During their first year, 60% to 80% of children will have an episode of otitis media and 17% of children will have 3 or more episodes.1,2 Otitis media is the illness for which infants and young children are most frequently taken to the doctor.3 It is also the leading reason for the dispensing of antibiotics, which in turn can lead to the development of antibiotic resistance.4 Medical costs associated with otitis media have been estimated at over $5 billion annually.5 Children who first experience acute otitis media at an early age (<6 months) are at increased risk for recurrent otitis media, and could also be at increased risk for hearing loss and delays in speech development.2,6,7
Although infectious in origin, the pathogenesis of otitis media is a complex, multifactorial process. Furthermore, rates of otitis media and recurrent otitis media have been increasing in recent years. A more complete understanding of the factors contributing to this disease could help in the design of prevention strategies involving risk factor modification, a control approach more desirable than wide-scale antibiotic use. The relationship between otitis media and environmental exposures such as tobacco smoke have been investigated,1,8–10 but few studies have examined the relationship between otitis media and other sources of air pollution. The purpose of this study was to assess the effect of environmental exposures from secondary home heating sources on otitis media and recurrent otitis media on infants in the first year of life.
Data used in this report were from a study whose primary purpose was to examine the effect of secondary home heating sources on the respiratory symptoms of infants during their first year of life. A full description of the methods has been provided elsewhere.11 Participants were recruited using a screening questionnaire administered to mothers who were delivering babies at 7 hospitals in Connecticut and 5 hospitals in Virginia between 1993 and 1996. Women with self-reported household exposures to kerosene heaters or gas stoves were identified. A random sample of women who did not report regular kerosene heater or gas stove use were identified using frequency matching at a ratio of 2:1 on race and hospital of delivery, and then were asked to participate. Respondents living in smoking households were excluded as a result of the association between environmental tobacco smoke and respiratory problems in children.12,13 Also excluded were infants admitted to newborn intensive-care units, infants placed for adoption, mothers less than 18 years of age, non-English-speaking mothers, families who anticipated moving out of the study area, and families who could not be reached by telephone or who were participating in the study with a previous child. Of the 17,447 women screened, 1540 were eligible and invited to participate. Of these, 918 (60%) enrolled and 299 (19%) of the remaining women declined to participate, 297 (19%) did not meet study eligibility at the time of the interview, and 26 (<1%) were not enrolled for other reasons. The Yale University Human Investigations Committee as well as institutional review boards at each participating hospital reviewed and approved the study.
Data from 904 of 918 infants who had at least 1 2-week reporting period of follow up were included in the analyses of the impact of secondary home heating sources on episodes of otitis media. We included data from a subset of 813 infants who had a full year of follow up in analyses examining recurrent otitis media.
Trained research assistants administered standardized questionnaires to study participants during home interviews when infants were between 2 and 5 months of age (mean = 3.1 month, standard deviation = 0.7). At that time, the research assistant described the study in detail to the infant’s mother and obtained informed consent. Household demographic data such as maternal race, education, and number of children in the household were collected. Detailed information regarding infant care (eg, breastfeeding and day care use) and health status of the mother (eg, self-reported history of allergies or physician-diagnosed asthma) were obtained. Additional questions included housing characteristics, self-reported presence of mold in the home, number and type of pets, and the use of secondary heating sources (eg, fireplaces, wood stoves, kerosene heaters, and gas space heaters).
Health Outcome and Exposure Assessment
On completion of the initial home interview, mothers were contacted approximately every 2 weeks for 1 year from the time of enrollment (median interval length 17 days; interquartile range = 14–19 days). During these telephone interviews, mothers were asked to report their infants’ respiratory symptoms and doctor or clinic visits (date, reason for visit, and mother’s report of the diagnosis). The mother’s report of a doctor’s diagnosis of ear infection counted as an episode of otitis media if it occurred more than 21 days after a previously reported ear infection, according to the methods of Niemela et al.14 We defined recurrent otitis media as 4 or more episodes of otitis media in 1 year.15
During the telephone interviews, mothers were also asked to report the number of hours each secondary heating source was used in the home during that reporting period. For each source, we classified use as a binary variable (use or nonuse during the reporting period) and as a continuous variable (average number of hours used per day, calculated as total hours of use divided by the number of days in the reporting period). Similar information was collected on use of air conditioners and attendance at day care during a reporting period. Biweekly telephone interviews were also used to ascertain whether a member of the household started smoking or if a smoker had joined the household; participants who responded affirmatively to this question were excluded from further follow up.
Our initial analysis considered any secondary heating source use during the year-long follow-up period and various study subject characteristics, as well as the unadjusted associations between the health outcome and subject characteristics using χ2 analyses on data from 904 study subjects. We conducted a more detailed examination of the relationship between episodes of otitis media and secondary heating sources by creating a set of daily heating source use observations for each subject. We estimated daily exposure using both a binary indicator for each source (ie, any day during a reporting period when the source had been used) and the average daily source use for that reporting period. Because of the uncertainty of the exact dates of daily exposure within reporting periods, an infant’s daily exposure history was estimated in 2 ways: as a binary variable indicating any exposure during the previous 7 days or as a continuous variable indicating the mean hours per day of source use over the previous 7 days. Daily data for the health outcome was entered as 0 for no new episode of otitis media on that date, 1 for a new episode diagnosed that day, or missing for dates when the infant was not at risk for a new episode (ie, the 21 days after an episode). An additional variable was created to indicate whether the observation date was during the heating season (October 15–April 15).
We estimated the association between episodes of otitis media, and exposure to secondary heating sources using generalized estimating equations (GEE, PROC GENMOD in SAS for Windows, 8th ed.; SAS Institute, Cary, NC) with an autoregressive (AR1) structure for repeated observations.16 All models controlled for air conditioning, day care, and other potential confounders (heating season, gas appliances in the home, infant sex, season of birth, race, mother’s education, other children in the household, duration of breastfeeding, mother’s allergies, mother’s asthma, reported mold in the home, pet cat or dog). We evaluated models with both binary and continuous exposure variables and found the results to be virtually identical.
We assessed the association of recurrent otitis media and heating sources using data from the subset of subjects (n = 813) who completed 1 year of follow up. For this analysis, exposure data were collapsed across the year and categorized as any use versus no use. Logistic regression was used to predict recurrent otitis media (yes or no) based on heating source use, controlling for the same potential confounders used in the repeated-measures model.
Overall, 25% of study participants used a fireplace, 18% used a kerosene heater, 17% used a wood stove (Table 1), and 3% used a gas space heater at least once during the study period. Few participants used gas space heaters; thus, we dropped this secondary heating source from further analyses. Similar numbers of boy and girl infants were included in the analysis (52% boys). Participants tended to be white (including fewer than 1% of study participants who classified themselves as Asian) and well educated. One third of children were not breastfed. Maternal allergies were common (44%) and 9% of the mothers reported a history of asthma. Approximately 22% of participants reported the presence of mold in the home and 44% reported owning a cat or dog. One third (34%) reported using gas appliances. Most infants in the study attended day care outside the home (67%). The majority of participants (87%) used an air conditioner at least once during the study period.
The percentage of fireplace users was higher among white or Asian study respondents (Table 1). Participants who used kerosene heaters were more likely to have no more than a high school degree, and those with more education were more likely to report fireplace use. Fireplace and wood stove users were more likely to breastfeed for more than 5 months. They were also more likely to report the presence of mold and to report having a cat or dog in the household. Wood stove users were less likely to report air conditioner use and were more likely to place their infants in day care outside of the home. Families using gas appliances were less likely to use fireplaces, kerosene heaters, or wood stoves.
Figure 1A shows the seasonal distribution of the use of fireplaces, kerosene heaters, wood stoves, air conditioning, and day care. There were a total of 311,830 person-days of observation; 813 subjects were followed for 365 to 373 days and 91 were followed for 11 to 364 days. Fireplaces, kerosene heaters, and wood stoves were used most in the winter months, and not used at all in June, July, and August. The percentage of person-days of use ranged from zero in July to a high in January of 12% for fireplaces, 11% for kerosene heaters, and 14% for wood stoves. In contrast, person-days of air conditioner use were low in winter and reached a high of 84% in July. Day care remained relatively constant, with an average 47% of person-days per month throughout the year.
Otitis media occurred most frequently during the winter months (Fig. 1B). A total of 1256 otitis media episodes were reported by 581 infants during the study period. Of these, 32% occurred during the winter months. The frequency of otitis media was lowest in summer months (17%) compared with 25% in the fall and spring months.
Approximately one third of mothers did not report any otitis media among their infants during the study period (Table 2). Almost 1 in 10 mothers reported that their infant experienced 4 or more otitis media episodes during their first year of life. Fireplace use, kerosene heater use, and wood stove use were only weakly associated with otitis media episodes. Stronger associations with otitis media were found for white or Asian race, maternal asthma, having a cat or dog in the household, day care outside of the home, and no air conditioner use (Table 2).
Table 3 shows the unadjusted associations between heating sources and single episodes of otitis media from repeated-measures logistic regression models. Fireplace use, woodstove use, and attendance in day care were associated with an episode of otitis media. The use of an air conditioner was protective for otitis media episodes in unadjusted models (Table 3). The relationship between otitis media and the winter heating season (October 15–April 15) was evaluated because otitis media episodes varied by season. Winter heating season was associated with more otitis media episodes (odds ratio [OR] =1.54; 95% confidence interval [CI] = 1.37–1.72).
The association between fireplace, kerosene heater, and air conditioning was substantially reduced or absent after adjustment for potential confounders (Table 3). Associations persisted for attendance in day care, winter heating season, white race, birth in the fall, additional children in the household, and a maternal history of allergy. Similarly, the association between secondary home heating source use and recurrent otitis media was negligible after adjustment (Table 4). Attendance in day care, white race, birth in the fall, and a maternal history of asthma and allergy were positively associated with recurrent otitis media in multivariate logistic regression models.
This prospective study examined the impact of exposure to secondary home heating sources on otitis media and recurrent otitis media. Although the use of fireplaces and wood stoves were associated with otitis media in unadjusted analyses, this association was not observed in multivariate models. Acute otitis media is widely considered to be a winter disease,15 in part because of the relationship between otitis media and respiratory viruses that are less common during the summer.17 The addition of heating season as a variable in multivariate models controlled for the season, and it is likely that the observed relationship between fireplace or wood stove use and otitis media occurred because the use of these heating sources was highest in winter months. Similarly, air conditioning appeared protective in unadjusted models because use was highest in summer months, a time of fewer otitis media episodes.
The lack of an association between secondary home heating source use and otitis media does not agree with a case-control study that has implicated wood-burning stoves as a risk factor for otitis media.18 A cross-sectional study of respiratory symptoms among children in Eastern Germany linked decreases in otitis media with decreases in ambient sulfur dioxide and total suspended particles in the years 1995–1996 compared with 1992–1993.19 Wood stove use and kerosene heater use have been linked to total days of cough and episodes of cough, respectively, in this cohort.11 Up to 93% of acute otitis media episodes are accompanied by respiratory symptoms such as cough or nasal congestion.15 The use of these secondary heating sources was intermittent in our study population, and the otitis media outcome was relatively rare compared with the frequency of respiratory symptoms.11 This could explain the lack of an association between exposures from these heating sources with otitis media and recurrent otitis media risk.
Attendance in day care is one of the most important risk factors for otitis media21; our study confirms these findings. The relationship between day care and otitis media is theoretically the result of children encountering more infectious agents at an earlier age. Within the first 6 months of life, the risk of otitis media is higher in large day care centers (5 or more children) and among children who attend more than 30 hours per week.20 A metaanalysis of the risk factors for acute otitis media in 22 studies indicated that child care outside of the home was one of the most important factors influencing acute otitis media.21 Furthermore, attendance in day care has been proposed as an explanation for increasing rates of recurrent otitis media.6
Our study found that, when compared with infants born in the summer, those born in the fall experienced an increased risk for otitis media (OR = 1.27; 95% CI = 1.02– 1.60). The effect was even more pronounced in the case of recurrent otitis media (OR = 2.97; CI = 1.51–5.83). Daly et al. also found that birth during the fall was associated with recurrent otitis media (relative risk [RR] = 2.6; 95% CI = 1.4–4.7).20 Children born in the fall could be susceptible to otitis media because they spend their early months in the winter and spring when exposure to respiratory viruses tends to be high. These children could be especially susceptible to recurrent otitis media risk because children who experience their first otitis media episode at an early age (<6 months) are up to twice as likely to experience recurrent otitis media.22
We found that white race was associated with otitis media and recurrent otitis media (OR = 1.63 and 4.50, respectively). Lanpear et al. also reported that black race was protective for recurrent otitis media.6 In contrast, a large prospective study of otitis media showed the opposite, ie, that black children experienced more otitis media than white children during the first 2 years of life.1
Maternal allergy was associated with otitis media, and both maternal allergy and asthma were positively associated with recurrent otitis media in our study population. A maternal history of asthma appears to predispose their children to asthma.23 Our findings of an association between a maternal history of asthma and recurrent otitis media are consistent with the results of other studies.10,24 Maternal asthma or allergy can serve as a marker for infants with an increased sensitivity to respiratory problems. Although the pathogenic mechanism remains unknown, children with allergies might experience more otitis media because cytokines released by inflammatory cells of the nasopharynx lead to eustachian tube obstruction similar to that observed in an upper respiratory viral infection. Alternatively, allergic responses can result in decreased mucociliary clearance.25–27 Pets and mold were not found to influence otitis media among the infants in our cohort. Dogs or cats in the home were not associated with otitis media in another prospective study of early otitis media.20
One limitation of this study was our inability to distinguish between recurrent acute otitis media and otitis media with effusion.15 Our data were collected by asking parents whether their child had a clinician-diagnosed medical condition. We did not ask parents to distinguish between the various types of otitis media. Because some risk factors for recurrent otitis media and otitis media with effusion can differ, some of our findings could have been diluted by grouping all types of otitis media together. The fact that the diagnosis was not verified by medical record review is another potential limitation of our study. However, recall bias should be low because we ascertained the diagnosis in biweekly interviews. Strengths of this study include the large sample size and prospective study design. This is a well-characterized population with extensive information regarding demographic and household information, and frequent prospective ascertainment of home heating sources. Because care was taken to ensure that all infants were from smoke-free households, we were able to analyze additional environmental exposures in the absence of this important potential confounder. This detailed study of secondary home heating sources with otitis media and recurrent otitis media and secondary home heating sources finds no risk from the intermittent use of these heating sources.
We thank the 917 women and their families in Virginia and Connecticut who participated in this study. We also thank the following hospitals from which the study population was selected: Danbury Hospital, Manchester Memorial Hospital, Middlesex Hospital, William W. Backus Hospital, Veterans Memorial Medical Center, Yale New Haven Hospital, Community Hospital, Danville Regional Medical Center, Martha Jefferson Hospital, University of Virginia Health Sciences Center, and Virginia Baptist Hospital.
1.Paradise JL, Rockette HE, Colborn DK, et al. Otitis media in 2253 Pittsburgh-area infants: prevalence and risk factors during the first two years of life. Pediatrics
2.Teele DW, Klein JO, Rosner B. Epidemiology of otitis media during the first seven years of life in children in greater Boston: a prospective, cohort study. J Infect Dis
3.Schappert SM. Office visits for otitis media: United States, 1975–90. Adv Data
4.McCaig L, Hughes J. Trends in antimicrobial drug prescribing among office-based physicians in the United States. JAMA
5.Klein JO. The burden of otitis media. Vaccine
6.Lanphear BP, Byrd RS, Auinger P, et al. Increasing prevalence of recurrent otitis media among children in the United States. Pediatrics
7.Daly KA, Giebink GS. Clinical epidemiology of otitis media. Pediatr Infect Dis J
8.Etzel RA, Pattishall EN, Haley NJ, et al. Passive smoking and middle ear effusion among children in daycare. Pediatrics
9.Alho OP, Laara E, Oja H. Public health impact of various risk factors for acute otitis media in northern Finland. Am J Epidemiol
10.Lieu JE, Feinstein AR. Effect of gestational and passive smoke exposure on ear infections in children. Arch Pediatr Adolesc Med
11.Triche E, Belanger K, Beckett W, et al. Respiratory symptoms in infants associated with indoor heating sources. Am J Respir Crit Care Med
12.Denson KW. Passive smoking in infants, children and adolescents: the effects of diet and socioeconomic factors. Int Arch Occup Environ Health
13.Burr ML. Indoor air pollution and the respiratory health of children. Pediatric Pulmonol
14.Niemela M, Uhari M, Mottonen M. A pacifier increases the risk of recurrent acute otitis media in children in day care centers. Pediatrics
15.Faden H, Duffy L, Boeve M. Otitis media: back to basics. Pediatr Infect Dis J
16.Allison P. Logistic Regression Using the SAS System: Theory and Application
. Cary, NC: SAS Institute Inc; 1999.
17.Vesa S, Kleemola M, Blomqvist S, et al. Epidemiology of documented viral respiratory infections and acute otitis media in a cohort of children followed from two to twenty-four months of age. Pediatr Infect Dis J
18.Daigler GE, Markello SJ, Cummings KM. The effect of indoor air pollutants on otitis media and asthma in children. Laryngoscope
19.Heinrich J, Hoelscher B, Wichmann HE. Decline of ambient air pollution and respiratory symptoms in children. Am J Respir Crit Care Med
20.Daly KA, Brown JE, Lindgren BR, et al. Epidemiology of otitis media onset by six months of age. Pediatrics
21.Uhari M, Mantysaari K, Niemela M. A meta-analytic review of the risk factors for acute otitis media. Clin Infect Dis
22.Teele DW, Klein JO, Rosner B. Epidemiology of otitis media in children. Ann Otol Rhinol Laryngol
23.Litonjua AA, Carey VJ, Burge HA, et al. Does parental history and the risk for childhood asthma: does mother confer more risk than father? Am J Respir Crit Care Med
24.Nafstad P, Magnus P, Jaakkola JJ. Early respiratory infections and childhood asthma. Pediatrics
25.Kvaerner KJ, Tambs K, Harris JR, et al. Otitis media: relationship to tonsillitis, sinusitis, and atopic diseases. Int J Pediatr Otorhinolaryngol
26.Lack G. Pediatric allergic rhinitis and comorbid disorders. J Allergy Clin Immunol
27.Fireman P. Otitis media and eustachian tube dysfunction: connection to allergic rhinitis. J Allergy Clin Immunol