Household Air Pollution from Coal and Biomass Fuels in China: Measurements, Health Impacts, and Interventions

As part of a random 3,500-household survey, a random subsample of 396 rural households in provinces of Shaanxi, Hubei, and Zhejiang were monitored over a 24-hr period for PM₄ (PM with aerodynamic diameter ≤4 μm) in kitchens and living rooms over (Edwards et al., in press). Among these 396 households, 159 were measured in both summer and winter. For nearly all household stove or fuel groupings, PM₄ levels were higher than China’s IAQ standard for PM₁₀ (SEPA 2007), even when the difference in particle size specified was ignored. Some of these higher levels were because of neighborhood pollution effects. Other higher levels were the result of the large variation in the complex fuel and stove situation in rural China where multiple types are used depending on season, room of the house, and purpose (cooking-water heating, space heating, heating beds). Many households change fuels according to daily and seasonal factors, resulting in different seasonal concentrations in living rooms, bedrooms, and kitchens. Indeed, 34 fuel–stove combinations were being used in some villages in the summer and 28 in the winter. The average household used 2.6 types of fuels. This complexity makes detailed assessment of the effects of different fuels, stoves, seasons, and household layout difficult without either large sample sizes in representative studies or more nonrepresentative studies in controlled settings. It illustrates that even though the “energy ladder” concept, in which households generally move from lower to higher quality household fuels with development (income), is well established at the macro level, the situation in individual households can be complex during the transition in a middle-income country like China (Sinton et al. 2004).

As part of a study funded by the World Bank, PM₄, CO, and SO₂ were monitored in 457 household-days in four poor provinces (Jin et al. 2005). The two provinces where biomass was the primary fuel had higher concentrations of PM₄ and CO than the two primarily coal-burning provinces. Among the two coal-burning provinces, Guizhou had lower concentrations of CO and SO₂ than Shaaxi. In addition to differences in local fuel quality and house ventilation conditions, heterogeneity in space heating contributes largely to the observed geographic differences in pollutant concentration within and between provinces. In the three northern provinces, indoor heating affected the level and spatial distributions of pollution inside homes, possibly to more of an extent than cooking. In an examination of temporal patterns of indoor pollution, day-today variability of concentrations within individual households, although substantial, were smaller than concentration variations across households. This temporal feature indicates that applying a crude proxy for exposure (e.g., whether or not using a biomass stove) would overlook large house-to-house differences in pollutant concentrations and exposures. It should be noted that although these studies employed a systematic measurement approach for all the measured households, the household selection procedure does not appear to be population-based or random; hence, the reported pollutant concentrations may not be generalizable to the general households in the four provinces. However, the spatial and temporal patterns observed in the study are expected to be typical for similar houses in the same regions (He et al. 2005; Jin et al. 2005).

The Sino–Dutch Project was a study of approximately 140 household kitchens in five counties of three provinces (Dong et al. 2006). The kitchens were monitored for 24 hr before and after introduction of a suite of improvements (improved biomass stoves, biogas) for PM_2.5, CO, SO₂, and ammonia. Measurements were performed a year apart to control for season and allowed for the improvements to “settle in.” Significant reductions occurred in all pollutants, with PM_2.5 levels over a 24-hr period averaging about 120 μg/m³ after a year (Dong et al. 2006). The levels found in other studies indicated a reduction of only about 40% from preintervention values (Dong et al. 2006). Most households still did not meet the Chinese IAQ standard for PM₁₀ (Dong et al. 2006).

A unique feature of coal smoke exposure in certain parts of China is the combination of inhalation exposure and ingestion exposure. Ingestion exposure occurs when foods are consumed that have been dried over coal or biomass fires. For example, in many “poisonous” coal endemic areas, drying corns and chili peppers over open fire pits is a common practice (He et al. 2005). During the drying process, the foods absorb the smoke and become enriched with toxic elements (e.g., fluoride and arsenic). The ingestion of contaminated foods is thought to be the predominant exposure route causing endemic arsenism and fluorosis in China (Chen et al. 1993; Finkelman et al. 1999; Jin et al. 2003). This potentially important exposure route, however, has not been examined for other toxic components (e.g., PAHs, nickel compounds) of coal or biomass smoke.

Associations between lung cancer and coal smoke exposure have been found in numerous epidemiologic studies conducted in China. Among these studies include the decades-long investigations in Xuanwei, the site of numerous studies of the relationship of coal smoke and lung cancer because of the unusually high rates of lung cancer in nonsmoking women using smoky coals in open fire pits (e.g., Chapman et al. 1989; Liu et al. 1991; Mumford et al. 1987). The odds ratios (ORs) for lung cancer associated with indoor coal use are summarized in a recent meta-analysis by Smith et al. (2004). For comparison, ORs were estimated with and without adjusting two important confounding factors. Tobacco smoking status was either adjusted or analyses were done solely in nonsmokers. Because chronic respiratory diseases such as chronic bronchitis, tuberculosis, asthma, and emphysema may increase the probability of developing lung cancer later in life (Luo et al. 1996), adjustment was also made for these diseases. These adjustments may result in underestimating the ORs of lung cancer, as some previous lung diseases may be on the intermediate path from exposure to lung cancer—in this case, they are not confounders and should not be adjusted. The OR estimate for women was 1.17 with 95% confidence interval (CI) of 1.02–1.35. However, when the analysis was restricted to studies that adjusted for smoking and chronic respiratory disease, the OR estimate for women substantially increased to 1.94 (95% CI, 1.09–3.47). The OR estimate for men was 1.79 (95% CI, 1.18–2.72) and slightly lower when including confounding by smoking and chronic airway disease with OR = 1.5 (95% CI, 0.97–2.46). The overall OR estimates for men and women combined were 1.86 (95% CI, 1.48–2.35) and 2.55 (95% CI, 1.58–4.10), respectively, both without and with the two confounding factors adjusted. In a more recent review, Zhao et al. (2006) found similar results.

The meta-analyses present strong epidemiologic evidence that exposure to indoor coal smoke significantly increases lung cancer risk. Recently, the International Agency for Research on Cancer (IARC) classified emissions of indoor combustion of coal as carcinogenic to humans (Group 1) on the basis of sufficient evidence both in humans and in animals (Straif et al. 2006). The role of certain genotypes and proteins in the development of lung cancer has been examined in Xuanwei County residents using smoky coal, suggesting that an individual’s susceptibility to lung cancer may be increased by the glutathione S-transferase 1 (GST1)-null genotype (Lan et al. 2000, 2001; Lan and He 2004; Mumford et al. 1999). There is also limited evidence that other cancers can be caused by exposure to household coal smoke, including esophageal (Pan et al. 1999) and cancers of the head and neck (Dietz et al. 1995). This may be due to not only direct respiration of airborne pollutants but also the contamination of food by coal smoke (Roth et al. 1998).

Indoor coal smoke exposure has been linked to various respiratory illnesses. A study conducted in Anhui Province shows that the prevalence rates of chest illness, cough, phlegm, and shortness of breath were significantly elevated in nonsmoking women living in homes with both smokers and coal heating (Pope and Xu 1993). A survey of 10,892 Xuanwei residents found that the OR estimates for smoky coal users compared with smokeless coal users, were 1.73 for shortness of breath, 3.30 for cough, and 4.23 for phlegm, and that the OR estimates for smokeless coal users compared with wood users were 1.35 for cough and 1.67 for phlegm (Zhou et al. 1995). A study of 5,051 seventh-grade students from 22 randomly selected schools in the greater metropolitan area of Wuhan found that coal burning for cooking/heating increased the risk of wheezing with colds (OR = 1.57; 95% CI, 1.07–2.29) and without colds (OR = 1.44; 95% CI, 1.05–1.97) (Salo et al. 2004). In a population-based case–control study of childhood asthma conducted in Shunyi County located in suburban Beijing, an increased risk was observed for use of coal for heating (OR =1.5; 95% CI, 1.1–1.9) and for use of coal for cooking without ventilation (OR = 2.3, 95% CI, 1.5–3.5) (Zheng et al. 2002). Indoor coal use was associated with increased incidence of rhinitis, faucitis, and tonsillitis in children living in Taiyan City, Shanxi Province (Cheng et al. 2002). The effects of household coal use were also observed in 624 infants and young children (1–3 years of age) in Nantong, Jiangsu Province, as the prevalence of cough and that of pneumonia were significantly higher in households using coal than in the “control” households using gas (Zhou et al. 1994). Chronic obstructive pulmonary diseases (COPD) is a major cause of ill heath in China, causing more than 1.3 million deaths annually (WHO 2002). A case–control study conducted in Shanghai showed that indoor coal use was more strongly associated with COPD than estimated exposure to outdoor SO₂ and PM₁₀ (Tao et al. 1992). A survey of 21,648 rural residents in Anhui Province showed that the COPD rate was significantly higher in individuals who used coal for heating than in those who did not (Li et al. 2002).

Exposure–response relationships have been examined in a study of 7,058 elementary school children living in four large Chinese cities. When lifetime exposures to coal smoke from heating were classified according to four ordinal levels (no, light, moderate, and heavy exposure), monotonic and positive exposure–response relationships were observed for OR estimates of phlegm, cough with phlegm, and bronchitis. In addition, OR estimates for cough, wheeze, and asthma were all > 1 in the exposed groups relative to the no-exposure group (Qian et al. 2004).

The effect of coal smoke exposure on lung function has been investigated in a few studies of children or adults. In school children living in the cities of Chengde (Hebei Province) and Shanghai, measurements made in winter showed reductions of 1.5–10.7% in forced vital capacity (FVC), forced expiratory volume in 1 sec (FEV₁), or peak expiratory flow rate (PEFR). These reductions were associated with the use of coal for cooking/heating compared with the use of natural gas or LPG (Shen et al. 1992). In adults, evidence of lung function impairment from coal smoke has been found in a few studies (Jin et al. 1995; Wang H et al. 2004; Xu et al. 1991). In 1986 when household coal use was still prevalent in Beijing, lung functions of 1,440 of the city’s adult never-smoking residents were measured (Xu et al. 1991). The authors found that heating with coal stoves was associated with reduced FEV₁ and FVC compared with radiator-based heating supplied from a centralized boiler. Simple and inexpensive PEFR measurement can be self-conducted by subjects, making a study possible to measure a large number of people, such as the one conducted in 10,892 adults living in 18 villages of Xuanwei County. This study compared the relative potencies of smoky coal, smokeless coal, and wood and examined the effect of these on PEFR. The results show that the strongest risk factor for lowered PEFR was smoky coal, followed by smokeless coal, then by wood. The study also found that the use of coal stoves with chimneys was associated with increased PEFR compared with the use of open fire pits (Jin et al. 1995). In a recent study conducted in Shanxi Province, women using coal for cooking had lower lung function values than those using gas for cooking (Wang H et al. 2004).

Direct exposure to coal smoke or coal smoke condensate has been examined for potential effects on the human immune system in about half-dozen studies, all published in Chinese journals (e.g., Jin et al. 2002). In nonsmoking Shanghai women, those who used coal for cooking had significantly lower serum IgG content, peripheral T-lymphocyte activity, E-rosette formation rate, and interleukin (IL)-2 induction activity than those who used gas for cooking (Wang et al. 1993). Similar findings were reported in another article that additionally reported decreased activity of natural killer cells in women using coal for cooking, although no significant association was found for IL-2 induction activity in T-lymphocyte cells (Mao et al. 1994). A study of 624 infants and young children found that serum IgG content was significantly lower in those whose households used coal for cooking than in those whose households used gas fuels (Zhou et al. 1994). These findings suggest that coal smoke exposure weakens the human immune system, making those exposed individuals more susceptible for developing illnesses (Jin et al. 2002).

In China there are approximately 100 counties (of approximately 1,500) that have been deemed “endemic” because local coal deposits have high contents of toxic elements (Sinton et al. 2004). The most noticeable coal-related endemics are arsenism and fluorosis, which are the result of chronic arsenic and fluoride poisoning. Burning arsenic-rich coals occurs widely at least in eight counties of two provinces (Guizhou and Shaanxi), affecting approximately 300,000 people (He et al. 2005; Jin et al. 2003). Reported illness includes symptoms of arsenicosis (Shraim et al. 2003).

It is known that high arsenic exposure via drinking water causes bladder, lung, and skin cancers (Boffetta 2004). However, cancer has not been studied in poisonous coal areas in China. It is estimated that more than 10 million people in Guizhou Province and surrounding areas suffer from dental and skeletal fluorosis (Cao 1991; Chen et al. 1993; Finkelman et al. 1999), mainly as a result of excess fluoride intake from eating foods dried over open fire pits (Wu and Li 1990; Yan 1990). In some fluorosis areas, almost all elementary and junior high school students had dental fluorosis, and osteosclerosis in the skeletal fluorosis patients was very serious (Ando et al. 1998, 2001; Watanabe et al. 1997). In addition, chronic selenium and possibly mercury poisoning has also been reported to result from household coal use in the affected areas (Fang et al. 2004; Finkelman et al. 1999; Horvat et al. 2003).

There have been numerous reported acute poisonings, including fatal cases, especially during heating seasons, resulting from indoor coal combustion under poor ventilation conditions. Under normal combustion and ventilation conditions, Zhang X et al. (1996) measured elevated blood levels of CO–hemoglobin adduct (COHb) in residents of households using coal and reported that the contribution to COHb from indoor coal combustion was larger than that from cigarette smoking. However, the health effects of chronic CO exposure at elevated levels but lower than those of acute poisoning have not been examined in China, although this type of CO exposure is typical in households using solid fuels (Zhang et al. 1999). CO is a known neurotoxin, and there is a potential for chronic exposure to exert neurologic effects. Furthermore, it has been associated with effects on prenatal and early postnatal mortality and low growth in children of women exposed during pregnancy. These effects are presumably due to oxygen deprivation.

Assignment of causality is difficult on the basis of observational studies alone, the interpretation of which is limited by potential unmeasured confounders, for example, socioeconomic status, which is often associated with disease outcome as well as use of dirty household fuels. No randomized intervention trials providing stronger evidence have yet to be published for household solid fuel use in China and would be difficult to conduct for chronic diseases such as COPD and lung cancer that are the results of many years of exposure. Using the rich database of studies in Xuanwei, however, two intervention studies (Chapman et al. 2005; Lan et al. 2002) have been published that take advantage of the “natural experiment” when improved coal stoves with chimneys were introduced in the late 1970s to replace open fire pits, thereby lowering indoor air concentrations by a factor of approximately 3 on average (Lan et al. 2002). A questionnaire was administered in 1992 to more than 10,000 farmers in the area upon which the following two studies were based.

Because pollutant emissions depend on both fuel quality and stove design, interventions can focus on fuel, stove, or both. To date, stove interventions have focused predominantly on households using biomass, and fuel interventions appear to have focused mainly on coal, although the reverse trends are also becoming apparent.

The most impressive organized rural energy intervention in human history was China’s NISP, through which China accomplished more than all other developing countries combined to improve household energy use by introducing more than 180 million improved stoves since the early 1980s. All introduced stoves had chimneys and some had manual or electric blowers to promote more efficient combustion. Unfortunately, the program ended in the mid-1990s, and now there is relatively little action to improve the present rural energy situation. The review of NISP by Sinton et al. (2004) found that the program improved IAQ, but not sufficiently to meet China’s IAQ standards. In addition, because the NISP focused mainly on biomass, rising coal use in rural areas, often in stoves without chimneys, is threatening to erode the benefits. Although the Chinese Ministry of Health has embarked on a program to introduce improved coal stoves to approximately 100 endemic arsenism and fluorosis counties, progress has been slow because of the lack of resources (Sinton et al. 2004).

Various formulated coals have been developed in China to reduce hazardous emissions. The most noticeable is the formulation of so-called honeycomb coals that have been used widely throughout the country for decades. The perforated shape of the coal allows for more uniform air supply, consequently leading to higher combustion efficiency. Some honeycomb coals are specially formulated, for example, with lime to react with fuel sulfur and retain it in ash, instead of emitting it as SO₂ (Ge et al. 2004). Because of different coals, different formulations, and different measurement methods, it is difficult to generalize about the impact of such coal fuels (Bond et al. 2002; Kasper et al. 1999; Perlack and Russell 1991; Yao et al. 1992).

In some areas, “clean” forms of coal are being promulgated for urban use, but their sustainability is uncertain. In addition, household coal fuels are not required to undergo standard testing, and there is even less regulation of what is actually sold in the market place. In rural China little attention is focused on clean coal, and household coal varies dramatically across the country according to the character of local coal deposits. In developing clean coal strategies, it is useful to consider some historical lessons from other parts of the world. Before actually banning coal in cities, the United Kingdom and other countries using household coal developed and deployed a range of clean coals for small-scale use. This might have helped lessen the pollution problem in the short-term but, eventually, these countries realized that in simple household combustion, even the processed forms of coal or cleaner natural forms, such as anthracite, could not be burned clean enough to use in urban areas and still meet health-based pollution standards. Perhaps the question now is how clean is today’s clean coal compared with those clean coals used half a century ago.

In rural China there is the potential for modernization of its rich biomass resources into clean energy sources. The generation rate for crop residues in the field plus agricultural processing residues amounts to about 790 million metric tons per year, or approximately 10 exajoule (EJ). For comparison, total coal use in China was 39 EJ in 2004 in all sectors (National Bureau of Statistics 2006). It has been estimated that about half the total crop residues might be available for energy use after accounting for other uses (e.g., fodder, fertilizer, and industrial feedstock). An alternative to traditional direct combustion of crop residues is the use of “gasifier” stoves that achieve high combustion efficiency through designs that promote secondary combustion. Reliably high combustion efficiency is easier to accomplish with small electric blowers, but some models attempt to do so with natural draft. A national competition was held in 2006 by the Chinese Association of Rural Energy Industries funded by the UK Shell Foundation to promote the development and dissemination of such low-emission biomass stoves and fuel cycles. To reliably achieve low emissions in field conditions, however, such stoves require more uniform fuel through, perhaps, development of small local biomass-processing enterprises. Stoves burning biomass pellets, for example, can achieve remarkable efficiencies.

Recently, new pollution problems are arising from crop residues because in more economically developed parts of the country, farmers are becoming reluctant to gather biomass residues from the field and store them for use throughout the year, preferring easier-to-handle fuels such as coal briquettes or LPG. This shift to modern fuels creates an excess of crop residues that are commonly burned in the field, leading to widespread ambient pollution in some seasons. A number of village-scale gasifiers have been built to make more efficient and cleaner use of these residues by distributing the gas to individual households. A Bio-energy Modernization Demonstration Project, carried out in Jilin Province, is designed to develop combined heat, electricity, and cooking fuel production (trigeneration) from corn stalks. The technology seems to be workable, but its widespread use, however, is limited by the lack of commercial viability and potential risks due to acute CO poisoning (United Nations Development Program et al. 2004).

Simple household biodigesters have long been available for converting animal waste into biogas containing methane, but these biogesters are limited to areas with sufficient dung, water, temperature, and financial capital. Biobriquette technology, in which biobriquettes are made of biomass, coal, and sulfur fixation agents through high-pressure manufacturing processes (Lu et al. 2000), is being investigated. Such a fuel has shown an acid-neutralizing capacity of the ash it produces (Dong et al. 2004) and a reduction in SO₂ emission (Isobe et al. 2005).

Advanced technologies are also being explored for converting crop residues into dimethyl ether (DME), a fuel with characteristics similar to those of LPG. DME, sold as a substitute for LPG, is currently made from natural gas but can only take on a large role if made from crop residues or coal. Finding ways to convert coal, the country’s most abundant energy resource, to DME has been attractive to researchers and the Chinese government (Han et al. 2004). With further cost reduction in coal-to-DME technologies and the rise in LPG costs, however, coal-derived DME for household use may become economic (Larson and Yang 2004). Produced with polluting and inefficient methods, the use of traditional coal gas in Chinese cities has been declining.

In large Chinese cities, policies of banning household coal use have been in place with the main goal of reducing outdoor air pollution but with the side benefit of reducing IAP levels. Although health benefits have not been directly reported or measured in Chinese cities, the findings from cities in other countries; for example, studies of the ban on coal sales in Dublin, Ireland, documented sudden decreases in atmospheric concentrations of PM and SO₂ as well as significant reductions in cardiovascular and respiratory death rates (Clancy et al. 2002).

Simple stove improvement is not expected to reduce total pollutant emissions because chimneys divert only a fraction of the emitted pollutants from the inside of households to the outside, rather, it worsens neighborhood and local air quality. A more permanent solution is to move to clean combustion through changes in fuel and/or stoves, which actually reduce emissions. An added benefit from clean combustion is improved atmospheric visibility and lower ambient pollution as well as potentially lower greenhouse emissions. PICs from low-efficiency biomass combustion in household stoves, for example, contain a relatively large fraction of methane many times more potent in radiative forcing than CO₂, implying that biomass stoves can have net global warming potentials, even with a renewable fuel cycle (Smith et al. 2000).