Prevalence and residential determinants of fungi within homes in Melbourne, Australia

Transcription

1 Clinical and Experimental Allergy, 1999, Volume 29, pages Prevalence and residential determinants of fungi within homes in Melbourne, Australia S. DHARMAGE*, M. BAILEY*, J. RAVEN, T. MITAKAKIS, F. THIEN, A. FORBES*, D. GUEST, M. ABRAMSON* and E. H. WALTERS Departments of *Epidemiology and Preventive Medicine, Respiratory Medicine, Allergy and Clinical Immunology, Monash Medical School and The Alfred Hospital and School of Botany, The University of Melbourne, Melbourne, Australia Summary Background Recent epidemiological studies suggest that the adverse respiratory health effects caused by the inhalation of fungal propagules are substantial. Knowledge of the prevalence and environmental determinants of indoor fungal levels is essential in designing effective avoidance measures. Aim To investigate the prevalence of fungi and the influence of residential characteristics on levels of fungi within homes in Melbourne, Australia. Methods Floor dust and air samples were collected from bedrooms in 485 houses over 1 year. The dust was analysed for ergosterol, a marker of cumulative fungal biomass exposure. Total and genera-specific fungal propagules were identified in air samples. Details of the relevant residential characteristics were documented using a questionnaire. Independent predictors (P < 0.05) of ergosterol and total fungal propagules were identified by multiple linear regression. Results Fifty-five percent of the houses had viable fungal propagules exceeding 500 CFU/m 3. Cladosporium and Penicillium were identified as the most prevalent and abundant fungal genera in indoor air. The median ergosterol level in bedroom floor was 3.8 mg/g of dust. Multivariate analysis showed that total fungal propagules in indoor air were lower in bedrooms with a ceiling fan, without visible mould, and those that were more frequently vacuumed, had a solid fuel fire, had windows closed at the time of the sampling or lacked pets. The presence of more than one cat had the greatest effect on total fungal propagules. Ergosterol levels were significantly lower in homes without old fitted carpets, visible mould or pets and those with frequent airing and regular use of an extractor fan in the kitchen. Old wall-to-wall carpets had the greatest effect on ergosterol. Conclusions High indoor fungal exposures were associated with infrequent ventilation or vacuuming, presence of pets, visible mould and old carpets. Keywords: ergosterol, fungi, house dust mites, indoor allergens, indoor environment Clinical and Experimental Allergy, Vol. 29, pp Submitted 25 September 1998; revised 10 November 1998; accepted 8 March Introduction Recent epidemiological studies suggest that adverse respiratory health effects caused by the inhalation of fungal propagules are substantial [1]. Observed home dampness Correspondence: S. Dharmage, Department of Epidemiology and Preventive Medicine, Monash Medical School, Alfred Hospital, Prahran, Vic. 3181, Australia Blackwell Science Ltd which correlates with higher fungal spore concentrations [2], and visible mould is associated with allergies and asthma in adults [3,4] and children [5,6]. Verhoeff et al. observed indoor dampness and visible mould to be associated with reported respiratory symptoms via sensitization to house dust mites and fungi [6]. In addition to their inherent allergenicity, metabolites of fungi have been suggested to adversely influence the respiratory system through 1481

2 1482 S. Dharmage et al. their immunosuppressant or irritant properties [1]. These findings highlight the need to critically evaluate the approaches to reduce indoor fungal exposure. Residential characteristics have an important influence on indoor fungal levels, although the occurrence of fungal propagules indoors largely reflects the outdoor air [7,8]. The prevalence of indoor fungi has been previously investigated, but only limited inconclusive information is available regarding the predictors of indoor fungal levels. Characteristics that have been investigated include damp housing [9 11], home construction [7,11], ventilation [7,11 14], household pets [10,13] and fitted carpets [7,10]. These factors are interrelated, as well as being related to other residential characteristics which may influence the levels of fungi. Many previous studies have only examined one or at the most a few relevant factors together, hence the results may have been confounded by other unmeasured factors. An intervention study to assess the effectiveness of frequent ventilation and cleaning in reducing indoor fungal levels has shown promising results, but firm conclusions cannot be drawn as the comparison was before and after the intervention without controls [14]. Furthermore the relationship between housing conditions and indoor fungal levels is likely to be affected by the geographical location under study. An understanding of the environmental determinants of the indoor fungal levels would clarify the changes to interior design necessary to reduce indoor exposure to fungi. Thus, we carried out a cross-sectional study to assess the prevalence of fungi and the influence of residential characteristics upon fungal levels in Melbourne homes. This study was nested within a longitudinal study in which a sample of young adults and their homes were investigated to assess the indoor environmental risk factors for asthma. Materials and methods Subjects The subjects were participants in a follow-up study to the European Community Respiratory Health Survey (ECRHS), as undertaken by our centre in Melbourne. The two-phase sampling procedure that was used has been described elsewhere [15,16]. Briefly, a questionnaire was mailed to a group of young adults randomly selected from the electoral roll during the first phase of the ECRHS in This comprised 4500 subjects aged between 20 and 44 years, of whom 3200 (72%) responded. The prevalence of symptoms of asthma among the respondents (i.e. having nocturnal dyspnoea, asthma within last 12 months or having been on medication for asthma) was 17.6%. A random sample (n ¼ 1642) and a symptomatic sample (n ¼ 433) of the respondents were invited to the laboratory for testing in the second phase. A total of 757 subjects attended between 1993 and This group was asked to participate in a follow-up study in 1996 and 485 (23%) of the subjects who were originally invited to the laboratory in 1993 complied. Among them 349 were from the random sample and 136 were from the symptomatic sample. In the follow-up, participants were visited at home once in a random order over a period of 1 year. The respective numbers visited during summer, autumn, winter and spring were 28, 146, 136 and 175. Dust and air samples were collected from their bedrooms, and a residential questionnaire was administered. The participants were requested to visit the laboratory to complete a detailed respiratory questionnaire, and to undergo skin prick and lung function tests, which will be reported separately. The current paper presents a cross-sectional analysis of the prevalence of fungi and the residential characteristics associated with fungal levels in the bedrooms of the participants. Questionnaire An interviewer-administered questionnaire was used to collect information about residential characteristics such as the age of the home, house construction, type of ventilation, heating, evidence of dampness and mould, floor covering, bedding and cleaning routines. Bedroom air temperature and relative humidity (i.e. absolute humidity of the sample as a percentage of fully saturated gas at the same temperature [17]) were recorded at the same visit using a digital thermohygrometer (Type THG-388 RS Components, China). The thermohygrometer was placed in the bedroom, 1 m above floor level and away from windows, beds and heat sources and read after equilibrating for 15 min. Absolute humidity (the amount of water vapour per unit of gas expressed as partial pressure for each house [17]) was calculated using the measurements of temperature and relative humidity. Collection of dust samples Samples of house dust were collected from the floor of the bedroom of the study subjects. These were collected using a modified Makita 4071D handy-vac (7.2 volt battery operated; Anjou, Aichi, Japan) with a 25-mm nylon filter and according to a standard protocol [18] with some modification. The modification was identified from a pilot study in which some of the floor dust samples collected from 1 m 2 were inadequate for the allergen assays. We found that at least 30 mg of total dust needed to be collected as each assay required around 10 mg of fine dust. Hence the bedroom floor was sampled according to two protocols. The standard protocol was to vacuum for a total of 2 min over an area of 2m 2 adjacent to the side of the bed on which the participant

3 Fungi in Australia 1483 slept. The extended protocol was followed if the dust collected according to the standard protocol seemed inadequate. In the extended protocol vacuuming was extended by 1m 2 per minute until an adequate sample was collected. Only 447 dust samples were sufficient to analyse for ergosterol as the dust samples were first analysed for house dust mite allergen the results of which have already been presented [19]. Measurement of cumulative fungal levels in floor dust The fungal membrane lipid ergosterol was used as an indicator of total fungal biomass in dust samples, using a modification of the technique of Martin et al. [20]. Dust samples (approximately 50 mg) were weighed, suspended in 0.5-mL cold ethanol and sonicated for 10 min in a cold water bath. After centrifugation the supernatant was decanted, and the pellet re-extracted with a further 0.5 ml of cold ethanol. The supernatants were combined, filtered (0.22 mm), and 100 ml of the ethanol extract was analysed by reverse-phase HPLC, using a 100GLC4-ODS2 8/5 C-18 cartridge (SGE Scientific Pty Ltd, Ringwood, Australia). The sample was eluted at 1.0 ml/minute with 100% methanol, and ergosterol was monitored by UV absorbance at 282 nm. The assay was calibrated using a pure sample of ergosterol (Sigma Chemical Co, St Louis, MO, USA; catalogue number E6510). Results are expressed as micrograms ergosterol per gram of fine dust and micrograms per square meter of floor area. Measurement of total and specific fungal levels in the air The identity and abundance of viable fungal propagules suspended in the air at the time of sampling was estimated using a two-stage Andersen sampler (Model , Graseby/Andersen, Atlanta, GA, USA) which divides particles into respirable and non-respirable fractions and deposits particles onto separate agar plates. Potato dextrose agar plates were exposed for 1 min in each subject s bedroom at an air-sampling rate of 28.3 L/min. Exposed plates with respirable fungal propagules were incubated in darkness at 25 C for 4 days. The resulting fungal colonies of Cladosporium, Alternaria, Epicoccum, Penicillium and Aspergillus were identified on the basis of colony and spore morphology and all the other colonies were totalled in one group as others. Counts were expressed as colony forming units per cubic metre of air. The five genera identified in this study have previously been reported as allergenic [21 25]. Statistical analysis The Statistical Analytical System package [26] and STATA [27] were used to analyse the data. The distributions of spore counts from the air sample and ergosterol levels were both positively skewed. Log transformation-normalized distributions of total fungal propagules, Cladosporium, other fungal propagules and ergosterol levels expressed per gram of fine dust. The two-phase sampling process over-sampled symptomatic subjects, and therefore a re-weighting (or poststratification) procedure [28] was applied to realign the sample to the estimated symptomatic prevalence of 17.6% in the population aged years to assess the prevalence of fungi. An additional consideration was that this prevalence estimate itself involved a degree of sampling error, which needed to be accounted for. Applying these corrections using the Stata software [27], yielded negligible differences to unweighted analyses, principally because the symptomatic and asymptomatic subjects in the sample were comparable across all studied factors, and the initial random sample was large enough that additional sampling variability from the symptomatic prevalence estimate was minimal. Therefore for simplicity, only the unweighted results are presented in the text and tables. Temperature and relative humidity in the bedroom, season, age of the house, type of construction, central heating, air conditioning, observed damp patches and mould in the bedroom, window coverings, fitted carpet, type and age of the carpet, ventilation, vacuuming practices for the floor and household pets were considered as possible predictors of indoor fungal levels. In the univariate analysis, the associations between different types of indoor fungi with temperature and relative humidity were examined by computing either Pearson or Spearman correlation coefficients, depending on the distribution of data. Total viable fungal levels in the air and ergosterol were considered as the final indicators of indoor fungal levels. Linear regression was carried out to assess the association between indicators of indoor fungal levels and residential characteristics. Multiple linear regression was carried out to identify the significant independent explanatory variables for the variance in fungal levels, after accounting for the effects of other relevant factors. Variables which were significant at P < 0.2 in the preliminary analysis, exhibited a large effect in the preliminary analysis (1.5-fold increase in the levels compared with the baseline category) or those previously described as important in influencing the fungal levels were included in the model. In the next step, variables which were not significantly associated with fungal levels (P > 0.05) were excluded manually in a sequential backward elimination manner. However, it was decided to control for the symptomatic status of the subject as defined at the second phase of sampling, age of the house, indoor humidity and season in all models irrespective of the statistical significance. The geometric means for the significant predictors adjusted for the other independent predictors, symptomatic

4 1484 S. Dharmage et al. status, age of the house, indoor humidity and season were computed by analysis of covariance (ANCOVA). Ethics The study was approved by the Ethics Review Committee at The Alfred Hospital. Written informed consent was obtained from all participants. Results Ergosterol levels in bedroom floor dust Ergosterol levels in 46 (10%) houses were below the detection limit. The median ergosterol level on the bedroom floor was 3.8 mg/g (range: ) of fine dust and 0.7 mg/m 2 (range: ) area. The correlation between the ergosterol levels per gram of fine dust and per square metre was 0.9 (P < 0.001). Hence we used only the ergosterol levels expressed as micrograms per gram of dust in further analysis. Fungal propagules in bedroom air samples Fifty-five percent of the houses (n ¼ 485) had viable airborne fungal propagules exceeding 500 CFU/m 3. The levels of viable airborne fungal propagules are described in Table 1. The most commonly identified fungi were Cladosporium and Penicillium. Cladosporium was isolated in 90% of the houses and Penicillium in 76% of houses. Among the five fungal genera identified, the largest single contribution to the total spore levels was from Cladosporium (36%). Fungi which were found only in less than 10% of the houses (i.e. Alternaria and Epicoccum) were incorporated into the group of other fungi for further analysis. Weak but statistically significant correlations were observed between the occurrence of some fungi. Penicillium correlated with Aspergillus (r ¼ 0.2, P < 0.001) while Cladosporium correlated with the group of other fungi (r ¼ 0.1, P ¼ 0.005). Interestingly, ergosterol levels were neither related to specific fungi nor to the total fungi in the entire sample, but were weakly correlated with the total fungal levels in the sample of houses which had closed windows at the time of sampling (r ¼ 0.1, P ¼ 0.05) Association between fungi and climatic indicators The mean temperatures and relative humidity in bedrooms were 18 C (SD¼ 3 C; range ¼ C) and 60% (SD ¼ 8.6%; range ¼ 32 80%), respectively. Small but statistically significant positive correlations were observed between viable fungal propagules in the air and interior temperature and humidity. Relative humidity was positively correlated with levels of total fungi (r ¼ 0.1, P < 0.001) and other fungi (r ¼ 0.2, P < 0.001). Temperature was positively correlated with levels of total fungi (r ¼ 0.2 P < 0.001) and Cladosporium (r ¼ 0.3, P < 0.001). Absolute humidity, which is a combination of temperature and relative humidity, was positively correlated with the levels of total fungi (r ¼ 0.3 P < 0.001), Cladosporium (r ¼ 0.3 P < 0.001) and other fungi (r ¼ 0.1 P < 0.001). Absolute humidity was used in further analysis, as it was found to better correlate with fungal levels and would be biologically more relevant. No correlation was observed between these climatic indicators and dust ergosterol levels. Levels of total fungal propagules (P < 0.001), Cladosporium (P < 0.001), and ergosterol (P ¼ 0.01) were found to be significantly associated with the season in which the sampling was carried out. Post hoc comparison of the total fungal propagules and Cladosporium revealed significantly higher geometric means (P < 0.01) in summer ( CFU/m 3 ) and autumn ( CFU/m 3 ) compared with winter ( CFU/m 3 ). In contrast ergosterol levels were significantly higher (P < 0.01) in winter compared with other seasons (geometric mean 4.5 vs 3.3 mg/g of dust). Open windows at the time of sampling were likely to be associated with the viable airborne fungal propagules and dust ergosterol levels as well as the season. Hence the Type of % of homes with Range Median % (specific/ fungi fungi (95% CI) (CFU/m 3 ) (CFU/m 3 ) total) Table 1. Prevalence of fungi in bedroom air from 485 homes Total fungi 100 (99-100) Cladosporium 90 (87-92) Penicillium 76 (72-80) Aspergillus 26 (23-31) Alternaria 5.1 ( ) Epicoccum 2.2 ( ) Other fungi 97 (96-99)

5 Fungi in Australia 1485 effects of season on total fungal propagules, Cladosporium and ergosterol were examined after accounting for whether windows were open at the time of sampling, and the associations remained significant. Interactions between season and open windows on the levels of ergosterol, Cladosporium and total fungal propagules were neither statistically significant (P > 0.05) nor practically important. We also observed an association between the amount of fine dust collected and season (P ¼ 0.007). Geometric mean fine dust levels were higher in winter (0.4 gm) and spring (0.4 gm) compared with autumn (0.3 gm) or summer (0.2 gm) (P < 0.001). Association between residential characteristics and fungi Tables 2 and 3 describe the geometric mean levels of total fungi and ergosterol, which differed significantly by residential characteristics (P < 0.05). In this univariate analysis, visible mould or damp patches and absence of mechanical ventilation by the bedroom ceiling fan or kitchen exhaust fan were associated with higher total fungal propagules in the air as well as higher ergosterol levels in the dust. In contrast, having open windows at the time of sampling and a cat in the home were associated with higher fungal propagules in the air, but paradoxically lower ergosterol levels in the floor dust. Figures 1 and 2 describe the geometric mean levels of the total viable propagules and ergosterol concentration for the residential characteristics identified as independent predictors in the final multivariate analysis, adjusted for the other significant predictors in the model, absolute humidity, and the age of the house and season. We also examined the significance of the identified predictors of ergosterol levels in Total viable spore levels (CFU/m 3 of air) Has a ceiling fan No ceiling fan No cats 1 cat > 1 cat No dogs 1 dog Solid fuel fire within last year No solid fuel fire No observed mould Observed mould-1 surface Observed mould-2 surfaces Observed mould- 3 surfaces Vacuumed within last week Not vacuumed within last week Windows closed at sampling Windows opened at sampling Fig. 1. Independent predictors of total viable fungal propagules in the multivariate analysis. Geometric means and 95% confidence intervals adjusted for absolute humidity, season, age of the house, symptomatic status and other predictors in the model. Table 2. Mean total fungal propagules by categories of the residential characteristics which were significant (P < 0.05) in the univariate analysis Residential Number Total fungal characteristic of homes propagules (CFU/m 3 ) Walls of the house? Double brick Brick veneer Timber Combination Ceiling fan present? Yes No Vertical blinds present? Yes No Sleep with windows open? Yes No Windows opened at the time of sampling? No Yes Time of vacuuming of the bedroom prior to sampling? Within the last week More than 1 week ago Number of surfaces with visible mould in the bedroom? Nil Observed damp patches? None In bedroom or adjoining room In both rooms Solid fuel fire used within house in the last 12 months? Yes No Number of cats? None Numbers do not add up to 485 in some variables due to missing information. bedroom dust while controlling for the sampling technique (i.e. standard vs extended) and the results were similar. These models explained 20% of the variance in each of ergosterol levels and fungal propagules. Presence of more than one cat in the home had the greatest independent effect (1.6-fold increase) on the total viable propagules. Old wall-to-wall carpets had the greatest effect on ergosterol (twofold increase).

6 1486 S. Dharmage et al. Table 3. Mean ergosterol (mg/g of dust) by categories of the residential characteristics which were significant (P < 0.05) in the univariate analysis Residential Number Geometric mean characteristic of homes ergosterol (mg/g) Fitted carpets? No < 5-year-old carpets year-old carpets Frequency of bedroom airing? Most days times a week /week < 1/week Sleep with windows open? Yes No Windows opened at the time of sampling? No Yes Number of surfaces with visible mould in the bedroom? Nil Observed damp patches? No In bedroom or adjoining room In both rooms Extractor fan over cook-top? None Not used regularly Used regularly Number of cats? None Numbers do not add up to 447 in some variables due to missing information. Absolute humidity was found to be a significant independent predictor of total viable spore numbers, but not of ergosterol levels. Higher viable fungal propagules in the air during summer and autumn were found to be confounded by the absolute humidity. However, season was a significant predictor of ergosterol levels in its own right with winter having significantly higher levels than other seasons. Visibly mouldy surfaces were found to be an independent predictor of both total viable fungal propagules and ergosterol levels with a trend for each to increase with the number of mouldy surfaces (Figs 1 and 2). Having damp patches was not a significant predictor in the multivariate analysis, as the higher levels of viable fungal propagules and ergosterol in damp houses were confounded by more mouldy surfaces. Total viable fungal propagules and ergosterol levels were 1.2- and 1.4-fold higher in houses where both bedroom and the adjoining room had damp patches compared with houses with no damp patches in either. One or more mouldy surfaces were observed in 90% of the houses with damp patches in both bedroom and the adjoining room, but only in 27% of the houses without damp patches in either. The different effects of open windows at the time of sampling on ergosterol and viable fungal propagules were confirmed in the multivariate analysis. Having a cat at home was found to be an independent predictor of higher total viable fungal propagules, but the observed effect of having a cat on the levels of ergosterol in the univariate analysis was found to be confounded by ventilation. Houses with cats were better aired and more likely to have open windows at the time of sampling. Ergosterol levels were lower in bedrooms with frequent airing and open windows. Having cats in the home did not have a significant effect on ergosterol levels, once adjustment was made for airing and open windows. Having a dog at home emerged as an independent predictor of viable fungal propagules in the multivariate analysis. This association was not found to be significant in the univariate analysis, as a result of the negative confounding effect of having cats at home. Only 28% of the houses with dogs had cats compared with 35% of the houses without dogs, while houses with cats had higher levels of viable fungal propagules. Discussion Propagules of Cladosporium and Penicillium dominated the fungal aerospora in the houses surveyed in Melbourne during Less frequently, propagules of Aspergillus, Alternaria, and Epicoccum were identified, while a range of unidentified species was present at low levels in most bedrooms. In this respect, our results confirm other studies of domestic environments that describe Cladosporium and Penicillium as the most prevalent and abundant aerospora [8,14,29,30]. We observed a correlation between Penicillium and Aspergillus which was statistically significant and biologically plausible. Penicillium and Aspergillus are wellknown as soil fungi and taxonomically related [31]. However the correlation between them was weak and unlikely to be clinically relevant. The prevalence and seasonal abundance of Cladosporium inside homes mirrors the pattern described for the outside aerospora of Melbourne [32]. Propagules of Alternaria which is considered to be a major allergenic fungal genus, were rare indoors, as they are outdoors in Melbourne [32].

7 Fungi in Australia 1487 Ergosterol levels ( µ g/g of fine dust) No carpet New carpet Old carpet Extractor fan used regularly Extractor fan not used regularly No extractor fan over cook-top Windows opened at sampling Windows closed at sampling Airing > 4 times 2-4 times 1/week Less than 1 week No observed mould Observed mould-1 surface Observed mould-2 surfaces Observed mould- 3 surfaces Summer Autumn Winter Spring Fig. 2. Independent predictors of ergosterol levels in the multivariate analysis. Geometric means and 95% confidence intervals adjusted for absolute humidity, season, age of the house, sympto- As in the outdoor environment [32], total propagule levels in indoor environments were highest in summer and autumn, and lowest in winter. The similar seasonal variation of indoor and outdoor aerospora highlights the influence of external contamination on the levels of indoor fungi [33,34]. However, one should not necessarily conclude that fungi indoors come entirely from outdoors. In the multivariate analysis, the seasonal variation in indoor viable spores was found to be related to the absolute humidity. The findings related to relative humidity, temperature and season should be interpreted while taking limitations such as the short duration of the measurement and the cross-sectional nature of the data into account. The total number of fungal propagules in the air of Melbourne homes was lower than found in the nearby LaTrobe Valley where the median over a year was found to be 812 CFU/m 3 [8]. The LaTrobe Valley is located in a rural area 200 km south-east of Melbourne, supporting evidence that homes in rural areas harbour higher levels of airborne fungal propagules than those in urban areas [35]. However, 55% of the Melbourne homes had high levels of indoor airborne fungi i.e. > 500 CFU/m 3 which are considered hazardous according to World Health Organization guidelines [36]. An independent assessment of fungal exposure is given by measuring ergosterol levels in house dust [20]. This technique provides an indication of cumulative exposure rather than the snapshot provided by brief air sampling, and should be less sensitive to the short-term fluctuations caused by open windows. Estimates of fungal biomass based on ergosterol assays also include non-viable fungal propagules and those of fungi that do not grow in culture, but which nevertheless contribute to the total fungal allergen load. However, this estimate of fungal biomass is affected by the amount of dust and the age of fungi in the dust, as well as variations in ergosterol content of different fungal species at different developmental stages. Because of these factors, our finding that viable propagule counts and ergosterol levels are only loosely related is not surprising. The influence of externally originating propagules on the internal environment is supported by our finding that the strongest correlation between viable propagule counts and ergosterol levels was found in rooms that had been closed for at least 24 h before sampling. Our results contrast with a study of Canadian homes in mid-winter, where a strong relationship between fungal propagule levels and ergosterol in house dust was established (r ¼ 0.9, P < 0.001) [37]. Presumably the few propagules present in the snow-covered external environment were excluded from the indoor samples by fundamental differences in home design and less frequent opening of windows and doors. A much closer relationship between the two estimates of fungi would be expected under these conditions. This highlights the importance of measuring ergosterol as a marker of cumulative exposure and identifying measures to reduce the levels, in addition to the levels of fungal propagules, especially where withinhouse variation is high. Having cats or dogs in the home was found to increase the total viable fungal propagules significantly. An association between household pets and total viable fungal propagules in the air has been observed previously [13]. This has been attributed to outdoor fungi being transported indoors via animal dander. In contrast, Verhoeff et al. [10] observed significantly lower viable fungal propagules in the floor dust with the presence of pets at home in a univariate analysis. Similarly we observed lower ergosterol levels in the presence of household cats, which was found to be confounded by frequent airing. Old wall-to-wall carpets had the greatest independent effect on ergosterol among the variables examined. Similarly, the levels of fungal propagules in the floor dust have been observed to be significantly higher with wall-to-wall carpets [7,10]. The influence of fitted carpets on indoor fungal levels in combination with the well-established association between fitted carpets and dust mite allergen levels [19,38] suggests a potentially beneficial effect of a carpet-free environment on asthma. Visible mould correlated strongly with both airborne propagule numbers and ergosterol levels in house dust, indicating the importance of indoor reservoirs on viable fungal propagules as well as ergosterol levels. Fungal propagule levels have been observed to be significantly higher in bed dust from rooms with damp and mouldy patches [10,11]. These results suggest that fungal propagules represent more than just external contamination and indicate the importance of indoor reservoirs on indoor fungal exposure. We observed both measures of fungal

8 1488 S. Dharmage et al. exposure to be higher in the presence of damp patches, but they were no longer important when we controlled for visible mould. Visible mould is more likely to be a step in the causal pathway between damp patches and high levels of fungi rather than a confounder. In our study, the regular use of an extractor fan and frequent airing of the bedroom were associated with lower ergosterol levels in dust. The presence of ceiling fans was associated with lower viable fungal propagules in air. Air cleaning devices [7,13] and extractor fans [11] have also been found to reduce the airborne viable fungal propagules. Ventilation through open windows has been shown to reduce the concentration of airborne total fungal propagules [11]. However, we observed higher estimates of airborne viable fungal propagules with open windows at the time of sampling, but not with frequent airing. These results suggest a major impact of the outdoor fungi on the levels of indoor airborne fungal propagules through open windows at the time of sampling, but the lack of an association between airborne fungal propagules and airing practices suggest that the effect of this external contamination is probably transient. On the other hand, ergosterol levels were lower with open windows at the time of sampling, which is likely to reflect frequent airing practices. These results indicate that the overall effect of natural ventilation on indoor fungal exposure is beneficial. Regular use of kitchen exhaust fans was associated with lower ergosterol levels. However, the levels of ergosterol in houses without an extractor fan were not significantly different from the levels of houses with an extractor fan. This contradiction is probably due to regular airing of the kitchen in some houses without an extractor fan, although we did not have information to examine this. We also observed lower levels of viable fungal propagules in houses that were vacuumed recently and those that had a solid fuel fire within 12 months prior to sampling which have not been observed before. Vacuuming may remove the fungal propagules in floor dust which is a reservoir of the viable fungal propagules in the air. Although the sample of subjects studied comprised a higher proportion of symptomatic subjects than in the reference population of adults aged years, our results were not altered when the sample was reweighted to match the reference population. However, the low participation rate among the original sample invited to the laboratory should be taken into account in generalizing the results to the young adult population of Melbourne. In conclusion, our results suggest that the indoor fungal levels in the majority of Melbourne homes are hazardous according to World Health Organization guidelines and indoor exposure to fungi could be reduced by modifying the home environment. This study highlights the need for further research in this field, especially interventions to modify the home environment and to establish causal relationships. From a public health perspective, we would encourage frequent airing, regular use of kitchen exhaust fans and bedroom ceiling fans, removal of old wall-to-wall carpets, exclusion of pets, frequent vacuuming and removing visible mould patches. These measures should significantly reduce exposure to fungal allergens in the home, although health benefits remain to be established. Acknowledgements We would like to thank the Victorian Health Promotion Foundation and Department of Human Services for funding this study. We acknowledge the assistance received in data collection from Tina Colagen and John Elliot. References 1 Miller JD. Fungi as contaminants in indoor air. Atmos Environ 1992; 26A: Wageningen Nv, Waegenmaekers M, Brunekreef B, Boleij J. Health complaints and indoor moulds in relation to moist problems in home. In: Seifert BEH, Fischer M, Riden M, Wegner J, eds. Fourth International Conference on Indoor Air Quality and Climate. Berlin: Institute of Water, Soil and Air Hygiene: Indoor Air, 1987: Dales RE, Burnett R, Zwanenburg H. Adverse health effects among adults exposed to home dampness and molds. Am Rev Respir Dis 1991; 143: Hu FB, Persky V, Flay BR, Richardson J. An epidemiological study of asthma prevalence and related factors among young adults. J Asthma 1997; 34: Li CS, Hsu LY. Home dampness and childhood respiratory symptoms in a subtropical climate. Arch Environ Health 1996; 51: Verhoeff AP, van Strien RT, van Wijnen JH, Brunekreef B. Damp housing and childhood respiratory symptoms: the role of sensitization to dust mites and molds. Am J Epidemiol 1995; 141: Wickman M, Gravesen S, Nordvall SL, Pershagen G, Sundell J. Indoor viable dust-bound micro-fungi in relation to residential characteristics, living habits, and symptoms in atopic and control children. J Allergy Clin Immunol 1992; 89: Garrett MH, Hooper BM, Cole FM, Hooper MA. Airborne fungal spores in 80 homes in the Latrobe Valley, Australia: Levels, seasonality and indoor-outdoor relationship. Aerobiologia 1997; 13: Pasanen AL, Niininen M, Kalliokoski P. Airborne Cladosporium and other fungi in damp versus reference residences. Atmos Environ 1992; 26B: Verhoeff AP, van Wijnen JH, van Reenen-Hoekstra ES, Samson RA, van Strien RT, Brunekreef B. Fungal propagules in house dust. II. Relation with residential characteristics and respiratory symptoms. Allergy 1994; 49: Garrett MH, Hooper BM, Rayment PR, Abramson MJ. Indoor airborne fungal spores, house dampness and associations with environmental factors and respiratory health in children. Clin Exp Allergy 1998; 28:

9 Fungi in Australia Reponen T, Nevalainen A, Raunemaa T. Bioaerosol and particle mass levels and ventilation in Finnish homes. Environ Int 1987; 15: Koster JAD, Thorne PS. Bioaerosol concentration in noncomplaint, complaint and intervention homes in the Midwest. Am Ind Hyg Assoc J 1995; 56: Prahl P. Reduction of indoor airborne mould spores. Allergy 1992; 47: Abramson MJ, Kutin J, Raven J, Lanigan A, Czarny D, Walters EH. Risk factors for asthma among young adults in Melbourne. Respirology 1996; 1: Burney PG, Luczynska C, Chinn S, Jarvis D. The European Community Respiratory Health Survey. Eur Respir J 1994; 7: Felder RM, Rousseau RW. Elementary Principles of Chemical Engineering 2nd edn. Toronto, Canada: John Wiley and Sons Inc, Platts-Mills TAE, De Weck AL. Dust mite allergens and asthma. A worldwide problem. J Allergy Clin Immunol 1989; 83: Dharmage S, Bailey M, Raven J et al. Residential characteristics influence Der p 1 levels in homes in Melbourne, Australia. Clin Exp Allergy 1999; 29: Martin F, Delaruelle C, Hilbert JL. An improved ergosterol assay to estimate fungal biomass in ectomycorrhizas. Mycol Res 1990; 94: Vijay HM, Burton M, Young NM, Copeland DF, Corlett M. Allergenic components of Cladosporium herbarum. Grana 1991; 30: Licorish K, Novey HS, Kozak P, Fairshter RD, Wilson AF. Role of Alternaria and Penicilliumspores in the pathogenesis of asthma. J Allergy Clin Immunol 1985; 76: Dixit AB, Lewis WH, Wedner WH. The allergens of Epicoccum nigrum. J Allergy Clin Immunol 1992; 90: Kurup VP, Resnick A, Scribner GH, Gunesekaran M. Enzyme profile and immunochemical characterization of Aspergillus fumigatus antigens. J Allergy Clin Immunol 1986; 78: Yunginger JW. Studies on Alternaria allergens. J Allergy Clin Immunol 1976; 57: SAS Institute Inc. Stat User s guide. Cary, NC.: SAS, 1988, release 6.03 edition. 27 Stata Corporation. Stata for windows. TX: Stata Corporation, 1997, Version Cochran WG. Sampling Techniques, 3rd edn. New York: John Wiley and Sons, Singh A, Gangal SV, Singh AB. Airborne fungi in the hospitals of Metropolitan Delhi. Aerobiologia 1994; 10: Li C, Hsu L. Fungus allergens inside and outside the residences of atopic and control children. Arch Environ Health 1995; 50: Su HJ, Rotnitzky A, Burge HA, Spengler JD. Examination of fungi in domestic interiors by using factor analysis: Correlations and associations with home factors. App Environ Microbiol 1992; 58: Mitakakis T, Ong EK, Stevens A, Guest D, Knox RB. Incidence of Cladosporium, Alternaria and total fungal spores in the atmosphere of Melbourne (Australia) over three years. Aerobiologia 1997; 13: Li D-W, Kendrick B. A year- round comparison of fungal spores in indoor and outdoor air. Mycologia 1995; 87: Pasanen AL, Reponen T, Kallioskoski P, Navalainen A. Seasonal variation of fungal spore levels in indoor and outdoor air in the subarctic climate. In: 5th International Conference Indoor Air Quality and Climate. Ottawa: Canada Mortgage and Housing Corporation, 1990: Pasenen AL. Airborne mesophilic fungal spores in various residential environments. Atmos Environ 1992; 26A: World Health Organization. Indoor Air Quality: Biological Contaminants. Copenhagen: WHO Regional Publications, 1990, European Series no Miller JD, Laflamme AM, Sobol Y, Lafontaine P, Greenhalgh R. Fungi and fungal products in some Canadian Houses. Int Biodeterior 1988; 24: Wickens K, Siebers R, Ellis I et al. Determinants of house dust mite allergen in homes in Wellington, New Zealand. Clin Exp Allergy 1997; 27:

10 This document is a scanned copy of a printed document. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material.