Investigating Associations Between Hazardous Environmental Exposures and Ovarian and Endometrial Cancer Risk
Open Access Deposited1.70, 95%CI=1.16-2.48) and Pennsylvania (HR=1.15, 95%CI=1.02-1.29), and a more attenuated NO2 association (HR=1.57, 95%CI=1.11-2.22 and HR=1.10, 95%CI=0.94-1.28, respectively). Associations with endometrial cancer were also suggested in the highest quintile of neighborhood deprivation (HR PM2.5=1.05
p-het=0.07) in association with the top tertile of the benzene AEI at 5km. DiscussionCollectively, this research helps elucidate the role that specific carcinogenic environmental pollutants may play in the development of endometrial and ovarian cancers. We discovered several gaps in the research of these relationships to-date, including relatively fewer studies of associations with endometrial cancer, an interesting observation given that the incidence of this disease has been increasing in the U.S. and globally. There was also a dearth of studies conducted outside the U.S. and limited studies of exposures other than outdoor air pollution or dietary factors. Although our meta-analyses were imprecise due to the limited number of studies, we found evidence of a positive relationship between PM2.5 and ovarian, and the relationship was stronger among studies that mutually adjusted for co-exposure to other pollutants. In risk analyses within the large prospective NIH AARP cohort, we demonstrated modest relationships between widespread ambient air pollutants PM2.5 and NO2 and the risk of ovarian and endometrial cancers for select geographies, neighborhood characteristics, and histologic subtypes. Our novel study of industrial emissions represents some of the only evaluation of the risk of these cancers at low levels of environmental exposure to a suite of known carcinogens from the occupational literature. The findings provide compelling new evidence that emissions of benzene may be associated with increased risk of both ovarian and endometrial cancers for people living near industrial sources. Expanding future research to include more geographic areas and a greater number of cancer cases would help clarify these relationships.
endometrial models only). For Aim 3, we pooled data from two large prospective cohorts in the U.S with geocoded residential addresses (the NIH-AARP study and the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial cohort
Background Endometrial and ovarian cancers are the number one and number two most diagnosed gynecologic cancers among women in the U.S., respectively. Research on these malignancies has so far primarily focused on modifiable risk factors, including body mass index and unopposed estrogen. Although there is biologic plausibility that environmental exposures may play a role in the incidence of both cancers, particularly through endocrine disrupting pathways, environmental risk factors for both tumor sites are greatly understudied. Methods This dissertation consists of three separate aims
p-trend=0.07). The positive association remained at 5km (HR Q4 vs. non-exposed=1.28, 95%CI=1.02-1.61
a systematic review and meta-analysis of epidemiologic studies that evaluated the association of non-lifestyle environmental exposures with endometrial and ovarian cancer incidence, mortality and survival (Aim 1)
p-trend=0.11) and the magnitude of association weakened with increasing distance between the residence and the source (5km HR Q4=1.39, 95%CI=1.03-1.88
58%), 30 (24%) focused on endometrial cancer only, and 23 (18%) evaluated both tumor sites. Cohort studies were the most commonly employed design (N=53
p-trend =0.02
p-trend =0.14) and a stronger association for non-epithelial tumors (HR T3 vs. non-exposed=1.72, 95%CI=1.06-2.77
p-trend=0.48). Ovarian cancer was also elevated in the highest emissions category of benzene at 3km (HR Q4 vs. non-exposed=1.45, 95%CI=0.92-2.28
17%) and dietary intakes of various factors (N=21
p-heterogeneity=0.33). For ovarian cancer, we found an elevated risk of serous tumors (HR T3 vs. non-exposed=1.31, 95%CI=0.89-1.92
and a pooled cohort analysis of the associations between carcinogenic industrial emissions and risk of these two malignancies (Aim 3). For Aim 1, we conducted a systematic literature search of five scientific databases for original epidemiologic studies published between January 1980 and October 2024 in English, developed inclusion and exclusion criteria, and presented descriptive syntheses of the final included studies. We also conducted a risk of bias assessment and meta-analysis of studies evaluating the association between PM2.5 and ovarian and endometrial cancer, the most commonly evaluated relationship in the literature. For the Aim 2 analysis, we used nationwide spatiotemporal models to estimate residential concentrations of PM2.5 and nitrogen dioxide (1980-2017) for female participants of the prospective NIH-AARP Diet and Health Study (N=140,457 enrolled 1995-1996), with follow-up through 2018. We used Cox proportional hazards regression to estimate hazard ratios (HR) and 95% confidence intervals (CIs) for associations between time-varying, 5-year average PM2.5 and NO2 concentrations (per 5µg/m3 and ppb
p-trend=0.07) and HR NO2=1.06, 95%CI=1.00-1.12
p-trend=0.03) and was evident but weaker at 10km (HR Q4 vs. non-exposed=1.09
42%), while 37 (30%) were ecologic, 33 (26%) were case-control (including four nested case-control studies), and 2 (2%) were cross-sectional studies. The environmental exposures most represented among studies included outdoor air pollution (N=21
endometrial models only), smoking status and intensity, and neighborhood-level socioeconomic deprivation (via the Yost Index). We evaluated co-pollutant adjustment and potential effect modification by state (per state-specific interquartile ranges), neighborhood deprivation, and body mass index (BMI
p-trend=0.04 and 10km HR Q4=1.32, 95%CI=1.07-1.63
p-trend=0.32). The positive association was also apparent but much less precise at 2km, with 18 exposed cases in Q4 (HR Q4 vs. non-exposed=1.21, 95%CI= 0.72-2.02
p-trend=0.02). We found positive, albeit imprecise, associations with ovarian cancer in Detroit, Michigan (PM2.5 HR=1.42, 95%CI=0.92-2.18 and NO2 HR=1.72, 95%CI=0.95-3.13). We observed some limited evidence of heterogeneity in pollutant associations across histologic subtypes. The impact of co-pollutant adjustment varied across state-specific associations, and relationships did not vary across strata of BMI. In Aim 3 analyses, endometrial cancer risk was elevated in association with the highest AEI category of benzene at 3km (HR Q4 vs. non-exposed=1.36, 95%CI=0.98-1.89
95%CI=0.93-1.28
15%), and pesticides (N=13
p-trend=0.41
p-trend=0.12) whereas no association was evident for endometrioid tumors (HR T3 vs non-exposed=0.91, 95%CI=0.72-1.14
10%). Most studies were conducted in North America (and in the U.S., specifically) and were published after the year 2009. We evaluated the risk of bias for six studies of the association between PM2.5 and either of these cancers for possible inclusion in meta-analyses, and all six were rated “low” or “probably low” risk of overall study bias. Most of these studies used a similar approach to exposure assessment based on regulatory monitoring data and were predominantly conducted within the U.S. The meta-association for PM2.5 and ovarian cancer based on four studies was positive (HRmeta per 5 µg/m3=1.14, 95%CI=1.06-1.23), and marginally weaker when limiting to U.S.-based studies (HRmeta=1.11, 95%CI=1.00-1.22). Among studies that mutually adjusted for co-pollutants, the meta-estimate was strengthened (HR per 5 µg/m3=1.19, 95%CI=1.05-1.35). We found a similar positive meta-association for endometrial cancer, also based on four studies (HRmeta=1.19, 95%CI=1.07-1.33), although we found that limiting to studies that implemented co-pollutant adjustment reduced the effect to the null (HRmeta per 5 µg/m3=0.89, 95%CI=0.55, 1.42). In Aim 2, we found no relationships between PM2.5 or NO2 and risk of either cancer overall. However, state-specific analyses yielded some positive relationships, including for PM2.5 and endometrial cancer among participants in Atlanta, Georgia (HR per IQR increase
N=167,617). We used the U.S. Environmental Protection Agency’s Toxics Release Inventory of industrial point sources to estimate historical exposure to airborne emissions (1987-1995) for chemicals and metals classified as known human carcinogens (Group 1) by the International Agency for Research on Cancer (IARC). For each carcinogen, we constructed inverse distance-and wind direction-weighted average emissions index (AEI) exposure metrics within 1,2, 3, 5, and 10 km of participants’ enrollment address. We used Cox regression to estimate HRs and 95% CIs for associations between tertiles or quartiles of exposure (compared to non-exposed) for each carcinogen with incident ovarian and endometrial cancer overall, by histologic subtype, and across categories of BMI (under/normal weight, overweight, and obese). Models were adjusted for state, age, race and ethnicity, BMI (endometrial models only), neighborhood deprivation, and smoking status and intensity. We also mutually adjusted for correlated exposures across all AEIs (Spearman’s ρ≥0.3) and evaluated the impact of adjustment for key reproductive risk factors. Results The systematic review in Aim 1 yielded 125 studies meeting inclusion criteria, of which over half were focused on investigation of ovarian cancer only (N=72
p-trend =0.01)
95%CI=0.90-1.22
) and incident endometrial (n=3,088) and ovarian (n=1,582) cancers, overall and by histologic subtype. Models were adjusted for state, age, race and ethnicity, body mass index (BMI
17%), followed by ultraviolet (UV) light (N=19
a cohort analysis evaluating the association between common outdoor air pollutants fine particulate matter (PM2.5, <2.5µm) and nitrogen dioxide (NO2) exposure and the risk of endometrial and ovarian cancers (Aim 2)
we were unable to evaluate associations at 2km for either cancer site. In analyses by histologic subtype, we found an elevated risk for non-endometrioid endometrial tumors in association with the asbestos AEI at 10km (HR T3 vs non-exposed=1.31, 95%CI=0.90-1.90
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.