Gas cooking appliances and indoor pollution

The contribution of the home environment and other lifestyle factors in the pathogenesis of allergic disease has attracted much attention, particularly the role of indoor pollution from gas cooking appliances and its association with asthma-like symptoms. Recognizing that the prevalence of gas cookers in developed countries is high at 50–70%, any evidence of a deleterious effect on health represents a major public health issue. The predominant sources of indoor pollutants are gas cooking (and heating) appliances and tobacco smoke, which produce complex mixtures of volatile organic compounds, sulphur dioxide, water vapour, particulates, carbon dioxide and oxides of nitrogen. The most extensively studied indoor pollutant is nitrogen dioxide, the concentrations of which are determined partly from outdoor concentrations, but are highest in kitchens with gas cookers, often at levels higher than the World Health Organization's guideline value for outdoor air of 150 μg/m3 in 24 h [ 1]. Despite such concerns, the strength of the epidemiological evidence linking the effects of gas cooking (and nitrogen dioxide exposure) to respiratory health has been inconsistent. Many of the epidemiological studies focused on children. Some studies have shown a small but significant effect of NO2 exposure on respiratory symptoms and lung function [ 234], while others have failed to confirm the association [ 567]. Samet and Utell [ 8] reviewed the inconsistencies in the epidemiological literature and concluded that methodological issues such as low statistical power, confounding or effect modification by other pollutants, misclassification of exposure and the use of insensitive measures of health outcome could explain the inability to derive definitive conclusions. In particular, many studies used the presence or absence of gas cookers as a proxy for exposure to NO2, and some measured the levels of NO2 in homes at fixed points. Monitoring analyses confirm that neither fixed site monitoring or use of categorical variables such as gas or electric cooking accurately predict personal NO2 exposure [ 1] as this will be influenced by outdoor levels, indoor sources, the rate of exchange between indoor and outdoor air, time spent in the microenvironment by the individual and other characteristics of the dwelling and its furnishings that influence NO2 dispersal and removal. There have been recent studies which have included direct measures of NO2 exposure, and have reported increased sore throats and colds in children with high hourly peak NO2 exposures [ 9], a small reduction in FVC with personal NO2 exposure [ 10], and increased respiratory symptoms in children with gas stove exposure even after adjusting for NO2 levels (OR 2.2 [1.0–4.8]) [ 11]. However, two other studies conducted in infants and children revealed no association between measured NO2 exposure and respiratory illness [ 12, 13]. In a landmark meta-analysis by Hasselblad et al. [ 14], the results of 11 studies were combined including those where a surrogate estimate of NO2 exposure, such as a presence of gas cooker was used, or NO2 levels were directly measured in relation to long-term exposure to a 30-μg/m3 increase in NO2 (comparable with the increase resulting from exposure to a gas stove). The conclusions after using four different statistical methods were all in agreement; the relative risk of lower respiratory illness in studies using gas stoves as a surrogate for NO2 exposure was 1.18, but when four studies including measured NO2 levels were combined, the risk increased to 1.27 with an overall combined risk of 1.20, i.e. exposure to a 30-μg/m3 increase in NO2 exposure would result in a 20% increase in respiratory illness. A further analysis of the evidence allowing for heterogeneity in study design was in agreement and concluded that according to age, the estimated risk is higher (1.29–1.60) in children aged 5–6 years and 6–12 years compared with infants particularly for health end-points that did not include asthma [ 15]. If children, the elderly and women tending to young families spend more time at home, it follows that these groups are likely to be exposed to the highest levels of NO2. There is little data specifically on the elderly population, while evidence in adults is equally fraught with inconsistency as in children, largely because of the same methodological limitations. Jarvis et al. described two cross-sectional studies; women using gas for cooking had an increased risk of current asthma-like symptoms (during the past 12 months) and reduced lung function (FEV1), and these associations were not observed in men [ 16]. Similarly using a meta-analysis approach as part of the European Community Respiratory Health Survey [ 17], data from 5561 males and 6029 females from 11 countries (23 centres) were analysed, and although there was no association between respiratory symptoms and gas cooking in males, the association varied between centres in females with an overall positive association with current wheeze and breathlessness. Gender differences in the effects of gas cooking have also been described in children [ 2, 4, 18, 19]. Few studies have directly compared measured personal NO2 exposure between children and adults who cook, but the data does not support the notion that children have consistently higher personal NO2 exposures than their mothers [ 202122]. While the pattern of exposures in cooking women may be different, children may equally be exposed to a similar 1-hr peak and weekly mean NO2 exposures [ 21]. The majority of epidemiological studies have focused on children largely because their period of exposure to pollution is recent, exposure to active smoking and occupational exposures are absent, and the frequency of respiratory virus infections is high, rather than any convincing evidence of increased biological susceptibility. Another important facet of gas cooking and allergic disease is whether exposure to gas stoves in childhood alters the prognosis of asthma. Given that recent controlled exposure studies suggest that NO2 exposure either alone or in combination with other pollutants may alter the early and late asthmatic response and potentiate specific bronchial hyperresponsiveness to inhaled allergen [ 23, 24], childhood exposure to gas stoves and effects on asthma incidence and prognosis have been postulated, but until recently not tested directly in an epidemiological study. Two studies specifically designed to address the question of gas cooking in childhood and prognosis of asthma later in life have been described, one of which is so far published only in abstract form [ 25]. The other is described in this issue of the Journal; Moran et al. [ 26] report the results of a cross-sectional study drawn from the British 1958 birth cohort that addresses the question of gas cooking in childhood and adulthood and the risk of asthma symptoms, allergic sensitization and lung function in early adulthood. From 5470 individuals who contributed data at birth and all five follow-ups ranging from ages 7–33 years, 1989 index cases with a history of asthma, wheezy bronchitis or pneumonia and 626 controls (a random 20% of the remainder) were selected, of which a final number of 1119 cases and 330 controls were interviewed (at ages 34–35 years). Assessment was made of the presence and severity of respiratory symptoms, measures of spirometric indices of lung function and atopic status by skin prick testing to common indoor and outdoor inhaled aeroallergens. Exposure to gas cooking in childhood at age 11 years (but not at 7 years) was assessed retrospectively at the interview, as cooking fuel use was not ascertained directly during childhood, while the prognosis of asthma was determined prospectively by the presence or absence of typical symptoms/wheeze at each cohort follow-up visit. Among those interviewed, the reported exposure to gas and electric cooking fuel alone at age 11 was high at 58 and 36%, respectively, while current exposure was lower at 45 and 35%, respectively. Furthermore, of those reporting current exposure to gas and electric cooking fuel, 34 and 51% were not exposed to the respective cooking fuel at age 11 years. The authors found no significant association between the use of gas cooking in childhood or adulthood and the incidence or prognosis of asthma/wheeze, current severity of symptoms or allergic sensitization. The results, however, indicated that individuals who developed asthma in childhood and were exposed to gas cooking currently or in childhood had a non-significant increased risk of persistent asthma at age 35 years. Of particular interest was the observation that subjects who were currently exposed to gas for cooking had a significantly reduced FEV1 (− 0.07 L, 95% CI ± 0.056) and FEV1/FVC ratio (− 1.0%, ± 0.8) compared with those who currently used electricity for cooking. These observed effects are small and of uncertain clinical significance. The authors also noted that the reduction in lung function was concentrated amongst males and those with current asthma/wheeze (but not those with past asthma/wheeze). Despite the methodological limitations, particularly in classification of exposure that are acknowledged by the investigators, the conclusions drawn are important. Firstly, the discrepancy in gender effects between this study and those described by Jarvis et al. [ 16, 17] are not new, and are in keeping with the observations of Viegi et al. [ 27] who noted increased cough and phlegm in males with use of bottled (but not natural) gas for cooking, and wheeze and shortness of breath in females with the use of a gas stove or forced-air circulation. Comstock et al. [ 28] also reported an increased frequency of respiratory symptoms and impaired ventilatory function only among men, particularly among those who never smoked. Overall the data suggest some differences in health risk between genders in both children and adults, but the evidence is not consistent in the groups expected to be at highest risk. One cannot conclude these are solely due to chance or variances in study design; further information is required to accurately reflect relevant exposures determined by factors such as length and intensity of cooking, personal exposures of individuals directly exposed to cooking irrespective of gender, kitchen ventilation characteristics and effective use and availability of extractor fans. In stratified analyses of women from East Anglia, Jarvis et al. [ 16] noted that among women who used gas cookers, they did not find a lower risk (as one might expect) in employed women or students compared with housewives or unemployed women, or in women who had extractor fans compared with those who did not. These latter observations likely reflect limitations in accurate pollutant exposure as one cannot ascertain whether employed women use gas cookers less often than housewives, or that the presence of extractor fans reflect their effective use. There is also little evidence that adult women are more susceptible than men to the effects of exposure to NO2 in controlled exposure studies [ 29], and very little data comparing measured NO2 exposures between genders. Dockery et al. [ 20] documented that the difference in weekly mean personal NO2 exposure between spouses was only 4 and 3 μg/m3 from homes with gas and electric cookers, respectively. A second important point from the study by Moran et al. [ 26] is that exposure to gas cooking in childhood or adulthood was not associated with the prognosis of asthma. A more pertinent observation was that the prevalence of individuals exposed to either gas or electric cooking fuel had fallen by 13 and 1%, respectively. Although this is subject to recall bias, it is supported by the observation that of the total domestic energy consumption by cooking in the UK in 1995, 68 and 30% was accounted for by gas and electricity, respectively, and that there has been a gradual fall in total energy consumption by gas cooking from 2.8 Million tonnes of oil equivalent (Mtoe) in 1970 to 2.1 Mtoe in 1995 paralleled by an increase in electrical energy consumption by electrical cooking appliances (including ovens and hobs) from 11.2 to 13.8 TWh between 1970 and 1994 [ 30 Department of Trade and Industry. Energy Consumption in the United Kingdom. Energy Paper no. 66. London: DoTI, ]. A decrease in use of electrical ovens has not been enough to offset the rapid proliferation in electrical energy consumption due to microwave ovens since the early 1980s. The consumption of residential energy due to gas and electrical cooking fuels show a similar trend in the US [ 31 US Bureau of the Census. Statistical Abstract of the United States. Washington, DC: US Bureau of the Census, ]. Given that the prevalence of individuals using gas cookers and energy consumption by gas cooking has fallen in real terms, the findings on the prognosis of asthma by Moran et al. (and the overall prevalence of allergic disease) are not surprising. The observation that current (but not childhood) gas exposure was associated with a small reduction in lung function in current asthmatics in the study by Moran et al. argues against any long-term effect of gas cooking on lung function. The decrement in FEV1 was small (0.070 L), and is supported by evidence from controlled exposure and epidemiological studies of NO2 [ 32]. Although there is some, but not consistent, evidence for a small increase in airway resistance with higher concentrations of NO2, the majority of studies have found no change in resting lung function in either normal or asthmatic subjects exposed to concentrations of NO2 of 1000 p.p.b. (≈ 1880 μg/m3) or less [ 33 Advisory Group on the Medical Aspects of Air Pollution Episodes. Report no. 3: Oxides of Nitrogen. London: Department of Health, ]. Bronchial responsiveness has been shown to increase slightly in normal subjects exposed to greater than 1880 μg/m3 NO2, and in only about a third of studies specifically in asthmatic subjects. Given that the major site of NO2-induced injury is the transitional zone (the area between the terminal bronchioles and alveoli), conventional pulmonary function tests most sensitive to changes in large airway function may be a relatively insensitive means of detecting the earliest responses in small peripheral airways, which may account for the relatively small changes in pulmonary function responses in both controlled exposure and epidemiological studies. Despite the methodological short comings, the evidence for significant effects on lung function in adult asthmatics from controlled exposure studies is limited, while the evidence on lung function and exposure to gas cooking in children and young women from epidemiological studies is inconsistent with some studies showing a reduction [ 3, 16, 28, 343536] while others have shown no effect [ 37383940]. So does gas cooking and exposure to NO2 cause asthma? If criteria of causality such as magnitude, consistency, specificity, coherence, temporal relationship and biological plausibility for the association are applied, then the evidence is absent. The wealth of literature on gas cooking and respiratory health does offer moderate evidence that some susceptible groups, particularly children and cooking women are at risk, and it is current exposure with acute but reversible respiratory effects that are likely to be important, particularly as there is little evidence of effects on long-term outcome [ 25, 26]. The challenge lies in identifying why some groups are susceptible and to accurately classify which aspects of gas cooking exposure are hazardous. Measurements of average NO2 levels by passive monitors have offered an improvement over surrogate information such as presence of gas cookers, but animal [ 41] and more recently human studies [ 9] have suggested that peaks of exposure which more closely mimic ‘real life’ domestic exposures to gas cooking may be more relevant to health effects. Some studies have suggested that averaged weekly NO2 exposures predict maximum 1-hr peak exposures [ 42] while others have argued for only a weak association at best [ 21], hence more effective and consistent characterization of these exposures are required. Accurate classification of exposure relies on measuring the right pollutant(s), and because gas cooking is associated with other indoor pollutants, of which nitric oxide (the predominant oxide of nitrogen which is quickly oxidized to NO2 soon after combustion) and particulates are important, the effects of these factors need to be accounted for in future studies. The epidemiological evidence for the effects of outdoor particulate pollution on respiratory morbidity and mortality are powerful [ 43], and the possibility that nitric oxide is also related to adverse outcome cannot be discounted [ 44, 45]. Garrett et al. [ 11] measured average indoor NO2 levels in the homes of 148 children and reported that gas cooker exposure was a significant risk factor for respiratory symptoms even after adjusting for NO2 levels (OR 2.2 [1.0–4.8]), suggesting an additional risk apart from the average NO2 exposure associated with gas cookers. The possibility that there are other confounders or interactions of other pollutants during gas cooking in addition to NO2 [ 11] could also explain some of the discrepancies in the epidemiological evidence but requires confirmation. Despite the inconsistencies, the current evidence would argue for a modest risk of gas cooking on health, with public health concerns due to the high prevalence of exposure resulting in a considerably increased attributable risk. What is missing from the evidence is knowing whether other pollutants in addition to NO2 are involved, what patterns of exposure are important and plausible biological mechanisms such as the interaction with respiratory infections [ 46].