Are Randomized Trials Necessary to Advance Epidemiologic Research on Household Air Pollution?

Facing the Realities of Pragmatic Design Choices in Environmental Health Studies: Experiences from the Household Air Pollution Intervention Network Trial

Objective: Household Air Pollution Intervention Network (HAPIN) investigators tested a complex, non-pharmacological intervention in four low- and middle-income countries as a strategy to mitigate household air pollution and improve health outcomes across the lifespan. Intervention households received a liquefied petroleum gas (LPG) stove, continuous fuel delivery and regular behavioral reinforcements for 18 months, whereas controls were asked to continue with usual cooking practices. While HAPIN was designed as an explanatory trial to test the efficacy of the intervention on four primary outcomes, it introduced several pragmatic aspects in its design and conduct that resemble real-life conditions. We surveyed HAPIN investigators and asked them to rank what aspects of the design and conduct they considered were more pragmatic than explanatory. Methods: We used the revised Pragmatic Explanatory Continuum Indicator Summary (PRECIS-2) to survey investigators on the degree of pragmatism in nine domains of trial design and conduct using a five-point Likert rank scale from very explanatory (1) to very pragmatic (5). We invited 103 investigators. Participants were given educational material on PRECIS-2, including presentations, papers and examples that described the use and implementation of PRECIS-2. Results: Thirty-five investigators (mean age 42 years, 51% female) participated in the survey. Overall, only 17% ranked all domains as very explanatory, with an average (±SD) rank of 3.2 ± 1.4 across domains. Fewer than 20% of investigators ranked eligibility, recruitment or setting as very explanatory. In contrast, ≥50% of investigators ranked the trial organization, delivery and adherence of the intervention and follow-up as very/rather explanatory whereas ≤17% ranked them as rather/very pragmatic. Finally, <25% of investigators ranked the relevance of outcomes to participants and analysis as very/rather explanatory whereas ≥50% ranked then as rather/very pragmatic. In-country partners were more likely to rank domains as pragmatic when compared to investigators working in central coordination (average rank 3.2 vs. 2.8, respectively; Wilcoxon rank-sum p < 0.001). Conclusion: HAPIN investigators did not consider their efficacy trial to be rather/very explanatory and reported that some aspects of the design and conduct were executed under real-world conditions; however, they also did not consider the trial to be overly pragmatic. Our analysis underscores the importance of using standardized tools such as PRECIS-2 to guide early discussions among investigators in the design of environmental health trials attempting to measure efficacy.

Effectiveness of Gas and Chimney Biomass Stoves for Reducing Household Air Pollution Pregnancy Exposure in Guatemala: Sociodemographic Effect Modifiers

Household air pollution (HAP) due to solid fuel use during pregnancy is associated with adverse birth outcomes. The real-life effectiveness of clean cooking interventions has been disappointing overall yet variable, but the sociodemographic determinants are not well described. We measured personal 24-h PM_2.5 (particulate matter <2.5 µm in aerodynamic diameter) thrice in pregnant women (n = 218) gravimetrically with Teflon filter, impactor, and personal pump setups. To estimate the effectiveness of owning chimney and liquefied petroleum gas (LPG) stoves (i.e., proportion of PM_2.5 exposure that would be prevented) and to predict subject-specific typical exposures, we used linear mixed-effects models with log (PM_2.5) as dependent variable and random intercept for subject. Median (IQR) personal PM_2.5 in µg/m³ was 148 (90-249) for open fire, 78 (51-125) for chimney stove, and 55 (34-79) for LPG stoves. Adjusted effectiveness of LPG stoves was greater in women with ≥6 years of education (49% (95% CI: 34, 60)) versus <6 years (26% (95% CI: 5, 42)). In contrast, chimney stove adjusted effectiveness was greater in women with <6 years of education (50% (95% CI: 38, 60)), rural residence (46% (95% CI: 34, 55)) and lowest SES (socio-economic status) quartile (59% (95% CI: 45, 70)) than ≥6 years education (16% (95% CI: 22, 43)), urban (23% (95% CI: -164, 42)) and highest SES quartile (-44% (95% CI: -183, 27)), respectively. A minority of LPG stove owners (12%) and no chimney owner had typical exposure below World Health Organization Air Quality guidelines (35 μg/m³). Although having a cleaner stove alone typically does not lower exposure enough to protect health, understanding sociodemographic determinants of effectiveness may lead to better targeting, implementation, and adoption of interventions.

Chemical composition and source apportionment of ambient, household, and personal exposures to PM2.5 in communities using biomass stoves in rural China

Fine particulate matter (PM_2.5) has health effects that may depend on its sources and chemical composition. Few studies have quantified the composition of personal and area PM_2.5 in rural settings over the same time period. Yet, this information would shed important light on the sources influencing personal PM_2.5 exposures. This study investigated the sources and chemical composition of 40 personal exposure, 40 household, and 36 ambient PM_2.5 samples collected in the non-heating and heating seasons in rural southwestern China. Chemical analysis included black carbon (BC), water-soluble components (ions, organic carbon), elements, and organic tracers. Source apportionment was conducted using organic tracer concentrations in a Chemical Mass Balance model. Biomass burning was the largest identified PM_2.5 source contributor to household (average, SD: 48 ± 11%) and exposures (31 ± 6%) in both seasons, and ambient PM_2.5 in winter (20 ± 4%). Food cooking also contributed to household and personal PM, reaching approximately half of the biomass contributions. Secondary inorganic aerosol was the major identified source in summertime ambient PM_2.5 (32 ± 14%), but was present in all samples (summer: 10 ± 3% [household], 13 ± 6% [exposures]; winter: 18 ± 2% [ambient], 7 ± 2% [household], 8 ± 2% [exposures]). Dust concentrations and fractional contribution to total PM_2.5 were higher in summer exposure samples (7 ± 4%) than in ambient or household samples (6 ± 1% and 2 ± 1%, respectively). Indoor sources comprised up to one-fifth of ambient PM_2.5, and outdoor sources (vehicles, secondary aerosols) contributed up to 15% of household PM_2.5. While household sources were the main contributors to PM_2.5 exposures in terms of mass, inorganic components of personal exposures differed from household samples. Based on these findings, health-focused initiatives to reduce harmful PM_2.5 exposures may consider a coordinated approach to address both indoor and outdoor PM_2.5 source contributors.

The Firepower Sweep Test: A novel approach to cookstove laboratory testing

Emissions from solid-fuel cookstoves have been linked to indoor and outdoor air pollution, climate forcing, and human disease. Although task-based laboratory protocols, such as the Water Boiling Test (WBT), overestimate the ability of improved stoves to lower emissions, WBT emissions data are commonly used to benchmark cookstove performance, estimate indoor and outdoor air pollution concentrations, estimate impacts of stove intervention projects, and select stoves for large-scale control trials. Multiple-firepower testing has been proposed as an alternative to the WBT and is the basis for a new standardized protocol (ISO 19867-1:2018); however, data are needed to assess the value of this approach. In this work, we (a) developed a Firepower Sweep Test [FST], (b) compared emissions from the FST, WBT, and in-home cooking, and (c) quantified the relationship between firepower and emissions using correlation analysis and linear model selection. Twenty-three stove-fuel combinations were evaluated. The FST reproduced the range of PM_2.5 and CO emissions observed in the field, including high emissions events not typically observed under the WBT. Firepower was modestly correlated with emissions, although the relationship varied between stove-fuel combinations. Our results justify incorporating multiple-firepower testing into laboratory-based protocols but demonstrate that firepower alone cannot explain the observed variability in cookstove emissions.

Developing a Clinical Approach to Air Pollution and Cardiovascular Health

Nearly 3 billion people are exposed to household air pollution emitted from inefficient cooking and heating stoves, and almost the entire global population is exposed to detectable levels of outdoor air pollution from traffic, industry, and other sources. Over 3 million people die annually of ischemic heart disease or stroke attributed to air pollution, more than from traditional cardiac risk factors such as obesity, diabetes mellitus, or smoking. Clinicians have a role to play in reducing the burden of pollution-attributable cardiovascular disease. However, there currently exists no clear clinical approach to this problem. Here, we provide a blueprint for an evidence-based clinical approach to assessing and mitigating cardiovascular risk from exposure to air pollution. We begin with a discussion of the global burden of pollution-attributable cardiovascular disease, including a review of the mechanisms by which particulate matter air pollution leads to cardiovascular outcomes. Next, we offer a simple patient-screening tool using known risk factors for pollution exposure. We then discuss approaches to quantifying air pollution exposures and cardiovascular risk, including the development of risk maps for clinical catchment areas. We review a collection of interventions for household and outdoor air pollution, which clinicians can tailor to patients and populations at risk. Finally, we identify future research needed to quantify pollution exposures and validate clinical interventions. Overall, we demonstrate that clinicians can be empowered to mitigate the global burden of cardiovascular disease attributable to air pollution.

Assessing Exposure to Household Air Pollution: A Systematic Review and Pooled Analysis of Carbon Monoxide as a Surrogate Measure of Particulate Matter

BACKGROUND

Household air pollution from solid fuel burning is a leading contributor to disease burden globally. Fine particulate matter (PM_2.5) is thought to be responsible for many of these health impacts. A co-pollutant, carbon monoxide (CO) has been widely used as a surrogate measure of PM_2.5 in studies of household air pollution.

OBJECTIVE

The goal was to evaluate the validity of exposure to CO as a surrogate of exposure to PM_2.5 in studies of household air pollution and the consistency of the PM_2.5-CO relationship across different study settings and conditions.

METHODS

We conducted a systematic review of studies with exposure and/or cooking area PM_2.5 and CO measurements and assembled 2,048 PM_2.5 and CO measurements from a subset of studies (18 cooking area studies and 9 personal exposure studies) retained in the systematic review. We conducted pooled multivariate analyses of PM_2.5-CO associations, evaluating fuels, urbanicity, season, study, and CO methods as covariates and effect modifiers.

RESULTS

We retained 61 of 70 studies for review, representing 27 countries. Reported PM_2.5-CO correlations (r) were lower for personal exposure (range: 0.22-0.97; median=0.57) than for cooking areas (range: 0.10-0.96; median=0.71). In the pooled analyses of personal exposure and cooking area concentrations, the variation in ln(CO) explained 13% and 48% of the variation in ln(PM_2.5), respectively.

CONCLUSIONS

Our results suggest that exposure to CO is not a consistently valid surrogate measure of exposure to PM_2.5. Studies measuring CO exposure as a surrogate measure of PM exposure should conduct local validation studies for different stove/fuel types and seasons. https://doi.org/10.1289/EHP767.