The complexities of PM_2.5, greenspace, and childhood cancer

Most studies of outdoor air pollution and cancer have observed associations between exposure to fine particulate matter (PM_2.5), an aerosol composed of small particles less than 2.5 µm in diameter, and risk of lung cancer. Evidence of this relationship from epidemiologic studies played a major role in the classification by the International Agency for Research on Cancer of outdoor air pollution, and PM_2.5 specifically, as carcinogenic to humans more than 10 years ago (1). This review, however, included few high-quality epidemiologic studies of childhood cancer. Biologic plausibility exists for pediatric cancers, as air pollutants can cross the placenta (2,3) and exposure can promote inflammation, oxidative stress, and cellular and DNA damage (1), all of which are tenable mechanisms. However, the epidemiologic literature on outdoor air pollution and childhood cancer remains limited (4). Most investigations have focused on the frequently diagnosed malignancies, including leukemias and central nervous system (CNS) tumors, and employed a range of exposure assessment approaches, including use of proxies like traffic density or directly measured or modeled concentrations of criteria air pollutants like PM_2.5. Results are mixed, with the most consistently observed association an increased risk of acute lymphoblastic leukemia (ALL) (4).

In this issue of the Journal, Williams et al. (5) reveal findings from a large study in Texas, demonstrating associations between ambient concentrations of PM_2.5 and several types of childhood cancers. Leveraging statewide birth and cancer registries over a 16-year period, a major contribution of this research was its broadening the scope from prior studies to evaluate less common cancers. Key observations include that estimated PM_2.5 exposure was associated with increased risk of ALL, retinoblastoma, ependymoma, and non-Hodgkin lymphoma, consistent with several prior investigations. They also report increased risk of 2 rare malignancies—thyroid carcinoma and Hodgkin lymphoma—which comprise roughly 6% (6) and 3% (7) of all childhood cancers in the United States, respectively, and have been little studied in relation to air pollution exposure. The ability to discern these latter associations is a testament to the size of the study, which is among the largest ever conducted in the United States. The study’s size also enabled evaluation of age-specific analyses that point to stronger associations for some cancers diagnosed before age 5 years, including Burkitt lymphoma, non-Hodgkin lymphoma, and osteosarcoma, all of which are less commonly identified in younger children. These findings could suggest that PM_2.5 exposures in early life may be most relevant to the development of these pediatric cancers. Because exposures were assessed on the basis of residence at birth, it will be useful for future studies to evaluate exposures for other time periods and residential locations.

Underlying this work was the premise that greenspace may mitigate exposure to air pollutants (8-10), thus theoretically reducing the risk of developing cancers associated with PM_2.5. Indeed, Williams et al. (5) observed independent, inverse relationships between greenspace exposures estimated from satellite imagery and 2 CNS tumors, including ependymoma, which has few known risk factors, and medulloblastoma, for which genetic predisposition is established (11). However, the opposite was also true: greenspace was positively associated with risk of several other cancers. Additionally, the authors found evidence of interaction between these 2 exposures on the risk of all pediatric cancers combined but not in the expected direction (ie, the risk associated with PM_2.5 was greater for children with high greenspace exposure). There was limited evidence that greenspace effects were stronger for young children, leaving the question of whether early life greenspace exposures are particularly relevant, as appeared to be the case for PM_2.5. Apart from potentially reducing air pollution exposure, it is unclear how other beneficial effects of greenspace thought to be relevant to cancer, such as reduced stress and increased physical activity, might influence the risk of childhood malignancies specifically. The literature on greenspace and adult cancers is nascent and has not demonstrated a consistent protective effect on either incidence or survival for any single malignancy (12), and 2 childhood cancer studies were conducted outside the United States (13,14). The lack of clear signal for greenspace, both independently and jointly with PM_2.5, likely reflects the complexity of how exposure to these factors could influence cancer risk. Despite the findings, evaluation of this potential interaction is a novel addition of the work. It demonstrates the authors’ recognition of the need to consider a broader context for the environment experienced by children, moving away from the conventional framework of evaluating only singular, adverse exposures.

Interesting patterns also emerged in selected subgroup analyses, including that the positive association between PM_2.5 and Hodgkin lymphoma was most apparent among those with high neighborhood-level socioeconomic status (SES). These findings are intriguing given the positive correlation between incidence of childhood Hodgkin lymphoma and SES (15) and in light of the fact that in the United States, a greater burden of exposure to air pollutants has consistently been found among lower SES populations (16). These increased risks may be due to PM_2.5 exposures but might also be driven by residual confounding. A positive relationship between greenspace and ALL was likewise observed weakly overall but was stronger among the high SES group; this is also challenging to interpret, because greenspace exposures also correlate with SES (17) and the lack of clarity around pathways by which they influence cancer development. These findings could also be spurious, and the authors are appropriately cautious with interpretations. At the least, these results again underscore the complexity of interactions between factors in the external environment on childhood cancer risk and warrant additional exploration.

Texas has one of the largest Hispanic and Latino populations in the United States (18,19), as reflected in this analysis, where 40% of the study population resided in a Hispanic enclave. In 2014-2018, the incidence rate for all cancers combined among US children aged 0-14 years was 17.8 cases per 100 000 persons, with rates stable among non-Hispanic Black children but increasing for Hispanic and non-Hispanic White children (20). The incidence of leukemia, the most commonly diagnosed malignancy in childhood, is highest among Hispanic and Latino children (20-23). Factors driving the excess of cancer in Hispanic children are largely unknown, and only 1 prior US study, in California, evaluated but did not observe differences in childhood cancer associations with air pollutants by Hispanic ethnicity (24). Williams et al. (5) found that the positive PM_2.5 association with pediatric ependymoma was apparent only among those who resided in a Hispanic enclave. This could be because of greater burden of PM_2.5 exposures in this group but could also implicate unique, unmeasured hazards and sociostructural factors that may act independently or jointly to contribute to cancer risk.

Other potential avenues for future work emerged from this study. The authors did not assess the chemical composition of PM_2.5, although they acknowledge the need. Heterogeneity in PM_2.5 composition across the United States and equivocal findings for childhood cancer studies make a case for evaluating PM_2.5 constituency to further the understanding of etiologic relationships. Major components of PM_2.5 vary geographically because of its diverse sources, and many show adverse cellular effects relevant to carcinogenesis (25). However, differential cancer associations by PM_2.5 constituents have rarely been evaluated for childhood cancers (26). Related, the overall air pollution mixture may better represent the actual exposure experience. Ambient PM_2.5 concentrations often correlate with other emissions like nitrogen dioxide, which imparts only weak mutagenic or genotoxic effects (27,28) but has been associated with childhood cancers, potentially because it serves as a marker for other, known carcinogens in traffic pollution, like benzene. It will be informative to formally assess the relative mixture of air pollutants in future evaluations of pediatric cancer associations.

Incidence rates of the predominant cancers diagnosed in childhood, including leukemia, brain, and other CNS cancers, have been increasing by approximately 1% per year in the United States over the last 2 decades (20), highlighting the need to extend research into their causes. Both genetic and environmental contributors are evident, but genetics explain only a small proportion, and so far, only a handful of environmental factors, including some parental occupational exposures and ionizing radiation, have been consistently linked with risk (11,29-31). Air pollution exposure is ubiquitous and for the most part involuntary, especially for children, further motivating public health concerns. The US Environmental Protection Agency recently lowered the annual standard for PM_2.5 from 12 µg/m³ to 9 µg/m³ in recognition of the body of research showing increased health risks at lower levels of exposure, especially for vulnerable population subgroups (32). One contribution from Williams et al. (5) is the demonstration of positive associations with PM_2.5 even after 2006, despite lower concentrations compared with earlier in the study period. This finding suggests that the detrimental effect of PM_2.5 is still observed at relatively low levels of exposure, a finding in sync with emerging research on adult cancers (33).

With robust analyses and a diverse study population, the work by Williams et al. (5) sheds new light on associations between PM_2.5 exposures and childhood cancer, especially for some rarer types. It offers findings for greenspace exposures that encourage further consideration of the joint impact of multiple environmental factors, both adverse and protective, on childhood cancer risk. More high-quality studies of this size and scope would continue to enhance our understanding of the burden of pediatric cancers driven by environmental factors.

Data availability

No new data were generated or analyzed for this editorial.

Author contributions

Rena R. Jones, PhD, MS (Conceptualization; Writing—original draft; Writing—review & editing).

Funding

No funding was used for this editorial. The author is supported by the Intramural Research Program of the Division of Cancer Epidemiology and Genetics, National Cancer Institute.

Conflicts of interest

Rena R. Jones has no disclosures.

Acknowledgements

The funder had no role in the writing of this editorial. The author thanks Drs. Debra Silverman and Mary Ward for their early review and feedback on the original draft of this manuscript. The opinions expressed herein are the opinion of the author and do not reflect those of the National Institutes of Health, the Department of Health and Human Services, or the US government.