Air Pollution as a Climate Forcing: A Workshop

Day 4 Presentations

Health Effects of Air Pollution from Household Fuels

Kirk R. Smith*, Sumi Mehta1*, Miriam Feuz+
* Environmental Health Sciences, University of California, Berkeley, CA, U.S.A.
+ Swiss Federal Office of Public Health, Basel, Switzerland

You may download a PDF version (760 MB) of this presentation.


Scientific and policy concerns about air pollution as a health hazard were focused by the "air pollution disasters" in London, Donora, Meuse Valley, and other industrial settings of the developed world in mid-20th century, although high-levels of ambient pollution had been common in many industrial areas for centuries. Such pollution tends to be a function primarily of large industrial and, more recently, widespread vehicular sources. Household fuels play a role, for example being the chief source of the London "smog" that killed some 4000 people in 1952, but attention had only focused on their contribution to outdoor pollution. It has been only relatively recently that the health impacts of household fuel use as a source of indoor pollution has also been recognized.

It is now recognized, however, that household use of simple solid fuels for heating and cooking, although not the major source of fuel-based emissions, are responsible for a large portion of human exposure to major health-damaging pollutants. About half the world's households use such fuels, often in open-fires or simple stoves that leak much of the emissions into the living area. This is because the intake fraction (also called dose effectiveness) of emissions, i.e. the fraction of the material that is actually breathed in, is typically 2-3 orders of magnitude higher than for outdoor sources. In other words, a gram released indoors will cause as much dose and consequent ill-health as a kilogram or even more released outdoors.

Because the fuel is not premixed with air before combustion, as with gas and liquid fueled stoves, a substantial portion of the fuel carbon is often diverted to products of incomplete combustion (PIC), sometimes more than 20%. These PIC include a range of health-relevant pollutants such as particulates, carbon monoxide, dozens of hazardous volatile organic compounds such as benzene, formaldehyde, and butadiene, and a range of polycyclic aromatic hydrocarbons such as benzo-alpha-pyrene. Depending on its composition, coal combustion can also result in emissions of sulfur oxides and a range of toxic elements such as arsenic, fluorine, and mercury. Unlike those coming from large-scale combustion devices, the particles produced by stoves are essentially all combustion-derived and thus small enough (below 1-2 microns in diameter) to penetrate deeply into the respiratory system. Unlike power plants or large boilers, for example, flue velocities are too small to entrain larger particles. Therefore, fuel ash content does not predict particulate emissions from stoves as it does in larger devices.

There are two general ways to approach the estimation of health effects of such emissions. The first is to measure human exposures to indicator pollutants and to the exposure-response functions derived from the hundreds of epidemiological studies done for urban outdoor pollution. It is generally agreed that small particles (below 10 or 2.5 µm) are best single indicator pollutant for this purpose. Such estimates have been done and usually derive quite high levels of impact from household use of solid fuels worldwide, well over 2 million premature deaths annually (out of a total of about 55 million). There are a number of important problems with this approach, however, and the confidence in such estimates is low.

The second method addresses many, but not all, of the problems of the first. It uses the growing epidemiological literature reporting studies done in developing-country rural settings only to derive exposure-response relationships and uses models estimating household fuel use derived from survey and other data to determine exposure levels. Estimates using this approach are part of WHO's program in Comparative Risk Assessment, in which global burdens of disease from some 30 major risk factors are being determined using compatible methods. They have been subject to a high degree of peer review and other forms of quality control.

The epidemiological evidence for different disease endpoints has been divided into three categories, depending on its strength, depth, consistency, coherence, physiological plausibility, and other factors:

I. Strong Evidence is available to quantify the enhancement of
a. Acute respiratory infections in young children, the largest source of ill-health in the world today;
b. Chronic obstructive lung disease in adult women, such as chronic bronchitis
c. Lung cancer in adult women, but only in coal-using households
II. Moderate evidence is available to quantify the enhancement of
a. Tuberculosis
b. Cataracts and perhaps other eye diseases
c. Asthma attacks
d. Aero-digestive system cancers
III. Highly suggestive, but to date, unquantifiable evidence for the enhancement of
a. Adverse pregnancy outcomes such as low birthweight, neonatal death, and stillbirth, one of the largest categories of ill-health in the developing world
b. Heart disease, which has been established as one of the major disease outcomes for outdoor particle pollution in the developed world.

In addition, there is substantial evidence, mainly in China, of disease caused by coal smoke contaminated by fluorine and arsenic.

Depending on their degree of conservatism (whether to include Category II as well as I), some 1.2-1.6 million premature deaths can thereby be attributed to indoor pollution from household solid fuel use worldwide in 2000. Most of this impact is found in Sub-Saharan Africa and South Asia. China, too, mostly relies on such fuels for household use, but has implemented a highly successful program of stoves with chimneys in recent decades, apparently lowering the burden substantially.

Workshop Homepage * Background
Summaries: Overview, Gases, Aerosols, Tech., Health, Agri./Eco.
Abstracts: Day 1, Day 2, Day 3, Day 4, Day 5 * Participants