© 2002 American Thoracic Society
Infant Respiratory Symptoms Associated with Indoor Heating SourcesCenter for Perinatal, Pediatric and Environmental Epidemology, Department of Epidemiology and Public Health, Yale University School of Medicine, New Haven, Connecticut; and Department of Environmental Medicine, University of Rochester School of Medicine and Dentistry, Rochester, New York Correspondence and requests for reprints should be addressed to Elizabeth W. Triche, Ph.D., Yale University School of Medicine, Department of Epidemiology and Public Health, 60 College Street, P.O. Box 208034, New Haven, CT 06520-8034. E-mail: elizabeth.triche{at}yale.edu ABSTRACT This study examined the effects of indoor heating sources on infant respiratory symptoms during the heating season of the first year of life. Mothers delivering babies between 1993 and 1996 at 12 hospitals in Connecticut and Virginia were enrolled. Daily symptom and heating source use information about their infant was obtained every 2 weeks during the first year of life. Heating sources included fireplace, wood stove, kerosene heater, and gas space heater use. Four health outcomes were analyzed by reporting period: days of wheeze, episodes of wheeze, days of cough, and episodes of cough. A large percentage of infants had at least one episode of cough (88%) and wheeze (33%) during the heating season of the first year of life. Wood stove, fireplace, kerosene heater, and gas space heater use was intermittent across the study period. In adjusted Poisson regression models controlling for important confounders, gas space heater use was associated with episodes and days of wheeze. Wood stove use was associated with total days of cough, and kerosene heater use was associated with episodes of cough. Fireplace use was not associated with any of the respiratory symptoms. Use of some heating sources appears related to respiratory symptoms in infants.
Key Words: cough • gas space heaters • infant respiratory symptoms • kerosene heaters • wheeze • wood stoves Infants spend the majority of their time breathing the indoor air of the home environment and have frequent respiratory symptoms. Although the long-term clinical prognosis of infants who experience symptoms such as wheeze and cough in the first year of life is uncertain (1), infant respiratory symptoms represent a health impact with enormous associated costs, including medical care visits, hospitalizations, and lost productivity for parents. Although infants have some passive immunity against infections, acquired transplacentally and through breast milk, the small caliber of the infant bronchial tree means obstruction to airflow may occur with relatively minor increase in nasal or tracheobronchial secretions or airway surface edema. Exposure to particulate matter has been linked to increases in hospital admissions for respiratory diseases (2–4), acute respiratory symptoms (5–9), lower respiratory illness (7–12), and decreases in pulmonary function (7, 10, 12–17). Some studies have focused on populations with existing respiratory disease, such as people with asthma (7, 9–11, 15, 16, 18–20), for whom exposures may exacerbate pre-existing disease. Often, these studies have focused on ambient air exposures. However, assessing exposures in these studies has proven challenging because of the complex nature of the pollutant mix, including both particle and gas-phase components, as well as the variability of concentrations in space and time. The complex nature of the ambient aerosol and strong correlations among ambient concentrations of particulate matter 10 µm or less in diameter (PM10), PM2.5, SO42-, and H+ make it difficult to untangle the role that different physical and chemical aerosol attributes may play in the observed effects. These studies have typically not controlled for the wide range of personal and household characteristics that may be potential confounders for outdoor exposures. Studying the indoor environment provides an opportunity to clarify some of these effects in infants. Heating sources such as gas, kerosene, wood, and tobacco products are important producers of particulate matter in the home. The nature of the chemical composition of particles varies by source of combustion. Environmental exposure chamber and field studies have identified unvented kerosene heaters as an important indoor source of PM2.5, SO42-, H+, SO2, and NO2 (21, 22). Unvented or poorly vented gas heaters produce high levels of NO2, which is released directly into the home. Wood stoves and fireplaces, even when vented to the outside, may emit high concentrations of pollutants into the indoor air, including respirable particles (23), carbon monoxide, and NO2 (24), especially during the winter when the home tends to be tightly closed. Studies of the health effects of these indoor heating sources have been inconsistent. Differences between findings could be due to small sample sizes, failure to adequately control for potential confounders, or differences in definitions of outcomes and levels of exposures. Further, the intermittent use of these heating devices in the home has been overlooked. None of the prior studies have accounted for intermittent use patterns of these sources over the course of the heating season. This study was specifically designed to examine in detail the effects of secondary home heating devices on respiratory symptoms. We test the hypothesis that use of unvented kerosene heaters, gas space heaters, fireplaces, or wood stoves in the home increases the rate of upper and lower respiratory symptom episodes and days during the heating season of the first year of life. METHODS
Study Population At enrollment, a standardized questionnaire was administered by a trained research assistant during a home interview when the infant was between 3 and 5 months of age. At the home interview, a research assistant described the study in detail to the respondent (infant's mother) and obtained informed consent. The home interview took about 50 minutes to administer and gathered detailed information about the health history of the infant (e.g., birth weight, respiratory illness), infant care (e.g., breastfeeding, daycare), household demographic data (e.g., maternal age, race, marital status, education, number of children in household, income), health status of mother (e.g., self-reported history of allergies or physician-diagnosed asthma), smoking history of mother, medication use of infant and mother, dwelling characteristics (e.g., type and age of residence, number of rooms, heating system), and use of secondary heating sources (i.e., fireplaces, wood stoves, gas space heaters, and kerosene heaters).
Respiratory Symptom Episodes and Symptom Days
Exposure Assessment
Data Analysis For each biweekly monitoring period, the number of days the infant experienced the symptom and the number of new episodes of the symptom were summed. An episode is defined as having reported a symptom for one or more days after having had two consecutive symptom-free days. Because our symptom data are in the form of counts (number of days of symptoms and number of episodes), we conducted Poisson regression analysis to examine the relationship between source use and respiratory symptoms, using the GENMOD procedure in SAS (see online data supplement for more detailed information). Separate models were run for the following outcomes: (1) episodes of wheeze, (2) days of wheeze, (3) episodes of cough, and (4) days of cough in each monitoring period. Medians and interquartile ranges (IQR, 25th–75th percentiles) were calculated as a measure of variability in nonnormally distributed data. Initial unadjusted models examined the association between use of each source separately and the respiratory symptom outcomes. Final Poisson regression models were chosen on the basis of a backward elimination strategy for each outcome. All source use variables—kerosene heater, gas space heater, fireplace, and wood stove use—were retained in final models as continuous variables. Other variables were retained in the model if they were independent risk factors (p < 0.10) or if their removal resulted in an increase or decrease in the parameter estimates of the exposure variables of 10% or more. RESULTS Mothers enrolled in this study tended to be college-educated, white women (Table 1) . Overall, 44% had allergies, but 9% had a history of physician-diagnosed asthma. The majority of families owned their homes and lived in one- to three-family dwellings. Nearly two-thirds of the study homes did not have a gas source. Most infants lived in families with one or two children. Approximately the same numbers of male and female infants were enrolled, some one-third was breastfed for more than 5 months, and a similar percentage was not breastfed at all. One-third of infants did not attend daycare, whereas one-quarter attended daycare an average of 4 hours/day or more. Only 6% of the infants were classified as being low birth weight (less than 2,500 g) based on mother report.
Wood source users (i.e., fireplace and wood stove) tended to be white, more highly educated, own their own homes, live in one- to three-family dwellings without a gas source, and breastfeed their infants (Table 1). The only factors related to gas space heater use were the presence of a gas source and being a low birth weight infant. Kerosene heater users tended to be less educated and own their own homes, but did not have a gas source. Twenty-five percent of participants used fireplaces at least once during the study period, 17% used wood stoves, 3% used gas space heaters, and 18% used kerosene heaters. Use of secondary heating sources was intermittent during the study period (data not shown). Of 219 fireplace users, the median percentage of monitoring periods in which they used the fireplace was 25%. Wood stove users tended to be more consistent, with use reported in 60% of their winter monitoring periods, followed by kerosene heater users, with reported use in 36% of their monitorings. There were only 25 gas space heater users and they were used in about 18% of the monitoring periods. The majority of the children had at least 1 day of cough during the winter heating season of the first year of their lives, and nearly one-third had at least 1 day of wheeze. However, most infants did not cough or wheeze during every monitoring period; while 88% (n = 784) of infants coughed at least once during the winter, these infants coughed in only about 27% of their monitoring periods. Similarly, 33% of the infants had at least 1 day of wheeze during the winter, but had only wheeze during 10% of their monitoring periods. The 890 infants contributed a total of 9,783 infant-reporting periods with symptom information available; 2,678 of the infant-monitoring periods had at least 1 day of cough and 503 had at least 1 day of wheeze. Among the infant-monitoring periods with at least 1 day of the symptom, the median days per 14 days of follow-up was 4.4 for cough (IQR, 2.3–7.8) and 3.7 for wheeze (IQR, 1.9–6.0), whereas the median number of episodes per 14 days at risk was 1.2 for cough (IQR, 0.8–2.0) and 1.1 for wheeze (IQR, 0.8–1.7). Table 2 displays unadjusted associations between use of heating sources, confounding variables, and respiratory symptoms. Wood stove use is the only source associated with any respiratory outcome in the unadjusted analyses. For every 8-hour per day increase in wood stove use, there was an 8% increase in the rate of cough days (e.g., 1% increase for every additional hour of use per day). Families in multifamily dwellings are also at increased risk of respiratory symptoms. Maternal asthma and allergies, as expected, are significantly associated with wheeze and cough days and episodes in the infants. Lower maternal education increases rates of wheeze and cough episodes, and wheeze days, but higher education increases rates of cough days. Boys have increased rates of all outcomes.
Results from the adjusted Poisson models are shown in Table 3 . Several of the source use variables are significant when controlling for important confounders. Gas space heater use, despite the small number of users, is associated with episodes and days of wheeze. Rates of wheeze episodes and days increase about 25% for every 8-hour per day increase in gas space heater use (e.g., 3% increase for every additional hour of use per day). An 8-hour per day increase in wood stove use increases rates of cough days by 10% (e.g., 1.2% increase for every additional hour of use per day), whereas kerosene heater use increases rates of cough episodes 7% for every 8-hour per day increase in use (e.g., about 1% for every additional hour of use per day)
DISCUSSION This study provides further evidence that use of secondary heating sources contributes to respiratory symptoms in infants. Use of wood stoves, gas space heaters, and kerosene heaters is associated with wheeze and cough in this cohort. This study is one of few prospective studies designed specifically to examine the relationships between sources of indoor air particulate matter and respiratory symptoms in infants, a potentially sensitive subgroup of the population. Major study strengths include extensive demographic and household characteristic information, prospective information about respiratory symptoms in the infant, frequent and regular ascertainment of source use information, and the large sample of infants. By limiting the population to nonsmoking households, we eliminated the effect of smoking, which has been found to be strongly associated with respiratory symptoms (25, 26). The study was conducted in the winter months, when subjects are likely to spend more time in close proximity to indoor combustion sources. An important contribution of this research is the ability of the analysis to account for the intermittent source use patterns. Use of secondary heating appliances varies widely over the heating season, even in the same homes, with periods of nonuse followed by periods of continuous use. This study was able to directly relate symptoms and exposures over shorter periods of time (biweekly reporting periods) rather than averaging use and symptoms over an entire heating season. Averaging source use over longer periods of time or using the presence/absence of a source as a measure of exposure, as prior studies have done, dilutes any effects from such an intermittent exposure. Respiratory symptoms were common in this large cohort of infants living in nonsmoking homes. Nearly one-half of the study population had at least one episode of wheeze and nearly all had at least one episode of cough during their first winter (October 15 to April 15). Gold and coworkers (12) found a similar prevalence of respiratory symptoms in their higher-risk cohort (i.e., at least one parent had asthma or allergies). The cohort in the current study is more representative of the asthma prevalence in the general population (i.e., only 9% of mothers had diagnosed asthma). The primary limitations of this study are related to potential selection and recall biases. Women who enrolled in the study tended to be better educated with stable living arrangements compared with those who were ineligible or chose not to participate. Although it is possible that some misclassification may have occurred because of the self-reported nature of some data, frequent follow-up of study participants limited the period of recall and, thus, the potential for misclassification. In addition, questions about infant symptoms were asked before questions about exposure to limit the potential bias from simultaneous reporting of exposure and symptoms. Gas appliances and gas space heaters are a major source of nitrogen dioxide (NO2) in residences, exposure to which has been associated with respiratory effects. Several epidemiologic studies have examined whether the presence of a gas appliance is associated with an increase in respiratory symptoms and respiratory infections and with decreases in lung function in adults and children (27–33). Some studies implicate NO2 in respiratory symptoms or asthma (27, 31, 33); others have failed to find an association (30, 32). Only one other study (34) examined whether unvented gas space heaters specifically were associated with respiratory effects. The cross-sectional study in Australia of 14,124 families with a child between 4 and 5 years of age (34) found an association between use of unvented gas heaters and respiratory symptoms in children. To overcome the bias caused by differences in access to natural gas and wood stove use, the authors restricted the analysis to the 8,154 families living in an area where natural gas was available. However, the study was cross-sectional and did not consider frequency and duration of source use. In addition, the only control for socioeconomic status was an "area measure" based on the proportion of families within a postal code region with low income. In our study, we found that use of gas space heaters was related to both new episodes of wheeze and total wheeze days in infants. We did not ask whether the gas space heaters used in our study were vented or unvented; however, our study confirms these findings in a prospective cohort controlling for several potential individual-level confounders. Residential concentrations of NO2 associated with gas cooking ranges and gas space heaters vary as a function of source characteristics, use patterns, and building characteristics (35). Gas ranges tend to be used sporadically and for short periods of time, resulting in an NO2 concentration gradient with higher levels in the kitchen and lower concentrations throughout other locations in a residence. Gas space heaters, while having NO2 emission rates similar to gas stoves, are typically used for longer periods of time and result in less of a spatial concentration gradient in a home. In our study homes, the median use of gas space heaters was 6.6 hours/day (IQR, 2–24 hours/day). Although demonstrating considerable variability, main living area residential levels of NO2 from gas cooking in winter months are typically in the range of 35 µg/m3, whereas unvented gas heaters are likely to result in residential concentrations four or more times higher and demonstrate less spatial variability within the home (35). Kerosene heater use increased the rate of new episodes of cough in the study infants. A possible explanation for the association with episodes, but not days, of cough may be that kerosene heater emissions trigger a more immediate, acute reaction rather than a longer-lasting reaction. These findings support the results of a study (36) examining the association between kerosene heater use and allergic symptoms in children and adults in an urban area of Ethiopia. The investigators found an increased likelihood of wheeze, rhinitis, and allergic sensitization among users of kerosene heaters compared with users of biomass fuel only, and proposed that the association was more likely explained by an adverse effect of nonbiomass fuels (kerosene) on allergy rather than on infection. We know of only one other published study (37) of the association between kerosene heater use and respiratory symptoms, which found that young children exposed to higher levels of NO2 (30 µg/m3) were at significantly higher risk of lower and upper respiratory symptoms than those who were not exposed. However, the small sample size led to unstable estimates and focused only on NO2 rather than a broad range of important emissions from the kerosene heaters, including particles and SO2. Extensive chamber studies were conducted to evaluate the nature and quantity of particulate emissions from kerosene heaters. These studies indicate that kerosene heaters are an almost exclusive indoor source of sulfate aerosol (SO42-) and acidic aerosol (H+) and an important source of fine particles 2.5 µm or less in diameter (PM2.5) in the indoor environment (22, 38). Numerous studies have found an association between ambient concentrations of PM2.5, SO42-, and H+ and respiratory symptoms, especially in at-risk populations (39). Our finding of an association helps to isolate the effects of SO42- and H+ by focusing on the indoor environment, where other pollutants and confounders could be more easily controlled. Several studies, particularly in developing countries, have found that use of kerosene heaters or stoves is associated with decreased pulmonary function (40–44). In our study, median use of kerosene heaters was 2.0 hours/day (IQR, 0.63–6.00 hours/day). Actual concentrations of air contaminants owed to kerosene heater use are related to heater type, use conditions, sulfur content of the fuel, home characteristics, and weather. Nonetheless, active pollutant monitoring in a subsample of 71 homes that participated in this study (38) suggests that the reported median of 2 hours/day of kerosene heater use added about 3.6 µg of PM2.5 per m3 and 1.7 µg of SO42- per m3 to existing background levels. Residential background levels of PM2.5 and SO42- measured in a subsample of 144 homes in our study without kerosene heaters during the winter months was 17.4 and 1 µg/m3, respectively, suggesting increases over background of 20 and 170%, respectively. Emissions of pollutants into residences from both fireplaces and wood stoves are difficult to characterize because of differences in design, use, and type of wood burned. In this cohort, fireplace use was not significantly related to cough or wheeze, which may be due to the short duration of exposure. Fireplaces tended to be used infrequently and for shorter periods of time (median of 0.61 hour of use per day averaged over the reporting periods), and so might offer less opportunity to contribute particulate matter into the indoor air. Wood stove use was related to days of cough in the study infants. Although it is generally accepted that wood stove emissions can irritate upper airways (45), epidemiologic evidence of an association between wood stove use and respiratory symptoms has been controversial. The association between indoor wood smoke and respiratory symptoms has been examined in more detail, particularly in developing countries, where wood burning for heating and cooking are more common. The effects of biomass burning in open hearths on respiratory symptoms and other outcomes in children and adults have been striking (24, 46–48). In studies of wood-burning stoves in developed countries, results have been more equivocal. Four of six studies involving children in developed countries found that use of wood-burning stoves increased the likelihood of respiratory symptoms or illness (49–52). However, one study (53) found a decreased risk of symptoms among children living in homes in which wood or coal was the primary source of heat compared with those with central heating. Another study (54) found no association between having a woodburning stove as a primary source of heat and respiratory illness in a group of U.S. schoolchildren. Our finding is consistent with those previous studies that found an association between wood stove use and respiratory symptoms in developed countries (49–52), but is inconsistent with findings of no association (54) or of a negative association (53). Differences in findings may be due to inadequate control for confounders, cross-sectional study designs, or differences in measures of exposure and outcomes. In addition, none of the prior studies took into account the intermittent nature of exposure to these sources in the way that we were able to do by frequent questioning of mothers about symptoms. The design of the current study allowed for the most detailed and largest prospective epidemiological evaluation of the association between secondary heating sources—home fireplaces, heating stoves, and kerosene heaters—and respiratory symptoms in infants, a population generally considered at risk. The findings suggest a significant, modest effect of these sources. Acknowledgments The authors thank the 918 families in Virginia and Connecticut who provided extensive information and permitted us to conduct the air measurements inside and outside of their homes. The authors also thank the following hospitals from which our study population was selected: In Connecticut—Danbury Hospital, Manchester Memorial Hospital, Middlesex Memorial Hospital, William W. Backus Hospital, Veterans Memorial Medical Center, and Yale-New Haven Hospital; in Virginia—Community Hospital (Roanoke), Danville Regional Medical Center, Martha Jefferson Hospital (Charlottesville), University of Virginia Health Sciences Center, and Virginia Baptist Hospital (Lynchburg). FOOTNOTES Supported by grant ES05410 from the National Institute of Environmental Health Sciences (NIEHS). This article has an online data supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form February 8, 2002; accepted in final form July 31, 2002 REFERENCES
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1-antitrypsin deficiency (PiZZ). Thorax 1998;53:939–943.
