Impact of improved stoves, house construction and child location on levels of indoor air pollution exposure in young Guatemalan children

Article metrics

Abstract

The goal of this study was to assess the impact of improved stoves, house ventilation, and child location on levels of indoor air pollution and child exposure in a rural Guatemalan population reliant on wood fuel. The study was a random sample of 204 households with children less than 18 months in a rural village in the western highlands of Guatemala. Socio-economic and household information was obtained by interview and observation. Twenty-four hour carbon monoxide (CO) was used as the primary measure of kitchen pollution and child exposure in all homes, using Gastec diffusion tubes. Twenty-four hour kitchen PM3.5 was measured in a random sub-sample (n=29) of kitchens with co-located CO tubes. Almost 50% of the homes still used open fires, around 30% used chimney stoves (planchas) mostly from a large donor-funded programme, and the remainder of homes used various combinations including bottled gas and open fires. The 24-h kitchen CO was lowest for homes with self-purchased planchas: mean (95% CI) CO of 3.09 ppm (1.87–4.30) vs. 12.4 ppm (10.2–14.5) for open fires. The same ranking was found for child CO exposure, but with proportionately smaller differentials (P<0.0001). The 24-h kitchen PM3.5 in the sub-sample showed similar differences (n=24, P<0.05). The predicted child PM for all 203 children (based on a regression model from the sub-sample) was 375 μg/m3 (270–480) for self-purchased planchas and 536 μg/m3 (488–584) for open fires. Multivariate analysis showed that stove/fuel type was the most important determinant of kitchen CO, with some effect of kitchen volume and eaves. Stove/fuel type was also the key determinant of child CO, with some effect of child position during cooking. The improved stoves in this community have been effective in reducing indoor air pollution and child exposure, although both measures were still high by international standards. Large donor-funded stove programmes need to aim for wider acceptance and uptake by the local families. Better stove maintenance is also required.

Introduction

Around two-thirds of the population of developing countries, around 3 billion people, still rely predominantly on biofuels (wood, dung and crop residues) for domestic energy (World Resources Institute, 1998; World Health Organisation, 2002). Burned in open fires or simple stoves, often indoors, this leads to high levels of air pollution, to which women and young children in particular are exposed (Smith, 1987; Bruce et al., 2000). This pollution increases the risk of pneumonia (Smith et al., 2000), chronic obstructive lung disease and a range of other conditions (Bruce, 2000; Boy et al., 2002). Recent estimates from the World Health Organisation indicate that indoor air pollution from solid fuels use accounts for 1.6 million excess deaths globally per year, and 3.7% of DALYs in the high-mortality developing countries (making this the fourth most important of the risk factors studied for those countries) (WHO, 2002). Most of this health burden arises from the increased risk of pneumonia in children.

Guatemala has large populations of rural people still primarily dependent on wood fuel. Bottled gas (LPG) is used in towns and the larger villages, much less so in the rural areas, and where electricity is available its use is generally restricted to lighting. Typical 24-h average PM10 levels recorded in Guatemalan homes with open wood fires are 800–1000 μg/m3 (Naeher et al., 2000). By contrast, the US Environmental Protection Agency (USEPA) 99th percentile standards for 24-h average PM10 and PM2.5 concentrations are 150 and 65 μg/m3, respectively (USEPA, 1997). Levels of carbon monoxide (CO) in homes using biomass fuels are reported in the range 2–50 ppm for 24-h averages, and 10–500 ppm during cooking (Boy et al., 2002). By comparison, the USEPA 8-h average CO standard is 9 ppm or 10 mg/m3.

Few studies have attempted to assess the population exposure of young children in such settings, or to study the impact of improved stoves, cleaner fuels or other factors that determine the level of exposure. Two recent reports from India examined respirable particulates in kitchens and living areas of 420 homes in Andhra Pradesh, India (Mehta et al., 2002; Balakrishnan et al., 2002). Fuel type (wood vs. kerosene/LPG) was found to be the most important determinant of pollution, with some contribution from the kitchen type (indoor vs. outdoor) and kitchen ventilation. No improved stoves with chimneys were included in the sample, however.

As with many poor countries, the implementation of interventions to reduce smoke pollution in rural Guatemala has been unsystematic. A variety of improved stoves have been installed through individual purchases by better-off families, and through programmes funded by NGOs, government, and external donors. Very few of these projects have been evaluated. One relatively large-scale recent programme in Guatemala was funded by the FIS (Fondo de Inversión Sociale — Social Investment Fund), typically installing between 100 and 150 stoves in rural villages of 500–600 households. The aim of this study was to study pollution levels and the exposure of young children in a typical rural setting where there had been substantial implementation of improved stoves. The specific objectives were to investigate:

  • The distribution of stove and fuel types in a rural village that had been subject to gradual uptake of improved stoves among better-off families, as well as a variety of donor-driven stove programmes including the FIS programme.

  • The levels of kitchen air pollution and exposure of children less than 18 months of age, associated with the different types of stove and fuel.

  • The influence of other household factors on kitchen pollution, and in addition the effect of child location during cooking on children's exposure.

Methods

The study site is the village of La Victoria, situated at an altitude of 1800 m in the western highlands of Guatemala. In this area, it is usual for the mother to carry the child on her back while cooking, usually until 15–18 months (Engel et al., 1997). Between 1996 and 1998, the FIS installed around 100 new stoves in the village. The most commonly used chimney stove, and that installed by the FIS, is the plancha (Figure 1) (McCracken and Smith, 1998; Boy et al., 2000). The key feature is a metal plate with (usually) three potholes that can be adjusted or closed by adding/removing concentric metal rings (Boy et al., 2000). Planchas use the same fuel as open fires (wood, agricultural residues), although wood is typically cut smaller. A survey, carried out during the dry (winter) season in January and February 1999, identified all houses in the village with children less than 18 months. From these a random sample of 250 was selected, and 204 (82%) agreed to participate. An interview survey was carried out by five trained bilingual (Spanish and Mam — the local Indian language) field workers to assess: stove type, fuel used, time since installation of stoves, smoking (mother and others in the home), lighting method, number of rooms, kitchen volume, size of eaves spaces and windows, and reported and observed position of the child during cooking.

Figure 1
figure1

The plancha wood stove, Guatemala.

Pollution and Exposure Assessment

The measurement of kitchen pollution and child exposure was based on earlier work in this area using CO as a proxy for particulates (Naeher et al., 2001). In this report of several studies, correlations (pooled for open fire and plancha homes) between 24-h CO and 24-h PM2.5 of 0.92 – 0.94 were found (Spearman rho), with coefficients of 0.50 – 0.70 for open fires and 0.89 – 0.90 for planchas). Building on this experience, the methods used for this study were as follows.

Area (Kitchen) Measurement

The 24-h kitchen CO concentration was measured in all homes using a Gastec 1D passive diffusion tube placed in the kitchen at a height of 1.25 m, 1 m from the edge of the fireplace or stove, at least 1.5 m from doors and windows and, where feasible, at least 1 m from any wall. This measurement was repeated using identical methods in a random sub-sample of 29 homes some weeks after the initial test period in order to estimate repeatability. Respirable particulate concentrations (PM3.5) were measured in the same random sub-sample over the same 24-h period using an SKC Aircheck Sampler and cyclones with 37 mm diameter Teflon filters, flow rate 2 l/min, co-located with the kitchen CO tubes. Pumps were programmed to sample 1 min out of every three over 24-h. Filters were pre- and post-weighed at the Department of Analytical Chemistry, University of San Carlos, Guatemala City.

Child Exposure Assessment

The 24-h child CO exposure concentration was measured using an identical Gastec 1D tube, placed on the child's clothing on the upper chest. The importance of keeping this tube on or near the child at all times was stressed to the mother, with advice on what to do when the child was being changed, washed or was sleeping. Tubes were left in place for 24 h, and read by the field worker at the time of collection. The 24-h child CO measurement was repeated on the same random sub-sample of homes as used for area monitoring, with readings on n=34 children (as in some homes, more than one child under 18 months was available). At interview, mothers were asked where the child was usually located during cooking. In every household, child location during cooking was also observed once or twice at setup, takedown, 12-h tube placement validation visits, or any additional visits made in order to observe the child at least once while the stove was in use.

Quality Assurance and Control

The five field workers were allocated randomly to homes in order to avoid systematic observer bias. Field supervisors validated data from 15% of the homes, which included both direct observation (80%) and repeat assessment (20%). On visits to the homes, observations were made of the position of the tube in the kitchen and on the child. CO tubes were read a second time, blind to the field worker's reading, by the research co-ordinator. This was done up to several days after collection, but it was later discovered that even though sealed with tape the mean values had risen substantially and were therefore unreliable. A number of methodological aspects of the use of CO tubes in these field conditions have been studied, including within- and between-observer variation in reading the tubes, and the effect on readings that might result from the child's CO tube being covered by the typical material (sabana) used to wrap and support the child when carried on the mother's back. To test observer variation, 30 exposed tubes were read twice by five observers, each observer being blind to their first and other observers' readings. To determine the effect of the sabana on CO tube readings, 24-h CO was measured in 23 kitchens using two Gastec 1D tubes, one in direct contact with the air, and the other next to the first but covered with two layers of the material. This was designed as an extreme test of the effect of the sabana, as it would be very unlikely that a tube on the child would be covered for the whole 24 h.

PM3.5 quality control in the field included co-located pumps and field blank filters in three houses. Validation of the true pump flow rate was carried out using the soap bubble technique by connecting a volumetric burette in series with the pump and cyclone. Laboratory procedures included the use of blanks and repeat weighing.

Data were entered into Epi-Info (Version 6), which was used for all descriptive analysis. Skewed data have been log-transformed where appropriate, although for clarity in presenting levels of pollution and exposure, means (and medians) are given and non-parametric tests are used for comparison. Regression analyses to determine (a) predicted values of PM3.5 and (b) those factors independently associated with (i) kitchen pollution and (ii) child exposure were carried out in SPSS.

Results

Results are reported on 203 of 204 homes with complete data. Almost half relied exclusively on an open fire, and around 30% on the plancha — the majority of which had been installed by the FIS project some 2–3 years previously. In all, 10% used gas (LPG), but in this community, use of LPG is usually combined with an open fire, the latter being used for longer cooking tasks and for space heating (this highland area is cold at night, light frosts being quite common in the cold, dry season).

CO Tube Readings

The investigation of observer variation in CO tube reading was reassuring. There was strong agreement between the first and second readings for all five observers (r ranged from 0.92 to 0.99), and there was no systematic difference between means for the two readings in paired comparison of all 150 observations (P=0.48). One-way ANOVA showed no significant differences between observers for either first (P=0.75) or second (P=0.99) readings. The association between 24-h CO (ppm) and 24-h PM3.5 (μg/m3) in the sub-sample of 24 homes yielded a Pearson correlation of 0.73 (P<0.001). Positively skewed CO distributions for both kitchen and child were effectively normalised by log transformation. PM3.5 was less skewed and not improved by log transformation. The Pearson correlation for log CO and PM3.5 yielded a very similar value of 0.74 (P<0.001). These findings are consistent with the earlier studies. For the whole sample (203 homes), log transformation was found to normalise both kitchen and child 24-h CO. There was a moderately strong association between the log 24 h kitchen CO and the child CO, with a Spearman correlation of r(s)=0.54 (P<0.001).

The sabana study showed that the material had little effect on recorded 24-h CO concentration: the mean CO for the uncovered tube (UT) was 6.70 ppm and that for the covered tube (CT) was 6.48 (difference 0.22 ppm, P=0.22). Correlation between UT and CT was 0.99 (P<0.0001). An Altman plot (methods comparison) showed no strong evidence of any relationship between CO concentration and the difference between the UT and CT readings (Figure 2), but a larger number of comparisons at higher concentrations are required to confirm this.

Figure 2
figure2

Methods comparison analysis (Altman plot) for uncovered (outside) and covered (inside) tubes in the study of the effect of sabana material on 24-h CO readings.

Pollution and Exposure Levels by Stove/Fuel Type

Table 1(a) shows the concentrations of 24-h kitchen CO and child CO for each stove/fuel type. Kitchen levels were markedly higher in homes using open fires, compared with those with a plancha (P<0.0001). The lowest mean (and median) was for households purchasing their own stove. Gas use yielded intermediate levels, consistent with the concurrent use of open fires for some tasks. These differences are reflected in a consistent way in the 24-h child CO levels. Although also highly significant (P<0.0001), these differences are proportionately less than those for the kitchen levels. Nevertheless, Table 1(b) shows that although there is some overlap in the distributions of child CO levels by stove type, the majority of values for the plancha homes (especially own plancha) are substantially lower than for homes with open fires. Table 1(b) also shows the mean (and median) ratios of child exposure:kitchen levels of CO for each type of stove, which suggest that the less polluting the stove, the higher the child's 24-h exposure as a proportion of the kitchen level.

Table 1 Levels and distributions of 24-h CO in kitchen and for child, by stove type.

Repeated Measurements of CO

Table 2 shows that for 24-h kitchen CO there was a significant difference between the first readings of CO and the repeats taken some weeks later. This difference was not the case for 24-h child CO. The standard deviation of the kitchen levels is of a similar order to that for the differences between the first and second readings — that is, — the variation within homes is as large as that between homes (based on two readings). A similar picture of variation is seen for child CO values, with coefficients of variation of the same order as for kitchen CO. This relatively high level of within-subject variation, relative to between-subject variation, has implications for the analysis and interpretation that are considered further in the discussion.

Table 2 Repeated measurements of 24-h kitchen and child CO.

Respirable particulate (PM3.5) levels measured in the random sub-sample of 24 homes were consistent with the differences in pollution reported in Table 1. Mean (± SD) values (μg/m3) of PM3.5 for open fires (n=11) were 1019 (547), planchas (all types, n=5) 351 (333), and gas/other (n=8) 579 (205). The difference in PM3.5 between stove types was significant (P=0.023, Kruskall Wallace), and the levels were consistent with earlier studies (Naeher et al., 2000; Albalak et al., 2001). An equation of (kitchen PM3.5=126+123(CO)) was obtained from regression of 24-h PM3.5 on CO in the sub-sample. Table 3 shows the predicted child 24-h PM3.5 exposure values based on 24-h child CO for the different stove/fuel types, derived from this equation.

Table 3 Means of individual predicted values (SD) for kitchen and child 24-h mean PM3.5 using the regression equation derived from mean CO levels.

Impact of Stoves and Other Factors

Table 4 shows the independent effects of stove/fuel type and other factors associated with (a) log kitchen CO and (b) log child CO in multiple regression. Exponentiated regression coefficients are shown to indicate the adjusted 24-h CO concentration for each variable level relative to the reference category. Thus, 0.43 for FIS plancha implies a 57% (95% CI 44–68%) lower 24-h CO concentration than for the open fire, etc. For kitchen CO, stove/fuel type was most important, although there was some effect of size of any eaves space and kitchen volume. Factors with no independent association were window size (window use during cooking was not recorded in this study), number of rooms, observer, and whether other people smoked in the house. Very few women in this community smoke.

Table 4 Multivariate analysis of determinants of log 24-h CO levels.

For child CO, stove/fuel type and observed position of the child on the two occasions were most important. The observer also appears to have had a significant effect; this was not seen for the higher kitchen levels, and may have resulted from the difficulty of reading the tubes precisely at lower levels. Factors with no independent association were reported position of the child, window size, number of rooms, eaves spaces, and whether other people smoked in the house.

Discussion

This community survey of a rural Guatemalan village community illustrates the impact of the mixed history of efforts at transition from an almost exclusive use of open fires to a quite widespread use of improved stoves and cleaner fuels. Nevertheless, almost 50% of homes with young children still use open fires. Of those using improved stoves, most received these through the FIS programme and only a minority (8%) had purchased the stove themselves. Gas is used, rarely exclusively, and just over half use electricity for lighting.

A key aspect of the methods described here is the use of mean 24-h CO as a proxy for mean 24-h particulate levels in the kitchen and exposure of children. The results from the sub-sample of homes confirm this relationship and are consistent with previous studies (Naeher et al., 2001). Although this association may not be sufficiently tight to predict an individual home or child level, it does appear useful for group means. The use of gas, particularly in combination with open fires, complicates the relationship due to very different patterns and ratios of CO and PM emissions with these two types of fuel. The studies of observer variation and effect of the sabana showed the method to be satisfactory, although multiple regression did identify an independent observer effect for the child readings, – possibly due to inconsistent reading at lower levels. There is evidence that within-house and within-child variation in CO concentrations is high relative to differences between houses and subjects respectively, consistent with a previous report on variability of indoor air pollution in rural Kenya (Boleij et al., 1989). This emphasises the extent of random variation when attempting to study between-group differences, and the value of taking repeat measurements.

Kitchen CO levels and measured PM3.5 (sub-sample) were highest for open fire homes and lowest for the plancha homes (particularly where the stove was obtained by the family) and intermediate for gas homes, consistent with previous studies in the area (Naeher et al., 2000; Albalak et al., 2001). Although it was observed in these studies that households using gas also used open fires, the amount and reasons for the continued use of open fires would be a useful subject for more detailed study. Young children's exposure reflected differences in kitchen pollution for each stove/fuel group, although proportionately the differentials were smaller. It was notable that children's exposure in homes using gas (in practice, in this community, usually a combination of gas and open fire) was similar to plancha homes. One possible explanation for this is that children are less likely to be in the kitchen during cooking tasks carried out with the open fire (for example, preparing animal feed, which is cooked slowly and requires less attention from the cook).

It was interesting that the mean ratios of child:kitchen 24-h CO increased progressively with less polluting stove/fuel groups. This may reflect protective behavioural responses and aspects of the wider environment in which the child lives. Mothers living in more polluted homes may be protecting their young children from the smoke by keeping them away from it more than in homes with a cleaner atmosphere. There was however no evidence in the study that the observed position of the child differed between stove/fuel types, or by kitchen CO concentration. Nevertheless, the observed position of the child was associated with the child's CO exposure, but without further investigation it is not possible to say whether this is cause or effect. In this setting, young children and their mothers move through a range of “micro-environments” each day, their own kitchen being only one. A child living in a very polluted home will, on balance, encounter less polluted micro-environments when outdoors and visiting other homes, and vice versa. In a community where around half the homes use open fires and the rest a mixture of improved stoves and LPG, there is ample opportunity for mixing. Despite this, the distributions of child CO for planchas and open fires are still reasonably distinct (Table 1b), indicating that the improved stoves do reduce exposure for the majority of users despite variations in stove use, condition, and the contributions that other sources of pollution (from other homes, etc.) make to daily mean exposure.

Although, as might be expected, physical characteristics of the house such as eaves spaces and kitchen volume did influence kitchen pollution levels, in this analysis these factors were not independently associated with children's exposure. The type of stove/fuel, mediated via kitchen pollution levels, was by far the most important factor, with the observed location of the child also having some influence (in this study, the reported usual location during cooking was not associated). As noted, it is possible that the location of the child is a behavioural response by the mother wishing to protect her child, but without further study it would be unwise to comment further. Furthermore, the large within-house and within-child variation in 24-h exposures found in this study will have reduced the statistical power of any analysis of determinants. A larger study and/or one with longer monitoring periods with allowance for within-individual variation may well identify the influence of house structure and ventilation on children's exposure. These findings are consistent with the Indian study by Mehta that, despite being carried out in another continent, with different house type and climate, also identified the stove/fuel type as being most important, with some effect of kitchen type and ventilation. Another recent study, from rural Kenya, where cleaner fuels were used very little, found that hoods (venting through the roof) were most effective in reducing kitchen 24-h PM3.5 and CO and personal (woman) CO, while enlarged eaves spaces also contributed (Bruce et al., 2002).

The predicted child 24-h PM3.5 exposures for all stove/fuel groups are high compared to USEPA guidelines (USEPA, 1997), even for homes using planchas. These results are higher than, although consistent with, levels measured in an earlier study directly measuring kitchen, mother and child average 22-h PM2.5 levels for homes with open fires, planchas and gas stoves, and with no fire alight, in this area of Guatemala (Naeher et al., 2000). Using direct measurement of personal particulate exposure with pumps, Naeher reported mean 10–12 h (daytime) PM2.5 levels for children less than 15 months of 279 μg/m3 (±SD of 19.5) for the open fire, 170 μg/m3 (±154) for the plancha and 149 μg/m3 (±69) for gas. The 22 h background (no fire alight) mean PM2.5 was 56.2 μg/m3. There may be a number of reasons for lower personal exposure levels in that study, including the different (smaller) particle diameter cut-off being used in the earlier study, sampling for only 10–12 h which may have selected a less heavily exposed portion of the day, and random variation (95% confidence intervals in both studies are wide). The presence of a significant background level of particles results from the widespread use of wood fuel in the community, and the resulting smoke entering homes. Nevertheless, the fact that children living in homes with relatively good quality chimney stoves and gas cookers (exclusive use) still record such high levels of exposure is of concern, and emphasises the need for community-wide strategies to improve access to clean fuels as a medium-term goal.

Donor-funded programmes such as the FIS can be criticised on grounds that these do little to stimulate local market-based sustainable conditions for wider distribution and uptake of stoves (von Schirnding et al., 2002). However, even though quite a number of the stoves were in quite poor condition, this study has shown that households receiving FIS planchas had substantially less-polluted kitchens and less heavily exposed infants, several years after the stoves were installed. The impact these lower levels of exposure would have on health outcomes, in particular the incidence of ALRI, remains to be firmly established through intervention studies such as that underway in Guatemala (Dooley 2003). However, some evidence is becoming available to suggest there would be some benefit at the level of exposures seen for open fires and planchas in the current study (Ezzati and Kammen, 2001). In the meantime, this study shows clearly that stove programmes such as the FIS — although limited in some respects — can have quite a substantial impact on indoor air pollution in poor rural populations, and there would seem to be the potential to increase this impact further, even before a transition to cleaner fuels becomes economically feasible and culturally appropriate. Although house construction and standards, and the location of the child, can contribute and should also receive greater attention, it is the type of stove and fuel that are most important.

References

  1. Albalak R., Bruce N., McCracken J.P., and Smith K.R. Indoor respirable particulate matter concentrations from an open fire, improved cookstove, and LPG/open fire combination in a rural Guatemalan community. Environ Sci Technol 2001: 35: 2650–2655.

  2. Balakrishnan K., Sankar S., Padmavathi R., Arnold J., Mehta S., Smith K.R., Rao C., Kumar S., Kumar P., Akbar S., and Lvovsky K. Respirable particulate levels in rural households of Andhra Pradesh, India — daily concentration and time activity data. Proceedings of the ninth Indoor Air Conference, Monterey, CA, 2002, pp. 560–565.

  3. Boleij J.S.M., Ruigewaard P., Hoek F., Thairu H., Wafula E., Onyango F., and de Koning H. Domestic air pollution from biomass burning in Kenya. Atmos Environ 1989: 23: 1677–1681.

  4. Boy E., Bruce N., Smith K.R., and Hernandez R. Fuel efficiency of an improved wood-burning stove in rural Guatemala: implications for health, environment and development. Energy Sustainable Dev 2000: 2: 21–29.

  5. Boy E., Bruce N., and Delgado H. Birthweight and exposure to kitchen wood smoke during pregnancy. Environ Health Perspect 2002: 110: 109–114.

  6. Bruce N., Perez-Padilla R., and Albalak R. Indoor air pollution in developing countries: a major environmental and public health challenge. Bull WHO 2000: 78: 1078–1092.

  7. Bruce N., Bates E., Nguti R., Gitonga S., Kithinji J., and Doig A. Reducing indoor air pollution through participatory development in rural Kenya. Proceedings of the ninth Indoor Air Conference, Monterey, CA, 2002, pp. 590–595.

  8. Dooley E. New stoves for better children's health? Environ Health Perspect 2003: 111: A33.

  9. Engel P., Hurtado E., and Ruel M. Smoke exposure of women and young children in highland Guatemala: predictions and recall accuracy. Hum Organ 1997: 56: 408–417.

  10. Ezzati M., and Kammen D. Quantifying the effects of exposure to indoor air pollution from biomass combustion on acute respiratory infections in developing countries. Environ Health Perspect 2001: 109: 481–488.

  11. McCracken J.P., and Smith K.R. Emissions and efficiency of improved wood burning cookstoves in highland Guatemala. Environ Int 1998: 7: 739–747.

  12. Mehta S., Smith K.R., Balakrishnan K., Sankar S., Padmavathi R., Kumar S., and Akbar S. Using household characteristics to predict respirable particulate levels in rural households of Andhra Pradesh, India. Proceedings of the ninth Indoor Air Conference, Monterey, CA, 2002, pp. 596–601.

  13. Naeher L.P., Leaderer B.P., and Smith K.R. Particulate matter and carbon monoxide in highland Guatemala: indoor and outdoor levels from traditional and improved wood stoves and gas stoves. Indoor Air 2000: 10: 200–205.

  14. Naeher L., Smith K.R., Leaderer B.P., Neufeld L., and Mage D.T. Carbon monoxide as a tracer for assessing exposures as particulate matter in wood and gas cookstove households of highland Guatemala. Environ Sci Technol 2001: 35: 575–581.

  15. Smith K. Biofuels, Air Pollution And Health: A Global Review. Plenum Press, New York, 1987.

  16. Smith K., Samet J., Romieu I., and Bruce N. Indoor air pollution in developing countries and acute respiratory infections in children. Thorax 2000: 55: 518–532.

  17. United States Environmental Protection Agency (USEPA). Revisions to the National Ambient Air Quality Standards for Particulate Matter. Federal Register July 18, 1997: 62(138).

  18. Von Schirnding Y., Bruce N., Smith K.R., Ballard-Tremeer G., Ezzati M., and Lvovsky K. Addressing the Impact of Household Energy and Indoor Air Pollution on The Health of the Poor: Implications For Policy Action and Intervention Measures. World Health Organisation, Geneva, 2002.

  19. World Health Organisation. The World Health Report 2002: Reducing Risks, Promoting Healthy Life. WHO, Geneva, 2002, pp. 69–70.

  20. World Resources Institute, UNEP, UNDP, and World Bank. 1998–99 World Resources: A Guide to The Global Environment, Oxford University Press, Oxford, UK, 1998, pp. 65–67.

Download references

Acknowledgements

This study was funded by a grant from the Department of Child and Adolescent Health and Development, WHO, Geneva. John McCracken was supported by a Fulbright Fellowship.

Author information

Correspondence to Nigel Bruce.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Bruce, N., McCracken, J., Albalak, R. et al. Impact of improved stoves, house construction and child location on levels of indoor air pollution exposure in young Guatemalan children. J Expo Sci Environ Epidemiol 14, S26–S33 (2004) doi:10.1038/sj.jea.7500355

Download citation

Keywords

  • developing countries
  • biomass fuels
  • indoor air pollution
  • improved stoves
  • children.

Further reading