Nitrogen dioxide (NO2) is a ubiquitous atmospheric pollutant due to the widespread prevalence of both natural and anthropogenic sources, and it can be a respiratory irritant when inhaled at elevated concentrations. Evidence for health effects of ambient NO2 derives from three types of studies: observational epidemiology, human clinical exposures, and animal toxicology. Our review focuses on the human clinical studies of adverse health effects of short-term NO2 exposures, given the substantial uncertainties and limitations in interpretation of the other lines of evidence. We examined more than 50 experimental studies of humans inhaling NO2, finding notably that the reporting of statistically significant changes in lung function and bronchial sensitivity did not show a consistent trend with increasing NO2 concentrations. Functional changes were generally mild and transient, the reported effects were not uniformly adverse, and they were not usually accompanied by NO2-dependent increases in symptoms. The available human clinical results do not establish a mechanistic pathway leading to adverse health impacts for short-term NO2 exposures at levels typical of maximum 1-h concentrations in the present-day ambient environment (i.e., below 0.2 ppm). Our review of these data indicates that a health-protective, short-term NO2 guideline level for susceptible (and healthy) populations would reflect a policy choice between 0.2 and 0.6 ppm. EXTENDED ABSTRACT: Nitrogen dioxide (NO2) is a ubiquitous atmospheric pollutant due to the widespread prevalence of both natural and anthropogenic sources, and it can be a respiratory irritant when inhaled at elevated concentrations. Natural NO2 sources include volcanic action, forest fires, lightning, and the stratosphere; man-made NO2 emissions derive from fossil fuel combustion and incineration. The current National Ambient Air Quality Standard (NAAQS) for NO2, initially established in 1971, is 0.053 ppm (annual average). Ambient concentrations monitored in urban areas in the United States are approximately 0.015 ppm, as an annual mean, i.e., below the current NAAQS. Short-term (1-h peak) NO2 concentrations outdoors are not likely to exceed 0.2 ppm, and even 1-h periods exceeding 0.1 ppm are infrequent. Inside homes, 1-h NO2 peaks, typically arising from gas cooking, can range between 0.4 and 1.5 ppm. The health effects evidence of relevance to ambient NO2 derives from three lines of investigation: epidemiology studies, human clinical studies, and animal toxicology studies. The NO2 epidemiology remains inconsistent and uncertain due to the potential for exposure misclassification, residual confounding, and co-pollutant effects, whereas animal toxicology findings using high levels of NO2 exposure require extrapolation to humans exposed at low ambient NO2 levels. Given the limitations and uncertainties in the other lines of health effects evidence, our review thus focused on clinical studies where human volunteers (including asthmatics, children, and elderly) inhaled NO2 at levels from 0.1 to 3.5 ppm during short-term ((1/2)-6-h) exposures, often combined with exercise, and occasionally combined with co-pollutants. We examined the reported biological effects and classified them into (a) lung immune responses and inflammation, (b) lung function changes and airway hyperresponsiveness (AHR), and (c) health effects outside the lungs (extrapulmonary). We examined more than 50 experimental studies of humans inhaling NO2, finding that such clinical data on short-term exposure allowed discrimination of NO2 no-effect levels versus lowest-adverse-effects levels. Our conclusions are summarized by these six points: For lung immune responses and inflammation: (1) healthy subjects exposed to NO2 below 1 ppm do not show pulmonary inflammation; (2) at 2 ppm for 4 h, neutrophils and cytokines in lung-lavage fluid can increase, but these changes do not necessarily correlate with significant or sustained changes in lung function; (3) there is no consistent evidence that NO2 concentrations below 2 ppm increase susceptibility to viral infection; (4) for asthmatics and individuals having chronic obstructive pulmonary disease (COPD), NO2-induced lung inflammation is not expected below 0.6 ppm, although one research group reported enhancement of proinflammatory processes at 0.26 ppm. With regard to NO2-induced AHR: (5) studies of responses to specific or nonspecific airway challenges (e.g., ragweed, methacholine) suggest that asthmatic individuals were not affected by NO2 up to about 0.6 ppm, although some sensitive subsets may respond to levels as low as 0.2 ppm. And finally, for extra-pulmonary effects: (6) such effects (e.g., changes in blood chemistry) generally required NO2 concentrations above 1-2 ppm. Overall, our review of data from experiments with humans indicates that a health-protective, short-term-average NO2 guideline level for susceptible populations (and healthy populations) would reflect a policy choice between 0.2 and 0.6 ppm. The available human clinical results do not establish a mechanistic pathway leading to adverse health impacts for short-term NO2 exposures at levels typical of maximum 1-h concentrations in the present-day ambient environment (i.e., below 0.2 ppm).