MS3 Symposium Title: RESPIRE—The Guatemala Randomized Intervention Trial. Symposium Organizers: Kirk R. Smith, Nigel Bruce, and Byron Arana. MS3-01 Objectives: The objectives of this study were to site and conduct a randomized, controlled trial (RCT) to examine the improvement in child acute lower respiratory infection (ALRI) by introduction of a chimney stove designed to reduce indoor exposures to biomass smoke in a rural developing country setting. Background: After unsuccessful efforts in the 1980s to fund the RCT in Nepal, an international committee was established by WHO, Geneva, in 1991 to locate the best site for the RCT. Based on criteria related to local health, exposure, logistics, and institutional conditions, a dozen sites were examined in Asia, Africa, and Latin America. Highland Guatemala was chosen and a range of pilot studies undertaken in the 1990s to develop the data needed to design and successfully fund the RCT, which began in 2001. Setting: The chosen site is home to a largely indigenous primarily agricultural Mayan population living at 2700 to 3000 m in the western Guatemalan highlands. Most of the population speaks a local language, Mam, with Spanish as a second language for many. Nearly all households in the area use only wood for all household fuel requirements. Preintervention ALRI rates in young children were estimated at 0.5 per child-year. Methods: A rapid assessment was undertaken among 5500 households in the area to screen for recruitment based on use of an open fire for cooking, presence of a pregnant women or child under 4 months, migration patterns, and willingness to participate. To meet statistical goals (two-sided α = 0.05, power = 0.80, 25% effect), 534 households were eventually recruited, 518 (97%) of which contributed to the final dataset. After baseline household and indoor air quality (IAQ) surveys, households were randomly assigned to receive an improved chimney stove, the plancha, at the start of the study or at the end, when the study child reached 18 months, or a household dropped out. Pilot work showed that planchas were popular, sturdy, and capable of substantially improving kitchen IAQ. After stove installation, weekly household visits were conducted by trained fieldworkers to identify potentially ill children, who were then referred to study physicians (see companion abstract for details). Planchas were repaired by the study when structural problems appeared. Anthropometry was conducted regularly and birth weights recorded within 48 hours after delivery. Verbal autopsies were conducted on all child deaths. Quarterly personal 48-hour CO exposures were obtained for all study children and mothers. Intensive IAQ monitoring for CO and PM was conducted in a 13% subsample along with outdoor monitoring and assessment of exposures from use of wood-fired saunas (temascals). Results: The 518 children followed contributed nearly 600 child-years of surveillance. Randomization was successful because there were no significant differences between the 2 groups for any of the dozens of household, demographic, and social variables examined, except for leaks in the roof. At 8%, the dropout rate was within the 20% projected in sample-size calculations and was due primarily to migration (50%) from the area. There was a slightly higher dropout rate (9.3% P = 0.39) for intervention households. Discussion: Although not easily quantified, absolutely essential for successful implementation were the hospitality and cooperative spirit of the local communities and the field workers recruited from them.Table

Few books about the environment have generated as much heated debate as Bjørn Lomborg's ‘The Skeptical Environmentalist: Measuring the Real State of the World’, published by Cambridge University Press in 2001. A flavor of the controversy can be gleaned from a series of reviews and rebuttals published in ‘Scientific American’ (Rennie 2002). In general, most positive reviews appeared in the popular press (e.g., ‘The Economist’, ‘Washington Post Book Review’, ‘The Wall Street Journal’) and most negative reviews appeared in the scientific press (e.g., ‘Science’, ‘Nature’, ‘Bioscience’). Although ‘The Skeptical Environmentalist’ (TSE) addresses a number of environmental health issues, voices from the environmental health community have not been prominent among the participants in this debate. Now that the dust from the initial stampede to praise and condemn the book has settled, we will explore lessons to be learned from TSE and the associated debate from an environmental health perspective.

ISEE-189 Introduction: Several reports have described the risk of burns to children from open wood fires in poor countries, but there is little information available on community incidence and none describing the impact of preventative interventions. A randomised controlled trial currently underway in Guatemala designed to assess impacts of an improved stove on pneumonia and other health outcomes provides an opportunity to study this issue. Methods: The study sample includes 507 homes using an open fire, recruited from a rural area of western Guatemala. Eligible homes had a child under 4 months or a pregnant woman (providing the index children for respiratory illness outcomes). Health surveillance for the index child is by weekly home visits until 18 months of age, with episodes of burns/scalds recorded. Household interview includes questions on burns/scalds in older siblings of the index child, repeated 6 monthly. Homes were randomly allocated to receive a plancha chimney wood stove immediately, or when the index child reaches 18 months. Results: Index children being very young and closely supervised initially had a very low rate of burns/scalds. Preliminary analysis of 34 cases in these children after 11 months of follow-up showed similar numbers of burns in plancha and control groups (although due to OF in 2 plancha homes), with a trend towards burns in open fire homes being in older children (9.7 vs 7.1 months, p=0.1). Among 1040 siblings aged up to 8 years in study homes, 22 reported a burn/scald in the 6 months prior to baseline (42 per 1,000 per year), all under 6 years. Eight resulted from falling into the fire, eight from scalds, and six from a hot object, pot or other cause. Fifteen left scars, seven larger than 2 cm diameter, one 3x14 cm. A further 40 episodes were recalled as having occurred prior to the last 6 months with similar causes, five leaving scars at least 5 cm in diameter. Discussion: An improved stove has the potential to prevent the majority of burns/scalds, particularly the most severe episodes resulting from falling into an open fire. Initial findings indicate that burns still occur with the plancha, but this could be in part lack of experience with young children near an unfamiliar, new type of stove, and a clearer picture may emerge as they get older. This study will provide information on incidence, severity and causes, and evidence on intervention effectiveness over 18 months in children up to around 10 years. Education in the safe use of improved stoves may well be required, as is the case for home accident prevention in any society.

ISEE-170 Introduction: There is accumulating evidence that indoor air pollution from household use of biomass fuels increases the risk of developing chronic obstructive pulmonary disease (COPD). Women receive the greatest exposures due to their activities in the home and kitchen. We will, in the first randomized control trial ever with air pollution, investigate the effect of reducing indoor air pollution (with improved stoves) on women’s lung health. Here we present preliminary results from the baseline assessment. This gives a picture of the lung health among women who experience indoor air pollution at the expected high end of the dose-response curve. Methods: This study is conducted in a Mayan Indian community in the Guatemalan highlands and includes only households using an open woodfire for cooking and spaceheating. It comprises 504 women of child-bearing age visited in their homes by trained fieldworkers who speak the local language. They were asked questions about chronic airway symptoms based on standard questions (MRC/IUATLD). The original format of key cough and phlegm questions were, however, not usable, and they were modified through extensive local piloting to versions for which the time patterns of symptoms were understood by the women. The women underwent spirometry testing and had their carbon monoxide (CO) levels in exhaled breath measured as a proxy for recent air pollution exposure. Measurements of CO from kitchens and personal monitors will also be available as will evidence of the relationship between CO and particle pollution in these households. Results: Prevalence of selected airway symptoms is presented in Table 1 and crude spirometry results in Table 2. Obstructive airway symptoms were frequently reported; 9.3% had persistent cough or phlegm for at least 3 months consecutively. Frequent phlegm experienced for at least three months was associated with a decrease in FEV1/ FVC ratio of 4.7% (p<0.001). Neither passive smoking (23.6%) nor pregnancy (32.5%) influenced lung function results. Only one woman smoked and two were ex-smokers, hence smoking was not included in the analyses. Adjusting for height and age, CO in exhaled breath was negatively associated with FEV1/FVC; one ppm increase in CO corresponded to a 0.15 unit fall in FEV1/FVC (p=0.021). The mean concentration of CO in exhaled breath was 7.5 ppm (sd 4.1).Table 1: Prevalence of selected airway symptoms among women in highland Guatemala using biomass fuels for cooking and spaceheating.Table 2: Age, height and lung function among the study women.Conclusion: Obstructive airway symptoms are frequent in this highly-exposed female population, and exposure to CO in indoor air is associated with a decrease in lung function.

Introduction In preparation for a randomized trial in highland Guatemala, a cross sectional survey was conducted in rural communities in the District of San Marcos, Guatemala to determine the distribution and proportion of households using indoor open fires. The randomized trial is designed to measure the effect of indoor air pollution on acute respiratory infections in children, in which improved chimney stoves will be introduced into houses using open fires for cooking. Methods Based on information from health authorities, village informants and staff field visits; we established the highlands of the Department of San Marcos as a suitable area to conduct the survey. Information on elevation, number of households/community, migration during harvest season, household and kitchen characteristics was collected to select potential study sites. Communities in five municipalities in the Department of San Marcos fulfilled criteria as potential study sites. A structured questionnaire, to collect data for final village selection, included the following main conceptual categories: Basic demographics (age of family members, gender, number of children < 7 months and pregnant women living in the house migration during harvest season), kitchen characteristics and cooking practices and family economic status based on households characteristics. Data was collected over a four week period. Frequency tables and histograms were prepared to present both a tabular and a graphic representation of the results. Cross tabulations were generated to obtain chi-square test results and p values. The p < 0.05 level was the criteria used thorough out to define statistical significance. Univariate and multinomial logistic regression were used to explore socioeconomic characteristic of the households to be studied. Results A total of 5,365 households were surveyed in 45 villages in five municipalities of the Department of San Marcos, Guatemala. Demographic characteristics and behaviors. The mean age of the interviewees was 39.27 years, (Sd = 14.86; range = 15–96). The mean number of persons/household was 6.26 (Sd = 2.7); 648 households (12.07%) reported a pregnant woman and 736 (13.71%) reported a child under 7 months of age. When caretakers were asked where they leave their children when cooking, 86.1% responded that they carried their infant on their back, while 8.5% reported they leave their baby outside the kitchen. The reported duration the baby is carried on the mother's back was variable: 31.3% reported up to 12 months, 42.2%, up to 18 months and 25.6% up to 24 months. Household characteristics. Overall, 52% of households use open fires for cooking Most kitchens, (84.0%), were located in a separate structure from the house. Households using open fires, however, were more likely to have the kitchen located inside the house (p = 0.002; 95% CI = 1.09–4.87). Households with open fires as compared with those using improved stoves were statistically significantly poorer, as evidence by living and cooking in the same room, less adequate housing, less likely to have electricity in the home and more frequent work-related migration (p < .0001). Among those using open-fires, households without electricity and with poorer housing conditions were more likely to be migrate laborers, OR = 1.5 (95% CI 1.3–1.8)

Introduction Pneumonia remains the principal cause of death of children under 5 in developing countries. Evidence from nearly 20 studies associates exposure to pollution from biomass fuels with increased risk of acute lower respiratory infections (ALRI). However, all these studies have been observational designs, very few with exposure assessment, and with inconsistent case ascertainment and definition. To strengthen this evidence, a randomised controlled trial (RCT) is currently underway in Guatemala to measure the effect of an improved stove on ALRI incidence. This provides the first opportunity to study health effects of air pollution in a randomised intervention, largely free of confounding which typically complicates interpretation in environmental epidemiology. This paper describes the methods used for standardised case-finding. Method The sample includes 500 children in a poor rural area of north-western Guatemala. At recruitment, children were aged less than 4 months, or in utero. Each child is being followed to age 18 months. Following baseline questionnaires, a random 50% of homes received plancha stoves-a well–accepted local product known to reduce indoor air pollution by up to 80–90%. Case finding is carried out at three levels: (a) household visits, (b) referral to a study physician, and (c) referral to the district hospital. Fieldworkers trained in IMCI case recognition visit each home weekly, using standard questions and examination (including respiratory rate) to identify possible ALRI cases. This process is very sensitive, but not very specific. Anthropological work was undertaken to identify symptom terms in the local (Mam) language with correct meanings. Children meeting ALRI criteria are referred to a physician, based in a community centre: this helps to maintain blindness of the physician to the intervention/control status of each child's home. Physician assessment utilises standardised clinical criteria, based on prior WHO studies and re-enforced through clinical training sessions. Such sessions are also carried out at regular intervals to maintain consistency. All children with signs of ALRI are referred for chest X-ray (CXR), and RSV antigen testing and pulse oximetry (PO) are carried out. The use of CXR and PO provide additional ‘objective’ assessment of each child, and RSV testing will allow comparison of IAP risk in bronchiolitis and pneumonia cases. Severely ill children are admitted to hospital, and additional clinical information collected at discharge. All deaths are being assessed by verbal autopsy. Conclusion Primary case definition is determined by the study physician, combining clinical findings, CXR, and hospital information, with PO providing a measure of severity. Independent, blind assessment of CXRs and verbal autopsy reports will also be available. Experience with this system of case finding will be reported, including numbers of ALRI cases treated at home and admitted, and %RSV positive. Also analysed will be the differences in numbers and characteristics of ALRI cases defined by (a) IMCI criteria at weekly visit, (b) physician examination with CXR, and (c) CXR findings only. Practical issues in maintaining standardised procedures, and referring children for CXR and admission in a rural population fearful of the hospital will also be reported.

Few books about the environment have generated as much heated debate as Bj ̆rn Lomborg's `The Skeptical Environmentalist: Measuring the Real State of the World', published by Cambridge University Press in 2001. A flavor of the controversy can be gleaned from a series of reviews and rebuttals published in `Scientific American' (Rennie 2002). In general, most positive reviews appeared in the popular press (e.g., `The Economist', `Washington Post Book Review', `The Wall Street Journal') and most negative reviews appeared in the scientific press (e.g., `Science', `Nature', `Bioscience'). Although `The Skeptical Environmentalist' (TSE) addresses a number of environmental health issues, voices from the environmental health community have not been prominent among the participants in this debate. Now that the dust from the initial stampede to praise and condemn the book has settled, we will explore lessons to be learned from TSE and the associated debate from an environmental health perspective.

Introduction More asthma and allergy studies from developing countries are needed both for public health reasons, and to increase our understanding of asthma in settings with non-western lifestyle, and different exposure to environmental factors and parasites. Methods We used standardized ISAAC (International Study of Asthma and Allergy in Childhood) written and video questionnaire with additional environmental questions. 4551 children from class 7 and 8 in 42 randomly selected schools in the Kathmandu Vally participated. Response rate was >90%. The initial analysis included chi-square tests, t-tests and Fisher's exact tests. In logistic regression models, we included sex, age, socio-economic status, urban/peri-urban home and ethnicity. All other relevant factors that were at least borderline significant in the univariate analysis were also included in the first models, and kept if considered independent risk factors (p < 0.10). Results The mean age was 13.8 years. The prevalence of current wheeze (defined as wheezing symptoms at least once during last 12 months) was 8.9% (95%CI: 8.0–9.7). The prevalence of “severe asthma” (defined as at least one of the following: Speech limitation to 1–2 words at a time between breaths due to wheeze last year, sleep distrbance >1 time per week in average due to wheeze last year or >12 wheezing episodes last year) was 7.3% (95% CI 6.5–8.1). The prevalence tended to be higher in the peri-urban population. Boys had significantly higher prevalence than girls in the urban but not the peri-urban population. In the univariate analysis “current wheeze” was significantly associated with farming animals (p < 0.001), traffic (p < 0.05), passive smoking (p < 0,001), modern mattress (p < 0,001), Whether the child had ever smoked (p < 0.01), sex (p < 0.01), whether the child had ever smoked (p < 0.01), sex (p < 0.01), whether the child had ever moved (p < 0.001) and electricity in the home (< 0.001). There were significant differences between the ethnic grous, with the tibeto-burmese groups having less asthma (p < 0.001). Stove type/fuel type used for cooking seems important,-details published elsewhere. Having one brother or sister was protective in the univariate analysis, compared to two, three, more than three or none brothers and sisters. Those who shared room with 1–3 other persons had significantly less asthma than children with their own room. “Severe wheeze” was associated with the same risk factors as current wheeze, but the association did not reach significance for number of people sharing the room. In the logistic regression model for current wheeze the following risk factors remained highly significant; passive smoking (OR = 1.5), modern mattress (OR = 1.5), cooking on open fire (OR = 2.2), moving ever (OR = 1.8). For severe wheeze: personal smoking ever (OR = 2.5), passive smoking (OR = 1.3), cooking on open fire (OR = 2.7), urban residence (OR = 0.45) and moving ever (OR = 1.4). Exposure to pets or farming animals was not significantly related to asthma. Conclusion The prevalence of asthma in Kathmandu is moderate, with a relatively high proportion of severe asthma.

Background The incidence of catracts is higher among women than men in developing countries. Risk factors for cataract include active cigarette smoking, exposure to intense sunlight, and severe diarrhea. Prior research suggests that exposure to smoke from household solid fuel may be a risk factor. Cooking and heating with solid fuels such as wood/wood-logs, dung-cake, coal and charcoal are predominant in developing countries where women are primary cooks. Methods An incidence-density casecontrol study was conducted to look at the relationship between exposure to smoke from household solid fuel and risk of cataracts among women visiting an eye hospital situated at the Nepal-India border. Cases (n = 206) were women patients, aged 35–65 years, who resided in neighboring districts in Nepal and India, and who visited the hospital for the first time with cataracts in one or both eyes. Controls (n = 203) were selected from among patients visiting the same hospital from same districts who had diagnoses of refractive error or astigmatism. Patients with a history of diabetes mellitus were excluded from both case and control groups. An interviewer-administered questionnaire included questions on education level, area of residence (urban v rural), cooking fuel, stove type, kitchen type and location, use of mosquito coil and incense, history of severe diarrhea, present food habits, smoking status, alcohol consumption, history of vitamin use, household crowding, time spent working outside in the sunlight, vaccination during childhood, house type, socio-economic status. Results Age was not a significant confounder because adjustment for age did not affect other relative risks in multivariate logistic models. In univariate analyses, the following variables were statistically significant as a risk factors for cataract: stove type (un-improved vs. improved: OR = 2.99, 95% CI: 1.93–4.63), fuel (solid vs. liquid: OR = 3.19, 95% CI: 1.99–5.1), education (illiterate vs. literate: OR =: 2.40, 95% CI: 1.44–4.00), area of residence (rural vs. urban: OR = 3.34, 95% CI: 1.76–6.25), kitchen location and ventilation (partially vs. fully: OR = 2.39 CI: 1.51–3.79), working in the sun (yes vs. no: OR = 1.51, 95% CI: 1.00–2.26). Variables such as smoking (ever vs. never: OR = 1.16, 95% CI: 0.73–1.85), history of severe diarrhea in the last five years (yes/sometime vs. no: OR = 1.22 CI: 0.81–1.82) were non-significant in the univariate analysis. In Multivariate logistic regression modeling, poor ventilation (OR: 1.89, 95% CI: 1.16–3.07), rural residency (OR: 2.15, 95% CI: 1.07–4.35), and solid fuel use with an unimproved stove (OR: 2.21, 95% CI: 1.29–3.82) were risk factors for cataracts. In our study 76% of cases and 79% of controls reportd that they did not smoke (p-value: 0.528) and only 3% of cases and controls reported that they were ex-smoker. Conclusions These preliminary analyses support the hypothesis that smokes from household solid fuel (wood, cow dung cake, crop residues, coal, coke, lignite), along with lack of ventilation in the kithcen, increases the risk of cataracts among women. In addition, rural residence seems to be a risk factor for cataracts.

Objectives: To estimate the prevalence and severity of asthma among indigenous children in Guatemala, and the association with cooking on open fires. Methods: In 1058 homes with children 4-6 years of age, we interviewed the mothers using standardized ISAAC procedures and questionnaire. Results: Only 3.4% reported wheezing symptoms in the last 12 months. In a logistic regression model, use of open fire for cooking was a significant risk factor for a number of asthma symptoms, with odds ratios varying from 1.81 to 3.21. For the different cooking technologies (1=improved stove with chimney, 2=mixture of gas and open fire, 3=open fire) we observed significant trends for some of the symptoms. Conclusions: The asthma prevalence is low among indigenous children in Guatemala, compared to other populations in Latin-America. Use of open fire for cooking, may be an important risk factor for asthma symptoms and severity. INDEX TERMS Allergies and asthma, Cooking technologies, Fuel and combustion pollutants, Surveys, Residences. INTRODUCTION The objectives of this study were to estimate the prevalence and severity of asthma among Native-American children in Guatemala, and to describe the relationship between cooking technology and the prevalence and severity of asthma. Asthma is the most common chronic disease among children worldwide, with recent increases in many countries. Despite a large body of information regarding the prevalence of asthma and allergic diseases, mostly from developed countries, the epidemiology of asthma is still in some respects similar to that of cancer epidemiology in the 1960s, when the international patterns of cancer incidences were studied (ISAAC -98). Those studies revealed striking international differences that gave rise to many new hypotheses, tested in further epidemiological studies that identified previously unknown risk factors. Whole populations may be exposed to risk factors for disease and the patterns may be apparent only when comparisons are made between, rather than within, populations. This underlines the importance of doing standardized asthma studies in countries like Guatemala. * Contact author email: mschei@uclink.berkeley.edu Proceedings: Indoor Air 2002

Data on all patients with acute myocardial infarction who were treated in Harstad District Hospital in 1995 were analysed. Of the 170 patients, 24% received thrombolytic treatment. Thrombolytics were withheld from 15% of the patients, although there were no contraindications present. Thrombolytics were administered two hours and 18 minutes (mean) after admission to hospital and seven hours after the onset of symptoms. 54% of the patients were admitted to hospital within six hours and 73% within 12 hours. In-hospital delay before the administration of thrombolytics is too long. In Norwegian hospitals this factor has only been analysed to a minor degree. Despite a fairly standardized treatment regimen for thrombolytics, how frequently it is used probably varies from hospital to hospital. There is great potential for improving thrombolytic treatment. The results of our analyses have resulted in an extensive change in routine in our hospital.