Prospective study of body mass index, height, physical activity and incidence of bladder cancer in US men and women

Authors

  • Crystal N. Holick,

    Corresponding author
    1. Department of Nutrition, Harvard School of Public Health, Boston, MA
    • Department of Nutrition, Harvard School of Public Health, 677 Huntington Avenue, Boston, MA 02115, USA
    Search for more papers by this author
  • Edward L. Giovannucci,

    1. Department of Nutrition, Harvard School of Public Health, Boston, MA
    2. Department of Epidemiology, Harvard School of Public Health, Boston, MA
    3. Channing Laboratory, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA
    Search for more papers by this author
  • Meir J. Stampfer,

    1. Department of Nutrition, Harvard School of Public Health, Boston, MA
    2. Department of Epidemiology, Harvard School of Public Health, Boston, MA
    3. Channing Laboratory, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA
    Search for more papers by this author
  • Dominique S. Michaud

    1. Department of Epidemiology, Harvard School of Public Health, Boston, MA
    2. Channing Laboratory, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA
    Search for more papers by this author

Abstract

We evaluated prospectively the association between body mass index (BMI), height, recreational physical activity and the risk of bladder cancer among US adults. Data were used from 2 ongoing cohorts, the Health Professionals Follow-up Study and the Nurses' Health Study, with 3,542,012 years of follow-up and 866 incident bladder cancer cases (men = 507; women = 359) for the anthropometric analysis and 1,890,476 years of follow-up and 706 incident bladder cancer cases (men = 502; women = 204) for the physical activity analysis. Cox proportional hazard models were used to estimate incidence rate ratios (RR) and 95% confidence intervals (CI) between BMI, height, physical activity and bladder cancer risk adjusting for age, pack-years of cigarette smoking and current smoking. Estimates from each cohort were pooled using a random-effects model. We observed no association between baseline BMI and bladder cancer risk, even when we compared a BMI of ≥30 kg/m2 to a BMI of 18–22.9 kg/m2 [pooled multivariate (MV) RR, 1.16; 95% CI: 0.89–1.52]. A weak, but statistically significant, association was observed for the same comparison after excluding bladder cancer cases diagnosed within the first 4 years of follow-up (pooled MV RR, 1.33; 95% CI: 1.01–1.76). Height was not related to bladder cancer risk (pooled MV RR, 0.82; 95% CI: 0.65–1.03, top vs. bottom quintile). Total recreational physical activity also was not associated with the risk of bladder cancer (pooled MV RR, 0.97; 95% CI: 0.77–1.24, top vs. bottom quintile). Our findings do not support a role for BMI, height or physical activity in bladder carcinogenesis. © 2006 Wiley-Liss, Inc.

An estimated 65% of adults in the US are currently overweight [a body mass index (BMI) between 25.0 and 29.9 kg/m2] or obese (BMI of 30.0 kg/m2 or greater).1 Overweight and obesity are associated with heart disease, stroke, diabetes, altered mental status, respiratory problems and various digestive and musculoskeletal disorders.2 There is growing evidence that obesity is associated with overall cancer and several site-specific cancer sites including breast (in postmenopausal women), colon, esophagus, kidney and endometrium2; however, the relation between obesity and the risk of bladder cancer remains speculative.

Data from studies examining weight or BMI,3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13 height,14, 15 physical activity8, 9, 16 and the subsequent risk of bladder cancer are inconsistent. The lack of consistency may be attributable to uncontrolled or residual confounding by cigarette smoking,3, 6, 7, 8, 14 potential recall or selection bias in case-control studies,4, 10, 12, 14 measurement error in the physical activity assessment7, 9 or small numbers of bladder cancer cases.3, 6, 8, 9, 11, 16 Furthermore, some studies that have investigated the association of these factors with bladder cancer risk have focused on men only7, 12, 16 or have included few women.3, 8, 11, 13, 14

Given the limitations of previous studies evaluating the role of anthropometric characteristics, physical activity and the risk of bladder cancer, including limited data on women, we conducted a prospective investigation in 2 large cohort studies of men and women.

Material and methods

Study populations

The Nurses' Health Study (NHS) was initiated in 1976 when 121,700 registered US female nurses, aged 30–55 years, returned a mailed questionnaire that assessed information on lifestyle factors, medical and smoking histories. Similarly, the Health Professionals Follow-Up Study (HPFS) is a cohort of 51,529 US male dentists, optometrists, osteopaths, podiatrists, pharmacists and veterinarians who were 40–75 years of age at enrollment in 1986. Follow-up questionnaires are mailed biennially to both cohorts to update information on lifestyle factors and newly diagnosed medical conditions. The follow-up rate for both cohorts for incidence of cancer was greater than 95% of the total possible person-years. Food frequency questionnaires (FFQs) were initially collected in 1980 for NHS and 1986 for HPFS and diet was generally updated every 4 years.

Assessment of anthropometric variables

Women participating in the NHS reported their height and current weight at baseline in 1976. An estimate of weight at 18 years old was collected in 1980. Similarly, in the HPFS at baseline in 1986, men reported their height, current weight and estimated weight at 21 years old. Current weight was reported every 2 years on subsequent questionnaires in both cohorts. To best represent total adiposity, the BMI (weight in kilograms divided by square of height in meters) was calculated from weight and height. The validity of self-reported weight for both men and women has been reported previously.17

Assessment of recreational physical activity

In 1986, participants from both cohorts reported the average time per week spent doing each of 8 moderate and vigorous recreational activities over the previous year: walking or hiking outdoors; jogging (<10 min/mile); running (≥10 min/mile); bicycling (including stationary machine); lap swimming; tennis, squash or racquetball and other aerobic exercise (calisthenics, aerobics, rowing machine). Ten prespecified responses were possible ranging from 0 to 11 or more hours per week. Participants also reported their usual walking pace (easy or <2 mph; average or 2–2.9 mph; brisk or 3–3.9 mph and very brisk or ≥4 mph) and the number of flights of stairs climbed daily. Calculations did not include household activities or occupational physical activity.

The reported time spent at each activity per week was multiplied by its typical energy expenditure requirements to derive an hours per week total recreational physical activity score expressed in metabolic equivalent tasks (METs).18 One MET-hr is the amount of energy expended when sitting quietly for 1 hr and is equivalent to 3.5 ml of oxygen uptake per kilogram of body weight per minute for a 154 pound adult. Body weight was excluded from the calculation of energy expenditure from physical activity to avoid confounding the expenditure variable by body weight. In accordance with the Centers for Disease Control and American College of Sports Medicine guidelines,18 total recreational physical activity was further classified as moderate-intensity (activities that burn ≤6 METs; includes walking or hiking outdoors and stair climbing) and vigorous-intensity (activities that burn >6 METs; includes all other activities) recreational physical activity. If a participant provided a response to at least one of the activities, the activities without responses were assumed to be 0. The validity of the physical activity assessment has been reported previously.19, 20

Smoking history and other risk factors

Information on smoking was obtained at baseline: past smokers reported when they last smoked and current smokers reported intensity of smoking. Current smoking status and intensity of smoking was updated biennially in subsequent questionnaires in both cohorts. We estimated pack-years (the equivalent of smoking 20 cigarettes a day for 1 year) by multiplying the number of packs smoked per day by the number of years of smoking; categorized as <10, 10–24, 25–44, 45–64 and ≥65 pack-years.

At baseline, participants in both cohorts reported their state of residence in the US (categorized as West, Midwest, South and Northeast) and total fluid intake (quintiles; from milk, coffee, juice, soda, alcohol, water, etc, but for NHS not including water as it was not assessed in the 1980 FFQ). Women were additionally asked about reproductive factors (status and age at menopause: premenopausal; postmenopausal, age <45; postmenopausal, age 45–49; postmenopausal, age 50–55 and postmenopausal, age >55).

Case ascertainment

On each biennial questionnaire, participants were asked whether they had been diagnosed with any cancer, heart disease or other medical conditions during the previous 2 years. When permission was received from the cases (or next of kin for decedents), medical records and pathology reports were obtained from hospitals and reviewed by study investigators, blinded to questionnaire exposure information. Nonrespondents were telephoned in an attempt to confirm the initial cancer report and date of diagnosis. Bladder cancer was confirmed by review of medical records in 87% of the cases. When we were unable to obtain medical records, we attempted to corroborate diagnoses of bladder cancer with additional information from the participant, next of kin or by death certificate. We only included cases (based on TNM classification system; stages CIS, Ta-T4) for which a medical record or other confirmation of the cancer was obtained. Vital status was ascertained through next of kin and the National Death Index; both methods identify at least 98% of deaths in the cohorts.21

In the NHS cohort, we identified 359 newly diagnosed bladder cancer cases between 1976 and 2002 among women eligible for analysis, with available data on anthropometric variables; 204 cases had provided data on physical activity in 1986. In the HPFS cohort, we identified 507 incident bladder cancer cases among men between 1986 and 2002 with data on anthropometric variables and 502 with data on physical activity.

Statistical analysis

Person-time of follow-up was calculated from the date for return of the baseline questionnaire until the date of bladder cancer diagnosis, date of death from any cause or the end of follow-up (December 31, 2002), whichever came first.

After excluding participants who reported a history of cancer other than nonmelanoma skin cancer or with missing information on weight data at baseline (1976 in NHS; 1986 in HPFS), the cohorts for analyses using anthropometric measurements included 47,914 men followed for up to 16 years (693,457 person-years of follow-up) and 114,621 women followed for up to 26 years (2,848,555 person-years of follow-up). Information on physical activity was obtained in 1986 for both the HPFS and NHS; analyses included all participants who completed the physical activity questions and who were free of cancer prior to baseline (48,951 men and 707,086 person-years of follow-up in HPFS; 76,922 women and 1,183,390 person-years of follow-up in NHS).

Cox proportional hazards models for failure-time data were used to estimate incidence rate ratios (RR) and 95% confidence intervals (CI) for bladder cancer risk and to simultaneously adjust for age (1-year), pack-years of cigarette smoking history (<10, 10–24, 25–44, 45–64 and ≥65) and current smoking status. Age and smoking status were updated every 2 years in all analyses. Additional adjustment for potential confounders, including quintile of total fluid intake, geographic region in the US and reproductive factors (i.e. status and age at menopause), did not change the associations between weight, height or physical activity and bladder cancer risk. Because of the relative homogeneity of the population of male health professionals and female nurses, we did not control for education or socioeconomic status. Participants in both cohorts were categorized into 5 groups of BMI using whole number cutpoints that included widely used WHO definitions of overweight and obesity.22 For height, whole number cutpoints were created to keep person-years fairly evenly distributed across the categories. In both cohorts we used quintiles of total, vigorous and moderate recreational physical activity. To minimize the possibility that baseline BMI or physical activity may have been influenced by the presence of preclinical disease, an analysis excluding the first 4 years of follow-up was performed. We also examined the relationship between recent BMI or physical activity and the risk of bladder cancer by updating baseline exposures with those from subsequent biennial questionnaires. Additional analyses were restricted to tumor stage (invasive stages T1–T4). Tests of linear trend were conducted by assigning the median values for each category of BMI, height or type of physical activity and treating those as a single continuous variable, using Cox proportional hazards regression. Tests for (multiplicative) interaction were assessed by examining stratum-specific estimates and formally by use of likelihood ratio tests.

Initial analyses assessed BMI, height and physical activity with the risk of bladder cancer in the HPFS and NHS cohorts separately. Before pooling, tests of heterogeneity of the main exposures by cohort (to test for effect modification by gender) were performed using the Q statistic.23 Data from the 2 cohorts were pooled using a random-effects model for the log of the RR.23 All reported p-values are two-tailed. Statistical analyses were performed using SAS software version 8.2 (SAS Institute, Cary, NC).

Results

At baseline, mean BMI was 25.6 kg/m2 for men and 23.8 kg/m2 for women. An estimated 45% of men and 21% of women were categorized as overweight (BMI 25.0–29.9 kg/m2) and approximately 8% of men and women were categorized as obese (BMI ≥ 30 kg/m2). As expected, heavier individuals were less likely to engage in recreational physical activity (Table I). Current smokers were more likely to be leaner and were less likely to engage in exercise. For both men and women, height did not vary across levels of BMI or physical activity. Less than 8% of men and women had missing data on weight or physical activity; we did not observe any difference in baseline characteristics for men or women eligible for the anthropometric or physical activity analyses when compared to those who were excluded at baseline. Compared to the NHS cohort in 1976, women eligible for the physical activity analysis at baseline in 1986 were older (mean age 52.9 vs. 43.5 years), less likely to be current smokers (21% vs. 32%) and had a slightly higher mean BMI (25.4 kg/m2vs. 23.8 kg/m2) and mean total fluid intake (2048 ml vs. 1162 ml). Selection bias is unlikely given that smoking rates are expected to have dropped (from women quitting) over time and similarly for other characteristics that may have changed, except for fluid intake; the difference is due to the inclusion of water on the 1986 FFQ.

Table I. Baseline Characteristics1 among Men and Women According to Category of Body Mass Index (BMI) and Recreational Physical Activity
CharacteristicBMI (kg/m2)
Men2Women2
18.0–22.923.0–29.9≥30.018.0–22.923.0–29.9≥30.0
  • 1

    Means or proportions; Standardized to the age distribution of the study populations.

  • 2

    Baseline: 1986 for men; 1976 for women (fluid intake from 1980 questionnaire).

  • 3

    Baseline: 1986 for both cohorts.

  • 4

    Pack-years are calculated for current and past smokers.

  • 5

    Among women with natural menopause.

  • 6

    Women who responded to the 1986 questionnaire are very similar to the overall cohort (with respect to smoking and other factors that changed over time). Of note, total fluid intake was larger in 1986 because of the addition of “water” to the 1986 FFQ.

No. of individuals8,60535,3423,96758,64146,2219,759
Age (yr)53.954.554.341.544.144.3
Height (in.)70.170.269.964.664.464.2
BMI (kg/m2)21.825.732.721.025.533.7
MET (hr/wk)325.520.813.215.712.99.9
Past smokers (%)37.244.549.023.023.625.0
Current smokers (%)10.89.79.536.230.424.7
Pack-years of cigarettes459.860.169.932.634.338.8
Age at menopause (yr)5   49.149.549.6
Fluid intake (ml/d)1,8071,9552,1601,4431,1831,196
 Total recreational physical activity (MET hr/wk)
Men3Women3, 6
≤5.96.0–20.7≥20.8≤3.63.7–13.5≥13.6
No. of individuals16,16116,32016,47025,59225,67925,651
Age (yr)55.154.753.652.952.853.0
Height (in)70.070.170.164.464.564.5
BMI (kg/m2)26.225.625.026.325.424.6
MET (hr/wk)2.412.347.61.77.733.0
Past smokers (%)41.644.344.730.934.638.6
Current smokers (%)13.39.67.126.619.717.4
Pack-years of cigarettes463.061.758.047.144.740.4
Age at menopause (yr)5   50.350.450.6
Fluid intake (ml/d)1,9451,9532,0121,9832,0502,113

After adjusting for age, pack-years of smoking and current smoking, we observed no association between BMI at baseline and risk of bladder cancer (Table II). In the multivariate (MV) analysis for height, we observed a statistically significant inverse association between height and bladder cancer risk for men (MV RR, 0.69, 95% CI: 0.50–0.95, >6 when compared to ≤5.6 ft) but not for women. Risk estimates did not change appreciably after additional inclusion of other potential risk factors including state of residence in the US, total fluid intake and menopausal status, as pack-years of smoking accounted for most of the difference between the age and MV adjusted analyses. There was no association between weight loss or gain and the risk of bladder cancer (data not shown). Additional analyses restricted to invasive stages T1–T4 bladder cancer cases showed similar results among women (cases = 34; RR, 1.31; 95% CI: 0.44–3.92) and men (cases = 56; RR, 0.95; 95% CI: 0.32–2.79), for a BMI of ≥30.0 when compared to 18.0–22.9 kg/m2.

Table II. Rate Ratio (RR) for Incident Bladder Cancer by Category of Body Mass Index (BMI) and Height Among Men (1986–2002) and Women (1976–2002)
 BMI, kg/m2p-value for trendLinear Per 1 kg/m2increase2
18.0–22.923.0–24.925.0–26.927.0–29.9≥30.0
  • 1

    RR are adjusted for age.

  • 2

    Multivariate (MV) RR are adjusted for age, pack-years of cigarette smoking and current smoking; data presented as RR (95% confidence interval, CI).

Men
 Cases7614314710239  
 Person-years124,646197,132188,477127,01956,183  
 RR11.001.161.201.221.150.34 
 MV RR21.00 (ref)1.11 (0.84–1.47)1.14 (0.86–1.51)1.12 (0.83–1.51)1.01 (0.68–1.50)0.871.00 (0.97–1.03)
Women
 Cases16671523436  
 Person-years1,467,236556,958335,214253,400235,747  
 RR11.001.001.181.011.170.36 
 MV RR21.00 (ref)1.04 (0.78–1.38)1.23 (0.90–1.69)1.09 (0.75–1.58)1.31 (0.91–1.89)0.111.02 (0.99–1.04)
 Pooled MV RR21.00 (ref)1.07 (0.88–1.31)1.18 (0.96–1.45)1.10 (0.87–1.40)1.16 (0.89–1.52)0.19 
 Height, ftp-value for trendLinear Per 1 ft increase2
≤5.605.61–5.805.81–5.905.91–6.00>6.00
Men
 Cases1081208913060  
 Person-years107,201157,151119,255189,567120,283  
 RR11.000.840.880.880.720.08 
 MV RR21.00 (ref)0.82 (0.63–1.07)0.85 (0.64–1.13)0.85 (0.65–1.10)0.69 (0.50–0.95)0.040.68 (0.45–1.01)
 ≤5.105.11–5.305.31–5.405.41–5.50>5.50  
Women
 Cases7650719666  
 Person-years644,237345,648483,204801,322574,144  
 RR11.001.231.271.061.050.87 
 MV RR21.00 (ref)1.18 (0.83–1.70)1.25 (0.90–1.73)1.02 (0.76–1.39)0.98 (0.70–1.37)0.560.91 (0.54–1.55)
 Pooled MV RR21.00 (ref)0.93 (0.76–1.16)1.00 (0.81–1.24)0.92 (0.75–1.12)0.82 (0.65–1.03)0.06 

Results using updated BMI (see Material and methods) were similar to those using the baseline questionnaire only. Using the same categories of BMI, the pooled MV RR for bladder cancer risk were 1.08 for 23.0–24.9 kg/m2, 1.16 for 25.0–26.9 kg/m2, 0.98 for 27.0–29.9 kg/m2 and 1.19 (0.94–1.50) for ≥30.0 kg/m2.

To reduce potential bias from the influence of preclinical cancer on baseline BMI, we conducted additional analyses that excluded all cases of bladder cancer diagnosed within the first 4 years of follow-up. Based on 399 bladder cancer cases (529,091 person-years) among men and 340 bladder cancer cases (2,365,643 person-years) among women, associations with BMI were strengthened (Table III); men and women with a BMI of 30.0 kg/m2 or higher had a statistically significant 33% increase in the risk of bladder cancer when compared with men and women with a BMI of 18.0–22.9 kg/m2.

Table III. Multivariate Rate Ratio (MV RR) for Incident Bladder Cancer by Category of Body Mass Index (BMI) among Men (1986–2002) and Women (1976–2002), Comparing Two Different Methods of Analysis
 BMI (kg/m2)p-value for trend
18.0–22.923.0–24.925.0–26.927.0–29.9≥30.0
  • 1

    Multivariate (MV) RR are adjusted for age, pack-years of cigarette smoking and current smoking; data presented as RR (95% confidence interval, CI).

  • 2

    Exclude first 4 years removes the first 4 years of follow-up (and uses baseline BMI).

Men
 Baseline
  MV RR11.00 (ref)1.11 (0.84–1.47)1.14 (0.86–1.51)1.12 (0.83–1.51)1.01 (0.68–1.50)0.87
 Exclude first 4 years
  Cases581171127537 
  Person-years95,634150,694143,85796,69842,208 
  MV RR1, 21.00 (ref)1.18 (0.86–1.62)1.14 (0.83–1.58)1.09 (0.77–1.54)1.33 (0.88–2.03)0.38
Women
 Baseline
  MV RR11.00 (ref)1.04 (0.78–1.38)1.23 (0.90–1.69)1.09 (0.75–1.58)1.31 (0.91–1.89)0.11
 Exclude first 4 years
  Cases16067483035 
  Person-years1,220,126462,359278,093210,019195,046 
  MV RR1, 21.00 (ref)1.08 (0.76–1.35)1.18 (0.85–1.64)1.00 (0.67–1.48)1.33 (0.92–1.92)0.17
Pooled
 Baseline
  MV RR11.00 (ref)1.07 (0.88–1.31)1.18 (0.96–1.45)1.10 (0.87–1.40)1.16 (0.89–1.52)0.19
 Exclude first 4 years
  MV RR1, 21.00 (ref)1.08 (0.88–1.34)1.16 (0.93–1.46)1.05 (0.81–1.36)1.33 (1.01–1.76)0.10

We examined the relation between body size at a younger age using BMI at ages 21 years (HPFS) and 18 years (NHS) and bladder cancer risk. There was no association between men with a BMI of 25.0 kg/m2 or greater at age 21 and bladder cancer risk when compared to men with a BMI less than 25.0 kg/m2 (MV RR, 1.05, 95% CI: 0.84–1.31). Similarly, there was no association between women with a BMI of 25.0 kg/m2 or greater at age 18 and bladder cancer risk when compared to women with a BMI less than 25.0 kg/m2 (MV RR, 0.79, 95% CI: 0.54–1.18). Risk estimates did not change appreciably with further adjustment for current BMI.

The association between BMI and risk of bladder cancer was examined across strata of smoking status (never, past and current smoker). Among both men and women, there was no evidence that the association was modified by smoking status (p-value for interaction for men = 0.52 and women = 0.62; data not shown).

Overall, no association was observed between baseline total, vigorous or moderate recreational physical activity and bladder cancer risk among men or women (Table IV). Similarly, there was no relation between walking or hiking and the risk of bladder cancer. Results using updated physical activity or excluding the first 4 years of follow-up were similar to those using the baseline questionnaire only.

Table IV. Rate Ratio (RR) for Incident Bladder Cancer by Category of Recreational Physical Activity among Men (1986–2002) and Women (1986–2002)
 Total recreational physical activity, MET, h/wkp-value for trend
≤2.72.8–7.77.8–16.917.0–34.0>34.0
  • 1

    RR are adjusted for age.

  • 2

    Multivariate (MV) RR are adjusted for age, pack-years of cigarette smoking and current smoking; data presented as RR (95% confidence interval, CI). Vigorous recreational activity models further adjusted for moderate recreational activity.

  • 3

    Vigorous recreational activity includes activities that burn >6 METs: jogging (<10 min/mile); running (≥10 min/mile); bicycling (including stationary machine); lap swimming; tennis, squash or racquetball and other aerobic exercise (calisthenics, aerobics, rowing machine).

  • 4

    Moderate recreational activity includes activities that burn ≤6 METs: walking or hiking outdoors and stair climbing.

Men
 Cases109107939994 
 Person-years136,054137,787143,081144,331145,833 
 RR11.000.920.820.890.890.62
 MV RR21.00 (ref)0.95 (0.73–1.25)0.88 (0.66–1.17)0.99 (0.75–1.30)1.01 (0.76–1.34)0.70
 ≤2.02.1–4.64.7–10.410.5–21.7>21.7 
Women
 Cases4550373537 
 Person-years223,646245,429238,170235,912240,233 
 RR11.001.010.760.730.770.18
 MV RR21.00 (ref)1.11 (0.74–1.67)0.87 (0.56–1.35)0.83 (0.53–1.30)0.91 (0.58–1.41)0.46
 Pooled MV RR21.00 (ref)1.00 (0.80–1.25)0.88 (0.69–1.11)0.94 (0.74–1.19)0.97 (0.77–1.24)0.96
 Vigorous recreational activity3 (MET, hr/wk) 
 00.1–3.43.5–12.412.5–29.9≥30.0 
Men
 Cases21886955350 
 Person-years254,124120,360118,480104,954109,168 
 RR11.000.981.210.810.840.15
 MV RR21.00 (ref)1.05 (0.82–1.36)1.34 (1.05–1.71)0.90 (0.67–1.23)0.99 (0.72–1.35)0.72
 00.1–1.51.6–5.96.0–14.9>14.9 
Women
 Cases10032212922 
 Person-years542,945153,049120,616173,568193,212 
 RR11.001.221.010.930.670.06
 MV RR21.00 (ref)1.38 (0.92–2.06)1.17 (0.72–1.88)1.04 (0.68–1.58)0.77 (0.48–1.24)0.18
 Pooled MV RR21.00 (ref)1.14 (0.92–1.41)1.30 (1.04–1.62)0.95 (0.74–1.22)0.92 (0.70–1.19)0.44
 Moderate recreational activity4 (MET, hr/wk) 
 <0.80.8–2.22.3–4.84.9–10.8>10.8 
Men
 Cases90898699138 
 Person-years133,127146,675144,314141,829141,141 
 RR11.000.900.770.871.020.23
 MV RR21.00 (ref)0.93 (0.69–1.25)0.79 (0.59–1.07)0.90 (0.68–1.21)1.05 (0.80–1.38)0.22
 ≤0.91.0–2.93.0–4.54.6–10.4>10.4 
Women
 Cases4345413243 
 Person-years230,406246,331230,276243,347233,030 
 RR11.001.010.960.710.950.70
 MV RR21.00 (ref)1.09 (0.72–1.67)1.07 (0.70–1.65)0.81 (0.51–1.29)1.07 (0.70–1.64)0.97
 Pooled MV RR21.00 (ref)0.98 (0.77–1.25)0.88 (0.68–1.12)0.88 (0.69–1.12)1.05 (0.84–1.33)0.28

We also examined the possibility of an interaction between physical activity (in tertiles) and BMI (18.0–25.0, 25.0–29.9, ≥30.0 kg/m2) and risk of bladder cancer among men and women. There was no evidence that the association between total recreational physical activity level and bladder cancer risk was modified by BMI (p-value for interaction, among men = 0.57 and among women = 0.24; data not shown).

Discussion

In 2 prospective studies, baseline BMI was not related to the risk of bladder cancer; however, after removing cases within the first 4 years of follow-up, the association between BMI and bladder cancer risk strengthened. An inverse association between height greater than 6 ft and the risk of bladder cancer was observed among men. No association was observed between total, vigorous or moderate recreational physical activity, walking or hiking and the risk of bladder cancer. There was no evidence that the associations between bladder cancer risk and BMI or recreational physical activity was modified by smoking status or BMI, respectively.

Case-control studies of BMI and bladder cancer risk have been inconsistent. An early study observed a nonstatistically significant inverse risk of bladder cancer among women (cases = 110) with a BMI of 27.3 kg/m2 or greater when compared to women with a BMI less than 23.8 kg/m2 (MV RR, 0.61; 95% CI: 0.33–1.14).13 A Canadian population-based case-control study showed a statistically significant elevated risk of bladder cancer among men (cases = 670) and women (cases = 359) with a BMI of 30.0 kg/m2 or greater when compared to adults with a BMI of less than 25.0 kg/m2 (MV RR, 1.27; 95% CI: 1.01–1.58; p-value for trend = 0.03).4 However, no significant differences in BMI were observed between cases and controls in 2 other case-control studies.10, 12

Seven prospective cohort studies have examined the association between weight or obesity and bladder cancer risk. In 1 study, a positive association was observed between body weight recorded at college physical examination and bladder cancer risk among men (MV RR, 3.0; 95% CI: 1.1–8.0, for a weight of 180 pounds or greater when compared to less than 140 pounds), but no association was observed with BMI.7 In a cohort of postmenopausal women, no association was observed between risk of bladder cancer and elevated BMI after adjusting for several potential bladder cancer risk factors (112 cases; MV RR, 0.63; 95% CI: 0.33–1.19, for ≥30.69 when compared to ≤22.89 kg/m2).9 Similarly, no association was observed between BMI and bladder cancer mortality among men (MV RR, 1.14; 95% CI: 0.88–1.46) or women (MV RR, 1.34; 95% CI: 0.91–1.95, for BMI of 30–34.9 when compared to 18.5–24.9 kg/m2), in the Cancer Prevention Study II (872 men and 297 women bladder cancer cases).5 In 3 obesity cohorts, individuals hospitalized and discharged with a diagnosis of obesity had statistically significant 13–20% elevated risks of bladder cancer when compared to the general population3, 6 or other hospitalized patients.24 One major limitation with these population-based cohort analyses, however, is that potential risk factors, including smoking, were not included in the analyses. After adjusting for age, smoking status and occupational group, a recent cohort study of Austrian adults observed no association between obesity and bladder cancer incidence among men (78 cases; MV RR, 0.74; 95% CI: 0.45–1.22) or women (21 cases; MV RR, 1.60; 95% CI: 0.76–3.36, for BMI of 30.0–34.9 when compared to 18.5–24.9 kg/m2).11

To date, only 2 studies have reported associations between height and bladder cancer. No association was observed between height and the risk of bladder cancer in a case-control study with 263 cases (RR, 0.60 for men and RR, 1.80 for women, highest when compared to lowest quintile; p-value > 0.05).14 Additional adjustment for education, social class, BMI and (for females only) age at menarche and parity did not modify the observed associations; smoking history was not considered in the MV analysis. A prospective study of adult height and cancer mortality (18 years of follow-up; number of cases not reported) observed no association between a 6 in. difference in height and bladder cancer mortality after adjusting for age, employment and current smoking.15

We observed a statistically significant inverse association between height and bladder cancer risk among men after adjusting for age, pack-years of cigarette smoking and current smoking. In contrast, we previously reported positive associations between height and the risk of several cancers including cancers of the colon,25 pancreas26 and prostate27 in our cohorts. Height may reflect hereditary factors, nutrition in the remote past or may be a marker for exposure levels to growth factors, including IGF-1, which is mitogenic and antiapoptotic. IGF-1 has been suggested to play a role in several of these cancers,28, 29, 30 including bladder cancer.31, 32 In light of this mechanism, height, as a marker of high IGF-1 levels, would be expected to increase the risk of bladder cancer. The inverse association of height with bladder cancer risk among men observed in the current study may be simply a chance finding or it may potentially suggest different etiologies among the genders or that poorer nutrition in adolescence or childhood may lead to a subsequent elevated risk of bladder cancer in adulthood.

Few studies have examined the relation between physical activity and risk of bladder cancer. Regular physical activity was associated with a suggested decrease in risk of bladder cancer among a cohort of postmenopausal women with 112 cases (MV RR, 0.66; 95% CI: 0.43–1.01).9 No association was observed between physical exercise (<5 when compared to ≥5 hr per week) reported at college exam and bladder cancer risk among men.7 Among a cohort of 7,588 British men, aged 40–59 years, 92 developed bladder cancer during a mean follow-up of 18 years.16 Compared to none to moderate physical activity, the RR for bladder cancer for sporting (vigorous) activity was 2.06 (95% CI: 1.08–3.95) after adjusting for age, cigarette smoking, BMI, alcohol intake and social class. Furthermore, a statistically significant increasing trend of bladder cancer risk with increasing frequency of sporting activity was observed (p-value for trend = 0.04). Regular walking was not related to bladder cancer risk (MV RR, 1.43; 95% CI: 0.57–3.58, for >60 when compared to <20 min per day).16 The lack of consistency between the studies may be attributable to different definitions of physical activity or chance findings due to small case numbers.

The strengths of our study include its large sample size, the prospective design with up to 26 years of follow-up, data from 2 separate cohorts and detailed information on smoking history. The prospective design precludes recall bias and selection bias is minimized by the very high rate of follow-up over a long period of time. In addition, we were able to tightly control for history of smoking and update smoking status in all analyses.

We cannot exclude measurement error as a contributor to the lack of associations in the current study. Obesity was measured by BMI, which was estimated from self-reported weight and height. We have previously demonstrated the accuracy of self-reported anthropometric measures in these cohorts, and several previous positive findings of health conditions with BMI in these cohorts further support that obesity is well-reported in our populations.25, 33, 34 Similarly, we used a validated self-reported measure of recreational physical activity. Even though some measurement error is likely to be present in the physical activity measure, especially since we have no data on occupational activity, associations with physical activity and various outcomes have been consistent with the literature in these cohorts. For example, physical activity is inversely associated with colon cancer in both cohorts.25, 34 Therefore, these previous studies indicate that our measure of physical activity is sufficiently accurate to detect true associations. Thus, the assessment of physical activity does appear to capture etiologically relevant variation in exercise levels.

In conclusion, in 2 large prospective cohorts of men and women, we observed no appreciable association between baseline BMI and the risk of bladder cancer; however, after excluding the first 4 years of follow-up, the positive association between BMI and bladder cancer risk strengthened. Overall, our findings suggest that obesity, height and physical activity are not likely to play a significant role in the etiology of bladder cancer.

Acknowledgements

We are indebted to Dr. Walter Willett for his expert help and support in this study.

Ancillary