Do Indoor Molds in Nonindustrial Environments Threaten Workers' Health? A Review of the Epidemiologic Evidence

Received for publication February 21, 2000; accepted for publication November 21, 2001. Since the early 1980s, mucosal, skin, and respiratory symptoms, in addition to general symptoms such as fatigue and headache, have been related to the indoor climate of nonindustrial workplaces. Mold growth has been suggested as a causal factor because the health complaints frequently have been related to indicators of microbial contamination: visible signs of mold growth, moisture, and water damage (1). Molds refer to growing colonies of different species of fungi. Fungi are nonphotosynthetic plant bodies that are ubiquitous in nature and decompose organic material. The species differ in size, but most are about 10 µm in diameter. Fungi are able to grow at a relative humidity of between 75 percent and 95 percent at normal room temperature. Fungi reproduce by spores that spread by air, depending on climatic factors, activity in the surrounding environment, and physiologic properties of the individual species (2, 3). Growing fungi may produce metabolites to protect a nutrient source from bacteria. Mycotoxins are metabolites that are able to initiate a toxic response in vertebrates when ingested, inhaled, or otherwise absorbed. Mold antigens may induce an immunoglobulin E–mediated response (4, 5). β-(1,3)-d-glucan is a polyglucose structure of mold cell walls that can induce inflammatory reactions through a specific receptor (6, 7). Ergosterol is the primary membrane sterol of filamentous fungi, and extracellular polysaccharides are stable carbohydrates secreted during fungal growth. At present, there is no evidence for a pathogenic role of ergosterol or extracellular polysaccharides in allergic or inflammatory reactions to fungal components (8). High-level exposure to airborne mold spores may cause allergic alveolitis (e.g., farmer’s lung) (9). Aerosols from air humidifiers heavily contaminated by bacteria, algae, or molds may cause inhalation fever (humidifier fever) (10). Single cases of occupational asthma have been attributed to mold exposure (11, 12), and asthma severity has been associated with outdoor mold spore levels in children and adolescents (13). Mold-contaminated food and feed have been held responsible for serious cases of poisoning in humans and livestock (14). It was discussed recently whether mycotoxins may be responsible for cases of acute idiopathic pulmonary hemorrhage in infants (15, 16). Among employees of nonindustrial workplaces with visible signs of mold growth, moisture, or water damage, there has been an increasing concern about possible health effects. This review evaluates this concern based on present epidemiologic literature. Recently, the topic was partly reviewed by Husman, Verhoeff and Burge, and Peat and Dickerson (17–19). However, they focused mainly on the health effects in children, which may not be relevant for the adult working population (20). Furthermore, several studies assessing mold exposure in the sick building syndrome were included to only a limited extent. This is a systematic review of reports on health effects related to mold growth (or indicators of mold growth) in nonindustrial work sites. Since comparable mold exposure is seen in dwellings, studies that related adult health effects to exposure at home were also reviewed (21). To assess the possible health effects of indoor molds, characterization of the presence and the degree of exposure is crucial. Use of questionnaire information on visible signs or the characteristic smell of molds or dampness is a simple method, but it gives little information about the degree of exposure and involves a risk of reporting bias (22). Inspection by surveyors can provide a standardized assessment of mold growth in terms of the area affected (23). Settlement plates, dust samples, and air samples can be used to determine viable mold spores. However, these methods have limitations concerning determination of concentrations of viable mold spores in the air representative of time-weighted exposure (24). Settlement plates have failed to adequately collect small spores and have shown little reproducibility, with average coefficients of variance of about 30 percent regarding the number of colony-forming units (CFUs) (2, 25). Dust sampling has resulted in coefficients of variance of between 11 percent and 27 percent (CFU/g dust), and the mean agreement for species isolated has varied from 37 percent to 60 percent (26). Comparable figures have been reported for air sampling (25). Viable mold counts may represent only a minor fraction of the total number of fungal particles in indoor air because of different survival rates and difficulties in detecting them on growth agar medium. Nonviable and viable spores can contain allergens and mycotoxins, and it may be of importance also to assess the total number of spores (viable and nonviable), even if a high correlation has been found between viable and total mold counts (27). Levels of β-(1,3)-d-glucan, ergosterol, and extracellular polysaccharides may be other appropriate measures of fungal biomass, even if only measures of extracellular polysaccharides provide information on the fungal species (8). Coefficients of variance of 25 percent and 16 percent have been reported for measurements of β-(1,3)-d-glucan and extracellular polysaccharides, respectively, while relative standard deviations for determining ergosterol have ranged from 5 percent to 27 percent (8, 28, 29). Total airborne mold counts (viable and nonviable) of 667–570,000 spores/m3 have been reported in dwellings (30, 31). Table 1 presents viable mold exposure levels expressed as CFU per cubic meter (only limited data exist for the other parameters) and shows considerable variation, ranging from 0 to 450,000 CFU/m3, between studies. Furthermore, counts differ by season, and there is a tendency toward higher counts at lower latitudes (in Asia). Generally, viable spore counts are higher outdoors than indoors. Significantly higher counts were reported in buildings with signs of mold growth or dampness than in reference buildings (32–34); however, no or small differences in mold levels have also been shown (21, 23, 25, 30, 35–37). During remediation of moldy buildings (homes or nonindustrial workplaces), mold exposure levels may increase significantly (38). The predominant mold genera are Penicillium, Cladosporium, and Aspergillus indoors as well as outdoors in all parts of the world. Fungi such as Trichoderma and Stachybotrys, which require high water activity (aw >0.90–0.95) for growth, are infrequent, but high exposure levels may be seen (34). Stachybotrys has been identified more frequently in buildings with mold problems (32, 39, 40). MEDLINE (National Institutes of Health, Bethesda, Maryland) (1968–June 2000) and NIOSHTIC (National Institute for Occupational Safety and Health Technical Information Center; database provided by SilverPlatter Information Ltd., London, United Kingdom) (1977–June 2000) were searched to identify human peer-reviewed studies published in English that related adult health effects to mold exposure in nonindustrial indoor environments. Additional snowball searches were conducted of the bibliographies in the original papers identified initially. A total of 47 articles were identified. The articles were initially classified as 1) case or cluster reports and 2) analytical studies in populations with no concerns about possible mold-related health effects. The analytical studies were then scored according to six quality parameters, as follows (score value in parentheses): 1) design: cross-sectional (0), longitudinal (1); 2) participation rate: <80 percent (0), ≥80 percent (1); 3) exposure assessment: qualitative (0), quantitative, nonindividual measurements (each measurement representing more than one study subject) (1), quantitative, individual measurements (2); 4) health outcome: self-reported symptoms or weakly defined health outcomes (0), self-reported physician-diagnosed disease or well-defined symptoms (1), physician-diagnosed disease or objective findings (2); 5) potential confounders controlled for: no or limited confounder control (0), comprehensive confounder control that included smoking (1); and 6) exposure-response assessment: no (0), yes (1). Finally, each study was assigned a total sum of scores. Weighted average odds ratios with 95 percent confidence intervals were computed when appropriate (41). The odds ratio values from each study were weighted by the inverse variance extracted from the confidence intervals, and 95 percent confidence intervals for the meta-odds ratios were computed from the sum of the weights. A total of 19 articles were identified that described patient cases with health effects attributed to mold contamination of the indoor environment or cross-sectional studies conducted in populations because of suspected health effects related to mold exposure in those populations. Three case reports described a teacher, an office worker, and a married couple, all diagnosed with allergic alveolitis attributed to indoor mold exposure at work or at home (42–44). High levels of viable airborne molds (approximately 5 × 103 CFU/m3) dominated by species from a contaminated were in the of the office exposure levels were for the married couple, but they species in home Cladosporium, Penicillium, and were shown in the teacher, but no relevant data were A case of organic dust toxic syndrome in a moldy was related with exposure to airborne molds viable molds and total A cluster of cases of disease in was related to levels in the work environment, but not with fungal levels of and symptoms, headache, and have been related to mold exposure symptoms were of and Stachybotrys from the of cases of allergic fungal mold species were identified in the indoor air as isolated in from the and of the was in employees of a that significantly exposure levels of β-(1,3)-d-glucan, but this was not by a study of between β-(1,3)-d-glucan exposure and inflammatory were Among office with symptoms to be related to the indoor environment, percent immunoglobulin to one or more of a of molds, percent immunoglobulin to but immunoglobulin to Penicillium, or Aspergillus however, these findings were to mold measures to and Trichoderma was in an in 25 with symptoms of sick building syndrome A number of in employees of an office contaminated with was but not in the employees with the exposure of and symptoms and of general symptoms were reported and of buildings with signs of indoor mold growth and viable mold levels of about or higher as well as than or they were related to β-(1,3)-d-glucan A total of articles based on analytical studies not by concerns about mold-related health effects (or health effects related to other of a sick in the study populations were identified. studies focused on occupational and samples of the general population for health problems related to at home 3). for longitudinal studies that in asthma and in all studies used a cross-sectional sampling and on reports or case the six studies The participation was high and varied between percent and percent The occupational exposure included and information frequently was based on the study reports of mold growth or of the this assessment 5 percent and 75 percent of the study or in with signs of mold growth or levels of viable molds were in air and dust samples were for each study on or at home studies of a limited number of measurements frequently exposure to several samples and samples however, individual sampling was also used study used sampling while the on varied from to The average of total viable molds ranged from to when in air and between and when in studies also reported total airborne mold counts (viable and nonviable) ranging from × to × β-(1,3)-d-glucan was in a study the presence of specific mold species in air or on building of asthma were defined by a or as physician-diagnosed asthma reported by the study A in with reference to mold exposure in with asthma was the in one study while or by in in 1 according to symptoms were defined by questionnaire or were based on of or as one of several symptoms for The of symptoms in the United 10 percent reported while percent of in reported of symptoms such as or or self-reported were included the health effects percent reported or during a in percent reported or during 1 of and in and measures of by were used in studies factor and cell counts in were in study occupational studies and of the studies on home information on a of and symptoms and on different general the studies more than different symptoms were different studies reported different symptoms for the symptoms were and headache, which affected more than percent of the was controlled for in several studies and dust in the indoor environment in were by at number of of or However, in confounder control was limited (e.g., no control for Generally, studies that mold levels also possible exposure-response by or by correlation however, this was not used by all or were conducted on a building and not an individual studies on qualitative exposure data classified exposure in high and from the of mold growth or dampness these were not by mold The quality are for each study in and that used a of self-reported and objective exposure or information were a quality to this A of was studies were scored with the of six while the was cases of as well as self-reported cases of physician-diagnosed were associated with signs of mold growth or dampness average odds ratio 95 percent confidence This was also the case for symptoms average odds ratio 95 percent differences were found if exposure characterization was based on or reported no with symptoms but not present the a with in the of of or because of in the However, if these of symptoms were no were for the increasing asthma severity by of visible mold growth as by or no with signs of mold or dampness in the dwellings of or symptoms were to counts of viable molds in air or dust reported of the and a was not The study that the of total airborne molds (viable and nonviable) no with or symptoms symptoms or or self-reported was associated with signs of mold growth or dampness in several studies but not all 5). A weighted average odds ratio of was computed percent Levels of viable molds in air or dust were associated with symptoms in studies but this was not the case in the other 5). was reported between total mold counts and symptoms of or 5). and no between β-(1,3)-d-glucan and symptoms 5). of of was related to signs of but a higher in was in employees of buildings with buildings however, this was not found for or levels of total molds (viable and nonviable) was related to expressed as the sum of the of the from the but not if from 0 to or if health was as the cross-sectional of the or as or concentrations of and during and to levels in of a with visible mold growth, but this was not seen for factor and cell were found between signs of mold growth or dampness and symptoms or symptoms or symptoms or and general symptoms fever and and 5). Weighted average odds ratio values of percent percent percent and percent respectively, were However, no were seen for case based on of or symptoms, lower respiratory symptoms, or case in with sick building of symptoms the or or of general symptoms by levels of viable molds in air or dust was 5). The study that total mold levels reported no with or lower respiratory symptoms study total mold levels but no 5). A was found between β-(1,3)-d-glucan and or fatigue but not for symptoms the or 5). in odds ratios (or other measures of was found for symptoms, or or symptoms or for general symptoms by study quality in the exposure of mold growth or viable molds, total molds, or were of an toward or no by increasing quality if the exposure were However, this was by studies that not exposure of symptoms was reported in office from a building in which was ratio 95 percent but symptoms were to Cladosporium, and species The presence of Aspergillus species 30 CFU/m3) was related to the presence of species was associated with and no was found for levels of Aspergillus species with and but this was not the case for levels of and species or in in pulmonary complaints related to levels of Penicillium, and Aspergillus species was reported in study A for species was associated with respiratory symptoms and growth of species at the but no such was suggested for and species However, specific immunoglobulin to was not more in asthma with signs of mold growth at home than in asthma such mold growth, but this was for mold levels or signs of indoor mold growth were to for molds Total immunoglobulin and were related to mold levels of not an between signs of mold growth and lower respiratory symptoms, which be if is of causal importance The analytical studies reviewed provide evidence that signs of mold growth or dampness in nonindustrial workplaces or dwellings are associated with asthma as diagnosed by a or on the of relevant symptoms and with and or However, no were found for objective measures of or parameters, but the studies were children, asthma also has been related to signs of mold growth or dampness The studies provide no evidence that measures of total or viable indoor molds, β-(1,3)-d-glucan, or specific mold species are associated with asthma symptoms, or findings are in with the that the of mold exposure on asthma in is even if cases of occupational mold asthma have been reported (11, and or objective were not associated with specific measures of fungal The in these may that the measures of mold exposure were that levels in the studies were or that these studies were because of The also to the studies on qualitative signs of mold growth or these one at a the studies different measures of molds in air or dust were total viable mold total mold counts (viable and nonviable), β-(1,3)-d-glucan and specific mold species or and dust are the relevant to or of is not to a and no in was when exposure assessment was based on air or dust However, one may the of viable mold counts as the of exposure higher levels are not in buildings by mold growth or water damage with buildings or outdoor air (21, 23, 25, 30, 35–37). This may also to total mold however, for this data are β-(1,3)-d-glucan levels have been associated with signs of mold growth in the indoor environment it may be that the measures of exposure not provide a characterization of mold the other they all the of molds from and in air or dust and are to to human health effects if a causal of indoor molds is if molds are causal for the health effects reported by working or in buildings with signs of mold growth or there are also other for dust dust to molds, a environment and are of allergic Furthermore, dust were related to symptoms in the study by (in to mold but not in the study by and of particles at of the symptoms, exposure the severity of asthma in allergic to dust however, the studies reviewed not provide evidence for this The of between levels of mold exposure and health effects may be to a limited number of mold measurements were used to exposure levels for several in buildings or sampling was This may have mold levels (25). there is a for mold-related health it may be that the populations in the studies were this However, no were seen in the studies reviewed with the average exposure levels It is that with or respiratory symptoms to in dwellings or workplaces that have mold and in cross-sectional which was the of all but of the studies a possible between mold exposure and health effects may be However, it is that such bias studies exposure according to signs of mold growth more than those on this can the in the of the studies It is also with to other of study because no systematic in quality confounder was seen for the studies with those on qualitative exposure in also be seen if lower exposure levels were found with the qualitative studies. However, this was not by the studies by and that the population by qualitative and exposure information The of the reviewed studies on of exposure is the risk of information this by that between self-reported home and symptoms when symptoms that be by indoor were from the have provided evidence of a of bias in studies of health effects to indoor climate but not all have of indoor mold growth may to this but not in studies with symptoms as the health symptoms, by are and mold growth visible to surveyors is also to be visible to study The studies that on of exposure comparable to those on A high of the studies reviewed were because of concerns about mold-related health effects in the study populations. studies may However, they not a for assessing causal effects even if an reference population is included in this these studies were than the analytical studies. the other these studies may provide information from exposure in systematic studies. The reports by and of cases of organic dust toxic syndrome as toxic or inhalation fever) and allergic alveolitis airborne mold exposure × 103 and CFU/m3, are in with well and in cases of exposure to humidifiers contaminated by levels of molds in indoor cause the respiratory and health effects immunoglobulin E–mediated is a possible However, no for this was provided by the studies that included information on mold health and or immunoglobulin inflammatory reactions and fatigue were the only symptoms that an with However, this was and was reported in only a Mycotoxins have been in buildings in which have about a sick However, reported health effects from this exposure have not been from other possible in the indoor environment and of the systematic studies reviewed that the of evidence that exposure to high mold levels may cause allergic alveolitis and inhalation and this may also to buildings by and mold growth. However, the studies reviewed provide no evidence that increasing levels of viable mold exposure in nonindustrial work or dwellings are related to an increasing of asthma or to and or in the adult which may be to or measures of mold were that related total fungal to health however, the studies were if the epidemiologic database is to whether mold exposure in nonindustrial has health effects on that workplaces be well and be and or to mold exposure in nonindustrial workplaces. However, these health problems more are reported by working or in buildings that have signs of indoor mold growth and To possible causal studies on populations mold exposure than those in nonindustrial environments. measures are to buildings that have signs of mold growth and may to molds and such as mycotoxins, but other also be as well as more appropriate exposure assessment objective methods to determine such as and be in to study exposure-response and to reporting bias in this area of The exposure also to the for control for confounders and the buildings to of Occupational Viable mold exposure levels in indoor and outdoor air, in buildings with and signs of mold growth, and by from exposure and epidemiologic colony-forming Viable mold exposure levels in indoor and outdoor air, in buildings with and signs of mold growth, and by from exposure and epidemiologic colony-forming studies of health effects related to mold exposure in nonindustrial the study were classified with different quality the and values are colony-forming sick building measurement of the of the the studies of health effects related to mold exposure in nonindustrial the study were classified with different quality the and values are colony-forming sick building measurement of the of the the studies of health effects related to mold growth at the study were classified with different quality the and values are colony-forming immunoglobulin studies of health effects related to mold growth at the study were classified with different quality the and values are colony-forming immunoglobulin between mold exposure in nonindustrial workplaces or dwellings and physician-diagnosed symptoms, and in of epidemiologic ratios confidence otherwise odds ratio from the original in 1 for mold levels for the study by which reported dust value in between mold exposure in nonindustrial workplaces or dwellings and physician-diagnosed symptoms, and in of epidemiologic ratios confidence otherwise odds ratio from the original in 1 for mold levels for the study by which reported dust value in between mold exposure in nonindustrial workplaces or dwellings and and symptoms, general symptoms, or of epidemiologic studies reporting several symptoms for each the most were as of one or more symptoms headache, of or or or or at while in the not in the odds ratio from the original mold levels in the study by that reported dust levels and and that reported air and dust as at one general and at one and one between mold exposure in nonindustrial workplaces or dwellings and and symptoms, general symptoms, or of epidemiologic studies reporting several symptoms for each the most were as of one or more symptoms headache, of or or or or at while in the not in the odds ratio from the original mold levels in the study by that reported dust levels and and that reported air and dust as at one general and at one and one 1 of Occupational of Occupational of of Institute of and Occupational