Hormonal and Reproductive Factors and the Risk of Bladder Cancer in Women

Abstract

Gender and cigarette smoking are among the most consistent predictors of bladder cancer risk. After adjustment for known risk factors, an excess risk remains for males, suggesting that other factors may be responsible for the gender differences. Given limited data on hormonal or reproductive factors and bladder cancer risk, the authors examined these factors among women in the US Nurses' Health Study cohort. During 26 years of follow-up (1976–2002), 336 incident cases of bladder cancer were diagnosed. Cox proportional hazards models were used to estimate incidence rate ratios and 95% confidence intervals between hormonal and reproductive factors and bladder cancer risk. Postmenopausal women, compared with premenopausal women, were at increased risk (incidence rate ratio = 1.93, 95% confidence interval: 0.99, 3.78). For postmenopausal women, early age at menopause (≤45 years) compared with late age at menopause (≥50 years) was associated with a statistically significant increased risk of bladder cancer (incidence rate ratio = 1.63, 95% confidence interval: 1.20, 2.23). The association between age at menopause and bladder cancer risk was modified by cigarette smoking status (p for interaction = 0.01). The authors observed no significant associations of age at menarche, parity, age at first birth, and exogenous hormone use with bladder cancer risk. Findings suggest that menopausal status and age at menopause may play a role in modifying bladder cancer risk among women.

Bladder cancer is the fourth most common cancer in US men and the ninth leading cancer in US women (1). Over 63,000 people will be newly diagnosed with bladder cancer in 2005, and more than 13,000 people will die from the disease (1). Tobacco smoking is the most established modifiable risk factor for bladder cancer (2–5). Additional risk factors include occupational exposures to industrial chemicals (6–8) and schistosomiasis infection (9). Furthermore, it is well established that, worldwide, men experience a two- to fourfold higher risk of bladder cancer compared with women (10–14). Gender differences in cigarette smoking patterns, occupational exposures, and other differences in known risk factor distributions cannot explain the excess bladder cancer risk observed for males (12, 15). It is plausible that hormonal factors may account for part of the differences in bladder cancer rates between men and women.

Data from animal and human studies suggest that sex steroids have important physiologic effects on the lower urinary tract, causing symptomatic and functional changes in the bladder (15–18). Pregnancy and menopause result in significant and potentially permanent changes in sex steroid hormone levels and can induce significant alterations in the lower urinary tract, such as incontinence, urgency, and recurring urinary tract infections (18). Risk factors that influence steroid hormone levels, such as parity, exogenous hormone use, and age at menarche and menopause, have been evaluated in relation to cancer risk, especially with respect to hormone-related cancers (e.g. breast, endometrial, and ovarian) (19–23). To date, few studies have investigated the associations of hormonal and reproductive risk factors with risk of bladder cancer in women, and findings have been inconsistent (15, 24–27). We examined the relation between several reproductive and hormonal factors and the risk of bladder cancer in a large prospective study of women.

MATERIALS AND METHODS

Study population

The Nurses' Health Study is an ongoing cohort study established in 1976 when 121,700 female US registered nurses aged 30–55 years completed a self-administered questionnaire on their medical histories and baseline health-related exposures. Detailed information on individual characteristics and behaviors was obtained from the questionnaires at baseline and biennially thereafter. After exclusions for prior cancer (except nonmelanoma skin cancer), the analytic cohort consisted of 116,598 women representing 1,341,190 person-years of follow-up. Follow-up rates averaged 90 percent for each 2-year follow-up cycle during the first 26 years of the study. Vital status was ascertained through next of kin and the National Death Index; both methods identified an estimated 98 percent of deaths in the cohort (28). This investigation was approved by the institutional review board of the Brigham and Women's Hospital, Boston, Massachusetts.

Menopause, reproductive factors, and exogenous hormone use

Menopausal status and age at onset of menopause were first determined with the initial questionnaire in 1976 and were subsequently updated with each biennial questionnaire. The questionnaires asked whether the participant's menstrual periods had ceased permanently and, if so, at what age and for what reason (e.g., natural, surgical, radiation). If menopause was due to surgery, the participant was asked to report the number of ovaries removed. Women were defined as postmenopausal from the self-reported time of natural menopause and surgical menopause. Surgical menopause was defined as undergoing a bilateral oophorectomy. For each time period, women of indeterminable menopausal status were those who underwent hysterectomy without bilateral oophorectomy, whose data on the extent of pelvic surgery were incomplete, and whose permanent cessation of menses was due to radiation therapy. Women of indeterminable menopausal status were also considered menopausal when they reached the age at which natural menopause occurred in 90 percent of the Nurses' Health Study cohort (54 years of age for smokers and 56 years of age for nonsmokers). Analyses investigating the association between menopausal status and bladder cancer risk included all women. Analyses investigating age at menopause and bladder cancer were limited to postmenopausal women who had undergone a natural or surgical menopause; therefore, the analytic cohort for this analysis consisted of 84,330 women representing 633,177 person-years of follow-up. Self-reported age at onset of menopause and type of menopause have been validated in the Nurses' Health Study (29).

Use and duration of hormone replacement therapy (HRT) were first determined with the initial questionnaire in 1976, and information on type of HRT use was subsequently collected beginning in 1978. Duration of HRT use was calculated from information provided on each of the subsequent questionnaires. Each biennial questionnaire asked participants whether they were currently using HRT (within the last month) and the number of months that HRT had been used during the past 24 months since the last biennial questionnaire was completed. Women for whom data on HRT use for a given 2-year period were missing were assigned to a missing category for that time period. This analysis included all postmenopausal women and women of indeterminable menopausal status.

We inquired about pregnancy and number of livebirths in 1976 and subsequently on the biennial mailed questionnaires until 1984, when only 329 participants (0.27 percent) reported new births. Parity was recorded as the total number of livebirths. In addition, participants responded to questions about their age at the time of their first pregnancy lasting at least 6 months and their age at menarche. In 1976, women were asked to record the “intervals of oral contraceptive use starting from first use and continuing until the present time.” These data were updated every 2 years until 1982, when fewer than 500 women in the cohort indicated use of oral contraceptives. We classified women as ever or never users of oral contraceptives and computed their total duration of use in months.

Cigarette smoking history and other risk factors

Cigarette smoking status and history of cigarette smoking were obtained at baseline and in all of the subsequent questionnaires. On each questionnaire, current smokers reported the average number of cigarettes smoked per day. We categorized participant cigarette smoking history as “current,” “former,” or “never” smoker at each time interval. In addition, we computed the cumulative total number of pack-years of smoking among ever smokers by multiplying the average reported number of packs smoked per day by the number of years of smoking in each time period, summing over all previous time periods. Since cigarette smoking is a well-established risk factor for bladder cancer (2–5, 10), participants for whom information on smoking status or pack-years of smoking was missing were excluded from the analysis for that time period. Four cases were excluded because of incomplete cigarette smoking information.

Baseline height and current weight were reported in 1976. Body mass index in 1976 was estimated by dividing baseline weight in kilograms by baseline height in meters squared. Each biennial questionnaire has subsequently asked about current weight, and body mass index has been recalculated for each 2-year time period. In a validation substudy, the correlation between self-reported weight and weight measured by a trained technician was 0.96 (30).

Information on dietary intake began to be collected in 1980. The 1980 dietary questionnaire listed 61 items, and subsequent questionnaires were expanded to include approximately 130 items. For each item, participants were asked to report their average use during the preceding year. Nine prespecified responses were possible, ranging from never or almost never to six or more times per day. Dietary intake was calculated by multiplying the reported frequency by a given serving size for each food item. Cruciferous vegetable intake in 1980 was determined by summing the average serving sizes of broccoli and cabbage. The reproducibility and validity of the food frequency questionnaires for this cohort have been described previously (31, 32).

Identification of bladder cancer cases

On each biennial questionnaire, participants were asked to report any cancer, heart disease, or other medical conditions that had been diagnosed during the previous 2-year period. Whenever a participant (or next of kin for decedents) reported a diagnosis of bladder cancer, we asked for permission to obtain related medical records. If permission to obtain records was denied, we attempted to confirm the self-reported cancer with an additional letter or telephone call to the participant. If the primary cause of death listed on a death certificate was previously unreported bladder cancer, we contacted a family member (subject to state regulations) to obtain permission to retrieve medical records or at least to confirm the diagnosis of bladder cancer. Hospital records and pathology reports that were obtained were reviewed by trained physicians, blinded to exposure information. We were able to confirm approximately 95 percent of self-reported cases of bladder cancer; 85 percent (n = 286) of those were confirmed by medical record review. After exclusion of prior cancers (other than nonmelanoma skin cancer) and missing cigarette smoking information, 336 confirmed incident bladder cancer cases were diagnosed between 1976 and 2002; 85 of the cases were never smokers, 133 former smokers, and 118 current smokers.

Statistical analysis

Person-time of follow-up was calculated for each participant from the return date of the baseline questionnaire to the date of bladder cancer diagnosis, death from any cause, or the end of follow-up (June 30, 2002), whichever occurred first. Those who reported cancer or who died were excluded from subsequent follow-up. Incidence rates of bladder cancer were computed by dividing the number of incident cases by the number of person-years in each category of exposure. Time-varying Cox proportional hazards models, treating age in months and 2-year questionnaire period as the time scale, were used to estimate multivariate incidence rate ratios and to calculate 95 percent confidence intervals, after adjusting simultaneously for additional confounding variables. Categories for the exposure variables and confounding factors were selected for consistency and comparability with previous analyses in this cohort and with prior published literature.

On the basis of previous analyses in this cohort (33, 34), participants were categorized into five groups of body mass index using whole-number cutpoints including widely used definitions of overweight and obesity (35). The following covariates were included in the final multivariate models: body mass index (<21, 21–22.9, 23–24.9, 25–29.9, ≥30 kg/m²), cigarette smoking status (never, former, current), and pack-years of cigarette smoking (i.e., number of packs of cigarettes smoked per day multiplied by the number of years of smoking (a continuous variable)). Cigarette smoking is an established risk factor for bladder cancer (2–5, 10) and has established antiestrogenic effects (36–39); therefore, it was considered a confounder and potential effect modifier. In postmenopausal women, the primary source of endogenous estrogen is from adipose tissue through the aromatization of androgens; thus, obesity was considered a potential confounder. Other factors considered were cruciferous vegetable intake, alcohol and total fluid intake, geographic region, and coffee consumption. These factors were selected on the basis of biologic plausibility and findings from previously published literature (2, 5, 8, 25, 27, 40–49). None of these variables altered the associations of reproductive and hormonal factors with bladder cancer risk. Therefore, these factors were not included in the final models. Age at menarche was analyzed by using baseline values only. All of the other variables in the multivariate analyses were treated as time varying and were updated biennially. Analyses for cigarette smokers only were adjusted for cigarette smoking status (former, current), pack-years of smoking (continuous), and body mass index (<21, 21–22.9, 23–24.9, 25–29.9, ≥30 kg/m²).

Tests for multiplicative interaction were performed by examining stratum-specific results and formally by using likelihood ratio tests comparing the model having both the main effects and the interaction terms with the model having the main effects only. Tests for linear trend were conducted by using continuous values for the different exposure variables. All of the statistical procedures were performed with SAS version 8 software (SAS Institute, Inc., Cary, North Carolina). All p values are based on two-sided tests.

RESULTS

We examined the baseline characteristics of participants in the Nurses' Health Study cohort by their reported menopausal status in 1976 (table 1). At baseline, approximately 21 percent of participants were postmenopausal; 63.7 percent of postmenopausal women experienced a natural menopause. Mean age at onset of menopause for women with a natural menopause was 47.9 years (standard deviation, 4.2)). Compared with postmenopausal women, premenopausal women were younger and were more likely to have ever used oral contraceptives (table 1). In addition, premenopausal women were also slightly leaner, smoked fewer cigarettes, and were younger at first birth. No other meaningful differences were found regarding other reproductive and hormonal factors between pre- and postmenopausal women.

Postmenopausal status, compared with premenopausal status, was associated with an increased risk of bladder cancer after adjusting for age, body mass index, cigarette smoking status, and pack-years of smoking (incidence rate ratio (IRR) = 1.93, 95 percent confidence interval (CI): 0.99, 3.78) (table 2). Among women experiencing a natural or surgical menopause, an increased risk of bladder cancer was also observed (adjusted IRR = 2.02, 95 percent CI: 1.06, 3.85). The effect estimates for natural and surgical menopause were not meaningfully different (p = 0.60).

We examined the association of age at onset of menopause among women who experienced a natural or surgical menopause with risk of bladder cancer (table 2). Earlier age at menopause (age ≤45 years) compared with later age at menopause (age ≥50 years) was associated with a statistically significant increased risk of bladder cancer (IRR = 1.63, 95 percent CI: 1.20, 2.23; p for trend = 0.08). A similar association was observed after restricting our analysis to those women experiencing a natural menopause only (IRR = 1.75, 95 percent CI: 1.17, 2.60; p for trend = 0.08). The associations between age at natural and surgical menopause and bladder cancer risk were not modified by postmenopausal hormone use (p for interaction = 0.70; early vs. later age at menopause IRR = 1.69, 95 percent CI: 0.88, 3.25 for never users and IRR = 1.65, 95 percent CI: 1.08, 2.53 for ever users).

We observed no overall associations of HRT use and oral contraceptive use with risk of bladder cancer (table 3). In addition, we did not observe any significant associations between duration and type of HRT use. Compared with that for never users of HRT, the incidence rate ratio for current users of HRT was 0.94 (95 percent CI: 0.68, 1.29) and for former users was 1.26 (95 percent CI: 0.99, 1.83). Current users of estrogen and progestin, compared with never users of HRT, had a nonsignificant decreased risk of bladder cancer (IRR = 0.75, 95 percent CI: 0.44, 1.26). There were also no significant associations found between past or current HRT use and risk of bladder cancer among postmenopausal women experiencing either natural or surgical menopause (data not shown). Further adjustment for age at menopause and/or duration of HRT use for these women did not modify the associations. Compared with that for women who had never used oral contraceptives, the adjusted incidence rate ratio for women who had ever used oral contraceptives was 0.84 (95 percent CI: 0.65, 1.08). However, we found no association of increasing duration of oral contraceptive use with bladder cancer risk.

In addition, we examined the association between reproductive factors known to influence hormone levels—namely, parity, age at first birth, and age at menarche—and risk of bladder cancer. We found no associations between these reproductive factors and risk of bladder cancer in this cohort of women (table 4).

Because cigarette smoking is antiestrogenic and, compared with nonsmokers, smokers experience menopause at an earlier age, we investigated whether cigarette smoking modified the relation between menopausal status and age at menopause and bladder cancer risk (table 5). We did not observe a significant interaction between menopausal status, risk of bladder cancer, and cigarette smoking status (p for interaction = 0.49). For nonsmokers, the adjusted incidence rate ratio for postmenopausal women compared with premenopausal women was 1.87 (95 percent CI: 0.64, 5.47); among female smokers, the adjusted incidence rate ratio was 1.97 (95 percent CI: 0.84, 4.62) for postmenopausal women compared with premenopausal women. However, we did observe effect modification by cigarette smoking status for age at onset of natural or surgical menopause and bladder cancer risk (p for interaction = 0.01). For cigarette smokers only, we observed a significant association with earlier age at natural or surgical menopause and bladder cancer risk (IRR = 1.84, 95 percent CI: 1.28, 2.64; p for trend = 0.02) compared with later age at menopause. Likewise, a significant association was found with age at natural menopause and bladder cancer risk for cigarette smokers only (IRR = 1.98, 95 percent CI: 1.28, 3.07; p for interaction = 0.01). Associations of other reproductive and hormonal factors with bladder cancer risk were not modified by cigarette smoking status (data not shown).

DISCUSSION

In this prospective cohort study of women, we assessed the relation between several reproductive and hormonal factors and the risk of bladder cancer. We observed an increased risk of bladder cancer for postmenopausal women compared with premenopausal women. For postmenopausal women who experienced a natural or surgical menopause, we also observed a statistically significant increase in bladder cancer risk with earlier age at menopause compared with later age at menopause; this association was significantly modified by cigarette smoking status. We did not observe any significant associations of type and duration of HRT use, age at first birth, parity, age at menarche, and oral contraceptive use and duration with bladder cancer risk. Several lines of evidence suggest a role of sex steroid hormones in bladder cancer risk.

Epidemiologic studies investigating the relation between reproductive and hormonal factors and bladder cancer risk are limited and inconsistent (15, 24–27). To our knowledge, an Italian case-control study with 110 cases and 298 controls is the only published study investigating menopausal status and bladder cancer risk, and the authors observed no significant associations with menopausal status and age at menopause; however, there were only nine premenopausal cases (25). Two small cohort studies (58 cases and 112 cases) reported no association between HRT use and bladder cancer risk (26, 27); in contrast, in a case-control study (110 cases), ever users of HRT had an elevated risk of bladder cancer, and a significant trend was observed for duration of use (25). Detection and/or recall bias may explain the strong, positive finding in the case-control study.

Similar to our findings, other studies observed no significant associations between age at menarche (25, 27) or oral contraceptive use (25) and bladder cancer risk. Two studies (15, 24) found a decreased bladder cancer risk for parous compared with nulliparous women, but two others (19, 21) did not observe such an association. With one exception (15), age at first birth was not associated with bladder cancer risk (25, 27). Many of these studies were limited by a small number of bladder cancer cases, were missing detailed information on cigarette smoking, and/or collected risk factor information after diagnosis.

We observed a statistically significant interaction between age at menopause, risk of bladder cancer, and cigarette smoking status. Cigarette smoking is the most well-established risk factor for bladder cancer (2–5, 10) and has been associated with an antiestrogenic effect on circulating estrogen levels and earlier age at natural menopause (36–39). On average, smokers experience menopause 1.0–1.5 years earlier than nonsmokers do (39). Cigarette smoking may also affect the absorption and metabolism of hormones used in postmenopausal hormone therapy (39). Effect modification by cigarette smoking status may reflect true underlying differences in the association between age at menopause and bladder cancer, or it may be the result of chance or represent residual confounding by cigarette smoking.

Findings from animal studies support a direct role for sex hormones in the etiology of bladder cancer because estrogens have been shown to inhibit and androgens shown to increase the growth and development of bladder cancer (50–55). A human bladder tumor cell line, R198, grew well in the presence of androgens, but not in the presence of estrogens (52). In male rats, the incidence of bladder cancer induced by N-butyl-N-(4-hydroxybutyl)nitrosamine (BBN) was increased in animals injected with testosterone compared with animals injected with estriol or estradiol (55). In a similar experiment, diethylstilboestrol inhibited bladder carcinogenesis induced by N-butyl-N-(4-hydroxybutyl)nitrosamine in male and female Wistar rats, whereas testosterone stimulated the growth of bladder tumors (53).

In postmenopausal women, decreases in circulating levels of estrogen have been associated with bladder dysfunction including incontinence, urgency, and recurrent urinary tract infections (17, 18), and administration of exogenous estrogen effectively reduces many of the symptoms associated with an overactive bladder (56, 57). In a number of epidemiologic studies, frequent and recurring urinary tract infections have been strongly associated with elevated bladder cancer risk (58, 59). Although speculative, it is conceivable that women who experience early menopause are at an increased risk of bladder cancer because they have recurrent urinary tract infections and concurrent inflammation starting at an earlier age, placing them at higher risk of subsequent bladder cancer. Because inflammation and cigarette smoking are likely to act synergistically to increase risk, this hypothesis would also explain our observed interaction with cigarette smoking. Furthermore, this proposed mechanism could explain why we did not observe associations with other reproductive factors, which are not linked to urinary tract infections.

The strengths of our study include a prospective design, a large number of cases for this cancer, high follow-up rates, and detailed information on potential risk factors for bladder cancer. Information was available on cigarette smoking habits from each biennial questionnaire, enabling tight control for the confounding effect of cigarette smoking. We adjusted for cigarette smoking status (never, former, and current) and used a continuous term for pack-years of smoking to minimize confounding by smoking. Self-reported age at onset of menopause may be subject to error; however, we excluded from these analyses all women whose age at onset of menopause was indeterminable; furthermore, in our cohort, age at onset of menopause has been observed to be highly accurate for women experiencing a natural menopause and women in our cohort who had a bilateral oophorectomy (29). Among those women who received a bilateral oophorectomy, age at menopause was reported within 1 year by 95 percent of the participants on the two subsequent questionnaires, which also asked the woman's age at menopause. Similarly, age at natural menopause was reported within 1 year by 82 percent of the women on the two follow-up questionnaires.

Although this study is the largest known prospective study to date to examine hormonal and reproductive risk factors in relation to bladder cancer risk, we had limited power to examine different HRT formulations among current users. Therefore, we cannot exclude the possibility that the lack of association in some analyses was due to small case numbers. In addition, we cannot completely rule out the possibility that residual confounding by smoking explains our positive findings for menopausal status and age at menopause. Finally, because this study is composed predominately of Caucasian women and the incidence of bladder cancer varies by ethnicity (60), generalizability to non-White women should be performed with caution.

To our knowledge, this is the largest prospective cohort study to investigate the associations of reproductive and hormonal risk factors with bladder cancer risk in women. We observed an increased risk of bladder cancer for postmenopausal women. For these women, we observed a statistically significant increased risk of bladder cancer with earlier age at onset of menopause compared with later age at onset of menopause. It is possible that differences in estrogen and androgen levels between the sexes may explain some of the excess risk observed in males. Additional studies should further investigate the biologic effects of sex hormones in the bladder and the role of hormones in bladder carcinogenesis.

This work is supported by National Institutes of Health grants T32 CA 09001-28 (M. M.), CA55075, and CA87969; and a grant from the American Cancer Society (RSG-00-061-04-CCE (I. D.)).

The authors thank Drs. Graham Colditz, Walter Willett, and Edward Giovannucci for their insightful comments.

Conflict of interest: none declared.

References