Several environmental carcinogens are found to be spread across wide geographic areas, and the exposed inhabitants are at risk of developing various types of cancers. Arsenic and disinfection by-products in drinking water, ultraviolet rays from the sun, and agricultural chemicals used in golf courses were found to be the possible cancer risks. The study aimed to estimate the risks of cancer due to exposure to environmental carcinogens known to be present in wide geographic areas in Newfoundland and Labrador (NL). The NL cancer care registry provided 2008–2017 data (histological diagnosis, age, sex, and six-digit postal code) on cancers relevant to arsenic, disinfection by-products, ultraviolet rays, and agricultural chemical exposures. The geographic distribution of environmental carcinogens was collected from government sources and previous studies. Risk ratios (RR) of annual prevalence rates of cancers in high-risk (exposed to environmental carcinogens) and low-risk populations. For ultraviolet rays, arsenic, disinfection by-products, and agricultural chemicals, the RR (95% CI) were 1.5 (1.4–1.6), 1.25 (1.03–1.51), 1.8 (1.67–1.94), and 1.49 (1.3–1.7), respectively. An excess number of cancers in high-risk areas was possibly associated with exposure to environmental carcinogens . Public health regulations, environmental monitoring, health promotion, and increased awareness in high-risk areas can prevent exposure to environmental carcinogens.
The International Agency for Research on Cancer has identified more than 200 agents as carcinogenic (Group 1) and probable carcinogens (Group 2A) to humans (IARC, 2019). Environmental carcinogens are broadly defined as compounds that are the subset of “known” and “reasonably anticipated” human carcinogens and are considered nongenetic exogenous factors that contribute to cancer risk (Sabo-Attwood et al., 2006; WHO, 2011; Wogan et al., 2004).
In Canada, neoplasms rank at the top for all-age disability-adjusted life year (DALY) counts and rates of age-standardized DALY per 100,000 (Lang et al., 2018). A broad estimate in Ontario has identified between 3,500 and 6,500 new cancer cases each year as a result of exposure to 23 environmental carcinogens (CCO, 2016). Some of these environmental carcinogens are present in nature and are spread across wide geographic areas, putting the entire exposed population at risk of developing cancer. However, Canada’s current cancer prevention strategies have yet to pay adequate attention to identifying and acting on the vulnerable population living in areas with potentially high environmental carcinogen exposure. Hence, health professionals have not been able to develop any environmental carcinogen-specific cancer prevention strategies in such high-risk areas. Also, local medical practitioners might not have any scope to alert local health authorities on the abnormally high prevalence of any cancer.
Newfoundland and Labrador (NL) has the highest age standardized incidence rate of cancer (587/100,000) (CCS, 2017). Based on available environmental contamination data for NL, four environmental carcinogens were found to be present in wide geographic areas. These were arsenic and disinfection by-products in drinking water, ultraviolet rays from the sun, agricultural chemicals (herbicides/fungicides/pesticides) in ambient air, and (or) dusts affecting households living in close proximity to a golf course (CAPE, 2016; CBC, 2019; de Leeuw, 2017; GoNL, 2012; Minnes & Kelly Vodden, 2017).
Arsenic is naturally present in underground sediments and contaminates well water (GoNL, 2019a). The municipalities have public water systems that treat raw water before supply and regularly monitor its quality after treatment (including arsenic). Therefore, it is unlikely that there is a high arsenic level in household taps supplying public water (Thomson et al., 2019). The communities affected by arsenic in drinking water are usually small in population size and do not have access to a public water supply and rely upon their own artesian wells, and the monitoring of water quality solely remains the responsibility of the individual well owners (Thomson et al., 2019). Thus, there are no official reports of arsenic levels in private wells (GoNL, 2019a).
Disinfection by-products are a very complex group of chemicals formed during the water-treatment process when disinfectants such as chlorine are added to untreated or partially treated raw water before removing organic matter (CDC, 2016). There are more than 600 disinfection by-products in chlorinated tap water, though only two types of disinfection by-products, i.e., trihalomethanes (THMs) and haloacetic acids (HAAs) are regularly tested (Bull et al., 2011; CDC, 2016). THMs and HAAs have been identified as weak carcinogens (Group 2B, possible human carcinogen), and they are often used as a proxy for cancer risk assessment (IARC, 2018; Nieuwenhuijsen et al., 2009; Salas et al., 2013; WHO, 2004).
According to the Canadian Cancer Society, sunlight in Canada is strong enough to cause skin cancer, one of the most common types of cancers (CCS, 2020). NL is known for prolonged foggy weather, and thus there is a misconception that there is a low risk of skin cancer due to a lack of direct sunlight (CBC, 2019).
Golf courses are known for using various types of agricultural chemicals (Golf ventures, 2019). In Canada, golf courses were exempted by municipal bylaws that restrict use of agricultural chemicals on private residential and municipal lands (CAPE, 2016). Agricultural chemicals are heavily used on golf courses, with four to seven times greater than the recommended doses meant for any agricultural farms (Feldman, 2020; Golf ventures, 2019). In 2012, NL banned the use and sale of some known carcinogenic agricultural chemicals (2,4-dichlorophenoxyacetic acid, carbaryl, and 2-methyl-4-chlorophenoxyacetic acid) on lawns, but golf courses were exceptions (Band et al., 2011; GoNL, 2012, 2019b). Golf courses are required to provide notice to all properties located just within 15 m of the proposed agricultural chemical application sites (GoNL, 2019b; VoPham et al., 2015). Several studies conducted elsewhere found that populations living within 500 m of agricultural farms are subject to airborne exposure to agricultural chemicals due to drift (Bernardi et al., 2015; Golf ventures, 2019; Ward et al., 2006). Agricultural chemicals are also transported via dust particles from farmlands and are carried away by strong winds. People are thus exposed to agricultural chemicals by the inhalation of contaminated air and dust. Furthermore, they are exposed to agricultural chemicals by ingestion after touching contaminated surfaces (by air and dust) in and around their residences. However, except for occupational (golfers, golf course maintenance workers) cancers, there is no published study that examined associations between residential exposure to agricultural chemicals in populations living in proximity to golf courses and a higher prevalence of the cancers caused by agricultural chemicals (Knopper & Lean, 2004; Kross et al., 1996; Murphy & Haith, 2007; Putnam et al., 2008).
We hypothesize that spatial distributions of environmental carcinogens are associated with prevalence of related cancers. The ecological study aimed to estimate the risks of cancers due to exposure to ultraviolet rays, arsenic, disinfection by-products, and agricultural chemicals.
Cancer data (histological and ICD code, age, sex, and geographic distribution (using six-digit postal code) from 2008 to 2017 were collected from the NL Cancer Care Registry (NLCCR). The registry contains data at the population level on cancer cases diagnosed in NL, and there was a near-complete case ascertainment. The registry was queried using the relevant data specifications described above. The cancers (histological types) known to have either of the four environmental carcinogens as risk factors were selected for our study (Table 1). However, the NLCCR did not provide any background information on the exposure history of the environmental carcinogens for the registered cancer cases, nor other risk factors such as smoking, diet, occupational exposure, etc.
Selection of communities for each environmental carcinogen
Daily ultraviolet index (UVI) monitoring data (1 March 2013 to 28 February 2019) for 37 meteorological centres of NL were collected from Environment and Climate Change Canada. UVI-6 was considered as high-risk level (protection required to prevent sun burn and skin damage) (Health Canada, 2018a, 2018b). The monitoring centres were ranked according to the number of days having UVI-6 (or more) during the data period (Health Canada, 2019). NL is known for its cold climate and summer is the most popular season for local outdoor activities. The UVI (for all the meteorological centres) were high in summer (end of May to beginning of September, i.e., ~100 days). The monitoring data show that the days with UVI-6 (or more) were essentially found during summertime, for both the high-risk and low-risk centres. The centres having UVI-6 (or more) for ~100 days (and above) per year were selected as high-risk centres, and the rest were selected as low-risk centres. The communities located within a 50-km radius of each centre were selected for the study (Figure 1A) (Daly, 2006; Fioletov et al., 2004).
Ten high-risk communities (Cormack, Campbellton, Baytona, Main Point, Fredericton, Deep Bay (Fogo Island), Bridge Port, Carter’s Cove, Moreton’s Harbour, and Valley Pond) were selected for the study. Arsenic in Cormack (population: 597) was accidentally discovered during a community-based research on the quality of private well water in 2011–2012. Arsenic was discovered in nine other small communities (population range 83–615) by the personal initiative of a local family physician, who suspected high incidences of some cancers potentially related to arsenic exposure (de Leeuw, 2017). The well owners and the community members voluntarily shared 96 water quality reports, showing 55 samples above the guideline value of 10 parts per-billion (ppb) (range 11–1,040 ppb, average 150 ppb) (GoNL, 2019a). As per the NL government’s policy, public water is regularly tested for quality, including arsenic, and the results are shared on its water portal (GoNL, 2020). Based on the report, two communities (Gander and Twillingate, very close to the high-risk communities), supplied by treated public water were selected as a control (low risk) population. The census data from 2016 show similar age and gender distribution in high-risk and low-risk populations. Since certain types of skin cancers (squamous cell and basal cell carcinomas) are also caused by ultraviolet rays (Table 1), the background UVI of the high-risk and low-risk communities were checked to ensure there was no overlapping. All selected communities (arsenic exposed and control) were low-risk UVI areas (Figure 1B).
THMs and HAAs are the disinfection by-products regularly tested four times a year by the NL government, and the reports (2010–2016) were made available on the same water portal (GoNL, 2020). Community-wise THM and HAA reports were transferred to an Excel spreadsheet. The communities having geometric averages of both THMs and HAAs above and below guideline values (THMs-0.1 mg/L and HAAs-0.08 mg/L) were selected as high-risk and low-risk areas, respectively (GoNL, 2019c). We have also identified the communities having either only high THMs or only high HAAs. Arsenic and disinfection by-products are the risk factors for cancers of urinary bladder (transitional cell carcinoma and urothelial cell carcinoma) and colon (adeno carcinoma) (Table 1). Since arsenic and disinfection by-products were found in private wells (communities not supplied by public water) and public water systems, respectively, there was no double exposure.
Out of a total 18 golf courses, nine have neighbourhoods surrounding them, and four are located within the St. John’s metropolitan area. In our study, the neighbourhoods located within 500 m of the boundary of nine golf course were selected as the high-risk population (Bernardi et al., 2015). With the help of the cartography department of the Memorial University library, the high-risk areas around the golf course were demarcated and each high-exposure risk area was further divided into six-digit postal codes (Figure 2). For low-risk areas, we selected the town of Conception Bay South (20 km from St. John’s, total population 26,199), and 15 small coastal fishing communities close to the town of Conception Bay South (total population 9,088; range 127 to 3,448) which have no golf course or agricultural land within 5–6 km from their boundary. Census data (2016) show no notable differences in age and gender distribution between high-risk and low-risk population. “Arsenic and agricultural chemicals” and “disinfection by-products and agricultural chemicals” were the common risk factors for renal cell carcinoma and acute myeloid leukemia (AML) and chronic myeloid leukemia (CML), respectively (Table 1). However, there was no golf course in the communities affected by groundwater arsenic. Also, disinfection by-product levels of the public water sources in the communities having golf course were lower than the guideline values (low-risk communities).
For arsenic, disinfection by-products, and ultraviolet rays, the total populations of the high-risk and low-risk communities were taken from 2016 census data produced by the Demography Division of Statistics Canada (Statistics Canada, 2019). The cancers (histological type) selected and analyzed for each carcinogen category are listed in Table 1. To count the total number of cancer cases in high-risk and low-risk communities, we first listed all the corresponding postal codes for each community. Then, we counted individual histological types of cancers (listed in Table 1) for each carcinogen (arsenic, disinfection by-products, and ultraviolet rays) and their demographic backgrounds (age and sex) from these postal codes and added them together.
To identify high-risk neighbourhoods around nine golf courses, all the postal codes within the high-risk areas were listed from the map (Figure 2). We mapped each high-risk area using a high-resolution Google satellite map and counted individual homes located in every postal code (Figure 3). For the postal code areas that extended beyond the 500-m boundary of the high-risk area, the entire postal code area was included for counting homes. Large buildings, such as apartment/condo complexes, were verified by browsing street view images, which allowed us to count the actual number of apartments or condos. The total number of houses in the high-risk areas was multiplied by 2.3 (average household size of NL) to generate the total population size (10,988) (Statistics Canada, 2019). Individual types of cancers due to agricultural chemicals (Table 1) and their age and sex were collected from the NLCCR according to the postal codes of the high-risk areas and added together. For low-risk areas (town of Conception Bay South and 15 small coastal fishing communities), the total population was obtained from the census, and from each postal code for the town and small communities, the selected cancer cases along with demographic backgrounds were collected.
For each environmental carcinogen category, the number of corresponding cancer cases was added together before calculating the average annual prevalence rates in high-risk and low-risk communities, and, subsequently risk ratios (RR). To measure significance of RR, 95% confidence intervals (CI) were calculated.
The excess number of cancer cases in the high-risk population associated with a specific environmental carcinogen was calculated by:
total number of cancer cases in the high-risk population – average annual prevalence rate in the low-risk population × total population in the high-risk area.
Table 2 shows that for all the environmental carcinogens, the annual prevalence rates of cancers are significantly higher in the high-risk populations. The prevalence rates of cancer among males were higher in all environmental carcinogen categories (both in high-risk and low-risk areas). There were no noticeable differences in average ages between high-risk and low-risk categories.
Since the sun’s rays become stronger as we move south, UVI also increases (Health Canada, 2018b). Figure 1A shows the high-risk areas only in the southern part of the province; 280,034 people (i.e., 54% of total NL population) were from high-risk areas. Estimated additional burden of cancer cases in high-risk areas were 3,043 in 10 years (2008–2017). Potential arsenic- exposed population in 10 communities was 2,876, i.e., almost 1,250 households (average household size of NL is 2.3 people) (Statistics Canada, 2019). There are an estimated 40,000 private wells in NL that are operating without any information on their arsenic profiles (Roche et al., 2013). If we go by the assumption that each household owns one well, our surveyed population covered only 3% of the well users. Nearly 412,000 people (79% of the total NL population) are served by the public water system, and around 15% of the serviced population are at risk of high disinfection by-products exposure (GoNL, 2016). It is important to note that cancer prevalence rates in the communities, exposed to either high THMs or high HAAs, were not significantly higher than the low-risk population (Table 3). It was the first evidence showing a high cancer prevalence rate in the population living near the golf course.
To the best of our knowledge, this is the first of this kind of population-based study in Canada that has tested hypothesis of spatial associations between exposure to environmental carcinogens and a higher prevalence of cancers. The major strength of the study is its wide population coverage (both rural and urban) and understanding of spatial distributions of potentially high-risk populations. While the studies of ultraviolet rays and disinfection by-products have covered almost the length and breadth of NL, the study on agricultural chemicals has covered all the golf courses located within communities.
Despite widely available environmental monitoring data on ultraviolet rays and disinfection by-products, there are few public health strategies addressing population vulnerabilities in high-risk populations in NL. Due to existing regulatory mechanisms, regardless of efforts by a rural physician to address the arsenic contamination of private wells, there is no effective mitigation strategy in NL (Greenham, 2018). Higher prevalence of cancers (specific to agricultural chemicals exposure) in the population living close to nine golf courses indicates significant association. There are 2,300 golf courses across Canada, and prohibition of cosmetic use of agriculture chemicals on the golf course is a contentious issue (Golf Canada, 2015; NGCOA). Other Canadian provinces currently do not have any provincial regulation controlling use of harmful agriculture chemicals for golf courses (CNLA). Hence, many Canadians who live close to a golf course are vulnerable to cancers and urgently need proper risk assessment.
Cancers are not attributable to a single cause, and there may be cumulative exposure to other risk factors. Therefore, to prove causal relations, future research should focus on testing biomarkers, analyzing the body burden of environmental carcinogens, examining genetic damage pertaining to specific environmental carcinogens and interviewing cancer survivors to explore other risk factors/confounders/effect modifiers relevant to particular cancers such as the duration of exposure and residence, demography, smoking, occupation, economic status, ethnicity, family history of cancer, diet and water consumption patterns, and co-exposure to other carcinogens (Madia et al., 2019).
A well-planned strategy of combining regulation (mandatory testing of private wells, improvement of public water treatment, banning of the use of carcinogenic agricultural chemicals at golf course), health promotion (application of sunscreen before outdoor activities in summer, low-cost water filters for arsenic, and disinfection by-products and environmentally friendly turf-care), and public awareness may effectively protect the high-risk population from further exposure (CCME, 2007; Hirst et al., 2012; Thomson et al., 2019).
The study has some limitations. First, the NLCCR data did not have any information on other risk factors such as smoking. Therefore, our analysis assumed that independent risk factors in both the high-risk and low-risk populations were the same. Second, the NLCCR data did not include any potential confounders/effect modifiers (mentioned above). Hence, our analysis was limited to spatial association only. We recommend the regional health authorities collect information on exposure to environmental risk factors relevant to any type of cancer while examining the patients and to incorporate the information to the existing electronic database. In this regard, proper orientation for the physicians are also needed to update knowledge on potential environmental carcinogens present in NL.
The project was supported by the Seed, Bridge, and Multidisciplinary Fund, Memorial University of Newfoundland (2018). The research was approved by the Health Research Ethics Board (HREB) (#2018:193) and the Research Proposals Approval Committee (RPAC) of Eastern Health, NL (dated 23 October 2018).
Conflicts of Interest
Andreotti G., Freeman L. E., Hou L., Coble J., Rusiecki J., Hoppin J. A., et al. 2009 Agricultural pesticide use and pancreatic cancer risk in the Agricultural Health Study Cohort Int J Cancer 124 10 2495 -2500
Bailey H. D., Infante-Rivard C., Metayer C., Clavel J., Lightfoot T., Kaatsch P., et al. 2015 Home pesticide exposures and risk of childhood leukemia: Findings from the childhood leukemia international consortium Int J Cancer 137 11 2644 -2663
Bernardi N., Gentile N., Mañas F., Méndez Á., Gorla N., and Aiassa D. 2015 Assessment of the level of damage to the genetic material of children exposed to pesticides in the province of Córdoba Arch Argent Pediatr 113 2 126 -131
Bull R. J., Reckhow D. A., Li X., Humpage A. R., Joll C., and Hrudey S. E. 2011 Potential carcinogenic hazards of non-regulated disinfection by-products: Haloquinones, halo-cyclopentene and cyclohexene derivatives, N-halamines, halonitriles, and heterocyclic amines Toxicology 286 1–3 1 -19
Fioletov V. E., Kimlin M. G., Krotkov N., McArthur L. J. B., Kerr J. B., Wardle D. I., et al. 2004 UV index climatology over the United States and Canada from ground-based and satellite estimates J Geophysical Res 109 1 -13
Gallagher R. P., Bajdik C. D., Fincham S., Hill G. B., Keefe A. R., Coldman A., and McLean D. I. 1996 Chemical exposures, medical history, and risk of squamous and basal cell carcinoma of the skin Cancer Epidemiol Biomark Prev 5 419 -424
Lang J. J., Alam S., Cahill L. E., Drucker A. M., Gotay C., Kayibanda J. F., et al. 2018 Global burden of disease study trends for Canada from 1990 to 2016 CMAJ 190 44 E1296–E1304 Available at: http://www.cmaj.ca/content/190/44/E1296 [accessed 25 January 2020]. doi: 10.1503/cmaj.180698
Melnick R. L., Nyska A., Foster P. M., Roycroft J. H., and Kissling G. E. 2006 Toxicity and carcinogenicity of the water disinfection byproduct, dibromoacetic acid, in rats and mice Toxicology 230 2–3 126 -36
NGCOA. Pesticides: Golf and the environment – A briefing report for Ontario M.P.P.s. National Golf Course Owners Association Canada. (undated). Available at: https://www.ngcoa.ca/chapters/1332/pesticides [accessed 25 January 2020].
Nieuwenhuijsen M. J., Grellier J., Smith R., Iszatt N., Bennett J., Best N., et al. 2009 The epidemiology and possible mechanisms of disinfection by-products in drinking water Philos Trans A Math Phys Eng Sci 367 1904 4043 -476
Presutti R., Harris S. A., Kachuri L., Spinelli J., Pahwa M., Blair A., et al. 2016 Pesticide exposures and the risk of multiple myeloma in men: An analysis of the North American Pooled Project Int J Cancer 139 8 1703 -1714
Rahman M. B., Cowie C., Driscoll T., Summerhayes R. J., Armstrong B. K., and Clements M. S. 2014 Colon and rectal cancer incidence and water trihalomethane concentrations in New South Wales, Australia BMC Cancer 14 445
Roche S. M., Jones-Bitton A., Majowicz S. E., Pintar K. D., and Allison D. 2013 Investigating public perceptions and knowledge translation priorities to improve water safety for residents with private water supplies: A cross-sectional study in Newfoundland and Labrador BMC Public Health 13 1225
Salas L. A., Cantor K. P., Tardon A., Serra C., Carrato A., Garcia-Closas R., et al. 2013 Biological and statistical approaches for modeling exposure to specific trihalomethanes and bladder cancer risk Am J Epidemiol 178 4 652 -660
Shah H. K., Bhat M. A., Sharma T., Banerjee B. D., and Guleria K. 2018 Delineating potential transcriptomic association with organochlorine pesticides in the etiology of epithelial ovarian cancer Open Biochem J 12 16 -28
Thomson K. K., Rahman A., Cooper T. J., and Sarkar A. 2019 Exploring relevance, public perceptions, and business models for establishment of private well water quality monitoring service Int J Health Plan Manage 34 2 e1098 -e1118
VoPham T., Wilson J. P., Ruddell D., Rashed T., Brooks M. M., Yuan J. M., et al. 2015 Linking pesticides and human health: A geographic information system (GIS) and Landsat remote sensing method to estimate agricultural pesticide exposure Appl Geogr 62 171 -181
Ward M. H., Lubin J., Giglierano J., Colt J. S., Wolter C., Bekiroglu N., et al. 2006 Proximity to crops and residential exposure to agricultural herbicides in Iowa Environ Health Perspect 114 6 893 -897
Arifur Rahman, Atanu Sarkar, Jinka Sathya, and Farah McCrate. 2020. Additional burden of cancers due to environmental carcinogens in Newfoundland and Labrador: a spatial analysis. Environmental Health Review.
63(3): 77-86. https://doi.org/10.5864/d2020-020
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.