Clean Fuels to Reduce Household Air Pollution and Improve Health. Still Hoping to Answer Why and How

The burning of biomass fuels, such as wood, animal dung, and dried crops, for cooking and heating has remained a mainstay energy source for centuries. Despite advances in energy delivery worldwide, access to clean fuels (i.e., natural gas, liquefied petroleum gas, electricity, solar, and biogas) is still lacking in many resource-poor settings of low- and middle-income countries in Africa and Southeast Asia (1). Moreover, it is just in the last few decades that we have gained a more nuanced understanding of the detrimental health effects of exposure to high levels of household air pollution (HAP) from burning of biomass fuels and our current gaps in knowledge (2). Multiple observational studies have shown that long-term exposure to high levels of HAP can lead to a variety of diseases across the lifespan, including chronic obstructive pulmonary disease, hypertension, child pneumonia, and low birth weight. Although we can unanimously agree that smoke is bad and that lower levels of HAP would be better for human health, making practical advances to reduce HAP has been far from simple. With this, the level of exposure reduction necessary to observe health benefits is unclear. Improved stoves, which use kerosene or biomass fuel with improved ventilation, still have high levels of HAP compared with clean fuel options. Does the improved stove model provide enough HAP reduction, or do we need to shift our focus exclusively to clean fuel options? Ongoing work must focus on evaluating the differences between improved and clean fuel options and directly correlating exposure levels to health outcomes. Work by Alexander and colleagues (pp. 1629–1639) reported in this issue of the Journal is a step in a new direction toward use of clean fuel for near elimination of HAP (3).

Alexander and colleagues present the first randomized intervention field trial investigating the effects of ethanol stove and fuel distribution, a clean cooking fuel alternative, on maternal perinatal blood pressure in Nigeria (3). HAP is a major risk factor for hypertension (4), which, during pregnancy, is well recognized to be associated with preterm birth and higher perinatal mortality (5). The investigators enrolled 324 pregnant women who used firewood or kerosene for cooking and randomized participants to either receive an ethanol stove or continue with their usual cooking practices. The authors found a 2.8–mm Hg reduction in diastolic blood pressure and a 4.5% reduction in the prevalence of systemic hypertension among pregnant women randomized to ethanol stoves when compared with those who were randomized to the control arm. They also performed subgroup analyses evaluating the results of kerosene or firewood users only against ethanol users and found significant improvement in blood pressure when kerosene users switched to ethanol but not when firewood users switched to ethanol. These discrepancies could be due to the fact that the study was not powered to see differences in subgroups or that subgroup analyses that break randomization are no longer causal and may be affected by unmeasured confounders. This may explain why blood pressure was reduced among pregnant women using kerosene but not reduced among those using firewood. Alternatively, as the authors speculated, this difference may highlight the potential that kerosene smoke is more pathogenic than firewood smoke. Kerosene is a similar petroleum distillate to diesel, and kerosene smoke is likely to be more toxic than biomass smoke.

Alexander and colleagues demonstrate that a clean fuel intervention can result in lower blood pressure outcomes in pregnant women (3). These findings are consistent with other cookstove studies that have found lower blood pressure outcomes after the implementation of improved stoves. One of the earliest intervention trials for evaluating health outcomes of reduced HAP, the RESPIRE (Randomized Exposure Study of Pollution Indoors and Respiratory Effects) trial, enrolled 534 households and evaluated blood pressure in older nonpregnant women in Guatemala. The authors found lower systolic and diastolic blood pressure in subjects randomized to biomass cookstoves with improved ventilation when compared with traditional biomass fuel–burning stoves (6). An observational study by Quinn and colleagues found an association between diastolic blood pressure and carbon monoxide exposure in pregnant women in Ghana (7). Two studies in Puno, Peru found higher blood pressure and increased markers of atherosclerosis in rural participants who use biomass fuels daily when compared with urban participants who used clean fuels, after adjusting for age, sex, body mass index, and insulin resistance and blood cholesterol markers (8, 9).

Although the results of the study by Alexander and colleagues are encouraging, it still leaves us with many questions (3). First, the clinical significance of an effect on diastolic but not systolic blood pressure is unclear. Although consistent with what Quinn and colleagues found in their observational study (7), their findings vary from what we have learned from ambient air pollution research. The two metaanalyses that examined the association between ambient air exposure and blood pressure in pregnant and nonpregnant women found that both systolic and diastolic blood pressure were affected (10, 11). The Sister Study evaluated 43,629 women and found ambient air pollution was associated with higher systolic but not diastolic blood pressure (12). Second, these findings have yet to be reproduced on a larger scale. Although no large clean fuel interventions have been published, several recent large randomized field trials evaluating improved biomass fuel stoves have failed to demonstrate improved clinical outcomes related to reductions in HAP. Mortimer and colleagues did not find a difference in the incidence of acute lower respiratory tract infections in 10,750 children from 8,626 households randomized to receive biomass-fueled cookstoves with improved ventilation when compared with a traditional open-fire cookstove (13). Similar negative findings were seen in a randomized cookstove trial in Rwanda looking at cases of acute lower respiratory tract infection in 566 households in rural areas (14) and a trial in Nepal looking at infant birth weight in 5,254 newborns from 3,376 households (15). With the use of improved biomass-burning cookstoves instead of clean fuel–burning cookstoves (like liquefied petroleum gas or ethanol) for the intervention, none of these studies reduced HAP to levels we believe are necessary to see a significant impact on clinical outcomes. With the use of ethanol stoves, Alexander and colleagues are taking the necessary step to focus on clean fuel alternatives (3).

Still, notably absent from this trial is the assessment of exposure to HAP. This additional information would serve to reduce misclassification of exposure regarding the intervention and further validate the role of HAP reduction on health outcomes. This is especially important to better understand the discrepancy in the findings of Alexander and colleagues when kerosene and firewood were evaluated independently against ethanol (3). Moreover, ambient and personal exposure data would allow us to further differentiate what components of HAP are associated with disease and identify an exposure–response relationship between pollutant concentrations and clinical outcomes. We need more data to support our current understanding of the exposure–response curve for HAP and clinical outcomes so we can finally say how much pollution is too much (3).

In addition, on the basis of what we know from a long history studying cigarette smoke exposure and ambient air pollution, there is a complex gene-by-environment interplay that further complicates how we can interpret the findings from field intervention trials (16, 17). To better define exactly who is at risk, we need carefully designed studies powered to evaluate exposure response, genetic interactions, and disease mechanisms (18, 19). The well-executed study by Alexander and colleagues has set the stage for larger studies that can build on these findings with both exposure data and opportunities for further mechanistic investigation (3). There are a number of studies on the horizon that are designed to take these next steps. Multiple studies supported by the Global Alliance for Clean Cookstoves are targeting cardiopulmonary outcomes and designed to determine personal exposure as it is related to clinical outcomes (20). The Household Air Pollution Investigation Network is a National Institutes of Health initiative to do this on a larger scale (21). This network plans to randomize 3,200 households in four countries to a liquefied petroleum gas cookstove intervention versus control and to evaluate personal exposure levels as they relate to various outcomes, including infant birth weight, incidence of acute lower respiratory tract infection, and cardiovascular health. These larger studies can help us take the next step in determining if clean fuel interventions are able to alleviate the global disease burden of HAP.