In collaboration with their partners, research teams at Health Canada's Safe Environments Programme conduct specific research projects related to air pollution. Research focusses mainly on the human health effects of chemical and biological agents present in air.
Health Canada scientists have played a leading role in understanding the effects of indoor and outdoor air pollution on human health in Canada. Studies completed to date include: analyses of the effects of home dampness and mould on respiratory health; the relationship between outdoor air pollution and lung function, respiratory symptoms, emergency department visits, hospital admissions and premature deaths; the biological mechanisms of air pollutant-induced effects; and analysis of the dose-effect relationships.
Studies have been developed to better understand the effects of indoor air quality on infant health; of short-term outdoor air pollution exposure on disability days and heart rate; of long-term outdoor air pollution exposure on the health of children and adolescents; and the toxicological effects of ozone and particles on the cardiovascular-resiratory system. Here is a list of some of these studies:
Particles from ambient air are not acutely toxic in terms of acute structural injury in the lungs, but will nevertheless exacerbate a pulmonary lesion or an inflammatory reaction. The nature and the slope of the dose-response relationship of the biologically effective dose of particles in the lungs will be closely related to host factors, such as a pre-existing pathology, or the existence of a primary lesion caused by a co-toxicant. It becomes essential to model the dose-response relationship for particles taking into account the interaction of host factors and toxicological interactions between contaminants.
With this end in mind, we are studying the interaction of particles and ozone in the lungs, for which we have shown a strong synergy. Ozone is a common environmental contaminant, but at the same time serves to create a limited primary lesion which mimics to some extent pre-disposing lung epithelial and inflammatory conditions. These experiments will allow us to characterize the intensity of cellular changes in the pulmonary tissues induced by particles under a continuum of biological sensitivity. The model will then serve to assess the potential interactions of oxides of nitrogen and carbon monoxide on the evolution and resolution of a pulmonary lesion. Part of this work is supported by TSRI (Toxic Substance Research Initiative) funds.
As part of our efforts to identify biologically plausible mechanisms of effects that relate exposure to particles with adverse cardiovascular effects, we are examining vasoactive peptides, such as endothelins, in the lungs and heart. We will use immunocytochemistry to locate and quantify these peptides at the tissue level. Endothelins are the most potent vasoconstrictors currently known and spillover from the lungs is the main source of circulating endothelins. The lungs also capture circulating endothelins. The action of endothelins towards the walls of blood vessels is well known, so it stands to reason that if the balance of vasoconstriction/relaxation is perturbed and the lungs represent a large source of endothelins, a potential for significant cardiovascular effects exists.
From our recent work, this in fact seems to be a plausible biological mechanism but the location of endothelin in the lungs and heart has yet to be observed in animals exposed to air pollutants. Immunocytochemical techniques are being employed to detect and locate endothelins in the lungs and heart. A number of pathological conditions are known to affect endothelins in the lungs and heart, for example, cardiac conditions, diabetes and pulmonary hypertension. In future years, we plan to investigate the effects of air pollutants on animal models of disease. Part of this work is supported by TRSI (Toxic Substance Research Initiative) funds.
We have shown the establishment of oxidative stress scenario with ozone exposure. Oxidative stress has been implicated in various disease processes and tissue injury. Reactive oxygen species (e.g. superoxide anion, hydrogen peroxide and hydroxyl radical) and reactive nitrogen species (e.g. nitric oxide and peroxynitrite) are formed under oxidative stress scenarios. Atmospheric particulate matter contains transition metals that can act as Fenton catalysts in the transformation of hydrogen peroxide to hydroxyl radical which is associated with tissue injury and thus can modify oxidative stress related toxic responses. We have observed a strong toxicological interaction between urban particulate matter and ozone in rats. This interaction can be due in part to oxidative stress mechanisms.
We have received financial support from the TSRI (Toxic Substances Research Initiative) program to study a specific aspect of oxidative stress, namely, the production of hydroxyl radicals under ozone/particle exposure scenario by means of endogenous (e.g. phenylalanine hydroxylation) and external markers (e.g. 5-amino salicylic acid). As oxidative stress mechanisms involve various reaction pathways, we need to identify a panel of sensitive markers. This will allow us to probe the toxic interactions of air pollutants in terms of oxidative stress.
Adverse acute responses to low levels of environmental contaminants in human populations is very sensitive to host status. This notion, deduced from epidemiological data, is receiving support from experimental work in animals and human subjects. We have found that one component of the pulmonary response to particles is the existence of a primary injury and an inflammatory state. This pro-inflammatory state can be created by exposure of animals to particles plus an irritant gas such as ozone, or alternatively can be created by over-expressing inflammatory cytokines in the lungs.
We are investigating the response of mice that over-express tumor necrosis factor alpha (TNF-alpha) in lung epithelia after exposure to particles. The TNF-alpha gene is regulated by the surfactant protein C promoter, which is active in bronchiolar and alveolar type 2 cells. These animals spontaneously develop a form of chronic obstructive pulmonary disease (emphysematous lesions).
This model also reproduces the high TNF-alpha phenotype of humans, which has been associated with increased risk for pneumoconiosis. Similarly, inflammatory cascades are regulated in part by antioxidant defenses, of which extracellular superoxide dismutase (EC-SOD) is the main component. A strain of mice in which the EC-SOD gene has been knocked out is also being investigated. The animals express an inactive form of the enzyme and display dysfunctional vasoregulation and sensitivity to oxidant stress. These transgenic mice models serve to mimic human sensitivity factors that should interact with mechanisms of action of air pollutants.
We have verified that inhalation of urban particulate matter by rats causes increases of plasma vasoconstrictors, namely endothelin-3 and endothelin-1 as well as increases in systemic blood pressure. To investigate this phenomenon in animal species as a potential pathway for the cardiovascular effects of ambient particles, we are continuing to look at these and other endpoints such as the vasodilator, NO (nitric oxide) and another vasoconstrictor, isoprostane. However, we have also begun to transfer these endpoints of effects to human studies by establishing a number of collaborations with other Canadian laboratories. Part of this work is supported by TSRI (Toxic Substances Research Initiative) funds.
These multi-year, and multi-disciplinary studies examine physiological, biochemical and clinical effects of ambient fine particles and ozone on respiratory and cardiovascular systems of healthy and asthmatic subjects, including asthmatic children. Fine particles are obtained from downtown Toronto ambient air. The subjects undergo an exposure of known amount of air pollutants which are expected to cause mild and transient effects. Scientists observe the subtle cellular and molecular changes in cardio-respiratory systems to elucidate the mechanism(s) of air pollutant-induced adverse effects and to determine why certain people are more vulnerable to air pollution, and potential intervention strategies. These controlled human exposure studies are critical in providing the biologically plausible evidence for epidemiological findings which have shown a significant association between increased ambient particulate and ozone pollution and population mortality and morbidity, and in support of evidence-based policy making.
These research projects are a collaborative effort of the scientists from Health Canada, the Gage Occupational and Environmental Health Unit of University of Toronto and St. Michael's Hospital, and Environment Canada. The projects are sponsored by the federal Toxic Substance Research Initiative and the Program on Energy Research and Development."
Analysis from the Childrens Health Study indicated an association between increased levels of acid particulates and decreased lung function among grades 4 and 5 children from 22 North American sites, including 6 sites in Canada. Data has been gathered and assembled into analytical data sets for 16 of the original sites (4 in Canada) for the follow-up Teen Lung Study.
Hospital admissions and deaths from cardiac disease (e.g. heart attacks) are higher on days of higher air pollution. The reasons for this association are unknown. We will compare personal air pollution exposure to changes in heart rate variability, arrhythmia and ischemia, as detected by Holter monitoring. The project was piloted in 1999 and 15 subjects are expected to be studied in 2000 and 2001.
Pharmaco-epidemiological methods are generally used to relate disease with hospital usage and administrative issues such as costs. The widespread use of computers and the extensive record keeping of drug dispensation could be used to explore an association between air pollution levels across regions/cities and within cities on a time series basis. The evidence is clear that there is an association between hospitalization and air pollution levels.
It would be useful to know if there is a similar association with the dispensation of medications associated with cardiorespiratory health. This evidence would be useful to more fully demonstrate the extent of the population at risk and to estimate or determine potential hidden costs of morbidity associated with air pollution levels.
People spend the majority of their time indoors yet the adverse health effects of indoor air are not well characterized. Infants are likely to be at increased risk of adverse health effects. In this study, infants will be followed for two years with daily symptom diaries and measurements will be made of the structure and performance of their homes and the quality of the indoor air. This is the fourth year of data collection and the sample is currently at 50% of the required sample size.
In the R-2000 Home Survey, occupants' health prior to and one year after taking up residence in an R-2000 energy-efficient house were assessed. To date, over 150 R-2000 and control new home owners have been identified, questionnaires administered and preliminary analysis conducted.*
*Funded in whole or in part by the Program of Energy Research and Development (PERD), Natural Resources Canada (NRCan).
The current air quality index is based on outdated health evidence, in particular as it relates to the existence of a threshold or safe level of exposure, and simultaneous effects of multiple pollutants. Current evidence necessitates a reworking of the index and associated health messages. Under this project, further analysis of health data will be conducted to support a new health-based index.
Physicians are a credible source of health information for the public, but education opportunities for physicians on air quality and health are limited. A unique opportunity was provided by the publication of a continuing education article on air pollution and health in the Canadian Medical Association Journal, as a platform for an interactive online continuing education exercise. This could serve as a model for programs on other environmental health issues. This project involves implementing an interactive e-mail discussion group on health effects of air pollution, and conducting an evaluation of the effectiveness of the program.
While there is considerable evidence linking short term air pollution exposure with adverse health effects, there is very little evidence regarding the consequences of long term exposure. This represents a major gap in risk assessments in terms of the whether existing standards protect the population from adverse effects of long term exposure. This gap has been identified in science assessment documents for ozone and particulate matter. Under this project, collaborating clinical and public health partners will be identified, design issues considered, data collection instruments developed, and procedures pilot tested.
While there is considerable evidence linking particulate matter and other pollutants with various acute health effects ranging from premature death to minor respiratory symptoms, there is limited evidence regarding the effects of long term exposure. A unique opportunity to fill this important gap in the base of scientific evidence is afforded by the availability of a relatively large cohort of Canadians (approximately 20,000) which is being followed as part of the National Population Health Survey (NPHS). This survey was initiated in 1994 and is repeated every two years. It includes questions on the presence of several chronic conditions, including asthma, chronic bronchitis and emphysema, and heart disease, all of which are of interest with respect to particulate matter exposure, as well as a numerous potentially confounding factors (eg. age, sex, income, smoking). The current survey instrument captures place of residence at the time of each survey wave. In order to refine assessment of long term air pollution exposure, a series of questions will be added to the 2002 survey cycle, which will determine place of residence from 1980, which is the year when widespread air pollution monitoring for multiple pollutants became consistently available across populated regions of the country. This will permit the assignment to each survey respondent pollution-years of exposure to selected air pollutants based on his/her place of residence from 1980 onwards. Once the 2002 data become available, a survival analysis will be conducted to assess the probability of development of disease in association with longitudinal air pollution exposure. This approach will be most statistically efficient in terms of capturing not just the development of new cases of cardiorespiratory disease, but also when they occur.Other opportunities afforded by the NPHS data include analysis of acute effects of air pollution on the frequency of reporting of days of restricted activity, as well as cross sectional analysis of air pollution and the prevalence of cardiorespiratory disease.
Episodes of ambient particulate matter have been associated with human health impacts, such as increased daily hospital admissions or mortality for both respiratory and cardiovascular causes. Associations are seen for low levels of contaminants, e.g. 30-50 ug/m3 (micrograms per cubic metre) particulate matter, or 30-50 ppb (parts per billion) ozone. There is a consensus that for such contaminant levels to cause acute adverse responses, individuals affected most likely suffer from an established pathology; i.e. the contaminants precipitate symptoms or alter the course of disease, but do not initiate a cardiac or lung disease in such a short time. Nevertheless, because of the apparent linear associations seen by epidemiology, it remains possible that the agents act inside general physiological processes, and not exclusively inside established pathophysiological processes. Our group is currently involved in developing and applying experimental models that allow us to define the mechanisms of effects of ambient air pollutants, and to identify the agents of toxicity.
We have demonstrated that a single inhalation exposure of rats to urban particles (Ottawa EHC-93 urban dust; 50 mg/m3, 4 hr) could stimulate cytokine production (TNF-alpha, MIP-2) in macrophages, but would not cause acute injury. This indicates that deposition of urban particles in intact lungs is relatively innocuous in terms of acute impact. However, when the integrity of the extracellular lining was compromised and an inflammatory response was initiated in the lungs by co-exposure of the rats to 0.8 ppm ozone, the deposition of particles in the lungs produced an injury several folds higher than what was measured for ozone alone. The data show that urban particles are actually intrinsically potent and can exacerbate epithelial injuries and interstitial inflammatory changes in sensitive lungs.
While inhalation of urban particles alone did not cause measurable acute lung lesions in the rats, we have found that plasma levels of endothelin-1 (ET-1) and endothelin-3 (ET-3) were elevated after exposure of the animals to urban particles. Endothelin-3 was elevated as early as 2 hr and as late as 48 hr after exposure. Elevation of circulating ET-3 (+100%, p<0.05) was associated with a vasopressor response (+3-4 mm Hg, p<0.05) on the second day after exposure. Inhalation of particles from which soluble components had been extracted did not significantly affect blood pressure and ET-3 plasma levels. Thus, chemical modification of the particles altered their potency. These observations reveal a novel mechanism for cardiovascular effects of inhaled particles in the absence of lung injury. It is known that elevation of ET-3 correlates with severity of disease and is a predictor of mortality in cardiac patients, but in principle, elevation of ET-3 should be rapidly compensated by homeostatic mechanisms in healthy subjects. This new biomarker of effects of urban particles is currently being applied in human studies. We are also probing into the markers of oxidative stress to understand other pathophysiologically relevant pathways.
Our current experimental findings appear to fall in line with the patterns of human impacts and explain why a few individuals could be severely affected, while most are apparently not acutely affected by ambient particles. Our observations contribute to the weight of evidence for a biological plausibility of acute health effects from exposure to ambient particulate matter.
Richard T. Burnett - Healthy Environments and Consumer Safety Branch, HealthCanada and Department of Epidemiology and Community Medicine, Faculty of Medicine, University of Ottawa.
The adverse health effects of both short-and long-term exposure to air pollution generated from the combustion of fossil fuels and industrial activity have now been well documented. New Canada-Wide Standards for ground-level ozone and particulate matter have been established which require air quality improvements in some regions of the country. Mitigation strategies need to target the most toxic components of the atmospheric mix in order to yield optimal benefits for population health. The most critical science needs are identified in the development of these strategies: placing air pollution related deaths in a population health perspective; identifying the toxic components of the pollution mix; and linking pollution exposure to the development of cardiopulmonary disease.
It has long been recognized that outdoor air pollution can play a role in exacerbating cardiopulmonary diseases and shortening life. The most dramatic evidence of this was given by a series of major pollution episodes in London, England in the 1950s and 1960s in which deaths and hospital admissions clearly increased in the days following the elevated pollution levels Figure A) (1). Several countries, including Canada, set more stringent standards for pollution emissions in an attempt to reduce ambient concentrations. Pollution levels have declined over the past several decades in Canada to a point where our National Air Quality Objectives are rarely violated (Figure B).
However, a series of new studies conducted in the 1990s have further implicated ambient air pollution as a risk factor for cardiopulmonary problems even at the lower concentrations experienced in Canada today (2). This new evidence prompted Federal and Provincial authorities to establish new Canada Wide Standards for particulate matter and ozone, two of the most prominent components of smog (2). Since several regions of the country do not currently comply with these standards, air pollution management strategies must be developed.
Although the link between exposure to ambient air pollution and adverse health effects has clearly been established, several scientific uncertainties remain that limit our ability to develop the most effective mitigation strategies in terms of improving population health. In this paper, three areas of scientific uncertainty are discussed which have a major influence on the development of such strategies: placing air pollution related deaths in a population health perspective; identifying the toxic components of the pollution mix; and linking pollution exposure to the development of cardiopulmonary disease.
A number of studies conducted worldwide have linked daily variations in the number of deaths and daily fluctuations in ambient pollution levels. These studies indicate that the daily number of deaths increases as outdoor pollution concentrations increase (Figure C). However, they are not able to yield estimates of the amount of life lost due to pollution exposure. If individuals who are severely ill and are expected to die shortly are the only people affected by current levels of air pollution, then reducing outdoor concentrations will not necessarily increase life expectancy. This so-called mortality displacement or harvesting effect has been examined by comparing mid-frequency or sub-seasonal cycles in both mortality and air pollution time series. The resulting positive associations between these variables indicate that there is more information in the data linking air pollution and death than would be predicted by the harvesting hypothesis alone. However, this approach cannot determine the amount of life lost and thus it is difficult to compare air pollution effects on population health with other health issues.
Several studies are now underway addressing this issue. Death certificate information was linked to drug, physician, and hospital billing data in the last five years of life for residents of the island of Montreal. The billing data was used to identify subjects with specific disease histories. For those groups, standard time series analyses were conducted linking air pollution and death. Daily mortality increased linearly as concentrations of particles increased for persons who had acute lower respiratory diseases, chronic coronary artery diseases (especially in the elderly), and congestive heart failure. Disease-specific life tables could then be developed for these groups and coupled with age-disease specific effect estimates from the time series analyses, estimates of life shortening could be obtained.
This approach assumes that the air pollution effect estimates, which represent increases in daily number of deaths, can be interpreted as relative risks. That is, there is some well-defined cohort of subjects who have been followed in time and there are subjects with varying degrees of exposure. In fact, the time series design does not strictly involve a cohort. Persons who live in a specified geographic area at their death may have moved there a few days previously, and other persons who lived in the area for many years, and thus have been exposed, may have moved out of the area. Thus it is not clear how well the effect estimates represent a true risk. A second limitation of time series studies is that in order to apply the effect estimates as relative risks in a life-table analysis, one needs to assume that the age-specific air pollution related deaths are for persons who would have had a normal life expectancy in the absence of exposure. This may not be the case if only the very ill are adversely affected by air pollution.
This issue could be examined by following a cohort of subjects in time within a community with respect to their formal interactions with the health care system through their billing records. This approach is similar to that used in the Montreal study except that the design is prospective instead of the retrospective approach used with the Montreal data. The adequacy of the time series studies to provide reasonable estimates of relative risk could be examined by comparing estimates of the relative risk of death based on air pollution exposure just prior to death from the prospective cohort study and the air pollution effect estimates from a time series version of the cohort data (i.e. comparing numbers of daily deaths with air pollution levels).
Another epidemiological design to assess the amount of life-shortening due to air pollution exposure is to follow subjects in time from several communities with varying levels of ambient air pollution concentrations. Typically, subjects are enrolled in the study with baseline information of several mortality risk factors (i.e. tobacco, food and alcohol consumption, occupational exposures, and social-economic status). Their vital status is ascertained after a specified time by linkage with death certificates. Survival analysis methods are used to determine the relative risk of air pollution related deaths. Estimates of changes in life span can be directly determined using this approach. In some cases, no additional information is obtained during follow-up other than vital status, however, some studies record residence changes and obtain longitudinal information of health status and risk factors. Data on ambient air pollution levels is recorded during follow-up and some historical information may be also available.
Estimates of Quality-Adjusted Life Expectancy (QALE) which incorporate age-specific quality of life or utility values based on health status may also be obtained. Such utility values can be obtained for the Canadian population from the National Population Health Survey (NPHS), a population-based longitudinal survey of over 20,000 households beginning in 1994. The survey consists of personal interview obtained information on health status and determinants of health. Subjects are reinterviewed every two years even if they have moved. The QALE will be less than life expectancy since the utility values range between zero and one.
A more advanced approach would incorporate the effects on QALE of air pollution both from mortality and from morbidity. This approach requires age-specific estimates of the association between air pollution exposure and the prevalence of disease (i.e. coronary heart disease, chronic obstructive lung disease) which could be determined from baseline information obtained in the cohort studies (i.e. linking air pollution levels to the prevalence of disease cross-sectionally). The prevalence of disease in the Canadian population could be obtained from the NPHS. Reductions in air pollution would then yield both increases in longevity and decreases in disease rates, thus expanding the length of time a subject would be a better health state, further translating into increased utility values. QALE determined using this latter approach could be larger than life expectancy.
Canada-Wide Standards have recently been set for ground-level ozone and particulate mass less than 2.5 microns in diameter (PM2.5) based on Health Canada's assessment that these pollutants pose an adverse health risk to Canadians. However, PM2.5 is composed of dozens of compounds including transition metals, organic, inorganic, and elemental carbon, volatile organic compounds, biological matter, and ions such as sulfate and nitrate. The chemical composition varies by particle size. The size of the particle can determine the probability of deposition in the lung, with differing toxic potential depending on the point of deposition.
The number of particles (number increases as size decreases) and their acidity have also been identified as potential toxic modifiers. Gas phase pollutants have also been linked to adverse health effects in Canadian cities, and the interactive role of gas and particle phase pollutants has been identified as a priority research issue. However, some questions have been raised as to the possibility that the associations between gas phase pollutants and health may be due to the gaseous pollutants acting as surrogates for ultra-fine particle phase contaminants. Toxicology has now identified several components of particulate matter that induce acute biological changes, which in turn may result in exacerbation of heart problems.
The relative toxicity of the urban pollution mix can be determined by comparing daily variations in mortality, hospital admissions, and emergency room visits for heart and lung problems and daily fluctuations in the components of the atmospheric mix. Environment Canada has established two "super stations", which will provide daily data on size fractionated elemental and carbon concentrations and particle counts in Toronto and Vancouver. Information from these population epidemiology studies can then be integrated with that from toxicology to better identify the most harmful components of the mix.
This process is important in order to develop mitigation strategies that target pollutant reductions that will yield the most benefit to population health.
The components of the atmospheric mix tend to be correlated over time since several pollutants are emitted from common sources with daily concentrations highly influenced by weather. It is thus desirable to have monitoring and health surveillance in several regions of the country with varying levels of pollution and atmospheric mixtures. It is also desirable to move from event-oriented health data to person-oriented information in which every formal contact with the health care system (physician and emergency room visits, hospital admissions, and death) are linked by the individual. Here, health profiles of individuals could be created and linked to air pollution exposures, thus potentially identifying those subjects who are most at risk.
Although there is some evidence that living in a polluted environment reduces life span, the association between long-term exposure and the development of heart and lung disease is much less clear. The NPHS provides an opportunity to examine this hypothesis. Disease status is ascertained by personal interview every two years. Linking residence with outdoor pollution monitors provides a means of determining exposure. This study also offers a unique opportunity to examine the effect of age-specific pollution exposure profiles on development of disease and longevity since subjects are tracked throughout the study period. Estimates of these relevant "exposure-time windows"can be made by examining subjects who live in varying air quality environments in different periods of their life.
However, additional information that is not routinely collected by the NPHS would be of value. For example, information on residence history would allow for improved estimates of long-term exposure to air pollution. Time activity patterns of study subjects would also contribute to improved exposure assessment. More detailed information on occupation and diet would also be of value, as would data on personal exposure and its relation to concentrations obtained from fixed site ambient monitors.
It is now widely recognized that ambient air pollution poses a health hazard for Canadians and that steps need to be taken to improve air quality. Canada-Wide Standards for particulate matter and ozone have now been established. Particulate matter has also been declared toxic under the Canadian Environmental Protection Act. Such a designation requires pollution management strategies to be developed. These strategies can be highly complex since particulate matter is generated from numerous sources, each with its own complex chemistry and potential toxicity. Science has an important role to play in the development of such strategies in order that the most toxic components are targeted.
You will find here a series of pictures of Health Canada research facilities located in the Environmental Health Center in Tunney's Pasture, Ottawa. Also included are the research facilities located in the Gage Occupational and Environmental Health Unit, University of Toronto and St. Michael's Hospital, research partners of Health Canada.
This controlled human exposure facility and the particle concentrator are housed at the Gage Occupational and Environmental Health Unit, University of Toronto. Health Canada scientists conduct human exposure studies in partnership with the Gage Occupational & Environmental Health Unit and St. Michael's Hospital. The facility is utilized to investigate clinical, physiological, and biochemical effects of urban particulate matter and gaseous air pollutants.
A volunteer sits in a controlled human exposure chamber, and inhales a known concentration of urban particulate matter for a short period of time. The concentration of the particulate matter is expected to cause some mild and transient effects on the individual. Working together, Health Canada scientists and the research team at the Gage Occupational and Environmental Health Unit examine the mechanisms and the dose-response relationship of urban particle-induced adverse health effects.
In order to recreate polluted urban air in the laboratory, investigators at Health Canada collect large quantities of particles from the outdoor air and analyse the particles for their chemical composition. The dust is then fractionated by size and stored. The fine dust, which contains metals and organic species from the environment, can then be suspended in air to recreate episodes of air pollution in the inhalation exposure chambers. Other contaminants such as ozone, oxides of nitrogen and carbon monoxide can be mixed in the air along with the particles in order to mimic a more complex atmosphere.