor CAG repeat length and
risk of testicular germ cell tumors. Increasing rates of testicular germ cells tumors (TGCTs) over
viii

the last 40 years suggest environmental factors are involved in disease etiology, but familial risk
studies indicate that genetic factors are also important.
I investigated whether variation in CAG trinucleotide repeat length in exon 1 of the
androgen receptor gene is associated with TGCTs using data from a population-based family
study. Analyses of 273 TGCT case-mother pairs revealed no association between androgen
receptor CAG repeat length and overall risk of TGCTs. Risk of seminoma was significantly
associated with shorter CAG repeat length, and the trend was highly significant over decreasing
CAG repeat length (CAG ≥20 versus CAG ≤19: OR=0.54 (95% CI: 0.31-0.93), p trend=0.003).
Case-case analyses comparing seminoma cases to nonseminoma and mixed germ cell tumors
confirmed that seminoma cases had shorter CAG repeat length than other histologies (CAG ≥20
versus CAG ≤19: OR=0.54 (95%CI: 0.29-1.01), p trend=0.003).
My findings indicate that the AR may be involved in progression from CIS to seminoma
and that genetically-determined differences in androgen action may be involved in TGCT
etiology. These results suggest further work on the role of the androgen receptor in carcinoma in
situ cells, and investigation of potential gene-environmental interactions between AR variants and
hormonally active agents that may together influence TGCT risk.
In my third paper, I changed my focus from male reproductive disorders, and investigated
the role of hormones in the etiology of female bladder cancer. It is known that men have two-to-
four times the rate of bladder cancer of women, and this difference in risk cannot be fully
explained by known bladder cancer risk factors. Hormones provide an intriguing research area
due to the fundamental hormonal differences between men and women, and biological evidence
of the effect of steroid hormones on bladder carcinogenesis. Investigating factors that reflect
hormone levels among women, including parity, age at menarche, age at menopause and use of
ix

exogenous hormones, provides an avenue to study the hypothesis that steroid hormones are
associated with risk of bladder cancer.
I analyzed data from women participating in two population-based studies: the Los
Angeles-Shanghai Bladder Cancer Study, in which 349 female case-control pairs were enrolled in
Los Angeles and 131 female cases and 138 frequency-matched controls were enrolled in
Shanghai; and the California Teachers Study (CTS), in which 196 incident cases of bladder
urothelial carcinoma were diagnosed between 1995 and 2005 in an analytic cohort of 120,857
women. I also conducted a meta-analysis summarizing data on these associations from our
primary analyses together with results in the published literature.
In primary data analyses, parous women experienced at least a 30% reduced risk of
bladder cancer compared with nulliparous women (Shanghai: OR=0.38, 95%CI: 0.14-1.07; CTS:
RR=0.69, 95%CI: 0.50-0.95) consistent with results of a meta-analysis of nine studies (summary
RR=0.73, 95%CI: 0.63-0.85). The CTS, which included data on formulation of menopausal
hormone therapy (HT), revealed a protective effect for use of combined estrogen and progestin
compared with no HT use (RR=0.60, 95%CI: 0.37-0.98). Meta-analysis of three studies provided
a similar effect estimate (summary RR=0.65, 95%CI: 0.48-0.88).
These results suggest that hormonal and reproductive factors influence risk of female
bladder cancer, and that processes mediated by steroid hormones may be important determinants
of the gender disparity in bladder cancer rates.

1

Chapter 1: Male infertility and variation in CAG repeat length in the androgen
receptor gene: A meta-analysis

1.1 Abstract

Many studies have investigated the association between male infertility and trinucleotide
repeat polymorphisms in the androgen receptor (AR) gene, but no comprehensive meta-analysis
of all published studies has been conducted. Our goals were to summarize published data on
associations between AR CAG and GGC repeat lengths and male infertility, and to investigate
sources of variation between study results.
We searched for reports published before October 2006 using Medline, PubMed and Web
of Science. All selected studies included the following: a case group with infertility as measured
by semen parameters, a control group of known or presumed fertile men, and measurement of
CAG and/or GGC repeat lengths among cases and controls. Thirty-nine reports were selected
based on these criteria, and 33 were ultimately included in the meta-analysis. One investigator
extracted data on sample size, mean and standard deviation of trinucleotide repeat length, and
study characteristics.
Estimates of the standardized mean difference (SMD) (95% confidence interval) were
0.19 (0.09-0.29) for the 33 studies, and 0.31 (0.14-0.47) for a sub-set of 13 studies that used more
stringent case and control selection criteria. Thus, in both groups, cases had statistically
significantly longer CAG repeat length than controls. Publication date appeared to be a
significant source of variation between studies.
This meta-analysis provides support for an association between increased androgen
receptor CAG length and idiopathic male infertility, suggesting that even subtle disruptions in the
androgen axis may compromise male fertility.

2

1.2 Introduction

Male factor infertility is poorly understood, and the etiology of nearly half of all cases is
unknown (1). It has been postulated that genetic factors may contribute to many cases of
idiopathic infertility, in particular those relating to defective spermatogenesis.
Androgens are required for male sex determination, development and spermatogenesis.
Androgen activity is mediated by the androgen receptor (AR), a member of the steroid receptor
superfamily. Receptor variants with diminished capacity to respond to androgens result in
androgen resistance, which compromises spermatogenesis. Additional features can also be
present, with severity depending upon the extent to which AR function is impaired. In the most
severe form, complete androgen insensitivity syndrome (CAIS), individuals with XY karyotype
have female phenotype, primary amenorrhea and markedly elevated levels of serum testosterone
and estradiol. In partial androgen insensitivity syndrome (PAIS, Reifenstein’s Syndrome),
patients have ambiguous genitalia (2). In the mildest form, patients with normal male phenotype
have abnormal spermatogenesis (3, 4). Based on androgen binding assays of fibroblasts from
infertile men, it has been estimated that androgen resistance may be present in 40% or more of
patients with idiopathic male infertility (5).
The AR is encoded by the androgen receptor gene (AR), located on chromosome Xq11-
12. The AR contains eight exons that encode three functional domains of the receptor:
transactivation domain (exon 1), DNA binding domain (exons 2 and 3) and ligand-binding
domain (exons 4-8) (6). Rare mutations that result in complete or partial androgen insensitivity
syndromes have been localized to the ligand-binding and DNA-binding domains (4). The
transactivation domain controls transcription of target genes. Two trinucleotide polymorphisms
in this domain vary in length in the population: a CAG repeat encoding a polyglutamine tract and
a GGC (GGN) repeat encoding a polyglycine tract.
3

Experimental research suggests that the number of repeats in the CAG tract is inversely
correlated with transcriptional activity of the AR protein (7). The usual range in repeat length is
nine to 36 repeats (8). Clinical findings have linked polyglutamine lengths of over 40 repeats
with reduced virilization and defective spermatogenesis among men affected by spinal bulbar
muscular atrophy, a fatal neuromuscular disease (9). Based on this evidence, it is postulated that
men with longer CAG repeats within the normal range may have subtle decreases in AR function
that result in reduced spermatogenesis.
Results of studies investigating this hypothesis are widely divergent. Some report
associations between infertility and longer repeats (1, 10-24), while others do not (25-46). It is
unknown whether differences between these studies, including race/ethnicity of study participants
and inconsistencies in case and control inclusion criteria, are responsible for conflicting findings.
This possibility can be investigated in meta-analyses that include statistical measures of
heterogeneity.
To our knowledge, no meta-analysis has been conducted to date analyzing results of all
published studies on this association. Two prior meta-analyses (19, 32) and one pooled analysis
(21) addressed sub-sets of published studies (12 studies, six studies and five studies, respectively)
and did not quantitatively investigate the impact of heterogeneity between studies on the overall
effect estimate. Our goals in preparing this report were to conduct a comprehensive meta-
analysis of the published literature summarizing data on associations between AR repeat length
polymorphisms and male infertility, and to investigate sources of heterogeneity that may have
influenced published results.
1.3 Materials and Methods
Study selection
We searched MEDLINE and PubMed for articles published in English until October
2006 describing associations between male infertility and CAG and/or GGC trinucleotide repeat
4

lengths in the AR. Search terms queried were: androgen receptor, male infertility, semen analysis,
polyglutamine, polyglycine, CAG, GGC and GGN. We screened identified publications by
reviewing titles and abstracts. Bibliographies of all original reports and review articles were
examined, and each was subjected to a citation search using Web of Science to identify additional
publications not retrieved through online searches.
Publications identified by any of the above procedures were reviewed, then selected for
possible inclusion in the meta-analysis if they fulfilled each of three criteria: (1) included a case
group with infertility as measured by semen parameters based on WHO guidelines (47), (2)
included a control group of known or presumed fertile men, and (3) reported measurement of
CAG and/or GGC repeat lengths among cases and controls. Thirty-nine reports met these
criteria.

Data extraction
A single reviewer extracted data from each of the 39 reports. The following qualitative
characteristics were noted: geographic location of the study population, demographic
characteristics of study participants (age and race/ethnicity), case and control definitions, case and
control exclusion criteria, and publication year. Quantitative data extracted were sample size and
mean and standard deviation (SD) of trinucleotide (CAG and/or GGC) repeat length for each
group of cases and controls. Data were either extracted directly from articles or calculated using
information provided in tables and figures. For several reports (1, 14, 16, 18, 23, 33, 36, 39, 43),
standard deviation was calculated from the standard error ) * ( SE n SD = . One report (28) did
not provide the data needed to calculate standard deviation, so it was estimated by using the P
value of the unpooled t test comparison of means between cases and controls: SD
cases
= SD
controls

5

=
controls cases
n n
Z
MD
1 1
* +
, in which Z represents the Z score of the P value from the unpooled t
test, and n
cases
and n
controls
represent the number of cases and controls (48). Among the 39
reports, seven did not present all information required to calculate or accurately estimate the
mean or standard deviation. We requested this information from authors, who provided detailed
data on three (12, 32, 41).

Data analysis
We implemented meta-analysis using Stata statistical software (Stata/SE 9.0, College
Station, TX). The overall standardized mean difference (SMD) and 95% confidence interval
were calculated to estimate differences in repeat length between cases and controls. To determine
the SMD, mean differences in number of repeats between cases and controls in each study were
weighted by sample size. A random effects model was used, taking into account within-study
and between-study variability. We graphically displayed the SMD along with mean differences
and confidence intervals from each study in a Forrest plot, and assessed the possibility of
publication bias using Egger’s unweighed regression asymmetry test (49).
To examine dispersion of data, we created Begg’s funnel plots, which display for each
study the SMD versus the standard error of the SMD. Results distributed within the “funnel”
defined by 95% confidence limits can be interpreted as variation due to sampling error. Variation
due to differences in design and conduct of the studies is termed statistical heterogeneity, and
may result in over-dispersion of results (e.g. outside the confidence limits). We used four
methods to investigate potential sources of heterogeneity.
First, to learn whether the use of stricter definitions of fertility influenced results of the
meta-analysis, we identified a sub-set of 13 studies (1, 13, 15, 16, 19-21, 23, 27, 29, 36, 38, 43)
6

that used more stringent case and control criteria. Cases with known causes of infertility
(including obstruction, infections, anatomic defects, defined genetic or endocrine disorders, and
chromosomal abnormalities) were excluded from these studies; and controls were confirmed to
have either sperm concentration >20 x 10
6
/mL and/or to have reported paternity of one or more
children by natural conception. We further restricted cases to those with semen concentration
<20 x 10
6
/mL (in accordance with WHO guidelines (47)), including in the sub-set only studies
that provided this information. We did not consider sperm motility and morphology because
these parameters were rarely reported. For this sub-set of studies, we calculated the overall SMD
and 95% confidence interval as described above, and created Forrest and Begg’s funnel plots.
Second, to explore possible effects of other study characteristics, we conducted a series
of analyses stratified individually on: race/ethnicity of study participants (Caucasian, Asian, study
population composed of several racial/ethnic groups (i.e. mixed), or unspecified), geographic
location of the study population (Europe, Asia, United States, or other), and type of control group
(proven fathers and/or normozoospermic men, fertile men (no evidence of fertility specified), or
unselected men). Stratified analyses were conducted on both the full set and the sub-set of 13
studies. In the sub-set, type of control group was stratified into fathers versus normozoospermic
men.
Using data from studies that provided mean repeat length of specific case groups, we
calculated SMDs to compare azoospermic cases (no sperm) and oligozoospermic cases (sperm
concentration >0 to <20 x 10
6
/mL) separately with controls. For each group of cases, the SMD
and 95% confidence interval were calculated.
Third, we quantified the degree of heterogeneity by calculating the I
2
statistic, which
estimates the proportion of variation in SMDs that is due to heterogeneity between studies, as
opposed to sampling variation (50). I
2
ranges from 0-100%, with higher values indicating greater
7

degrees of heterogeneity (0-30%, mild heterogeneity; 30-50%, moderate heterogeneity; 50-100%,
notable heterogeneity) (50). I
2
was calculated from the Q-statistic, a χ
2
statistic used to test for
the presence of heterogeneity in meta-analyses (I
2
= (Q – degrees of freedom)/Q) (50). We
calculated I
2
statistics for the overall analyses of the full set and the sub-set of 13 studies, and for
the stratified analyses.
Fourth, we conducted meta-regression analyses on the full set and the sub-set of 13
studies to investigate effects of individual study characteristics on the SMD while controlling for
effects of other study characteristics. The SMD was modeled as the outcome weighted on the
standard error of the SMD, and study characteristics that may influence heterogeneity were
included as covariates in each of two models. In model I covariates were: race/ethnicity
(Caucasian versus other), geographic location (Europe versus other), and type of control (fathers
versus all others). In model II publication date was added to the covariates in model I. We
considered covariates with p<0.05 to be modifiers of the effect of trinucleotide repeat length on
the risk of infertility, and therefore to be possible sources of heterogeneity.
To investigate trends in case and control repeat length over time, we conducted separate
linear regression analyses of case repeat length and control repeat length on publication date.

1.4 Results

Study characteristics
Of the 39 articles identified, 38 reported data on the CAG repeat (1, 10-46), five on both the
CAG and GGC repeats (10, 28, 38, 41, 46), and one on the GGC repeat (51). Among studies
conducted on the GGC repeat, none reported statistically significant associations between GGC
repeat length and infertility. Only two provided data required for the meta-analysis (38, 51), and
data for a third was provided by the author (41). Due to the scant data available, no formal meta-
8

analysis was conducted on the GGC repeat. Among articles addressing the CAG repeat, four
were excluded because they did not provide the required data and no additional information was
received from authors (10, 25, 30, 46).
In all, data from 33 independent studies on the CAG repeat were included in this meta-
analysis (1, 11-24, 27-29, 32-45). One article reported on two independent study groups, one
from the United States and one from Singapore, so these data were included as two separate case-
control series (14). Two articles compared the same control group to each of two cases series
(19, 31). Data reported in these articles were analyzed as follows: in most analyses, data from the
larger series (19) were used; however, data from the smaller series (31) were used in the analyses
of case sub-groups (azoospermic and oligozoospermic) because this information was not reported
for the larger series. Two additional articles (26, 28) presented data on the same case-control
series, so data from only one report (28) were included. Altogether, data for 3,027 cases and
2,722 controls extracted from 33 reports were included in these analyses.
Characteristics of all 39 articles selected for possible inclusion are shown in Table 1.1.
Publication dates ranged from 1997 to 2006. Among the 33 studies included in the meta-analysis,
racial/ethnic backgrounds of study participants were diverse: 17 studies enrolled Caucasian men,
seven enrolled Asian men, five enrolled men of mixed races, and four did not specify
race/ethnicity of men enrolled. Study participants were enrolled in numerous geographic
locations: 15 studies were conducted in Europe, seven in Asia, four in the United States, and
seven in other countries. In most reports, the authors specified that cases and controls were of
similar racial/ethnic background and age.

9

Associations between CAG Repeat Length and Infertility
Analysis of the full set of 33 studies revealed statistically significantly longer CAG repeat length
among cases compared with controls (SMD = 0.19, 95% CI: 0.09-0.29) (Table 1.2A), as
illustrated by the Forrest plot of results (Figure 1.1A). The corresponding funnel plot shows
over-dispersion of the data (Figure 1.1B), an indication of greater differences between studies
than expected from sampling variation alone. Egger’s test for publication bias was significant for
the full set of studies (p = 0.04).
In the sub-set of 13 studies that used stringent definitions of case and control status, the SMD
was larger than in the full set (SMD = 0.32, 95% CI: 0.14-0.50) (Table 1.2B). The corresponding
Forrest plot suggests a decrease over time in the mean difference between cases and controls
(Figure 1.2A). The funnel plot reveals greater dispersion of the data than expected from sampling
variation alone (Figure 1.2B), but Egger’s test for publication bias was not statistically significant
(p = 0.40).
Among the complementary sub-set of 20 studies that did not use stringent case and control
definitions, the SMD was notably smaller (SMD = 0.12, 95% CI: -0.005-0.24). The difference
between SMDs estimated for the two sub-sets was highly significant (p = 0.0007), indicating that
case and control definitions likely influenced study results.
Statistical assessment of heterogeneity
Stratified analyses of the full set of studies revealed differences in SMDs between some sub-
groups defined by race, geographic location and control type. SMDs were slightly larger for
Asian and mixed race populations than Caucasian populations, but differences were not
statistically significant (p = 0.68) (Table 1.2A). There were statistically significant differences
between the SMDs calculated for studies conducted in Europe, Asia, the United States and other
countries (p = 0.02). Studies using proven fathers or confirmed normozoospermic men as
10

controls found greater differences between cases and controls than studies that used other control
types, but these differences were not statistically significant (p = 0.15). Among the sub-set of 13
studies, no significant differences in SMDs were detected between strata defined by race or
geographic location (p = 0.82 and 0.76, respectively), but marginally significant differences were
found for control type (p = 0.06) (Table 1.2B).
Specific data on azoospermic and oligozoospermic cases were provided by 20 (1, 13-15, 18,
21, 22, 27, 29, 31-33, 36-41, 43) and 15 (1, 14, 15, 22, 24, 27, 32, 36-41, 43) studies,
respectively. Among both azoospermic and oligozoospermic cases repeat lengths were
significantly longer than among controls. However, SMD estimates for both types of cases were
similar in magnitude to the overall SMD for all 33 studies (Table 1.2A). Results were similar
when data were restricted to studies that used more stringent case and control definitions (Table
1.2B).
I
2
statistics calculated for unstratified analyses of the full set and the sub-set of studies
were 69% and 64%, respectively, indicating that more than half of the variation in SMDs may be
due to between-study heterogeneity (Table 1.2). In analyses stratified on race/ethnicity,
geographic location, control type and case type, I
2
statistics ranged from 14 to 88%, indicating
that a notable amount of heterogeneity remained within strata.
Meta-regression analyses addressing joint effects of multiple study characteristics
identified race and geographic location as significant modifiers of the SMD in all 33 studies (p =
0.001 and 0.03, respectively using Model I; p = 0.02 and 0.001, respectively using Model II), but
not in the sub-set of 13 (Table 1.3). Modification by publication date was highly significant in
the sub-set of 13 studies (p = 0.005). To better understand the influence of publication date in
this sub-set, we conducted separate linear regression analyses of case and control repeat length on
11

publication date. There was a highly significant decrease in repeat length over time among cases
(p = 0.009), but no apparent time trend among controls (p = 0.70) (data not shown).

Table 1.1 Published studies on associations between male infertility and length of the CAG and GGC (GGN) trinucleotide tracts in the
androgen receptor gene. Geographic location of the study population, race/ethnicity of study participants, and the mean and standard
deviation of CAG and/or GGC repeat lengths among case and control groups are provided for each study.
Study Location Race/Ethnicity Cases Controls
# of men Mean CAG Mean GGC # of men Mean CAG Mean GGC
length ± SD length ± SD length ± SD length ± SD
Tut et al. 1997
1
(10) Singapore Asian 153 ND ND 72 ND ND
Giwercman et al. 1998
1
(25) Sweden Caucasian 33 ND 294 ND
Dowsing et al. 1999 (1) Australia Mixed 30 23.2 ±3.8 32 20.5 ±1.7
Komori et al. 1999 (11) Japan Asian 59 21.2 ±4.2 36 21.4 ±3.5
Legius et al. 1999
2
(12) Belgium Caucasian 223 21.8 ±2.6 181 21.3 ±2.4
Yoshida et al. 1999 (13) Japan Asian 41 26.5 ±3.5 48 23.9 ±2.9
Dadze et al. 2000 (27) Germany Caucasian 119 22.0 ±3.2 22 20.8 ±3.3
Hiort et al. 2000
1,3
(28) Germany Caucasian 180 23.0 ±3.3 22.0 ±ND 53 24.0 ±3.3 23.0 ±ND
Mifsud et al. 2001 (USA) (14) USA Mixed 95 22.0 ±3.0 55 20.7 ±3.9
Mifsud et al. 2001 (Sing.) (14) Singapore Asian 120 23.1 ±3.1 87 22.4 ±3.0
Patrizio et al. 2001 (15) USA Caucasian 69 23.5 ±3.4 45 22.0 ±2.8
Sasagawa et al. 2001 (29) Japan Asian 30 23.4 ±2.9 51 23.7 ±3.2
Wallerand et al. 2001 (16) France Caucasian 37 23.9 ±3.0 50 22.2 ±2.8
Kukuvitis et al. 2002
1
(30) Greece Caucasian 109 ND 64 ND
von Eckardstein et al. 2001
4
(31)Germany Unspecified 43 20.5 ±2.8 131 19.9 ±3.1
Madgar et al. 2002 (17) Israel Mixed 61 18.6 ±3.0 50 16.6 ±2.7
Pan et al. 2002 (18) Taiwan Asian 48 23.0 ±4.2 47 21.0 ±2.7
Rajpert De-Meyts et al. 2002
2
(32) Denmark Caucasian 113 21.5 ±2.8 87 21.5 ±3.4
Thangaraj et al. 2002 (33) India Unspecified 280 21.7 ±3.0 201 22.4 ±2.7
Van Golde et al. 2002 (34) Netherlands Unspecified 75 22.2 ±3.1 70 21.7 ±3.4
Asatiani et al. 2003
5
(19) Germany Caucasian 99 21.6 ±3.0 131 19.9 ±3.1
Casella et al. 2003
6
(20) USA Mixed 70 22.0 ±3.2 55 21.0 ±3.9
Dhillon et al. 2003 (35) India Unspecified 183 22.2 ±1.5 59 21.5 ±1.4
Erasmuson et al. 2003 (36) New Zealand Caucasian 105 21.5 ±3.1 93 21.0 ±2.7
Lund et al. 2003 (37) Finland Caucasian 90 21.9 ±2.6 149 22.4 ±2.8
12

Table 1.1 (Continued)
Mengual et al. 2003 (21) Spain Caucasian 102 23.2 ±2.8 96 22.4 ±2.8
Tse et al. 2003
7
(22) China Asian 85 23.1 ±3.9 45 23.0 ±3.1
Ferlin et al. 2004 (38) Italy Caucasian 163 21.7 ±2.8 17.2 ±1.9 115 21.6 ±3.3 17.0 ±1.7
Hadjkacem et al. 2004
8
(39) Tunisia Unspecified 65 20.9 ±3.0 98 21.1 ±3.1
Jeong et al. 2004 (40) Korea Asian 135 21.6 ±3.3 206 21.2 ±2.9
Milatiner et al. 2004
9
(23) Israel Mixed 61 21.6 ±3.0 111 21.5 ±2.6
Ruhayel et al. 2004
2, 10
(41) Sweden Caucasian 85 22.3 ±2.6 23.0 ±2.1 223 21.9 ±3.1 23.0 ±2.3
Lavery et al. 2005 (42) Ireland Caucasian 66 23.3 ±2.4 77 23.1 ±2.3
Tufan et al. 2005(43) Turkey Caucasian 30 22.3 ±2.3 32 22.4 ±3.1
Canale et al. 2006 (44) Italy Caucasian 29 21.4 ±2.0 91 21.5 ±1.7
Dakouane et al. 2006 (45) France Caucasian 15 21.9 ±2.2 13 22.8 ±3.0
Katagiri et al. 2006 (24) USA Caucasian 64 22.2 ±3.0 13 19.3 ±5.0
Rajender et al. 2006
1
(51) India Mixed 395 ND 21.51 ±1.2 200 ND 21.51 ±1.0
Singh et al. 2006
1
(46) India Unspecified 399 ND ND 100 ND ND
1
ND = No data reported or not enough information provided to calculate mean and/or standard deviation
2
Data obtained through correspondence with the authors
3
Hiort et al. 1999 presented the identical data; standard deviation for CAG repeat was estimated using the method of Zeegers et al. 2004, but no estimation could
be made for the GGC repeat since no p value was provided.
4
The two case groups and two control groups were combined and a weighted mean and standard deviation for each case and control group were calculated;
same control group as Asatiani et al. 2003
5
The two control groups were combined, and weighted mean and standard deviation were calculated. Same control group as Von Eckardstein et al. 2001
6
Median used in place of mean
7
The two case groups were combined, and weighted mean and standard deviation were calculated.
8
The case group excludes normozoospermic men and provides a weighted mean and standard deviation of the azoospermic and oligozoospermic men.
9
Cases and controls were defined by semen concentration (<20x 10
6
/mL = cases, ≥ 20 x 10
6
/mL = controls).
10
85 cases were genotyped for the CAG repeat and 81 for the GGN repeat.
13

Table 1.2 Results of overall and stratified meta-analyses. Standardized mean differences (SMDs) are summary estimates of the mean
difference in repeat length between cases and controls. I
2
statistics estimate the proportion of variation in SMDs that is due to
heterogeneity between studies. SMDs, their 95% confidence intervals (95% CIs) and I
2
statistics are provided for overall analyses, and for
analyses conducted within strata defined by selected study characteristics. (A). Results for the full set of 33 studies. (B). Results for the
sub-set of 13 studies that used more stringent case-control criteria. (C). Results for the sub-set of 20 studies that did not use stringent
case-control criteria.

Description # Studies #Cases # Controls SMD (95% CI)
1
I
2
(%)
2

A. All Studies 33 3,027 2,722 0.19 (0.09, 0.29) 69
Strata

Race
Caucasian 17 1,589 1,471 0.14 (0.01, 0.28) 64
Asian 7 518 520 0.22 (0.009, 0.43) 61
Mixed/Unspecified 9 920 731 0.26 (0.018, 0.51) 81
p = 0.68
3

Mixed only 5 317 303 0.43 (0.15, 0.71) 64
Unspecified only 4 603 428 0.07 (-0.26, 0.40) 83

Geographic Location
Europe 15 1,426 1,390 0.10 (-0.03, 0.23) 61
Asia 7 518 520 0.22 (0.009, 0.43) 61
USA 4 298 168 0.42 (0.22, 0.61) 14
Other 7 785 644 0.25 (-0.05, 0.54) 85
p = 0.02
3

Type of Control Group
Proven Fathers and/or
Normozoospermic Men 22 1,784 1,724 0.23 (0.10, 0.35) 67
Fertile Men (no details) 8 855 598 0.15 (-0.09, 0.40) 77
Unselected Men 3 388 400 0.06 (-0.20, 0.31) 48
p = 0.15
3
14

Table 1.2 (Continued)
Sperm Concentration of Case Group
4

Azoospermic 20 897 1,863 0.21 (0.02, 0.39) 75
Oligozoospermic 15 911 1,323 0.18 (0.05, 0.30) 40

B. Sub-set of 13 studies that used
stringent case-control criteria
5
13 956 881 0.31 (0.14, 0.47) 64
Strata

Race
Caucasian 8 724 584 0.28 (0.09, 0.46) 58
Asian 2 71 99 0.36 (-0.53, 1.26) 88
Mixed 3 161 198 0.38 (-0.07, 0.82) 74
p = 0.82
3
Geographic Location
Europe 6 550 446 0.27 (0.03, 0.51) 67
Asia 2 71 99 0.36 (-0.53, 1.26) 88
USA 2 139 100 0.37 (0.11, 0.63) ---
6

Other 3 196 236 0.32 (-0.09, 0.74) 75
p = 0.76
3

Type of Control Group
7

Proven Fathers 7 457 332 0.37 (0.14, 0.60) 53
Normozoospermic Men 4 370 367 0.23 (-0.06, 0.51) 71
p = 0.06
3
Sperm Concentration of Case Group
4
Azoospermic 10 317 665 0.33 (0.05,0.60) 70

Oligozoospermic 6 401 339 0.29 (0.01, 0.56) 64
15

Table 1.2 (Continued)
C. Sub-set of 20 studies that did not
use stringent case-control criteria 20 2,071 1,841 0.12 (-0.005, 0.24) 68

1
SMD = Standardized mean difference, 95% CI = 95% Confidence interval
2
I
2
is the proportion of variability that may be attributed to between-study variation
3
P values represent tests for differences in the SMDs between stratum (heterogeneity tests)
4
Only studies that provided specific data on sub-groups of cases are included; no significance test could be conducted because the same control groups
were used for comparison in several of the included studies.
5
The sub-set of studies includes those with a case definition of idiopathic infertility and case semen concentration of <20 x 10
6
/mL, and controls defined
as fathers and/or with normal semen concentration based on WHO criteria ( ≥ 20 x 10
6
/mL).
6
The value of the I
2
statistic was undefined for this sub-group due to the extremely small value from the Q test
7
Data from two studies, Asatiani et al. 2003 and Sasagawa et al. 2001, were excluded from this analysis due to overlap in control
16
17

Table 1.3 Results of multivariate meta-regression analyses implemented to investigate effects of
individual study characteristics on the standardized mean difference while controlling for effects
of other study characteristics. Results are presented for the full set of 33 studies and the sub-set of
13 studies that used more stringent case-control criteria.
Description # Studies Model I Model II
P values
1
P values
2

All studies 33
Race 0.03* 0.02
*

Geographic Location 0.001* 0.001
*

Control Type 0.56 0.42
Publication Date ---- 0.08

Sub-set of studies that used
stringent case-control criteria 13
Race 0.88 0.67
Geographic Location 0.19 0.20
Control Type 0.84 0.53
Publication Date ---- 0.005
*

*
Covariate statistically significantly modifies the effect of CAG repeat length on infertility
1
Model I includes the following covariates: race (Caucasian versus other), geographic location (Europe
versus other), and control type (fathers versus other control types)
2
Model II includes all covariates in Model I and publication date (continuous)

18

Figure 1.1A. Forrest plot of the full set of 33 studies showing differences in CAG repeat length
between cases and controls for each study, and the overall standardized mean difference
determined from the meta-analysis. For each study, the mean difference in CAG repeat length
between cases and controls is displayed as a box, and corresponding 95% confidence intervals are
displayed as a horizontal line. Mean differences to the right of zero (solid vertical line) represent
longer CAG repeat length in cases than controls. The size of each box is proportional to the
sample size of the corresponding study. Closed boxes represent the sub-set of 13 studies that
used stringent case and control criteria; open boxes represent the remaining studies. The overall
estimate of the SMD from the meta-analysis is displayed as a diamond (point estimate
represented by the top and bottom points of the diamond; 95% confidence interval represented by
the left and right points of the diamond).

19

Figure 1.1B. Begg’s Funnel plot showing standardized mean difference (SMD) estimates for
each of the 33 studies plotted against the standard error of each SMD (circles). The overall SMD
is presented as a horizontal line and estimated 95% confidence limits as a funnel defined by
diagonal lines. Results distributed within the 95% confidence limits can be interpreted as
variation due to sampling error; results distributed outside may result from statistical
heterogeneity between studies. The sub-set of 13 studies that used stringent case and control
criteria are depicted as closed circles; those for the remaining studies are depicted as open circles.

20

Figure 1.2A. Forrest plot of the sub-set of 13 studies that used stringent case and control criteria
showing differences in CAG repeat length between cases and controls for each study, and the
overall standardized mean difference calculated for this sub-set.

Figure 1.2B. Begg’s Funnel plot showing standardized mean difference (SMD) estimates for
each of the 13 studies from the sub-set that used more stringent case-control criteria plotted
against the standard error of each SMD (closed circles). The overall SMD for this sub-set is
presented as a horizontal line and estimated 95% confidence limits as a funnel defined by
diagonal lines.

21

1.5 Discussion

Results of this comprehensive meta-analysis provide support for the hypothesis that
longer AR CAG repeat lengths are associated with reduced male fertility. Since these variants
reportedly encode receptor protein with diminished function, this finding is consistent with the
suggestion first made in the pre-genome era that limited function of the AR may contribute to
idiopathic infertility. However, androgen action is required for both male sexual morphogenesis
and spermatogenesis following puberty (52), and men with idiopathic infertility have a normal
male phenotype. Therefore, if the association we report is causal, functional deficits encoded by
longer CAG tracts must interfere with androgen action required for spermatogenesis without
disrupting male sexual morphogenesis.
Spermatogenesis is regulated by androgens in a largely paracrine fashion. Leydig cells of
the adult testis secrete testosterone, but adult germ cells reportedly do not express the AR.
Therefore, AR-mediated effects of androgens on spermatogenesis must involve the action of
somatic cells. Experimental research has shown that targeted disruption of AR expression only
in Sertoli cells creates mouse models with the key features of idiopathic male infertility:
phenotypically normal males with severely disrupted spermatogenesis (53, 54). It is therefore
reasonable to speculate that AR variants with limited Sertoli cell function may contribute to
spermatogenetic deficits in men with idiopathic infertility. Moreover, because longer
polyglutamine tracts appear to reduce AR function far less than mutations that cause defined
androgen insensitivity syndromes, our results suggest that other determinants of subtle variation
in androgen response may also influence male fertility.
This meta-analysis not only substantiates an association between CAG repeat length and
infertility, but also identifies sample size and differences in study design as sources of variation
22

between earlier reports. To achieve 80% power to detect an SMD of magnitude estimated by the
meta-analysis (SMD=0.20, standard deviation of repeat length=3.0), 3,533 cases and 3,533
controls are needed (55). Although the aggregate data addressed in the meta-analysis approach
this sample size, samples used in each of the 33 individual studies were extremely small by
comparison.
Stringency of case and control definitions is an important determinant of differences in
repeat length between cases and controls, as estimated by the SMD. Meta-analysis revealed a
steady increase in the SMD as we examined data sets defined by increasingly strict definitions:
among 20 studies that did not use stringent definitions, there was no statistical evidence of a
difference between cases and controls. When these data were combined with those from 13
studies that used more stringent definitions, cases were found to have significantly longer CAG
repeat length than controls. Even larger SMDs were observed when the sub-set of 13 studies was
analyzed separately, particularly when controls were restricted to proven fathers (SMD = 0.37,
95% CI: 0.14-0.60). We anticipate that even this value under-represents the difference in CAG
repeat length that influences male infertility, because among men with idiopathic infertility there
is inevitably an unknown proportion whose infertility does not involve this polymorphism.
Stratified and meta-regression analyses identified only publication date as an additional
source of variation within the sub-set of 13 studies, with estimated SMDs tending to increase
over time (Figure 1.2A). Repeat lengths among controls were nearly constant, suggesting that
investigators sampled controls from similar populations over time. However, average repeat
length among cases declined during the interval 1999-2005. This decline may be attributable to
changing patterns of referral to infertility clinics during this period, with the introduction of new
23

therapies such as intracytoplasmic sperm injection influencing men with a wider array of
conditions to seek treatment.
To bring results of this meta-analysis to clinical decision-making, answers to three
questions are desired: 1) What range of AR CAG repeat lengths predisposes to idiopathic
infertility? 2) What risk of infertility is associated with each length in this range? (3) Will AR-
associated predisposition to infertility be transmitted to offspring conceived by in vitro
fertilization using sperm of infertile men with longer repeats? The summary nature of published
data included in the meta-analysis does not permit us to address questions 1 and 2 in this
analysis. Therefore, collection of data required to answer these questions is now a priority. As a
refinement to envisioned research we recommend measurement of additional genotypic variants
in the AR, including single nucleotide polymorphisms and the GGC repeat sequence. These data
will allow investigators to address the possibility that multiple variants in the AR may act in
conjunction to influence fertility, and to rule out the possibility that the association reported here
is substantially influenced by unmeasured variants in linkage disequilibrium with longer CAG
repeats. Because the AR is located on the X chromosome, a man’s copy of the AR is normally
transmitted to all of his daughters but none of his sons. Therefore, any predisposition to
infertility encoded by the AR is predicted to be transmitted by a man to none of his sons, and on
expectation, to one-quarter of his grandsons.
In conclusion, results of this comprehensive meta-analysis suggest that variation in the
AR polyglutamine tract may be a determinant of infertility in otherwise healthy men. Since
longer polyglutamine tracts are far more common than mutations associated with complete or
partial androgen insensitivity syndromes, this polymorphism may influence fertility in a much
24

larger proportion of men. In light of this result, studies providing empiric estimates of the risk of
infertility associated with individual tract lengths are now a pressing priority.

25

Chapter 1 References

Aiman J, Griffin JE, Gazak JM, et al. Androgen insensitivity as a cause of infertility in otherwise
normal men. New Engl J Med 1979;300:223-7. (3)

Aiman J, Griffin JE The frequency of androgen receptor deficiency in infertile men. J Clin
Endocrinol Metab 1982;54:725-32. (5)

Andrew SE, Goldberg YP, Hayden MR Rethinking genotype and phenotype correlations in
polyglutamine expansion disorders. Hum Mol Genet 1997;6:2005-10. (8)

Asatiani K, von Eckardstein S, Simoni M, et al. CAG repeat length in the androgen receptor gene
affects the risk of male infertility. Int J Androl 2003;26:255-61. (19)

Canale D, Caglieresi C, Moschini C, et al. Androgen receptor polymorphism (CAG repeats) and
androgenicity. Clin Endocrinol (Oxf) 2005;63:356-61. (44)

Casella R, Maduro MR, Lipshultz LI, et al. Significance of the polyglutamine tract polymorphism
in the androgen receptor. Urol 2001;58:651-6. (6)

Casella R, Maduro MR, Misfud A, et al. Androgen receptor gene polyglutamine length is
associated with testicular histology in infertile patients. J Urol 2003;169:224-7. (20)

Chamberlain NL, Driver ED, Miesfeld RL The length and location of CAG trinucleotide repeats
in the androgen receptor N-terminal domain affect transactivation function. Nucleic Acids Res
1994;22:3181-6. (7)

Dadze S, Wieland C, Jakubiczka S, et al. The size of the CAG repeat in exon 1 of the androgen
receptor gene shows no significant relationship to impaired spermatogenesis in an infertile
Caucasoid sample of German origin. Mol Hum Reprod 2000;6:207-14. (27)

Dakouane-Giudicelli M, Legrand B, Bergere M, et al. Association between androgen receptor
gene CAG trinucleotide repeat length and testicular histology in older men. Fertil Steril
2006;86:873-7. (45)

De Gendt K, Swinnen JV, Saunders PT, et al. A Sertoli cell-selective knockout of the androgen
receptor causes spermatogenic arrest in meiosis. PNAS 2004;101:1327-32. (53)

Denolet E, De Gendt K, Allemeersch J, et al. The effect of a sertoli cell-selective knockout of the
androgen receptor on testicular gene expression in prepubertal mice. Mol Endocrinol
2006;20:321-34. (54)

Dhillon VS, Husain SA Cytogenetic and molecular analysis of the Y chromosome: absence of a
significant relationship between CAG repeat length in exon 1 of the androgen receptor gene and
infertility in Indian men. Int J Androl 2003;26:286-95. (35)

Dowsing AT, Yong EL, Clark M, et al. Linkage between male infertility and trinucleotide repeat
expansion in the androgen-receptor gene. Lancet 1999;354:640-3. (1)
26

Erasmuson T, Sin IL, Sin FY Absence of association of androgen receptor trinucleotide
expansion and poor semen quality. Int J Androl 2003;26:46-51. (36)

Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis detected by a simple,
graphical test. BMJ 1997;315:629-34. (49)

Ferlin A, Bartoloni L, Rizzo G, et al. Androgen receptor gene CAG and GGC repeat lengths in
idiopathic male infertility. Mol Hum Reprod 2004;10:417-21. (38)

Giwercman YL, Xu C, Arver S, et al. No association between the androgen receptor gene CAG
repeat and impaired sperm production in Swedish men. Clin Genet 1998;54:435-6. (25)

Hadjkacem L, Hadj-Kacem H, Boulila A, et al. Androgen receptor gene CAG repeats length in
fertile and infertile Tunisian men. Ann Genet 2004;47:217-24. (39)

Higgins JP, Thompson SG Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539-
58. (50)

Hiort O, Horter T, Schulze W, et al. Male infertility and increased risk of diseases in future
generations. Lancet 1999;354:1907-8. (26)

Hiort O, Holterhus PM, Horter T, et al. Significance of mutations in the androgen receptor gene
in males with idiopathic infertility. J Clin Endocrinol Metab 2000;85:2810-5. (28)

Jeong YM, Park JH, Choi JS, et al. The study of trinucleotide repeat length in the androgen
receptor gene of Korean infertile males. Korean J Genet 2004;26:269-275. (40)

Katagiri Y, Neri QV, Takeuchi T, et al. Androgen receptor CAG polymorphism (Xq11-12) status
and human spermatogenesis: a prospective analysis of infertile males and their offspring
conceived by intracytoplasmic sperm injection. Int J Mol Med 2006;18:405-13. (24)

Kelsey JL, Whittemore AS, Evans AS, et al. Methods in Observational Epidemiology, Second ed.
Oxford University Press, New York, 1996. (55)

Komori S, Kasumi H, Kanazawa R, et al. CAG repeat length in the androgen receptor gene of
infertile Japanese males with oligozoospermia. Mol Hum Reprod 1999;5:14-6. (11)

Kukuvitis A, Georgiou I, Bouba I, et al. Association of oestrogen receptor alpha polymorphisms
and androgen receptor CAG trinucleotide repeats with male infertility: a study in 109 Greek
infertile men. Int J Androl 2002;25:149-52. (30)

La Spada AR, Wilson EM, Lubahn DB, et al. Androgen receptor gene mutations in X-linked
spinal and bulbar muscular atrophy. Nature 1991;352:77-9. (9)

Lavery R, Houghton JA, Nolan A, et al. CAG repeat length in an infertile male population of
Irish origin. Genetica 2005;123:295-302. (42)

27

Legius E, Vanderschueren D, Spiessens C, et al. Association between CAG repeat number in the
androgen receptor and male infertility in a Belgian study. Clin Genet 1999;56:166-7. (12)

Lund A, Tapanainen JS, Lahdetie J, et al. Long CAG repeats in the AR gene are not associated
with infertility in Finnish males. Acta Obstet Gynecol Scand 2003;82:162-6. (37)

Madgar I, Green L, Kent-First M, et al. Genotyping of Israeli infertile men with idiopathic
oligozoospermia. Clin Genet 2002;62:203-7. (17)

McLachlan RI, O'Donnell L, Meachem SJ, et al. Identification of specific sites of hormonal
regulation in spermatogenesis in rats, monkeys, and man. Recent Progress Hormone Res
2002;57:149-79. (52)

Mengual L, Oriola J, Ascaso C, et al. An increased CAG repeat length in the androgen receptor
gene in azoospermic ICSI candidates. J Androl 2003;24:279-84. (21)

Mifsud A, Sim CK, Boettger-Tong H, et al. Trinucleotide (CAG) repeat polymorphisms in the
androgen receptor gene: molecular markers of risk for male infertility. Fertil Steril 2001;75:275-
81. (14)

Milatiner D, Halle D, Huerta M, et al. Associations between androgen receptor CAG repeat
length and sperm morphology. Hum Reprod 2004;19:1426-30. (23)

Pan H, Li YY, Li TC et al. Increased (CTG/CAG)(n) lengths in myotonic dystrophy type 1 and
Machado-Joseph disease genes in idiopathic azoospermia patients. Hum Reprod 2002;17:1578-
83. (18)

Patrizio P, Leonard DG, Chen KL, et al. Larger trinucleotide repeat size in the androgen receptor
gene of infertile men with extremely severe oligozoospermia. J Androl 2001;22:444-8. (15)

Quigley CA, De Bellis A, Marschke KB, et al. Androgen receptor defects: historical, clinical,
and molecular perspectives.[erratum appears in Endocr Rev 1995 16:546]. Endocr Rev
1995;16:271-321. (4)

Rajender S, Rajani V, Gupta NJ, et al. No association of androgen receptor GGN repeat length
polymorphism with infertility in Indian men. J Androl 2006;27:785-9. (51)

Rajpert-De Meyts E, Leffers H, Petersen JH, et al. CAG repeat length in androgen-receptor gene
and reproductive variables in fertile and infertile men. Lancet 2002;359:44-6. (32)

Ruhayel Y, Lundin K, Giwercman Y, et al. Androgen receptor gene GGN and CAG
polymorphisms among severely oligozoospermic and azoospermic Swedish men. Hum Reprod
2004;19:2076-83. (41)

Sasagawa I, Suzuki Y, Ashida J, et al. CAG repeat length analysis and mutation screening of the
androgen receptor gene in Japanese men with idiopathic azoospermia. J Androl 2001;22:804-8.
(29)

28

Singh R, Deepa SR, Madhavi S, et al. Male infertility: no evidence of involvement of androgen
receptor gene among Indian men. J Andrology 2006;27:102-5. (46)

Thangaraj K, Joshi MB, Reddy AG, et al. CAG repeat expansion in the androgen receptor gene
is not associated with male infertility in Indian populations. J Androl 2002;23:815-8. (33)

Tse JY, Liu VW, Yeung WS, et al. Molecular analysis of the androgen receptor gene in Hong
Kong Chinese infertile men. J Assist Reprod Genet 2003;20:227-33. (22)

Tufan AC, Satiroglu-Tufan NL, Aydinuraz B, et al. No association of the CAG repeat length in
exon 1 of the androgen receptor gene with idiopathic infertility in Turkish men: implications and
literature review. Tohoku J Exp Med 2005;206:105-15. (43)

Tut TG, Ghadessy FJ, Trifiro MA, et al. Long polyglutamine tracts in the androgen receptor are
associated with reduced trans-activation, impaired sperm production, and male infertility. J Clin
Endocrinol Metab 1997;82:3777-82. (10)

Van Golde R, Van Houwelingen K, Kiemeney L, et al. Is increased CAG repeat length in the
androgen receptor gene a risk factor for male subfertility? J Urol 2002;167:621-623. (34)

von Eckardstein S, Syska A, Gromoll J, et al. Inverse correlation between sperm concentration
and number of androgen receptor CAG repeats in normal men. J Clin Endocrinol Metab
2001;86:2585-90. (31)

Wallerand H, Remy-Martin A, Chabannes E, et al. Relationship between expansion of the CAG
repeat in exon 1 of the androgen receptor gene and idiopathic male infertility. Fertil Steril
2001;76:769-74. (16)

WHO Laboratory Manual for the Examination of Human Semen and Sperm-Cervical Mucus
Interaction, 4th ed. Cambridge University Press, Cambridge, 1999. (47)

Wilson JD Roosters B Reifenstein's syndrome and hormone resistance. New Engl J Med 1977;
297:386-7. (2)

Yoshida KI, Yano M, Chiba K,et al. CAG repeat length in the androgen receptor gene is
enhanced in patients with idiopathic azoospermia. Urol 1999; 54:1078-81. (13)

Zeegers MP, Kiemeney LA, Nieder AM, et al. How strong is the association between CAG and
GGN repeat length polymorphisms in the androgen receptor gene and prostate cancer risk?
Cancer Epidemiol Biomarkers Prev 2004;13:1765-71. (48)

29

Chapter 2: Length of the polymorphic CAG tract in the androgen receptor gene
varies by histologic type in testicular germ cell tumor cases from California

2.1 Abstract

Increasing rates of testicular germ cells tumors (TGCTs) over the last 40 years suggest
environmental factors are involved in disease etiology, but familial risk studies indicate that
genetic factors are also important.
We investigated whether variation in CAG trinucleotide repeat length in exon 1 of the
androgen receptor gene is associated with TGCTs using data from a population-based family
study. Analyses of 273 TGCT case-mother pairs revealed no association between androgen
receptor CAG repeat length and overall risk of TGCTs. Risk of seminoma was significantly
associated with shorter CAG repeat length, and the trend was highly significant over decreasing
CAG repeat length (CAG ≥20 versus CAG ≤19: OR=0.54 (95% CI: 0.31-0.93), p trend=0.003).
Case-case analyses comparing seminoma cases to nonseminoma and mixed germ cell tumors
confirmed that seminoma cases had shorter CAG repeat length than other histologies (CAG ≥20
versus CAG ≤19: OR=0.54 (95%CI: 0.29-1.01), p trend=0.003).
Our findings indicate that the AR may be involved in progression from CIS to seminoma
and that genetically-determined differences in androgen action may be involved in TGCT
etiology. These results suggest further work on the role of the androgen receptor in carcinoma in
situ cells, and investigation of potential gene-environmental interactions between AR variants and
hormonally active agents that may together influence TGCT risk.

30

2.2 Introduction

Testicular germ cell tumors (TGCTs) are the most common malignancies affecting young
men between the ages of 15 and 35. Although incidence rates of these tumors are low, ranging
from 6.4 per 100,000 among non-Hispanic whites to 1.2 per 100,000 among African-Americans
(1), rates have doubled worldwide over the last 40 years (2). Survival rates exceed 95% (3), but
long-term consequences include increased risk of a second malignancy, cardiovascular disease,
and infertility (4).
There are few known risk factors for TGCTs, but both environmental and genetic factors
are believed to play a role. The early peak in age-specific incidence suggests that factors acting
early in life predispose to tumor development. It is hypothesized that during embryonic
development, fetal germ cells are transformed into a precursor lesion called carcinoma in situ
(CIS), which later progresses into the main histologic types of invasive TGCTs: seminomas,
nonseminomas and mixed germ cell tumors (GCT), which contain elements of both seminoma
and nonseminoma (5).
Well established risk factors for TGCTs are family history of TGCT, a prior diagnosis of
a TGCT, undescended testes, also called cryptorchidism, as well as additional urogenital
malformations (6). Familial risk studies have found significantly increased risk of TGCTs in
family members of cases, with an 8-10-fold risk for brothers of TGCT cases and 4-6-fold risk for
fathers (7, 8). Three recent genome-wide association studies have identified several
susceptibility alleles associated with risk of TGCTs (9-11), but it is estimated that these alleles
account for only a small portion of heritability (9). TGCT cases have a greatly increased risk of
developing a second primary tumor; men with one primary TGCT have 25 times the risk of
developing a tumor in the contralateral testicle (12). Cryptorchidism is associated with a 2-4-
fold increased risk of TGCT (13, 14).
31

Evidence of increasing incidence rates of TGCTs over time and of substantial variation in
rates geographically and by racial/ethnic group (2) point toward environmental causes, and the
involvement of hormones has been postulated (6). Due to the vital role of androgens in male
sexual differentiation and function, and the fact that sex hormone and gonadotropin levels
dramatically increase during the pubertal period when the age at risk of adult TGCT begins,
androgens have been suggested as an etiologic factor in TGCTs (15).
The effects of androgens are mediated by the androgen receptor (AR), a member of the
nuclear receptor superfamily (16). Adult germ cells apparently do not express the AR (17), but
AR expression has been demonstrated in human and rodent fetal germ cells (18, 19), CIS cells
(20) and most recently in both seminoma and nonseminoma TGCTs (21), which may suggest a
role for androgen activity in initiation or progression of TGCTs. Loss of function mutations in
the androgen receptor gene (AR) can result in varying degrees of androgen insensitivity, manifest
as a spectrum of phenotypes ranging from complete sex reversal to mild urogenital
malformations, and are associated with greatly increased risk of TGCTs (22). However,
mutations in the AR are rare, thus not likely to be related to most cases of TGCTs.
Two polymorphic trinucleotide tracts are located in the transactivation domain of the AR,
a CAG
n
(glutamine) repeat and GGC
n
(glycine) repeat. Experimental research indicates that
CAG repeat length in the normal range of 9 to 39 repeats is inversely correlated with
transactivation capability of the AR (16, 23). Since genetic variants are suspected as etiologic
factors in development of TGCTs, we postulated that lower AR activity resulting from longer
CAG repeat length may be associated with increased TGCT risk. Previous research has
addressed associations between CAG tract length and male reproductive disorders. Our recent
meta-analysis (24) indicated that increased CAG repeat length may be associated with risk of
32

male infertility, a possible risk factor for TGCT (25). Less is known about the role of the GGC
polymorphism, but limited experimental evidence suggests it also influences AR function (26).
Thus far, four case-control studies have investigated the association of TGCT risk and AR CAG
tract length, all conducted in hospital-based settings (27-30). None found significantly longer
CAG repeat length among TGCT cases, but results of one indicated that nonseminoma cases had
significantly more CAG>25 repeats compared with seminoma cases (28), and another found
significantly more TGCT cases with the combination of CAG=20/GGC=17 (29). We sought to
expand upon the existing body of knowledge by using an alternative study design composed of
population-based TGCT cases and their mothers as controls, by including cases with several
TGCT phenotypes, and by assessing associations within categories of tumor histology. For this
analysis we investigated associations between CAG repeat length and risk of TGCTs among
unilateral and bilateral cases with and without cryptorchidism using data from a population-based
family study of TGCT cases identified through the California Cancer Registry.

2.3 Materials and Methods

Study Population
Individuals analyzed are part of an on-going population-based study at the University of
Southern California that began in 2000. Using data provided by population-based regional cancer
registries in California and the California Cancer Registry (CCR), which contribute data to the
Surveillance, Epidemiology and End Results (SEER) program, we attempted to contact men
diagnosed with a testicular germ cell tumor in California from the inception of each regional
cancer registry (as early as 1972) through the end of 2006. Cancer registries provided contact
information on all living cases diagnosed with a TGCT, identified by site code: C62 and ICD-O-3
histology/behavior codes: 9060-9085 and 9100-9102.
33

This study is composed of two enrollment phases. For the first phase, 16,963 eligible
cases for whom we had a recent address, were sent a questionnaire querying information on
cases’ personal history of two testicular conditions: cryptorchidism and diagnosis with a second
primary TGCT (“bilateral TGCT”). The questionnaire asked whether any family members of the
case had been diagnosed with TGCT or cryptorchidism, and requested additional demographic
information on first-degree family members of the cases, including year of birth, year of death, if
deceased, and year of diagnosis of testicular cancer, if applicable.
A total of 5,719 cases (34%) participated in phase one and completed a questionnaire as
of January 2011. From these, 1,112 have been contacted for participation in the second phase of
the study, with over-sampling of factors indicating possible genetic pre-disposition: 1) personal
history bilateral TGCT (70 enrolled/92 contacted, 76%), 2) personal history of cryptorchidism (86
enrolled/138 contacted, 62%), 3) a family member diagnosed with a TGCT or cryptorchidism
(374 enrolled/475 contacted, 79%), 4) no personal history of bilateral TGCT or cryptorchidism,
and no family history of either condition, but two living parents (106 enrolled/159 contacted,
67%), or 5) any combination of the previous four criteria (184 enrolled/248 contacted, 74%). In
all, 820 cases (74%) have thus far provided consent for further participation.
For cases participating in the second phase, a family history interview was conducted and
additional family members were asked to participate. Male family members who had a history
of a TGCT or cryptorchidism were enrolled as additional cases. Unaffected male and female
family members were selected and enrolled as controls based on their likely contribution to
genetic family-based analyses. DNA specimens were collected on enrolled participants in the
form of blood (2,855) or saliva (281); a total of 3,136 DNA specimens, including 717 registry-
based TGCT cases, have been collected as of January 2011. For the present analysis, case-mother
34

pairs were selected in which the case was diagnosed with a TGCT and a DNA specimen had been
isolated for each member of the pair at the time of laboratory analysis, 588 pairs.
This study was approved by the Institutional Review Board of the University of Southern
California.

Clinical and Covariate Data
California cancer registries provided clinical and demographic data on cases including
date of birth, date of diagnosis, race, ethnicity and tumor histology. Enrolled cases completed a
medical history questionnaire that queried date of birth, race, ethnicity, and history of TGCT and
cryptorchidism, including laterality. For all cases who signed a medical records release form, we
attempted to collect medical records to confirm diagnoses of TGCT and cryptorchidism and
verify tumor histology.
Ethnicity (Hispanic, non-Hispanic) was designated for all cases using the following data
sources, in order: 1) self-report from the medical history questionnaire, if available, 2) cancer
registry data, if available, and 3) the 1980 United States Census list of Spanish surnames, which
was used to assign ethnicity by surname.

Analysis of the number of CAG Repeats in the AR Gene
Genomic DNA was extracted from peripheral blood cells (553 individuals) or buccal
epithelial cells (35 individuals) using standard protocols. Genotyping of the AR CAG
microsatellite polymorphism was performed using PCR in combination with fluorescently labeled
oligonucleotide primers, as described previously (31).

35

Statistical Analysis
Androgen receptor CAG repeat length was compared between 273 TGCT cases and their
mothers (546 individuals) assigning length of the CAG allele transmitted from the mother to the
case as the transmitted allele (“case”) and the allele not transmitted from the mother as the non-
transmitted allele (“mother”). A paired t-test was used to compare mean CAG repeat length
between cases and mothers. Conditional logistic regression was used to test for associations
between CAG repeat length and risk of TGCT. Associations were measured by odds ratios (ORs)
and corresponding 95% confidence intervals (CIs). CAG repeat lengths were categorized into
four groups ( ≤19, 20-21, 22-23 and ≥24) and into two groups ( ≤19 and ≥20) based upon
functional knowledge of the linear association of increasing CAG repeat length with reduced AR
transactivation (23) and the distribution of data among mothers. Wald tests for trend were
conducted with CAG repeat length coded as an ordinal variable. Case-mother pairs were
excluded from the analysis if genotype data was missing on either individual (6 pairs), or the pair
exhibited Mendelian incompatibility (15 pairs).
For all analyses, histologic groups were defined using ICD-O-3 histology/behavior codes:
seminoma: 9060-9064, nonseminoma: 9065-9084, 9100-9102, and mixed germ cell tumors
(GCT): 9085. Mixed GCT, representing tumors containing elements of seminoma and
nonseminoma have been designated as a separate group since 1990 but were previously
categorized as nonseminoma.
To address possible differences in genetic risk by histology, analyses were conducted
using the following sub-groups of pairs based on case histology: all TGCT cases (273 case-
mother pairs), seminoma cases (140 case-mother pairs), nonseminoma and mixed GCT cases
(121 case-mother pairs), mixed GCT cases only (53 case-mother pairs), and pure nonseminoma
cases only (68 case-mother pairs). Unconditional logistic regression was used for case-case
36

analyses comparing seminoma cases to the following other histologic groups: non-seminoma and
mixed GCT cases, mixed GCT cases only, and pure nonseminoma cases only. Analyses were
adjusted for age at diagnosis ( ≤25, 26-30, 31-35, 36-40, >40); 5-year age groups were used when
numbers were sufficient, but for ages 25 and younger and older than age 40, data were collapsed
due to small numbers. All analyses were conducted among all races/ethnicities, with and without
adjustment for race/ethnicity, and among non-Hispanic whites only.
Analyses were conducted using SAS statistical software, version 9.1 (SAS Institute,
Cary, NC).

Age-Specific Incidence Curve
Using SEER Stat software provided by the Surveillance, Epidemiology and End Results
(SEER) program, age-adjusted incidence rates of TGCTs were output for SEER 13 registries in
five-year age groups, by histology, over the period 1992-2007. Incidence rates for seminoma,
nonseminoma and mixed GCT histologies were produced using the above described scheme of
ICD-O-3 histology/behavior codes. The age-specific incidence curve was plotted using Microsoft
Excel software.
2.4 Results

Characteristics of TGCT cases analyzed are shown in Table 2.1A, over all and within
strata defined by histology. Characteristics of all eligible cases from the California Cancer
Registry (CCR), cases who completed the questionnaire in phase one, cases who provided
consent for participation in phase two, and cases that provided a DNA sample are provided in
Table 1B for comparison with cases analyzed. Seminoma cases were older at diagnosis than
mixed GCT and pure nonseminoma cases among all sets of cases (Tables 2.1A and 2.1B). The
distribution of diagnosis date is similar between cases analyzed and all other cases. The majority
37

of cases from all sets are white and non-Hispanic, though there are slightly more white and Non-
Hispanic men among cases analyzed compared with all CCR cases (white: 95.2% and 92.4%,
respectively; Non-Hispanic 94.1% and 80.0%, respectively). The distribution of histologic types
is similar between cases analyzed and all other cases. A majority of cases analyzed (64%) had a
family history of TGCT and/or cryptorchidism because cases with a family history were over-
sampled by design, while 13.7% of cases who completed the questionnaire reported a family
history of TGCT and/or cryptorchidism. The percentage of cases analyzed reporting a history of
cryptorchidism or bilateral TGCT was slightly larger than the percentage among cases that
completed the questionnaire (cryptorchidism: 17.0% versus 10.8%, respectively; bilateral TGCT:
10.6% versus 3.8%, respectively).
Analyses of differences in mean CAG repeat length between alleles transmitted to cases
and mothers non-transmitted alleles revealed mothers have significantly longer mean CAG repeat
length than seminoma cases (p = 0.01) and pure nonseminoma cases have marginally
significantly longer mean CAG repeat length than their mothers (p = 0.06) (Table 2.2A). Case-
case comparisons revealed nonseminoma and mixed GCT cases had significantly longer mean
CAG length (mean± standard deviation) 21.69±2.99 than seminoma cases 20.71±2.92 (p = 0.006)
(comparison not shown in Table 2.2A).
Logistic regression analyses comparing cases to mothers (Table 2.2B) indicate no
association between CAG repeat length and overall TGCT risk, among all races and ethnicities
and non-Hispanic whites only. In sub-groups defined by tumor histology, shorter CAG repeat
length (CAG ≤19) was significantly associated with risk of seminoma (CAG ≥20 versus CAG
≤19: OR=0.54 (95% CI: 0.31-0.93, p trend =0.003)). This association was somewhat stronger in
analyses restricted to non-Hispanic whites (CAG ≥20 versus CAG ≤19: OR=0.43 (95% CI: 0.23-
0.79), p trend < 0.001)) (Table 2.2B). Although estimated associations between CAG repeat
38

length and risk of nonseminoma and mixed GCT (together or separately) were not significant, the
effect was in the opposite direction consistent with increased risk of nonseminoma and mixed
GCT with longer CAG repeat length. This association was strongest among men with pure
nonseminoma (CAG ≥20 versus CAG ≤19: OR=1.80 (95% CI: 0.83-3.90), p trend = 0.17))
(Table 2.2B). Among men with mixed GCT tumors, point estimates of association were between
those estimates for seminoma and nonseminoma (CAG ≥20 versus CAG ≤19: OR=1.10 (95%
CI: 0.47-2.59)). Results were similar among all races/ethnicities and among non-Hispanic whites.
We conducted a number of sensitivity analyses with exclusions on particular criteria.
From the analysis of seminoma, we excluded the small numbers of pairs in which cases had
anaplastic seminoma or spermatocytic seminoma, and results were not materially altered (data not
shown). Mixed germ cell tumors have only been designated as a separate histologic group since
1990. To determine whether histologic designations affected the results of analyses of mixed
GCTs, we conducted analyses limited to date of diagnosis after 1990, and results were unchanged
(data not shown). Because personal history of cryptorchidism and family history of TGCT or
cryptorchidism could modify genetic associations with risk of TGCT, we repeated analyses
limited to: pairs in which cases were without cryptorchidism and without a family history of
TGCT, and associations were consistent with overall results (data not shown).
Based on the findings of a significant difference in mean CAG repeat length between
seminoma and nonseminoma and mixed GCT cases, we conducted a case-case analysis
comparing seminoma cases to the other histologic groups (Table 2.3). Comparing seminoma
cases to nonseminoma and mixed GCT cases, analyses confirmed that, after adjusting for age at
diagnosis, seminoma cases have shorter CAG repeat length than nonseminoma and mixed GCT
cases (CAG ≥20 versus CAG ≤19: OR=0.54 (95%CI: 0.29-1.01), and the trend over increasing
CAG repeat length is highly significant (p trend=0.003). Among non-Hispanic white cases, the
39

association and test of trend remained statistically significant (CAG ≥20 versus CAG ≤19:
OR=0.49 (95% CI: 0.25-0.94), p trend = 0.002). When seminoma cases were compared with
mixed GCT and nonseminoma cases separately, the associations were similar (Table 2.3).
A graph of age-specific incidence rates for TGCTs by histology based on SEER 13
registry data is shown in Figure 2.1. The graph displays age-adjusted incidence rates of
seminoma, nonseminoma and mixed GCTs over the period 1992-2007. The age-incidence curves
of the three groups are similar, but seminoma tumors are diagnosed about 10 years later than
nonseminoma and mixed GCT. Trends for mixed germ cell tumors and nonseminoma are nearly
identical using this data, which covers diagnosis dates post-1990.

Table 2.1 Characteristics of all testicular germ cell tumor cases

(A.) Cases analyzed
*
: all cases, seminoma cases, mixed germ cell tumor
cases, and pure nonseminoma cases (B.) All eligible cases from the California Cancer Registry, cases who completed the questionnaire in
phase one, cases consented to participate in phase two, and cases that provided a DNA sample
_______________________________________________________________________________________________________________
All Cases
†
Seminoma Mixed germ cell tumors Pure Nonseminoma
(N= 273) (N = 140) (N = 53) (N = 68)
N (%) N (%) N (%) N (%)
_______________________________________________________________________________________________________________
A.
Age at diagnosis
‡

≤25 44 (8.8) 9 (6.5) 9 (17.0) 24 (35.3)
26-30 44 (16.1) 16 (11.4) 12 (22.6) 16 (23.5)
31-35 63 (23.1) 35 (25.0) 13 (24.5) 13 (19.1)
36-40 53 (19.4) 34 (24.3) 12 (22.6) 7 (10.3)
>40 45 (16.5) 38 (27.1) 3 (5.7) 4 (5.9)
Unknown 24 (16.5) 8 (5.7) 4 (7.6) 4 (5.9)

Diagnosis date
1974-1984 19 (7.0) 11 (7.9) -- 8 (11.8)
1985-1989 35 (12.8) 17 (12.1) 3 (5.7) 13 (19.1)
1990-1994 54 (19.8) 24 (17.1) 11 (20.8) 18 (26.4)
1995-2000 79 (28.9) 49 (35.0) 18 (34.0) 12 (17.7)
2000-2006 63 (23.1) 32 (22.9) 17 (32.1) 14 (20.6)
Unknown 23 (8.4) 7 (5.0) 4 (7.6) 3 (4.4)

Race
White 260 (95.2) 133 (95.1) 51 (96.2) 65 (95.6)
African-American 4 (1.5) 1 (0.7) 1 (1.9) 2 (2.9)
Asian 5 (1.8) 3 (2.1) 1 (1.9) 1 (1.5)
Unknown 4 (1.5) 3 (2.1)

Ethnicity
Non-Hispanic 257 (94.1) 132 (94.3) 50 (94.3) 63 (92.6)
40

Table 2.1 (Continued)
Hispanic 16 (5.9) 8 (5.7) 3 (5.7) 5 (7.4)

TGCT laterality
Unilateral 241 (88.3) 125 (89.3) 41 (77.4) 64 (94.1)
Bilateral 29 (10.6) 14 (10.0) 12 (22.6) 3 (4.4)
Laterality unknown 3 (1.1) 1 (0.7) 1 (1.5)

Personal history of cryptorchidism
None 219 (80.2) 112 (80.0) 44 (83.0) 53 (77.9)
Unilateral 40 (14.7) 19 (13.6) 6 (11.3) 13 (19.1)
Bilateral 6 (2.2) 4 (2.9) 2 (3.8) 0
Laterality unknown 8 (2.9) 5 (3.6) 1 (1.9) 2 (3.0)

Family history
None 98 (35.9) 40 (28.6) 26 (49.0) 31 (45.6)
TGCT only 76 (27.8) 45 (32.1) 11 (20.8) 12 (17.7)
CO only
†
80 (29.3) 47 (33.6) 14 (26.4) 19 (27.9)
TGCT and CO
†
19 (7.0) 8 (5.7) 2 (3.8) 6 (8.8)

Histology
§

Seminoma 140 (51.3) 140 (100) -- --
Pure seminoma 134 (95.7) 134 (95.7) -- --
Anaplastic seminoma 4 (2.9) 4 (2.9) -- --
Spermatocytic seminoma 2 (1.4) 2 (1.4) -- --
Mixed germ cell tumor 53 (19.4) -- 53 (100) --
Pure Nonseminoma 68 (24.9) -- -- 68 (100)
Embryonal carcinoma 36 (13.2) -- -- 36 (52.9)
Teratocarcinoma 18 (26.5) -- -- 18 (26.5)
Teratoma 6 (8.8) -- -- 6 (8.8)
Yolk sac 3 (1.1) -- -- 3 (4.4)
Choriocarcinoma 5 (1.8) -- -- 5 (7.4)
41

Table 2.1 (Continued)
______________________________________________________________________________________________________________
All Registry Cases
†
Questionnaire Consented Cases DNA Specimen Cases
Completed Cases
(N= 16,963) (N = 5,719) (N = 820) (N = 717)
N (%) N (%) N (%) N (%)
_______________________________________________________________________________________________________________
B.
Age at diagnosis
‡

≤25 2913 (17.2) 761 (13.3) 120 (14.6) 100 (14.0)
26-30 2980 (17.5) 883 (15.4) 137 (16.7) 114 (15.9)
31-35 3308 (19.5) 1142 (20.0) 170 (20.7) 148 (20.6)
36-40 2755 (16.2) 986 (17.3) 150 (18.3) 138(19.3)
>40 4028 (23.8) 1691 (29.6) 210 (25.6) 190 (26.5)
Unknown 979 (5.8) 256 (4.5) 33 (4.0) 27 (3.8)

Diagnosis date
1974-1984 1623 (9.6) 611 (10.7) 66 (8.1) 58 (8.1)
1985-1989 2087 (12.3) 698 (12.2) 116 (14.2) 105 (14.6)
1990-1994 3371 (19.9) 1173 (20.6) 160 (19.5) 142 (19.8)
1995-2000 3845 (22.7) 1441 (25.2) 189 (23.1) 173 (24.1)
2000-2006 5364 (31.6) 1574 (27.5) 255 (31.1) 211 (29.4)
Unknown 673 (4.0) 222 (3.9) 34 (4.2) 28 (3.9)

Race
White 15669 (92.4) 5350 (93.5) 767 (93.5) 672 (93.7)
African-American 288 (1.7) 64 (1.1) 6 (0.7) 6 (0.8)
Asian/ Pacific Islander/Other 618 (3.6) 171 (3.0) 22 (2.7) 19 (2.7)
Unknown 388 (2.3) 134 (2.3) 25 (3.0) 20 (2.8)

Ethnicity
Non-Hispanic 13557 (80.0) 5035 (88.0) 719 (87.7) 630 (87.9)
Hispanic 3407 (20.0) 684 (12.0) 101 (12.3) 87 (12.1)
42

Table 2.1 (Continued)
Self-reported as bilateral TGCT case
Yes NA 216 (3.8) 108 (13.2) 96 (13.4)
No NA 5436 (95.1) 706 (86.1) 618 (86.2)
Missing 67 (1.1) 6 (0.7) 3 (0.4)

Self-reported history of cryptorchidism
Yes NA 616 (10.8) 169 (20.6) 152 (21.2)
No NA 4918 (86.0) 625 (76.2) 544 (75.9)
Missing 185 (3.2) 26 (3.2) 21 (2.9)

Family history of TGCT or cryptorchidism
Yes NA 784 (13.7) 476 (58.1) 428 (59.7)
No NA 4818 (84.3) 330 (40.2) 279 (38.9)
Missing 117 (2.0) 14 (1.7) 10 (1.4)

Histology
§

Seminoma 9371 (55.1) 3259 (56.9) 426 (52.0) 376 (52.4)
Pure seminoma 8848 3075 411 362
Anaplastic seminoma 439 152 11 11
Spermatocytic seminoma 84 32 4 3
Mixed germ cell tumor 2518 (14.8) 805 (14.1) 130 (15.9) 105 (14.6)
Pure Nonseminoma 4162 (24.5) 1341 (23.4) 211 (25.7) 189 (26.4)
Nonseminomatous 73 18 5 4
Embryonal carcinoma 1948 662 101 88
Teratocarcinoma 1120 369 56 52
Teratoma 317 100 16 13
Yolk sac 244 68 8 8
Choriocarcinoma 460 124 25 24
Unknown 912 (5.4) 314 (5.5) 53 (6.5) 45 (6.6)
43

Table 2.1 (Continued)
_______________________________________________________________________________________________________________
*
Cell counts may not add up across all histology categories because of missing data
†
TGCT = testicular germ cell tumor, CO = cryptorchidism
‡
Age at first diagnosis is used for cases with two primary testicular germ cell tumors (bilateral cases)
§
Histology details on 29 bilateral cases (first TGCT/second TGCT=number of cases): Seminoma: seminoma/seminoma=8, seminoma/spermatocytic
seminoma=1, seminoma/ unknown=5; Mixed germ cell tumor: mixed GCT/mixed GCT =1, seminoma/mixed GCT =5, seminoma/teratoma=4, mixed
GCT/embryonal=1, mixed GCT /unknown=1; Pure nonseminoma: teratoma/teratoma=1, teratoma/unknown=2
44

Table 2.2 Androgen receptor CAG repeat lengths comparing the allele transmitted from mothers to testicular germ cell tumor (TGCT)
cases (“cases”) to the allele not transmitted (“mothers”) (A.) Distribution of CAG repeat lengths among cases and mothers (B.) Odds
ratios and 95% confidence intervals for the association of CAG repeat length and risk of TGCT among all cases, seminoma cases, mixed
germ cell tumor cases and pure nonseminoma cases, comparing the allele transmitted from mothers to TGCT cases to the allele not
transmitted
_______________________________________________________________________________________________________________
Range Mean ± S.D. Median p value
*

_______________________________________________________________________________________________________________
A.
Mothers of all TGCT cases (N=273) 9-33 21.33±3.30 21.0
All TGCT cases (N=273) 12-30 21.12±3.01 21.0 0.45

Mothers of seminoma cases (N=140) 12-30 21.61±3.19 22.0
Seminoma cases (N=140) 12-29 20.71±2.92 20.0 0.01

Mothers of nonseminoma and mixed cases (N=121) 9-33 21.01±3.42 21.0
Nonseminoma and mixed cases (N=121) 14-30 21.69±2.99 21.0 0.10

Mothers of mixed GCT cases (N=53) 13-33 21.66±3.66 22.0
Mixed GCT cases (N=53) 17-30 21.92±3.00 22.0 0.69

Mothers of pure nonseminoma cases (N=68) 9-29 20.51±3.15 20.0
Pure nonseminoma cases (N=68) 14-29 21.51±3.00 21.0 0.06
45

Table 2.2 (Continued)
_______________________________________________________________________________________________________________
Number of Cases by Type/Number of Mothers
All cases/ Seminoma cases
*
/ Nonseminoma and Mixed GCT cases
*
/ Pure Nonseminoma cases
*
/
Mothers Mother Mixed GCT cases
*
/ Mothers Mothers
Mothers
_______________________________________________________________________________________________________________
B.
I. All races/ethnicities:

No. CAG repeats (N=546) (N=280) (N=242) (N=106) (N=136)
≤19 83/73 48/31 29/38 12/13 17/25
≥20 190/200 92/109 92/83 41/40 51/43

20-21 80/73 45/36 33/32 14/13 19/19
22-23 56/61 24/37 30/23 13/12 17/11
≥24 54/66 23/36 29/28 14/15 15/13

OR (95% CI)
†
≤19 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
≥20 0.84 (0.58-1.21) 0.54 (0.31-0.93) 1.45 (0.82-2.56) 1.10 (0.47-2.59) 1.80 (0.83-3.90)

20-21 0.96 (0.61-1.49) 0.82 (0.43-1.58) 1.34 (0.67-2.67) 1.13 (0.40-3.20) 1.53 (0.61-3.87)
22-23 0.81 (0.50-1.29) 0.39 (0.19-0.83) 1.62 (0.81-3.21) 1.13 (0.40-3.20) 2.11 (0.83-5.34)
≥24 0.72 (0.44-1.17) 0.42 (0.20-0.86) 1.39 (0.67-2.86) 1.01 (0.35-2.94) 1.79 (0.66-4.85)
p trend = 0.13 p trend = 0.003 p trend = 0.30 p trend = 0.98 p trend = 0.17
46

Table 2.2 (Continued)
II. Non-Hispanic whites:

No. CAG repeats (N=488) (N=250) (N=216) (N=96) (N=120)
≤19 75/63 44/24 25/35 11/13 14/22
≥20 169/181 81/101 83/73 37/35 46/38

20-21 73/66 40/33 31/29 14/11 17/18
22-23 49/54 22/33 26/20 11/11 15/9
≥24 47/61 19/35 26/24 12/13 14/11

OR (95% CI)
†
≤19 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref) 1.0 (ref)
≥20 0.79 (0.53-1.17) 0.43 (0.23-0.79) 1.56 (0.86-2.81) 1.22 (0.51-2.95) 1.89 (0.84-4.24)

20-21 0.92 (0.57-1.49) 0.65 (0.32-1.34) 1.46 (0.71-3.02) 1.47 (0.49-4.44) 1.49 (0.56-3.95)
22-23 0.77 (0.47-1.26) 0.36 (0.17-0.78) 1.68 (0.82-3.46) 1.10 (0.38-3.26) 2.30 (0.85-6.22)
≥24 0.65 (0.39-1.09) 0.29 (0.13-0.65) 1.51 (0.72-3.19) 1.11 (0.37-3.35) 1.97 (0.70-5.53)
p trend = 0.10 p trend < 0.001 p trend = 0.24 p trend = 0.98 p trend = 0.12
_______________________________________________________________________________________________________________
*
P value from paired t-test comparing mean length of transmitted and non-transmitted alleles between case and mother
*
Histologic subgroups include all TGCT cases (unilateral and bilateral) without cryptorchidism
†
Logistic regression (comparing transmitted allele to non-transmitted allele)
47

Table 2.3 Case-case analyses displaying odds ratios and 95% confidence intervals for the association of androgen receptor CAG repeat
length with testicular germ cell tumor histology comparing seminoma cases with: nonseminoma and mixed germ cell tumor cases, mixed
germ cell tumor cases only and pure nonseminoma cases only
_______________________________________________________________________________________________________________

Seminoma cases/ OR (95% CI)
†
Seminoma Cases/ OR (95% CI)
†
Seminoma Cases/ OR (95% CI)
†

Nonseminoma and Mixed Mixed GCT Cases
*
Pure Nonseminoma
GCT cases
*
Cases
*

_______________________________________________________________________________________________________________
I. All races/ethnicities:

No. CAG repeats (N=242) (N=178) (N=193)
≤19 44/27 1.0 (ref) 44/10 1.0 (ref) 44/17 1.0 (ref)
≥20 85/86 0.54 (0.29-1.01) 85/39 0.46 (0.20-1.06) 85/47 0.58 (0.27-1.24)

20-21 43/29 0.95 (0.44-2.01) 43/12 0.79 (0.30-2.11) 43/17 1.11 (0.44-2.81)
22-23 19/28 0.35 (0.15-0.83) 19/13 0.32(0.11-0.93) 19/15 0.28 (0.10-0.79)
≥24 23/29 0.36 (0.16-0.82) 23/14 0.32 (0.11-0.87) 23/15 0.41 (0.15-1.13)
p trend=0.003 p trend=0.008 p trend=0.02
II. Non-Hispanic whites

No. CAG repeats (N=214) (N=158) (N=171)
≤19 40/23 1.0 (ref) 40/9 1.0 (ref) 40/14 1.0 (ref)
≥20 75/76 0.49 (0.25-0.94) 75/34 0.45 (0.19-1.07) 75/42 0.50 (0.23-1.11)

20-21 38/27 0.81 (0.37-1.78) 38/12 0.68 (0.25-1.88) 38/15 0.97 (0.37-2.54)
22-23 18/23 0.35 (0.14-0.86) 18/10 0.36 (0.11-1.12) 18/13 0.25 (0.08-0.76)
≥24 19/26 0.30 (0.13-0.72) 19/12 0.29 (0.10-0.87) 19/14 0.32 (0.11-0.91)
p trend=0.002 p trend=0.01 p trend=0.008
_______________________________________________________________________________________________________________
*
Includes unilateral and bilateral TGCT cases with histology, race, ethnicity and age at diagnosis available
†
Adjusted for age at diagnosis ( ≤25, 26-30, 31-35, 36-40, >40)
48
49

Figure 2.1. Age-specific incidence rates of testicular germ cell tumors by histologic type, 1992-
2007, SEER 13

50

2.5 Discussion

Due to the unique age-incidence patterns of testicular germ cell tumors, including a small
peak in the perinatal period, and associations with congenital conditions (6) and birth weight (32),
it has been widely postulated that these tumors have fetal origins, which may be influenced by
genetic and/or environmental factors. A model proposed in 1983 by Henderson and others
suggests two critical phases in the development of TGCTs (33, 34). The model posits that in the
first phase, during embryogenesis, fetal germ cells undergo arrested development due to the
influence of hormonal factors, and fail to enter the normal pathway to become pre-spermatgonia.
Instead, these altered gonocytes are hypothesized to proliferate and develop into carcinoma in situ
(CIS), the presumed precursor to TGCTs (5). The second phase proposed is later in life, after the
pubertal surge in hormones and gonadotropins; CIS cells that have remained dormant throughout
childhood are postulated to progress to develop into invasive TGCTs, manifest as tumors
diagnosed in early adulthood. It is thus hypothesized that hormones may be involved in both
impaired development of gonocytes during embryogenesis and progression of CIS cells to
invasive tumors during young adulthood.
Androgens are fundamentally involved in male sexual differentiation and
spermatogenesis. The androgen receptor (AR) reportedly is not expressed in adult male germ
cells; however, AR expression has been demonstrated in rodent (19, 35) and human (18) fetal
germ cells. It has been shown that gonocyte proliferation is enhanced among testicular feminized
mice, suggesting AR deficiency leads to an increase in gonocyte numbers in fetal life (19). Thus,
lower AR activity has a plausible role in increasing fetal germ cell proliferation. We postulated
that variants in the androgen receptor gene (AR) that influence AR activity may be involved in
initiation of TGCTs through altered gonocyte development. Functional work on the CAG repeat
length polymorphism in the AR has identified a linear association between longer CAG repeat
51

length and reduced AR transcription activity, which likewise may reduce activity of the AR. This
has led us and other investigators to examine whether longer AR CAG repeat length is associated
with increased risk of TGCT.
The results from this population-based case-mother pair study indicate no overall
association between CAG repeat length and risk of TGCTs. However, analyses of the main
histologic types were more revealing. Though non-significant, we found the predicted
association of increased risk of non-seminoma and mixed GCT with longer CAG repeat length.
Unexpectedly, the association observed among the seminoma group was in the opposite direction,
as shorter CAG repeat length was significantly associated with seminoma TGCT. Associations
for mixed GCTs were in between those of seminoma and nonseminoma, which could indicate of
a mix of effects. Direct comparisons of seminoma and nonseminoma/mixed GCT cases in case-
case analyses revealed that seminoma cases had significantly shorter CAG repeat length than both
nonseminoma and mixed germ cell tumor cases, which was stronger after adjusting for age at
diagnosis. The results were similar when restricted to non-Hispanic whites.
Although we did not find CAG repeat length to be predictive of over all risk for TGCTs,
the difference in repeat length by histologic type may be of significance biologically. Seminomas
and nonseminomas have been designated into two broad groups due to differences in their
histology, age at presentation and clinical features. The median age at diagnosis of seminoma
tumors is 35 years, while nonseminomas and mixed GCT present a decade earlier (Figure 1).
Seminomas are more slow-growing, respond better to treatment, and have a better prognosis than
other histologies (4).
Both seminomas and nonseminomas are believed to develop from pre-cursor CIS cells.
Seminomas are more similar histologically to CIS cells than are nonseminomas, and it is
postulated that a default pathway of testicular carcinogenesis is for seminoma to arise from CIS,
52

with activation of pluripotency additionally required for development of nonseminoma (5). It
remains unclear what factors drive the transformation of CIS cells to invasive tumors, but
interestingly, AR expression has been demonstrated in CIS cells (20). Our finding of an
association of shorter CAG repeat length with seminoma risk could indicate increased AR
activity, as a result of shorter CAG repeat length, may be important during the time when CIS
lesions progress to invasive germ cell tumors. It is possible that among men with dormant CIS
and shorter CAG repeat lengths, increased AR activity may lead CIS cells to proliferate into
seminomatous tumors.
Furthermore, an effect of shorter CAG repeat length in increasing risk of seminoma
suggests a pathway through which environmental agents, such as exogenous hormones or
hormone agonists, or so-called “endocrine-disruptors” could act to influence development of
TGCTs (6). In this scenario, men with shorter CAG length may be genetically pre-disposed to a
higher risk of progression to seminoma due to exposure to exogenous hormonal agents, which
may have an impact on hormone levels and the action of androgens mediated by the AR.
Two alternative explanations for the association we observed between shorter CAG
repeat length and seminoma are chance, and confounding by an unmeasured genetic variant in
linkage disequilibrium with the CAG repeat polymorphism. We believe our finding is unlikely to
be a result of chance for several reasons: 1) there was a clear a monotonic trend of increasing risk
of seminoma over decreasing CAG repeat length, 2) the association with risk of seminoma
persisted when we made exclusions based on several different factors, including specific
seminoma histologies, history of cryptorchidism and family history of TGCT or cryptorchidism,
and 3) the mothers used as a reference constitute a genetically ideal control group. Findings
from familial risk studies indicate greater risk of developing a TGCT among brothers of TGCT
cases than among fathers, which accords with a maternal inheritance pattern. We were did not
53

analyze the AR GGC polymorphism in this study, but we must also consider the possibility that
that alleles of that polymorphism or other X-chromosome variants in linkage disequilibrium with
CAG
n
genotypes could influence TGCT risk.
Recently, studies have identified variants at other loci associated with TGCT risk.
Candidate gene studies reported the Y chromosome gr/gr deletion (36) and variants in the
PDE11A gene (37) to be associated with TGCT risk. Genome-wide association studies have
indentified variant alleles with strong associations with TGCT risk, implicating the KITLG,
SPRY4 and BAK1 genes (9-11). In these investigations, however, no notable differences were
reported for any of the identified risk variants by histologic type. The present study is the first to
identify a genetic variant associated with risk of TGCT that significantly differs by histologic
type. This is of importance because of the demonstrated differences in age at onset, tumor
presentation and behavior by these histologic types, and furthermore, because as of yet no
mechanisms have been identified to explain what distinguishes the progression of CIS to
seminoma versus nonseminoma.
Our overall result is in agreement with four case-control studies that did not find CAG
repeat length to be predictive of overall TGCT risk (27-30). There are limited published data on
AR CAG repeat length among tumors of specific histologic types. Consistent with our result, a
Swedish study found a significantly greater percentage of CAG>25 in nonseminoma cases than
seminoma cases (28), but two other studies (27, 30), reported no difference in CAG length by
histologic type. Also consistent with our finding, data from two of the three studies (27, 28)
observed seminoma cases with shorter mean CAG repeat length than nonseminoma and mixed
GCT cases.
Any differences between our findings and those from previous studies may be due in part
to unique features of our investigation – the use of population-based cases and mothers as
54

controls. The advantage of using mothers as a reference is that, genetically, the cases are derived
directly from their mothers, which makes them an ideal reference group for a genetic study of an
X chromosome variant. One potential disadvantage of using mother controls in genetic studies is
the concern that parental genotypes associated with risk of disease may interfere with the ability
to reproduce. Indeed, previous research suggests longer CAG repeat length may be associated
with reduced fertility among men (24). However, there is no evidence that variation in CAG tract
length influences female fertility, thus mothers were an appropriate control group in this
investigation.
The strengths of our analysis include recruitment of population-based TGCT cases with
various phenotypes, and the collection of detailed information on cases including family history
and personal history of cryptorchidism so that those factors could be taken into account. In
addition, cases included in the analysis were similar to cases from the base-population on features
that we did not over-sample by design, and among features upon which we over-sampled,
sensitivity analyses revealed consistency of results. A limitation is that we were able to enroll
only surviving TGCT cases, which may have excluded cases with most severe disease. However,
survival rates are exceedingly high for TGCTs, so this seems a minor concern.
In summary, we found that shorter CAG repeat length in the androgen receptor gene is associated
with risk of seminomatous TGCT. Our results suggest that the AR may be involved in
progression from CIS to seminoma and that genetically-determined differences in androgen
action may be involved in TGCT etiology. Further work is warranted to confirm and interrogate
these findings in additional epidemiologic studies; in particular to investigate effects of among
mixed GCTs, which may represent a combination of effects of seminoma and nonseminoma
elements. In addition, we recommend further mechanistic studies to understand the role of the AR
in the development of fetal germ cells and in tissue derived from CIS cells and invasive TGCTs,
55

and genetic research to investigate potential gene-environmental interactions between AR variants
and hormonally active agents that may together influence TGCT risk.
56

Chapter 2 References

Akimoto C, Ueda T, Inoue K, et al. Testis-specific protein on Y chromosome (TSPY) represses
the activity of the androgen receptor in androgen-dependent testicular germ-cell tumors. PNAS
2010;107:19891-6. (21)

Altekruse SF, Kosary CL, Krapcho M, et al. (eds). SEER Cancer Statistics Review, 1975-2007,
National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2007/, based on
November 2009 SEER data submission, posted to the SEER web site, 2010. (1)

Brinkmann AO. Molecular basis of androgen insensitivity. Mol Cell Endocrin 2001;179: 105-9.
(22)

Chamberlain NL, Driver ED, Meisfeld RL. The length and location of CAG trinucleotide repeats
in the androgen receptor N-terminal domain effect transactivation function. Nucleic Acids Res
1994;22:3181-6. (16)

Cortessis VK. Epidemiologic insights into the occurrence and causes of testicular cancer. In:
Raghavan D, ed. American Cancer Society Atlas of Clinical Oncology, Germ Cell tumors.
London: BC Decker 2003,16-29. (34)

Davis-Dao CA, Tuazon ED, Sokol RZ, et al. Male infertility and variation in CAG repeat length
in the androgen receptor gene: a meta-analysis. J Clin Endocrin Metab 2007;92: 4319-26. (24)

Diekmann K-P, Pichlmeier U. Clinical epidemiology of testicular germ cell tumors. World J Urol
2004;22:2-14. (6)

Forman D, Oliver RT, Brett AR, et al. Familial testicular cancer: a report of the UK family
register, estimation of risk and an HLA class I sib-pair analysis. Br J Cancer 1992;65: 255-62. (7)

Giwercman A, Lundin KB, Eberhard J et al. Linkage between androgen receptor gene CAG
trinucleotide repeat length and testicular germ call cancer histological type and clinical stage. Eur
J Cancer 2004;40:2152-8. (28)

Gao T, Marcelli M, McPhaul MJ. Transcriptional activation and transient expression of the
human androgen receptor. J Steroid Biochem Mol Biol 1996;59: 9-20. (26)

Garolla A, Ferlin A, Vinanzi C, et al. Molecular analysis of the androgen receptor gene in
testicular cancer. Endocr-Related Cancer 2005;12:645-55. (29)

Garner MJ Turner MC, Ghadirian P, et al. Epidemiology of testicular cancer: an overview. Int J
Cancer 2005;116:331-39. (13)

Heimdal K, Olsson H, Tretli S, et al. Risk of testicular cancer in relatives of testicular cancer
patients. Br J Cancer 1996;73:970-73. (8)

Henderson BE, Ross RK, Pike MC, et al. Epidemiology of testis cancer. In: Skinner DG, ed.
Urological Cancer. New York: Grune and Stratton 1983,237-50. (33)
57

Horvath A, Korde L, Greene MH, et al. Functional phosphodiesterase 11A mutations may
modify the risk of familial and bilateral testicular germ cell tumors. Cancer Res 2009;69:5301-6.
(37)

Horwich A, Shipley J, Huddart R. Testicular germ cell cancer. Lancet 2006;367:754-65. (4)

Hotaling JM, Walsh TJ Male infertility: a risk factor for testicular cancer. Nat Rev Urol
2009;6:550-6. (25)

Huyghe E, Matsuda T, Thonneau P. Increasing incidence of testicular cancer worldwide: a
review J Urol 2003;170:5-11. (2)

Irvine RA, Ma H, Yu MC, et al. Inhibition of p160-mediated coactivation with increasing
androgen receptor polyglutamine length. Hum Mol Genet 2000; 9: 267-74. (23)

Kanetsky PA, Mitra N, Vardhanabhuti S, et al. Common variation in KITLG and at 5q31.3
predisposes to testicular germ cell cancer. Nat Genet 2009;41:811-15. (11)

Merlet J, Racine C, Moreau E, et al. Male fetal germ cells are targets for androgens that
physiologically inhibit their proliferation. PNAS 2007;104:3615-20. (19)

Nathanson KL, Kanetsky PA, Hawes R, et al. The Y Deletion gr/gr and susceptibility to testicular
germ cell tumor. Am J Hum Genet 2005;77:1034-43. (36)

Oosterhuis JW, Looijenga LHJ. Testicular germ-cell tumors in a broader perspective. Nat Rev
Cancer 2005;5:210-22. (5)

Rajpert-De Meyts E, Skakkabaek NE. Immunohistochemical identification of androgen receptors
in germ cell neoplasia. J Endrocrinol 1992;135:R1-4. (20)

Rajpert-De Meyts E, Leffers H, et al. Analysis of the polymorphic CAG repeat length in the
androgen receptor gene in patients with testicular germ cell cancer. Int J Cancer 2002;102: 201-
4. (27)

Rajpert-De Meyts E. Development model for the pathogenesis of testicular carcinoma in situ:
genetic and environmental aspects. Hum Reprod Update 2006;12:303-23. (15)

Rapley EA, Turnbull C, Amin Al Olama A, et al. A genome-wide association study of testicular
germ cell tumor. Nat Genet 2009;41:807-10. (10)

Richiardi L, Pettersson A, Akre O. Genetic and environmental risk factors for testicular cancer.
Int J Androl 2007;30:230-40. (32)

Schnack TH, Poulsen G, Myrup C, et al. Familial coaggregation of cryptorchidism, hypospadias,
and testicular germ cell cancer: a nationwide cohort study. J Nat Cancer Instit 2009; 102:187-92.
(14)

58

Shapiro E, Huang H, Masch RJ, et al. Immunolocalization of androgen receptor and estrogen
receptors alpha and beta in human fetal testis and epididymis. J Urol 2005;174:1695-8. (18)

Travis LB, Beard C, Allan JM et al. Testicular cancer survivorship: research strategies and
recommendations. J Nat Cancer Instit 2010;102:1114-30. (3)

Turnbull C, Rapley EA, Seal S et al. Variants near DMRT1, TERT and ATF7IP are associated
with testicular germ cell cancer. Nat Genet 2010;42:604-7. (9)

Vastermark A, Lundberg Giwercman Y, et al. Polymorphic variation in the androgen receptor
gene: association with risk of testicular germ cell cancer and metastatic disease. Eur J Cancer
2010; in press. (30)

Wanderas EH, Fossa SD, Tretli S. Risk of a second germ cell cancer after treatment of a primary
germ cell cancer in 2201 Norwegian male patients. Eur J Cancer 1997;33: 244-52. (12)

Wang W, John EM, Ingles SA. Androgen receptor and prostate-specific antigen gene
polymorphisms and breast cancer in African-American women. Cancer Epidemiol Biomarkers
Prev 2005;14:2990-4. (31)

Zhou X, Kudo A, Kawakami H, et al. Immunohistochemical localization of androgen receptor in
mouse testicular germ cells during fetal and postnatal development. Anat Rec 1996;245:509-18.
(35)

Zhou Q, Nie R, Prins GS et al. Localization of androgen and estrogen receptors in adult male
mouse reproductive tract. J Androl 2002;23:870-81. (17)
59

Chapter 3: Lower risk in parous women suggests hormonal factors important in
bladder cancer etiology

3.1 Abstract

Urinary bladder cancer is two-to-four times more common among men than women, a
difference in risk not fully explained by established risk factors. Our objective was to determine
whether hormonal and reproductive factors are involved in female bladder cancer.
We analyzed data from two population-based studies: the Los Angeles-Shanghai Bladder
Cancer Study, with 349 female case-control pairs enrolled in Los Angeles and 131 female cases
and 138 frequency-matched controls enrolled in Shanghai; and the California Teachers Study
(CTS), a cohort of 120,857 women with 196 incident cases of bladder urothelial carcinoma
diagnosed between 1995 and 2005. We also conducted a meta-analysis summarizing associations
from our primary analyses together with published results.
In primary data analyses, parous women experienced at least 30% reduced risk of bladder
cancer compared with nulliparous women (Shanghai: OR=0.38, 95%CI: 0.13-1.10; CTS:
RR=0.69, 95%CI: 0.50-0.95) consistent with results of a meta-analysis of nine studies (summary
RR=0.73, 95%CI: 0.63-0.85). The CTS, which queried formulation of menopausal hormone
therapy (HT), revealed a protective effect for use of combined estrogen and progestin compared
with no HT (RR=0.60, 95%CI: 0.37-0.98). Meta-analysis of three studies provided a similar
effect estimate (summaryRR=0.65, 95%CI: 0.48-0.88).
A consistent pattern of reduced bladder cancer risk was found among parous women and
those who used estrogen and progestin for HT. These results suggest that hormonal and
reproductive factors influence risk of female bladder cancer. Thus, processes mediated by steroid
hormones may contribute to the gender disparity in bladder cancer rates.
60

3.2 Introduction
Urinary bladder cancer is the fifth most common malignancy in industrialized nations. In
the US, 70,530 incident cases and 14,680 bladder cancer deaths were anticipated in 2010 (1), over
90% urothelial carcinoma (UC). Cigarette smoking and occupational exposure to several
arylamine compounds are established UC risk factors (2).
Occupational exposure to carcinogenic arylamines was dramatically reduced in the US by
banning use of 2-napthylamine, ceasing large-scale use of benzidine, and substituting other
compounds for 4-aminobiphenyl (3). Approximately half of UC diagnoses are now attributed to
cigarette smoke (4), which contains 4-aminobiphenyl and 2-napthylamine (3), but etiology among
non-smokers remains largely unknown.
Incidence is notably greater among men; age-adjusted US rates (US standard population,
2000) were 37.2/100,000 for men and 9.2/100,000 for women during 2003-2007 (1). Established
risk factors do not explain this disparity: absent exposure to cigarettes, occupational hazards and
urinary tract infections, men experience an estimated 2.7 times the risk of women (5). Postulated
explanations include gender differences in lifestyle (2), anatomy (5), and hormones (6).
Mechanisms involving steroid hormones seem plausible because there are fundamental
gender differences in production and response to these compounds, and the androgen receptor
(AR), estrogen receptors (ERs), and progesterone receptors (PRs) are expressed in human bladder
(7-12). Reportedly in rodent models, estrogens inhibit and androgens promote bladder tumor
growth (13, 14), incidence of chemically-induced bladder tumors is significantly greater among
male animals (15), and parous females have significantly smaller bladder tumors than nulliparous
females (16).
Hormonal and reproductive factors have not been a major focus of human bladder cancer
research, possibly due to small numbers of women enrolled in studies (17-29). Two studies
61

reported reduced risk among parous women (21, 27); additional studies addressing parity are
consistent with this effect, although not statistically significant (17-19, 26, 29). To further
investigate associations between hormonal and reproductive factors and UC risk among women,
we analyzed data from two large studies – the Los Angeles-Shanghai Bladder Cancer Study, and
the California Teachers Study, placing results in context with published reports.

3.3 Materials and Methods
The Los Angeles-Shanghai Bladder Cancer Study
Study Population
This population-based case-control study was conducted at sites representing high (Los
Angeles County (LA), California) and low (city of Shanghai, China) UC risk as described (4, 30).
In LA, the Los Angeles County Cancer Surveillance Program (31), a National Cancer Institute
Surveillance, Epidemiology and End Results (SEER) registry, identified incident UC cases
diagnosed between 1987 and 1996 among non-Asians aged 25-64 years. Of 2,384 eligible cases;
210 died before being contacted or were too ill to participate, physicians denied permission to
contact 99, 404 declined participation, and 1,671 (71%) were interviewed. Controls were
identified by standard procedure within cases’ neighborhood of residence (4), individually
matched to cases on sex, age (within five years), race/ethnicity (non-Hispanic white, Hispanic
white, African-American), and neighborhood. Of 1,586 interviewed; 1,090 (69%) were first, 325
(20%) were second, 111 (7%) were third, and 60 (4%) were ≥fourth eligible controls. In
Shanghai, the Shanghai Cancer Registry identified Han Chinese residents of Shanghai aged 25-74
years when diagnosed with bladder cancer between 1995 and 1998. Of 749 cases identified, 56
died before being contacted or were too ill to participate, 29 declined participation, 42 were not
located, and 622 (83%) were interviewed. Population-based controls were Han Chinese, selected
62

from Shanghai residents by established algorithm (32), and frequency-matched to cases by five-
year age groups. Among 726 randomly selected controls, 72 declined participation, 44 were not
located, and 610 (84%) were interviewed. All participants signed consent forms. This analysis
is limited to women: 349 case-control pairs from LA, 131 cases and 138 controls from Shanghai.

Data Collection
Participants were interviewed at home by trained interviewers using structured questionnaires
standardized across study sites. Cases provided information up to two years before cancer
diagnosis, and controls up to two years before the matched case’s diagnosis (LA) or two years
prior to interview (Shanghai). Questionnaires covered demographic characteristics, diet, alcohol
intake, tobacco use, medical history, and hormonal and reproductive history. Use of hair dyes
and non-steroidal anti-inflammatory drugs were collected only in LA.
In LA, women were asked number of pregnancies and whether they had undergone
hysterectomy. In Shanghai, women were asked numbers of sons and daughters to whom they had
given birth. At both sites, women were asked about use of hormones for contraception. In LA
women were asked about estrogen use for menopausal hormone therapy (HT). In Shanghai,
women were asked about use of estrogen injections for menopausal symptoms. Progestin use
was not queried at either site. Body mass index (BMI) was calculated and categorized using
WHO guidelines (33).

Statistical Analysis
Associations were measured by odds ratios (ORs) and corresponding 95% confidence
intervals (CIs). Most analyses of LA data used matched pairs conditional logistic regression,
although analyses stratified on smoking history used unconditional logistic regression. Shanghai
63

data were analyzed using unconditional logistic regression with adjustment for age. All models
were adjusted for smoking status in reference year (current, former, never), average number of
cigarettes smoked per day, years of smoking, and BMI (16-19.9, 20-24.9, 25-29.9; ≥30 <16 or
≥55 coded unknown); pack-years of smoking was included for analyses among ever smokers.
Estimates of main effects did not change by ≥10% with further adjustment of Shanghai data for
parity, or LA data for NSAID use, carotenoid intake, hair dye use, or pregnancy history, so these
variables were not retained. Analyses were performed using SAS version 9.1 (SAS Institute,
Cary, NC). Wald tests for trend were conducted with exposures coded as ordinal variables.
Reported P-values are two-sided.

The California Teachers Study (CTS)

Study Population
A detailed description of the CTS cohort is published (34). The cohort is composed of
current, former and retired female public school professionals who were members of the
California State Teachers Retirement System in 1995 when the study began. Cohort members
completed a detailed, mailed questionnaire that queried information on many factors including
menstrual and reproductive history, medical history, menopausal HT, diet, physical activity,
alcohol intake, and smoking. All collaborating institutions received institutional review board
approval for the study. The cohort is composed of 133,479 women. For this analysis sequential
exclusion for prior history of bladder cancer (n=130), unknown prior history of cancer (n=662),
residing outside California at baseline (n=8,867), consenting to participate only in breast cancer
research (n=18), baseline age ≥85 (n=2,199), unacceptable questionnaire (n=2), and unknown
smoking history (n=731) yielded a potential analytic cohort of 120,870.
64

Case Ascertainment and Follow-up
Incident cases of invasive bladder cancer (International Classification of Diseases for
Oncology ICD-O-2 site codes C67.0-C67.9) were identified through linkage to the California
Cancer Registry, which receives information on diagnoses in California based on state mandate
established in 1985, estimated to be 99% complete (35). During follow-up (1995-2005), 209
incident, invasive bladder tumors (including in situ) were diagnosed among the analytic cohort.
After excluding 13 diagnosed with non-UC bladder tumors, 120,857 women remained in this
analysis, 196 with UC (ICD-O-2 histology codes 8120 and 8130).
Person-time accrued from date of completion of baseline questionnaire until date of first
diagnosis with UC or first censoring event (relocation out of California for >four months; death;
end of follow-up, December 31, 2005; or ≥85 years of age). Residence in California was
determined though annual mailings of newsletters, linkage with US Postal Service National
Change of Address database, and change-of-address postcards submitted by participants. Date
and cause of death were obtained through California state mortality files, the Social Security
Administration death master file, and the National Death Index.

Exposure Assessment
Exposure measures are based upon participants’ responses to the baseline questionnaire.
Pregnancy history included all pregnancy outcomes, whereas parity was restricted to full-term
pregnancies (live births and stillbirths).
Menopausal status was determined using answers to detailed questions about history of
menstrual periods, hysterectomy, and ovarian surgeries with categories (premenopausal,
perimenopausal, postmenopausal, unknown menopausal status) defined as described (36).
65

HT use was categorized by duration, formulation (estrogen alone, estrogen-progestin
combination (E+P)), and for never, past, or current use as described (36).
Age at menarche, use of oral contraceptives (OCs), and breastfeeding history were
categorized as described (36).
The baseline questionnaire collected race/ethnicity. BMI was calculated and categorized
(33). History of cigarette smoking, cigarette smoking status, and number of cigarettes smoked
per day were collected. Smoking status was categorized as never, former, or current. Ever
smokers reported smoking ≥100 cigarettes during their lives. Pack-years of smoking were
calculated for ever smokers.

Statistical Analysis
Multivariate Cox proportional hazards regression was used to estimate associations. Hazard
rate ratios, presented as relative risks (RRs) with 95% confidence intervals (CIs) were estimated
using age in days at baseline as the time metric, stratified on age at baseline (in single years).
Models estimating associations with UC risk were adjusted for race/ethnicity (Non-Hispanic
white, African-American, Hispanic white, Asian/Pacific Islander, other/mixed race, unknown),
smoking status (never, former, current) and BMI (16-19.9, 20-24.9, 25-29.9, ≥30 kg/m
2
, values
<16 or ≥55 were coded unknown). We adjusted for smoking status as the only measure of
smoking history, as additional inclusion of pack-years did not alter inference. Other potential
confounders included alcohol intake, use of NSAIDs, history of diabetes and physical activity,
parity and HT use. These were not included in the final model because, with a single exception
(noted in text), adjustment did not change estimates of main effects by ≥10%. Wald tests for
trend were conducted with exposures coded as ordinal variables. Missing values were included
as indicator variables, and in all instances were not found to be associated with UC risk. Tests of
66

significance were two-sided. The proportional hazards assumption was assessed for each key
hormonal and reproductive variable by examining Kaplan-Meier curves and plotting scaled
Schoenfeld residuals to test for zero slope. No evidence of violation of the proportional hazards
assumption was detected.
Analyses were performed on eligible women with the following exclusions: 31,511 women
with no full-term pregnancy excluded from analyses of pregnancy and breastfeeding; 47,750
premenopausal and 2,476 perimenopausal women excluded from analyses of age at menopause;
47,750 premenopausal women excluded from analyses of HT. Associations between hormonal
and reproductive exposures and UC were further assessed in separate strata defined by ever or
never having been a smoker, with pack years of smoking included as a covariate in analyses
among ever smokers. Analyses were conducted using SAS statistical software, version 9.1 (SAS
Institute, Cary, NC).

Meta-Analysis
We conducted a meta-analysis of the associations between hormonal and reproductive factors
and bladder cancer risk combining our results with those from published studies. We searched
Medline and PubMed for articles published in English through December 2010, selecting
publications that 1) included a case group of women diagnosed with bladder cancer, 2) analyzed
associations between hormonal and/or reproductive exposures and bladder cancer risk, and 3)
addressed effects of smoking. Twelve articles met these criteria (17-25, 27-29). Two provided
data on the same cohort (18, 20); we retained the more recently published article (18).
We analyzed effects of ever-versus-never exposure to: parity, use of OCs, use of any HT, use
of E+P for HT, use of estrogen alone for HT. One study (18) did not provide risk estimates for
ever-versus-never parity, but did compare nulliparous women to each of several parous categories
67

defined by number of births. To include data from these studies, we first calculated RR and
variance estimates for ever-versus-never parity by weighing reported RR and variance estimates
for each parous category by corresponding number of person-years.
Analyses were performed using Stata statistical software (Stata/SE 9.0, College Station, TX).
For each analysis we estimated summary RR and corresponding 95% CI, and graphically
displayed estimates from each study and the summary estimate in a Forrest plot.
Variation due to differences in design and conduct of studies may manifest as between-study
heterogeneity, which we assessed in each analysis by calculating between-study heterogeneity p-
values (38) and I
2
statistics (37) and creating Begg’s funnel plots (39). I
2
range is 0-100%, higher

values indicating greater heterogeneity (0–30%,

mild; 30–50%, moderate; 50–100%,

notable)
(37). Begg’s funnel plots display RR estimate versus standard error of the RR for each study;
absent between-study heterogeneity, sampling variation alone tends to distribute results within the
“funnel” defined by pseudo 95% confidence limits. In the single analysis in which some studies
were outside these limits, we repeated the meta-analysis excluding outlying studies, and report
results for full and restricted sets of studies.
For a single analysis stratified on smoking history, we assessed heterogeneity of effects
between smokers and non-smokers by estimating a between-strata heterogeneity p-value.

3.4 Results
The Los Angeles-Shanghai Bladder Cancer Study
Characteristics of cases and controls appear in Table 3.1. Estimates of risk factor associations
appear in Table 3.2; only nonsmoking women from Shanghai are presented as there were few
smokers from this site (n=26), and results for Shanghai non-smokers are similar to those for all
Shanghai participants. Among all women in Shanghai, risk was lower among those who were
68

parous (OR=0.38 (95%CI: 0.13-1.10)) (not shown). Similarly, among all women in LA, risk was
lower among those who were never pregnant (OR=0.58 (95% CI: 0.33-0.98)). However, there
was no apparent trend over number of births (or pregnancies) or age at first birth (or pregnancy).
In analyses stratified on smoking history, effects of parity were greater among non-smokers in
Shanghai (non-smokers: OR=0.35 (95%CI: 0.11-1.08); smokers: OR=0.87 (95%CI: 0.05-15)
(data not shown)), as were effects of pregnancy in LA (non-smokers: OR=0.30 (95%CI: 0.13-
0.65); smokers: OR=1.29 (95%CI: 0.71-2.35)).
Bladder cancer risk was not statistically significantly associated with ever using OCs (LA
non-smokers: OR=0.83 (95%CI: 0.55-1.24); Shanghai nonsmokers: OR=0.82 (95% CI: 0.38-
1.79).
In LA, HT use was not associated with bladder cancer risk (OR=0.96 (95%CI: 0.65-1.42)).
However, data were not collected on use of progestin for HT, and data from Shanghai were too
sparse to estimate effects of using estrogen for HT.
Among additional reproductive history factors, hysterectomy, assessed only in LA, was not
significantly associated with risk (OR=1.16 (95%CI: 0.79-1.69)). Among Shanghai participants
there was a distinct association of risk with increasing BMI (p
trend
=0.008) among non-smokers;
however, no such pattern was observed among LA participants. Among non-smokers in LA, no
factor other than pregnancy history was associated with risk.

Table 3.1 Selected characteristics of women from the Los Angeles-Shanghai Bladder Cancer Study
*
______________________________________________________________________________________________________________
Los Angeles Shanghai

Cases (N = 349) Controls (N = 349) Cases (N = 131) Controls (N = 138)
N (%) N (%) N (%) N (%)
____________________________________________________________________________________________________________
Age at reference year
≤45 46 (13.2) 51 (14.6) 13 (9.9) 10 (7.2)
45-49 38 (10.9) 39 (11.2) 9 (6.9) 6 (4.3)
50-54 69 (19.8) 69 (19.8) 16 (12.2) 14 (10.1)
55-59 91 (26.1) 96 (27.5) 14 (10.7) 6 (4.3)
60-64 104 (29.8) 70 (20.1) 27 (20.6) 36 (26.1)
≥65 1 (0.30) 24 (6.9) 52 (39.7) 66 (47.8)
Race/ethnicity
Non-Hispanic white 319 (91.4) 320 (91.7) -- --
Hispanic white 16 (4.6) 19 (5.4) -- --
African-American 14 (4.0) 10 (2.9) -- --
Han Chinese -- -- 131 (100) 138 (100)

Body Mass Index
16-19.9 70 (20.0) 39 (11.2) 14 (10.7) 25 (18.1)
20-24.9 194 (55.6) 207 (59.3) 79 (60.3) 91 (65.9)
25-29.9 54 (15.5) 66 (18.9) 35 (26.7) 18 (13.0)
≥30 30 (8.6) 36 (10.3) 3 (2.3) 2 (1.5)
Unknown/out of range 1 (0.3) 1 (0.3) 0 2 (1.5)

Education
High school or less 164 (47.0) 140 (40.1) 101 (77.1) 118 (85.5)
At least one year of college 185 (53.0) 209 (59.9) 30 (22.9) 20 (14.5)
69

Table 3.1 (Continued)
Smoking status
Never 75 (21.5) 160 (45.8) 113 (86.3) 130 (94.2)
Former 94 (26.9) 111 (31.8) 4 (3.0) 1 (0.7)
Current 180 (51.6) 78 (22.4) 14 (10.7) 7 (5.1)
Parity
Nulliparous -- -- 15 (11.5) 6 (4.3)
Parous -- -- 116 (88.5) 132 (95.7)

Number of Children
1 -- -- 28 (24.2) 23 (17.4)
2 -- -- 28 (24.2) 25 (18.9)
3 -- -- 20 (171) 29 (22.0)
≥4 -- -- 40 (34.5) 55 (41.7)

Pregnancy
Never 45 (12.9) 37 (10.6) -- --
Ever 304 (87.1) 312 (89.4) -- --

Number of pregnancies
1 44 (14.4) 32 (10.3) -- --
2 64 (21.1) 85 (27.2) -- --
3 69 (22.7) 72 (23.1) -- --
≥4 127 (41.8) 123 (39.4) -- --

Oral contraceptive use
Never user 170 (48.7) 170 (48.7) 112 (85.5) 118 (85.5)
Ever user 177 (50.7) 179 (51.3) 19 (14.5) 20 (14.5)
Unknown 2 (0.6)

Hormone injection for contraception
Never user -- -- 128 (97.7) 133 (96.4)
Ever user -- -- 3 (3.3) 5 (3.6)
70

Table 3.1 (Continued)
Menopausal estrogen therapy use
Never user 193 (55.3) 194 (55.6) 129 (98.5) 138 (100)
Ever user 156 (44.7) 154 (44.1) 2 (1.5) 0
Unknown 1 (0.3)
Hysterectomy
No 233 (66.8) 243 (69.6) -- --
Yes 116 (33.2) 106 (30.4) -- --
_______________________________________________________________________________________________________________
*
N = number
71

Table 3.2 Adjusted odds ratios and 95% confidence intervals for associations between hormonal and reproductive factors and risk of
bladder cancer among women in the Los Angeles-Shanghai Bladder Cancer Study
*
_______________________________________________________________________________________________________________

Los Angeles Shanghai
_____________________________________________________ ________________________________
Cases/Controls All Women Nonsmokers Smokers Cases/Controls Nonsmokers
(All Women) OR (95% CI)
†
OR (95% CI)
‡
OR (95% CI)
‡
(Nonsmokers) OR (95% CI)
§

_______________________________________________________________________________________________________________
Parity
Nulliparous -- -- -- -- 11/5 1.0 (ref)
Parous -- -- -- -- 102/125 0.35 (0.11-1.08)

Number of Children
1 -- -- -- -- 27/22 1.0 (ref)
2 -- -- -- -- 25/25 0.63 (0.23-1.74)
3 -- -- -- -- 17/28 0.47 (0.16-1.33)
≥4 -- -- -- -- 33/50 0.56 (0.21-1.48)
p trend = 0.31
Pregnancy
Never 45/37 1.0 (ref) 1.0 (ref) 1.0 (ref) -- --
Ever 304/312 0.58 (0.33-0.98) 0.30 (0.13-0.65) 1.29 (0.71-2.35) -- --

Total number of pregnancies
1 44/32 1.0 (ref) 1.0 (ref) 1.0 (ref) -- --
2 64/85 0.41 (0.19-0.87) 0.42 (0.14-1.22) 0.71 (0.35-1.44) -- --
3 69/72 0.63 (0.29-1.34) 0.42 (0.12-1.37) 0.75 (0.37-1.51) -- --
≥4 127/123 0.70 (0.35-1.39) 0.79 (0.29-2.18) 0.84 (0.43-1.62) -- --
p trend = 0.73 p trend = 0.82 p trend = 0.93
Oral contraceptive use
Never 170/170 1.0 (ref) 1.0 (ref) 1.0 (ref) 95/110 1.0 (ref)
Ever 177/179 0.81 (0.55-1.19) 1.40 (0.75-2.62) 0.83 (0.55-1.24) 18/20 0.82 (0.38-1.79)
72

Table 3.2 (Continued)
Duration of oral contraceptive use
Non-user 170/170 1.0 (ref) 1.0 (ref) 1.0 (ref) 95/110 1.0 (ref)
<4 years 84/86 0.74 (0.47-1.18) 1.54 (0.70-3.36) 0.76 (0.48-1.22) 12/11 0.92 (0.36-2.37)
≥4 years 88/93 0.79 (0.50-1.25) 1.11 (0.53-2.32) 0.88 (0.54-1.43) 6/9 0.68 (0.22-2.16)
p trend = 0.28 p trend = 0.51 p trend = 0.54 p trend = 0.54
Menopausal estrogen therapy use
Never 193/194 1.0 (ref) 1.0 (ref) 1.0 (ref) -- --
Ever 156/154 0.96 (0.65-1.42) 1.00 (0.54-1.86) 1.03 (0.70-1.54) -- --

Duration of estrogen use
(Any type – pill, patch, injection or cream)
Non-user 193/194 1.0 (ref) 1.0 (ref) 1.0 (ref) -- --
< 4 years 76/70 1.11 (0.69-1.80) 1.02 (0.48-2.17) 1.13 (0.69-1.86) -- --
≥4 years 80/84 0.86 (0.54-1.38) 1.00 (0.47-2.14) 0.96 (0.60-1.54) -- --
p trend = 0.61 p trend = 0.98 p trend = 0.93
Hysterectomy
Never 233/243 1.0 (ref) 1.0 (ref) 1.0 (ref) -- --
Ever 116/106 1.16 (0.79-1.69) 0.83 (0.44-1.58) 1.23 (0.82-1.86) -- --

Body Mass Index
16-19.9 70/39 2.25 (1.29-3.92) 1.90 (0.78-4.64) 1.84 (1.08-3.11) 13/23 0.67 (0.31-1.47)
20-24.9 194/207 1.0 (ref) 1.0 (ref) 1.0 (ref) 67/87 1.0 (ref)
25-29.9 54/66 0.72 (0.44-1.22) 0.79 (0.36-1.75) 0.82 (0.49-1.37) 30/18 2.31 (1.14-4.47)
≥30 30/36 1.46 (0.77-2.78) 0.55 (0.23-1.33) 2.73 (1.06-7.03) 3/2 2.11 (0.33-13)
p trend = 0.12 p trend = 0.10 p trend = 0.52 p trend = 0.008
_______________________________________________________________________________________________________________
*
OR = odds ratio; CI = confidence interval
†
Conditional logistic regression adjusted for age at reference year, race/ethnicity (Non-Hispanic white, Hispanic-white, or African-American), smoking
status (never, former, or current), number of cigarettes per day, number of years of smoking and body mass index (<20 20-24.9, 25-29.9, ≥30, unknown)
‡
Unconditional logistic regression adjusted for age at reference year, race/ethnicity (Non-Hispanic white, Hispanic-white, or African-American), and
body mass index (<20 20-24.9, 25-29.9, ≥30, unknown)
§
Unconditional logistic regression adjusted for age at reference year and body mass index (<20 20-24.9, 25-29.9, ≥30, unknown)

73
74

The California Teachers Study (CTS)
Characteristics of cohort members appear in Table 3.3. Estimates of risk factor associations
appear in Table 3.4. Ever pregnant and parous women had significantly lower UC risk than never
pregnant or nulliparous women (ever pregnant, RR=0.60 (95%CI: 0.43-0.83); parous, RR=0.69
(95%CI: 0.50-0.95)) (Table 3.4). Among parous women, risk was not associated with age at first
full-term pregnancy, number of full-term pregnancies, or history of breastfeeding. Among all
women, neither age at menarche nor history of OC use was associated with risk.
Post-menopausal women did not have increased UC risk compared with pre- and peri-
menopausal women (RR=1.02 (95%CI: 0.48-2.19) for natural menopause; RR=1.39 (95%CI:
0.63-3.09) for menopause due to bilateral oophorectomy). No significant association was found
for age at menopause, history of oophorectomy or hysterectomy, or BMI.
Among peri-menopausal and postmenopausal women, those who used E+P for HT
experienced significantly lower risk than those who used no HT (RR=0.60, (95%CI: 0.37-0.98)).
Use of estrogen alone was not associated with risk (RR=1.18 (95%CI: 0.83-1.70)). No case
reported using only progestin.
Among non-smoking women, parity (RR=0.61 (95%CI: 0.37-1.00)) and use of E+P for HT
appeared protective (RR=0.49 (95%CI: 0.22-1.10)), although associations were not statistically
significant. The association with history of hysterectomy was not significant when HT use
(never, ever estrogen alone, ever other formulation) was included in the model (RR=1.65
(95%CI: 0.96-2.81).
Estimates of pregnancy-UC associations were similar to estimates of parity-UC associations
in all analyses.

Table 3.3 Selected characteristics of the California Teachers Study cohort, by smoking status
*

________________________________________________________________________________________________
Characteristic Non-Smokers Smokers
N (%) N (%)
________________________________________________________________________________________________
All participants 79,886 (66.1) 40,971 (33.9)

Age at baseline
<35 10,949 (13.7) 1,763 (4.3)
35-44 16,358 (20.5) 5,021 (12.3)
45-54 23,340 (29.2) 12,994 (31.7)
55-64 13,077 (16.4) 10,328 (25.2)
65-74 10,242 (12.8) 7,544 (18.4)
75-84 5,920 (7.4) 3,321 (8.1)

Race/ethnicity
Non-Hispanic White 67,847 (84.9) 36,700 (89.6)
African-American 2,122 (2.7) 1,134 (2.8)
Hispanic White 4,079 (5.1) 1,112 (2.7)
Asian/Pacific Islander 3,470 (4.3) 896 (2.2)
Other/Mixed Race 1,722 (2.2) 789 (1.9)
Unknown 646 (0.8) 340 (0.8)

Body Mass Index
16-19.9 8,938 (11.2) 3,670 (9.0)
20-24.9 38,526 (48.2) 19,559 (47.7)
25-29.9 18.689 (23.4) 10,320 (25.2)
≥30 10,724 (13.4) 5,646 (13.8)
Unknown/out of range 3,009 (3.8) 1,776 (4.3)
75

Table 3.3 (Continued)
Age at menarche
<12 17,878 (22.4) 9,081 (22.2)
12-13 44,766 (56.0) 22,956 (56.0)
≥14 16,107 (20.2) 8,325 (20.3)
Unknown/no first menstrual period 1,135 (1.4) 609 (1.5)

Parity
Never pregnant 17,026 (21.3) 7,433 (18.1)
Ever pregnant, no full-term 4,153 (5.2) 2,899 (7.1)
Ever pregnant, having any full-term 57,774 (72.3) 30,118 (73.5)
Unknown 933 (1.2) 521 (1.3)

Total number full-term pregnancies
†

1 12,141 (20.7) 6,469 (21.1)
2 25,808 (44.1) 13,070 (42.7)
3 12,491 (21.3) 6,744 (22.0)
≥4 6,617 (11.3) 3,576 (11.7)
Unknown 1,512 (2.6) 761 (2.5)

Age at first pregnancy
‡

<20 2,715 (4.6) 1,776 (5.8)
20-25 22,098 (37.7) 12,730 (41.6)
26-30 22,213 (37.9) 10,390 (33.9)
>30 10,030 (17.1) 4,962 (16.2)
Unknown 1,513 (2.6) 762 (2.5)

Oral contraceptive use
Never user 25,463 (31.9) 12,945 (31.6)
Ever user 52,018 (65.1) 26,872 (65.6)
Unknown 2,405 (3.0) 1,154 (2.8)
76

Table 3.3 (Continued)
Menopausal status
Premenopausal 36,425 (45.6) 11,325 (27.6)
Perimenopausal 1,585 (2.0) 920 (2.3)
Postmenopausal 35,388 (44.3) 24,977 (61.0)
Unknown/all other categories 6,488 (8.1) 3,749 (9.1)

Hysterectomy
Never 60,875 (76.2) 29,274 (71.5)
Ever 17,305 (21.7) 10,849 (26.5)
Unknown 1,706 (2.1) 848 (2.0)

Ovary Removed
None 66,549 (83.3) 32,436 (79.2)
One Removed 3,360 (4.2) 2,250 (5.5)
Both Removed 8,840 (11.1) 5,630 (13.7)
Unknown 1,137 (1.4) 655 (1.6)

Menopausal hormone therapy use
§

Never user 10,379 (23.9) 6,334 (21.4)
Estrogen alone only user 12,668 (29.1) 8,288 (28.0)
E+P only user 11,669 (26.8) 8,855 (29.9)
Estrogen alone and E+P user 3,914 (9.0) 3,096 (10.4)
Progestin alone or progestin and
other type of HT user 1,363 (3.2) 814 (2.7)
Unknown 3,468 (8.0) 2,259 (7.6)
______________________________________________________________________________________________
*
N = Number; E+P = Estrogen plus progestin; HT = menopausal hormone therapy
†
Includes pregnancies resulting in live birth or still birth, 89,189 women remain
‡
Among women with full-term pregnancies, 89,189 women remain
§
Premenopausal women excluded, 73,107 women remain
77

Table 3.4 Adjusted relative risks and 95% confidence intervals for associations between selected menstrual, hormonal and reproductive
factors and risk of bladder cancer in the California Teachers Study
*
_______________________________________________________________________________________________________________
All Women Nonsmoking Women Smoking Women

Characteristics N Total N Cases RR (95% CI)
†
N Total N Cases RR (95% CI)
‡
N Total N Cases RR (95% CI)
‡
120,857 196 79,886 82 40,971 115
_______________________________________________________________________________________________________________
Age at menarche
<12 26,959 41 1.0 (ref) 17,878 22 1.0 (ref) 9,081 19 1.0 (ref)
12 32,606 55 1.08 (0.72-1.62) 21,503 19 0.68 (0.37-1.27) 11,103 36 1.51 (0.87-2.64)
13 35,116 52 0.91 (0.60-1.38) 23,263 21 0.67 (0.37-1.22) 11,853 31 1.16 (0.65-2.06)
14+ 24,432 41 0.94 (0.61-1.46) 16,107 17 0.71 (0.37-1.35) 8,325 24 1.18 (0.64-2.16)
Unknown or missing 1,744 7 1,135 3 609 4
p trend = 0.62 p trend = 0.42 p trend = 0.98
Pregnancy
Never 24,459 48 1.0 (ref) 17,026 22 1.0 (ref) 7,433 26 1.0 (ref)
Ever 94,944 141 0.60 (0.43-0.83) 61,927 58 0.53 (0.32-0.87) 33,017 83 0.65 (0.42-1.01)
Unknown or missing 1,454 7 933 2 521 5

Parity
§

Nulliparous 31,511 52 1.0 (ref) 21,179 23 1.0 (ref) 10,332 29 1.0 (ref)
Parous 87,735 137 0.69 (0.50-0.95) 57,636 57 0.61 (0.37-1.00) 30,099 80 0.78 (0.47-1.29)
Unknown or missing 1,611 7 1,071 2 540 5

Age at first full-term pregnancy
**
<20 4,491 5 0.78 (0.31-1.95) 2,715 4 1.77 (0.61-5.21) 1,776 1 0.24 (0.03-1.78)
20-25 34,828 62 1.0 (ref) 22,098 21 1.0 (ref) 12,730 41 1.0 (ref)
26-30 32,603 51 1.10 (0.76-1.62) 22,213 22 1.22 (0.66-2.25) 10,390 29 1.04 (0.64-1.68)
>30 14,992 19 1.00 (0.59-1.71) 10,030 10 1.30 (0.58-2.82) 4,962 9 0.86 (0.41-1.80)
Unknown or missing 2,432 7 1,651 2 781 5
p trend = 0.56 p trend = 0.92 p trend = 0.49
78

Table 3.4 (Continued)
Total number of full-term pregnancies
**

1 18,610 24 1.0 (ref) 12,141 11 1.0 (ref) 6,469 13 1.0 (ref)
2 38,878 60 1.14 (0.71-1.86) 25,808 25 0.98 (0.49-2.00) 13,070 35 1.28 (0.68-2.43)
3 19,235 27 0.73 (0.42-1.27) 12,491 9 0.52 (0.22-1.26) 6,744 18 0.91 (0.44-1.89)
4+ 10,193 26 1.06 (0.61-1.87) 6,617 12 1.04 (0.46-2.56) 3,576 14 1.06 (0.49-2.30)
Unknown or missing 2,430 7 1,650 2 780 5
p trend = 0.60 p trend = 0.68 p trend = 0.73
Breastfeeding
**

Never breastfed 19,684 46 1.0 (ref) 11,622 19 1.0 (ref) 8,062 27 1.0 (ref)
Ever breastfed 67,003 91 0.93 (0.64-1.33) 45,311 38 0.79 (0.45-1.37) 21,692 53 1.05 (0.66-1.67)
Unknown or missing 2,659 7 1,774 2 885 5

Oral contraceptive use
Never user 38,408 107 1.0 (ref) 25,463 48 1.0 (ref) 12,945 59 1.0 (ref)
Ever user 78,890 79 1.05 (0.73-1.50) 52,018 29 0.92 (0.51-1.67) 26,872 50 1.09 (0.70-1.69)
Unknown or missing 3,559 10 2,405 5 1,154 5

Menopausal status
Pre- or peri-menopausal 50,226 15 1.0 (ref) 37,989 10 1.0 (ref) 12,237 5 1.0 (ref)
Post-menopausal, natural 35,045 100 1.02 (0.48-2.19) 20,202 35 0.71 (0.25-2.04) 14,843 65 1.49 (0.46-4.82)
Post-menopausal, surgical
††
10,172 30 1.39 (0.63-3.09) 6,200 14 1.25 (0.42-3.69) 3,972 16 1.70 (0.50-5.80)
Unknown
§§
or missing 25,414 51 15,495 23 9,919 28

Age at menopause
‡‡
53 or older 11,097 35 1.0 (ref) 6,614 11 1.0 (ref) 4,483 24 1.0 (ref)
47-52 21,776 64 1.02 (0.68-1.55) 12,580 26 1.41 (0.69-2.86) 9,196 38 0.85 (0.50-1.42)
44-46 5,889 15 0.94 (0.51-1.73) 3,409 7 1.44 (0.56-3.76) 2,480 8 0.71 (0.32-1.58)
43 or younger 6,455 16 1.07 (0.59-1.93) 3,799 5 1.11 (0.38-3.23) 2,656 11 1.02 (0.50-2.10)
Unknown
§§
or missing 25,414 51 15,495 23 9,919 28
p trend = 0.93 p trend = 0.72 p trend = 0.83
79

Table 3.4 (Continued)
Ever use of menopausal hormone therapy (HT)
***
Never HT user 16,713 49 1.0 (ref) 10,379 24 1.0 (ref) 6,334 25 1.0 (ref)
Ever HT user 50,928 120 0.93 (0.66-1.31) 29,781 45 0.78 (0.47-1.28) 21,147 75 1.01 (0.64-1.60)
Unknown or missing 5,466 16 3,301 6 5,165 10

Type of HT used
***

Never HT user 16,713 49 1.0 (ref) 10,379 24 1.0 (ref) 6,334 25 1.0 (ref)
Estrogen alone only user 20,956 75 1.18 (0.83-1.70) 12,668 30 0.98 (0.57-1.68) 8,288 45 1.31 (0.80-2.14)
E+P only user 20,524 26 0.60 (0.37-0.98) 11,669 9 0.49 (0.22-1.10) 8,855 17 0.65 (0.35-1.24)
E alone and E+P user 7,010 16 0.75 (0.42-1.32) 3,914 6 0.70 (0.28-1.73) 3,096 10 0.76 (0.36-1.58)
P alone and E+P user 2,177 3 0.74 (0.23-2.41) 1,363 0 -- 814 3 1.41 (0.42-4.72)
Unknown or missing 5,727 16 3,468 6 2,259 10

Past or current estrogen and progestin use
***
Never HT user 16,713 49 1.0 (ref) 10,379 24 1.0 (ref) 6,334 25 1.0 (ref)
Past E+P user 3,063 7 0.93 (0.42-2.06) 1,696 3 0.98 (0.29-3.33) 1,367 4 0.87 (0.30-2.51)
Current E+P user 17,046 19 0.55 (0.31-0.95) 9,740 6 0.40 (0.15-1.02) 7,306 13 0.62 (0.31-1.25)
E alone user 20,956 75 1.18 (0.83-1.70) 12,668 30 0.98 (0.57-1.67) 8,288 45 1.31 (0.80-2.14)
E alone and E+P 9,863 19 0.72 (0.42-1.22) 5,677 6 0.55 (0.22-1.35) 4,186 13 0.82 (0.42-1.60)
or P alone and E+P user
Unknown or missing 5,466 16 3,301 6 2,165 10

Duration of estrogen and progestin use
***
Never HT user 16,713 49 1.0 (ref) 10,379 24 1.0 (ref) 6,334 25 1.0 (ref)
E+P only user, <1-2 years 7,196 9 0.73 (0.35-1.54) 4,312 4 0.70 (0.23-2.11) 2,884 5 0.72 (0.27-1.93)
E+P only user, 3+ years 12,448 17 0.61 (0.35-1.08) 6,865 5 0.45 (0.17-1.22) 5,583 12 0.69 (0.34-1.40)
p trend = 0.11 p trend = 0.13 p trend = 0.35
E alone, or E alone and E+P
or P alone and E+P user 31,284 94 1.04 (0.73-1.47) 18,604 36 0.86 (0.52-1.44) 12,680 58 1.14 (0.71-1.82)
Unknown or missing 5,466 16 3,301 6 2,165 10
80

Table 3.4 (Continued)
Hysterectomy
Never 90,149 106 1.0 (ref) 60,875 40 1.0 (ref) 29,274 66 1.0 (ref)
Ever 28,154 81 1.29 (0.97-1.74) 17,305 39 1.61 (1.02-2.55) 10,849 42 1.09 (0.74-1.62)
Unknown or missing 2,554 9 1,706 3 848 6

Ovary removed
None 98,985 138 1.0 (ref) 66,549 56 1.0 (ref) 32,436 82 1.0 (ref)
One removed 5,610 14 1.13 (0.66-1.97) 3,360 5 1.10 (0.74-2.76) 2,250 9 1.17 (0.58-2.33)
Both removed 14,470 40 1.14 (0.81-1.63) 8,840 18 1.26 (0.74-2.16) 5,630 22 1.07 (0.66-1.71)
Unknown or missing 1,792 4 1,137 3 655 1

Body Mass Index
16-19.9 12,608 14 0.80 (0.46-1.42) 8,938 6 0.83 (0.35-1.98) 3,670 8 0.82 (0.39-1.73)
20-24.9 58,085 94 1.0 (ref) 38,526 38 1.0 (ref) 19,559 56 1.0 (ref)
25-29.9 29,009 44 0.77 (0.54-1.10) 18,689 19 0.84 (0.48-1.45) 10,320 25 0.73 (0.45-1.16)
≥30 16,370 26 0.97 (0.64-1.50) 10,724 12 1.09 (0.58-2.11) 5,646 14 0.85 (0.47-1.53)
Unknown or missing 4,785 18 3,009 7 1,776 11
p trend = 0.85 p trend = 0.82 p trend = 0.49
_______________________________________________________________________________________________________________
*
N = number; RR = relative risk; CI = confidence interval; HT = menopausal hormone therapy; E+P = estrogen plus progestin; E = estrogen; P =
progestin
†
Stratified on age at baseline and adjusted for smoking status (never, former, current), race/ethnicity (Non-Hispanic white, African-American, Hispanic
white, Asian/Pacific Islander, Mixed or other race, unknown), body mass index (<25, 25-29.9, 30+, unknown)
‡
Stratified on age at baseline and adjusted for race/ethnicity (Non-Hispanic white, African-American, Hispanic white, Asian/Pacific Islander, Mixed or
other race, unknown) and body mass index (<25, 25-29.9, 30+, unknown)
§
Nulliparous includes women who were never pregnant or did not have a full-term pregnancy
**
Excluding women with no full-term pregnancies
††
Surgical menopause was defined as undergoing a bilateral oophorectomy before occurrence of natural menopause
‡‡
Pre- and peri-menopausal women excluded
§§
Includes women who underwent a hysterectomy before age 56 and were less than 56 years old at baseline, or had menopause due to chemotherapy or
radiation, or had menopause due to other reasons, or had unknown menopausal status due to hormone therapy.
***
Pre-menopausal women excluded
81
82

Meta-analysis
Reports included in meta-analyses are enumerated in 3.5. Seven provided data on parity,
of which two also provided parity data stratified by smoking history; five provided data on OC
use; eight provided data on any use of HT, of which two specified HT formulation. Associations
between these factors and bladder cancer estimated in individual studies, and summary estimates
are displayed as Forrest plots (Figures 3.1 and 3.2), with summary estimates provided for
subgroups of like study design and overall. Summary results appear in Table 3.6.
There was significantly reduced risk among parous women (summaryRR=0.73 (95%CI:
0.63-0.85), Figure 3.1A), with no indication of heterogeneity between studies. Summary
estimates within strata defined by smoking history reveal the parity-UC association to be greater
among nonsmokers (summaryRR=0.51 (95%CI: 0.37-0.69)) than smokers (summaryRR=0.82
(95%CI: 0.56-1.21); between-group p
heterogeneity
=0.05)) (Figure 3.1B).
Data on OCs show no effect of ever use (summaryRR=0.94 (95%CI: 0.81-1.09)) (Figure 3.2A),
with no indication of between-study heterogeneity.
Data on any use of HT provide no indication of association with risk (summaryRR=1.01
(95% CI 0.90-1.13)) (Figure 3.2B). Results of two studies (23, 24) were clearly outside 95%
pseudo-confidence limits (not shown) (between-study p
heterogeneity
=0.06). After excluding these
outlying studies, no indication of heterogeneity remained (between-study p
heterogeneity
=0.95), and
the association remained null (summaryRR=0.96 (95%CI: 0.85-1.07)).
Data on HT formulation, available for three studies, suggested little or no increase in risk
following use of E alone (summaryRR=1.14 (95% CI: 0.92-1.40)) (Figure 3.2C), but clearly
suggest a protective effect of using E+P for HT (summaryRR=0.65 (95%CI: 0.48-0.88)) (Figure
3.2D), the association persisting in summary analysis limited to the two published studies
(summaryRR=0.68 (95%CI: 0.46-1.00), not shown).

Table 3.5 Published reports on hormonal and reproductive exposures and risk of bladder cancer in women
*
_______________________________________________________________________________________________________________
Study Year Study Cases/Controls Exposures Meta-analyzed trend p value
Design or cohort size Parity OC HT E E+P Age at first birth N births
_______________________________________________________________________________________________________________
A. Data included in the meta-analysis

Prizment et al. (18) 2006 Cohort 192/37,459 X X X 0.25 0.38

Cantwell et al. (19) 2006 Cohort 167/54,308 X X X X X NA
†
NA
†

McGrath et al. (17) 2006 Cohort 336/116,598 X X X X X 0.50 0.30

Olsson et al. (28) 2003 Cohort 22/29,508 X

Persson et al. (24) 1996 Cohort 58/22,597 X

Huang et al. (27) 2009 Hosp. ca-ctrl 152/166 X 0.74 NA
†

Fernandez et al. (23) 2003 Hosp. ca-ctrl 106/6976 X

Pelucchi et al. (22) 2002 Hosp. ca-ctrl 110/298 X X

Cantor et al. (21) 1992 Pop. ca-ctrl 317/833 X 0.15 0.56

La Vecchia et al. (25)

1993 Hosp ca-ctrl 68/5619 X NA
†
NA
†

Dietrich et al. (29) 2010 Pop. ca-ctrl 207/463 X X X NA
†
0.63

B. Data excluded from the meta-analysis

Tripathi et al. (20)
‡
2002 Cohort 112/37,459

Miller et al. (26)
§
1980 Cohort 113/11,127
83

Table 3.5 (Continued)
_______________________________________________________________________________________________________________
*
Parity = ever-versus-never completion of full term pregnancy; OC = ever-versus-never use of oral contraceptives; HT = ever-versus-never use of
menopausal hormone therapy; E = ever-versus-never use of estrogen alone for menopausal hormone therapy; E+P = ever-versus-never use of estrogen
plus progestin for menopausal hormone therapy; X = exposure was analyzed in the meta-analysis for the corresponding study; N = number; Hosp. ca-
ctrl = Hospital-based case-control study; Pop. ca-ctrl = population-based case-control study
†
NA = p value for trend not available
‡
Same cohort has Prizment et al. 2006
§
Data could not be extracted for meta-analysis due to absence of any measure of variance
84

Table 3.6 Contributing data and summary relative risks and 95% confidence intervals from the meta-analysis of hormonal and
reproductive exposures and risk of bladder cancer in women
*

_______________________________________________________________________________________________________________
Exposure Studies contributing N cases Summary 95% CI p
†
heterogeneity
I
2
N studies
RR outside pseudo
95%
confidence limits
_______________________________________________________________________________________________________________
Parity CTS, Shanghai, 17-19, 21,
25, 27, 29 1698 0.73 0.63-0.85 0.64 0 0
Among nonsmokers CTS, Shanghai, 21, 27 468 0.51 0.37-0.69 0.69 0 0
Among smokers CTS, Shanghai, 21, 27 326 0.82 0.56-1.21 0.95 0 0
p
‡

= 0.05
Parity excluding study CTS, Shanghai, 17-19, 21 1630 0.71 0.61-0.83 0.76 0 0
with no smoking adjustment 27. 29

Oral contraceptive use CTS, LA, Shanghai, 17-19,
22, 29 1688 0.94 0.81-1.09 0.63 0 0

Any HT, all studies CTS, LA, 17-19, 22-24,
28, 29 1743 1.01 0.90-1.13 0.06 45 2
HT excluding outlying studies CTS, LA, 17-19, 24, 28, 29 1527 0.96 0.85-1.07 0.95 0 0

HT excluding outlying studies CTS, LA, 17-19, 28, 29 1469 0.98 0.86-1.13 0.96 0 0
and study with no smoking adjustment

Estrogen alone for HT CTS, 17, 19 699 1.14 0.92-1.40 0.88 0 0

E+P for HT CTS, 17, 19 699 0.65 0.48-0.88 0.79 0 0
______________________________________________________________________________________________________________
85

Table 3.6 (Continued)

*
N = number; Summary RR = Summary relative risk; CI = confidence interval; I
2
= percentage of variation in summary estimate due to heterogeneity
between studies; CTS = California Teachers Study; LA = Los Angeles bladder cancer study; Shanghai = Shanghai bladder cancer study; HT =
menopausal hormone therapy; E+P = estrogen plus progestin
†
Between-study heterogeneity p value (studies contributing to each summary RR)
‡
P value for nonsmokers versus smokers, stratified by study

86
87

Figure 3.1A. Forrest plots displaying contributing data and results of meta-analyses relating
parity to risk of bladder cancer, comparing all parous and nulliparous women. Summary
relative risk (RR) is displayed as top and bottom points of diamond, and 95% CI of summary
RR as left and right points of diamond. Open diamond represents stratum-specific summary
RR; filled diamond represents overall summary RR. For each contributing study, RR
estimate is displayed as a point, and 95% confidence interval (CI) as a horizontal line; box
size is proportional to study’s relative weight in summary RR.

88

Figure 3.1B. Forrest plots displaying contributing data and results of meta-analyses relating
parity to risk of bladder cancer, comparing parous and nulliparous women within strata of
smoking history.

89

Figure 3.2A. Forrest plots displaying contributing data and results of meta-analyses relating
history of exogenous hormone use to risk of bladder cancer, comparing never and ever users
of oral contraceptives.

90

Figure 3.2B. Forrest plots displaying contributing data and results of meta-analyses relating
history of exogenous hormone use to risk of bladder cancer, comparing never and ever users of
any menopausal hormone therapy (HT).

Figure 3.2C. Forrest plots displaying contributing data and results of meta-analyses relating
history of exogenous hormone use to risk of bladder cancer, comparing never and ever users of
estrogen alone for HT.

91

Figure 3.2D. Forrest plots displaying contributing data and results of meta-analyses relating
history of exogenous hormone use to risk of bladder cancer, comparing never and ever users of
estrogen plus progestin for HT.

92

3.5 Discussion

These analyses revealed a completely consistent pattern of lower UC risk among parous
women in the CTS, Shanghai case-control data, and published epidemiologic studies addressing
parity. A similar pattern was evident in the LA case-control data, in which births were not
measured, but history of any pregnancy was associated with lower risk. Published studies
examining the parity-UC association (17-19, 21, 25, 27, 29) reported estimates consistent with
lower risk among parous women, but most were not statistically significant, likely because of
limited numbers of females in individual studies. Most or all reduction in risk may be related to
the first birth (or pregnancy), because risk does not appear to depend on number of births (or
pregnancies) beyond the first, or on age at first birth (or pregnancy).
The parity-UC association appears more pronounced among women who never smoked. In
Shanghai case-control and CTS cohort data, effects of parity appeared to be stronger among never
smokers. This pattern persisted in summary estimates of four studies stratified on smoking
history, with significant heterogeneity between never- versus ever-smokers, and was reinforced
by stronger pregnancy-UC association among non-smokers in LA. It is not clear whether the
mechanism whereby parity is associated with reduced risk operates primarily among nonsmokers,
or effects of smoking on risk are simply so great that effects of parity are not apparent among
smokers. Nonetheless, dramatic effects of parity among non-smokers suggest that future efforts to
understand the biological basis of the parity-UC association may provide long-awaited insights
into UC causes among non-smoking women.
During pregnancy, the bladder undergoes dramatic alterations in structure, function,
histology, and gene expression; steroid hormones may govern these changes, since they can be
recapitulated in rodents by treatment with estrogen and progesterone (40), and estrogen and
93

progesterone rise dramatically during pregnancy achieving levels unequalled at any other time of
a woman’s life.
Epidemiologic studies previously established that parous women tend to experience lesser
risk of cancers of the breast, endometrium, and ovary (41). Studies seeking the biologic basis for
reduced breast cancer risk years after pregnancy have demonstrated patterns of gene expression
that differ between healthy breast tissue of parous and nulliparous women (42, 43). Some
differences persist at least a decade after pregnancy, and include expression steroid hormone
targets: ER α, ER β (43), and progesterone receptor membrane component 2 (42). Thus, persistent
parity-related changes in gene expression may plausibly influence malignant potential of the
breast. Although we are unaware of research addressing persistence of pregnancy-related gene
expression in the bladder, consistently and substantially reduced risk of bladder cancer among
parous women strengthens our suspicion that hormone-mediated carcinogenesis may also be
important in the human bladder.
An alternate mechanism whereby pregnancy may reduce risk of other cancers is cessation of
hormonal cycling, thus lower lifetime number of menstrual cycles. This seems a less likely
mechanism in bladder cancer, not only because men -- who do not have menstrual cycles --
experience greater risk, but also because in women proxies for cycle number appeared in our
analyses to be unrelated to risk. In the CTS we observed no association with age at menarche,
age at menopause, history of breastfeeding, or number of full term pregnancies; in the CTS, LA,
Shanghai, and summary data (17-19, 22, 29) history of OC use was not associated with risk.
However, we did not model lifetime number of cycles, and three cohort studies previously
reported increased bladder cancer risk among women with menopause by age 42 (18) or 45 (17,
19), one result not statistically significant (19).
94

Two non-hormonal effects of pregnancy on the bladder warrant consideration. First, urinary
incontinence, particularly stress incontinence, is reportedly more prevalent among parous women
(44), and resulting increases in frequency of urination may in theory reduce bladder cancer risk
(45). To address this possibility, we examined frequency of daytime urination as a possible
modifier of the parity-UC association in LA case-control data, finding no such modification. A
second possibility is that an unrecognized common cause of both UC and infertility could create a
spurious association with parity. This may be plausible because the bladder forms from the
urogenital sinus, which also gives rise to much of the reproductive system. However, in analyses
of CTS data in which we excluded from the nulliparous group women who reported inability to
achieve pregnancy, a robust association persisted between parity and reduced UC risk.
We found inconsistent associations of BMI with bladder cancer risk, detecting no association
among California residents in the CTS or LA component of the case-control study. However,
among Shanghai participants there was a dramatic association of risk with increasing BMI. We
could not determine whether this trend represents a true BMI-bladder cancer association that is
more apparent in populations with predominantly low BMI, other important but unrecognized
differences between women of Shanghai versus California, or a spurious association. Higher
BMI generally indicates greater stores of adipose tissue, which in the postmenopausal period can
elevate levels of bioactive estrogen by aromatization of androgens. Thus, the trend observed
among women in Shanghai may represent an association between postmenopausal estrogen levels
and UC risk, a possibility warranting investigation by more direct methods.
Exogenous hormone exposure was measured by reported use of OCs and menopausal HT. In
CTS, LA and Shanghai data and meta-analysis of five data sets (17-19, 22, 29), ever-versus-never
OC use was not associated with risk. Ever-versus-never use of any HT was not associated with
risk in CTS or LA data, or meta-analysis of eight studies (17-19, 22-24, 28, 29). Statistically
95

significant OR estimates from two hospital-based case-control studies (22, 23) exceeded 1.0, but
may reflect substantial bias, as articulated by authors of one original report (22); these studies
little influenced the summary estimate.
To examine in primary data effects of estrogens alone and E+P, we relied on the CTS cohort,
because use of progestin for HT was not measured in the case-control study. CTS data and meta-
analysis of data from two cohorts (17, 19) suggest that use of E+P for HT is associated with a 35-
40% reduction in bladder cancer risk. The same studies suggest that use of estrogen alone may be
associated with somewhat increased risk, but individual and summary estimates were not
statistically significant. Nonetheless, these results cause us to question whether null associations
estimated for ever-versus-never use of HT of unspecified formulation may represent a mixture of
protective effects of progestin and harmful effects of estrogens, a pattern previously demonstrated
for ovarian cancer (46).
Clearer understanding of any role of exogenous hormones on bladder cancer risk may follow
detailed analysis addressing duration, formulation, and schedule of HT use, and analyses of OC
use by these quantitative measures together with timing of OC use relative to pregnancy and
childbirth. Pooled analysis of extant studies may provide considerable insight.
In premenopausal women, progesterone levels change over the menstrual cycle, but are
highest by far late in pregnancy. At menopause, endogenous levels of estrogen and progesterone
fall sharply, but HT provides continued exposure to exogenous estrogen and/or progestin.
Although PRs are expressed in the human bladder (11, 12), little is known about their function –
or that of progesterone or progestins – in this organ. However, one study found that bladder
expression of PRs was significantly higher in premenopausal women and postmenopausal women
taking HT than in postmenopausal women not using HT (47). It is intriguing to postulate that
action of progesterone and progestins, mediated by PRs, may influence malignant potential of the
96

bladder. However, whether effects of parity are mediated by progesterone, and whether any
influence of progesterone during reproductive years and progestins in menopause involve
common biological processes, are questions awaiting mechanistic studies. The possibility that
use of progestin as HT may be associated with delayed bladder cancer detection, as has been
proposed for colorectal cancer (48), also warrants investigation.
Strengths of our primary analyses include use of two large, well-designed studies of distinct
data structure, with cases limited to UC. Strengths of the case-control study are population-
based design, matching of cases and controls on key characteristics, and enrollment of
participants from separate populations characterized by high versus low UC incidence. Major
strengths of the CTS are the prospective cohort of women followed since 1995, and detailed
information on hormonal and reproductive factors.
Each study has several limitations. CTS case numbers were small. Also, hormonal and
reproductive data are self-reported, and thus subject to misclassification; however, since data
were collected prospectively, any misclassification is likely to be non-differential, with any
resulting bias in the direction of no effect. Finally, the cohort consists of public school
professionals limiting generalizability of results. In the case-control study, an analysis of HT
constituents was not possible. In Shanghai, there were only 131 cases. In LA, history of any
pregnancy was measured, rather than history of term pregnancies; we therefore did not include
LA data in summary estimates of parity-bladder cancer associations, which were nonetheless
robust and supported by the pregnancy-bladder cancer associations from LA data. Recall bias is
not likely to have influenced results on parity or pregnancy, since these were not previously
regarded as bladder cancer risk factors. As with all analyses, it is possible that some results
could be due to chance, but consistency of associations with parity and E+P in all studies
addressing these factors is encouraging.
97

In conclusion, consistent results of epidemiology studies suggest that parous women
experience substantially reduced risk of bladder cancer. Protective effects of parity may arise
from first pregnancy, and are particularly evident among nonsmokers. Women who use E+P for
HT may also experience reduced risk. Research is now needed to understand the basis of the
parity-bladder cancer association, and a possible role of steroid hormones in bladder
carcinogenesis. Resulting insights may explain why rates among men greatly exceed those
among women; have implications for bladder cancer prevention strategies among non-smokers,
who comprise nearly half of incident cases (4); and inform efforts to develop targeted therapies
(49).
98

Chapter 3 References

American Cancer Society. Cancer Facts and Figures 2010. Atlanta: American Cancer Society;
2010. (1)

Balogh GA, Heulings R, Mailo DA, et al. Genomic signature induced by pregnancy in the human
breast. Int J Oncol 2006;28:399-410. (43)

Bernstein L, Ross RK. Cancer in Los Angeles County. Los Angeles (CA): University of Southern
California, 1991. (31)

Bernstein L, Allen M, Anton-Culver H, et al. High breast cancer incidence rates among
California teachers: results from the California Teachers Study (United States). Cancer Causes
Control 2002;13:625-35. (34)

Blakeman PJ, Hilton P, Bulmer JN. Oestrogen and progesterone receptor expression in the
female lower urinary tract, with reference to oestrogen status. BJU Int 2000;86: 32-8. (47)

Cantor KP. Lynch CF. Johnson D. Bladder cancer, parity, and age at first birth. Cancer Causes
Control 1992;3:57-62. (21)

Cantwell MM. Lacey JV Jr. Schairer C. Schatzkin A. Michaud DS. Reproductive factors,
exogenous hormone use and bladder cancer risk in a prospective study. Int J Cancer
2006;119:2398-401. (19)

Castelao JE, Yuan J-M, Skipper PL, et al. Gender- and smoking-related bladder cancer risk.
JNCI 2001;93: 538-45. (4)

Chaliha C and Stanton SL. Urological problems in pregnancy. BJU Int 2002;89:469-76. (44)

Cortessis VK, Yuan JM, Van Den Berg D, et al. Risk of urinary bladder cancer is associated with
8q24 variant rs9642880 [T] in multiple racial/ethnic groups: results from the Los Angeles-
Shanghai case-control study. Cancer Epidemiol Biomark Prevention 2010;19:3150-6. (30)

DeLellis Henderson K, Duan L, Sullivan-Halley J, et al. Menopausal hormone therapy and risk of
invasive colon cancer. The California Teachers Study. Am J Epidemiol 2010;171: 415-25. (36)

Dietrich K, Demidenko E, Schned A, et al. Parity, early menopause and the incidence of bladder
cancer in women: A case-control study and meta-analysis. Eur J Cancer 2010 Nov 9. [Epub
ahead of print]. (29)

Egger M, Davey Smith G, Schneider M, Minder C Bias in meta-analysis detected by a simple,
graphical test. BMJ 1997;315:629-34. (39)

Fernandez E. Gallus S. Bosetti C. Franceschi S. Negri E. La Vecchia C. Hormone replacement
therapy and cancer risk: a systematic analysis from a network of case-control studies. Int J Cancer
2003;105:408-12. (23)

99

Hartge P, Harvey EB, Linehan WM, et al. Unexplained excess risk of bladder cancer in men.
JNCI 82: 1990;1636-40. (5)

Higgins JPT, Thompson SG Quantifying heterogeneity in a meta-analysis. Stat Med
2002;21:1539-58. (37)

Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions
4.2.6. In: The Cochrane Library, Issue 4, 2006. Chichester, UK: John Wiley & Sons, Ltd. (38)
Huang AT, Kogevinas M, Silverman DT, et al. Bladder cancer and reproductive factors among
women in Spain. Cancer Causes Control 2009;20:1907-13. (27)

Imada S, Akaza H, Ami Y, Koiso K, Ideyama Y, Takenaka T. Promoting effects and mechanism
of androgen in bladder cancer carcinogenesis in male rats. Eur Urol 1997;31: 360-4. (13)

Johnson AM, O’Connell MJ, Messing EM, Reeder JE. Decreased bladder cancer growth in
parous mice. J Urol 2008;72: 470-3. (16)

Kirkali Z, Chan T, Manoharan M, et al. Bladder cancer: epidemiology, staging and grading, and
diagnosis. J Urology 2005;66 (Suppl 6A): 4-34. (6)

Kwong SL, Perkins CI, Morris CR, et al. Cancer in California: 1988-1999. Sacramento, CA:
Cancer Surveillance Section, California Department of Health Services, 2001: 336. (35)

La Vecchia C. Negri E. Franceschi S. Parazzini F. Long-term impact of reproductive factors on
cancer risk. Int J Cancer 1993;53:215-9. (25)

McGrath M. Michaud DS. De Vivo I. Hormonal and reproductive factors and the risk of bladder
cancer in women. Am J Epidemiol 2006;163:236-44. (17)

Miller AB, Barclay TH, Choi NW, et al. A study of cancer, parity and age at first pregnancy. J
Chronic Diseases 1980;33(10):595-605. (26)

Miyamoto H, Yang Z, Chen YT, et al. Promotion of bladder cancer development and progression
by androgen receptor signals. JNCI 2007;99: 558-568. (14)

Murta-Nascimento C, Schmitz-Drager BJ, Zeegers MP, et al. Epidemiology of urinary bladder
cancer: from tumor development to patient’s death. World J Urol 2007;25: 285-95. (2)

Okajima E, Hiramatsu T, Iriya K, Ijuin M, Matsushima S. Effects of sex hormones of
development of urinary bladder tumors induced by N-butyl-N-(4-hydroxybutyl)-nitrosamine.
Urol Res 1975;3:73-77. (15)

Olsson H, Bladstrom A, Ingvar C. Are smoking-associated cancers prevented or postponed in
women using hormone replacement therapy? Obstet Gynecol 2003;102: 565-70. (28)

Pearce CL, Chung K, Pike MC, Wu AH. Increased Ovarian Cancer Risk Associated with
Menopausal Estrogen Therapy is Reduced by Adding a Progestin. Cancer 2009;115:531-9. (46)

100

Pelucchi C. La Vecchia C. Negri E. Dal Maso L. Franceschi S. Smoking and other risk factors for
bladder cancer in women. Prev Med 2002;35:114-20. (22)

Persson I. Yuen J. Bergkvist L. Schairer C. Cancer incidence and mortality in women receiving
estrogen and estrogen-progestin replacement therapy--long-term follow-up of a Swedish cohort.
Int J Cancer 1996;67:327-32. (24)

Pike MC. Pearce CL. Wu AH. Prevention of cancers of the breast, endometrium and ovary.
Oncogene. 2004;23:6379-91. (41)

Prentice RL, Pettinger M, Beresford SA, et al. Colorectal cancer in relation to postmenopausal
estrogen and estrogen plus progestin in the Women's Health Initiative clinical trial and
observational study. Cancer Epidemiol Biomarkers Prev 2009;18:1531-7. (48)

Prizment AE. Anderson KE. Harlow BL. Folsom AR. Reproductive risk factors for incident
bladder cancer: Iowa Women's Health Study. Int J Cancer 2006;120:1093-8. (18)

Report on Carcinogens. Eleventh Edition; U.S. Department of Health and Human Services,
Public Health Service, National Toxicology Program, 2005. (3)

Rizk DEE, Raaschou T, Mason N, Berg B. Evidence of progesterone receptors in the muscosa of
the urinary bladder. Scand J Urol Nephrol 2001;35: 305-9. (12)

Rodriguez LV, Wang B, Shortliffe LMD. Structural changes in the bladder walls of pregnancy
and hormone-treated rats: correlation with bladder dynamics. BJU Int 2004;94:1366-72. (40)

Russo J, Balogh GA, Russo IH. Full-term Pregnancy Induces a Specific Genomic Signature in the
Human Breast Cancer. Epidemiol Biomarkers Prev 2008;17:51-66. (42)

Shan Y, Li L, Yan C, Yin F. Expression of estrogen receptors and progesterone receptors in
transitional cell carcinoma of bladder. Chin Med J (Engl) 1998;111: 191-2. (10)

Shen SS, Smith CL, Hsieh JT, et al. Expression of estrogen receptors- α and - β in bladder cancer
cell lines and human bladder tumor tissue. Cancer 2006: 106: 2610-6. (8)

Silverman DT, Alguacil J, Rothman N, et al. Does increased urination frequency protect against
bladder cancer? Int J Cancer 2008;123:1644-6. (45)

Teng J, Wang ZY, Jarrard DF, Bjorling DE. Roles of estrogen receptor α and β in modulating
urothelial cell proliferation. Endocrine-Related Cancer 2008;15: 351-64. (9)

Tincello DG, Taylor AH, Spurling SM, Bell SC. Receptor isoforms that mediate estrogen and
progestagen action in the lower urinary tract. J Urol 2009;181: 1474-82. (11)

Tripathi A. Folsom AR. Anderson KE. Iowa Women's Health Study. Risk factors for urinary
bladder carcinoma in postmenopausal women. The Iowa Women's Health Study. Cancer
2002;95:2316-23. (20)

101

WHO. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert
Committee. WHO Technical Report Series 854. Geneva: World Health Organization, 1995. (33)

Wu JT, Han BM, Yu SQ, Wang HP, Xia SJ. Androgen Receptor is a potential therapeutic target
for bladder cancer. Urology 2010;75:820-7. (49)

Yuan MY, Wang XL, Xiang YB, Gao YT, Ross RK, Yu MC. Preserved foods in relation to risk
of nasopharyngeal carcinoma in Shanghai, China. Int J Cancer 2000;85: 358-63. (32)

Zhuang YH, Blauer M, Tammela T, Tuohimaa P. Immunodetection of androgen receptor in
human urinary bladder cancer. Histopathology 1997;30: 556-62. (7)

102

Comprehensive References

Aiman J, Griffin JE, Gazak JM, et al. Androgen insensitivity as a cause of infertility in otherwise
normal men. New Engl J Med 1979;300:223-7.

Aiman J, Griffin JE The frequency of androgen receptor deficiency in infertile men. J Clin
Endocrinol Metab 1982;54:725-32.

Akimoto C, Ueda T, Inoue K, et al. Testis-specific protein on Y chromosome (TSPY) represses
the activity of the androgen receptor in androgen-dependent testicular germ-cell tumors. PNAS
2010;107:19891-6.

Altekruse SF, Kosary CL, Krapcho M, et al. (eds). SEER Cancer Statistics Review, 1975-2007,
National Cancer Institute. Bethesda, MD, http://seer.cancer.gov/csr/1975_2007/, based on
November 2009 SEER data submission, posted to the SEER web site, 2010.

American Cancer Society. Cancer Facts and Figures 2010. Atlanta: American Cancer Society;
2010.

Andrew SE, Goldberg YP, Hayden MR Rethinking genotype and phenotype correlations in
polyglutamine expansion disorders. Hum Mol Genet 1997;6:2005-10.

Asatiani K, von Eckardstein S, Simoni M, et al. CAG repeat length in the androgen receptor gene
affects the risk of male infertility. Int J Androl 2003;26:255-61.

Balogh GA, Heulings R, Mailo DA, et al. Genomic signature induced by pregnancy in the human
breast. Int J Oncol 2006;28:399-410.

Bernstein L, Ross RK. Cancer in Los Angeles County. Los Angeles (CA): University of Southern
California, 1991.

Bernstein L, Allen M, Anton-Culver H, et al. High breast cancer incidence rates among
California teachers: results from the California Teachers Study (United States). Cancer Causes
Control 2002;13:625-35.

Blakeman PJ, Hilton P, Bulmer JN. Oestrogen and progesterone receptor expression in the
female lower urinary tract, with reference to oestrogen status. BJU Int 2000;86: 32-8.

Brinkmann AO. Molecular basis of androgen insensitivity. Mol Cell Endocrin 2001;179: 105-9.

Canale D, Caglieresi C, Moschini C, et al. Androgen receptor polymorphism (CAG repeats) and
androgenicity. Clin Endocrinol (Oxf) 2005;63:356-61.

Cantor KP. Lynch CF. Johnson D. Bladder cancer, parity, and age at first birth. Cancer Causes
Control 1992;3:57-62.

103

Cantwell MM. Lacey JV Jr. Schairer C. Schatzkin A. Michaud DS. Reproductive factors,
exogenous hormone use and bladder cancer risk in a prospective study. Int J Cancer
2006;119:2398-401.

Casella R, Maduro MR, Lipshultz LI, et al. Significance of the polyglutamine tract polymorphism
in the androgen receptor. Urol 2001;58:651-6.

Casella R, Maduro MR, Misfud A, et al. Androgen receptor gene polyglutamine length is
associated with testicular histology in infertile patients. J Urol 2003;169:224-7.

Castelao JE, Yuan J-M, Skipper PL, et al. Gender- and smoking-related bladder cancer risk.
JNCI 2001;93: 538-45.

Chaliha C and Stanton SL. Urological problems in pregnancy. BJU Int 2002;89:469-76.

Chamberlain NL, Driver ED, Miesfeld RL The length and location of CAG trinucleotide repeats
in the androgen receptor N-terminal domain affect transactivation function. Nucleic Acids Res
1994;22:3181-6.

Cortessis VK. Epidemiologic insights into the occurrence and causes of testicular cancer. In:
Raghavan D, ed. American Cancer Society Atlas of Clinical Oncology, Germ Cell tumors.
London: BC Decker 2003,16-29.

Cortessis VK, Yuan JM, Van Den Berg D, et al. Risk of urinary bladder cancer is associated with
8q24 variant rs9642880 [T] in multiple racial/ethnic groups: results from the Los Angeles-
Shanghai case-control study. Cancer Epidemiol Biomark Prevention 2010;19:3150-6.

Dadze S, Wieland C, Jakubiczka S, et al. The size of the CAG repeat in exon 1 of the androgen
receptor gene shows no significant relationship to impaired spermatogenesis in an infertile
Caucasoid sample of German origin. Mol Hum Reprod 2000;6:207-14.

Dakouane-Giudicelli M, Legrand B, Bergere M, et al. Association between androgen receptor
gene CAG trinucleotide repeat length and testicular histology in older men. Fertil Steril
2006;86:873-7.

Davis-Dao CA, Tuazon ED, Sokol RZ, et al. Male infertility and variation in CAG repeat length
in the androgen receptor gene: a meta-analysis. J Clin Endocrin Metab 2007;92: 4319-26.

De Gendt K, Swinnen JV, Saunders PT, et al. A Sertoli cell-selective knockout of the androgen
receptor causes spermatogenic arrest in meiosis. PNAS 2004;101:1327-32.

DeLellis Henderson K, Duan L, Sullivan-Halley J, et al. Menopausal hormone therapy and risk of
invasive colon cancer. The California Teachers Study. Am J Epidemiol 2010;171: 415-25.

Denolet E, De Gendt K, Allemeersch J, et al. The effect of a sertoli cell-selective knockout of the
androgen receptor on testicular gene expression in prepubertal mice. Mol Endocrinol
2006;20:321-34.

104

Dhillon VS, Husain SA Cytogenetic and molecular analysis of the Y chromosome: absence of a
significant relationship between CAG repeat length in exon 1 of the androgen receptor gene and
infertility in Indian men. Int J Androl 2003;26:286-95.

Diekmann K-P, Pichlmeier U. Clinical epidemiology of testicular germ cell tumors. World J Urol
2004;22:2-14.

Dietrich K, Demidenko E, Schned A, et al. Parity, early menopause and the incidence of bladder
cancer in women: A case-control study and meta-analysis. Eur J Cancer 2010 Nov 9. [Epub
ahead of print].

Dowsing AT, Yong EL, Clark M, et al. Linkage between male infertility and trinucleotide repeat
expansion in the androgen-receptor gene. Lancet 1999;354:640-3.

Erasmuson T, Sin IL, Sin FY Absence of association of androgen receptor trinucleotide
expansion and poor semen quality. Int J Androl 2003;26:46-51.

Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis detected by a simple,
graphical test. BMJ 1997;315:629-34.

Ferlin A, Bartoloni L, Rizzo G, et al. Androgen receptor gene CAG and GGC repeat lengths in
idiopathic male infertility. Mol Hum Reprod 2004;10:417-21.

Fernandez E. Gallus S. Bosetti C. Franceschi S. Negri E. La Vecchia C. Hormone replacement
therapy and cancer risk: a systematic analysis from a network of case-control studies. Int J Cancer
2003;105:408-12.

Forman D, Oliver RT, Brett AR, et al. Familial testicular cancer: a report of the UK family
register, estimation of risk and an HLA class I sib-pair analysis. Br J Cancer 1992;65: 255-62.

Giwercman YL, Xu C, Arver S, et al. No association between the androgen receptor gene CAG
repeat and impaired sperm production in Swedish men. Clin Genet 1998;54:435-6.

Giwercman A, Lundin KB, Eberhard J et al. Linkage between androgen receptor gene CAG
trinucleotide repeat length and testicular germ call cancer histological type and clinical stage. Eur
J Cancer 2004;40: 2152-8.

Gao T, Marcelli M, McPhaul MJ. Transcriptional activation and transient expression of the
human androgen receptor. J Steroid Biochem Mol Biol 1996;59: 9-20.

Garolla A, Ferlin A, Vinanzi C, et al. Molecular analysis of the androgen receptor gene in
testicular cancer. Endocr-Related Cancer 2005;12: 645-55.

Garner MJ Turner MC, Ghadirian P, et al. Epidemiology of testicular cancer: an overview. Int J
Cancer 2005;116: 331-39.

Hadjkacem L, Hadj-Kacem H, Boulila A, et al. Androgen receptor gene CAG repeats length in
fertile and infertile Tunisian men. Ann Genet 2004;47:217-24.
105

Hartge P, Harvey EB, Linehan WM, et al. Unexplained excess risk of bladder cancer in men.
JNCI 82: 1990;1636-40.

Heimdal K, Olsson H, Tretli S, et al. Risk of testicular cancer in relatives of testicular cancer
patients. Br J Cancer 1996;73: 970-73.

Henderson BE, Ross RK, Pike MC, et al. Epidemiology of testis cancer. In: Skinner DG, ed.
Urological Cancer. New York: Grune and Stratton 1983,237-50.

Higgins JP, Thompson SG Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539-
58.

Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions
4.2.6. In: The Cochrane Library, Issue 4, 2006. Chichester, UK: John Wiley & Sons, Ltd.

Hiort O, Horter T, Schulze W, et al. Male infertility and increased risk of diseases in future
generations. Lancet 1999;354:1907-8.

Hiort O, Holterhus PM, Horter T, et al. Significance of mutations in the androgen receptor gene
in males with idiopathic infertility. J Clin Endocrinol Metab 2000;85:2810-5.

Horvath A, Korde L, Greene MH, et al. Functional phosphodiesterase 11A mutations may
modify the risk of familial and bilateral testicular germ cell tumors. Cancer Res 2009;69:5301-6.

Horwich A, Shipley J, Huddart R. Testicular germ cell cancer. Lancet 2006;367:754-65.

Hotaling JM, Walsh TJ Male infertility: a risk factor for testicular cancer. Nat Rev Urol
2009;6:550-6.

Huang AT, Kogevinas M, Silverman DT, et al. Bladder cancer and reproductive factors among
women in Spain. Cancer Causes Control 2009;20:1907-13.

Huyghe E, Matsuda T, Thonneau P. Increasing incidence of testicular cancer worldwide: a
review J Urol 2003;170:5-11.

Imada S, Akaza H, Ami Y, Koiso K, Ideyama Y, Takenaka T. Promoting effects and mechanism
of androgen in bladder cancer carcinogenesis in male rats. Eur Urol 1997;31: 360-4.

Irvine RA, Ma H, Yu MC, et al. Inhibition of p160-mediated coactivation with increasing
androgen receptor polyglutamine length. Hum Mol Genet 2000; 9: 267-74.

Jeong YM, Park JH, Choi JS, et al. The study of trinucleotide repeat length in the androgen
receptor gene of Korean infertile males. Korean J Genet 2004;26:269-275.

Johnson AM, O’Connell MJ, Messing EM, Reeder JE. Decreased bladder cancer growth in
parous mice. J Urol 2008;72: 470-3.

106

Kanetsky PA, Mitra N, Vardhanabhuti S, et al. Common variation in KITLG and at 5q31.3
predisposes to testicular germ cell cancer. Nat Genet 2009;41:811-15.

Katagiri Y, Neri QV, Takeuchi T, et al. Androgen receptor CAG polymorphism (Xq11-12) status
and human spermatogenesis: a prospective analysis of infertile males and their offspring
conceived by intracytoplasmic sperm injection. Int J Mol Med 2006;18:405-13.

Kelsey JL, Whittemore AS, Evans AS, et al. Methods in Observational Epidemiology, Second ed.
Oxford University Press, New York, 1996.

Kirkali Z, Chan T, Manoharan M, et al. Bladder cancer: epidemiology, staging and grading, and
diagnosis. J Urology 2005;66 (Suppl 6A): 4-34.

Komori S, Kasumi H, Kanazawa R, et al. CAG repeat length in the androgen receptor gene of
infertile Japanese males with oligozoospermia. Mol Hum Reprod 1999;5:14-6.

Kukuvitis A, Georgiou I, Bouba I, et al. Association of oestrogen receptor alpha polymorphisms
and androgen receptor CAG trinucleotide repeats with male infertility: a study in 109 Greek
infertile men. Int J Androl 2002;25:149-52.

Kwong SL, Perkins CI, Morris CR, et al. Cancer in California: 1988-1999. Sacramento, CA:
Cancer Surveillance Section, California Department of Health Services, 2001: 336. (35)

La Spada AR, Wilson EM, Lubahn DB, et al. Androgen receptor gene mutations in X-linked
spinal and bulbar muscular atrophy. Nature 1991;352:77-9.

La Vecchia C. Negri E. Franceschi S. Parazzini F. Long-term impact of reproductive factors on
cancer risk. Int J Cancer 1993;53:215-9.

Lavery R, Houghton JA, Nolan A, et al. CAG repeat length in an infertile male population of
Irish origin. Genetica 2005;123:295-302.

Legius E, Vanderschueren D, Spiessens C, et al. Association between CAG repeat number in the
androgen receptor and male infertility in a Belgian study. Clin Genet 1999;56:166-7.

Lund A, Tapanainen JS, Lahdetie J, et al. Long CAG repeats in the AR gene are not associated
with infertility in Finnish males. Acta Obstet Gynecol Scand 2003;82:162-6.

Madgar I, Green L, Kent-First M, et al. Genotyping of Israeli infertile men with idiopathic
oligozoospermia. Clin Genet 2002;62:203-7.

McGrath M. Michaud DS. De Vivo I. Hormonal and reproductive factors and the risk of bladder
cancer in women. Am J Epidemiol 2006;163:236-44.

McLachlan RI, O'Donnell L, Meachem SJ, et al. Identification of specific sites of hormonal
regulation in spermatogenesis in rats, monkeys, and man. Recent Progress Hormone Res
2002;57:149-79.

107

Mengual L, Oriola J, Ascaso C, et al. An increased CAG repeat length in the androgen receptor
gene in azoospermic ICSI candidates. J Androl 2003;24:279-84.

Merlet J, Racine C, Moreau E, et al. Male fetal germ cells are targets for androgens that
physiologically inhibit their proliferation. PNAS 2007;104:3615-20.

Mifsud A, Sim CK, Boettger-Tong H, et al. Trinucleotide (CAG) repeat polymorphisms in the
androgen receptor gene: molecular markers of risk for male infertility. Fertil Steril 2001;75:275-
81.

Milatiner D, Halle D, Huerta M, et al. Associations between androgen receptor CAG repeat
length and sperm morphology. Hum Reprod 2004;19:1426-30.

Miller AB, Barclay TH, Choi NW, et al. A study of cancer, parity and age at first pregnancy. J
Chronic Diseases 1980;33(10):595-605.

Miyamoto H, Yang Z, Chen YT, et al. Promotion of bladder cancer development and progression
by androgen receptor signals. JNCI 2007;99: 558-568.

Murta-Nascimento C, Schmitz-Drager BJ, Zeegers MP, et al. Epidemiology of urinary bladder
cancer: from tumor development to patient’s death. World J Urol 2007;25: 285-95.

Nathanson KL, Kanetsky PA, Hawes R, et al. The Y Deletion gr/gr and susceptibility to testicular
germ cell tumor. Am J Hum Genet 2005;77:1034-43.

Okajima E, Hiramatsu T, Iriya K, Ijuin M, Matsushima S. Effects of sex hormones of
development of urinary bladder tumors induced by N-butyl-N-(4-hydroxybutyl)-nitrosamine.
Urol Res 1975;3:73-77.

Olsson H, Bladstrom A, Ingvar C. Are smoking-associated cancers prevented or postponed in
women using hormone replacement therapy? Obstet Gynecol 2003;102: 565-70.

Oosterhuis JW, Looijenga LHJ. Testicular germ-cell tumors in a broader perspective. Nat Rev
Cancer 2005;5:210-22.

Pan H, Li YY, Li TC et al. Increased (CTG/CAG)(n) lengths in myotonic dystrophy type 1 and
Machado-Joseph disease genes in idiopathic azoospermia patients. Hum Reprod 2002;17:1578-
83.

Patrizio P, Leonard DG, Chen KL, et al. Larger trinucleotide repeat size in the androgen receptor
gene of infertile men with extremely severe oligozoospermia. J Androl 2001;22:444-8.

Pearce CL, Chung K, Pike MC, Wu AH. Increased Ovarian Cancer Risk Associated with
Menopausal Estrogen Therapy is Reduced by Adding a Progestin. Cancer 2009;115:531-9.

Pelucchi C. La Vecchia C. Negri E. Dal Maso L. Franceschi S. Smoking and other risk factors for
bladder cancer in women. Prev Med 2002;35:114-20.

108

Persson I. Yuen J. Bergkvist L. Schairer C. Cancer incidence and mortality in women receiving
estrogen and estrogen-progestin replacement therapy--long-term follow-up of a Swedish cohort.
Int J Cancer 1996;67:327-32.

Pike MC. Pearce CL. Wu AH. Prevention of cancers of the breast, endometrium and ovary.
Oncogene. 2004;23:6379-91.

Prentice RL, Pettinger M, Beresford SA, et al. Colorectal cancer in relation to postmenopausal
estrogen and estrogen plus progestin in the Women's Health Initiative clinical trial and
observational study. Cancer Epidemiol Biomarkers Prev 2009;18:1531-7.

Prizment AE. Anderson KE. Harlow BL. Folsom AR. Reproductive risk factors for incident
bladder cancer: Iowa Women's Health Study. Int J Cancer 2006;120:1093-8.

Quigley CA, De Bellis A, Marschke KB, et al. Androgen receptor defects: historical, clinical,
and molecular perspectives.[erratum appears in Endocr Rev 1995 16:546]. Endocr Rev
1995;16:271-321.

Rajender S, Rajani V, Gupta NJ, et al. No association of androgen receptor GGN repeat length
polymorphism with infertility in Indian men. J Androl 2006;27:785-9.

Rajpert-De Meyts E, Skakkabaek NE. Immunohistochemical identification of androgen receptors
in germ cell neoplasia. J Endrocrinol 1992;135:R1-4.

Rajpert-De Meyts E, Leffers H, Petersen JH, et al. CAG repeat length in androgen-receptor gene
and reproductive variables in fertile and infertile men. Lancet 2002;359:44-6.

Rajpert-De Meyts E, Leffers H, et al. Analysis of the polymorphic CAG repeat length in the
androgen receptor gene in patients with testicular germ cell cancer. Int J Cancer 2002;102: 201-
4.

Rajpert-De Meyts E. Development model for the pathogenesis of testicular carcinoma in situ:
genetic and environmental aspects. Hum Reprod Update 2006;12:303-23.

Rapley EA, Turnbull C, Amin Al Olama A, et al. A genome-wide association study of testicular
germ cell tumor. Nat Genet 2009;41:807-10.

Report on Carcinogens. Eleventh Edition; U.S. Department of Health and Human Services,
Public Health Service, National Toxicology Program, 2005.

Richiardi L, Pettersson A, Akre O. Genetic and environmental risk factors for testicular cancer.
Int J Androl 2007;30:230-40.

Rizk DEE, Raaschou T, Mason N, Berg B. Evidence of progesterone receptors in the muscosa of
the urinary bladder. Scand J Urol Nephrol 2001;35: 305-9.

Rodriguez LV, Wang B, Shortliffe LMD. Structural changes in the bladder walls of pregnancy
and hormone-treated rats: correlation with bladder dynamics. BJU Int 2004;94:1366-72.
109

Ruhayel Y, Lundin K, Giwercman Y, et al. Androgen receptor gene GGN and CAG
polymorphisms among severely oligozoospermic and azoospermic Swedish men. Hum Reprod
2004;19:2076-83.

Russo J, Balogh GA, Russo IH. Full-term Pregnancy Induces a Specific Genomic Signature in the
Human Breast Cancer. Epidemiol Biomarkers Prev 2008;17:51-66.

Sasagawa I, Suzuki Y, Ashida J, et al. CAG repeat length analysis and mutation screening of the
androgen receptor gene in Japanese men with idiopathic azoospermia. J Androl 2001;22:804-8.

Schnack TH, Poulsen G, Myrup C, et al. Familial coaggregation of cryptorchidism, hypospadias,
and testicular germ cell cancer: a nationwide cohort study. J Nat Cancer Instit 2009; 102:187-92.

Shan Y, Li L, Yan C, Yin F. Expression of estrogen receptors and progesterone receptors in
transitional cell carcinoma of bladder. Chin Med J (Engl) 1998;111: 191-2.

Shapiro E, Huang H, Masch RJ, et al. Immunolocalization of androgen receptor and estrogen
receptors alpha and beta in human fetal testis and epididymis. J Urol 2005;174:1695-8.

Shen SS, Smith CL, Hsieh JT, et al. Expression of estrogen receptors- α and - β in bladder cancer
cell lines and human bladder tumor tissue. Cancer 2006: 106: 2610-6.

Silverman DT, Alguacil J, Rothman N, et al. Does increased urination frequency protect against
bladder cancer? Int J Cancer 2008;123:1644-6.

Singh R, Deepa SR, Madhavi S, et al. Male infertility: no evidence of involvement of androgen
receptor gene among Indian men. J Andrology 2006;27:102-5.

Teng J, Wang ZY, Jarrard DF, Bjorling DE. Roles of estrogen receptor α and β in modulating
urothelial cell proliferation. Endocrine-Related Cancer 2008;15: 351-64.

Thangaraj K, Joshi MB, Reddy AG, et al. CAG repeat expansion in the androgen receptor gene
is not associated with male infertility in Indian populations. J Androl 2002;23:815-8.

Tincello DG, Taylor AH, Spurling SM, Bell SC. Receptor isoforms that mediate estrogen and
progestagen action in the lower urinary tract. J Urol 2009;181: 1474-82.

Travis LB, Beard C, Allan JM et al. Testicular cancer survivorship: research strategies and
recommendations. J Nat Cancer Instit 2010;102:1114-30.

Tripathi A. Folsom AR. Anderson KE. Iowa Women's Health Study. Risk factors for urinary
bladder carcinoma in postmenopausal women. The Iowa Women's Health Study. Cancer
2002;95:2316-23.

Tse JY, Liu VW, Yeung WS, et al. Molecular analysis of the androgen receptor gene in Hong
Kong Chinese infertile men. J Assist Reprod Genet 2003;20:227-33.

110

Tufan AC, Satiroglu-Tufan NL, Aydinuraz B, et al. No association of the CAG repeat length in
exon 1 of the androgen receptor gene with idiopathic infertility in Turkish men: implications and
literature review. Tohoku J Exp Med 2005;206:105-15.

Turnbull C, Rapley EA, Seal S et al. Variants near DMRT1, TERT and ATF7IP are associated
with testicular germ cell cancer. Nat Genet 2010;42:604-7.

Tut TG, Ghadessy FJ, Trifiro MA, et al. Long polyglutamine tracts in the androgen receptor are
associated with reduced trans-activation, impaired sperm production, and male infertility. J Clin
Endocrinol Metab 1997;82:3777-82.

Van Golde R, Van Houwelingen K, Kiemeney L, et al. Is increased CAG repeat length in the
androgen receptor gene a risk factor for male subfertility? J Urol 2002;167:621-623.

Vastermark A, Lundberg Giwercman Y, et al. Polymorphic variation in the androgen receptor
gene: association with risk of testicular germ cell cancer and metastatic disease. Eur J Cancer
2010; in press.

von Eckardstein S, Syska A, Gromoll J, et al. Inverse correlation between sperm concentration
and number of androgen receptor CAG repeats in normal men. J Clin Endocrinol Metab
2001;86:2585-90.

Wallerand H, Remy-Martin A, Chabannes E, et al. Relationship between expansion of the CAG
repeat in exon 1 of the androgen receptor gene and idiopathic male infertility. Fertil Steril
2001;76:769-74.

Wanderas EH, Fossa SD, Tretli S. Risk of a second germ cell cancer after treatment of a primary
germ cell cancer in 2201 Norwegian male patients. Eur J Cancer 1997;33: 244-52.

Wang W, John EM, Ingles SA. Androgen receptor and prostate-specific antigen gene
polymorphisms and breast cancer in African-American women. Cancer Epidemiol Biomarkers
Prev 2005;14:2990-4.

WHO Laboratory Manual for the Examination of Human Semen and Sperm-Cervical Mucus
Interaction, 4th ed. Cambridge University Press, Cambridge, 1999.

WHO. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert
Committee. WHO Technical Report Series 854. Geneva: World Health Organization, 1995.

Wilson JD Roosters B Reifenstein's syndrome and hormone resistance. New Engl J Med 1977;
297:386-7.

Wu JT, Han BM, Yu SQ, Wang HP, Xia SJ. Androgen Receptor is a potential therapeutic target
for bladder cancer. Urology 2010;75:820-7. (49)

Yoshida KI, Yano M, Chiba K,et al. CAG repeat length in the androgen receptor gene is
enhanced in patients with idiopathic azoospermia. Urol 1999; 54:1078-81.

111

Yuan MY, Wang XL, Xiang YB, Gao YT, Ross RK, Yu MC. Preserved foods in relation to risk
of nasopharyngeal carcinoma in Shanghai, China. Int J Cancer 2000;85: 358-63.

Zeegers MP, Kiemeney LA, Nieder AM, et al. How strong is the association between CAG and
GGN repeat length polymorphisms in the androgen receptor gene and prostate cancer risk?
Cancer Epidemiol Biomarkers Prev 2004;13:1765-71.

Zhou X, Kudo A, Kawakami H, et al. Immunohistochemical localization of androgen receptor in
mouse testicular germ cells during fetal and postnatal development. Anat Rec 1996;245:509-18.

Zhou Q, Nie R, Prins GS et al. Localization of androgen and estrogen receptors in adult male
mouse reproductive tract. J Androl 2002;23:870-81.

Zhuang YH, Blauer M, Tammela T, Tuohimaa P. Immunodetection of androgen receptor in
human urinary bladder cancer. Histopathology 1997;30: 556-62.

Abstract (if available)

Abstract One of the main goals of my doctoral training has been to gain experience working with a variety of types of epidemiologic data. The three projects that compose this dissertation have provided me with the opportunity to work with several different data structures: (1) published summary data, (2) population-based family data, (3) prospective cohort data, and (4) population-based case-control data.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Breast cancer in the multiethnic cohort study: Genetic (prolactin pathway genes) and environmental (hormone therapy) factors

PDF

Hormone therapy timing hypothesis and atherosclerosis

PDF

Risk factors of testicular germ cell tumor

PDF

The role of inflammation in non-Hodgkin lymphoma etiology

PDF

Genes and hormonal factors involved in the development or recurrence of breast cancer

PDF

Bilateral testicular germ cell tumors: distribution of histologic types

PDF

Antibiotic resistance in a large ICU cohort: 1995-2002

PDF

Factors that influence mammographic density: role of estrogen metabolism genes, biomarkers of inflammation, and lifestyle

PDF

Hormonal and genetic risk factors of endometrial cancer and trends in incidence and survival of adult acute lymphoblastic leukemia

PDF

Vision epidemiology and the impact of vision loss on vision-specific quality of life

PDF

Age related macular degeneration in Latinos: risk factors and impact on quality of life

PDF

The impact of visual field impairment on health related quality of life: a population-based study

PDF

The role of heritability and genetic variation in cancer and cancer survival

PDF

Body size and the risk of prostate cancer in the multiethnic cohort

PDF

The multiethnic nature of chronic disease: studies in the multiethnic cohort

PDF

Arm lymphedema in a multi-ethnic cohort of female breast cancer survivors

PDF

Chronic eye disease epidemiology in the multiethnic ophthalmology cohorts of California study

PDF

Assessing the impact of air pollution on adverse birth outcomes in a low resource setting

PDF

Characterizing the genetic and environmental contributions to ocular and central nervous system health

PDF

Disease risk estimation from case-control studies with sampling

Asset Metadata

Creator Davis-Dao, Carol Ann (author)

Core Title The role of steroid hormones in the etiology of urologic diseases

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Publication Date 02/07/2011

Defense Date 12/15/2010

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag bladder cancer,Hormones,male infertility,OAI-PMH Harvest,testicular germ cell tumors

Place Name California (states), China (countries), Los Angeles (counties), Shanghai (city or populated place)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Cortessis, Victoria K. (committee chair), Azen, Stanley Paul (committee member), Coetzee, Gerhard A. (committee member), McKean-Cowdin, Roberta (committee member), Siegmund, Kimberly D. (committee member)

Creator Email caroldav@med.usc.edu,lorac2201@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3644

Unique identifier UC1204671

Identifier etd-DavisDao-4256 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-427923 (legacy record id),usctheses-m3644 (legacy record id)

Legacy Identifier etd-DavisDao-4256.pdf

Dmrecord 427923

Document Type Dissertation

Rights Davis-Dao, Carol Ann

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu