Article Text
Abstract
Objectives: Air pollution has been associated with an increased risk for lung cancer. We examined whether long-term air pollution is associated with bladder cancer risk.
Methods: Information from a case-control study in Spain that included 1219 incident cases and 1271 hospital controls was used. Information on residential history including several indicators of exposure to air pollution and other potential risk factors was collected in a face-to-face computerised personal interview. Odds ratios (OR) and 95% confidence intervals (95% CI) were adjusted for age, gender, region, smoking, occupation, water contaminants and diet.
Results: Living more than 40 years in a city with a population of more than 100 000 was associated with an increased risk for bladder cancer overall (OR 1.30, 95% CI 1.04 to 1.63). Emissions of polycyclic aromatic hydrocarbons and diesel from industries near the residence, as evaluated by experts, were associated with an increased risk (OR 1.29, 95% CI 0.85 to 1.98), while lower or no excess risks were observed for other pollution-related variables. Odds ratios among never smokers tended to be higher than among smokers.
Conclusions: The small to moderate positive associations found for several indices of air pollution and bladder cancer, while suggestive of excess risk, require further evaluation in other settings.
Statistics from Altmetric.com
Particulate air pollution has been associated with increased lung cancer mortality in the American Cancer Society Cancer Prevention II study.1 In the EPIC prospective study in Europe, indices for air pollution, such as residential proximity to heavy traffic, have been associated with an increased incidence of lung cancer.2 Several air pollutants, including polycyclic aromatic hydrocarbons (PAHs) and diesel engine exhaust, have been shown to be carcinogenic and have been associated with an increased risk for bladder cancer in the occupational setting.3
An increased risk of bladder cancer for exposure to total PAHs and benzo(a)pyrene (B(a)P) was observed among workers in a pooled analysis of 11 European case-control studies4 and has been observed among occupations in industries with high exposure to PAHs such as aluminium production, coal gasification and tar distillation.5 A higher risk has also been observed for occupations with exposure to diesel motor exhaust, especially truck drivers.3 6 1-Nitropyrene is one of the compounds found in the particulate emissions of diesel engines, and has been classified as possibly carcinogenic to humans by the International Agency for Research on Cancer.7 PAH levels in occupational settings, however, may be considerably higher compared to levels in the ambient air. For example, in an aluminium plant in Quebec,8 average exposure to B(a)P in 1970–1988 was 8.5 μg/m3. In the city of Barcelona, Spain, the level of B(a)P sampled in the air during the winter of 1990 reached 6.5 ng/m3 with a total concentration of PAHs of 69 ng/m3.9 Among specific PAHs, the highest levels observed were for benzo(ghi)perylene. The presence of this compound denotes that gasoline-powered vehicle emissions are among the main sources of PAHs, since it is one of the individual PAHs with highest emission rates from these vehicles.10
We examined if long-term exposure to air pollution, particularly industrial emissions and surrogate indices of vehicle emissions, was associated with an increased risk of bladder cancer in the Spanish Bladder Cancer Study. This is one of the first studies to examine risk of bladder cancer and exposure to air pollution.
METHODS
We recruited 1219 incident bladder cancer cases and 1271 hospital controls between 1998 and 2001 from 18 hospitals in five areas of Spain: Barcelona, Vallès/Bages, Alicante, Tenerife and Asturias (table 1). Subjects were 21–80 years of age and lived in the catchment area of the participating hospitals. Cases were histologically confirmed by a panel of expert pathologists which also ensured uniformity of classification criteria according to the World Health Organization system.11
Controls were selected from patients admitted to the participating hospital with diseases not a priori related to bladder cancer or known risk factors for the disease. The main diagnoses in hospital admissions were hernias (37%), other abdominal surgery (11%), fractures (23%), other orthopaedic conditions (7%), hydrocele (12%), circulatory conditions (4%), dermatological conditions (2%) and ophthalmologic conditions (1%). Controls were individually matched to cases on age at interview within 5-year categories, gender, ethnicity and region of residence at the time of hospital admission. Only six subjects were non-Caucasian and were not further examined in this analysis. A total of 84% of cases and 88% of controls completed a face-to-face computer-assisted personal interview that included information on socio-demographic data, smoking, occupational history, lifetime residential history and environmental exposures, medication, family history of cancer and other factors. Nearly all interviews were conducted while subjects were inpatients at the hospital. Overall, 20% of the subjects completed a critical items questionnaire that was an abbreviated version of the complete interview: it included lifetime residential history but not specific information on surrogates of air pollution exposures. Written informed consent was obtained from all patients. The study was approved by the ethics committees of all participating institutes.
Several indices were used to assess exposure to air pollution at each residence occupied for more than 1 year during the patient’s lifetime: proximity to industries, windows facing traffic and size of the city of residence. Questions about the type and quantity of traffic (number of lanes and intensity of traffic defined as <20, 20–49 or ⩾50 trucks or buses per day on the traffic street which the residence faced) were asked about each residence with windows facing a traffic street. The size of the city of residence was obtained from the National Institute of Statistics, Spain for the year 2000. For each of these indices, we classified subjects according to air pollution exposure information and the number of years spent at each residence during their lifetime as low exposure or high exposure (table 2). Subjects who did not fit in either category were classified as intermediate exposure. For instance, a subject who had lived all of their life in residences without windows facing a heavy traffic street was classified as low exposure; by contrast, a subject who had lived more than 30 years in residences with windows facing a traffic street was classified as high exposure.
Subjects reporting the presence of an industry within 1 km of their residence were asked to report the type of industrial plant (ie, textile, coal mine, etc) and, if they knew, the main products (ie, cars, electric equipment, etc). Subjects reported a broad spectrum of industry types and primary products (488 different entries). An industrial hygienist (SG) evaluated the possible emissions of PAHs and diesel particles from each of these industries, blinded as to case-control status. These evaluations were then reviewed by a second expert (MD) and exposure to PAHs and diesel particles was assigned to 737 subjects who reported having an industry in the vicinity of their home.
Results are reported for lifetime exposure; introducing a lag in the exposure windows did not change the observations, and results are not shown. We also analysed the air pollution indices for the longest residence only; results were similar to those obtained when using lifetime residential history, and are not shown here. Residential history was available for a total of 137 217 person-years (84% of the total person-years of subjects in the study from birth until interview), with a mean duration of 19 years at each residence, ranging from 15 years in the area of Barcelona to 22 years in the Alicante region.
Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated using logistic regression with the statistical package Stata 8.2. Results were adjusted for age, gender, region, smoking (pack-years) and a priori high risk occupations (metal workers, hairdressers and barbers, painters, and chemical, leather, transport and rubber industry workers). Additional adjustment for consumption of fruits and vegetables and for exposure to disinfection by-products in water did not modify results and are not presented. Tests for linear trend were performed.
RESULTS
Most subjects (87%) were males (table 1). Smoking status differed between cases and controls and also between genders, with 13% of males being never smokers compared with 79% of females. The average time of residence in the same home was 18 years for cases and 19 years for controls. A total of 34% of cases and 33% of controls had ever lived in residences having industries in their proximity (table 3), and more than half of the subjects did not have windows facing heavy traffic.
Living for more than 40 years in a city with a population above 100 000, compared with living more than 25 years in a small village/town, was associated with an increased risk of bladder cancer (OR 1.30, 95% CI 1.04 to 1.63; never smokers OR 1.17, 95% CI 0.71 to 1.94) (table 3). Subjects with intermediate exposures had also an increased risk (OR 1.23, 95% CI 1.00 to 1.53; never smokers OR 1.25, 95% CI 0.75 to 2.07), with a significant increasing risk for all subjects (ptrend = 0.027). Lower, non-significant excess risks were observed for residential proximity to an industry (OR for >25 years 1.14, 95% CI 0.90 to 1.44; never smokers OR 1.18, 95% CI 0.70 to 2.00) and for number of traffic lanes (OR 1.11, 95% CI 0.65 to 1.92; never smokers OR 1.24, 95% CI 0.44 to 3.52). A small non-statistically significant excess risk for windows facing heavy traffic streets was seen only among never-smokers (OR 1.30, 95% CI 0.80 to 2.12). Combinations of exposure indices to take into account different sources of air pollution did not substantially modify most results (table 3). Similar results were observed for the longest residence, with an OR of 1.07 (95% CI 0.94 to 1.21) for having windows facing a heavy traffic street and an OR of 1.21 (95% CI 0.95 to 1.56) for living in a city.
Among subjects living close to industries, experts assessed possible emissions of PAHs and diesel exhaust from the reported industries. For example, repair shops and metallurgy industries were classified as having emissions of both pollutants; on the other hand, steel works, sulfur or coal industries were considered to have only PAH emissions. Only subjects who had a score from the industrial hygienist evaluation were included in the exposed group of subjects. The referent in the analysis was the group of subjects who reported never having industries in the proximity of their residences during their lives. A higher but not statistically significant risk was observed among subjects with residences within 1 km of an industry with either PAH or diesel emissions (OR 1.21, 95% CI 0.94 to 1.55) and for exposure to both pollutants (OR 1.29, 95% CI 0.85 to 1.98); ptrend>0.05 (table 4). Length of time lived in the residence did not modify these results substantially, nor did adjustment by occupational history or size of the city. This analysis was not carried out separately for never smokers because of small numbers.
In a stratified analysis by gender, ORs for all air pollution indices tended to be slightly higher in women than in men, but differences were not statistically significant. For example, having spent most of one’s life in a city versus a small town was associated with only a slightly higher risk in women (n = 311, OR 1.34, 95% CI 0.74 to 2.42) than in men (n = 2129, OR 1.30, 95% CI 1.02 to 1.67). The OR for male subjects (n = 2125) who had lived for more than 25 years close to an industrial plant compared with those who had never lived close to industries was 1.09 (95% CI 0.85 to 1.40). In women (n = 309) the corresponding OR was 1.51 (95% CI 0.79 to 2.89). Differences between genders were not pronounced for size of the city of residence.
DISCUSSION
We examined the association of bladder cancer risk with air pollution indicators for traffic and industry in a large case-control study. We found small to moderate positive associations for several indices of air pollution. Our findings are supported by studies showing an elevated risk of bladder cancer among workers with occupational exposure to PAHs and diesel engine exhaust.5 The strength of associations of bladder cancer risk with ambient air pollution exposures are likely to be considerably lower than those observed in the occupational environment, similar to the results found for environmental lung cancer.2 12–14
Similar to other studies examining the long-term effects of air pollution, the exposure indicators used in this analysis were based on both individual and grouped data.14–16 There was a high likelihood of misclassification in estimates of lifetime exposure, particularly since exposure estimates were mainly based on self-reports. A statistically significant association of bladder cancer risk was found with size of the city, and risk increased with time lived in large cities. Lower excess risks were observed among subjects who lived close to an industry or had lived on a street with more than two lanes of traffic in both directions for 25 years or more. City size is an ecological variable that correlates well with air pollution measurements, and it has been used as an indicator of air pollution.15 17 Nonetheless, subjects living in different sized cities may also differ in life-style, and residual confounding may occur after adjusting for major bladder cancer risk factors.
Smoking is the main risk factor associated with bladder cancer, and might confound the relationship between air pollution and bladder cancer. The associations we observed tended to be stronger among never smokers. More than 50% of bladder cancer in men and about 30% in women has been attributed to smoking.18 A higher risk among non-smokers, even if not significant, indicates that the overall excess risk for all study subjects cannot be attributed to residual confounding by smoking, and supports the observed associations. Residual confounding by other risk factors for which we adjusted in the analysis such as diet, social class or other environmental contaminants, could also be present. These risk factors are only weakly associated with bladder cancer risk and it is unlikely that incomplete adjustment would have affected results for air pollution in any considerable way.
Accurate exposure assessment is a key step in the study of air pollution-related disease. For long-term air pollution, individual measures are not available, and imputing fixed-site air monitoring measurements to an individual may lead to errors due to heterogeneity in the distribution of traffic-related compounds in the air.19 Levels of traffic intensity were not available to validate the traffic-related exposure to air pollution. In a recent study of exposure to PAHs, self-reported traffic density through a survey was comparable (Kruskal-Wallis one-way ANOVA, p = 0.001) with GIS-measured traffic density values.20 With the exception of the city of Barcelona (1 505 000 inhabitants), very few measurements of airborne PAHs are available in Spain. During the period 1987–88, the mean level of the sum of 12 PAHs in the air in the city of Barcelona was 50 ng/m3, with much higher levels in winter than in summer (85 vs 24 ng/m3)21; for the same period of time, the mean level of total suspended particles (TSP) was 175 μg/m3. In the beginning of the 1990s, TSP levels dropped to a mean level of 56 μg/m3 and PAHs levels to 5.80 ng/m3 (communication from the Highway and Urban Air Pollution Symposium, Barcelona, May 2002) mainly due to the shift from gasoline/diesel to natural gas as the power source of most industries, and to the more stringent regulation of vehicular exhaust. Data from the cities of Sabadell (185 000 inhabitants; Vallès region) and Manresa (64 000 inhabitants; Bages region) were available only from 1995 to 1998, with a mean TSP level of 124 and 91 μg/m3, respectively (data from the Catalan Air Pollution Monitoring Network webpage; see http://mediambient.gencat.net/cat/el_medi/atmosfera/immissions/inici.jsp?ComponentID = 28451&SourcePageID = 23429#1, and the Ministry of the Environment of the government of Catalonia). Data on TSP variables from the main cities in the Asturias region were also available for the period 1993–98 with mean levels for Oviedo (201 000 inhabitants), the main city of the region, being 76 μg/m3 (data from the government of the Principado de Asturias). Although no data are available, it can be expected that air pollution levels decreased in all regions of Spain as they did in Barcelona. The limited amount of data does not allow a retrospective ecological exposure model to be constructed. Some indicators, such as living near a major road, are likely surrogates for exposure at least for some contaminants22 23 and have been associated with mortality,24 and in the EPIC study in Europe with lung cancer incidence.2
Main messages
Several air pollutants, such as polycyclic aromatic hydrocarbons, have been shown in the occupational setting to increase bladder cancer risk.
Small to moderate positive associations were found for several indices of air pollution and bladder cancer.
Associations between air pollution and bladder cancer risk were stronger among non-smokers.
Policy implications
Air pollution is being associated with an increasing number of health outcomes.
Risk assessment for the evaluation of levels of exposure to air pollution should also consider bladder cancer as one of the potentially associated health outcomes.
In addition to traffic, industrial emissions also contribute to airborne PAHs. From the 498 types of industries or combinations of these industries reported, the most common were metallurgy and textile industries, foundries and refineries, representing 25% of the total. All primary metals industries and petroleum-related industries have relatively high airborne emissions of PAHs. The Toxic Release Inventory of the Environmental Protection Agency in the USA estimates that more than 60% of the total PAHs emissions during the year 2000 were from these industries. The textile industry ranked third, with 9% of total emissions (Toxics Release Inventory Program, US Environmental Protection Agency). Therefore, we evaluated the specific emissions of PAHs and diesel fumes from these industries which were located within 1 km of places of residence of cases and controls. Diesel fumes contain both PAHs and nitro-PAHs,25 the latter clearly associated with mutagenic activity.26
It is unlikely that the selection of hospital controls from the catchment areas of the hospitals has biased our results. None of the hospitals included in the study is a referral hospital for cancer and both cases and controls were selected from the catchment areas of the hospitals. In addition, the regions subjects originated from were large enough to ensure heterogeneity of environmental exposures. Finally, cases and controls were selected on the basis of current residence, but information on air pollution was derived from lifetime residential history. We consider that exposure misclassification was non-differential for cases and controls, and if any, we would expect the bias to be towards the null.
Our study is the first to examine the risk of bladder cancer and exposure to air pollution. Small to moderate positive associations were found for several indices of air pollution, with slightly stronger associations found among never smokers. However, most results were not statistically significant. An association of several air pollutants with bladder cancer is supported by findings from high exposure occupational settings, even though the increased risk for environmental exposures would not be expected to be high.
Acknowledgments
We thank Nino Kuenzli for his comments on this paper. We thank Robert C Saal from Westat, Rockville, MD, Leslie Carroll and Eric Boyd from IMS, Silver Spring, MD and Paco Fernández, IMIM, Barcelona, for their support with study and data management; Dr Maria Sala from IMIM, Barcelona for her work in data collection; Silvia González (SG) for the evaluation of industrial emissions; Jan-Paul Zock, Hans Kromhout and Ferran Ballester for helping with exposure assessment; and physicians, nurses, interviewers (Ana Alfaro, Cristina Villanueva, Cristina Pipó, Joan Montes, Iolanda Velez, Pablo Hernández, Ángeles Pérez, Carmen Benito, Adela Castillejo, Elisa Jover, Natalia Blanco, Avelino Menéndez, Cristina Arias, Begoña Argüelles) and all participants in the study for their efforts during field work.
Footnotes
Funding:This research was primarily supported by the Intramural Research Program of the NIH, National Cancer Institute, Division of Cancer Epidemiology and Genetics (NCI contract no. NO2-CP-11015). This project was also funded by the Spanish Ministry of Health (FIS 2001–2002), EPICUR-red (ISIII-GO3/174) and the European Union (Environment and genetic factors in bladder cancer: a multicentric case-control study in Europe. BIOMED. 1998–2001).
Competing interests: None.