Evaluating Health Effects of Single Pollutants in a Multipollutant World: Results of Recent Integrated Science Assessments for CO and NO2

TOM LONG (1), Tom Luben (1), Dennis Kotchmar (1), Mary Ross (1), Jennifer Richmond-Bryant (1), Doug Johns (1), Barbara Buckley (1), Elizabeth Oesterling Owens (1), James Brown (1), Ellen Kirrane (1)

(1) USEPA National Center for Environmental Assessment

     Abstract Number: 135
     Last modified: November 6, 2009

Abstract
Recent Integrated Science Assessments (ISAs) prepared by EPA’s National Center for Environmental Assessment (NCEA) have evaluated the health effects attributed to the traffic-related pollutants carbon monoxide (CO) and nitrogen dioxide (NO2). This integration process is a form of qualitative analysis that considers the evidence from atmospheric sciences, exposure assessment, dosimetry, and health studies (e.g., toxicological, controlled human exposure, and epidemiologic studies) to determine the likelihood of a causal relationship for the independent health effects of ambient air pollutants. NCEA has developed a causality framework for integrating evidence from multiple disciplines which considers the weight of evidence along with associated uncertainties. During the evaluation of CO and NO2 studies for their respective ISAs, several sources of uncertainty were identified that complicated the quantitative interpretation of evidence from controlled human exposure, epidemiologic, and toxicological studies. Many of these uncertainties are related to the fact that CO and NO2 exist as part of a mixture of combustion- and traffic-related gaseous and particulate pollutants. Examples of these uncertainties include the ability of community fixed-site monitors to represent spatially variable ambient concentrations and personal exposures, the impact of nonambient exposure on health outcomes, the possibility that a pollutant is acting as a surrogate for a mixture of combustion-related pollutants, and the biological plausibility of effects occurring at low ambient concentrations. Such uncertainties make it difficult to interpret the results observed across epidemiologic studies. New research summarized in recent ISAs has reduced uncertainty regarding the influence of nonambient exposure on effect estimates in epidemiologic studies. In addition, studies have provided evidence for consistency of effect estimates across studies and robustness of effect estimates in copollutant models for specific health outcomes. Furthermore, the coherence between outcomes observed in epidemiologic studies and the results of controlled human exposure or toxicological studies lends biological plausibility for CO- and NO2-induced health effects. For example, the ISA for CO presents the results of new epidemiologic studies that provide evidence for an association between short-term ambient CO exposure and hospitalizations for coronary artery disease (CAD). This is consistent with evidence from controlled human exposure studies demonstrating cardiovascular effects in individuals with CAD following CO exposures resulting in carboxyhemoglobin (COHb) concentrations as low as 2.0 – 2.4%. The known role of CO in limiting O2 availability lends biological plausibility to CAD outcomes following CO exposure. Recent toxicological studies suggest that CO also may act through other mechanisms by initiating or disrupting cellular signaling. When the evidence from multiple disciplines is aggregated and evaluated against the causality framework, it supports the likely causal determination made for CO and cardiovascular effects in the ISA. Evidence is combined in a similar manner to formulate the likely causal determination for short-term ambient NO2 exposure and respiratory morbidity. This integrated causality framework will be presented to elucidate the most recent evidence available regarding the independent health effects of CO and NO2, along with uncertainties that complicate quantification of the magnitude of health effects at current ambient concentrations.