Indoor Air Quality and Health: Validity and

International Journal of Epidemiology © International Epidemiological Association 1997

Vol. 26, No. 1 Printed in Great Britain

Indoor Air Quality and Health: Validity and Determinants of Reported Home Dampness and Moulds ROBERT E DALES,* DAVID MILLER** AND ED MCMULLEN† Dales R E (Health Effects-Research Section, Health Canada), Miller D and McMullen E. Indoor air quality and health: Validity and determinants of reported home dampness and moulds. International Journal of Epidemiology 1997; 26: 120–125. Background. Questionnaire-based surveys from several countries have consistently detected adverse health associated with home dampness and mould growth. Methods. To test the validity of questions commonly used to indicate the presence of indoor mould, questionnaires were administered in 403 homes where dust samples were taken for viable fungi and air samples for ergosterol. Results. Geometric mean concentrations of the total viable fungi were 255 (SE 116) × 103 CFU/g when mouldy odours were reported and 155 (SE 55) when odours were not reported (P = 0.01). Similarly, reported water damage was associated with a 50% increase (P = 0.06). Geometric mean concentrations of the predominantly indoor-source fungi, Aspergillus plus Penicillium, were twice as high when mould or mildew was reported than when not mentioned (P = 0.01). The presence of reported mould or water damage was unrelated to the presence of detectable levels of ergosterol. There was evidence for reporting bias: in the presence of low concentrations of viable fungi in dust, respondents reporting allergies were more likely to report visible mould growth (odds ratio [OR] = 1.8, 95% confidence interval [CI] : 0.9–3.5, P = 0.10. In the presence of elevated concentrations of dust fungi, respondents who smoked were less likely to report visible mould growth, (OR = 0.4, 95% CI : 0.2–0.7, P = 0.005). Conclusions. Reported mould, water damage, and mouldy odours were associated with elevated levels of indoor fungi. However, inaccuracy was high and there was evidence of a systematic reporting bias. Future research should concentrate on developing accurate objective measures of exposure to fungi, and then use this information to develop valid questionnaires. Currently, objective measures, not questionnaires, are recommended to clarify the health effects of indoor fungi. Keywords: respiratory health, epidemiology, questionnaires, air pollution, fungi

Fungi are known to cause a spectrum of clinical diseases.1–5 Potentially pathogenic fungi are not uncommon in the indoor environment6–8 and illness possibly caused by mycotoxins has been documented in case reports.9–13 However, the burden of illness in the population attributable to fungi in private homes and public buildings is not yet known. Many epidemiological studies from several countries have consistently detected an association between respiratory symptoms and reported home dampness and mould growth,14–22 but causality in these studies has not been established. A major reason for this uncertainty is that exposure to fungi has usually been assessed by the occupant’s response to questions concerning flooding, water damage, and visible moulds in the house. Fungal contamination

however may not be visible to the occupant if it is within walls or ceilings, or under carpets. Although questionnaires are able to reach large segments of the population at relatively low cost, the ability to accurately measure mould growth is uncertain. Reporting biases are a potential problem with all questionnaire results, and underreporting of home dampness by parents of symptom-free children has been demonstrated.23 Uncertainty about the accuracy of questionnaire reports limits acceptance of a causal relation between fungi and health. Testing the validity of questionnaire reports as an indicator of the burden of indoor fungi is the subject of the present study. Occupants’ reports of moulds were compared to simultaneous measures of home fungal biomass.

* Health Effects-Research Section, Health Canada. ** Mycotoxin Research Unit, Agriculture Canada. † Biostatistics section, Health Canada. Reprint requests to: Dr Robert Dales, Room 203, Environmental Health Centre, Tunney’s Pasture, Ottawa, Canada, K1A OL2.

MATERIALS AND METHODS Study Population The study was carried out in Wallaceburg, Ontario, a city of approximately 14 000 people. Field work was 120

121

INDOOR AIR QUALITY AND HEALTH

carried out during the winter of 1993–1994 in order to minimize indoor contamination from outdoor fungi. Using elementary school lists, a census was performed which indicated that there were 1438 families with elementary schoolchildren. Some 670 families could not be reached because they had moved or had incorrect or unlisted phone numbers (n = 337), or could not be reached by telephone after multiple calls (n = 333). In all, 403 families and their homes were studied. The Questionnaire Questions concerning visible mould growth and conditions conducive to mould growth were as follows: 1. Wet or damp spots. ‘Within the past year … have you ever had wet or damp spots on surfaces inside your present home other than in the basement (e.g. on walls, wallpaper, ceilings or carpets)?’ 2. Mould or mildew. ‘Have you ever had mould or mildew growing on any surface inside your present home? (e.g. on walls, wallpaper, ceilings, carpets, shower curtain, etc.) … in the basement? … in the shower area(s)? … in other areas of your house … (within) the past year?’ 3. Water damage. ‘Have you ever had a leak, flooding, or water damage in your basement … (within) the past year?’ These questions had been previously administered to several thousand residents22 and reproducibility of the questionnaire, previously tested, was found to be moderate to good.24 Test-retest agreement was 87% for visible mould growth and 92% for damp interior surfaces.24 For the purposes of the current study, questions one through three were duplicated using the present tense. In addition, questions concerning mouldy odours were added: ‘Have you ever experienced “mouldy” odours in any area of your home?’ and ‘Do you currently experience mouldy odours in any area of your house?’ Fungal Measurements Viable fungi. Dust was collected from the entire main living area floor over 10 minutes using a Euroclean HEPA-filtered UZ 930 vacuum®. DACI cotton filters (normally used for dust mite and cat antigen collection) were attached at the entry of the collection orifice of the vacuum to avoid cross-contamination of samples. Vacuum utensils were decontaminated between homes. One, ten, and a thousand-fold dilutions in 2% peptone broth were made from each of two 0.5 g dust samples from the same home. Nutrient media were Martin’s Rose Bengal Agar with and without 25% glycerine to favour the isolation of both mildly xerophilic and hydrophilic species.25 Ergosterol. Measurement of ergosterol (24B-methlcholesta-5,7,trans 22-trien-3B-ol), the predominant fungal

TABLE 1 Characteristics of the questionnaire respondents stratified by the presence or absence of reported home dampness, mould, or water damage in the past 12 months Characteristics

age, mean (SD) gender, % female education, % no post-secondary residence, % owned allergies % asthma % usual coughs % smoke %

Reported mould, dampness, or water damage Present (n = 237)

Absent (n = 164)

36 (5) 51 43 84 27 13 24 28

37 (6) 51 51 87 26 12 20 29

sterol, provides a quantitative measure of total fungal matter.26–28 Ergosterol values for air samples were used in the present study to complement the viable fungi measured in the dust. Air samples from both the main living area and the child’s bedroom were collected over a 14–20 h period at a flow of 1.7–2.0 lpm. Weekly calibration was carried out and sampling time was recorded. Ergosterol was measured using high pressure liquid chromatography and gas chromatography mass spectrometry was used to confirm all values greater than zero.29 The detection limit was approximately 500 pg of ergosterol, or the equivalent of about 200 spores.29 Statistical Analysis The distribution of fungal colony forming units (CFU) per gram of dust was skewed to the right, necessitating logarithmic transformations prior to t-testing differences. The test for normality of the transformed data was not rejected.30 All summary statistics for CFU data appear as geometric means and geometric errors. T-tests were used to check the statistical significance of differences in means between questionnaire responses. The ergosterol values could not be normalized because of the large number of values below the detection limit; 79% of the bedroom and 90% of the living room values. Thus, these data were dichotomized into detectable and undetectable. Odds ratios (OR) and 95% confidence intervals (CI) were used to examine the relation between questionnaire reports and ergosterol values.

RESULTS In all, 403 questionnaires were completed, fungal dust samples were available for 362 homes and ergosterol values were available for 373 homes. Table 1 describes

122

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

TABLE 2 The association between questionnaire reports of mould and dampness occurring in the main living area within the past year, and objective measures of indoor fungia Dust fungi (103 CFUb/g)

Questionnaire report

Air ergosterol (10–3 µg/l)

total

Aspergillus plus Penicillium

mean

% detectable

mould or mildew absent present

(n = 332) (n = 25)

162 (54) 237 (168)

21 (8) 45 (36)

0.8 (0.4) 3.0 (3.0)

10 10

wet or damp spots absent present

(n = 325) (n = 37)

158 (53) 213 (127)

21 (9) 33 (22)

0.8 (0.4) 2.5 (2.4)

10 11

water damage absent present

(n = 285) (n = 63)

150* (52) 238 (124)

22 (9) 27 (19)

0.9 (0.4) 1.6 (1.6)

10 5

mouldy odours absent present

(n = 269) (n = 71)

155† (55) 255 (116)

22 (9) 28 (19)

1.2 (0.6) 0.2 (0.2)

11 10

a The mean (standard error) are geometric for fungi and arithmetic for ergosterol. One exceedingly high ergosterol value, 1150 × 10–3µg/l, was excluded because it markedly skewed all averages. T-tests of mean differences in CFU/g, and χ2 tests of differences in the percentage of detectable ergosterol were not statistically significant at the 0.05 level. b Colony forming units. *P = 0.057; † P = 0.012.

the characteristics of the questionnaire respondent. The majority were female, with some post-secondary school education. One-quarter reported allergies, and usually coughed. There were no significant differences between those reporting dampness or mould in the previous 12 months and those not. The geometric mean CFU/g of both total fungi, and the subset of fungi with a predominantly indoor source, Aspergillus plus Penicillium, were contrasted between homes with and without detectable levels of ergosterol. The Spearman rank correlation coefficient was 0.84 for the association between total CFU/g obtained from the glycerol media and the non-glycerol media, and was 0.86 for the association between CFU/g of Aspergillus plus Penicillium between the two media. Only results obtained using the glycerol-based media are presented. Validation of questionnaire reports was done separately for all three fungus indicators: ergosterol, total viable fungi obtained from glycerol-containing medium, and Aspergillus plus Penicillium obtained from glycerol-containing medium. The associations between living area fungus measures and questionnaire-reported indicators of mould growth in the past year are presented in Table 2. The geometric mean concentrations of total viable fungi were 50% greater when visible mould, water damage, or mouldy odours were present. The sum of the predominant indoor air fungi, Aspergillus

plus Penicillium, were best indicated by visible mould growth. The geometric mean concentrations were 45 000 CFU/g when mould or mildew was reported and 21 000 CFU/g when not reported. Although not presented here, the fungal counts using non-glycerol medium had a similar association with the reported mould and dampness. Mean airborne ergosterol was greater when mould, moisture or water damage was reported. The data were highly skewed and the percentage of homes with detectable ergosterol was unrelated to the questionnaire reports (Table 2). In the bedroom (Table 3), neither mean levels of ergosterol nor the percentage of samples with detectable ergosterol were related to subjective reports of moulds or dampness. Bedroom ergosterol values were neither correlated with living area ergosterol values nor CFU/g of main living area dust. The results obtained using questions concerning current mould growth were similar in direction and magnitude to those addressing the past year which are shown in Tables 2 and 3. Selected characteristics of the questionnaire respondent were examined for their influence on the reporting of current wet or damp spots, and mould or mildew (Table 4). Dust concentrations of viable fungi were dichotomized into low and high using the median value of 175 000 CFU/g (glycerol medium) as a cutpoint. There was evidence for reporting bias: in the presence

123

INDOOR AIR QUALITY AND HEALTH

TABLE 3 The association between questionnaire reports of mould and dampness occurring in the child’s bedroom within the past year, and objective measures of indoor mouldsa Air ergosterol (10–3 µg/l)

Questionnaire report

mean (SE)

% detectable

mould or mildew absent (n = 330) present (n = 57)

11.6 (6.4) 37.1 (31.4)

22 23

wet or damp spots absent (n = 364) present (n = 29)

15.8 (7.6) 1.0 (0.5)

22 14

water damage absent present

(n = 313) (n = 66)

12.4 (6.7) 3.0 (1.6)

22 17

mouldy odours absent present

(n = 294) (n = 74)

16.7 (9.1) 13.1 (8.5)

21 22

a The mean (standard error) are arithmetic. T-tests of mean differences in CFU/g and χ2 tests of differences in the percentage of detectable ergosterol were not statistically significant at the 0.05 level.

of low concentrations of viable fungi in dust, respondents reporting allergies were more likely to report visible mould growth (OR = 1.8, 95% CI : 0.9–3.5, P = 0.10). In the presence of elevated concentrations of dust fungi, respondents who smoked cigarettes were less likely to report visible mould growth, (OR = 0.4, 95% CI : 0.2–0.7, P = 0.005). The parents’ age, gender, and education did not influence the relation between measured and reported mould indicators.

DISCUSSION Viable fungi in dust was greater in living areas when occupants reported mould or mildew, mouldy odours, wet or damp spots, and water damage. However, the ability to detect airborne ergosterol was unrelated to occupant’s reports. Prior to the present study, there had been very little information concerning the validity of these frequently used questions to indicate fungus growth. Waegemaekers et al. studied 328 adults and 190 children from 185 homes in a Dutch coastal town.18 Homes were classified as damp if two or more of the following indicators were present: visible mould growth, damp spots, silver-fish, sowbugs, a stale odour, or a wet crawl space. Homes were classified as dry if the respondent’s opinion of their home was dry or rather dry. Viable airspora was measured over 10–15 minutes using an Anderson sampler. Geometric mean (SD) concentrations were 214 (2.8) CFU/m3 in the damp homes and 91 (1.9) in homes classified as dry. Contributions of the present study are the use of two different nutrient media and measurement of ergosterol over a 12- to 18-h period. In 88 Edinburgh homes, Strachan et al. compared questionnaire information to 3-minute air samples of viable spores using an Anderson sampler.15 Median fungus concentrations ranged from 200 to 294 CFU/m3, unassociated with reported visible mould growth. The authors suggested that the short duration of sampling may not accurately represent true exposure, which varies considerably with physical activity in the room. To reflect longer term exposure, we chose dust samples of viable fungi and 18-h sampling of airborne ergosterol. Platt et al. studied 597 households

TABLE 4 The association between self-reported symptoms and the relative accuracy of reporting visible mould growth Fungal CFUa/g ù medianb

Characteristic of respondent

Mould reported

cough asthma allergies smoking

a

absent present absent present absent present absent present

Odds ratio (95% CI)

yes

no

62 21 73 11 59 25 67 17

73 18 81 9 71 22 55 37

1.4 (0.7–2.8) 1.4 (0.5–3.4) 1.4 (0.7–2.7) 0.34 (0.2–0.7)**

Fungal CFU/g ù median Mould reported yes

no

60 19 68 11 57 24 59 21

75 23 86 13 80 19 71 28

Colony forming units. The median value was 175 000 CFU/g of living area dust obtained using glycerol-containing medium *P = 0.10; **P = 0.005. b

Odds ratio (95% CI)

1.0 (0.5–2.1) 1.1 (0.5–2.5) 1.8 (0.9–3.5)* 0.9 (0.5–1.8)

124

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

in areas of public housing in Glasgow, Edinburgh and London.16 The correlation coefficient between selfreported dampness/mould and air spore count was 0.14, P , 0.001. Recently, in a 52-home study by Verhoeff et al.31 fungal CFU/g of house dust were approximately 20% higher in the setting of observed damp spots and mould growth. The lack of a stronger association between measured fungal biomass and reported moulds can theoretically be explained by problems both with fungal measures and respondent reports. Non-speciated CFU represent only viable fungi and do not distinguish those growing indoors from those brought in from outdoors. The latter were minimized in the present study by collecting the field data in the winter when outdoor levels are at their nadir. In the present study, there was no correlation between bedroom ergosterol and living area fungal counts. Fungal counts in the living area thus do not represent the average levels found throughout the house, which may be what the respondents are attempting to estimate. To measure total indoor exposure, sampling from several areas in the home over longer time periods are needed. Further, poor reproducibility of fungal measures has recently been reported.32 The reliability coefficients of floor dust CFU/g over a 6-week period ranged from 0.01 to 0.32. This high variability would tend to minimize any correlation with the true average fungal burden and occupant’s reports. However, to enhance the accuracy of fungal exposure in the present study, dust and air measures were taken, with the latter collected over several hours to average out transient fluctuations. Apart from fungal measures, self-reports are complex, requiring observation, interpretation, judgement as to degree of severity, and finally verbal expression. Variability is theoretically possible at all levels of this process. Fungal growth can go unnoticed by the occupant if it is beneath carpets, behind wall coverings, or in humidifier reservoirs or refrigerator drainage pans. This could account for underreporting. Further, total fungal counts combine variable quantities of different species, some of which perhaps may be more apparent to the residents than others, leading to inaccurate reporting. For example, small quantities of fungi around bathtubs and window sills are easily seen and perhaps frequently reported, but may not importantly influence the total levels found in the floor dust. This would tend to cause overreporting. We have also provided evidence for a systematic reporting bias; parents with chronic cough, asthma or allergies, may overreport relative to asymptomatic parents. This type of bias could artifactually create a positive association between mould growth and health. Although these odds ratios were approximately

1.4, this is similar in magnitude to the odds ratios between symptoms in adults and reported mould growth in previous studies. If this bias exists unrecognized it would be assumed that fungus exposure leads to symptoms or unreported asthma, but the present study suggests that having symptoms can lead to increased reported mould exposure. In summary, questionnaire reports are associated with the levels of viable indoor fungi in dust, but not airborne ergosterol. Larger sample sizes, prolonged monitoring, and measurements in each room of the house would help to overcome the poor precision in fungus measures and in personal exposures. Future research should concentrate on developing accurate objective measures of exposure to fungi, and then using this information to develop valid questionnaires. Currently, objective measures and not questionnaires are recommended to clarify the health effects of indoor fungi. Relying on questionnaire information alone for studying the health effects of indoor fungi should be abandoned because of inaccuracy and possible systematic biases which would not be corrected by large sample sizes.

ACKNOWLEDGEMENTS The authors thank Mr Jim White of Canada Mortgage and Housing Corporation for help in designing the study, Magellan Engineering, Ottawa, Canada, for performing the field work, and Dr David Malloch, University of Toronto, Canada, for enumerating and speciating the fungi. REFERENCES 1

Flannigan B, McCabe E M, McGarry F. Allergenic and toxigenic micro-organisms in houses. J Appl Bacteriol Symp Ser 1991; 20: 61S–73S. 2 Burge H. Bioaerosols: Prevalence and health effects in the indoor environment. J Allergy Clin Immunol 1990; 86: 687–701. 3 Tobin R S, Baranowski E, Gilman A, Kuiper-Goodman T, Miller J D, Giddings M. Significance of fungi in indoor air: report from a working group. Can J Public Health 1987; 8 (Suppl.): S1–S30. 4 Jarvis B B. Mycotoxins and indoor air quality. In: Morey P R, Feeley J C Jr, Otten J A (eds). Biological Contaminants in Indoor Environments. ASTM special technical publication 1071. American Society for Testing and Materials Committee D-22 on Sampling and Analysis of Atmospheres. Section .06 on Biological Aerosols. 1990. 5 Broder I. An Instructional Review of Airborne Microorganisms for Safety Officers of Human Resources Development Labour Component. Final report to Technical Services Division, Occupational Safety and Health Branch, Labour Component, Human Resources Development, Government of Canada, 29 March 1994. Contract: 1993, File YR828-93-049.

INDOOR AIR QUALITY AND HEALTH 6

Smith J E, Anderson J G, Lewis C W, Murad Y M. Cytotoxic fungal spores in the indoor atmosphere of the damp domestic environment. FEMS Microbiol Letters 1992; 79: 337–43. 7 Miller J D, Laflamme A M, Sobol Y, Lafontaine P, Greenhalgh R. Fungi and fungal products in some Canadian houses. Int Biodet 1988; 24: 103–20. 8 Smoragiewicz W, Cossette B, Boutard A, Krzystyniak K. Tricothecene mycotoxins in the dust of ventilation systems in office buildings. Int Arch Occup Environ Health 1993; 65: 113–17. 9 Croft W A, Jarvis B B, Yatawara C S. Airborne outbreak of trichothecene toxicosis. Atmos Environ 1986; 20: 549–52. 10 Di Paolo N, Guarnieri A, Loi F, Sacchi G, Mangiarotti A M, Di Paolo M. Acute renal failure from inhalation of mycotoxins. Nephron 1993; 64: 621–25. 11 Gordon K E, Masotti R E, Waddell W R. Tremorgenic encephalopathy: a role of mycotoxins in the production of CNS disease in humans? Can J Neurol Sci 1993; 20: 237–39. 12 Yoshida K, Ando M, Araki S. Acute pulmonary edema in a storehouse of moldy oranges: a severe case of the organic dust toxic syndrome. Arch Environ Health 1989; 44: 382–84. 13 Johanning E, Biagini R, Hull D, Morey P R, Jarvis B, Landsbergis P. Health and immunology study following exposure to toxigenic fungi (Stachybotrys chartarum) in a water-damaged office environment. Int Arch Occup Environ Health 1996; 68: 207–18. 14 Strachan D P, Elton R A. Relationship between respiratory morbidity in children and the home environment. Fam Pract 1986; 3: 137–42. 15 Strachan D P, Flannigan B, McCabe E M, McGarry F. Quantification of airborne moulds in the homes of children with and without wheeze. Thorax 1990; 45: 382–87. 16 Platt S D, Martin C J, Hunt S M, Lewis C W. Damp housing, mould growth, and symptomatic health state. BMJ 1990; 298: 1673–78. 17 Hyndman S J. Housing dampness and health amongst British Bengalis in East London. Soc Sci Med 1990; 30: 131–41. 18 Waegemaekers M, van Wageningen N, Brunekreef B, Boleij J. Respiratory symptoms in damp homes. A pilot study. Allergy 1989; 44: 192–98. 19 Brunekreef B, Dockery D W, Speizer F E, Ware J H, Spengler J D, Ferris B G. Home dampness and respiratory morbidity in children. Am Rev Respir Dis 1989; 140: 1363–67. 20 Brunekreef B. Damp housing and adult respiratory symptoms. Allergy 1992; 47: 498–502.

21

125

Verhoeff A P, van Strien R T, van Wijnen J H, Brunekreef B. Damp housing and childhood respiratory symptoms: the role of sensitization to dust mites and molds. Am J Epidemiol 1995; 141: 103–10. 22 Dales R E, Zwanenburg H, Burnett R, Franklin C A. Respiratory health effects of home dampness and molds among Canadian children. Am J Epidemiol 1991; 134: 196–203. 23 Brunekreef B, van Strien R, Verhoeff A P, van Wijnen J H. Responder bias in assessment of home dampness. Am Rev Respir Dis 1992; 145 (Suppl.): A534. 24 Dales R E, Schweitzer I, Bartlett S, Raizenne M, Burnett R. Indoor air quality and health: reproducibility of respiratory symptoms and reported home dampness and molds using a self-administered questionnaire. Indoor Air 1994; 4: 2–7. 25 Malloch D W. Moulds: Their Isolation, Cultivation and Identification. Toronto: University of Toronto Press, 1981. 26 Newell S Y. Estimating fungal biomass and productivity in decomposing litter. In: Carroll G C, Wicklow D T (eds). The Fungal Community. Its Organization and Role in the Ecosystem. 2nd edn. New York: Marcel Dekker Inc., 1992, pp. 521–61. 27 Newell S Y, Arsuffi T L, Falleon R D. Fundamental procedures for determining ergosterol content of decaying plant material by liquid chromatography. Appl Environ Microbiol 1988; 54: 1876–79. 28 West A W, Grant W D, Sparling G P. Use of ergosterol, diaminopimelic acid and glucosamine contents of soils to monitor changes in microbial populations. Soil Biol Biochem 1987; 19: 607–12. 29 Miller J D, Young J C. The use of ergosterol to measure exposure to fungal propagules in indoor air. J Indust Hyg Assoc (In Press), 1997. 30 Shapiro S S, Wilk M B. An analysis of variance test for normality (complete samples). Biometrika 1965; 52: 591–611. 31 Verhoeff A P, van Wijnen J H, van Reenen-Hoekstra E S, Samson R A, van Strien R T, Brunekreef B. Fungal propagules in house dust. II. Relation with residential characteristics and respiratory symptoms. Allergy 1994; 49: 540–47. 32 Verhoeff A P, van Reenen-Hoekstra E S, Samson R A, van Strien R T, Brunekreef B, van Wijnen J H. Fungal propagules in house dust. I. Comparison of analytic methods and their value as estimators of potential exposure. Allergy 1994; 49: 533–39.

(Revised version received July 1996)