Respiratory and Allergic Health Effects of Dampness, Mold, and Dampness-Related Agents: A Review of the Epidemiologic Evidence

Objectives Many studies have shown consistent associations between evident indoor dampness or mold and respiratory or allergic health effects, but causal links remain unclear. Findings on measured microbiologic factors have received little review. We conducted an updated, comprehensive review on these topics. Data sources We reviewed eligible peer-reviewed epidemiologic studies or quantitative meta-analyses, up to late 2009, on dampness, mold, or other microbiologic agents and respiratory or allergic effects. Data extraction We evaluated evidence for causation or association between qualitative/subjective assessments of dampness or mold (considered together) and specific health outcomes. We separately considered evidence for associations between specific quantitative measurements of microbiologic factors and each health outcome. Data synthesis Evidence from epidemiologic studies and meta-analyses showed indoor dampness or mold to be associated consistently with increased asthma development and exacerbation, current and ever diagnosis of asthma, dyspnea, wheeze, cough, respiratory infections, bronchitis, allergic rhinitis, eczema, and upper respiratory tract symptoms. Associations were found in allergic and nonallergic individuals. Evidence strongly suggested causation of asthma exacerbation in children. Suggestive evidence was available for only a few specific measured microbiologic factors and was in part equivocal, suggesting both adverse and protective associations with health. Conclusions Evident dampness or mold had consistent positive associations with multiple allergic and respiratory effects. Measured microbiologic agents in dust had limited suggestive associations, including both positive and negative associations for some agents. Thus, prevention and remediation of indoor dampness and mold are likely to reduce health risks, but current evidence does not support measuring specific indoor microbiologic factors to guide health-protective actions.

volume 119 | number 6 | June 2011 • Environmental Health Perspectives Review Dampness and mold exposures in buildings are common, with estimates ranging from 18% to 50% of buildings (Gunnbjornsdottir et al. 2006;Mudarri and Fisk 2007). A large num ber of studies in many geographical regions have found consistent associations between evi dent indoor dampness or mold and respiratory or allergic health effects in infants, children, and adults [Institute of Medicine (IOM) 2004; World Health Organization (WHO) Europe 2009]. A review by the IOM (2004) reported documented associations, but not documented causal relationships, between indoor damp ness and upper respiratory tract symptoms, cough, wheeze, and asthma symptoms in sensi tized persons, but not for asthma development. A more recent review by WHO up to 2007 expanded the observed associations to include asthma development, current asthma, dysp nea, and respiratory infections (WHO Europe 2009). Associations were found in both atopic and nonatopic individuals. Other published reviews or opinion pieces on this topic are available (e.g., Bornehag et al. 2004;Douwes 2005;Mudarri and Fisk 2007).
The consistent associations between evident dampness or mold and health may represent underlying causal relationships between fungal exposures and health. However, conventional quantitative measurements of fungi or other microbiologic exposures, such as counts of cul turable airborne fungi, have shown less consis tent associations with health effects than have qualitative assessments of visible dampness or water damage, visible mold, or mold odor. Thus, although a causal role for microbiologic exposures is plausible and likely, the evidence for this is still weak (Douwes and Pearce 2003). This is likely attributable in part to the lack of valid exposure assessment methods for the still unknown causal agents, microbial and pos sibly nonmicrobial, that increase with damp ness and directly cause adverse respiratory and allergic effects.
Much additional epidemiologic research on qualitative and quantitative assessments of dampness and dampnessrelated agents has become available in the last few years. The present review combines findings of the IOM review of findings up to 2003 (IOM 2004) and a new assessment of later published stud ies. In this review we provide a) an updated, comprehensive review of available epidemio logic evidence on qualitative assessments of dampness or mold factors, and b) a new synthesis of evidence on quantitative meas ure ments of microbiologic factors. Earlier work on this review (summarizing literature through 2007) was originally done to support the WHO Guidelines for Indoor Air Quality related to dampness and mold (2009).

Methods
The online database PubMed (National Library of Medicine 2010) was searched using three groups of keywords such as dampness, damp, "water damage," moisture, humid ity, fungi, fungus, mold, mould, bacteria, or microorganisms, crossed with health, asthma, allergy, eczema, wheeze, cough, respiratory, "respiratory infection," lung, skin, nasal, nose, "hypersensitivity pneumonitis," alveolitis, bronchial, hypersensitivity, or inflammation and with building, house, home, residence, dwelling, office, school, or "daycare center." A similar search was run in the ISI/Web of Knowledge database (Thomson Reuters 2010). We identified additional publications from reference lists and personal databases. Some indoor exposures/conditions were not included, for example, humidity, mattress moisture, and dust mites.
Inclusion of a primary study required the following characteristics: • Publication in a peer-reviewed journal by November 2009 • Reporting of original data from one of the following study designs: intervention (qua siexperimental intervention), prospective (prospective cohort), retrospective (retro spective cohort or nested case-control), or crosssectional (crosssectional or prevalence case-control) • No minimum study size, but if exposure was characterized only at the building level, inclusion of > 10 buildings • Including risk factors related to dampness or microbiologic organisms/components/ products, other than allergens (dust mites, cockroaches, mice) • Including allergic or respiratory health effects • Providing adequate control, in study design or analysis, of selection bias and confound ing from key variables: sex, smoking (active in adults, passive in children), and socio economic status (SES; control for SES not required if SES shown not to confound in study, if adjusted for race when race highly correlated with SES, if study conducted within specific occupational groups, or if study from Nordic countries or Holland). We gave primary consideration to associa tions between specific health outcomes (e.g., wheeze) and one or more qualitative assess ments of indoor dampness or mold (e.g., vis ible dampness, visible mold, water damage, or mold odor), with the latter grouped for review. We refer to this set of factors collectively as evi dent dampness or mold, qualitatively assessed dampness or mold, or simply dampness or mold. Each study generally reported multi ple findings (for example, four findings from a study reporting estimates for associations between visible dampness and daytime wheeze, visible dampness and nighttime wheeze, mold odor and daytime wheeze, and mold odor and nighttime wheeze). Based on all currently available evidence, including studies reviewed in the IOM report, new studies included in this review, and findings from available quan titative metaanalyses, we drew conclusions about associations between specific health out comes and qualitatively assessed dampness or mold (excluding quantitative assessments of microbiologic factors). In "Results," we gener ally refer to all ratio estimates of effect as odds ratios (ORs), although a few studies used other types of ratio estimates.
In this review we classified strength of evidence using the same categories as the IOM review on dampness and health (IOM 2004) (box ES1, p. 8): sufficient evidence of a causal relationship, sufficient evidence of an association, limited or suggestive evi dence of an association, and inadequate or insufficient evidence to determine whether an association exists. For each relationship considered, we classified the evidence using professional judgment on its persuasiveness, based on reported findings plus the strength, quality, diversity, and number of studies. Findings from quantitative meta analyses were also considered. We placed increasing weight in the review on studies of stronger design. The strongest epidemiologic evidence was considered to come from individually randomized controlled experimental/interven tion trials that added or removed risk factors. Studies considered next strongest were pro spective (cohort), followed by retrospective (cohort or nested case-control), observational studies. We considered crosssectional obser va tional studies (including prevalence casecontrol studies) to provide the weakest evidence included. A set of strongly designed human studies of different designs and in different populations, with findings generally consistent in direction and magnitude, especially if mag nitudes of effect were large or dose-response relations were found, was considered to provide the most persuasive overall evidence.
We drew separate conclusions, more pre liminary because evidence was sparse, about associations between specific health outcomes and specific quantitatively assessed micro biologic factors. For findings on associations between a specific outcome and a specific measured indoor microbiologic factor, our criteria for evidence suggestive of associations required at least 80% consistency of estimates either ≤ 1.0 or > 1.0 (with no minimum change from the null required) among at least five estimates available from three or more studies. This rough tally of findings above or below the null did not consider magnitude of effects, precision, statistical significance, study design, or age of subjects.

IOM review.
The IOM review of epidemio logic evidence to 2003 on dampnessrelated health effects found no demonstrated causal associations (IOM 2004). Sufficient evidence of association was reported for four outcomes (upper respiratory tract symptoms, cough, wheeze, and asthma symptoms in sensitized persons, i.e., asthma exacerbation) for the two kinds of risk factor considered: exposure to damp indoor environments and presence of molds or other agents in damp indoor environ ments. Sufficient evidence of asso ciation was also reported between hypersen sitivity pneumonitis in susceptible persons and mold or other agents in damp environ ments. The 45 studies included in the IOM review are summarized in Supplemental Material, Tables A1.1-A1.6 (doi:10.1289/ ehp.1002410). Table 1  as ORs and 95% confidence intervals (CIs), but occasionally was reported as other ratio estimates such as relative risks (RRs) or inci dence rate ratios (IRRs) or as linear regression coefficients or proportions.
Considering all current evidence, most published findings involved qualitative assess ments of dampness or mold, including visible water damage, visible moisture, dampness, leaks, flooding, visible condensation on win dows, visible mold or mildew, and moldy or musty odor. Fewer findings were available on quantitatively measured microbiologic factors, including specific or total culturable fungi or bacteria; microscopically enumerated, noncul tured fungi or bacteria; ergosterol (a structural component of fungi, used as a marker for total fungal biomass); extracellular polysaccharides (produced by fungi and used as a marker for specific fungal groups); (1→3)βdglucans (a cell wall compound with immunomodu lating properties found in fungi but also in some bacteria and pollens); endotoxin or lipopolysaccharide (a cellwall compound of Gramnegative bacteria with proinflamma tory properties, associated with dampness but also with many other sources); and markers of endotoxin such as 3 hydroxyl fatty acids.
Meta-analyses. Three available quantitative metaanalyses combined multiple qualitative dampness or mold factors into a single set of dampnessrelated risk factors. Findings, sum marized in Table 2, are described for specific outcomes below. Two metaanalyses using the same methods estimated summary ORs and 95% CIs for associations of dampness or mold in residences with respiratory effects: upper respiratory tract symptoms, cough, wheeze, asthma development, current asthma, and everdiagnosed asthma , and respiratory infections and bronchitis (Fisk et al. 2010). Antova et al. (2008) estimated summary ORs for dampnessrelated factors and everdiagnosed asthma, bronchitis, allergic sensitization, hay fever, cough, and wheeze.
Results for qualitative dampness or mold. We considered no health outcomes to have sufficient evidence to document a causal rela tionship with indoor dampness or mold.
We considered four health outcomes to have sufficient evidence for association with indoor dampness or mold that were already so classified in the IOM review: asthma exacer bation, cough, wheeze, and upper respiratory tract symptoms (Table 3).
For asthma exacerbation and dampness or mold, we consider current evidence sufficient to document association and strongly sugges tive of causality. Among 31 currently available studies [see Supplemental Material, Tables A1.2 and A2.2 (doi:10.1289/ehp.1002410)], quali tative dampnessrelated factors were consis tently associated with asthma exacerbation, with ORs consistently exceeding 1.0 in both volume 119 | number 6 | June 2011 • Environmental Health Perspectives adults [100% of findings in retrospective stud ies (ORs from 1.7 to 2.6) and 100% of find ings in crosssectional studies (ORs from 1.02 to 4.2)] and in children [100% of findings in intervention studies (protective associations, not reported as ORs), 100% of findings in prospective studies (ORs from 3.8 to 7.6), 100% of findings in retrospective studies (ORs from 1.5 to 4.9), and 95% of findings in cross sectional studies (ORs from 1.0 to 7.6)]. Most notably, Kercsmar et al. (2006) conducted a controlled experimental intervention study on asthma exacerbation in houses of highly symp tomatic asthmatic children. Comprehensive removal of dampness sources and visible mold caused dramatic reductions in asthma exacer bations. Acute care visits at 6-12 months after intervention were 90% fewer in those remedi ated versus controls (p = 0.003). This study (although of necessity unblinded), because of the implausibility of noncausal explanations for the findings and in conjunction with other available studies, strongly suggests a causal asso ciation between indoor dampness or mold and exacerbations in children with asthma.
We classified eight health outcomes as having sufficient evidence for association with indoor dampness or mold that were not so classified or not evaluated in the IOM review: asthma development, current asthma, asthma ever, dyspnea, respiratory infections, bronchi tis, allergic rhinitis, and eczema (Table 3).
Asthma development is a health out come of special public health importance. Five studies included in the IOM report [of the eight listed there for asthma develop ment; see Supplemental Material, Table A1.1 (doi:10.1289/ehp.1002410)] explicitly exam ined associations between dampness or mold and asthma development (Jaakkola et al. 2002;Nafstad et al. 1998;Oie et al. 1999;Thorn et al. 2001;Yang et al. 1998). Eight new studies were identified (Supplemental Material, Table A2.1) (CoxGanser et al. 2009;Gunnbjornsdottir et al. 2006;Hyvarinen et al. 2006;Iossifova et al. 2009;Jaakkola et al. 2005;Matheson et al. 2005;Park et al. 2008;Pekkanen et al. 2007). Among all currently available studies (five studies in Supplemental Material, Table A1.1; all studies in Supplemental Material, Table A2.1), 78% of findings exceeded 1.0. In retrospective casecontrol studies of adults, 60% of ORs exceeded 1.0 (range, 0.8-2.2). Among children, 80% of prospective or retrospective/case-control ORs exceeded 1.0 (range, 0.6 to 4.1). The three studies in infants (Iossifova et al. 2009;Nafstad et al. 1998;Oie et al. 1999) reported ORs all exceeding 1.0 (range, 1.7-7.1); however, as asthma cannot be reliably assessed in infants, these findings should be interpreted with cau tion. Infant studies were excluded from the metaanalysis of Fisk et al. (2007), which reported a summary OR (95% CI) of 1.3 (0.9-2.1) for asthma development and damp ness factors. One of the strongest reported studies, by Pekkanen et al. (2007), showed in an incident case-control study of asthma cases that dampness or mold in the main living area of houses was related in a dose-response fashion to asthma development in infants and children. Multivariateadjusted ORs (95% CIs) for asthma incidence, for baseline and two increasing levels of maximum severity of moisture damage (assessed by civil engineers), were 1.0, 2.8 (1.4-5.4), and 4.0 (1.6-10.2). This welldesigned study provides the strongest evidence (e.g., incident case-control, large and statistically significant effects, dose-response relation, unbiased exposure assessment), within a body of generally consistent other findings, that dampnessrelated exposures may cause asthma development in infants and children.
For dyspnea, considered in the IOM report to have limited or suggestive evidence of association with dampness, the number of available studies for adults and children has increased from 4 to 16, all crosssectional except 1 controlled intervention study. The intervention study found significant improve ments in perceived breathing after mold removal, fungicide application, and ventilation increase (Burr et al. 2007). Among the 14 other available studies of dampness or mold [Supplemental Material, Tables A1.3 and A2.5 (doi:10.1289/ehp.1002410)], meas ures of association for dampness or mold with dys pnea were predominantly (84%) > 1.0, with ORs ranging from 0.7 to 9.4 in adults and from 0.4 to 2.3 in children.
We included findings on current asthma when defined as either asthma diagnosis in prior 12 months, asthma diagnosis ever plus asthmatic symptoms in prior 12 months, or recent prescription of asthma medication. Current asthma, not specifically evaluated in the IOM review, was consistently associated in available studies [Supplemental Material, Table A2.4 (doi:10.1289/ehp.1002410)] with dampness or mold. In these crosssectional studies of adults, children, or both, almost all ORs (94%) exceeded 1.0 (ranging from 0.3 to 13.0). Fisk et al. (2007) reported, as a sum mary effect estimate, an OR (95% CI) of 1.6 (1.3-1.9) for current asthma and qualitative dampness factors.
Everdiagnosis with asthma [Supplemental Material, Table A2.3 (doi:10.1289/ ehp.1002410)] was associated consistently with dampness or mold (91% of ORs; range, 0.6-2.6) in both adults and children. Most studies were crosssectional and in children. All studies in adults and the single prospective study in children found completely consis tent positive associations. Both available meta analyses found increased summary ORs for everasthma diagnosis and residential damp ness or mold, with ORs (95% CIs) of 1.37 (1.23-1.53) for children and adults in Fisk et al. (2007) and 1.35 (1.20-1.51) for children in Antova et al. (2008).
Dampness or mold was associated consis tently with allergic rhinitis (92% of findings, all in children), with ORs ranging from 0.7 to 3.5 [Supplemental Material, Table A2.14 (doi:10.1289/ehp.1002410)]. We included only allergic rhinitis outcomes defined as either medically diagnosed allergic rhinitis or the combination of rhinitis symptoms with documented atopy. The strongest single study (prospective) found dose-response increases in allergic rhinitis associated with visible mold, with ORs to 3.2 (Biagini et al. 2006).
Dampness or mold was associated consis tently with eczema [Supplemental Material, Table A2.12 (doi:10.1289/ehp.1002410)], with 89% of ORs > 1.0 (range 0.2 to 2.9). The strongest study, a prospective study in children, found consistently increased ORs up to 2.9 for prenatal mold exposure to infants with no parental atopic history.
Other outcomes evaluated here but not in the prior IOM report include common cold, allergy/atopy, and altered lung func tion (Table 1). Common cold [Supplemental Material, Table A2.11 (doi:10.1289/ ehp.1002410)] was positively associated with dampness or mold in 71% of reported find ings. However, the methodologically strongest single study, a prospective study in children, found only 4 of 9 estimates elevated, with ORs ranging from 0.6 to 1.8. Therefore, we consider this association only suggestive.
Increase in allergy/atopy (excluding allergic rhinitis and eczema) [Supplemental Material, Table A2.13 (doi:10.1289/ ehp.1002410)] in association with dampness or mold was found in 77% of reported assess ments in the available studies; ORs ranged from 0.6 to 2.4. Findings in the strongest studies, two prospective studies in children, were overall somewhat inconsistent, as were the other studies. This association is made more plausible by the increased summary ORs in the metaanalysis by Antova et al. (2008) for sensitivity to inhaled antigens and for hay fever, as well as by the consistent association found in this review between dampness or mold and both allergic rhinitis and eczema. However, the overall evidence linking allergy/ atopy and dampness or mold was inconsis tent enough that we currently consider it only (strongly) suggestive.
The evidence associating altered lung func tion with dampness or mold [Supplemental Material, Table A2.8 (doi:10.1289/ehp. 1002410)] was considered too inconsistent to draw conclusions. No eligible epidemio logic studies were found on hypersensitivity pneumonitis and dampness or mold (but see "Discussion" regarding overall evidence).

Results for measured microbiologic factors.
Findings on health risks associated with quantitatively assessed microbiologic factors were sparse across specific health outcomes and 53 specific types of microbial measure ments. Suggestive associations (as defined in "Methods" for conclusions about quantitatively assessed microbiologic factors: requiring at least 80% consistency of estimates either ≤ 1.0 or > 1.0, among at least five estimates available from three or more studies) were not seen for measurements in air but were apparent for some measurements in dust (Table 4). Higher concentrations of ergosterol in dust were associated with increases in current asthma. Higher concentrations of endotoxin in dust were associated with increases in wheeze. For (1→3)βdglucan in dust, although medium concentrations were associated with increases in wheeze, the highest concentrations were associ ated with decreases in wheeze. We consider these associations with quantitative microbio logic assessments to be only suggestive. Other microbial measurements used in reviewed stud ies [listed in Supplemental Material,

Discussion
Epidemiologic evidence from primary stud ies and quantitative metaanalyses shows evi dent indoor dampness or mold to be associated consistently with a wide range of respiratory or allergic health effects, including asthma develop ment and exacerbation, current and ever diagnosis of asthma, dyspnea, wheeze, cough, respiratory infections, bronchitis, aller gic rhinitis, eczema, and upper respiratory tract symptoms. In addition to the consistently posi tive associations across many study designs, populations, ages, and health outcomes, doseresponse relations with observed dampness and mold were often reported (e.g., Biagini et al. 2006;Cummings et al. 2008;Park et al. 2004;  Iossifova et al. 2007Iossifova et al. , 2009Park et al. 2001Park et al. , 2006Zhao et al. 2008;Schram-Bijkerk et al. 2005;Bolte et al. 2003;Campo et al. 2006;Douwes et al. 2006;Gehring et al. 2008;Gillespie et al. 2006;Litonjua et al. 2002;Perzanowski et al. 2006 (1→3)-β-d-glucans in dust, medium levels Wheeze 7 of 8 (88%) 3 0.89-6.05 Douwes et al. 2006;Iossifova et al. 2007Iossifova et al. , 2009-β-d-glucans in dust, highest levels Wheeze 10 of 11 (91%) 4 0-1.25 Douwes et al. 2006;Iossifova et al. 2007Iossifova et al. , 2009Schram-Bijkerk et al. 2005 a A suggestive association required, among reported findings on associations between a specific measured indoor microbiologic factor and a specific respiratory or allergic health outcome, at least 80% consistency of estimates either ≤ 1.0 or > 1.0, among at least five estimates available from three or more studies. This assessment did not consider magnitude of effects, precision, statistical significance, study design, or age of subjects. b Measured microbiologic factors with inadequate or insufficient evidence to determine whether an association exists with any specific health outcome are listed in Supplemental Material,  Pekkanen et al. 2007). Although available epi demiologic evidence does not yet establish that indoor dampness or mold causes human health effects, findings from one strong epidemio logic intervention study (Kercsmar et al. 2006), in conjunction with other available studies, strongly suggest causation of asthma exacerba tion in children by dampness or mold. Several studies provide evidence for temporal asso ciation of dampness/mold and health effects by demonstrating increased incidence density of new asthma diagnosis among occupants of waterdamaged buildings compared with periods before water damage (CoxGanser et al. 2005;Laney et al. 2009). It is well accepted that hypersensitiv ity pneumonitis (HP), a granulomatous, cell mediated lung inflammation, is caused by inha lation of antigens from microorganisms or other sources, although causal exposures often can not be determined (Fink et al. 2005). Current knowledge is based on outbreak investigations and limited epidemiology, mostly in indus trial and agricultural settings, but also in office buildings (CoxGanser et al. 2005;Kreiss 1989;Park et al. 2004) and, in both adults and chil dren, in homes (Venkatesh and Wild 2005). One specific dampnessrelated mold exposure (Trichosporon cutaneum) is documented to cause HP in homes (Ando et al. 1995 Few studies included objective, replicable assessments of dampness. Both Karvonen et al. (2009) andPark et al. (2004), using scales combining area of water damage or area of water stains with subjective assessments, found exposure-response relations with multiple health outcomes. Williamson et al. (1997), using a scale based only on moisture meter readings from walls, also found positive adjusted associations, for example, ORs (95% CIs) for asthma and any dampness of 3.03 (1.65-5.57), exceeding ORs for subjective inspectordetermined visible mold. Williamson et al. (1997) also found positive correlations between total moisture meter dampness score and both asthma severity (p = 0.0006) and predicted FEV 1 (forced expiratory volume in 1 sec) (p = 0.006). One potential advantage of quantitative dampness measurements as indi cators of exposure, relative to specific quan titative microbial measurements, is that they can be proxies for various dampnessrelated causal agents, whether microbial or chemical. Quantifying visible mold may also prove use ful; however, Dales et al. (2010) found no sig nificant relationship between measured area of visible mold and respiratory health outcomes.
Although evidence is limited that links any quantitative microbial measurements to specific health effects, in this review we have identified some preliminary associations (Table 4), all for measurements in dust: increased ergosterol with increased current asthma; increased endo toxin with increased wheeze; and for (1→3) βdglucans, medium concentrations with increased wheeze but the highest concentra tions with decreased wheeze. We consider these associations to be only suggestive, because of the limited number of studies, the limited fac tors considered in summarizing them, and the demonstrated complexity of some of these rela tionships, such as for endotoxin and (1→3)β dglucans, each associated in multiple studies with both adverse and protective associations (Douwes et al. 2004(Douwes et al. , 2006. Current findings thus cannot define causal microbiologic exposures or dose-response relations sufficiently to define safe levels of exposure to dampnessrelated agents. At pres ent, subjectively assessed dampness or mold has the most consistently documented asso ciations with respiratory and allergic disease. Quantifying dampness objectively has shown promise (Karvonen et al. 2009;Park et al. 2004;Williamson et al. 1997), but findings are few. For quantifying microbiologic factors, concentrations of culturable airborne organ isms have fared poorly in empirical health research. Some assessments in dust, such as ergosterol as an indicator of total biomass of fungi, are more promising; others, such as endotoxin and glucans, have relationships with health too complex for simple interpreta tion. Polymerase chain reaction (PCR) assays for specific fungi in dust also have promise, but no studies using PCR met inclusion criteria for this review, and a standard scale now used to group PCR findings across fungi seems premature (e.g., Vesper et al. 2007 Difficulties in finding clear relationships with measured microbiologic exposures may be attributable to measurement errors in expo sure assessment, including measurement of noncausal factors; to effects that change with intensity and duration of exposure or age at exposure; or to interaction effects occurring with multiple exposures. Endotoxin, tradi tionally associated with nondampnessrelated exposures such as farm animals and pets and with potential protection against atopy, has now been shown to be associated in water damaged office buildings with observed damp ness, fungal spores, and increased building related asthmatic symptoms (Park et al. 2006;Rao et al. 2005). Adverse effects from endo toxin may be increased by other dampness associated agents and vice versa (Park et al. 2006). In addition, moisture in buildings can increase nonbiologic emissions not measured in most dampness research, including form aldehyde [associated with increased asthma (McGwin et al. 2009;Mendell 2007)] from composite wood products (Matthews et al. 1986) and 2ethyl1hexanol from moisture related degradation of plasticizer in vinyl floor ing (Norbäck et al. 2000).
Based on available evidence, dampness and mold may have enormous health and social costs worldwide. A northern European study found an 18% prevalence of indoor damp ness (Gunnbjornsdottir et al. 2006). The IOM review (IOM 2004), using European and North American data, estimated that at least 20% of buildings had problems with dampness. Mudarri and Fisk (2007) estimated a 50% prevalence of dampness or fungi in U.S. houses. Fisk et al. (2007) concluded that "building dampness and mold are associated with approximately 30-50% increases in a variety of respiratory and asthmarelated health outcomes." Mudarri and Fisk (2007) esti mated that 21% of current U.S. asthma cases were potentially attributable to dampness and mold in housing, for an annual national cost of $3.5 billion. Fisk et al. (2010) estimated that residential dampness or mold is associated with 8-20% of U.S. respiratory infections.
With regard to practical implications of these findings, we did not evaluate health bene fits of specific strategies for remediation of dampness or mold. However, a recent expert review has concluded that the intervention of "combined elimination of moisture intrusion and leaks and removal of moldy items" had sufficient evidence of effectiveness for reducing respiratory symptoms from asthma and aller gies and was ready for widespread implemen tation (Krieger et al. 2010).
Limitations. Much of the epidemiology on dampness, mold, and health has used sub jective reports for assessing exposure or health and thus has potential for reporting bias. Two reviews have considered whether biased subjective response by building occupants in dampness studies might have positively biased findings. On the basis of comparison of results in six studies from occupant reports versus inspectorreported dampness and clinically determined illness, Fisk et al. (2007) con cluded that observed associations of respiratory health effects with dampnessrelated exposures were unlikely to be explained by over reporting. Bornehag et al. (2001) reported that findings of studies with independent assessment of both dampness and health effects were similar to findings of studies with more subjective information sources. Additionally, avoidance behavior (prior exposure reductions by persons with asthma) may be a source of past exposure misclassification with assessment of only cur rent or recent exposure. However, this is not a concern in prospective or intervention studies, which have generally confirmed dampness/ health associations.
Quantitative measures of exposure used in the reviewed studies also have important limitations. Measured airborne concentrations of culturable microorganisms have substantial volume 119 | number 6 | June 2011 • Environmental Health Perspectives errors, for example, from shortterm estimation of airborne concentrations with large and rapid variations over time; from differential abilities of organisms to grow on specific culture media; and from nondetection by culture assays of most bioactive microbial materials, whether intact spores or fragments. Most important, culturebased or nonculturebased microbial measurements used in many studies may not target actual causal factors. All these reasons may explain the lack of consistent associations between reported microbial measurements and health. And as with glucans or endo toxins, even prior demonstration in many studies that a substance causes inflammation does not implicate it as consistently harmful, because both glucans and endotoxins have also demon strated healthprotective associations (Douwes et al. 2006;Iossifova et al. 2007). However, subjectively assessed dampness or mold has not shown protective associations, even in infants.
Finally, definitions of respiratory health effects are not standardized, potentially caus ing bias. In population studies, asthma is usu ally defined by selfreported (or parentally reported) asthma symptoms. Selfreports of doctordiagnosed asthma are also often used. An alternative approach to questionnaires has been to use more objective measures, either alone or in combination with questionnaires. As with measures of home dampness or fun gal exposures, differences in asthma definition are likely to result in differences in estimates of RRs. In addition, as mentioned above, sev eral studies (Nafstad et al. 1998;Oie et al. 1999) focused on infants at an age where the diagnosis of asthma is uncertain. Most of these potential sources of bias are expected to underestimate any true association between indoor dampness and health effects.
The restricted scope of this review led to further limitations. The method of evaluating published evidence was largely nonquantitative. Results of available quantitative metaanalyses, however, are consistent with qualitative sum maries. Publication bias in this review is likely to have inflated associations of risk factors with health effects. A formal application of available statistical methods for assessing presence of this bias was not feasible for this broad review. A search for unpublished findings, which may decrease publication bias, was not performed.
Conclusions drawn from this review should thus be considered provisional until the pro duction of quantitative summary estimates of RRs based on more thorough consideration of all available findings, with formal evaluation for publication bias.

Evidence for plausible biologic mechanisms of health effects from dampness-related agents.
Toxicologic evidence suggests plausible bio logic mechanisms for the respiratory health effects associated epidemiologically with damp ness or mold (WHO Europe 2009). In vitro and in vivo studies have demonstrated diverse inflammatory, cytotoxic, and immuno suppressive responses after exposure to the spores, metabolites, and components of specific microbial species found in damp buildings. Repeated immune activation and prolonged inflammation by microbiologic exposures may contribute to inflammationrelated dis eases such as asthma. The immunosuppressive response demonstrated in animals exposed to fungal spores associated with damp buildings may explain a link to respiratory infections.
The wide variety of health effects associated with dampness and mold cannot be explained by a single mechanism. Epidemiologic evi dence suggests involvement of both allergic and nonallergic mechanisms, as both atopic and nonatopic individuals are susceptible to adverse effects of dampness or mold (e.g., CoxGanser et al. 2005;Dales et al. 2006;Douwes et al. 2006;Kuyucu et al. 2006). The inflammatory responses demonstrated in many microbiologic exposures include histamine release by nonimmunoglobulin E-mediated mechanisms, providing plausible mechanisms for the occurrence of allergylike symptoms in nonsensitized individuals. Increased human susceptibility to severe asthma exacerbation from fungal exposures has been demonstrated with genetic polymorphisms related to chi tinase, suggesting mechanisms involving fun gal chitin (Wu et al. 2010).
Some available evidence is consistent with involvement of fungal toxins in some health effects associated with damp environments, although this has been debated extensively in the literature (Bennett and Klich 2003;Jarvis and Miller 2005). Recently, animal models with curdlan (a specific triplehelical form of fungal glucan) and several toxic fun gal metabolites have demonstrated inflam matory, nonallergic respiratory health effects consistent with the epidemiology of damp ness Rand et al. 2010). Observed synergistic interactions in toxico logic studies among microbial agents present in damp buildings, including specific fungi, actinomycetes, and amoebae (Penttinen et al. 2006;YliPirila et al. 2007) suggest that immunotoxic effects of fungal and bacterial strains typically found in damp buildings may be potentiated during joint exposures. Such potentiation could explain difficulties in iden tifying specific causal exposures for health effects in damp buildings.
Many limitations of culturebased micro bial assessments for investigating causes of dampnessrelated health effects have long been evident. Additional support for the need to investigate nonculturebased microbial assessment methods has been provided by the demonstration (Gorny et al. 2002) that fungi and actinomycetes can emit large numbers of airborne particles smaller than spores and not detectable by culture but with demon strated immunogenic properties. These find ings provide additional plausibility for health effects associated with microbial growth but not measurable with culture assays.
The hygiene hypothesis. As summarized in this review, indoor dampness or mold is consistently associated with increased respi ratory health risks, and microbial exposures have been suggested (but not proven) to play a causal role. On the other hand, an increas ing number of studies suggest that earlylife microbiologic exposures to endotoxin or spe cific fungal agents may protect against atopy and allergic disease. This potentially protective effect is consistent with the "hygiene hypothe sis," which postulates that growing up in a more microbiologically hygienic environment may increase the risk of developing respiratory allergies (e.g., Douwes et al. 2004Douwes et al. , 2006Liu and Leung 2006).
However, the evidence for protective effects of microbial exposures has not been consis tent, as increased health risks have been associ ated with some specific measured exposures (e.g., Bolte et al. 2003;Dharmage et al. 2001;Michel et al. 1996). Some of these inconsisten cies, found for endotoxin, (1→3)βdglucans, and fungi, may be related to timing or dose of exposure, as has been recently hypothesized (Douwes et al. 2007), but evidence is still weak. For instance, Iossifova et al. (2007Iossifova et al. ( , 2009, in prospective data, identified nonmonotonic relationships between (1→3)βdglucans in dust and recurrent wheeze, wheeze with atopy, and an index for future asthma: Risks increased at increasing low concentrations, reached a maximum at 60 µg/g dust, and then decreased at increasing high concentrations. Similar pat terns have also been observed with dust mite antigen (Tovey et al. 2008).
At present, modest exposure to some microbial exposures under certain circum stances appears to protect against allergies and allergic asthma but not wheeze; however, as indicated previously, the overall evidence is inconsistent. Damp or moldy buildings seem only to increase, not decrease, the develop ment of respiratory disease, both in allergic and nonallergic subjects including infants.
Suggested research. A focused research program in this area might include a) studies to identify and improve objective tools and metrics that, in assessing either dampness or specific related factors (microbiologic or non biologic), optimally predict disease; b) stud ies to characterize dose-response relations, to determine safe levels and identify ageor dose related protective effects; and c) strong stud ies (intervention or prospective) designed in the aggregate to document causality between dampness or mold and key health effects such as asthma or respiratory infections. Genetic epidemiology may enhance abilities to detect causal exposures and identify mechanisms (Wu et al. 2010). Indoor occupational set tings and schools, with multiple advantages for study efficiency and logistics, have been under utilized. Good examples to follow include the strong disease prediction by an objective and easily interpreted tool, the electronic resistance type moisture meter (Williamson et al. 1997), and the welldesigned and extremely effective remediation study by Kercsmar et al. (2006). Although future findings will improve health protective policies, healthprotective actions need not await further etiologic research.

Conclusion
Based on the material reviewed here, there is sufficient evidence of an association between indoor dampnessrelated factors and a wide range of respiratory or allergic health effects (Table 3), including asthma development, asthma exacerbation, current asthma, ever asthma, dyspnea, wheeze, cough, respiratory infections, bronchitis, allergic rhinitis, eczema, and upper respiratory tract symptoms. There is suggestive evidence of associations with health effects for several nonculturebased measurements related to fungi and bacte ria in dust, although some of these associa tions seem equivocal. No evidence suggests protective effects of evident dampness and mold. Mechanisms seem likely to be both allergic and nonallergic. Available quantita tive metaanalyses have estimated consistently and significantly increased risks for multiple outcomes associated with dampness or mold, including OR ranges of 1.30-1.75.
Substantial increases in a number of impor tant respiratory health outcomes, including a 50% increase in current asthma, are associated with dampnessrelated risk factors in residences ). These estimates, based on limited data, broad lumping of diverse risk factors, and multiple unverified assumptions, should be interpreted cautiously; however, they indicate that dampnessrelated risk factors may contribute substantially, but preventably, to the burden of respiratory disease.
In agreement with the IOM report (2004), we consider that there is not sufficient epide miologic evidence of a causal relationship for any of the reviewed health outcomes, although for asthma exacerbation in children we con sider the evidence strongly suggestive of cau sality by dampnessrelated agents. Although it is plausible that microbial exposures may play a causal role, specific causative agents have not been established. In fact, limited and inconsis tent evidence suggests that moderate exposures to certain microbial agents, especially at early ages, may prevent allergies and allergic asthma.
Based on available evidence, the presence of dampness, water damage, visible mold or mold odors or a history of water damage pro vides more reliable indicators of dampness or moldrelated health risks than do current quan titative microbiologic assessments. As reduc tion of indoor dampness and mold is likely to have benefits for respiratory and allergic health of occupants, this level of knowledge should guide practical prevention and remedia tion now. Still, available research does not yet indi cate the amount of water damage, mold, or mold odor meriting concern nor document the relative magnitude of health benefits from different environmental remediations.
Although Williamson et al. (1997) pub lished findings of strong, doserelated asso ciations of asthma severity with systematic moisture measurements in walls 13 years ago, research use of quantified dampness metrics has not been reported since. Future research, generally, should develop objective metrics for dampnessrelated and microbial (or non microbial) risk factors that predict health effects. This will help in identifying specific causal dampnessrelated agents and character izing exposure-response relationships.
Challenges to progress include the wide variety of currently plausible micro organisms (fungi, bacteria, amoebae/protozoans) and microbial components and products eligible to be causal factors; the potentially non monotonic effects of some of these compo nents (e.g., glucans and endotoxin); the potential synergistic actions of some organ isms, including actinomycetes and amoebae; the possible involvement of nonbiological, chemical agents released from damp indoor materials; and the modification of microbial effects by human age at exposure or by genetic or other host susceptibility factors. However, although their effectiveness may ultimately be improved, prevention and remediation actions to reduce indoor dampness are important and urgently needed in a large proportion of our building stock. These measures are likely to significantly reduce the current global burden of respiratory and allergic disease.