Scientists at the United States Environmental Protection Agency (EPA) have estimated that, in the United States, 130,000 deaths each year are attributable to ambient (i.e., outdoor) particulate matter (PM) exposure. Neal Fann et al., Estimating the national public health burden associated with exposure to ambient PM(2.5) and ozone, 32 Risk Analysis 81, 88 (2012). Given these types of mortality estimates, EPA and other public health agencies frequently recommend that people stay indoors during high PM days to reduce their exposure to ambient particles. Ironically, scientific findings demonstrate that levels of indoor PM often exceed those outdoors due to common everyday indoor PM generation sources. That exposure is further exacerbated by the fact that people in the United States spend the majority of their time indoors where they are often in close proximity to indoor PM sources. Yet, relatively few epidemiology and toxicology studies have addressed the potential health effects of indoor PM as compared to ambient PM. It has been generally assumed that indoor PM is somehow less toxic than ambient PM.
EPA continues to tighten its National Ambient Air Quality Standards (NAAQS) for ambient PM, and scientists and regulators have expressed renewed interest in better understanding exposures and potential health risks posed by indoor PM. In fact, at the request of EPA, the National Academies of Sciences, Engineering, and Medicine (NASEM) organized a scientific workshop focused on the health risks of indoor PM in February 2016. In this article, we begin to fill in the knowledge gaps by examining what is currently known about indoor PM exposures and health risks and by discussing the importance of additional research on this topic, including the implications of EPA’s regulation of ambient PM for indoor PM. The focus of our article is residential indoor PM in developed countries—rather than occupational PM or indoor PM in developing countries, where unvented, open combustion of solid fuels is prevalent for household cooking and heating purposes.
Sources and Properties of Indoor PM
Indoor PM can be classified as deriving from two distinct sources: namely, ambient particles that have infiltrated indoors and particles generated indoors. Ambient fine particulate matter (referred to as PM2.5 and generally defined as consisting of tiny particles with diameters of 2.5 micrometers and smaller) can penetrate readily through the building envelope and reach indoor spaces, even in tightly sealed buildings. Because of this penetration and the major proportion of time people typically spend indoors, research has focused on understanding the infiltration of ambient particles into indoor spaces to inform the regulation of ambient PM. Studies show that levels of ambient-derived indoor PM can be reduced substantially compared to outdoor concentrations, because significant particle losses can occur both during penetration through the building shell and because of deposition of ambient-derived PM on indoor surfaces.
Researchers have demonstrated that ambient PM infiltration can be highly variable depending on a building’s characteristics (e.g., building tightness, air conditioner usage, and presence of an air filtration system), human behavior (e.g., whether windows and doors are left open), and ambient PM properties (e.g., particle size and composition). See, e.g., Evangelia Diapouli et al., Estimating the concentration of indoor particles of outdoor origin: A review, 63 J. Air Waste Mgmt. Ass’n 1113, 1113–29 (2013). Because these factors can limit the ability of ambient particles to infiltrate indoors, the size distribution and chemical composition of ambient particles found indoors are generally different compared to those found outdoors. For example, due to greater infiltration losses, coarse-mode particles (which encompass particles with diameters between 2.5 and 10 micrometers and are known as PM10-2.5) and ultrafine particles (UFP, which are defined as having diameters of 100 nanometers (0.1 micrometers) and smaller) of ambient origin are typically at significantly reduced levels indoors compared to their outdoor concentrations. Studies have demonstrated that variability in PM infiltration factors may be an important source of the variability in the health effect associations reported by ambient PM epidemiology studies. See, e.g., Jeremy A. Sarnat et al., Spatiotemporally resolved air exchange rate as a modifier of acute air pollution-related morbidity in Atlanta, 23 J. Exposure Sci. Envtl. Epidemiology 606, 606–15 (2013).
It is now well-recognized that a number of sources of indoor-generated PM frequently can cause overall indoor PM concentrations to exceed both outdoor concentrations and health-based ambient air quality standards, such as the EPA 24-hour PM2.5 NAAQS of 35 micrograms per cubic meter (μg/m3). As we discuss below, indoor generation is often a greater source of coarse-mode PM and UFPs, given that infiltration losses are greatest for ambient particles of these sizes. As with ambient PM, combustion can yield significant amounts of indoor-generated PM. Indoor combustion sources, including cooking activities, natural gas-fueled ovens and stoves, smoking, incense, candle-burning, and domestic heating sources (e.g., wood and natural gas fireplaces, wood stoves, kerosene heaters, and gas-fired furnaces), can be significant sources of both UFPs and PM2.5 in residential homes. In fact, cooking activities are known to be major sources of indoor particles across a range of different sizes. UFPs and PM2.5 derive from natural gas fuel combustion, heating of cooking oils, and burning of food; and coarse-mode PM derives from mechanical cooking processes like the sautéing and frying of food. See, e.g., Christopher M. Long et al., Characterization of indoor particle sources using continuous mass and size monitors, 50 J. Air Waste Mgmt. Ass’n 1236, 1236–50 (2000). Moreover, researchers recently reported evidence that UFPs can be emitted from the heating of cooking pans, electric coil burners in stoves and ovens, and other surfaces containing organic matter and detergent residues. Lance A. Wallace et al., Ultrafine particles from electric appliances and cooking pans: Experiments suggesting desorption/nucleation of sorbed organics as the primary source, 25 Indoor Air 536, 536–46 (2015). This research has explained why significant quantities of UFPs can be emitted during cooking with electric as well as gas heat.
Cleaning activities are also well-established as a major source of indoor PM, in particular for coarse-mode PM. For example, dusting and vacuuming can re-suspend house dust, which can contain tracked-in soil; clothes fibers; plastic-wear particles; a variety of bioaerosols, such as bacterial-laden skin flakes, animal dander, insect parts, plant pollens, mold spores, bacteria, and viruses; and semi-volatile organic compounds such as pesticides and fire retardants. Although vacuuming is widely viewed as beneficial for reducing indoor dust exposure, research has demonstrated that this activity can actually be a source of several different kinds of airborne particles including UFPs and viable bacteria. Indeed, vacuum cleaner motors have been shown to be sources of UFPs and PM2.5, and while high-efficiency particulate air (HEPA) filters can reduce these emissions, vacuum cleaner bags or collection chambers can release viable bacteria as well as coarse-mode PM. Luke D. Knibbs et al., Vacuum cleaner emissions as a source of indoor exposure to airborne particles and bacteria, 46 Envtl. Sci. Tech. 534, 534–42 (2012). The very act of vacuuming can re-suspend coarse particles from rugs and other surfaces, thereby increasing the potential for inhalation.
UFPs can be generated indoors through a number of other everyday, but often overlooked, sources. For example, UFPs are formed via indoor chemical reactions, such as the gas-phase reactions of ozone with terpenes, which are a class of volatile unsaturated hydrocarbons found in lemon- and pine-scented products and foods as well as plant oils. Similar to UFPs emitted from electric coil burners and cooking pans, the heating of organic matter deposits on some consumer products, such as hairdryers and printers, also has been identified as a source of indoor UFPs. Secondhand smoke is a well-recognized source of UFPs (and PM2.5) in homes with smokers, and recent studies have demonstrated how it can be transported from apartments with smokers to smoke-free apartments. See, e.g., Elizabeth T. Russo et al., Comparison of indoor air quality in smoke-permitted and smoke-free multiunit housing: Findings from the Boston Housing Authority, 17 Nicotine & Tobacco Research 316, 316–22 (2015). On a related note, e-cigarettes are an emerging source of both indoor UFPs and PM2.5. Esteve Fernández et al., Particulate matter from electronic cigarettes and conventional cigarettes: a systematic review and observational study, 2 Current Envtl. Health Rep. 423, 423–29 (2015).
Coarse-mode PM from house dust, tracked-in soils, and other deposited particles can be re-suspended via any physical activity of people and pets inside their homes, such as simply walking across a room or collapsing onto a sofa. Finally, other lesser-known indoor sources of coarse-mode particles (and PM2.5) include showers and humidifiers, where water-soluble salts are left behind after the evaporation of airborne water droplets.
Aside from our homes, indoor-generated PM is common in other places we visit on an everyday basis. Not surprisingly, restaurants have been shown to contain high levels of indoor PM emitted during cooking. A study of indoor particle levels in California office buildings indicated that significant fractions of the PM2.5 concentrations inside dental offices, hair salons, and gyms/daycares were attributable to indoor sources. Xiangmei Wu et al., Indoor particle levels in small- and medium-sized commercial buildings in California, 46 Environ. Sci. Technol. 12355, 12355–63 (2012). In a study of particle levels in 14 retail stores in Pennsylvania and Texas, other researchers reported that indoor sources accounted for 83 percent and 53 percent (as medians) of overall PM10 and PM2.5 indoor concentrations, respectively. Marwa Zaatari and Jeffrey Siegel, Particle characterization in retail environments: Concentrations, sources, and removal mechanisms, 24 Indoor Air 350, 350–61 (2014).
One often can distinguish indoor sources of PM from background levels of ambient PM inside by their contributions to short-term peak PM events. A large number of studies have characterized the episodic PM2.5 and UFP levels from cooking in residential homes, demonstrating the frequent occurrence of short-term peak concentrations of indoor PM2.5 and UFP at levels ranging from tens to hundreds of times higher during cooking as compared to indoor concentrations during non-activity periods. Moreover, in a study of PM2.5 and UFP levels in nonsmoking restaurants in California, researchers reported that simply going into a restaurant can increase one’s UFP exposures to more than 100,000 particles/cm3. Although more modest increases in PM2.5 concentrations were typical of most sampled restaurants (restaurant PM2.5 mass concentrations averaged 36.3 μg/m3), they reported PM2.5 concentrations exceeding 100 μg/m3 in a small number of restaurants and a maximum PM2.5 concentration of 454 μg/m3. Wayne R. Ott et al., Fine and ultrafine particle exposures on 73 trips by car to 65 nonsmoking restaurants in the San Francisco Bay Area, 27 Indoor Air 205, 215 (2017). These investigators also measured PM2.5 and UFP concentrations in cars during travel to and from restaurants, and concluded that restaurants represent a greater source of both PM2.5 and UFP exposures than cars. Id. at 215. For vacuuming and for walking on a carpeted basement floor, short-term peak PM5 concentrations of 81 and 165 μg/m3, respectively, have been reported. (PM5 refers to particles with diameters of 5 micrometers and smaller, and thus encompasses PM2.5 as well as a fraction of coarse-mode particles.) Andrea R. Ferro et al., Elevated personal exposure to particulate matter from human activities in a residence, 14 J. Exposure Analysis and Envtl. Epidemiology S34, S34–S40 (2004). As part of a study of PM2.5 levels inside homes using wood burning for heat, researchers reported a median one-minute PM2.5 peak concentration of 386 μg/m3 during stove loadings after change outs to new EPA-certified woodstoves. Curtis A. Noonan et al., Residential indoor PM2.5 in wood stove homes: Follow-up of the Libby changeout program, 22 Indoor Air 492, 492–500 (2012). Because of these and other indoor PM sources, we are all exposed to short-term peak levels of PM on an everyday basis.
It is difficult to precisely separate indoor PM into ambient and non-ambient components because indoor particles typically contain many of the same chemical constituents as ambient particles, including sulfates, elemental carbon (a main component of combustion soot), organic matter, trace elements, and crustal materials. Nevertheless, some studies have successfully demonstrated major differences in the composition of indoor-generated PM versus ambient PM. In particular, there is evidence that indoor-generated PM can be more enriched with organic matter content as compared to ambient PM. For example, Polidori et al. reported mean particulate organic matter concentrations that were two times mean ambient concentrations in one of the largest studies conducted to date of residential PM chemical composition, namely the Relationship of Indoor, Outdoor, and Personal Air (RIOPA) study. Andrea Polidori et al., Fine organic particulate matter dominates indoor-generated PM2.5 in RIOPA homes, 16 J. Exposure Sci. and Envtl. Epidemiology 321, 328 (2006). Based on data for 173 homes in Houston, Texas; Los Angeles County, California; and Elizabeth, New Jersey; researchers reported evidence that the predominant species in indoor-generated PM2.5 is organic matter, with upward of 71 percent to 76 percent of indoor organic PM being emitted or formed indoors. Id. Indoor organic matter can contain thousands of chemical species, including synthetic organic chemicals such as plasticizers, organophosphates, and fire retardants. Nat’l Acad. of Sci., Eng’g, and Med. (NASEM), Health Risks of Indoor Exposure to Particulate Matter: Workshop Summary, at 33 (2016).
Other characteristics of indoor and outdoor PM also emerged from the RIOPA study. Consistent with studies demonstrating major losses of nitrates during indoor infiltration, the RIOPA study also suggested significantly lower nitrate concentrations for indoor PM2.5 than outdoor PM2.5. A. Polidori et al. at 329. However, similar levels of elemental carbon were observed for indoor and outdoor PM2.5 samples, likely due in part to both indoor-generation sources as well as the ability of ambient soot combustion particles to readily infiltrate into indoor spaces. A. Polidori et al. at 326. Finally, indoor PM is widely recognized to often contain larger amounts of biological components that can be potent allergens, including bacteria, fungi, molds, and animal dander. NASEM at 16.
In summary, we are exposed on an everyday basis to a number of different kinds of indoor-generated particles. Indoor PM sources are well-known to increase indoor PM concentrations above outdoor concentrations, in particular over short-term periods; to modify particle size distributions, increasing the fractions of both ultrafine and coarse-mode particles; and to enrich the organic matter content of indoor PM.
Toxicity of Indoor-Generated PM
There is a prevailing belief that indoor PM is less toxic than ambient PM and thus does not pose the same degree of risk to human health as ambient PM. This viewpoint has persisted due in large part to the relatively small number of studies to investigate the toxicity of indoor-generated PM. It also has persisted despite some of the research findings regarding indoor PM properties and concentrations discussed above, which would suggest a toxic potential that is similar to or possibly greater than ambient PM. Among the key findings supporting the similar or greater toxicity of indoor PM as compared to ambient PM are (1) the significant contributions of indoor combustion sources to indoor PM; (2) evidence that indoor sources give rise to UFP exposures and the growing concerns regarding the toxicity of UFPs; (3) the frequent occurrence of higher PM2.5 levels inside homes as compared to outdoors; and (4) indoor PM commonly contains significant amounts of organic matter and biological components (e.g., viable pathogens).
Among the earlier studies to investigate the relative toxicity of indoor and outdoor PM2.5 is a cell culture (in vitro) study conducted by one of the authors of this article. Christopher M. Long et al. 2001, A pilot investigation of the relative toxicity of indoor and outdoor fine particles: In vitro effects of endotoxin and other particulate properties. Envtl. Health Persp. 109 (10):1019–26. For this pilot study, a rat cell line was dosed with liquid suspensions of particles derived from 14 paired indoor and outdoor PM2.5 samples collected in nine Boston-area homes. The findings indicated greater inflammatory responses for the indoor PM2.5 samples than the outdoor PM2.5 samples. The inflammatory responses were due in part to the indoor PM2.5 samples containing larger amounts of endotoxins, which are organic toxins associated with bacterial cell walls. Id. at 1023.
Results from other in vitro studies provide evidence of the toxic potential of indoor-generated PM2.5. Two recent studies, one conducted in Germany (Sebastian Oeder et al., Airborne indoor particles from schools are more toxic than outdoor particles, 47 Am. J. Respiratory Cell Molecular Biology 4575, 575–82 (2012)) and the other in Singapore (Mei Ling Chua et al., Particulate matter from indoor environments of classroom induced higher cytotoxicity and leakiness in human microvascular endothelial cells in comparison with those collected from corridor, Indoor Air (Sept. 2016), available at https://www.researchgate.net/publication/308531381_Particulate_matter_from_indoor_environments_of_classroom_induced_higher_cytotoxicity_and_leakiness_in_human_microvascular_endothelial_cells_in_comparison_to_those_collected_from_corridor), reported evidence that indoor PM from school classrooms elicited greater cellular responses for a range of different biomarkers, including cell death, allergenicity, inflammatory response, and acceleration of blood coagulation, as compared to either outdoor or hallway PM samples. Both of these studies linked the endotoxin content of the indoor PM samples with many of the observed responses. It bears mentioning that although these studies are useful for generating hypotheses and exploring the plausibility of toxicological mechanisms, in vitro studies, by their very nature, are simplifications that may not reflect actual exposure conditions or account for the complex whole-organism processes that could mitigate or amplify responses seen in isolated cells.
Human studies also are revealing of the toxicity of indoor-generated PM. Some of these studies are aimed at evaluating the beneficial health effects of interventions that reduce indoor PM exposures, including in particular air filtration devices. Several different intervention studies have demonstrated beneficial impacts of air filtration on markers of adverse cardiovascular and respiratory health effects. Although suggestive of the toxicity of indoor-generated PM, most of these studies have not been able to separate contributions of indoor-generated PM versus indoor PM of ambient origin. One notable exception is an intervention study that assessed the impact of air filtration devices on indoor PM2.5 concentrations and markers of endothelial dysfunction and inflammation among healthy adults living in wood-heated homes in British Columbia. (Endothelial dysfunction refers to the impaired function of the endothelium, which is the inner lining of blood vessels; it is well-established as a predictor of cardiovascular disease.) Ryan W. Allen et al., An air filter intervention study of endothelial function among healthy adults in a woodsmoke-impacted community, 183 Am. J. Respiratory Critical Care Med. 1222, 1222–30 (2011). As part of this study, these researchers separated indoor PM2.5 concentrations into indoor source and ambient contributions, reporting that for both filtration and non-filtration periods, an average of 67 percent of the total indoor PM2.5 concentration was associated with indoor sources. Id. at 1223. They reported evidence that both total indoor PM2.5 and indoor-generated PM2.5 were more strongly associated with pro-inflammatory markers than ambient PM. Id. at 1225.
Other observational studies of small panels of human volunteers have reported mixed findings regarding the biological activity of indoor-generated PM. For example, a recent study of healthy middle-aged Danish subjects from 58 residential homes reported statistical associations between number concentrations of indoor-generated particles (primarily associated with candle burning and bioaerosols) and lung function, markers of systemic inflammation, and diabetes. Dorina G. Karottki et al. Cardiovascular and lung function in relation to outdoor and indoor exposure to fine and ultrafine particulate matter in middle-aged subjects, 73 Envtl. Int’l 372, 372–81 (2014). Interestingly, they did not observe similar associations for cooking-related particles, which they concluded may be due to factors such as the relatively short duration of cooking events, the use of fume hoods in kitchens, and the separation of kitchens from the main living areas of the homes. Id. at 379. By comparison, another study of Danish adults conducted by many of the same investigators reported findings suggesting that indoor-generated PM may have smaller effects on blood vessel function than ambient PM. Yulia Olsen et al, Vascular and lung function related to ultrafine and fine particles exposure assessed by personal and indoor monitoring: A cross-sectional study 13 Envtl. Health 1, 1–10 (2014).
Finally, a controlled human exposure study of 55 healthy adult volunteers in Germany investigated the relationship between several lung function measures and indoor PM emissions from burning candles, frying sausage, and toasting bread. Vanessa J. Soppa et al., Respiratory effects of fine and ultrafine particles from indoor sources—a randomized sham-controlled exposure study of healthy volunteers, 11 Int’l J. Envtl. Res. Public Health 6871, 6871–89 (2014). In contrast to the observational studies described above, this study was conducted in an exposure chamber with well-defined, source-specific exposure concentrations and durations. Although no associations were found between toasting bread and impaired lung function, the authors reported associations between PM2.5 concentrations from candle burning and frying sausage with small negative changes in multiple lung function variables. Id. at 6872. PM2.5 concentrations averaged approximately 50 to 80 μg/m3 for candle burning experiments and 80 to 240 μg/m3 for frying sausage experiments, and thus were in the range of levels reported for these activities in residential homes. Id. at 6877.
Almost all of the human observational studies, including those described above, have focused on short-term exposures to indoor PM. There remains a general absence of epidemiological studies examining the health implications of longer-term, chronic exposures to indoor PM. One exception in this regard is the human health risk posed by long-term exposure to secondhand smoke. In addition, a recent study reported that current incense users in Singapore, the majority of whom had been burning incense daily for more than 20 years, had a 12 percent higher risk of cardiovascular mortality as compared to non-users. An Pan et al., Incense use and cardiovascular mortality among Chinese in Singapore: The Singapore Chinese Health Study, 122 Envtl. Health Persp. 1279, 1279–84 (2014). In addition to generating PM2.5 and UFPs, incense burning can emit carcinogenic gases such as benzene and carbonyls, but this study provides suggestive evidence of the potential human health risks of long-term exposure to indoor PM from incense burning.
In summary, more investigation of potential health risks posed by indoor-generated PM is needed, but increasingly, studies support the conclusion that indoor-generated PM has a toxic potential that is similar to or possibly greater than ambient PM.
Because of the relative scarcity of health effects studies focused on indoor PM, large gaps remain in our understanding of the specific health risks posed by indoor-generated PM. This dearth is in stark contrast to the wealth of information on ambient PM, for which there are now thousands of health effects studies, including short-term and long-term epidemiology studies, human clinical studies, laboratory animal studies, and in vitro studies.
EPA has determined that both short-term and long-term exposures to ambient PM2.5 are causally related to increased risk of premature mortality and cardiovascular health effects, and that there is a “likely to be causal” relationship between short-term ambient PM2.5 exposures and respiratory health effects. EPA, Office of Research and Development, National Center for Environmental Assessment (NCEA)—RTP Division, Integrated Science Assessment for Particulate Matter (Final), at 2–32 (2009). For long-term ambient PM2.5 exposure and reproductive and developmental effects, cancer, mutagenicity, and genotoxicity, EPA has concluded that there is “suggestive” evidence of a causal relationship. And EPA has made causal determinations for two other PM size intervals, namely PM10-2.5 and UFPs. For these PM size intervals, EPA identified a “suggestive” causal relationship between short-term exposures and some health endpoints (e.g., cardiovascular and respiratory effects for both; mortality for just PM10-2.5), but found that there was inadequate evidence for all other health effects.
The regulatory framework established by the Clean Air Act has been a major factor driving the extensive health effects research on ambient PM, and the absence of regulation for indoor air likely has contributed to the research for indoor PM being relatively sparse. Nonetheless, regulators recognize indoor PM as a public health issue, as EPA and other regulatory agencies, such as the California Air Resources Board, provide information on their websites on measures that can be taken to reduce exposures to common sources of indoor-generated PM. Examples of recommended practices include installing and using vented kitchen range hoods whenever cooking, limiting indoor smoking and burning of candles and incense, and using medium- or high-efficiency filters in central forced air systems or high-efficiency portable air cleaners for homes without central systems.
Congress is not likely to give EPA or other agencies authority to regulate indoor PM anytime soon, but there are several reasons why indoor PM should be receiving greater regulatory scrutiny. In particular, it has been previously mentioned that EPA and other regulatory agencies often recommend that people stay indoors on high air pollution days. However, as discussed in this article, indoor environments can frequently have higher PM concentrations, especially for the UFP and coarse-mode PM fractions. Moreover, it has been hypothesized that the impacts of global climate change may serve to prompt further efforts to tighten homes and reduce air conditioning costs, in turn decreasing ventilation to the outdoors and allowing more indoor-generated PM to build up and remain confined in indoor spaces. NASEM, supra, 17. Regulatory bodies will need to both recognize these countervailing tendencies and find ways to balance them if they are to protect the public health.
In December 2016, EPA released its “Integrated Review Plan for the National Ambient Air Quality Standards [NAAQS] for Particulate Matter,” and the agency is in the early stages of reviewing the ambient PM NAAQS. It is far too early to predict whether EPA will determine that the scientific evidence is supportive of further reductions to the ambient PM NAAQS, but a key public health question is whether the risks posed by indoor PM may overshadow those from ambient PM, and therefore may lead to marginal health benefits from additional reductions to the ambient PM NAAQS. In other words, further reductions in the ambient PM NAAQS may have diminishing returns in limiting people’s total PM exposures and health risks, if, as this article suggests, such exposures are dominated by PM from indoor sources. Based on the frequent references to UFPs in EPA’s Integrated Review Plan document, it is likely the agency will evaluate again the need for a separate NAAQS for UFPs. Yet, given the well-recognized contributions of indoor generation sources to total UFP exposures, it is unclear how development of a UFP NAAQS could proceed without accounting for indoor UFP exposures.
In conclusion, current research findings demonstrate that indoor PM sources contribute in a major way to people’s total PM exposures and also suggest that indoor-generated PM should be investigated further for potential health risks. As shown by the recent NASEM workshop, there appears to be a growing consensus among regulators and scientists regarding the need to better understand exposures and potential health risks posed by indoor PM sources.
"Take me IN to the ballgame where the air is clear (or is it?)"