Domestic airborne pollutants and asthma and respiratory symptoms in middle age
- (Associate Editor: Chi Chiu Leung).
Background and objective
The role of indoor air pollution as a risk factor for asthma and respiratory symptoms in middle age is unclear. We investigated associations between indoor air pollution sources and (i) asthma phenotypes and (ii) asthma-related respiratory symptoms in middle-aged adults.
Subjects (n = 5729) who participated in the 2004 survey of the Tasmanian Longitudinal Health Study completed respiratory and home environment questionnaires. Associations between indoor air pollution sources, and asthma phenotypes and asthma-related respiratory symptoms were estimated.
Recent mould in the home was associated with current asthma (odds ratio (OR) 1.26; 95% confidence interval 1.06–1.50), wheeze (OR 1.34; 1.17–1.54) and nocturnal chest tightness (OR 1.30; 1.12–1.51). Stratified by atopy and gender, recent mould was associated with current non-atopic asthma only in males (OR 3.73; 1.29–10.80). More rooms affected by mould were associated with significant trends for current asthma, wheeze and nocturnal chest tightness. Home environmental tobacco smoke was associated with doctor-diagnosed asthma (OR 1.25; 1.02–1.53), wheeze (OR 1.69; 1.41–2.03), nocturnal chest tightness (OR 1.54; 1.26–1.88), with current asthma only in non-smokers (OR 2.09; 95%: 1.30–3.35) and with current asthma only in males (OR 1.74; 95%: 1.25–2.42). Among heating appliances, reverse cycle air conditioning was negatively associated with doctor-diagnosed asthma (OR 0.84; 0.70–1.00). Neither electric nor gas stove use was associated with either asthma phenotype or with asthma-related respiratory symptoms.
In middle age, reducing home exposure to mould and environmental tobacco smoke might reduce asthma and asthma-related respiratory symptoms.
environmental tobacco smoke
Tasmanian Longitudinal Health Study.
It is now well established that asthma has increased substantially in recent decades especially in the developed world, and environmental factors have been suggested as important contributors to this increase.
Among domestic factors studied, visible mould, high fungal spore levels and dampness in the home have been associated with asthma morbidity,[1-3] and we reported previously that visible mould indoors was associated with increased peak flow variability in current asthmatics sensitized to fungi. Similarly, exposure to environmental tobacco smoke (ETS) in the home is a well-documented risk factor for asthma and related symptoms.[5-10] Relationships between cooking devices and combustion by-products, and asthma or wheeze have been reported,[11-14] and combustion by-products including nitrogen dioxide and particulate matter provide a plausible mechanism for this association.[15, 16]
However, studies of domestic airborne pollutants in atopic and non-atopic asthma phenotypes are sparse. Also, risk factors for atopic and non-atopic asthma in various populations have been shown to differ.[17, 18] Our recent work examining early-life risk factors for adult atopic and non-atopic asthma uncovered important associations prompting the research questions examined in the current manuscript.
We hypothesized that home exposure to mould, indoor ETS and domestic heating and cooking appliances was associated with asthma and asthma-related respiratory symptoms and that these associations differed according to atopic asthma phenotypes. Using the Tasmanian Longitudinal Health Study (TAHS) data, we investigated the effect of indoor air pollutants on atopic and non-atopic asthma and asthma-related respiratory symptoms in middle age.
The TAHS commenced in 1968 when parents of 8583 Tasmanian school-children born in 1961 (98.9% of those eligible) completed a questionnaire on their children's respiratory health. The methodology and results from previous surveys have been reported elsewhere. The most recent survey started in 2004 when 7312 participants (85.2%) from 1968 were traced and sent a detailed respiratory questionnaire. Of these, 5729 (78.4%) returned a completed questionnaire. A subset (n = 1383), enriched for a history of asthma and chronic bronchitis, attended a laboratory study where lung function tests were performed and skin-prick tests obtained to a panel of common aero-allergens.
The 2004 survey questionnaire included items on home heating and cooking appliances. These validated questions were used to assess exposure to combustion by-products.
Mould exposure was ascertained from a self-report of ever having visible mould on any household surface except food, particularly in the previous 12 months. Participants also identified affected rooms.
Smoking was classed as current, former and never smoking. Exposure to ETS was ascertained by a participant reporting at least one other person in the household smoking regularly.
Asthma ever was ascertained by a participant's affirmative response to the question ‘Have you ever had asthma or wheezy breathing?’ Those with ‘asthma ever’ identified whether symptoms were current (within the last 12 months). Participants who reported ‘asthma ever’ but no current symptoms were classified as non-current asthma. The 2004 survey contained questions on hay fever, doctor-diagnosed asthma, and wheeze and nocturnal chest tightness in the last 12 months.
Atopy was identified by a positive skin-prick test (wheal diameter ≥3 mm than the negative control) for any allergen. Socioeconomic status was defined by the participant's occupation coded according to the Australian Standard Classification of Occupations.
Maternal and paternal allergic disease was identified by the participant's parents reporting asthma and/or allergic rhinitis in the 1968 enrolment survey.
Associations between exposure to combustion appliances or mould, and asthma and asthma-related respiratory symptoms were evaluated using χ tests. Multivariable logistic regression models examined the independent effects of these exposures on asthma and asthma-related respiratory symptoms. Potential confounders examined included place of residence, gender, hay fever, maternal and paternal allergic disease, smoking, and socioeconomic status. Interactions between each exposure and gender, and personal smoking were examined and retained in the model if statistically significant. Model fit was tested by the Hosmer and Lemeshow goodness-of-fit test. Multinomial logistic regression examined mould exposure in the previous 12 months, and current or non-current, atopic or non-atopic asthma. Two-sided P-values <0.05 were considered statistically significant. Analyses were performed using Stata Statistical Software: Release 11.1 (Stata Corporation, College Station, TX, USA).
The study was approved by the Human Research Ethics Committees at the Universities of Melbourne and Tasmania (Melbourne University Human Research Ethics Committee approval number 010626X). Written informed consent was obtained from all participants.
Participant characteristics are shown in Table 1. Just over half were male, nearly 30% were current smokers, and current asthma prevalence was 11.6%. Wheeze in the last 12 months (23.3%) and nocturnal chest tightness (16.8%) were relatively common. At least one regular smoker other than the participant was present in 14.7% of households.
Table 1. Participant characteristics (n = 5729)
|Current asthma||663 (11.6)|
|Wheeze in last 12 monthsa||1335 (23.3)|
|Nocturnal chest tightness in last 12 monthsb||961 (16.8)|
|Doctor-diagnosed asthma||1150 (20.1)|
|Male gender||2953 (51.5)|
|Smoking 2004c|| |
|Exposure to indoor tobacco smoked|| |
|Educational qualificationse|| |
|Year 9 or less||440 (7.7)|
|Years 10–12 (or equivalent)||2472 (43.2)|
|University degree||993 (17.3)|
Prevalence of household heating and cooking appliances and mould exposure
The commonest types of heating were wood or coal fire (43.2%), or electric heating (38.2%); (Table 2). Reverse-cycle air conditioning accounted for nearly 20% of heating appliances, while gas room heaters and/or ducted gas central heating were used by approximately 17% of participants. Over one third (33.7%) of respondents used more than one heating appliance. Electric (76.7%) and gas stoves (22.1%) accounted for almost all cooking facilities.
Table 2. Prevalence of heating and cooking appliances and mould exposure
|Heating appliancea|| |
|Ducted gas central heating||339 (5.9)|
|Wood or coal fire||2477 (43.2)|
|Gas room heater||632 (11.0)|
|Electric heater||2188 (38.2)|
|Other central heating||363 (6.3)|
|Reverse cycle air conditioning||1141 (19.9)|
|Cooking applianceb|| |
|Coal, coke or wood||38 (0.7)|
|>One appliance||13 (0.2)|
|Mould ‘ever’ in the house||2794 (48.8)|
|Mould in the last 12 months||1985 (34.7)|
Just under half the participants reported ever having mould on any home surface, while over a third reported mould within the last 12 months.
Domestic heating and cooking appliances, and asthma or respiratory symptoms
Table 3 shows the relationship between heating and cooking appliances, and asthma or asthma-related symptoms. Reverse-cycle air conditioning use was associated with less wheeze in the last 12 months (P = 0.02) and less doctor-diagnosed asthma (P = 0.05). Central heating other than ducted gas central heating was associated with less nocturnal chest tightness (P = 0.04). None of the cooking appliances was associated with current or doctor-diagnosed asthma or with either asthma-related symptom.
Table 3. Household heating and cooking appliances, asthma and asthma-related symptoms
|Ducted gas central heating|| || || || || |
|Yes||339||45 (13.3)||76 (22.4)||64 (18.9)||82 (24.2)***|
|No||5390||618 (11.5)||1259 (23.4)||897 (16.6)||1068 (19.8)|
|Wood or coal fire|| || || || || |
|Yes||2477||289 (11.7)||581 (23.5)||415 (16.8)||483 (19.5)|
|No||3252||374 (11.5)||754 (23.2)||546 (16.8)||667 (20.5)|
|Gas room heater|| || || || || |
|Yes||632||84 (13.3)||165 (26.1)||114 (18.0)||132 (20.9)|
|No||5097||579 (11.4)||1170 (23.0)||847 (16.7)||1018 (20.0)|
|Electric heater|| || || || || |
|Yes||2188||254 (11.6)||520 (23.8)||379 (17.4)||456 (20.8)|
|No||3541||409 (11.6)||815 (23.1)||582 (16.5)||694 (19.6)|
|Other central heating|| || || || || |
|Yes||363||49 (13.5)||70 (19.3)||47 (13.0)**||76 (21.0)|
|No||5366||614 (11.4)||1265 (23.6)||914 (17.1)||1074 (20.0)|
|Reverse cycle air conditioning|| || || || || |
|Yes||1141||115 (10.1)||236 (20.7)*||178 (15.6)||205 (18.0)***|
|No||4588||548 (11.9)||1099 (24.0)||783 (17.1)||945 (20.6)|
|Gas stove|| || || || || |
|Yes||1276||151 (11.8)||285 (22.3)||207 (16.2)||270 (21.2)|
|No||4439||508 (11.4)||1048 (23.6)||752 (16.9)||877 (19.8)|
|Electric stove|| || || || || |
|Yes||4405||509 (11.6)||1039 (23.6)||750 (17.0)||873 (19.8)|
|No||1310||153 (11.7)||294 (22.4)||209 (16.0)||274 (20.9)|
|Coal, coke or wood stove|| || || || || |
|Yes||40||4 (10.0)||11 (27.5)||4 (10.0)||7 (17.5)|
|No||5675||658 (11.6)||1322 (23.3)||955 (16.8)||1140 (20.1)|
In a logistic model adjusted for gender, smoking, ETS, maternal atopy, socioeconomic status and hay fever, there was a modest association between reverse-cycle air conditioning and less doctor-diagnosed asthma (odds ratio (OR) 0.84; 0.70–1.00, P = 0.05). In a similar model, there was no association between reverse-cycle air conditioning and wheeze in the last 12 months (P = 0.25). None of the other heating appliances was associated with either asthma phenotype or related respiratory symptom (results not shown). Neither electric nor gas stove use was associated with either current or doctor-diagnosed asthma or respiratory symptoms (results not shown).
Indoor mould, exposure to ETS, and asthma or asthma-related respiratory symptoms
Mould ‘ever’ in the house and mould in the last 12 months were associated with current and doctor-diagnosed asthma, and with wheeze and nocturnal chest tightness (Table 4).
Table 4. Visible mould and associated symptoms/diseases
|Mould ever in housea|| || || || || |
|Yes||2794 (48.9)||363 (13.0)*||737 (26.4)**||521 (18.7)**||604 (21.7)***|
|No||2921 (51.1)||298 (10.2)||595 (20.4)||438 (15.0)||543 (18.7)|
|Mould in last 12 monthsb|| || || || || |
|Yes||1985 (34.8)||272 (13.7)**||534 (26.9)**||386 (19.5)**||428 (21.6)****|
|No||3712 (65.2)||386 (10.4)||795 (21.4)||571 (15.4)||714 (19.2)|
In a model mutually adjusted for exposure to indoor mould in the last 12 months and exposure to ETS and also adjusted for gender, smoking, socioeconomic status, maternal atopy and hay fever, mould in the last 12 months was associated with current asthma and related respiratory symptoms but not with doctor-diagnosed asthma. There was no effect modification of these associations by smoking (results not shown). As the number of mould-affected rooms increased, the odds of current asthma and related respiratory symptoms increased (Table 5).
Table 5. Visible indoor mould, home exposure to environmental tobacco smoke, asthma and asthma-related respiratory symptoms
|Number of rooms in the house affected by mould ‘at any time’a|| || || || |
|0 rooms affected||1.0||1.0||1.0||1.0|
|1 room affected||1.16 (0.96–1.42)||1.37 (1.18–1.59)||1.19 (1.01–1.41)||1.09 (0.93–1.28)|
|2 rooms affected||1.28 (0.98–1.67)||1.36 (1.10–1.69)||1.32 (1.04–1.66)||1.12 (0.90–1.39)|
|3 rooms affected||1.19 (0.83–1.69)||1.53 (1.17–2.01)||1.52 (1.14–2.03)||1.17 (0.87–1.55)|
|4 rooms affected||1.76 (1.08–2.87)||1.96 (1.31–2.94)||1.73 (1.12–2.67)||0.97 (0.61–1.54)|
|Test for trend||P = 0.010||P < 0.001||P < 0.001||P = 0.278|
|Mould in the house in the last 12 monthsb||1.26 (1.06–1.50)||1.34 (1.17–1.54)||1.30 (1.12–1.51)||1.05 (0.91–1.21)|
|Exposure to indoor ETS at home (no interaction terms)b||ND||1.69 (1.41–2.03)||1.54 (1.26–1.88)||1.25 (1.02–1.53)|
|Exposure to indoor ETS at homec|| || || || |
|Among non-smokers||2.09 (1.30–3.35)||NA||NA||NA|
|Among current smokers||1.22 (0.90–1.66)||NA||NA||NA|
| ||Pinteraction = 0.06|| || || |
|Among males||1.74 (1.25–2.42)||NA||NA||NA|
|Among females||1.10 (0.80–1.53)||NA||NA||NA|
| ||Pinteraction = 0.045|| || || |
From the same model, exposure to ETS was associated with doctor-diagnosed asthma, wheeze and nocturnal chest tightness (Table 5). There was evidence (Pinteraction = 0.045) that this association was modified by gender, being stronger in males. There was modest evidence that the association between exposure to ETS and current asthma was modified by smoking, being stronger in non-smokers than current smokers (Pinteraction = 0.06). Neither smoking nor gender modified the effect of exposure to ETS on either asthma-associated respiratory symptom or doctor-diagnosed asthma.
Mould and atopic versus non-atopic asthma
There was no difference in the prevalence of mould sensitization in those exposed to mould in the last 12 months (12.9%) compared with those not so exposed (12.6%).
A multinomial logistic regression model examined the relationship between mould exposure, and atopic and non-atopic asthma. There was evidence (Pinteraction = 0.034) that mould exposure in the last 12 months was associated with current non-atopic asthma in males (OR 3.73; 1.29–10.8) but not females (OR 0.93; 0.46–1.89) when compared with participants with neither asthma nor atopy. No association was found between mould exposure and any other combination of asthma and atopy (Table 6).
Table 6. Multinomial regression model showing the relationship between ‘visible mould in the last 12 months’ and asthma phenotypes
|Never asthma, non-atopic (ref)||308||117 (38.0)||1.0||ND|
|Never asthma, atopic||238||86 (36.1)||0.87 (0.60–1.26)||0.458|
|Non-current asthma, non-atopic||246||103 (41.9)||1.18 (0.83–1.71)||0.346|
|Non-current asthma, atopic||329||123 (37.4)||0.84 (0.59–1.20)||0.330|
|Current asthma, non-atopicc|| || || || |
|In males||19||12 (63.2)||3.73 (1.29–10.80)||(Pinteraction = 0.034)|
|In females||45||20 (44.4)||0.93 (0.46–1.89)|
|Current asthma, atopic||198||80 (40.4)||0.94 (0.63–1.41)||0.764|
Our study is novel in that it examined the effect of indoor air pollutants on asthma phenotypes and respiratory symptoms in a cohort of middle-aged adults, contrasting with similar studies that enrolled only children[25-27] or young adults,[28, 29] and did not differentiate asthma phenotypes.
We found that recent mould exposure in the home was associated with a near fourfold increase in the odds of current non-atopic asthma but no other phenotype. Recent mould exposure was also associated with asthma-related respiratory symptoms. A dose-response effect was suggested where the greater the number of rooms affected by mould, the greater the odds of current asthma and related respiratory symptoms.
We confirmed previous findings that ETS in the home was associated with doctor-diagnosed asthma as well as asthma-related respiratory symptoms.[13, 30] In addition, we found that the association between ETS in the home and current asthma was modified by personal smoking and gender.
Our finding of a strong association between mould exposure and current non-atopic asthma is novel and underlines the importance of exploring risk factors for asthma according to asthma phenotypes. Identification of atopic and non-atopic asthma phenotypes was a key step in our earlier work, while our current finding lends support to the concept of a non-atopic mechanism being important in adult asthma.
Over one-third of our cohort reported mould on any surface in the last 12 months, somewhat similar to the European Community Respiratory Health Survey that reported a 22% overall prevalence of mould or mildew in the house in the last 12 months. In the European Community Respiratory Health Survey, mould exposure in the previous year was associated with both current asthma (OR 1.28; 1.13–1.46) and wheeze and breathlessness (OR 1.34; 1.18–1.51), findings remarkably similar to ours. A meta-analysis of studies examining mould in homes and respiratory outcomes found an association with current asthma in adults and children combined (OR 1.56; 1.30–1.86) and with wheeze in adults (OR 1.39; 1.04–1.85). Similar findings in adults on associations between indoor mould, and asthma or asthma-related symptoms have been published with OR ranging from 1.5 (1.0–2.4) to 2.9 (1.6–5.3).[30, 33-35] However, none of these studies reported a dose-response effect between mould exposure and asthma.
We found that the number of rooms affected by mould was associated with an increasing trend for current asthma and respiratory symptoms. Having more than one room affected by mould is likely to increase the airborne spore concentration, creating a more allergenic environment. We previously reported that higher exposure to airborne fungal spores was associated with current asthma and increased bronchial hyperresponsiveness in the Australian arm of the European Community Respiratory Health Survey 2.
The possibility of reverse causation explaining the relationship between mould exposure and current asthma cannot be ruled out if those with current asthma were more likely to report mould. However, this association being restricted to non-atopic asthmatics argues against this possibility. Our finding that mould had no independent association with doctor-diagnosed asthma must cause the association between mould exposure and current non-atopic asthma to be regarded with caution.
The mechanism by which mould exposure could induce asthma or asthma-related respiratory symptoms is undetermined. Mould has long been recognized as an allergen and shown to induce immunoglobulin E production and airway hyperresponsiveness in mice. However, our finding of an association between mould exposure and current non-atopic asthma supports a non-allergic mechanism. Volatile organic compounds, ergosterol, glucans and mycotoxins from mould could be relevant to a non-atopic mechanism. These data were not collected in our study preventing exploration of such factors. Our finding that the association between indoor mould exposure and current non-atopic asthma was stronger in males than females is not easily explained and might be an incidental finding related to multiple comparisons in the analysis.
The association between ETS and respiratory disease, particularly doctor-diagnosed asthma, has been well characterized.[8, 39-41] It might be expected that the association between ETS and asthma would be stronger in smokers, but our findings were opposite. Nonetheless, our findings were supported by Eisner et al. who reported that higher measured ETS in a cohort of non-smoking asthmatics was associated with worse asthma severity and greater risk of hospitalization for asthma.
In a middle-aged population, asthma is frequently more common in females and in those exposed to ETS. Our finding of an association between ETS in the home and current asthma in males is both in keeping with previous findings and somewhat unexpected with males being more affected than females.
Cooking and heating appliances
There was no association between domestic cooking appliances, and current or doctor-diagnosed asthma. Notably, we found no association between exposure to a solid-fuel cooking appliance, and asthma or related symptoms. It is possible that our finding was due to lack of measured concentrations of indoor combustion products.
Exposure to reverse cycle air conditioning was weakly associated with reduced odds of doctor-diagnosed asthma, a finding contrary to European Community Respiratory Health Survey but consistent with other studies[43-45] that found little or no impact of this appliance on respiratory symptoms. The protective effect of reverse cycle air conditioning is likely to be due to a better controlled indoor environment with less damp and conditions that favour mould growth. However, we cannot exclude the possibility that our finding might have been confounded by another unmeasured environmental factor.
A limitation to our study was the use of a questionnaire to assess exposures to indoor combustion by-products and mould. This arose from the need to estimate exposures for large numbers of subjects where individual measurements were not feasible. Nonetheless, we have found self-reported exposure to indoor pollutants was a reliable and valid method of assessing the indoor home environment. In addition, we have found that questionnaire data on surface mould in the home correlated strongly with measured airborne fungal spores and total ergosterol in floor dust, suggesting that self-reported visible mould in our current study would correlate well with measured data. Our study would be strengthened by mould exposure data from earlier TAHS surveys, allowing examination of an association between mould exposure and incident asthma. However, such data were not collected, and we were unable to explore such longitudinal associations.
The ascertainment of asthma and asthma-related respiratory symptoms was by questionnaire and not objective test. However, the questions used were almost identical to validated questions on asthma and asthma-related symptoms in the European Community Respiratory Health Survey 1 questionnaire and in the original TAHS survey. We believe that our survey questions accurately identified those with asthma and asthma-related symptoms.
In conclusion, exposure to mould and ETS in the home were shown to be risk factors for asthma and asthma-related respiratory symptoms in our middle-aged participants. We acknowledge that unmeasured confounders, including other pollutants, could have influenced our findings. Studies that directly measure indoor concentrations of combustion by-products and mould spores would be needed to investigate further the association between specific indoor pollutants and respiratory disease in older populations.
The authors gratefully acknowledge the contributions of Mrs Cathryn May, BApplSci, Victorian Cervical Cytology Registry, East Melbourne, Victoria, Australia, who was largely responsible for tracing participants for the follow up survey in 2004, and Dr Buddhini Ekanayake, MB.BS, Doutta Galla Community Health, Kensington, Victoria, Australia, who undertook some preliminary data analysis for this study. We also acknowledge and thank all previous investigators in the TAHS and gratefully acknowledge the continued support and participation of the original 1968 study cohort. The TAHS is supported by grants from the National Health and Medical Research Council of Australia, the Victorian, Queensland and Tasmanian Asthma Foundations, the Clifford Craig Medical Research Trust, the Royal Hobart Hospital Research Foundation, and the University of Melbourne. Dr Melanie Matheson, Dr John Hopper and Dr Shyamali Dharmage are supported by the National Health and Medical Research Council of Australia.