Abstract
Purpose:
Investigators have reported inconsistent findings regarding associations between body mass index (BMI) and bladder cancer risk, and they have postulated that sex steroids mediate such associations. We assessed the impact of BMI on the relationship between bladder cancer risk and combinations of age at first childbirth, parity, and age at menopause, among Egyptian women.
Methods:
We used data from our multicenter case-control study of 419 cases and 786 controls in logistic regression models to estimate adjusted odds ratios (AORs) and 95% confidence intervals (CIs) of such associations.
Results:
Age >18y at first childbirth and parity <6 were significantly associated with bladder cancer risk, which was higher when both factors (AOR =2.31, 95% CI = 1.55–3.43) and age at menopause <45y (AOR = 3.51, 95% CI= 1.88–6.55) were present. Early menopause was associated with higher bladder cancer risk in obese (AOR = 2.90, 95% CI = 1.40–5.98) but not normal weight women (AOR = 0.98, 95% CI = 0.58–1.65; Pinteraction = 0.11), and the risk was greatest when both first childbirth at age >18y and parity ≤6 were present (AOR = 7.60, 95% CI = 1.84–31.35); although overweight and obesity were associated with significantly lower bladder cancer risk (AOR = 0.59, 95% CI = 0.43–0.81, and AOR = 0.26, 95% CI = 0.18–0.38, respectively).
Conclusion:
BMI appears to modify bladder cancer risk in Egyptian women after menopause by slightly enhancing the risk associated with low estrogen exposure among the obese only. Longitudinal studies of the BMI role in bladder malignancy in this distinctive population are required.
Introduction
Urinary bladder cancer is the 9th most common malignancy worldwide, and the disease disproportionately affects more men (~75%) than women [1]. Both animal models [2,3] and human epidemiological studies have explored the role of sex hormones in this disparity, presenting evidence of estrogens [4–7] and anti-androgens [8] as protective factors against this malignancy, particularly the urothelial type.
Investigators have used early age at menarche and/or at first childbirth, late menopause, and parity (number of babies delivered) as proxy indicators of high exposure to estrogens [4–7,9,10], and examine their associations with bladder cancer risk; they found inconsistent results. An association between bladder cancer incidence and age at first childbirth was found among Swedish [9] but not among American women [5], and bladder cancer mortality was reported to rise with increasing age at first childbirth among Taiwanese women [10]. Some investigators found lower bladder cancer risk among parous compared to nulliparous women in the US [7,11] and in Spain [12], but others reported no significant associations between parity and bladder cancer risk [7,10]. Similarly, some researchers detected associations between early age at menopause and increased bladder cancer risk [4–6], but not others [12].
On the other hand, increases in body mass index (BMI), used to determine the presence or absence of overweight and obesity, have been reported to be associated with several diseases, including some cancers [13]. Obesity is potentially associated with increased risks of bladder cancer progression and/or mortality [14], based on mixed reports of positive associations [15,16], while other reports have indicated no correlation [17,18]. The BMI and cancer risk relationship has been found to differ by population, sex, and ethnic origin [19]. Mechanisms postulated to underlie carcinogenesis [20], and more specifically, bladder cancer development in obese individuals [21], involve sex steroids that also impact reproductive outcomes among women.
In Egypt, urinary bladder carcinomas are diagnosed among women in their fifties and are highly prevalent [6,22], including both the urothelial (UC) and the squamous cell (SCC) type that is rare in western countries. In a preliminary analysis of data from a case-control study of sex differences in bladder cancer risk factors [6], we found significant associations between estrogen exposure proxies and bladder cancer risk. We conducted the present study of a larger sample of Egyptian women than the preliminary analysis [6] to 1) assess the relationships between bladder cancer risk and a gradient of cumulative estrogen exposure not previously examined, and 2) evaluate the effect of BMI on such associations.
Methods
We analyzed data from our multicenter case-control study conducted in Egypt between 2006 and 2014. The parent study was approved by the institutional review boards of the 3 cancer referral centers (the National Cancer Institute in Cairo, the Oncology Center at Minia University, and the South Egypt Cancer Institute in Assiut), Egypt’s Ministry of Health, the University of Maryland, Baltimore, and Georgetown University.
Study population
We previously described the study population in detail [6,22]. Briefly, cases were recruited from the above-mentioned 3 cancer referral centers; to be eligible, they had to have been diagnosed with presumed bladder cancer within 12 months and self-identified as capable of completing a 20-minute interview. No proxy interviews were conducted. The diagnosis of primary urinary bladder cancer was ascertained by either one of the study’s pathologists.
Healthy controls were randomly selected from the general population to frequency-match the cases by sex, 5-year age-group, governorate (province) of current residence, and urban/rural place of residence.
After cases and controls signed consent forms, trained interviewers administered the in-person questionnaire to collect data on sociodemographic characteristics; environmental exposure histories, including exposure to tobacco smoke; and medical histories, including history of schistosomiasis, weight, and height. Questions for women also asked about reproductive health history, including age at first childbirth, number of babies delivered, menopausal status, and age at menopause.
Variables
The outcome variable was bladder cancer status: all cases, SCC or UC case, or control. The independent variables included 1) age at first childbirth in years, 2) parity (number of babies delivered), and 3) age at menopause in years. Based on the distribution of these variables among the controls (median) and what is known about early versus late menopause, we coded categories as follows: age at first childbirth >18y=1 versus ≤18y=0; parity ≤6=1 versus >6=0; and age at menopause <45y=1 versus ≥45y=0. We used different combinations of these reproductive factors (estrogen exposure proxy variables) to generate two new categorical variables, Composites A and B. As detailed in Table 1, Composite A consisted of level 1 with age at first childbirth ≤18y and parity >6; level 2, with either one of the level 1 characteristics; and level 3, with age at first childbirth >18y and parity ≤6. Composite B consisted of level 1, with age at first childbirth ≤18y, parity >6, and age at menopause ≥45y; level 2, with at least two of the level 1 characteristics; level 3, with at least one of the level 1 characteristics; and level 4, with age at first childbirth >18y, parity ≤6, and age at menopause <45y. Level 1 represented the highest cumulative estrogen exposure for each composite variable and the reference category in the statistical analysis, while levels 3 and 4 represented the lowest cumulative estrogen exposure for Composite A and Composite B, respectively. We used the whole sample to analyze age at first childbirth, parity, and Composite A. For age at menopause and Composite B, we restricted the analyses to only those women who reported menopause.
Table 1.
Description of Composite Variables (and Levels) Reflecting Cumulative Estrogen Exposure among Egyptian Women, Based on Age (y) at First Childbirth, Parity (Number of Babies Delivered), and Age (y) at Menopause
| Variable name and level | Age (y) at first childbirth ≤18 = 0 >18 = 1 |
Parity >6 = 0 ≤6 = 1 |
Age (y) at menopause ≥45 = 0 <45 = 1 |
Cumulative estrogen exposure |
|---|---|---|---|---|
| Composite A (used for the entire study sample) | ||||
| Level 1 | 0 | 0 | NA | Highest* |
| 2 | 0 | 1 | NA | |
| 1 | 0 | NA | ||
| 3 | 1 | 1 | NA | Lowest |
| Composite B (used only for the subset who reported menopause) | ||||
| Level 1 | 0 | 0 | 0 | Highest* |
| 2 | 0 | 0 | 1 | |
| 0 | 1 | 0 | ||
| 1 | 0 | 0 | ||
| 3 | 1 | 1 | 0 | |
| 1 | 0 | 1 | ||
| 0 | 1 | 1 | ||
| 4 | 1 | 1 | 1 | Lowest |
The statistical analysis used the highest cumulative estrogen exposure category (Level 1 for each composite variable) as the reference for comparison to each of the other levels.
BMI was calculated for each participant from self-reported weight in kilograms divided by the square of self-reported height in meters. We used the conventional categories of underweight (BMI <18.5), normal (18.5 ≤ BMI ≤ 24.9), overweight (24.9 < BMI ≤ 29.9), and obese (BMI >29.9) to describe the study population. Because of the small number of underweight participants, we combined them with those of normal weight and used BMI as a trilevel categorical variable, with normal as the reference.
Other examined covariates included 1) marital status, regrouped as currently married versus all other categories (never married, widowed, separated, or divorced); 2) education (some versus none); 3) history of schistosomiasis (yes versus no); 4) environmental tobacco smoke (ETS) (exposure either in or outside the home versus none); and 5) the matching variables (age and residence location). Only 14 cases and 14 controls reported smoking cigarettes, waterpipes, or both; therefore, we restricted the analyses to women who never smoked.
Statistical analyses
We used t-test and Chi-square statistics for continuous and categorical variables, respectively, to compare cases to controls. We used unconditional logistic regression models of the associations between each variable and bladder cancer risk to estimate unadjusted (OR) and adjusted (AOR) odds ratios with 95% confidence intervals (CIs).
We tested any variable that was significantly associated with the outcome (p-value ≤ 0.05) for its potential effect modification or confounding of the association between cancer and each of the estrogen exposure proxies. Using a step-wise approach, we built multivariable unconditional logistic regression models including covariates that remained significantly associated with the outcome or modified the regression coefficient of the main effect by ≥10%.
We looked for effect modification by adding each main predictor variable, one covariate, and the interaction term for the two to the model one at a time and testing for its statistical significance. We also conducted analyses after stratification by the potential modifier.
The final models included significant covariates and the matching variables, age, and area of residence (North versus South and urban versus rural). We also conducted separate analyses of all cases, SCC cases, and UC cases. We used SAS version 9.4 software (SAS Institute, Cary, NC, USA) to conduct all statistical analyses.
Results
The study sample consisted of 433 cases (200 SCC, 187 UC, and 46 adenocarcinoma) and 800 controls. The mean ages for cases and controls were 56.1y and 54.4y, respectively (Table 2). Most participants were not educated and lived in rural areas in south Egypt. More cases than controls had a first child at age >18y (53% versus 42%), had ≤6 children (this study population’s median) (66% versus 63%), and reached menopause at age <45y (34% versus 27%) (Table 3). More controls than cases had high BMI (>24.9); thus, they were overweight or obese (~70% versus 49%). More cases (23%) than controls (11%) reported history of schistosomiasis.
Table 2.
Socio-demographic Variables among Egyptian Women Controls Compared to Urinary Bladder Cancer Cases
| Variable | Controls n = 800 n (%) |
Alla cases n = 433 n (%) |
p-value | SCCb cases n = 200 n (%) |
p-value | UCc cases n = 187 n (%) |
p-value |
|---|---|---|---|---|---|---|---|
| Age (SD) (y) | 54.4 (12.2) | 56.1 (11.2) | 0.02 | 53.2 (10.7) | 0.19 | 59.3 (10.7) | <0.0001 |
| Age-group | |||||||
| ≤45 | 209 (26.1) | 80 (18.5) | 0.01 | 46 (23.0) | 0.002 | 21 (11.2) | <0.0001 |
| >45 to ≤55 | 221 (27.6) | 136 (31.4) | 82 (41.0) | 43 (23.0) | |||
| >55 to ≤65 | 238 (29.8) | 130 (30.0) | 48 (24.0) | 69 (36.9) | |||
| >65 | 132 (16.5) | 87 (20.1) | 25 (12.0) | 54 (28.9) | |||
| Education | |||||||
| None | 593 (74.3) | 389 (89.8) | <0.0001 | 183 (91.5) | <0.0001 | 166 (88.8) | <0.0001 |
| Some | 205(25.7) | 44 (10.2) | 17 (8.5) | 21 (11.2) | |||
| Residence | |||||||
| North | 80 (10.0) | 57 (13.2) | 0.09 | 24 (12.0) | 0.4 | 26 (13.9) | 0.12 |
| South | 720 (90.0) | 376 (86.8) | 176 (88.0) | 161 (86.1) | |||
| Urban | 100 (12.5) | 55 (12.7) | 0.91 | 20 (10.0) | 0.3 | 31 (16.6) | 0.13 |
| Rural | 700 (87.5) | 378 (87.3) | 180 (90.0) | 156 (83.4) | |||
| Marital status | |||||||
| Never married | 10 (1.2) | 9 (2.1) | 0.003 | 7 (3.5) | 0.006 | 1 (0.5) | 0.004 |
| Married | 538 (67.3) | 252 (58.2) | 119 (59.5) | 102 (54.6) | |||
| Widowed | 241 (30.1) | 157 (36.2) | 66 (33.0) | 79 (42.2) | |||
| Divorced | 11 (1.4) | 15 (3.5) | 8 (4.0) | 5 (2.7) | |||
| Smoking | |||||||
| Never | 786 (98.2) | 419 (96.8) | 0.06 | 103 (96.5) | 0.14 | 182 (97.3) | 0.14 |
| Waterpipe | 11 (1.4) | 6 (1.4) | 4 (2.0) | 2 (1.1) | |||
| Cigarettes | 3 (0.4) | 7 (1.6) | 3 (1.5) | 3 (1.6) | |||
| Both | 0 | 1 (0.2) | 0 | 0 | |||
| ETSd | |||||||
| No | 328 (41.0) | 184 (42.5) | 0.6 | 76 (38.0) | 0.4 | 89 (47.6) | 0.1 |
| Yes | 472 (59.0) | 249 (57.5) | 124 (62.0) | 98 (52.4) |
All cases included SCC, UC, and adenocarcinoma;
squamous cell carcinoma;
urothelial cell carcinoma;
environmental tobacco smoke; because of missing data, the sum of the n in each column may not add to the total.
Boldface indicates statistical significance.
Table 3.
Health and Endogenous Estrogen Exposure Variables among Egyptian Women Controls Compared to Urinary Bladder Cancer Cases
| Variable | Controls n = 786 n (%) |
Alla cases n = 419 n (%) |
p-value | SCCb cases n = 193 n (%) |
p-value | UCc cases n = 182 n (%) |
p-value |
|---|---|---|---|---|---|---|---|
| Age at first childbirth | |||||||
| ≤18y | 463 (57.9) | 201 (46.4) | 0.0001 | 92 (46.0) | 0.002 | 86 (46.0) | 0.003 |
| >18y | 337 (42.1) | 232 (53.6) | 108 (54.0) | 101 (54.0) | |||
| Parity n (%) | |||||||
| >6 children | 293 (36.6) | 145 (33.5) | 0.2 | 65 (32.5) | 0.2 | 64 (34.2) | 0.5 |
| ≤6 children | 507 (63.4) | 288 (66.5) | 135 (67.5) | 123 (65.8) | |||
| Menopause | |||||||
| No | 249 (31.2) | 56 (12.9) | <0.0001 | 32 (16.0) | <0.0001 | 15 (8.0) | <0.0001 |
| Yes | 550 (68.8) | 377 (87.1) | 168 (84.0) | 172 (92.0) | |||
| Age at menopaused | |||||||
| ≥45y | 402 (73.1) | 246 (65.2) | 0.01 | 107 (63.7) | 0.01 | 113 (65.7) | 0.06 |
| <45y | 148 (26.9) | 131 (34.7) | 61 (36.3) | 59 (34.3) | |||
| BMI | |||||||
| Continuous | 28.7 (6.2) | 25.0 (5.2) | <0.0001 | 24.7 (4.7) | <0.0001 | 25.3 (5.8) | <0.0001 |
| <18.5 | 21 (2.8) | 27 (6.9) | <0.0001 | 14 (7.7) | <0.0001 | 12 (7.1) | <0.0001 |
| ≥18.5 to ≤24.9 | 202 (27.3) | 172 (44.0) | 78 (42.8) | 71 (42.3) | |||
| >24.9 to ≤29.9 | 235 (31.8) | 125 (32.0) | 62 (34.1) | 52 (30.9) | |||
| >29.9 | 281 (38.1) | 67 (17.1) | 28 (15.4) | 33 (19.7) | |||
| History of schistosomiasis | |||||||
| No | 661 (82.6) | 301 (69.5) | <0.0001 | 139 (69.5) | 0.0002 | 134 (71.7) | 0.0008 |
| Yes | 93 (11.6) | 100 (23.1) | 42 (21.0) | 41 (21.9) | |||
| Unknown | 46 (5.8) | 32 (7.4) | 19 (9.5) | 12 (6.4) | |||
| Composite Ae level | |||||||
| 1 | 206 (25.8) | 82 (18.9) | 0.005 | 39 (19.5) | 0.02 | 33 (17.7) | 0.05 |
| 2 | 344 (43.0) | 182 (42.0) | 79 (39.5) | 84 (44.9) | |||
| 3 | 250 (31.2) | 169 (39.1) | 83 (41.0) | 70 (37.4) | |||
| Composite Bde level | |||||||
| 1 | 122 (22.2) | 44 (11.7) | 0.0001 | 21 (12.5) | 0.0005 | 16 (9.3) | 0.001 |
| 2 | 214 (38.9) | 134 (35.5) | 54 (32.1) | 68 (39.5) | |||
| 3 | 166 (30.2) | 146 (38.7) | 66 (39.3) | 67 (40.0) | |||
| 4 | 48 (8.7) | 53 (14.1) | 27 (16.1) | 21 (12.2) |
All cases included SCC, UC, and adenocarcinoma;
squamous cell carcinoma;
urothelial cell carcinoma;
analysis among women who reported menopause;
Composite A and Composite B used different combinations of the three endogenous estrogen exposure variables (age at first birth, parity, and age at menopause), with level 1 representing the highest endogenous estrogen exposure for either variable, while levels 3 and 4 represented the lowest for Composites A and B, respectively (see Table 1 for details).
Boldface indicates statistical significance.
For the rest of the analyses we used 419 cases and 786 controls, all of whom never smoked. Table 4 depicts the unadjusted and adjusted associations between bladder cancer risk and different proxies of estrogen exposure. The odds of having bladder cancer was were significantly high for women aged >18y at first childbirth and those aged <45y at menopause, and high but not statistically significantly for those who had <6 children. The combination of either two (Composite A) or all three (Composite B) of these variables, which represented proxies for cumulative estrogen exposure, revealed significant positive trends (p < 0.05) for higher bladder cancer risk with lower cumulative estrogen exposure (Table 4).
Table 4.
Unadjusted and Adjusted Associations between Single and Composite Reproductive Health Variables and Urinary Bladder Cancer Risk among Egyptian Women
| Variable | Alla cases | SCCb cases | UCc cases | |||
|---|---|---|---|---|---|---|
| Unadjusted OR (95% CI)d | Adjustede OR (95% CI) | Unadjusted OR (95% CI) | Adjusted OR (95% CI) | Unadjusted OR (95% CI) | Adjusted OR (95% CI) | |
| Age at first childbirth | ||||||
| ≤18y | Ref | Ref | Ref | Ref | Ref | Ref |
| >18y | 1.56 (1.23–1.98) | 1.76 (1.33–2.33) | 1.57 (1.15–2.16) | 1.87 (1.29–2.71) | 1.60 (1.16–2.21) | 1.71 (1.18–2.49) |
| Parity | ||||||
| >6 children | Ref | Ref | Ref | Ref | Ref | Ref |
| ≤6 children | 1.23 (0.95–1.59) | 1.44 (1.07–1.94) | 1.18 (0.85–1.66) | 1.36 (0.91–2.03) | 1.17 (0.83–1.64) | 1.52 (1.02–2.28) |
| Age at menopausef | ||||||
| ≥45y | Ref | Ref | Ref | Ref | Ref | Ref |
| <45y | 1.44 (1.08–1.93) | 1.29 (0.92–1.81) | 1.58 (1.09–2.29) | 1.12 (0.71–1.78) | 1.39 (0.96–2.02) | 1.40 (0.91–2.15) |
| Composite Ae | ||||||
| level 1 | Ref | Ref | ref | Ref | Ref | Ref |
| 2 | 1.37 (0.99–1.88) | 1.56 (1.07–2.28) | 1.22 (0,80–1.87) | 1.37 (0.82–2.29) | 1.60 (1.02–2.50) | 1.92 (1.13–3.25) |
| 3 | 1.73 (1.25–2.40) | 2.31 (1.55–3.43) | 1.69 (1.10–2.59) | 2.25 (1.33–3.80) | 1.83 (1.15–2.91) | 2.51 (1.45–4.36) |
| p-value for trend | 0.001 | |||||
| Composite Bfg | ||||||
| level 1 | Ref | Ref | Ref | Ref | Ref | Ref |
| 2 | 1.74 (1.16–2.63) | 1.85 (1.14–3.01) | 1.44 (0.83–2.51) | 1.51 (0.77–2.97) | 2.60 (1.42–4.75) | 3.04 (1.47–6.29) |
| 3 | 2.40 (1.58–3.63) | 2.67 (1.62–4.38) | 2.19 (1.27–3.79) | 2.33 (1.10–4.59) | 3.17 (1.72–5.82) | 3.78 (1.80–7.94) |
| 4 | 3.24 (1.91–5.49) | 3.51 (1.88–6.55) | 3.38 (1.74–6.57) | 2.85 (1.25–6.50) | 3.68 (1.75–7.75) | 4.84 (1.97–11.89) |
| p-value for trend | 0.003 | |||||
All cases included SCC, UC, and adenocarcinoma;
squamous cell carcinoma;
urothelial cell carcinoma;
odds ratio (95% confidence interval);
adjusted for matching variables (age and residence location), BMI, history of schistosomiasis, education, and marital status;
analysis among women who reported having reached menopause; Composite A and Composite B used different combinations of the three endogenous estrogen exposure variables (age at first childbirth, parity, and age at menopause), with level 1 representing the highest endogenous estrogen exposure for either variable, while levels 3 and 4 represented the lowest for Composites A and B, respectively (see Table 1 for details).
Boldface indicates statistical significance.
BMI was inversely associated with bladder cancer risk. Compared to normal weight women, those overweight and obese women had lower odds of having bladder cancer (OR = 0.58, 95% CI = 0.43–0.78, and OR = 0.26, 95% CI = 0.19–0.36; respectively).
Among the other covariates assessed, schistosomiasis history (OR = 2.27, 95% CI = 1.64–3.13), education (OR = 0.32, 95% CI = 0.22– 0.46), and marital status (OR = 0.64, 95% CI = 0.50–0.82) were significantly associated with bladder cancer risk.
After adjustment for the matching variables (age and residence location), as well as the covariates listed above, including the categorical BMI, the positive trend in the associations between low cumulative estrogen exposure proxies and bladder cancer risk remained for all cases, as well as for the SCC and the UC types (Table 4). Indeed, the Composite B combination variable, level 4, with age at menopause <45y, age at first childbirth >18y, and parity ≤6, was significantly associated with higher bladder cancer risk than each factor separately (AOR = 3.51, 95% CI = 1.88–6.55) (Table 4). Analyses conducted separately for SCC and UC revealed similar patterns, except the risk for those with the lowest estrogen exposure was slightly higher for UC than SCC (AOR = 4.84, 95% CI = 1.97–11.89, and AOR = 2.85, 95% CI = 1.25–6.50; respectively) (Table 4). Furthermore, the significant unadjusted OR (95% CI) [1.44 (1.08–1.93)] for early menopause became statistically non-significant [1.29 (0.92–1.81)] after adjustment for covariates, specifically the BMI categories.
In the fully adjusted models of the whole study sample, the AORs (95% CIs) were 0.59 (0.43–0.81) for overweight and 0.26 (0.18–0.38) for obese: not different from the unadjusted ORs. None of the interaction terms was statistically significant when introduced into the model. We then conducted logistic regression analyses after stratification of the study sample by BMI level (Table 5). Bladder cancer risks associated with age at first childbirth and with parity were not significantly different among the normal weight, overweight, and obese BMI strata.
Table 5.
Adjusted Associations between Single and Composite Reproductive Health Variables and Urinary Bladder Cancer Risk among Egyptian Women after Stratification by Different Categories of Body Mass Index (BMI)
| Variable | Adjusteda Odds Ratio (95% confidence interval) | ||
|---|---|---|---|
| BMI ≤24.9 | 24.9< BMI ≤29.9 | BMI >29.9 | |
| Age at first childbirth | |||
| ≤18y | Ref | Ref | Ref |
| >18y | 2.01 (1.30–3.11) | 1.55 (0.95–2.52) | 1.87 (1.03–3.39) |
| Parity | |||
| >6 children | Ref | Ref | Ref |
| ≤6 children | 1.56 (0.99–2.45) | 1.16 (0.69–1.96) | 1.83 (0.93–3.58) |
| Age at menopauseb | |||
| ≥45y | Ref | Ref | Ref |
| <45y | 0.98 (0.58–1.65) | 1.19 (0.65–2.17) | 2.90 (1.40–5.98) |
| Age at menopause*BMI interaction term | |||
| P-value | 0.76 | 0.11 | |
| Composite Ac level | |||
| 1 | Ref | Ref | Ref |
| 2 | 1.44 (0.81–2.57) | 1.69 (0.88–3.24) | 1.62 (0.69–3.80) |
| 3 | 2.78 (1.50–5.14) | 1.79 (0.90–3.55) | 2.84 (1.20–6.75) |
| Composite Bbc level | |||
| 1 | Ref | Ref | Ref |
| 2 | 1.80 (0.86–3.77) | 1.88 (0.81–4.37) | 2.47 (0.75–8.13) |
| 3 | 2.14 (1.02–4.50) | 2.71 (1.11–6.60) | 4.49 (1.38–14.58) |
| 4 | 3.51 (1.32–9.35) | 2.85 (0.98–8.32) | 7.60 (1.84–31.35) |
Adjusted for matching variables (age and residence location), schistosomiasis history, education, and marital status;
analysis among women who reported having reached menopause;
Composite A and Composite B were generated using different combinations of the three endogenous estrogen exposure variables (age at first childbirth, parity, and age at menopause), with level 1 representing the highest endogenous estrogen exposure for either variable, while levels 3 and 4 represented the lowest for Composites A and B, respectively (see Table 1 for details).
Boldface indicates statistical significance.
However, menopause at age <45y was significantly associated with higher bladder cancer risk in obese (AOR = 2.90, 95% CI = 1.40–5.98), but not normal weight group (AOR = 0.98, 95% CI = 0.58–1.65; Pinteraction = 0.11), and the risk was greatest when both first childbirth at age >18y and parity ≤6 were present (AOR = 7.60, 95% CI = 1.84–31.35) (Table 5); although overweight and obesity were associated with significantly lower bladder cancer risk (AOR = 0.59, 95% CI =0.43–0.81, and AOR = 0.26, 95% CI = 0.18–0.38, respectively).
To further understand the contribution of the BMI to the post-menopausal risk of bladder cancer, we conducted regression analyses of the BMI categories restricted to menopaused women, before and after stratification by age at menopause. As in Table 6 and in the analyses of the whole sample, the higher the BMI, the lower the risk; however, the statistically significant lower risk (AOR = 0.58, 95% CI = 0.43–0.78) among the overweight subgroup of the whole sample increased and became borderline significant in the sample restricted to menopaused women (AOR = 0.71, 95% CI = 0.50–1.01) (Table 6). Furthermore, we noted significantly low AORs for both overweight and obese categories in the age at menopause ≥45y stratum, while we found no significant effect in the age at menopause <45y stratum (Table 6).
Table 6.
Associations between BMI Categories and Bladder Cancer Rsk among Menopaused Women before and after Sratification by Age at Menopause
| All menopaused women | Menopause at ≥45y | Menopause at <45y | |||||
|---|---|---|---|---|---|---|---|
| BMI | Case/Control n | Unadjusted aOR (95% CI) | bAOR (95% CI) | Case/Control n | AOR (95% CI) | Case/Control n | AOR (95% CI) |
| ≤24.9 | 167/169 | Ref | Ref | 118/118 | Ref | 49/51 | Ref |
| >24.9 to ≤29.9 | 106/148 | 0.72 (0.52–1.01) | 0.71 (0.50–1.01) | 70/108 | 0.64 (0.42–0.98) | 36/40 | 0.91 (0.47–1.74) |
| >29.9 | 55/195 | 0.28 (0.20–0.41) | 0.26 (0.17–0.40) | 31/148 | 0.19 (0.11–0.32) | 24/47 | 0.50 (0.24–1.03) |
Odds ratio and 95% confidence interval;
Adjusted for matching variables (age and residence location), schistosomiasis history, education, and marital status.
Discussion
Our current study is the first to examine the impact of cumulative endogenous estrogen exposures by BMI level on bladder cancer risk among postmenopausal women. We found significant dose-response associations between bladder cancer risk and indicators of cumulative endogenous estrogen exposure (age at first childbirth, parity, and age at menopause) in Egyptian women; the lower the exposure, the higher the risk. We also found BMI to be significantly inversely associated with bladder cancer risk independently of the other variables included in the regression models; however, the BMI effect appeared to be slightly attenuated among menopaused women (Table 6) and to modify the risk associated with early menopause. Indeed, we found greater odds of having bladder cancer for obese (BMI >29.9) (AOR = 2.90, 95% CI =1.40–5.98) than for normal weight (AOR = 0.98, 95% CI = 0.58–1.65) women with early menopause.
Our findings differ from the results of a dose-response meta-analysis, which suggested a nonlinear positive association between BMI and bladder cancer risk based on 14 prospective studies [23]. Although the meta-analysis authors explored biologic mechanisms contributing to the association between excess weight and cancer risk, including possible increased bioavailability of steroid hormones, they did not address sex differences and impact on risk from endogenous estrogen exposures. Our results also diverge from those of Cantiello et al. [18], who conducted a systematic review examining metabolic syndrome and each of its components separately (high blood pressure, obesity, and diabetes) in relation to bladder cancer risk, and reported equivocal findings; they also did not consider the impact of the assessed risk factors among women at different life stages. Further, a systematic review [14] based on 31 investigations that included diverse designs found a potential association of obesity and increased bladder cancer risk, as well as increased risk of recurrence and mortality; the authors acknowledged that sex disparities in bladder cancer may involve obesity-linked hormonal changes, considering estrogen production by adipose tissue [24] and evidence of estrogen receptors in bladder cancer tissues [25,26]. In postmenopausal women, adipose tissue is the main source of circulating estrogen [24]; and an increase in BMI was reported to be associated with a decrease in estradiol concentrations in premenopausal women but an increase in postmenopausal women [27]. Therefore, one would expect obesity to counteract the effect of early menopause on bladder cancer risk, which was not the case in our study. The effect of early menopause (OR = 1.44, 95% CI = 1.08–1.93) was attenuated after adjustment for covariates driven by BMI categories (AOR = 1.29, 95% CI = 0.92–1.81) (Table 4), but in the stratified analysis by BMI, the AORs were significantly higher (AOR= 2.90, 95% CI = 1.40–5.98) among obese than among normal weight women (AOR = 0.98, 95% CI = 0.58–1.65). It is possible that in postmenopausal women, sex steroids, including estrogen, are produced and metabolized in target tissues; they do not function as the circulating hormones, and hence, their involvement in the development of metabolic syndrome [28]. Our finding that obese women (BMI >29.9) with menopause at age <45y or with overall low cumulative estrogen exposure had higher odds of bladder cancer than those of normal weight further raises the question of the roles of different types of estrogen (estradiol, estrone, or estriol) and other sex steroids in bladder cancer risk.
Although we used BMI at the time of diagnosis, our findings of an inverse association between BMI and bladder cancer risk were consistent with findings from longitudinal studies of BMI and other cancers. In a prospective pooled analysis of two large Nurses’ Health Study cohorts, significant inverse associations were found between BMI at age 18y and during childhood, and risk of most subtypes of breast cancer [29]. A recent multicenter analysis of pooled individual-level data from 758,592 premenopausal women, aged 18 to 54 years, from 19 prospective cohorts of breast cancer, found increased adiposity to be associated with a reduced risk of premenopausal breast cancer, and the strongest inverse associations were observed for BMI in early adulthood [30].
The present study relied on self-reported reproductive history for the endogenous estrogen exposures and for the height and weight used to calculate BMI; therefore, potential for bias exists. Tendencies for self-reported height to be overestimated and weight underestimated compared to measured values would potentially lead to bias that is greater among overweight and obese participants than those of normal weight; such biases in our study would result in an underestimation of our finding.
Where potential for bias remains, we believe that non-differential versus differential recall error between cases and controls is more likely. The women in our study were unlikely to be biased by knowledge of previous research findings on possible associations of bladder cancer with the examined risk factors.
While assessing the impact of endogenous estrogen exposures on bladder cancer risk by BMI from a single point in time could be problematic because of the uncertain exposure window of relevance, our findings nevertheless suggest the need for further studies to determine preventive potential, whether related to BMI as a marker of hormonal activity suppressive of malignancy, better nutrition or immune function, or other protective factors in early life or at other points in the life course.
Nonetheless, our study addressed BMI and reproductive health variables in a population that is distinct from other previously studied populations; Egyptian women had their first babies at relatively early ages (median 18y) and for some as early as 12y; and the median number of born babies was 6. The cases were diagnosed with bladder cancer in their fifties, and nearly half were SCC and not the UC type reported predominantly in western countries. Our finding that obesity (BMI >29.9) enhanced the association between bladder cancer risk and low cumulative estrogen exposure in postmenopausal women with distinctive characteristics is novel and supports the need for longitudinal studies of BMI, reproductive factors, and cancer risk among different populations.
Acknowledgments:
The authors thank Sameera Ezzat, Tamer El Hifnawy, and Doa’a Salah, who supervised the recruitment of subjects for this study. Iman Gouda, Iman Loay, and Bhaskar Kallakury performed the expert pathology review of the cases. Nabiel N. Mikhail provided data management services.
Funding: A grant from the National Cancer Institute, National Institutes of Health (R01-CA115618 to CAL) supported this study.
Footnotes
Publisher's Disclaimer: Disclaimer: Coauthors BJW and IJ are scientists with the U.S. FDA Center for Food Safety and Applied Nutrition. The findings and conclusions of this study are those of the authors and do not necessarily represent the official position of the FDA or the United States.
Ethical approval
Written informed consent was obtained from all participants and the institutional review boards of the 3 cancer referral centers (the National Cancer Institute in Cairo, the Oncology Center at Minia University, and the South Egypt Cancer Institute in Assiut), Egypt’s Ministry of Health, the University of Maryland, Baltimore, and Georgetown University approved the protocols.
Conflict of interest
The authors declare that they have no conflict of interest.
References
- 1.Antoni S, Ferlay J, Soerjomataram I, Znaor A, Jemal A, Bray F (2017) Bladder cancer incidence and mortality: a global overview and recent trends. Eur Urol 71:96–108. 10.1016/j.eururo.2016.06.010. [DOI] [PubMed] [Google Scholar]
- 2.Miyamoto H, Yang Z, Chen YT, Ishiguro H, Uemura H, Kubota Y, Nagashima Y, Chang YJ, Hu YC, Tsai MY, Yeh S, Messing EM, Chang C (2007) Promotion of bladder cancer development and progression by androgen receptor signals. J Natl Cancer Inst 99:558–568. 10.1016/j.eururo.2016.06.010. [DOI] [PubMed] [Google Scholar]
- 3.Zhang Y (2013) Understanding the gender disparity in bladder cancer risk: the impact of sex hormones and liver on bladder susceptibility to carcinogens. J Environ Sci Health Part C 31:287–304. 10.1080/10590501.2013.844755. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.McGrath M, Michaud DS, De Vivo I (2006) Hormonal and reproductive factors and the risk of bladder cancer in women. Am J Epidemiol 63:236–244. 10.1093/aje/kwj028. [DOI] [PubMed] [Google Scholar]
- 5.Prizment AE, Anderson KE, Harlow BL, Folsom AR (2007) Reproductive risk factors for incident bladder cancer: Iowa Women’s Health Study. Int J Cancer 120:1093–1098. 10.1002/ijc.22418. [DOI] [PubMed] [Google Scholar]
- 6.Wolpert BJ, Amr S, Ezzat S, Saleh DA, Gouda I, Loay I, Hifnawy T, Mikhail NN, Abdel-Hamid M, Zhan M, Zheng Y-L, Squibb K, Abdel-Aziz MA, Zaghloul M, Khaled H, Loffredo CA (2010) Estrogen Exposure and Bladder Cancer Risk in Egyptian Women. Maturitas 67:353–357. 10.1016/j.maturitas.2010.07.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Daugherty SE, Lacey JV Jr, Pfeiffer RM, Park Y, Hoover RN, Silverman DT (2013) Reproductive factors and menopausal hormone therapy and bladder cancer risk in the NIH-AARP Diet and Health Study. Int J Cancer 133:462–472. 10.1002/ijc.28022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Morales EE, Grill S, Svatek RS, Kaushik D, Thompson IM Jr, Ankerst DP, Liss MA (2016) Finasteride Reduces Risk of Bladder Cancer in a Large Prospective Screening Study. Eur Urol 69:407–410. 10.1016/j.eururo.2015.08.029. [DOI] [PubMed] [Google Scholar]
- 9.Weibull CE, Eloranta S, Altman D, Johansson AL, Lambe M. (2013) Childbearing and the risk of bladder cancer: a nationwide population-based cohort study. Eur Urol 63:733–738. 10.1016/j.eururo.2013.01.005. [DOI] [PubMed] [Google Scholar]
- 10.Chiu HF, Chen BK, Yang C (2016) Parity, Age at First Birth, and Risk of Death from Bladder Cancer: A Population-Based Cohort Study in Taiwan. Int J Environ Res Public Health 13: pii: E1197. 10.3390/ijerph13121197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Davis-Dao CA, Henderson KD, Sullivan-Halley J, Ma H, West D, Xiang YB, Gago-Dominguez M, Stern MC, Castelao JE, Conti DV, Pike MC, Bernstein L, Cortessis VK (2011) Lower risk in parous women suggests that hormonal factors are important in bladder cancer etiology. Cancer Epidemiol Biomarkers Prev 20:1156–1170 DOI: 10.1158/1055-9965.EPI-11-0017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Huang AT, Kogevinas M, Silverman DT, Malats N, Rothman N, Tardon A, Serra C, Garcίa-Closas R, Carrato A, Cantor KP (2009) Bladder cancer and reproductive factors among women in Spain. Cancer Causes Control 20:1907–1913. 10.1007/s10552-009-9384-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.De Ridder J, Julian-Almarcegui C, Mullee A, Rinaldi S, Van Herck K, Vicente-Rodrίguez G, Huybrechts I (2016) Comparison of anthropometric measurements of adiposity in relation to cancer risk: a systematic review of prospective studies. Cancer Causes Control 27:291–300. 10.1007/s10552-015-0709-y. [DOI] [PubMed] [Google Scholar]
- 14.Noguchi JL, Liss MA, Parsons JK (2015) Obesity, physical activity and bladder cancer. Curr Urol Rep 16: 74 10.1007/s11934-015-0546-2. [DOI] [PubMed] [Google Scholar]
- 15.Qin Q, Xu X, Wang X, Zheng XY (2013) Obesity and Risk of Bladder Cancer: a Meta-analysis of Cohort Studies. Asian Pac J Cancer Prev 14:3117–3121. DOI : 10.7314/APJCP.2013.14.5.3117. [DOI] [PubMed] [Google Scholar]
- 16.Sun JW, Zhao LG, Yang Y, Ma X, Wang YY, Xiang YB (2015) Obesity and risk of bladder cancer: a dose-response meta-analysis of 15 cohort studies. PLoS One 10:e0119313 10.1371/journal.pone.0119313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Häggström C, Stocks T, Rapp K, Bjørge T, Lindkvist B, Concin H, Engeland A, Manjer J, Ulmer H, Selmer R, Tretli S, Hallmans G, Jonsson H, Stattin P (2011) Metabolic syndrome and risk of bladder cancer: prospective cohort study in the metabolic syndrome and cancer project (Me-Can). Int J Cancer 128:1890–1898. 10.1002/ijc.25521 [DOI] [PubMed] [Google Scholar]
- 18.Cantiello F, Cicione A, Salonia A, Autorino R, De Nunzio C, Briganti A, Gandaglia G, Dell’Oglio P, Capogrosso P, Damiano R (2015) Association between metabolic syndrome, obesity, diabetes mellitus and oncological outcomes of bladder cancer: a systematic review. Int J Urol 22:22–32. 10.1111/iju.12644. [DOI] [PubMed] [Google Scholar]
- 19.Renehan AG, Tuson M, Egger M, Heller RF, Zwahlen M (2008) Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. Lancet 371:569–578. 10.1016/S0140-6736(08)60269-X. [DOI] [PubMed] [Google Scholar]
- 20.De Pergola G, Silvestris F (2013) Obesity as a Major Risk Factor for Cancer. J Obes 2013:291546 10.1155/2013/291546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Dobruch J, Daneshmand S, Fisch M, Lotan Y, Noon AP, Resnick MJ, Shariat SF, Zlotta AR, Boorjian SA (2016) Gender and Bladder Cancer: A Collaborative Review of Etiology, Biology, and Outcomes. Eur Urol 69:300–310. 10.1016/j.eururo.2015.08.037. [DOI] [PubMed] [Google Scholar]
- 22.Zheng YL, Amr S, Saleh DA, Dash C, Ezzat S, Mikhail NN, Gouda I, Loay I, Hifnawy T, Abdel-Hamid M, Khaled H, Wolpert B, Abdel-Aziz MA, Loffredo CA (2012) Urinary bladder cancer risk factors in Egypt: a multicenter case-control study. Cancer Epidemiol Biomarkers Prev 21:537–546. DOI: 10.1158/1055-9965.EPI-11-0589. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhao L, Tian X, Duan X, Ye Y, Sun M, Huang J (2017) Association of body mass index with bladder cancer risk: a dose-response meta-analysis of prospective cohort studies. Oncotarget 8:33990–34000. doi: [ 10.18632/oncotarget.16722]. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kershaw EE, Flier JS (2004) Adipose Tissue as an Endocrine Organ. J Clin Endocrinol Metab 89:2548–2556. 10.1210/jc.2004-0395. [DOI] [PubMed] [Google Scholar]
- 25.Tan W, Boorjian S, Advani P, Farmer S, Lohse C, Cheville J, Kwon E, Leibovich B (2015) The Estrogen Pathway: Estrogen Receptor-a, Progesterone Receptor, and Estrogen Receptor-P Expression in Radical Cystectomy Urothelial Cell Carcinoma Specimens. Clin Genitourin Cancer 13:476–484. 10.1016/j.clgc.2015.04.001. [DOI] [PubMed] [Google Scholar]
- 26.Miyamoto H, Yao JL, Chaux A, Zheng Y, Hsu I, Izumi K, Chang C, Messing EM, Netto GJ, Yeh S (2012) Expression of androgen and oestrogen receptors and its prognostic significance in urothelial neoplasm of the urinary bladder. BJU Int 109:1716–1726. 10.1111/j.1464-410X.2011.10706.x. [DOI] [PubMed] [Google Scholar]
- 27.Randolph JF Jr, Sowers MF Bondarenko IV, Harlow SD, Luborsky JL, Little RJ (2004) Change in Estradiol and Follicle-Stimulating Hormone across the Early Menopausal Transition: Effects of Ethnicity and Age. J Clin Endocrinol Metab 89:1555–1561. 10.1210/jc.2003-031183. [DOI] [PubMed] [Google Scholar]
- 28.Simpson ER, Misso M, Hewitt KN, Hill RA, Boon WC, Jones ME, Kovacic A, Zhou J, Clyne CD (2006) Estrogen—the Good, the Bad, and the Unexpected. Endocr Rev 26:322–330. DOI: 10.1210/er.2004-0020. [DOI] [PubMed] [Google Scholar]
- 29.Warner ET, Hu R, Collins LC, Beck AH, Schnitt S, Rosner B, Eliassen AH, Michels KB, Willett WC, Tamimi RM (2016) Height and body size in childhood, adolescence, and young adulthood and breast cancer risk according to molecular subtype in the Nurses’ Health Studies. Cancer Prev Res 9:732–738. DOI: 10.1158/1940-6207.CAPR-16-0085 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.The Premenopausal Breast Cancer Collaborative Group (2018) Association of Body Mass Index and Age with Subsequent Breast Cancer Risk in Premenopausal Women. JAMA Oncol. Published online June 21. doi: 10.1001/jamaoncol.2018.1771. [DOI] [PMC free article] [PubMed] [Google Scholar]