Using Parametric g-Computation to Estimate the Effect of Long-Term Exposure to Air Pollution on Mortality Risk and Simulate the Benefits of Hypothetical Policies: The Canadian Community Health Survey Cohort (2005 to 2015)

The association between long-term PM2.5 exposure and risk for pancreatic cancer: an application of social informatics

There is a profound need to identify modifiable risk factors to screen and prevent pancreatic cancer. Air pollution, including fine particulate matter (PM2.5), is increasingly recognized as a risk factor for cancer. We conducted a case-control study using data from the electronic health record (EHR) of Duke University Health System, 15-year residential history, NASA satellite fine particulate matter (PM2.5), and neighborhood socioeconomic data. Using deterministic and probabilistic linkage algorithms, we linked residential history and EHR data to quantify long-term PM2.5 exposure. Logistic regression models quantified the association between a 1 interquartile range (IQR) increase in PM2.5 concentration and pancreatic cancer risk. The study included 203 cases and 5027 controls (median age of 59 years, 62% female, 26% Black). Individuals with pancreatic cancer had higher average annual exposure (9.4 μg/m3) as compared to an IQR increase in average annual PM2.5, which was associated with greater odds of pancreatic cancer (odds ratio = 1.20; 95% CI, 1.00-1.44). These findings highlight the link between elevated PM2.5 exposure and increased pancreatic cancer risk. They may inform screening strategies for high-risk populations and guide air pollution policies to mitigate exposure. This article is part of a Special Collection on Environmental Epidemiology.

Neighborhood socioeconomic disparities in cancer incidence following a hypothetical intervention to increase residential greenspace cover in the UK Biobank cohort

BACKGROUND

Higher greenspace exposure has been associated with lower risk of certain cancers. However, few studies have evaluated potential benefits of increasing population-level exposure to greenspace on cancer disparities. We estimated the impact of a hypothetical intervention to increase residential greenspace cover on neighborhood socioeconomic disparities in total, breast, colorectal, lung, and prostate cancer incidence.

METHODS

Our study included 411,787 cancer-free UK Biobank participants. Percentage of greenspace around baseline residential addresses (300m, 1000m distance buffers) was derived by combining domestic gardens and greenspace cover from the Generalized Land Use Database. We categorized neighborhood socioeconomic deprivation using the Index of Multiple Deprivation (2010). We estimated hazard ratios (HR) and 95% confidence intervals (CI) of each cancer associated with greenspace, adjusting for sociodemographic and lifestyle factors. We additionally adjusted for air pollution in supplementary analyses as we a-priori hypothesized that it was on the causal pathway between greenspace and cancer. Further, we used parametric g-computation to calculate the standardized 10-year risk of each cancer, comparing the least to most socioeconomically disadvantaged participants, both without any hypothetical greenspace intervention and under a hypothetical intervention to increase residential greenspace cover to a favorable threshold (75th percentile amongst the least socioeconomically deprived participants).

RESULTS

We documented 40,519 incident cases of cancer over 4,210,008 person-years follow-up. An interquartile range increase in greenspace cover within 300m was associated with lower incidence of total (HR 0.98; 95% CI 0.97, 1.00) and lung (HR 0.96; 95% CI 0.92, 0.99) cancer, and was suggestively associated with lower prostate and breast cancer incidence, but not colorectal cancer. Additional adjustment for fine particulate matter air pollution (PM2.5) weakened lung cancer associations but strengthened breast and prostate cancer associations (e.g., greenspace 1000m breast cancer HR 0.94; 95% CI 0.89 0.99; 1000m prostate cancer HR 0.91; 95% CI 0.86, 0.95). The hypothetical intervention to increase greenspace (300m) resulted in 1.3 fewer total cancer cases per 1000 (95% CI 1.0, 1.6) in the most compared to least deprived group, a 23% reduction in the socioeconomic disparity gap.

DISCUSSION

Higher residential greenspace cover was associated with lower total and lung cancer incidence, and suggestively associated with lower breast and prostate cancer incidence. Policies to increase residential greenspace cover may reduce the risk of certain cancers, particularly among socioeconomically disadvantaged groups.

Impact of a 10-year shift in ambient air quality on mortality in Canada: a causal analysis of multiple pollutants

BACKGROUND

The impact of past air quality improvements on health and equity at low pollution levels near the revised WHO air quality guidelines remains largely unknown. Less is known about the influence of simultaneous reductions in multiple major pollutants. Leveraging real-world improvements in air quality across Canada, we sought to directly evaluate their health benefits by quantifying the impact of a joint shift in three criteria pollutants on mortality in a national cohort.

METHODS

In this population-based cohort study, we assembled a cohort of 2·7 million adults living in Canada in 2007 who were followed up through 2016. Annual mean concentrations of fine particulate matter (PM2·5), nitrogen dioxide (NO2), and ozone (O3) were assigned to participants' residential locations. For each pollutant individually and combined, we conducted a causal analysis of the impact of the decadal shift in annual exposure from the pre-baseline level (2004-06) on the risk of non-accidental mortality using the parametric g-formula, a structural causal model. To check the robustness of our results, we conducted multiple sensitivity analyses, including exploring alternative exposure scenarios. We also evaluated differential benefits across regions and socio-demographic subgroups.

FINDINGS

Between 2007 and 2016, annual mean exposures to PM2·5 and NO2 decreased (from 7·1 μg/m3 [SD 2·3] to 5·5 μg/m3 [1·9] for PM2·5 and from 11·1 ppb [SD 6·6] to 8·0 ppb [4·9] for NO2), whereas O3 declined initially and then rebounded (from 38·6 [SD 8·3] ppb to 36·0 [6·0] ppb and then 38·1 [5·4] ppb). Compared to pre-baseline (2004-06) levels, the joint change in the pollution exposures beginning in 2007 resulted in, per million population, 70 (95% CI 29-111) fewer deaths by 2009, 416 (283-549) fewer deaths by 2012, and 609 (276-941) fewer deaths by 2016, corresponding to a -0·7% change in mortality risk over the decade. Stratified analyses showed greater beneficial impacts in men, adults aged 50 years and older, low income-earners, and residents in regions undergoing substantial air quality improvements. Had all regions experienced pollution reductions similar to the most improved region, approximately three times as many deaths would have been averted (2191 fewer deaths per million). Conversely, if the observed air quality improvements had been delayed in all regions by 3 years, there would have been 429 more deaths per million by 2016.

INTERPRETATION

In Canada, substantial health gains were associated with air quality improvements at levels near the revised WHO guidelines between 2007 and 2016, with notable heterogeneity observed across socio-demographic subgroups and regions. These findings indicate that modest declines in air pollution can considerably improve health and equity, even in low-exposure environments.

FUNDING

Health Canada.

Improved Air Quality and Asthma Incidence from School Age to Young Adulthood: A Population-based Prospective Cohort Study

Rationale: The benefits of improved air quality on asthma remain understudied. Objectives: Our aim was to investigate associations of changes in ambient air pollution with incident asthma from school age until young adulthood in an area with mostly low air pollution levels. Methods: Participants in the BAMSE (Swedish abbreviation for Children, Allergy, Environment, Stockholm, Epidemiology) birth cohort from Stockholm without asthma before the 8-year follow-up were included (N = 2,371). We estimated the association of change in individual-level air pollutant exposure (particulate matter with an aerodynamic diameter ≤ 2.5 μm [PM2.5] and ≤ 10 μm [PM10], black carbon [BC], and nitrogen oxides [NOx]) from the first year of life to the 8-year follow-up with asthma incidence from the 8-year until the 24-year follow-up. Multipollutant trajectories were identified using the group-based multivariate trajectory model. We also used parametric G-computation to quantify the asthma incidence under different hypothetical interventions regarding air pollution levels. Results: Air pollution levels at residency decreased during the period, with median reductions of 5.6% for PM2.5, 3.1% for PM10, 5.9% for BC, and 26.8% for NOx. A total of 395 incident asthma cases were identified from the 8-year until the 24-year follow-up. The odds ratio for asthma was 0.89 (95% confidence interval [CI], 0.80-0.99) for each interquartile range reduction in PM2.5 (equal to 8.1% reduction). Associations appeared less clear for PM10, BC, and NOx. Five multipollutant trajectories were identified; the largest reduction trajectory displayed the lowest odds of asthma (odds ratio, 0.55; 95% CI, 0.31-0.98) compared with the lowest reduction trajectory. If the PM2.5 exposure had not declined up to the 8-year follow-up, the hypothetical asthma incidence was estimated to have been 10.9% higher (95% CI, 0.8-20.8%). Conclusions: A decrease in PM2.5 levels during childhood was associated with a lower risk of incident asthma from school age to young adulthood in an area with relatively low air pollution levels, suggesting broad respiratory health benefits from improved air quality.

Using Parametric g-Computation for Time-to-Event Data and Distributed Lag Models to Identify Critical Exposure Windows for Preterm Birth: An Illustrative Example Using PM2.5 in a Retrospective Birth Cohort Based in Eastern Massachusetts (2011-2016)

BACKGROUND

Parametric g-computation is an attractive analytic framework to study the health effects of air pollution. Yet, the ability to explore biologically relevant exposure windows within this framework is underdeveloped.

OBJECTIVES

We outline a novel framework for how to incorporate complex lag-responses using distributed lag models (DLMs) into parametric g-computation analyses for survival data. We call this approach "g-survival-DLM" and illustrate its use examining the association between PM 2.5 during pregnancy and the risk of preterm birth (PTB).

METHODS

We applied the g-survival-DLM approach to estimate the hypothetical static intervention of reducing average PM 2.5 in each gestational week by 20% on the risk of PTB among 9,403 deliveries from Beth Israel Deaconess Medical Center, Boston, Massachusetts, 2011-2016. Daily PM 2.5 was taken from a 1 -km grid model and assigned to address at birth. Models were adjusted for sociodemographics, time trends, nitrogen dioxide, and temperature. To facilitate implementation, we provide a detailed description of the procedure and accompanying R syntax.

RESULTS

There were 762 (8.1%) PTBs in this cohort. The gestational week-specific median PM 2.5 concentration was relatively stable across pregnancy at ∼ 7 μ g / m 3 . We found that our hypothetical intervention strategy changed the cumulative risk of PTB at week 36 (i.e., the end of the preterm period) by - 0.009 (95% confidence interval: - 0.034 , 0.007) in comparison with the scenario had we not intervened, which translates to about 86 fewer PTBs in this cohort. We also observed that the critical exposure window appeared to be weeks 5-20.

DISCUSSION

We demonstrate that our g-survival-DLM approach produces easier-to-interpret, policy-relevant estimates (due to the g-computation); prevents immortal time bias (due to treating PTB as a time-to-event outcome); and allows for the exploration of critical exposure windows (due to the DLMs). In our illustrative example, we found that reducing fine particulate matter [particulate matter (PM) with aerodynamic diameter ≤ 2.5 μ m (PM 2.5 )] during gestational weeks 5-20 could potentially lower the risk of PTB. https://doi.org/10.1289/EHP13891.

Differential Participation, a Potential Cause of Spurious Associations in Observational Cohorts in Environmental Epidemiology

Differential participation in observational cohorts may lead to biased or even reversed estimates. In this article, we describe the potential for differential participation in cohorts studying the etiologic effects of long-term environmental exposures. Such cohorts are prone to differential participation because only those who survived until the start of follow-up and were healthy enough before enrollment will participate, and many environmental exposures are prevalent in the target population and connected to participation via factors such as geography or frailty. The relatively modest effect sizes of most environmental exposures also make any bias induced by differential participation particularly important to understand and account for. We discuss key points to consider for evaluating differential participation and use causal graphs to describe two example mechanisms through which differential participation can occur in health studies of long-term environmental exposures. We use a real-life example, the Canadian Community Health Survey cohort, to illustrate the non-negligible bias due to differential participation. We also demonstrate that implementing a simple washout period may reduce the bias and recover more valid results if the effect of interest is constant over time. Furthermore, we implement simulation scenarios to confirm the plausibility of the two mechanisms causing bias and the utility of the washout method. Since the existence of differential participation can be difficult to diagnose with traditional analytical approaches that calculate a summary effect estimate, we encourage researchers to systematically investigate the presence of time-varying effect estimates and potential spurious patterns (especially in initial periods in the setting of differential participation).