The challenge for microbial measurements in buildings.

Abstract

Dampness and mold (D/M) in buildings are consistently linked to poor respiratory health of occupants. D/M-related health effects include the development of asthma, asthma exacerbations, respiratory infections, and upper and lower respiratory symptoms.1, 2 Despite the long understanding that living in a damp or moldy home is bad for your health—Leviticus 14:33-54 advises the destruction of homes with recurring mold—the specific agents that cause the health effects have still not been identified.1, 2 When D/M occur, current public health guidelines typically recommend steps to remediate the building: identify and eliminate sources of building moisture, dry or remove wet materials, and clean or remove moldy materials.3 Fundamental questions remain: how much D/M is too much and should trigger immediate remedial actions; after remediation, was it good enough? We want to be able to set protective threshold levels that would specify both what to measure and how much is unacceptable for health. To achieve this, we first need to identify measurable D/M factors for which we can show that the more exposure to the factor, the greater the health risks; that is, factors correlated strongly with health effects in a “dose-related” way. Ideally, the measured factors would be the actual dampness-related causal agent(s) of disease. However, because we do not know what those agents are, until they are identified we could use other, easy-to-measure indicators of D/M as surrogates, if they were strongly associated with the health effects. Then, we would need to quantify the dose-response relationships of these agents or surrogates with the most important health effects, so that we could set maximum exposures acceptable for health. Figure 1 summarizes a simplified version of this process. The ways that D/M have been assessed in building studies fall into three broad categories: (a) quantitative microbiological measurements of either microbial organisms, communities, components, or products1, 2, 4; (b) measured moisture in building materials, for example, using 2-pin conductance meters1, 5; and (c) observed D/M factors (assessed by researchers or occupants) such as visible mold, mold odor, water damage, and moisture/dampness.1, 2, 4, 6 Figure 2 briefly summarizes, for each category, the strongest relationships (as odds ratios or risk ratios) reported with adverse health effects. As shown in Figure 2, the non-microbiological assessments of D/M, including measured moisture and observed D/M, have shown much stronger associations with health effects (including estimated odds ratios or risk ratios as large as 14) than have any of the microbiological measurements studied thus far (all estimates below 3). The best predictors we have now to use for D/M guidelines are the observed D/M factors, including mold odor, visible mold, water damage, or moisture. These assessments in Figure 2 were all conducted by trained researchers, not by occupants.6 Although these are only qualitative or, at best, semi-quantitative,6 no specific quantitative microbiological measurement has predicted adverse health effects as successfully and consistently as these non-microbiological measures. It is perplexing that our quantitative microbiological measurements are not showing strong associations with the health effects being studied. This suggests that we are not yet measuring the correct microbial factors, or not measuring them well, or perhaps just not measuring them at the right location or time. Meanwhile, the non-microbiological factors being assessed, because they are related so much more strongly to the health effects, often with dose-response, are apparently linked more closely to the underlying causal agents. For this reason, these factors would be considered more suitable (compared to weakly associated measurements) to use for now as a basis for health-related guidelines and standards. The results in Figure 2 define a challenge and a set of goals for microbiological methods development: to identify microbial agents or products, and ways to measure them, that are at least as effective at identifying elevated health risks as are the more qualitative D/M observations. Methods to measure microbiological agents have advanced over the years. Culturing microorganisms has been an invaluable scientific tool, but the limitations of culture-based methods for comprehensive microbial exposure assessment have long been recognized. Microbial cell components (eg, glucans in fungi and endotoxins in bacteria) have had conflicting associations (ie, both positive and negative) with health effects, limiting the utility of these immunologically-based assays for estimation of health risks.2 These components are not included in Figure 2. Volatile compounds emitted by microorganisms (microbial volatile organic compounds or MVOCs), identified by chemical analyses at low levels, have not been correlated with human health effects.20 Most current efforts in indoor microbiology focus on detecting the genetic material of microbial cells. Of the quantified microbiological measurements used to date, assessing the presence and quantity of specific fungi by quantitative polymerase chain reaction (qPCR) has been the most successful at predicting health effects.10 The most recently available methods, called next-generation sequencing (NGS), are DNA-based tools allowing comprehensive microbial identification with greatly increased speed and reduced cost. The potential of NGS approaches to comprehensively characterize indoor microbiomes and then identify the dampness-related causal agents of disease, however, has thus far not been fulfilled. This is perhaps because NGS tools have been rarely applied to this question, but also because of limitations of the current methods, which are only semi-quantitative, suffer from various biases, and do not differentiate live from dead organisms. Currently, observed dampness and mold factors are the closest to providing practical strategies for estimating dampness-related health risks in buildings, but these assessments still need further refinement for setting explicit health-relevant guidelines. Specifically, these assessments should be dose-related to key health effects and reproducible enough to be suitable for setting health-protective thresholds. Promising possibilities include metrics that combine water damage, visible mold, and mold odor,6, 17, 18 and moisture measurement instruments with readings that are more standardized and interpretable.5 In parallel with that approach, it is still important to identify the dampness-related microbial agents of disease, or microbial factors strongly associated with them, and to develop practical ways of measuring these. Ultimately, the most precise exposure guidelines would be based on levels of the actual causal agents. Toward that goal, we focus here on the challenge of microbial measurement in buildings. We suggest improvements in several aspects of research methods, including exposure assessment, microbial identification, and analytical approaches. Current exposure assessment strategies to quantify microbial products may inadequately describe relevant exposures. Even if quantified microbiological approaches were measuring the responsible agent, that signal could be lost if our sampling techniques were inefficient or biased, or missed the appropriate location or time. For instance, many current measurements are based on either dust samples (floor or elevated settled dust), which may not adequately represent inhalation exposures, or brief grab samples of air, which poorly represent air concentrations over time. These limitations point to a need in the exposure sciences to refocus on improved measurement of microbiological agents.21 For microbial identification, three improvements in methods could help identify health-relevant microbial exposures. One would be to take measurements that indicate either active growth or building conditions allowing active growth, in the present or the past, just as the successful non-microbiological measurements shown in Figure 2 have done. In contrast, the quantified microbiological measurements used presently can detect the presence of microorganisms in building environments, but not necessarily whether they grew in the building or if building conditions could have supported this growth. Another improvement would be to make NGS methods, which are the most comprehensive at detecting organisms, fully quantitative at the species level, which would increase the ability to identify the relevant exposures. In addition, even successfully identifying and quantifying microbial species may not be enough to provide the needed answers. Because available evidence suggests that multiple dampness-related agents may cause health effects through different pathways, research should include the many ways in which microbes can interact with the immune system, including living as well as non-living microbes, microbial organisms as well as various configurations of microbial components or products, and allergenic as well as inflammatory or toxic microbial agents. Regarding analytical approaches, NGS methods detect DNA signatures so sensitively and broadly that new tools are needed to identify the truly health-relevant exposures within the vast amount of data produced. Finding the signal within the noise is also a challenge in research on MVOCs and health effects. Regarding analytical modeling specifically, the studies with the “successful” non-microbiological measurements used exposure indices with multi-categorical or continuous values6 (Figure 2). Rather than binary exposure assessments, these measured moisture and observation-based metrics have multiple outcome levels to compare with health outcomes. Analyzing quantitative microbiological measurements as multi-categorical or continuous indices may reveal associations that are less apparent with simple presence/absence or high/low indicators, and can assess dose-response relationships, providing a better basis for setting protective guidelines. Overall, we see promise in parallel efforts to improve both non-microbiological and microbiological assessments. For the latter, we need to develop methods that consider the possible involvement of multiple microbial agents, and that quantify the exposures of microorganisms currently or previously active in the building, or their microbial products. Future research to find D/M-related microbiological measurements strongly related with health effects will be more effective if it includes collaboration between microbiology, epidemiology, and exposure science. We hope that several thousand more years do not pass before we understand why damp and moldy buildings are bad for our health, and how much dampness and mold is too much. The authors state that they have no conflicts of interest related to the material in this editorial. This work was supported entirely by the California Department of Public Health.