Indoor Air 2016; 26: 61–78 © 2014 The Authors. Indoor Air published by John Wiley & Sons Ltd wileyonlinelibrary.com/journal/ina Printed in Singapore. All rights reserved INDOOR AIR doi:10.1111/ina.12174 Keynote: Indoor Air 2014

Indoor bioaerosol dynamics

Abstract Inhaling indoor air is the primary means by which humans are William W Nazaroff exposed to bioaerosols. Considering bacteria, fungi, and viruses, this study reviews the dynamic processes that govern indoor concentrations and fates of Department of Civil and Environmental Engineering, biological particulate material. Bioaerosol behavior is strongly coupled to University of California, Berkeley, CA, USA – particle size; this study emphasizes the range 0.1 10 lm in aerodynamic Key words: Bacteria; Deposition; Emission; Fungi; diameter. The principle of material balance allows concentrations to be Ventilation; Virus. determined from knowledge of important source and removal processes. Sources reviewed here include outdoor air introduced by air exchange plus indoor emission from occupants, occupant activities, and moldy materials. W. W. Nazaroff Important mechanisms that remove bioaerosols from indoor air include air Department of Civil and Environmental exchange, deposition onto indoor surfaces, and active filtration. The review Engineering, University of California, Berkeley, summarizes knowledge about size-dependent particle deposition in different CA 94720-1710, USA regions of the respiratory tract, techniques for measuring indoor bioaerosols, Tel.: +1 510 642 1040 e-mail: [email protected] and evidence for diseases caused by airborne exposure to bioaerosols. Future research challenges and opportunities are highlighted. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any med- ium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

The copyright line for this article was changed on 27 March 2015 after original online publication.

Received for review 26 September 2014. Accepted for publication 3 December 2014.

Practical Implications Bioaerosol exposure indoors is an important factor influencing human health. Understanding the dynamic behavior of biological particles can contribute to more incisive research to characterize exposure and associated health conse- quences. Such understanding is also essential in the design of effective engineering controls to reduce concentrations and to limit exposures.

Introduction seek to summarize what is known or what can reason- ably be inferred about an important subtopic of the In our daily lives, we humans move through a sea of microbiology of the built environment. The title microbial life that is seldom perceived except in the con- ‘indoor bioaerosol dynamics’ is meant to imply that we text of potential disease and decay (Feazel et al., 2009). seek a mechanistic, quantitative description of the pro- It is by now well established that most humans spend cesses and outcomes regarding the microbiology of the most of their time indoors. Furthermore, it is also well built environment, centering on what is, was, or will be known that the indoor environments occupied by airborne. humans contain abundant material of microbial origin. There are many reasons to be interested in the However, we have gained until now only partial under- microbiology of the built environment. The context for standing of the complexity and richness of the indoor this study derives from three primary concerns. First, microbiome and its significance for human well-being. exposure to bioaerosol material can cause or can con- While acknowledging the limitations, in this article, I tribute to many important diseases. Second, adverse

61 Nazaroff respiratory symptoms correlate well with indicators of bioaerosol control, airborne growth and decay, and indoor dampness (Fisk et al., 2007). It is a reasonable indoor transport and mixing. hypothesis, although not yet well established, that the In the conclusion, important challenges facing fur- underlying cause of these associations is microbial. ther studies are described along with several opportuni- Third, the microbiology of spaces we occupy may influ- ties for near-term progress advancing knowledge about ence the human microbiome in ways that could confer indoor bioaerosol dynamics. Such progress is funda- health benefits. Evidence for this last point is admit- mental to efforts to better understand how the microbi- tedly thin. There is growing evidence that people influ- ome of indoor environments interacts with human ence the microbiology of the spaces they inhabit (e.g., health and well-being. Lax et al., 2014). Emerging evidence also indicates that some aspects of the human microbiome are important for health (Grice and Segre, 2012). We further know Framing the issues that aspects of our personal microbiome can be influ- Scope, limitations, and approach enced by individual factors such as diet and illness. So, it is not a large stretch to imagine that elements of our Bioaerosols are usually defined as aerosols or particu- personal microbiome that matter for health might in late matter of microbial, plant or animal origin.... fact be influenced by attributes of the spaces we Bioaerosols ... may consist of pathogenic or non- inhabit, including their microbiology generally and pathogenic live or dead bacteria and fungi, viruses, their bioaerosol aspects specifically. Hanski et al. high molecular weight ... allergens, bacterial endo- (2012) reported evidence supporting an analogous sug- toxins, mycotoxins, peptidoglycans, b(1?3)-glucans, gestion: ‘rapidly declining biodiversity may be a con- pollen, plant fibers, etc.—Douwes et al. (2003) ... tributing factor to the rapidly increasing prevalence ‘Bioaerosol’ is a contraction of ‘biological aerosol’, of allergies and other chronic inflammatory diseases and ‘aerosol’ refers to a suspension of particles in a ... among urban populations .’ gas. The exact boundary for what should be included Following the introduction, the study is organized or excluded in the term ‘bioaerosol’ is challenging to into two main sections. In ‘Framing the issues,’ I define. For the purposes of this study, the central focus define a framework for considering indoor bioaero- will be defined more narrowly. All microbes are sol dynamics. That section also defines the scope included: viruses, bacteria, and fungi. Also included and limitations of this study. It presents some are microbe-associated chemicals such as endotoxins empirical evidence about indoor bioaerosols to help and mycotoxins. However, although they are part of establish a context for studying dynamics. The sec- bioaerosols, this article shall not explicitly address tion outlines the use of a material balance as a either pollen or pet- or pest-associated allergens such first-principles method for linking process to out- as cat dander or fecal pellets of dust-mites. come for indoor bioaerosols. The size of particles is This review also emphasizes the human–bioaerosol described as a primary variable of concern. Regio- nexus in indoor environments. The indoor environ- nal deposition of bioaerosol particles in the respira- ments to be considered are those ordinarily and com- tory tract is described as a process linking monly occupied by humans: residences, offices, concentrations and exposure to dose. Opportunities schools, and other settings that are occupied a high and limitations for progress in understanding proportion of the time, or in which occupant density is indoor bioaerosol dynamics are strongly influenced high. Not included are industrial environments that by measurement technology, so some of the key have high associated occupational exposure potential, measurement methods are summarized. Evidence such as those associated with food systems. regarding infectious disease transmission by airborne Much of the literature concerning microbiology in routes is also assembled. indoor environments focuses on dust as the sampled The second primary section (‘Dynamic processes’) matrix (Rintala et al., 2012). Favoring this approach is summarizes available information on many important that dust reflects a longer term average condition in the processes that affect the concentrations and fates of environment, whereas bioaerosols are more dynami- indoor bioaerosol particles. This section begins by cally variable and therefore more challenging to sample reviewing the state of knowledge regarding building air in a manner that is representative of time-averaged exchange, an important removal mechanism for indoor conditions. The primary counterargument is that the bioaerosols and also a means by which outdoor bio- relationship between dust and human exposure is much aerosol particles are brought indoors. Then, the article less clear than is the relationship between bioaerosol proceeds to discuss several additional processes that particles and inhalation exposure. Notwithstanding its can affect indoor bioaerosol levels: deposition onto prominence in the literature, the microbiological com- room surfaces, bioaerosol intrusion from outdoor air, position of indoor dust is not a primary focus of this indoor emission sources, and other factors, including study.

62 Indoor bioaerosol dynamics

The overall approach employed in this review is reported on the sampling protocol; however, it seems to adapt and apply concepts from indoor aerosol likely that the measurements were based on short-term science (e.g., as described in Nazaroff, 2004). In par- sample collection. Also, little information was reported ticular, material balance is utilized as a core princi- about the buildings in which these measurements were ple. Particle size is a primary determinant of indoor taken. The cumulative probability distribution func- bioaerosol behavior. We seek a mechanistic under- tions (Figure 1) show that the data are well fit by log- standing because it provides a powerful basis for normal distributions. The geometric mean outdoor extrapolating from limited empirical evidence. We concentration (493 CFU/m3) is about 6 9 higher than also seek to develop quantitative insight, because the geometric mean indoor concentration (79 CFU/ knowledge of the scale of processes and outcomes is m3). The indoor levels show considerably broader essential as a basis for separating the important range of concentration than do the outdoor levels from the trivial. Some prior scholarly reviews cover (GSD of 5.5 compared to 3.3). These data support an similar topics (Burge, 1990; Douwes et al., 2003; important general finding that ‘in the majority of cases, Gregory, 1971; Spendlove and Fannin, 1983). How- indoor fungal aerosols are controlled by outdoor con- ever, none of these earlier reviews is as strongly centrations’ (Burge, 2002). Note that the ratio of geo- grounded in indoor aerosol science as the present metric mean concentrations, 0.16, lies well within the article. range 0.04–0.46 predicted by Riley et al. (2002) to apply as the fraction of coarse-mode (2.5–10 lm) particles that would be found in indoor air when out- Some empirical evidence door air was the sole source. To provide context for the process-oriented discussion Culture-based analyses measure only a small portion to follow, it is instructive to consider some of the of bioaerosols. Figure 2 presents data from a personal important empirical evidence concerning indoor bio- monitoring study of elementary school teachers in Fin- aerosols. In the summaries to follow, I highlight several land (Toivola et al., 2004). For each of the 81 subjects, field-sampling studies whose results provide important air was sampled through filters throughout two 24-h clues about bioaerosol concentrations, associated par- periods using personal sampling pumps. Particles were ticle-size distributions, and potential influencing fac- extracted from the filters and analyzed for both fungi tors. and bacteria, using culture-based methods and also The largest published survey of indoor bioaerosol using microscope-assisted visual counting. A key point levels involved more than 12 000 fungi samples mea- is that the total number concentrations of fungal sured indoors and outdoors in ~ 1700 buildings in the spores and bacterial cells determined microscopically United States (Shelton et al., 2002). The analysis was were a few orders of magnitude higher than the corre- culture based, so the results reflect number concentra- sponding number concentrations of colony forming tions of viable airborne spores. Little information was units, as determined by culture-based assessment. If

3000 10 5

GM × GSD 1000 10 000 GM

GM/GSD Viable bacteria 300 1000 Total bacteria Viable fungi 100 100 Total fungi

30 10 Outdoor (GM = 493; GSD = 3.3) Concentration (per cubic meter) Fungi: Total/viable ~ 400 Culturable fungi (CFU per cubic meter) Indoor (GM = 79; GSD = 5.5) Bacteria: Total/viable ~ 130 10 1 2550 75 95 Cumulative frequency (%) Fig. 2 Results from personal sampling of 81 elementary school teachers in eastern Finland (Toivola et al., 2004). For each sub- Fig. 1 Cumulative probability distribution of indoor and out- ject, 24-h filter samples were collected twice during winter. Total door airborne concentrations of culturable fungi measured in fungi and bacteria were determined by microscopic analysis. 1717 US buildings (Shelton et al., 2002). In all, 9619 indoor and Viable fungi and bacteria were evaluated by culture-plate assay. 2407 outdoor samples were collected during indoor air quality The ratios in the lower right box are based on geometric means investigations (GM)

63 Nazaroff culture-based sampling methods only measure a tiny samples are considered. The ten houses are divided fraction of indoor bioaerosols and if what they do mea- into two equal-sized groups, sorted according to sure is not representative of what is airborne, then cul- degree of occupancy (Chen and Hildemann, 2009b). ture-based investigations cannot completely Occupancy is used here as a surrogate measure of the characterize indoor bioaerosol attributes that might intensity of human activity. Across all measures dis- influence human health and well-being. played in the figure—that is, airborne concentrations Although it is common for fungi concentrations to of particulate matter, protein, endotoxin, and glucans be higher in outdoor air than indoors, in occupied —the geometric mean level in the high activity homes spaces the reverse is commonly true for bacteria (Bart- was considerably higher than in the low activity lett et al., 2004; Fox et al., 2005; Hospodsky et al., homes. 2015). Figure 3 (Chen and Hildemann, 2009a) illus- Measurement of airborne viruses in indoor envi- trates this point. The plotted results are geometric ronments has lagged behind measurement of bacteria mean concentrations based on filter samples of 9–12 h and fungi. The recent development of quantitative duration collected inside and outside of ten occupied polymerase chain reaction (qPCR) and other DNA- residences in California. Summing across all particle based measurement technologies has facilitated stud- sizes, the GM level of endotoxin (associated with Gram- ies that measure pathogenic viruses in indoor air. negative bacteria) was about 50% higher indoors than Data from one study targeting the influenza A virus outside, whereas the GM level of (1–3)-b-D-glucans are presented in Figure 5. During the flu season, size- (a marker of fungi) in outdoor air was almost twice the resolved particle samples were collected on filters in corresponding indoor level. three different environment types: a day care center, An important point is that human occupancy and a health center, and (three) airplanes. In all, 16 sam- activities are major factors influencing indoor microbi- ples were collected by means of sampling at a rate of ology. Humans are important primary sources of cer- 9 l/min for periods of 6–8 h. Eight of these samples tain bacteria and viruses. Even for fungi, for which (50%) contained influenza A virus, with concentra- humans do not appear to be a major primary source, tions ranging from 5800 to 37 000 genomes per m3, human activities play an important role, for example, and a substantial proportion of the detected virus in shedding particulate matter from our clothing or in was associated with fine particles (< 2.5 lm) that can suspending settled dust that can contain materials of remain airborne for extended periods and that can fungal origin. Figure 4 illustrates this point using sam- also penetrate and deposit deeply in the respiratory ples collected in the same ten houses as discussed in tract when inhaled. connection with Figure 3. In this case, only indoor air

Particle size: ≤ 2.5 µm2.5–10 µm > 10 µm Particle size: ≤ 2.5 μm 2.5–10 µm> 10 µm 6 4 PM Protein Endotoxin Glucans PM Protein Endotoxin Glucans (× 10 µg/m3) (µg/m3) (EU/m3) (ng/m3) (× 10 µg/m3) (µg/m3) (EU/m3) (ng/m3) 3.5 5

3 4 2.5

3 2

1.5 2 Concentration, GM Concentration, Concentration, GM 1 1 0.5

0 0 In Out In Out In Out In Out LH LH LH LH Indoor or outdoor measurement Indoor Activity Level (L ≤ 0.3 person-h/h, H ≥ 0.4 person h/h)

Fig. 3 Aerosol and bioaerosol parameters sampled inside and Fig. 4 Aerosol and bioaerosol parameters sampled inside ten outside ten homes in northern California. At each house, 4 day- homes in northern California. At each house, 4 daytime samples time samples were obtained (9–12 h each), using size-segregated, were obtained (9–12 h each), using size-segregated, filter-based filter-based collection. Plotted results are the geometric mean collection. Houses were sorted into two groups according to the concentrations for indoors (N = 39–40) and outdoors (N = 20). degree of occupancy in the living room during sampling. Data Data source: Chen and Hildemann (2009a) source: Chen and Hildemann (2009b)

64 Indoor bioaerosol dynamics

45 (a) Q L V Particle size 40 µ C p > 2.5 m o 1–2.5 µm Day-care (3/4) C µ Q 35 0.5–1 m N Q Q 0.25–0.5 µm N + L 30 < 0.25 µm E Airplane (2/3) β 25 Health Center (3/9) η 20 R Q

ion (1000 GCN per cubic meter) ion (1000 GCN per cubic R 15 (b) Q 10 L V C p 5 o Virus concentrat C 0 Q Q + Q η M M L Fig. 5 Size-resolved influenza A virus concentrations measured M E in filter samples collected in indoor environments during the β 2009–2010 flu season (Yang et al., 2011). In all, eight of sixteen samples collected (3 of 9 in a health center, 3 of 4 in a day care η R Q center, and 2 of 3 in an airplane) tested positive for influenza A. R Abbreviation: GCN = genome copy number Fig. 6 Schematic representation of indoor environments for quantitatively relating influencing processes to resulting concen- trations of bioaerosol parameters: (a) residence; (b) commercial Material balance building. Symbols (associated units) are as follows: Co—out- 3 — A fundamental principle that is usefully applied in door concentration of bioaerosol parameter (quantity/m ); C indoor concentration of bioaerosol parameter (quantity/m3); V quantitative, mechanistic studies of indoor environ- 3 3 —interior volume (m ); QN—natural ventilation rate (m /h); 3 mental quality is material balance: Stuff is conserved. QL—ventilation rate associated with infiltration (m /h); QR—re- 3 On a time-averaged basis, the sum of the rates of sup- circulation airflow rate (m /h); QM—mechanical ventilation ply of a bioaerosol component to the indoor air must supply flow rate (m3/h); p—penetration efficiency of bioaerosol — — balance the sum of the rates of removal. This quantita- parameter associated with infiltration ( ); gR single-pass fil- tration plus deposition efficiency in recirculating airflow (—); tive balance provides a basis for connecting rates of gM—single-pass filtration plus deposition efficiency in mechani- processes to concentrations of bioaerosol components. cal ventilation supply flow (—); b—rate coefficient for deposi- Figure 6 presents schematic representations of tion on indoor surfaces (h1); and E—emission rate from indoor indoor environments that can be used to formulate sources (quantity/h) material balance equations. Consider a residential space (Figure 6a). In this representation, three pro- applicability are that some of the parameters on the cesses can add bioaerosol material to the indoor air: (i) right-hand side are time dependent. If applied over natural ventilation through designed openings (at rate short time intervals, one must also be concerned that Q 9 C ); (ii) infiltration through leaks in the building N o accumulation, that is, the increase or decrease in the envelope (p 9 Q 9 C ); and (iii) direct indoor L o amount of bioaerosol in the indoor space, is not incor- emissions (E). (To the extent that tracked-in microbes porated in equation (1). Also, on a time-averaged basis, contribute to the indoor bioaerosol load, this process the equation is only approximately correct because it would be included in the direct indoor emission term, does not account for the possibility that time-varying E.) Three processes can remove bioaerosol material: (i) parameters may correlate in such a way that the aver- ventilation airflow out of the building ([Q + N age of the product is not the same as the product of the Q ] 9 C); (ii) filtration in the recirculating airflow L averages. A second key point is that some of the pro- (g 9 Q 9 C); and (iii) deposition onto room sur- R R cesses exhibit strong dependence on particle size. This faces (bVC). aspect is addressed in detail in subsequent sections of the study. A third feature of this equation is that it ðÞþ þ QN pQL Co E : ð Þ implicitly assumes that the indoor space can be repre- C e þ þ þ 1 QN QL gRQR bV sented as well mixed. That is not always the case. Finally, the equation is specific to the particular sche- Some important observations should be made about matic representation of the indoor environment equation (1). First, the symbol ‘~’ is used instead of an depicted in Figure 6a. This configuration accommo- equal sign because the expression is only approximately dates some common conditions in residences, such as true. Among the considerations that limit its strict air exchange by a combination of natural ventilation

65 Nazaroff

(QN) and infiltration (QL), and the potential presence of (2014) found that the geometric mean size of fungal a central air distribution system (with flow rate QR) for DNA associated with particular taxa was considerably heating and cooling. On the other hand, the schematic larger than individual cultured spores. and the resulting equation would need to be modified Theoretical and experimental evidence support an to accommodate mechanical ventilation. expectation that microbial airborne particles behave For a commercial building space, the schematic rep- similarly to abiotic particles of the same aerodynamic resentation in Figure 6b is more common, with size. Consequently, powerful tools and theories from mechanical ventilation (QM) and no natural ventila- aerosol mechanics can be applied to study indoor bio- tion. The appropriate material balance in this case is aerosols. presented in equation (2). Similar caveats as in equa- tion (1) apply. Respiratory tract deposition ½Þð þ þ For bioaerosol particles, arguably the most important 1 gM QM pQL Co E : ð Þ C e þ þ þ 2 exposure pathway is inhalation followed by deposition QM QL gRQR bV in the respiratory tract. The probability of deposition varies with particle size, with lung morphology, and with breathing characteristics. Figure 7, which is based Particle size on semi-empirical modeling originally developed for Most indoor airborne microbial material is found on radiological protection (Yeh et al., 1996), illustrates particles in the diameter range 0.1–10 lm. This range some of these features. In these model calculations, the of diameters corresponds to six orders of magnitude in respiratory tract is divided into three zones: the head particle mass. In part because of the large range, size is region (NOPL), the tracheobronchial or conducting a major determinant of indoor airborne particle behav- airways (TB), and the pulmonary or gas-exchange ior (Nazaroff, 2004). The behavior of larger particles is region (P). The information presented in this figure strongly influenced by their mass. Gravitational set- reflects two dominant characteristics of the system. tling and inertial impaction are important deposition First, the three regions of the respiratory tract are mechanisms. The smaller particles in this size range fol- exposed to bioaerosol particles sequentially. For the low airstreams more closely. Transport mechanisms largest particles considered, the high deposition effi- leading to the departure of smaller particles from air streamlines include Brownian motion. Generally, the efficiency of filtration, the likelihood of deposition 0.8 somewhere in the respiratory tract, and the rate of NOPL (head) region deposition onto indoor surfaces are all considerably 0.6 smaller for particles with diameters in the range 0.1–1 lm as compared with particles with diameters in 0.4 – the range 1 10 lm. 0.2 Whole microbial agents vary widely in size and mainly follow this pattern: Viruses are much smaller 0 0.2 than bacterial cells or endospores, which are smaller TB (tracheobronchial) region than fungal spores. For example, the influenza A virion 0.1 ~ is 0.1 lm in size (Noda, 2012). The cells of Staphylo- 0.3 0 coccus aureus, a common opportunistic pathogenic Deposition fraction (—) P (pulmonary) region species, have diameters of ~ 1 lm (http://textbookof- 0.2 bacteriology.net/staph.html). Common indoor fungal spores have been measured to have aerodynamic diam- 0.1 ~ eters of 1.8 lm(Cladosporium cladosporioides), 0 ~ 2.2 lm(Aspergillus fumigatus), and ~ 2.7 lm(Peni- 0.1 0.3 1 3 10 cillium melinii) (Reponen et al., 1996). Spores of other Particle diameter (µm) species such as Alternaria alternata and Epicoccum Fig. 7 Fractional particle deposition in different regions of the nigrum are considerably larger (McGinnis, 2007). respiratory tract in relation to particle size. The results are from Bioaerosol particles can be larger or smaller than the the NCRP/ITRI semi-empirical model (Yeh et al., 1996). Parti- size of whole microbial agents. For example, bacteria cle density is assumed to be 1 g/cm3. The results assume nose have been observed to occur in clusters or attached to breathing, tidal volume of 0.77 l, breathing frequency of 13/ other material such as fragments of human skin min, and a functional residual capacity of 3 l. In each frame, the deposition fraction is referenced to particle concentrations in (Davies and Noble, 1962). Fungal fragments have also the inhaled air. Total deposition fraction in the respiratory tract been measured in indoor air (Gorny et al., 2002). for a given particle size would be obtained by summing results Based on samples of outdoor air, Yamamoto et al. for the three regions

66 Indoor bioaerosol dynamics ciency in the head protects the distal airways from suring bioaerosol attributes, the availability of suit- exposure. Second, two different mechanism classes able methods remains an important constraint on control particle deposition. The behavior of the larger research progress. particles is dominated by their inertia. Larger particles Table 1 provides a summary of many methods that have a higher mass-to-drag ratio, and so the larger the have been used for bioaerosol sampling and analysis. particle, the more efficient the deposition. However, (See also the reviews by Henningson and Ahlberg, for submicron particles, inertial processes are relatively 1994 and Verreault et al., 2008). The most widely unimportant. For the smallest particles in this figure, applied methods have been culture based. These are Brownian diffusion is the dominant transport mecha- subject to the limitations noted in connection with Fig- nism. This is a slow process, important only in the ure 2. Culture-based methods offer the virtues of being smallest airways: Deposition efficiency of 0.1–1 lm relatively inexpensive, well developed, quantitative, particles is small in the head region, yet substantial in and taxon specific. Disadvantages include that only the pulmonary region. Similarly, the deposition effi- viable organisms are measured and only a subset of air- ciency increases with decreasing particle size when borne organisms is culturable. A commonly employed Brownian diffusion dominates. method of direct impaction onto an agar substrate uti- Worth noting is that the combination of these effects lizes a short sampling period, which provides only a leads to two important modes of particle deposition in snapshot of the viable organism concentration at the the pulmonary region. Not only are the smallest parti- time of sampling. Relating these results to longer term cles deposited with reasonable efficiency, but there is exposure conditions is challenging because of the high also an important mode that peaks in efficiency at degree of temporal variability, for example, associated about 3 lm in diameter, a size that is important for with occupancy and occupant activities. bacterial and fungal bioaerosol particles. Effective bioaerosol sampling and analysis for stud- ies of indoor environmental conditions must address two key challenges: specificity and temporality. No Measurement technologies method is well suited to address both of these chal- Measure what is measurable, and make measurable lenges well. The nature and significance of these issues what is not so.—Galileo Galilei varies according to the specific concern. In studies of airborne infection, specificity is essential. Pathogenic Many important aspects of indoor bioaerosols strains may be closely related to nonpathogenic organ- must be determined by experiment rather than from isms of the same species. theory. Experimental capabilities are intrinsically Some of the chemical analytes that can be used in linked to technologies available for measurement. bioaerosol studies are of interest because of their direct While many methods have been developed for mea-

Table 1 Analytical methods applied for studying indoor bioaerosols

Method/Analyte Comment References

Culture based Most indoor bioaerosol data have been collected Shelton et al. (2002); Tsai and Macher (2005); in this manner; limited to culturable species Tsai et al. (2007) and (typically) short-term sampling Microscopy Labor intensive but can generate high-quality results; Hernandez et al. (1999) combine with staining to highlight features such as metabolic competence Quantitative PCR Suitable for quantification of total fungal DNA or total Hospodsky et al. (2010) bacterial DNA using universal probes and primers High-throughput DNA sequencing Characterize bacterial and fungal taxa using PCR-amplified Pitk€aranta et al. (2008); Noris et al. (2011); DNA abundance employing universal probes and primers Hospodsky et al. (2012); Hoisington et al. (2014) Metagenomic DNA sequencing Construct genetic sequencing information for bioaerosol Yooseph et al. (2013) without amplification; has required large volumes of air to collect sufficient DNA Endotoxin Lipopolysaccharide cell wall component of the outer membrane Park et al. (2001) of Gram-negative bacteria Beta glucans Glucose polymers found in cell wall of most fungi and some bacteria Douwes (2005) Ergosterol Analog of cholesterol, found in fungi, but not plants or animals; Miller and Young (1997) suitable for total airborne characterization of fungi, not species specific 3-Hydroxy fatty acids Chemical marker of lipopolysaccharide in Gram-negative bacteria Fox et al. (2005); Liu et al. (2000) N-acetylhexosaminidase (NAHA) Fungal enzyme Rylander et al. (2010) Muramic acid Derivative of glucosamine, found in bacterial cell walls Fox et al. (2003) Fluorescence of airborne particles Real-time capability, but not biologically specific Bhangar et al. (2014) Quantitative reverse-transcriptase polymerase chain reaction Airborne viruses such as influenza A Fabian et al. (2009); Yang et al. (2011)

67 Nazaroff potential for adverse health consequences. Examples Dynamic processes include endotoxin and (1?3)-b-D-glucans. Other ana- lytes, such as ergosterol or muramic acid, are not of This section summarizes evidence concerning the direct health concern, but rather can be valuable as dynamic processes that influence indoor bioaerosol lev- quantitative indicators of broad bioaerosol categories. els. The emphasis is on those processes depicted sche- Dynamic processes can influence indoor bioaerosol matically in Figure 6. concentrations by an order of magnitude or more over time scales that are as short as minutes. Consequently, Air exchange: source and sink it is important to have measurement tools that permit sampling and analysis with high time resolution. There Building air exchange is the replacement of indoor air are no methods for bioaerosol sampling and analysis with outdoor air. Air exchange is needed to limit the that are suitable for routine research application, that accumulation of carbon dioxide and other bioeffluents are highly specific, and that exhibit good time resolu- from human occupants. It is also used to limit the con- tion. Because of these limitations, process-oriented centrations of pollutants emitted from inanimate studies that are discussed in the next section have indoor sources. When outdoor air is uncontaminated, largely used abiotic particles as surrogates. The recent then increasing the air exchange rate consistently advent of real-time fluorescence-based instruments improves indoor air quality. However, when climate enhances capabilities for studying dynamic behavior of conditions are not comfortable, then the air exchange bioaerosols. These instruments offer excellent time and rate is limited to avoid excessive energy use. In many size resolution, but typically provide no biological circumstances, outdoor air is polluted to levels that specificity, even at the level of differentiating between exceed health-based standards. In these cases, by intro- bacteria and fungi. ducing pollutants from outdoor air, air exchange can be an important contributor to indoor pollution. Building air exchange may be divided into three Diseases associated with bioaerosol exposure modes: Infiltration refers to the uncontrolled leakage ...there is a growing body of data in support of the of air through the building envelope; natural ventila- conclusion that air transmission within enclosed tion occurs through windows and other designed open- spaces plays an important role in the communication ings; and fans induce mechanical ventilation. Buildings of many bacterial and viral diseases, especially those generally leak, so infiltration is regularly a contributor of the respiratory tract.—Robertson (1943) to air exchange. Although practices vary and are changing with time, it is common in the United States There are two noteworthy points to make about for air exchange in single-family dwellings to occur by the quote from Robertson: (i) It was published more a combination of natural ventilation plus infiltration. than 70 years ago and (ii) the role of airborne routes Mechanical ventilation plus infiltration is common in in the transmission of disease remains controversial commercial buildings. today for many infectious agents. Table 2 provides a The air exchange rate (AER) is an important metric list of many diseases for which there is published evi- for characterizing air quality aspects of buildings. This dence indicating that airborne exposure indoors measure is the volume-normalized flow rate of air from makes a meaningful contribution to the occurrence the building to outdoors. As depicted in Figure 6, the or spread of disease. AER would be (QN+QL)/V for the residential sche-

Table 2 Some diseases for which a contribution to transmission is associated with indoor bioaerosols

Disease Microbial agent Taxa Reference

Chickenpox Varicella zoster virus Virus Gustafson et al. (1982) Cold (common) Rhinovirus Virus Heikkinen and J€arvinen (2003) Gastroenteritis Norovirus Virus Marks et al. (2003) Influenza Influenza virus A Virus Tellier (2009) Legionnaires’ disease Legionella pneumophila Bacteria Fields et al. (2002) Measles Measles virus Virus Bloch et al. (1985) Pneumonia Streptococcus pneumoniae Bacteria Hoge et al. (1994) Pulmonary disease Non-tuberculous mycobacteria (NTM) Bacteria Thomson et al. (2013) SARS SARS coronavirus Virus Yu et al. (2004) Smallpox Variola major or Variola minor Virus Wehrle et al. (1970) Staphylococcal infection Staphylococcus aureus Bacteria Mortimer et al. (1966) Tuberculosis Mycobacterium tuberculosis Bacteria Riley (1974) Whooping cough Bordetella pertussis Bacteria Warfel et al. (2012)

68 Indoor bioaerosol dynamics matic (i) and (QM+QL)/V for the commercial schematic range that spans about an order of magnitude, from (ii). Figure 8 presents a summary of AER data from 0.2 to 3 per hour. The BASE study of ~ 100 commer- two large studies conducted in the United States. The cial buildings (Figure 8b) shows a similar central ten- measurements in residences (Figure 8a) show a median dency and a somewhat larger range, especially at the in the approximate range of 0.5–1 per hour. Consider- high end of the distribution. Each of these data sets ing individual households, most of the data lie within a conforms reasonably well to a lognormal distribution. The AER sets a time scale for one of the main removal processes for impurities in indoor air. An (a) 10 AER of 0.5–1 per hour means that any airborne species Los Angeles, CA removed primarily by air exchange will have a charac- Elizabeth, NJ Houston, TX teristic residence time of one to two hours in the build- ing air.

3 Bioaerosol deposition onto room surfaces Across the aerodynamic diameter range of 0.1–10 lm, particle deposition onto room surfaces is an important 1 fate. In equations (1) and (2), the deposition loss rate is parameterized by a size-dependent rate constant, b. The importance of deposition as a removal mechanism

Air-exchange rate (per h) Air-exchange for airborne bioaerosol particles can be explored by 0.3 comparing b to the air exchange rate. Lognormal fits: Figure 9 presents some data on size-dependent parti- CA: GM = 1.00; GSD = 2.11 NJ: GM = 1.09; GSD = 2.09 cle loss rates by means of deposition to room surfaces. TX: GM = 0.59; GSD = 2.68 These data show that, for particles in the size range – ~ – 0.1 0.5 1 lm, the deposition loss rate coefficient is 0.2 1 2 5 10 25 50 75 90 9598 99 0.3 per hour, a value that is comparable to the lower Cumulative frequency (%) end of range of air exchange rates discussed in the pre- (b) vious subsection. Under low air exchange conditions, deposition of these submicron particles is competitive 10 average measured air-exchange rate (96 sites) lognormal fit (GM = 1.14 per h; GSD = 2.89) with air exchange as a removal process, but in indoor 6

3

10 Furnished room 1 Fan off 5.4 cm/s 0.6 14.2 cm/s 3 19.1 cm/s 0.3 Outdoor air-exchange rate (per h)

0.1 1 2 5 10 25 50 75 90 95 98 Cumulative percentile (%)

Fig. 8 (a) Air exchange rates (AER) measured in samples of 0.3 houses in Los Angeles, CA (N = 105), Elizabeth, NJ (N = 96), and Houston, TX (N = 100). For each house, one or two mea-

surements of air exchange rate were taken over a few-day period coefficient rate (per h) Deposition loss using perfluorocarbon tracers. Analysis is based on treating each 0.1 house as a single unit; in cases where two AER measurements 0.5 1 3 10 were taken, the results were averaged. Yamamoto et al. (2010) Particle diameter (µm) have also published an analysis of these data. (Data source: https://riopa.aer.com/login.php.). (b) Air exchange rates (AER) Fig. 9 Particle deposition loss rate coefficient (b) measured in a measured in 96 US commercial buildings during the US Envi- 14-m3 room as a function of particle diameter. Three of four ronmental Protection Agency’s BASE Study. At each site, up to experiments were conducted with four small fans operating to four determinations of outdoor AER were made over periods of induce different degrees of air motion. As noted in the key, the several hours each, while the building ventilation system was mean measured speed in the room with the fans on ranged operating. All measurements for any given site were averaged; between 5.4 and 19.1 cm/s. The room was furnished with carpet- the distribution of the averaged results is presented in the figure. ing, chairs, table, bookcase, and curtains. Source: Thatcher (Data source: Persily and Gorfain, 2008) et al. (2002)

69 Nazaroff spaces with high air exchange rates, deposition of these 80 9 0.03 = 2.4 CFU/m2/h. This process would repre- smaller particles is less important than air exchange. sent a small contribution to the total loss rate of fungi For larger particles, in the range 3–10 lm in diameter, from indoor air, but could be an important step in the the deposition loss rate coefficient is much higher, in process of fungal colonization of walls. Such coloniza- the range 2–10 per hour. For these larger particles, tion would also require sufficient quality and quantity deposition is an important mechanism influencing the of nutrients, appropriate surface pH, and sufficient fate of bioaerosols even for buildings with relatively water activity in the substrate (Pasanen et al., 1992). high air exchange rates. The data in Figure 9 also illustrate that enhanced air Bioaerosol sources: outdoor air movement increases the rate of particle deposition, at least over the range of conditions studied (up to air ...the atmosphere is thronged with travellers: speeds of about 20 cm/s). The effect of such air move- microbes using the wind, speaking teleologically, as ment is to more rapidly deliver particles to surfaces to a convenient transport from one place to another. which they deposit. These air speeds are too low to Travellers mostly performing quite short journeys.— cause particles to be resuspended from surfaces onto Gregory (1971) which they had previously deposited. It is worthwhile to differentiate among the sources Most of the particle loss reflected in Figure 9 is that contribute to indoor bioaerosols. Such differentia- attributable to gravitational settling onto upward fac- tion can improve the basis for understanding concen- ing surfaces. However, some deposition also occurs to trations, exposures, and influencing factors. It also vertical and even downward facing surfaces. Figure 10 serves as a basis for engineering interventions to alter shows results from a study that measured particle exposures. deposition to chamber walls under stirred conditions. Among the major categories that can contribute to The deposition velocity, plotted on the vertical axis of indoor bioaerosols is air exchange-induced supply the figure, multiplied by the airborne concentration from outdoor air. From equations (1) and (2), the rate yields the deposition flux. So, for example, if we of supply of bioaerosol material from outdoors is rep- consider as a typical indoor air value for viable fungal resented either by the term (Q + pQ )C (for the resi- spores of 80 CFU/m3 (Figure 1) and assume they are N L o dential schematic, Figure 6a) or by the term [(1g ) associated with 6-lm particles for which the deposition M Q + pQ ]C (for the commercial building schematic, velocity is 30 mm/h (= 0.03 m/h), then the resulting M L o Figure 6b). We have discussed the various air exchange deposition flux to vertical walls would be flow rates and—to an extent—the outdoor bioaerosol concentration, Co. The new terms to address here are the fractional penetration of bioaerosols along with 300 infiltration flow, p, and the efficiency of bioaerosol par- Model line (fit) ticle removal in the mechanical ventilation supply flow, V = min (V , V ) g . An important point is that both of these efficiency 100 dd1d2 M V d 2.20 terms are expected to exhibit significant particle-size d1 = 0.75 p V d –0.51 dependence. Consequently, the relationship between d2 = 7.5 p d µ V outdoor bioaerosol concentrations and the source of 30 p [=] m; d [=] mm/h indoor bioaerosols can vary with particle size. The main principles that govern p and gM are well 10 understood. However, evaluation of proper values of these parameters for any particular building remains a challenge because uncontrolled and unknown details Deposition velocity (mm/h) velocity Deposition 3 of construction and operation can influence the out- comes. Consider particle penetration through leakage paths. 1 1 35 10 As air flows into a building through a leak in the enve- Particle diameter (µm) lope, particles suspended in that airstream may contact a surface bounding the leak, adhere, and be lost from Fig. 10 Deposition velocity (vd) measured to the vertical wall of 3 the airstream. The penetration factor, p, represents that a 2-m chamber as a function of particle diameter (dp) (Lai and portion of particles in the outside air that make it Nazaroff, 2005). The deposition velocity is linked to the loss rate through the leaks to enter the indoor environment. coefficient as follows: bw = vd Sw/V, where bw is the contribution to the total loss rate coefficient attributable to deposition on the Large particles may deposit because of gravitational 1 walls (h ), Sw is the wall area, and V is the room volume. The settling or inertial impaction. Small particles may model equations are linear regressions to the log-transformed deposit because of Brownian motion. Figure 11 pre- data, utilizing three measured points (vd1) and six measured sents the results of model calculations showing how the points (vd2), respectively

70 Indoor bioaerosol dynamics

1 mm x 4 Pa 1 mm x 10 Pa 100 Crack height x 0.25 mm x 4 Pa 0.25 mm x 10 Pa Particle diameter pressure difference 90 0.1 mm x 4 Pa 0.1 mm x 10 Pa 3–10 µm 80 1–3 µm

1 70 0.3–1 µm

60 0.8 50

0.6 40

30 Single-pass efficiency (%) 0.4 20

10 Penetration fraction (—) 0.2 0 Crack length MERV 4 MERV 7 MERV 11 MERV 13 MERV 16 = 3 cm 0 Minimum efficiency reporting value

0.1 0.3 1 3 10 Fig. 12 Specified minimum single-pass particle removal effi- ciency of fibrous filters used in ventilation systems as a function Particle diameter (µm) of particle size (three shading styles) and MERV rating (hori- zontal axis). Source: Azimi and Stephens (2013) Fig. 11 Model prediction for size-dependent particle penetra- tion (p) through a crack in the building envelope. Source: Liu and Nazaroff (2001) degree of protection of the indoor environment from outdoor bioaerosols. In mechanical ventilation systems, some bioaerosol penetration factor varies with particle size for idealized deposition can occur on surfaces other than the filters, crack geometry. Different parameter values are including ducts and heat-exchanger fins. Evidence sug- assumed for the indoor–outdoor pressure drop (4 or gests that such deposition is size dependent and much 10 Pa) and the height of the crack (0.1, 0.4, and higher for the larger bioaerosol particles than for the 1.0 mm). An important message from this figure is that smaller ones (Sippola and Nazaroff, 2003; Waring and cracks must be quite fine for any meaningful attenua- Siegel, 2008). This deposition process might contribute tion of the airborne particles during infiltration. Specif- to meaningful rates of removal from airstreams in ically, penetration is essentially complete across the full some circumstances. However, a more significant con- diameter range 0.1–10 lm for any crack whose mini- cern is the risk of fouling and the degradation of sys- mum dimension exceeds ~ 1 mm (given a 4 Pa or tem hygiene. higher pressure drop and assuming that the flow chan- nel through the crack is no longer than 3 cm). The dis- Bioaerosol sources: indoor emissions tribution of leak dimensions in any real building is not known. However, it seems likely that a normal case An important and challenging feature of indoor bio- would feature most of the infiltrating air flowing aerosol dynamics is characterizing indoor emission through cracks larger than 1 mm in minimum dimen- sources. From a systematic research perspective, a core sion. Hence, there is an expectation that p ~ 1 for bio- advantage of source characterization is that it is likely aerosol particles. to provide more generalizable information than would In the case of a mechanically ventilated building, a phenomenological studies of concentrations or other second important parameter is the efficiency of a parti- outcome variables. The information sought in source cle filter in the flow path connecting outdoor air to the characterization would include these factors for any air supply registers. Figure 12 illustrates two important particular bioaerosol analyte: the quantity emitted per points about filter efficiency. First, it is highly variable time, the size distribution of particles with which the with filter quality, ranging from low for filters with a emitted analyte is associated, and the important MERV 4 rating to high for MERV 13 or MERV 16 parameters that influence the emission rate. Depending filters. Second, filter efficiency can vary markedly with on the particular source, experiments to investigate particle size. Across the range that is pertinent for bio- emissions might suitably be conducted in a bench-scale aerosols, the filtration efficiency tends to be higher for laboratory apparatus, in a room scale chamber, or larger particles than for smaller particles. A mechani- through controlled field monitoring. cally ventilated building with a high mechanical venti- There are many potential indoor sources of bioaero- lation to infiltration flow rate ratio (QM/QL ≫ 1) and a sols. Research that characterizes emissions is still in a high-efficiency filter (1gM 1) can provide a high relatively early stage of development, with limited 71 Nazaroff quantitative information available for most sources. In 10 9 more intensely. Bernard et al. (1965) observed the following paragraphs, several studies that have that the shedding of airborne bacteria from humans investigated indoor bioaerosol sources are highlighted. was markedly elevated during a period 10–45 min Primary goals include indicating the breadth of source after showering. The application of lanolin or alcohol types that have been investigated and providing entry to the skin reduced the effect, as did wearing a tightly points into the literature for those interested in deeper woven fabric. Hall et al. (1986) did not find this show- study. ering effect to occur. However, they did observe that Among the merits of quantifying emissions from ‘men dispersed many more bacteria than women’ and interior sources are these. First, such quantification that the emissions rate could be considerably lowered allows for the assessment of the relative importance of through the application of skin lotion. They also indoor vs. outdoor sources as contributors to the noted that friction between skin and clothing indoor burden. Examine the numerators of equa- appeared to be an important factor inducing the tions (1) and (2), and note that in each case, there is a release. term that is proportional to the outdoor concentration A second mechanism by which human activities may and a term (E) that reflects indoor emissions. If, for a contribute to bioaerosol loading is through resuspen- particular bioaerosol component of concern, the sion of biological particles that had previously settled indoor emissions term is small compared to the out- on flooring or on other upward facing surfaces (Qian door source, then we can safely focus our attention on et al., 2014). Using a mechanical stepping apparatus the outdoor environment transported via air exchange that mimicked walking combined with artificially as the dominant contributor to indoor levels. Con- seeded flooring materials, Tian et al. (2014) found that versely, if the indoor emission source greatly exceeds the fractional resuspension of abiotic particles was the term associated with the outdoor level, then we can ~ 105 and ~ 104 per step for particles with diameters focus on indoor emissions and scale down our atten- 1–3 lm and 5–10 lm, respectively. Goebes et al. tion to the outdoor air as a significant contributor. (2011) documented that foot traffic contributes to Second, if the indoor emissions are well-character- measureable concentrations of airborne Aspergillus; ized, then (provided we have adequate additional infor- resuspension from flooring was demonstrated to be an mation about the indoor environment) we can estimate important mechanism. the contribution of the indoor emissions to airborne A few recent studies have aimed to quantify size- concentrations. Equations (1) and (2) illustrate the resolved biological particle emission rates associated relationship, and the additional required information with human occupants using modern analytical meth- is that needed to quantify the appropriate denomina- ods. Qian et al. (2012) studied a university classroom tor. with bioaerosol sampling using a cascade impactor fol- lowed by quantitative PCR applied with universal bac- Human occupants. Adult man carries 1012 microbes terial and fungal primers. By applying material balance associated with his epidermis and 1014 microbes in his to the indoor and outdoor data under room-occupied alimentary tract. ... The 1013 cells in his body are a and unoccupied conditions, they inferred the per-occu- distinct numerical minority of the total being that we pant effective emission rates of bacteria and fungi. For call man. If we abandon anthropomorphism for the example, the bacterial emission rates were determined microbic view, we must admire the efficiency of these to average 37 million genome copy numbers per hour microbes in using man as a vehicle to further their per occupant. Particle size conformed reasonably well own cause.—Luckey (1972) to a lognormal distribution with a geometric mean In the context of better understanding and control- aerodynamic diameter of 4.4 lm and a geometric stan- ling airborne infection, bioaerosol emissions from dard deviation of 1.39. Bhangar et al. (2014) used real- humans have been a topic of concern since at least the time particle detection to quantify size-resolved emis- 1940s. For example, Duguid (1946) experimentally sions of fluorescent biological aerosol particles also in assessed the size distribution of particles and droplets a university classroom. They concluded that the modal emitted by sneezing, coughing, and talking. The likeli- size was in the range 3–4 lm and that the average emis- hood that such particles would contain bacteria was sion rates were 1.6 million particles per person per hour estimated based on their prevalence in respiratory flu- during the main portion of lecture classes plus 0.8 mil- ids. Duguid and Wallace (1948) experimentally inves- lion particles per person emitted during transitions tigated the ‘bacterial contamination of air produced between classes. by liberation of dust from the skin and personal cloth- Advanced analytical methods and ongoing con- ing during bodily movement’. Using culture-based cerns about the spread of infectious respiratory dis- analysis methods, they found that dust particles carry- eases have motivated a renewed effort to study ing bacteria were liberated at a rate of about 1000 per bioaerosol release from the nose and mouth. Fennelly minute from a ‘person making slight movements’ and et al. (2004) measured the size distribution of cough- that marching liberated culturable bacteria about generated particles containing culturable Mycobacte-

72 Indoor bioaerosol dynamics rium tuberculosis emitted from patients with pulmo- in house dust in homes in which household organic nary tuberculosis. Stelzer-Braid et al. (2009) collected waste was separately stored indoors. respiratory emissions from 50 subjects while breath- Floor dust or settled dust can be richly microbial ing, talking, and coughing. They detected one or (Rintala et al., 2012), and this can serve as a bioaero- more of nine respiratory viruses in 21 of 33 subjects sol source through resuspension (Qian et al., 2014). who had symptoms of upper respiratory tract infec- In a laboratory study, substantial bacteria and mold tions and only in four among 17 asymptomatic sub- emission rates were observed from the operation of jects. Gralton et al. (2013) conducted an analogous household vacuum cleaners (Veillette et al., 2013). study that focused on breathing and coughing, This finding is not a surprising: Vacuum cleaner use included children among the subjects, and investi- was already demonstrated to be a potent short-term gated the size distribution of the emitted particles generator of airborne particulate matter, both in the and droplets. They concluded that ‘individuals with exhaust air (because of incomplete filtration) and symptomatic respiratory viral infections produce both through mechanical agitation of the floor (Corsi large and small particles carrying viral RNA on et al., 2008; Knibbs et al., 2012). The net effect of coughing and breathing’. vacuuming on indoor bioaerosol levels is not clear, Overall, human occupants are important contribu- however, as presumably vacuum cleaner use would tors to the bioaerosol burden of indoor environments. also reduce the floor load of microbes available for They shed bacteria along with their skin; they emit later resuspension. viruses from their respiratory tract; and they resuspend Indoor surfaces that are frequently moist can harbor particulate material that contains biological agents microbial growth; examples related to household from floors and other surfaces that they contact. hygiene include shower curtains (Kelley et al., 2004) and sink drains (Adams et al., 2013). Bollin et al. Moldy materials. A second potentially important emis- (1985) showed that showerheads and hot-water faucet sion source category for indoor bioaerosols is moldy use could produce aerosols containing Legionella pneu- building materials. Dampness and mold is common in mophila. Thomson et al. (2013) found non-tuberculous buildings. For example, Spengler et al. (1994) reported mycobacteria (NTM) in shower aerosols in homes of that half of surveyed households in 24 US and Cana- patients with pulmonary disease caused by NTM, a dian cities had a dampness-related condition (water finding broadly consistent with evidence from Feazel damage, water in basement, and/or mold or mildew). et al. (2009) of enhanced NTM prevalence in shower- Several laboratory studies have investigated emis- head biofilms. Toilet flushing can produce aerosol sions of bioaerosols from moldy materials. For exam- droplets, and as fecal material is rich in microbes, the ple, Gorny et al. (2001) characterized the release of toilet is a potentially important source of indoor bio- fungal spores—Aspergillus versicolor, Cladosporium aerosols, especially in connection with diarrheal dis- cladosporioides, and Penicillium melinii—from ceiling eases (Johnson et al., 2013). tiles in relation to the air speed above the surface and the vibration of the contaminated material. Seo et al. Other factors (2008) investigated the release of (1?3)-b-D-glucan from moldy ceiling tiles and gypsum board. In many Bioaerosol exposure control. Let us acknowledge that buildings, moisture intrusion or condensation occurs the goal should not be to make indoor environments in wall cavities or in other hidden spaces that may be sterile. At the same time, elevated levels of airborne coupled by airflow pathways to the occupied building pathogens are to be avoided, as are excessive levels of interior. Muise et al. (2010) demonstrated experimen- many bioaerosol attributes. And in certain circum- tally that mold spores could penetrate effectively stances, we should be particularly concerned about through wall service outlets. That finding is consistent protecting vulnerable people from even ordinary bio- with expectations (see Figure 11) as most fungal aerosol exposure, such as individuals who are immuno- spores are smaller than 10 lm in diameter and as compromised. cracks and gaps between a wall cavity and the indoor How can bioaerosol control be achieved? Conceptu- space would commonly be larger than 1 mm in mini- ally, in the context of the material balance described mum dimension. earlier, there are two broad options: (i) reduce sources and/or (ii) increase removal rates. Equations (1) and Housekeeping and hygiene. A third bioaerosol source (2) provide a basis for quantitatively estimating the category that is potentially important indoors is related benefit of a control measure. to housekeeping and hygiene. Several examples are Among the options for source control are to keep briefly mentioned here. Davies and Noble (1962) dem- indoor environments dry, to maintain good hygienic onstrated that bedmaking increased the airborne con- conditions in ventilation systems, to apply effective fil- centrations of skin scales and bacteria. Wouters et al. tration on mechanical supply ventilation, and to use (2000) reported increased levels of microbial markers masks in the event of respiratory illness.

73 Nazaroff

Regarding removal processes, the primary control mechanical ventilation practice uses air diffusers alternatives are three, which appear in the denominator designed to promote rapid mixing, other concepts aim of equations (1) and (2): (i) increase the outdoor air to deliberately exploit incomplete mixing as a basis to ventilation rate (typically either QN or QM); (ii) use re- improve efficiency. Such methods include displacement circulating air filtration (i.e., introduce or enhance gR ventilation (Novoselac and Srebric, 2002) and person- QR, where gR is the filtration efficiency, as illustrated in alized ventilation (Melikov, 2004). Understanding Figure 12, and QR is the recirculating flow rate how sources relate to concentrations in the breathing through the filter; or (iii) enhance the rate of deposition zone of occupants in cases like these cannot be accu- or degradation of the bioaerosol attribute (thereby rately accomplished using a well-mixed analysis frame- increasing b). Miller-Leiden et al. (1996) explored the work; instead, more sophisticated methods are effectiveness of in-room recirculating filtration for con- required, such as approaches based on computational trolling the transmission of tuberculosis. Cheng et al. fluid dynamics (Chen, 2009). Experimentally, investi- (1998) have evaluated the use of portable air cleaners gations in such conditions require methods that can for reducing indoor concentrations of fungal spores. accommodate spatially varying contaminant Ultraviolet germicidal irradiation can be applied to concentrations (Bjørn and Nielsen, 2002; Brohus and reduce the infectivity of air without actively removing Nielsen, 1996). the bioaerosol particles (Reed, 2010). The control measures described in the preceding paragraphs all aim to reduce the airborne concentra- Conclusion tion of the bioaerosol agent, denoted C in equa- Although study of the normal human lung microbiome tions (1) and (2). A complementary approach is to is still in its early stages, the bulk of published evi- provide susceptible individuals with personal protec- dence demonstrates that phylogenetically diverse — — tive equipment, which if performed well can reduce microbial communities in the lungs of healthy humans the inhalation intake by an order of magnitude or more can be detected using high throughput sequencing.— for a given airborne concentration. Nicas (1995) Beck et al. (2012) presents an illustrative example for the case of respira- tory protection of healthcare workers against Myco- In a recent review, Grice and Segre (2012) articulate bacterium tuberculosis bacilli. important ways in which microorganisms modulate human health: ‘The human microbiome is a source of Airborne growth and decay. The embodiment of the genetic diversity, a modifier of disease, an essential com- material balance principle in equations (1) and (2) does ponent of immunity, and a functional entity that influ- not account for microbial reproduction or for any ences metabolism and modulates drug interactions’. other bioaerosol growth process during the period of They then summarize what has recently been learned suspension in the indoor air. There is some evidence about the microbiology of the human gastrointestinal supporting the possibility of airborne microbial life in tract, the oral cavity, the reproductive tract, and the skin. the atmosphere, as reviewed by Womack et al. (2010). However, they do not comment on the microbiology of However, indoors, where the airborne residence time is the respiratory tract. Beck et al. (2012), in their review, limited to about a few hours or less, such processes note that, ‘although the lungs were classically believed to have not been demonstrated and seem unlikely to be be sterile, recently published investigations have identi- important. fied microbial communities in the lungs of healthy Infectious agent viability may decay at a significant humans’. Lax et al. (2014) document that the microbes rate when airborne (Weber and Stilianakis, 2008). found in a home are distinctively related to the people Incorporating the effects of such processes into model who live in that home and that ‘after a house move, the equations can be achieved through the decay parame- microbial community in the new house rapidly con- ter, b, in equations (1) and (2). verged on the microbial community of the occupants’ former house, suggesting rapid colonization by the fam- Transport and mixing. Throughout this study, it has ily’s microbiota’. What is not yet clear, but seems plausi- been assumed that an indoor environment can be rep- ble, is whether aspects of the human microbiome are resented as a well-mixed space. At the level of an indi- strongly influenced by microbiological and other condi- vidual room, the size of a typical bedroom or private tions in inhabited indoor spaces. office, that description is often but not always reason- Indoor bioaerosol behavior might play an important able. An entire residence, or a large building, might be role in these stories. Almost certainly, the most impor- appropriately represented as a network of well-mixed tant exposures of the human lung to environmental rooms or zones, interconnected by airflow paths (Feu- microorganisms occur via inhalation of bioaerosols. Fur- stel, 1999). thermore, most of the air that is inhaled by humans is For some situations, the well-mixed conceptualiza- indoor air. Bioaerosols also are important vectors trans- tion is inappropriate. For example, while much of porting microorganisms from outdoors to indoors and

74 Indoor bioaerosol dynamics from one indoor surface to another. It is feasible that air- and evolving knowledge of microbial ecology. We can borne transport to and deposition on the human enve- anticipate continuing opportunities from cooperation lope influences skin microbiota (Adams et al., 2013). between scholars from these domains. As we gain Relative to the complexity and importance of the empirical knowledge, it will be important to seek gen- subject of indoor bioaerosol dynamics, our under- eralizable understanding from the specifics of particu- standing is not yet mature. One might anticipate fun- lar investigations. We will never measure everything! damental paradigm shifts to occur as our knowledge Research that focuses on processes and that is framed grows. Our ability to ask and answer incisive questions in the context of well-established mechanistic knowl- should improve. edge can be a valuable way to proceed. In this review, Although the gap between what we know and the principle of material balance has served to structure what we would like to know is quite large, our a relationship between bioaerosol processes and indoor current knowledge is substantial. For example, in concentrations, an important intermediate outcome. considering the dynamic behavior of airborne bio- Further studies can be fruitfully pursued to better logical particles in buildings, indoor aerosol science understand each of the input parameters that appear in provides a good starting point. Mechanistically, bio- the material balance equations. Research could also be aerosol particles behave like their abiotic counter- undertaken to test the accuracy of and to refine as nec- parts. Particle size is a primordial determinant of essary the model equations themselves. Benefits would behavior, and bioaerosol particles are mainly found especially be anticipated from studies to better charac- in the aerodynamic diameter range 0.1–10 lm. Aer- terize and quantify indoor bioaerosol emission sources osol science has developed powerful tools and theo- and the influencing factors. This particular process has ries regarding emission processes, airborne behavior, especially large influence on outcomes, it is difficult to and fate. Much of the understanding developed characterize without direct experimental measurement, from aerosol science can be applied to indoor bio- and it is subject to enormous variability. aerosol dynamics. The ultimate goal for improving knowledge about As we proceed in studying the microbiology of the indoor bioaerosol dynamics is to contribute to a stronger indoor environment, we should maintain a central focus knowledge base for the design, construction, operation, on people. Human occupants are a major source of and maintenance of healthful buildings. This review has indoor bacteria. Our activities influence the emissions and focused on the natural science and engineering aspects of fate of other bioaerosols as well. The outcomes of primary the theme. This review has not addressed the complemen- concern are centered on human health and welfare. tary and comparably complex theme of how human Technological advances enable the acquisition and health is influenced by bioaerosol exposure. Because we analysis of microbial data of phenomenal richness. As are at an early stage of understanding both themes and we conduct research with the new tools, it is important because of their complementarity, progress toward the that we not lose sight of the knowledge gained by prior ultimate goal would benefit from the development of syn- generations of scholars. Public health engineering stud- ergistic interactions between the respective research com- ies conducted in the 1940s through 1970s, for example, munities. That is a challenging proposition for several on the theme of infection control in healthcare settings, reasons, including largely independent educational tracks contain particularly important insights that remain rel- and funding agencies that—at least in the United States evant to our current research agendas. —do not embrace the nexus between the built environ- The diversity and complexity of the system will con- ment and health. The first steps to overcoming any chal- tinue to pose great challenges for studies of indoor bio- lenge are recognizing its importance and making a aerosol dynamics, especially in efforts to link commitment to the attempt. Perhaps the recent efforts to microbiological abundance to exposure and to health improve knowledge about the indoor microbiome and outcomes. The measurement limitations continue to be about the relationship between the human microbiome daunting. Dust is an attractive sampling medium of and health are establishing fertile ground for future syn- questionable exposure relevance. Culture-based analysis ergistic efforts to better understand the relationship methods have limited scope. Microscopic methods, between the indoor microbiome and health. DNA-based analysis, and methods using chemical mark- ers are best suited for making time-integrated measure- Acknowledgements ments with sampling periods of hours. These methods are not well suited for studying dynamic processes. The Partial support for this work was provided by a grant fluorescent particle-sampling and analysis methods have from the Alfred P. Sloan Foundation in support of the the advantage of offering excellent time and particle-size Berkeley Indoor Microbial Ecology Research Consor- resolution. However, these methods lack specificity. tium (BIMERC). Additional support was provided by In recent years, we have seen benefits from efforts to the Republic of Singapore’s National Research Foun- fuse concepts and approaches from the indoor environ- dation through a grant to the Berkeley Education Alli- mental sciences with the rapidly developing techniques ance for Research in Singapore (BEARS) for the

75 Nazaroff

Singapore-Berkeley Building Efficiency and Sustain- research and education in Singapore. Thanks to Nao- ability in the Tropics (SinBerBEST) Program. BEARS michi Yamamoto, Jordan Peccia, and two anonymous has been established by the University of California, reviewers for constructive comments and suggestions Berkeley, as a center for intellectual excellence in that contributed to improvements in the final article.

References Adams, R.I., Miletto, M., Taylor, J.W. and side of homes, Aerosol Sci. Technol., 43, Fields, B.S., Benson, R.F. and Besser, R.E. Bruns, T.D. (2013) The diversity and dis- 699–713. (2002) Legionella and Legionnaires’ dis- tribution of fungi on residential surfaces, Chen, Q. and Hildemann, L.M. (2009b) The ease: 25 years of investigation, Clin. PLoS One, 8, e78866. effects of human activities on exposure to Microbiol. Rev., 15, 506–526. Azimi, P. and Stephens, B. (2013) HVAC fil- particulate matter and bioaerosols in resi- Fisk, W.J., Lei-Gomez, Q. and Mendell, tration for controlling infectious airborne dential homes, Environ. Sci. Technol., 43, M.J. (2007) Meta-analyses of the associa- disease transmission in indoor environ- 4641–4646. tions of respiratory health effects with ments: predicting risk reductions and Cheng, Y.S., Lu, J.C. and Chen, T.R. (1998) dampness and mold in homes, Indoor Air, operational costs, Build. Environ., 70, Efficiency of a portable indoor air cleaner 17, 284–296. 150–160. in removing pollens and fungal spores, Fox, A., Harley, W., Feigley, C., Salzberg, Bartlett, K.H., Kennedy, S.M., Brauer, M., Aerosol Sci. Technol., 29,92–101. D., Sebastian, A. and Larsson, L. (2003) van Netten, C. and Dill, B. (2004) Evalu- Corsi, R.L., Siegel, J.A. and Chiang, C. Increased levels of bacterial markers and ation and determinants of airborne bacte- (2008) Particle resuspension during the CO2 in occupied school rooms, J. Envi- rial concentrations in school classrooms, use of vacuum cleaners on residential ron. Monitor., 5, 246–252. J. Occup. Environ. Hyg., 1, 639–647. carpet, J. Occup. Environ. Hyg., 5, Fox, A., Harley, W., Feigley, C., Salzberg, Beck, J.M., Young, V.B. and Huffnagle, 232–238. D., Toole, C., Sebastian, A. and Larsson, G.B. (2012) The microbiome of the lung, Davies, R.R. and Noble, W.C. (1962) Dis- L. (2005) Large particles are responsible Transl. Res., 160, 258–266. persal of bacteria on desquamated skin, for elevated bacterial marker levels in Bernard, H.R., Speers, R., O’Grady, F.W. Lancet, 280, 1295–1297. school air upon occupation, J. Environ. and Shooter, R.A. (1965) Airborne bacte- Douwes, J. (2005) (1 ? 3)-b-D-glucans and Monitor., 7, 450–456. rial contamination: investigation of respiratory health: a review of the Goebes, M.D., Boehm, A.B. and Hilde- human sources, Arch. Surg., 91, 530–533. scientific evidence, Indoor Air, 15, 160– mann, L.M. (2011) Contributions of foot Bhangar, S., Huffman, J.A. and Nazaroff, 169. traffic and outdoor concentrations to W.W. (2014) Size-resolved fluorescent Douwes, J., Thorne, P., Pearce, N. and Hee- indoor airborne Aspergillus, Aerosol Sci. biological aerosol particle concentrations derik, D. (2003) Bioaerosol health effects Technol., 45, 352–363. and occupant emissions in a university and exposure assessment: progress and Gorny, R.L., Reponen, T., Grinshpun, S.A. classroom, Indoor Air, 24, 604–617. prospects, Ann. Occup. Hyg., 47, 187– and Willeke, K. (2001) Source strength of Bjørn, E. and Nielsen, P.V. (2002) Dispersal 200. fungal spore aerosolization from moldy of exhaled air and personal exposure in Duguid, J.P. (1946) The size and the dura- building material, Atmos. Environ., 35, displacement ventilated rooms, Indoor tion of air carriage of respiratory droplets 4853–4862. Air, 12, 147–164. and droplet nuclei, J. Hyg. Cambridge, Gorny, R.L., Reponen, T., Willeke, K., Bloch, A.B., Orenstein, W.A., Ewing, W.M., 44, 471–479. Schmechel, D., Robine, E., Boissier, M. Spain, W.H., Mallison, G.F., Herrmann, Duguid, J.P. and Wallace, A.T. (1948) Air and Grinshpun, S.A. (2002) Fungal frag- K.L. and Hinman, A.R. (1985) Measles infection with dust liberated from cloth- ments as indoor air biocontaminants, outbreak in a pediatric practice: airborne ing, Lancet, 252, 845–849. Appl. Environ. Microbiol., 68, 3522–3531. transmission in an office setting, Pediat- Fabian, P., McDevitt, J.J., Lee, W.M., Gralton, J., Tovey, E.R., McLaws, M.L. and rics, 75, 676–683. Houseman, E.A. and Milton, D.K. Rawlinson, W.D. (2013) Respiratory Bollin, G.E., Plouffe, J.F., Para, M.F. and (2009) An optimized method to detect virus RNA is detectable in airborne and Hackman, B. (1985) Aerosols containing influenza virus and human rhinovirus droplet particles, J. Med. Virol., 85, Legionella pneumophila generated by from exhaled breath and the airborne 2151–2159. shower heads and hot-water faucets, environment, J. Environ. Monitor., 11, Gregory, P.H. (1971) The Leeuwenhoek Lec- Appl. Environ. Microbiol., 50, 1128–1131. 314–317. ture, 1970: airborne microbes: their sig- Brohus, H. and Nielsen, P.V. (1996) Personal Feazel, L.M., Baumgartner, L.K., Peterson, nificance and distribution, Proc. R. Soc. exposure in displacement ventilated K.L., Frank, D.N., Harris, J.K. and Lond. B Biol. Sci., 177, 469–483. rooms, Indoor Air, 6, 157–167. Pace, N.R. (2009) Opportunistic patho- Grice, E.A. and Segre, J.A. (2012) The Burge, H. (1990) Bioaerosols: prevalence and gens enriched in showerhead biofilms, human microbiome: our second genome, health effects in the indoor environment, Proc. Natl Acad. Sci. USA, 106, Annu. Rev. Genomics Hum. Genet., 13, J. Allergy Clin. Immunol., 86, 687–701. 16393–16399. 151–170. Burge, H. (2002) An update on pollen and Fennelly, K.P., Martyny, J.W., Fulton, Gustafson, T.L., Lavely, G.B., Brawner, fungal spore aerobiology, J. Allergy Clin. K.E., Orme, I.M., Cave, D.M. and Hei- E.R. Jr, Hutcheson, R.H. Jr, Wright, Immunol., 110, 544–552. fets, L.B. (2004) Cough-generated aero- P.F. and Schaffner, W. (1982) An out- Chen, Q. (2009) Ventilation performance sols of Mycobacterium tuberculosis: a new break of airborne nosocomial varicella, prediction for buildings: a method over- method to study infectiousness, Am. J. Pediatrics, 70, 550–556. view and recent applications, Build. Envi- Respir. Crit. Care Med., 169, 604–609. Hall, G.S., Mackintosh, C.A. and Hoffman, ron., 44, 848–858. Feustel, H.E. (1999) COMIS — an interna- P.N. (1986) The dispersal of bacteria and Chen, Q. and Hildemann, L.M. (2009a) Size- tional multizone air-flow and contami- skin scales from the body after showering resolved concentrations of particulate nant transport model, Energ. Buildings, and after application of a skin lotion, J. matter and bioaerosols inside versus out- 30,3–18. Hyg. Cambridge, 97, 289–298.

76 Indoor bioaerosol dynamics

Hanski, I., von Hertzen, L., Fyhrquist, N., Lai, A.C.K. and Nazaroff, W.W. (2005) Su- sampling mechanism for indoor micro- Koskinen, K., Torppa, K., Laatikainen, permicron particle deposition from tur- bial communities, Build. Environ., 45, T., Karisola, P., Auvinen, P., Paulin, L., bulent chamber flow onto smooth and 338–346. Makel€ a,€ M.J., Vartiainen, E., Kosunen, rough vertical surfaces, Atmos. Environ., Novoselac, A. and Srebric, J. (2002) A criti- T.U., Alenius, H. and Haahtela, T. 39, 4893–4900. cal review on the performance and design (2012) Environmental biodiversity, Lax, S., Smith, D.P., Hampton-Mercell, J., of combined cooled ceiling and displace- human microbiota, and allergy are inter- Owens, S.M., Handley, K.M., Scott, ment ventilation systems, Energ. Build- related, Proc. Natl Acad. Sci. USA, 109, N.M., Gibbons, S.M., Larsen, P., Sho- ings, 34, 497–509. 8334–8339. gan, B.D., Weiss, S., Metcalf, J.L., Ursell, Park, J.H., Spiegelman, D.L., Gold, D.R., Heikkinen, T. and Jarvinen,€ A. (2003) The L.K., Vazquez-Baeza, Y., Van Treuren, Burge, H.A. and Milton, D.K. (2001) common cold, Lancet, 361,51–59. W., Hasan, N.A., Gibson, M.K., Colwell, Predictors of airborne endotoxin in the Henningson, E.W. and Ahlberg, M.S. (1994) R., Dantas, G., Knight, R. and Gilbert, home, Environ. Health Perspect., 109, Evaluation of microbiological aerosol J.A. (2014) Longitudinal analysis of 859–864. samplers: a review, J. Aerosol Sci., 25, microbial interaction between humans Pasanen, A.L., Juutinen, T., Jantunen, M.J. 1459–1492. and the indoor environment, Science, and Kalliokoski, P. (1992) Occurrence Hernandez, M., Miller, S.L., Landfear, 345, 1048–1052. and moisture requirements of microbial D.W. and Macher, J.M. (1999) A com- Liu, D.L. and Nazaroff, W.W. (2001) growth in building materials, Int. Biode- bined fluorochrome method for quantita- Modeling pollutant penetration across ter. Biodegr., 30, 273–283. tion of metabolically active and inactive building envelopes, Atmos. Environ., 35, Persily, A.K. and Gorfain, J. (2008) airborne bacteria, Aerosol Sci. Technol., 4451–4462. Analysis of Ventilation Data from the 30, 145–160. Liu, L.J.S., Krahmer, M., Fox, A., Feigley, U.S. Environmental Protection Agency Hoge, C.W., Reichler, M.R., Dominguez, C.E., Featherstone, A., Saraf, A. and Building Assessment Survey and E.A., Bremer, J.C., Mastro, T.D., Hen- Larsson, L. (2000) Investigation of the Evaluation (BASE) Study, NISTIR dricks, K.A., Musher, D.M., Elliott, J.A., concentration of bacteria and their cell 7145-Revised. National Institute of Facklam, R.R. and Breiman, R.F. (1994) envelope components in indoor air in two Standards and Technology, Washing- An epidemic of pneumococcal disease in elementary schools, J. Air Waste Man- ton, DC. an overcrowded, inadequately ventilated age., 50, 1957–1967. Pitkaranta,€ M., Meklin, T., Hyvarinen,€ A., jail, N. Engl. J. Med., 331, 643–648. Luckey, T.D. (1972) Introduction to intesti- Paulin, L., Auvinen, P., Nevalainen, A. Hoisington, A.J., Maestre, J.P., King, M.D., nal microecology, Am. J. Clin. Nutr., 25, and Rintala, H. (2008) Analysis of fungal Siegel, J.A. and Kinney, K.A. (2014) 1292–1294. flora in indoor dust by ribosomal DNA Impact of sampler selection on the char- Marks, P.J., Vipond, I.B., Regan, F.M., sequence analysis, quantitative PCR, and acterization of the indoor microbiome via Wedgwood, K., Fey, R.E. and Caul, E.O. culture, Appl. Environ. Microbiol., 74, high-throughput sequencing, Build. Envi- (2003) A school outbreak of Norwalk- 233–244. ron., 80, 274–282. like virus: evidence for airborne transmis- Qian, J., Hospodsky, D., Yamamoto, N., Hospodsky, D., Yamamoto, N. and Peccia, sion, Epidemiol. Infect., 131, 727–736. Nazaroff, W.W. and Peccia, J. (2012) J. (2010) Accuracy, precision, and McGinnis, M.R. (2007) Indoor mould devel- Size-resolved emission rates of airborne method detection limits of quantitative opment and dispersal, Med. Mycol., 45, bacteria and fungi in an occupied class- PCR for airborne bacteria and fungi, 1–9. room, Indoor Air, 22, 339–351. Appl. Environ. Microbiol., 76, Melikov, A.K. (2004) Personalized ventila- Qian, J., Peccia, J. and Ferro, A.R. (2014) 7004–7012. tion, Indoor Air, 14(Suppl. 7), 157–167. Walking-induced particle resuspension in Hospodsky, D., Qian, J., Nazaroff, W.W., Miller, J.D. and Young, J.C. (1997) The use indoor environments, Atmos. Environ., Yamamoto, N., Bibby, K., Rismani-Ya- of ergosterol to measure exposure to fun- 89, 464–481. zdi, H. and Peccia, J. (2012) Human gal propagules in indoor air, Am. Ind. Reed, N.G. (2010) The history of ultraviolet occupancy as a source of indoor airborne Hyg. Assoc. J., 58,39–43. germicidal irradiation for air disinfection, bacteria, PLoS One, 7, e34867. Miller-Leiden, S., Lobascio, C., Nazaroff, Public Health Rep., 125,15–27. Hospodsky, D., Yamamoto, N., Nazaroff, W.W. and Macher, J.M. (1996) Effective- Reponen, T., Willeke, K., Ulevicius, V., W.W., Miller, D., Gorthala, S. and Pec- ness of in-room air filtration and dilution Reponen, A. and Grinshpun, S.A. (1996) cia, J. (2015) Characterizing airborne fun- ventilation for tuberculosis infection con- Effect of relative humidity on the aerody- gal and bacterial concentrations and trol, J. Air Waste Manage., 46, 869–882. namic diameter and respiratory deposi- emission rates in six occupied children’s Mortimer, E.A., Wolinsky, E., Gonzaga, tion of fungal spores, Atmos. Environ., classrooms, Indoor Air, 25, 641–652. A.J. and Rammelkamp, C.H. (1966) Role 30, 3967–3974. Johnson, D.L., Mead, K.R., Lynch, R.A. of airborne transmission in staphylococ- Riley, R.L. (1974) Airborne infection, Am. J. and Hirst, D.V.L. (2013) Lifting the lid cal infections, Br. Med. J., 1, 319–322. Med., 57, 466–475. on toilet plume aerosol: a literature Muise, B., Seo, D.C., Blair, E.E. and Riley, W.J., McKone, T.E., Lai, A.C.K. and review with suggestions for future Applegate, T. (2010) Mold spore penetra- Nazaroff, W.W. (2002) Indoor particu- research, Am. J. Infect. Control, 41, tion through wall service outlets: a pilot late matter of outdoor origin: importance 254–258. study, Environ. Monit. Assess., 163, of size-dependent removal mechanisms, Kelley, S.T., Theisen, U., Angenent, L.T., 95–104. Environ. Sci. Technol., 36, 200–207, 1868 St. Amand, A. and Pace, N.R. (2004) Nazaroff, W.W. (2004) Indoor particle Rintala, H., Pitkaranta,€ M. and Taubel,€ M. Molecular analysis of shower curtain bio- dynamics, Indoor Air, 14(Suppl. 7), (2012) Microbial communities associated film microbes, Appl. Environ. Microbiol., 175–183. with house dust, Adv. Appl. Microbiol., 70, 4187–4192. Nicas, M. (1995) Respiratory protection and 78,75–120. Knibbs, L.D., He, C., Duchaine, C. and the risk of Mycobacterium tuberculosis Robertson, O.H. (1943) Air-borne infection, Morawska, L. (2012) Vacuum cleaner infection, Am. J. Ind. Med., 27, 317–333. Science, 97, 495–502. emissions as a source of indoor exposure Noda, T. (2012) Native morphology of influ- Rylander, R., Reeslev, M. and Hulander, T. to airborne particles and bacteria, Envi- enza virions, Front. Microbiol., 2, 269. (2010) Airborne enzyme measurements to ron. Sci. Technol., 46, 534–542. Noris, F., Siegel, J.A. and Kinney, K.A. detect indoor mould exposure, J. Environ. (2011) Evaluation of HVAC filters as a Monitor., 12, 2161–2164.

77 Nazaroff

Seo, S.C., Reponen, T., Levin, L., Borchelt, Toivola, M., Nevalainen, A. and Alm, S. Trans. R. Soc. Lond. B Biol. Sci., 365, T. and Grinshpun, S.A. (2008) Aerosoli- (2004) Personal exposures to particles 3645–3653. zation of particulate (1?3)-b-D-glucan and microbes in relation to microenviron- Wouters, I.M., Douwes, J., Doekes, G., from moldy materials, Appl. Environ. mental concentrations, Indoor Air, 14, Thorne, P.S., Brunekreef, B. and Heeder- Microbiol., 74, 585–593. 351–359. ik, D.J.J. (2000) Increased levels of mark- Shelton, B.G., Kirkland, K.H., Flanders, Tsai, F.C. and Macher, J.M. (2005) Concen- ers of microbial exposure in homes with W.D. and Morris, G.K. (2002) Profiles of trations of airborne culturable bacteria in indoor storage of organic household airborne fungi in buildings and outdoor 100 large US office buildings from the waste, Appl. Environ. Microbiol., 66, environments in the United States, Appl. BASE study, Indoor Air, 15(Suppl. 9), 627–631. Environ. Microbiol., 68, 1743–1753. 71–81. Yamamoto, N., Shendell, D.G., Winer, Sippola, M.R. and Nazaroff, W.W. (2003) Tsai, F.C., Macher, J.M. and Hung, Y.Y. A.M. and Zhang, J. (2010) Residential air Modeling particle loss in ventilation (2007) Biodiversity and concentrations of exchange rates in three major US metro- ducts, Atmos. Environ., 37, 5597–5609. airborne fungi in large US office buildings politan areas: results from the Relation- Spendlove, J.C. and Fannin, K.F. (1983) from the BASE study, Atmos. Environ., ship among Indoor, Outdoor, and Source, significance, and control of 41, 5181–5191. Personal Air study 1999–2001, Indoor indoor microbial aerosols: human health Veillette, M., Knibbs, L.D., Pelletier, A., Air, 20,85–90. aspects, Public Health Rep., 98, 229–244. Charlebois, R., Lecours, P.B., He, C., Yamamoto, N., Nazaroff, W.W. and Peccia, Spengler, J., Neas, L., Nakai, S., Dockery, Morawska, L. and Duchaine, C. (2013) J. (2014) Assessing the aerodynamic diam- D., Speizer, F., Ware, J. and Raizenne, M. Microbial contents of vacuum cleaner eters of taxon-specific fungal bioaerosols (1994) Respiratory symptoms and hous- bag dust and emitted bioaerosols and by quantitative PCR and next-generation ing characteristics, Indoor Air, 4,72–82. their implications for human exposure DNA sequencing, J. Aerosol Sci., 78,1–10. Stelzer-Braid, S., Oliver, B.G., Blazey, A.J., indoors, Appl. Environ. Microbiol., 79, Yang, W., Elankumaran, S. and Marr, L.C. Argent, E., Newsome, T.P., Rawlinson, 6331–6336. (2011) Concentrations and size distribu- W.D. and Tovey, E.R. (2009) Exhalation Verreault, D., Moineau, S. and Duchaine, C. tions of airborne influenza A viruses mea- of respiratory viruses by breathing, (2008) Methods for sampling of airborne sured indoors at a health centre, a day- coughing, and talking, J. Med. Virol., 81, viruses, Microbiol. Mol. Biol. Rev., 72, care centre and on aeroplanes, J. R. Soc. 1674–1679. 413–444. Interface, 8, 1176–1184. Tellier, R. (2009) Aerosol transmission of Warfel, J.M., Beren, J. and Merkel, T.J. Yeh, H.C., Cuddihy, R.G., Phalen, R.F. and influenza A virus: a review of new studies, (2012) Airborne transmission of Bordetel- Chang, I.Y. (1996) Comparisons of calcu- J. R. Soc. Interface, 6, S783–S790. la pertussis, J. Infect. Dis., 206, 902–906. lated respiratory tract deposition of parti- Thatcher, T.L., Lai, A.C.K., Moreno-Jack- Waring, M.S. and Siegel, J.A. (2008) Particle cles based on the proposed NCRP model son, R., Sextro, R.G. and Nazaroff, loading rates for HVAC filters, heat and the new ICRP66 model, Aerosol Sci. W.W. (2002) Effects of room furnish- exchangers, and ducts, Indoor Air, 18, Technol., 25, 134–140. ings and air speed on particle deposi- 209–224. Yooseph, S., Andrews-Pfannkoch, C., Ten- tion rates indoors, Atmos. Environ., 36, Weber, T.P. and Stilianakis, N.I. (2008) ney, A., McQuaid, J., Williamson, S., 1811–1819. Inactivation of influenza A viruses in Thiagarajan, M., Brami, D., Ziegler- Thomson, R., Tolson, C., Carter, R., Coul- the environment and modes of transmis- Allen, L., Hoffman, J., Goll, J.B., ter, C., Huygens, F. and Hargreaves, M. sion: a critical review, J. Infect., 57, Fadrosh, D., Glass, J., Adams, M.D., (2013) Isolation of nontuberculous myco- 361–373. Friedman, R. and Venter, J.C. (2013) A bacteria (NTM) from household water Wehrle, P.F., Posch, J., Richter, K.H. and metagenomic framework for the study of and shower aerosols in patients with pul- Henderson, D.A. (1970) An airborne out- airborne microbial communities, PLoS monary disease caused by NTM, J. Clin. break of smallpox in a German hospital One, 8, e81862. Microbiol., 51, 3006–3011. and its significance with respect to other Yu, I.T.S., Li, Y., Wong, T.W., Tam, W., Tian, Y., Sul, K., Qian, J., Mondal, S. and recent outbreaks in Europe, Bull. World Chan, A.T., Lee, J.H.W., Leung, D.Y.C. Ferro, A.R. (2014) A comparative study Health Organ., 43, 669–679. and Ho, T. (2004) Evidence of airborne of walking-induced dust resuspension Womack, A.M., Bohannan, B.J.M. and transmission of the severe acute respira- using a consistent test mechanism, Indoor Green, J.L. (2010) Biodiversity and bio- tory syndrome virus, N. Engl. J. Med., Air, 24, 592–603. geography of the atmosphere, Philos. 350, 1731–1739.

78