2. How does Nitrogen Dioxide (NO₂) affect human health?

2.1 Effects of long-term exposure to levels of NO₂ observed currently in Europe
          2.1.1 Chronic effects
          2.1.2 Effects on mortality

2.2 Is NO₂ per se responsible for effects on health?
2.3 Are health effects of NO₂ influenced by the presence of other air pollutants?
2.4 Characteristics of individuals that may influence how NO₂ affects them
2.5 Is there a threshold below which nobody’s health is affected by NO₂?

2.1 Which effects can be expected of long-term exposure to levels of NO₂ observed currently in Europe?

2.1.1 Chronic effects at current NO2 levels
2.1.2 Effects on mortality at current NO₂ levels

2.1.1 Chronic effects at current NO₂ levels

WHO states:

"Answer:

The epidemiological studies provide some evidence that long-term NO₂ exposure may decrease lung function and increase the risk of respiratory symptoms.

Rationale:

Although there are fewer epidemiological studies on long-term respiratory effects of NO₂ than those of particulate matter, new evidence has been provided in recent years. Both cross-sectional and longitudinal studies indicate an association between NO₂ and lung function. The Southern California Children’s Study showed that lung function levels among 9 to 16 year old children were lower in communities with higher NO₂ concentration (236). Lung function growth, evaluated in a longitudinal study, was also impaired among these children (22, 23). The NO₂ effect in the cohort study was robust when other pollutants (e.g. PM₁₀ and O₃) were included in the statistical model, but weakened when acid vapours (including NO₂ derived nitric acid) were simultaneously considered. The cross-sectional SAPALDIA Study in Switzerland (93, 392) gives support to the association of NO₂ exposure and lung function decrements among adults.

Two cross-sectional studies among children (79, 393) provide some evidence of an association between NO₂ and acute bronchitis, while the Southern California Children’s Study suggested that chronic respiratory symptoms (cough and phlegm) were more frequent among children with asthma in communities with higher NO₂ exposure (88). Two cross-sectional studies found an association between NO₂ and cough and phlegm symptoms in adults (94, 394).

In most of these studies, NO₂ concentrations at the community level were correlated with PM and ozone, making it difficult to disentangle an effect of NO₂ per se. In the SAPALDIA study (392), however, there was a clear association of personal exposure to NO₂ with lung function (FVC and FEV1) within the same communities of presumably rather homogeneous PM concentrations."

Source & © : WHO Regional Office for Europe "Health Aspects of Air Pollution" (2003),
Chapter 7 Nitrogen dioxide, Section 7.2 Answers and rationales, Question 2

2.1.2 Effects on mortality at current NO₂ levels

To what extent is mortality being accelerated by long and short-term exposure to NO₂?

WHO states:

"Answer:

Methodological limitations constrain identification of harvesting due to NO₂ itself. The few long-term studies have not shown evidence for association between NO₂ and mortality. Associations have been observed between NO₂ and mortality in daily time-series studies, but on the basis of present evidence these cannot be attributed to NO₂ itself with reasonable certainty.

Rationale:

The number of cohort studies investigating associations of long-term exposure to NO₂ in ambient air with mortality is small. Neither the Six Cities Study with data from a cohort of 8 111 individuals followed over 15 years (82) nor the American Cancer Study with data from 500 000 individuals followed over 15 years (13) nor the ASHMOG study with 6 338 non-smoking Adventists in California (9) found an association of mortality risks with outdoor measured NO₂. A recent publication from the Netherlands Lung Cancer Study with a sample of 5 000 people followed over 8 years investigated mortality associations with long-term exposure to black smoke and nitrogen dioxide concentrations at the subjects’ home addresses. Cardiopulmonary mortality was associated with background concentration of both pollutants but more consistently with living near a major road (12). These results point to the importance of road traffic exhaust, for which NO₂ may be an indicator.

A Czech case-control study on air pollution and infant respiratory mortality showed associations with TSP and NO_x (75). The findings are possibly mediated by a higher susceptibility for infections in more polluted air as observed in time-series studies with hospital admissions and mortality data of infants and cannot be attributed to NO_x alone.

Many recent publications on time-series studies report associations between one hour or 24-hour average concentrations of outdoor NO₂ and mortality modelling ambient NO₂ in single pollutants models. The NMMAPS Study with data for NO₂ from 19 cities in the United States of America did not find such an association (27). The European study on short-term exposure to air pollution and mortality and morbidity, APHEA, investigated data from 29 cities and found heterogeneity between the cities. In addition, higher signals for PM₁₀ were detected in cities with higher mean NO₂ levels (29). The reason for this finding is currently unclear. It could, for example, be due to interactions in atmospheric chemistry or at the pathophysiological level.

A Canadian research team published meta analyses on several gaseous pollutants and particles with mortality data from 109 studies published between 1982 and 2000 summing up 32 different estimates for NO₂ from single pollutant models. They found an overall effect estimate for NO₂ of similar magnitude as for PM₁₀, ozone or carbon monoxide (391). A larger effect size was observed for respiratory mortality. In the multi-pollutant model the effect size for total mortality dropped to one third and became non-significant, supporting the view that the concentration- response signal generated for NO₂ may largely be the consequence of exposure to other pollutants related to NO₂."

Source & © : WHO Regional Office for Europe "Health Aspects of Air Pollution" (2003),
Chapter 7 Nitrogen dioxide, Section 7.2 Answers and rationales, Question 5

2.2 Is NO₂ per se responsible for effects on health?

WHO states:

"Answer:

The evidence for acute effects of NO₂ comes from controlled human exposure studies to NO₂ alone. For the effects observed in epidemiological studies, a clear answer to the question cannot be given. Effects estimated for NO₂ exposure in epidemiological studies may reflect other traffic related pollutants, for which NO₂ is a surrogate. Additionally there are complex interrelationships among the concentrations of NO₂, PM and O₃ in ambient air.

Rationale:

Controlled human exposure studies have been used to investigate the effects of NO₂ per se and were used as the basis for establishing the current 1-hour guideline value.

In epidemiological studies, NO₂ concentrations are often highly correlated with levels of other ambient pollutants either being emitted by the same sources or related through complex atmospheric reactions. NO₂ has been found to be an indicator for (often unmeasured) traffic related pollutants such as organic and elemental carbon or freshly emitted primary ultrafine particles. This view is supported by analyses of sources of PM including gaseous pollutants where NO₂ is found to correlate well with traffic (406). In addition, NO₂ might be a marker for the contribution of NO_x to the formation of secondary pollutants such as secondary particles and O₃.

It is important to note that measurement errors with regard to population average exposure will be important when comparing effects of different pollutants. In time-series analyses, pollutants with homogenous within-city distributions will be inherently favoured over pollutants with an inhomogeneous distribution. Thus while fine particulates may have a relatively homogenous distribution over wider distances, NO₂, NO, CO and ultrafine particulates may be strongly
“disadvantaged” as exposure variables in time-series analyses due to their much lager spatial variability (145). As already stated in response to question 5, a meta analysis of time-series investigations on mortality which included 109 studies published between 1982 and 2000 was conducted by Stieb et al (391). This analysis included 32 effect estimates for NO₂ from single- pollutant models and 15 from multi-pollutant models. Over a range of 24-hours average NO₂ concentration from 20 to 103 µg/m³, the overall effect estimate from the single pollutant model for all-cause mortality was 2.8+0.3% (mean + SEM) per 44 µg NO₂/m³, which fell to 0.9+0.5% in multi-pollutant models. The multi-pollutant models included particle measures and sometimes in addition O₃ or other gaseous pollutants, further supporting the view that the concentration response signal generated for NO₂ is largely, but not necessarily entirely, the consequence of other pollutants emitted by the same source or being derived from NO₂.

In long-term studies, the spatial inhomogeneity of NO₂ levels within a city might weaken the ability to detect effects of NO₂ based on measurements from urban background sites such as in the Harvard Six City Study or the American Cancer Society Study. However, NO₂ has been shown to be an appropriate indicator for assessing long-term exposure to traffic related air pollution. An example of a long-term study using this inhomogeneity is a recent European cohort study on mortality (12). The within-community variability was also used in the SAPALDIA Study (392) to capture the effects of NO₂ pollution beyond the variability that would exist for PM alone. However, associations found between ambient NO₂ concentration and health outcomes could also be explained by exposure to road traffic exhaust (79, 394, 407) and/or when it was available by exposure to particulate matter (79, 93, 94).

Therefore, the interpretation of the short-term as well as long-term epidemiological studies is that these results are not primarily due to NO₂ per se but to other unmeasured traffic related pollutants or to secondary pollutants, which have complex interrelationships with NO₂. Potential pollutants for which NO₂ might be an indicator include black smoke, organic and elemental carbon and ultrafine particles (see also section on PM)."

Source & © : WHO Regional Office for Europe "Health Aspects of Air Pollution" (2003), Chapter 7 Nitrogen dioxide, Section 7.2 Answers and rationales, Question 6

2.3 Are health effects of NO₂ influenced by the presence of other air pollutants?

WHO states:

"Answer:

There have been few controlled human exposure studies on interactions with other chemical pollutants, although several studies show that NO₂ exposure enhances responses to inhaled pollens. Some epidemiological studies have explored statistical interactions of NO₂ with other pollutants, including particles, but the findings are not readily interpretable.

Rationale:

The number of new studies using human controlled exposures to assess the effect of multiple pollutants including NO₂ is limited. The combination of NO₂ and O₃ was addressed by (408) showing a reduction in cardiac output which was largest for NO₂ in combination with O₃. Human controlled exposure studies have suggested that single and multiple controlled exposures to concentrations of NO₂ in excess of those normally achieved in ambient air can sensitize the airways of adult asthmatic subjects to the bronchoconstrictor effect of an inhaled allergen to which they are specifically sensitized (387, 388, 389, 409, 410, 411). The late response defined as 3 to 10 hours after administration of sub-acute allergen concentrations seemed to be more affected than the immediate response. Short term exposure to air pollutants (including NO₂) in a road tunnel has also been shown to enhance the asthmatic response to allergen (207). The mechanistic basis for these interactions has not been elucidated. The studies might suggest that NO₂ can exhibit a “priming” effect, by for example affecting epithelial function.

Epidemiological time-series analyses addressing possible synergisms between NO₂ and other pollutants have focused on the role of NO₂ as an effect modifier of the association between PM and health outcomes. In the APHEA2 study, PM associations with daily mortality were larger in areas with high NO₂ concentrations (29). In contrast in the NMMAPS study, no evidence for effect modification of the association between PM and daily mortality by NO₂ was found (27). A possible interpretation of the APHEA2 results could be that NO₂ might serve as an indicator for the presence of more toxic particles such as traffic related particles. In the light of the NMMAPS results however, the assumption needs to be made that in European cities NO₂ is a better traffic indicator than in North American cities. Generally, the statistical analyses assessing effect modification carry with them the inherent limitation that in complex mixtures the pollutants themselves are indicators and that measurement error further reduces the ability to disentangle their contributions as discussed in the rationale on question 6. With the level of evidence available it is not yet possible to state with a sufficient degree of confidence whether NO₂ is able to synergize with PM or to state that NO₂ is a valid indicator for more toxic PM present under these conditions."

Source & © : WHO Regional Office for Europe
"Health Aspects of Air Pollution" (2003), Chapter 7 Nitrogen dioxide,
Section 7.2 Answers and rationales, Question 8

2.4 Which characteristics of individuals may influence how Nitrogen Dioxide affects their health?

Are effects of NO₂ dependent upon the subjects’ characteristics such as age, gender, underlying disease, smoking status, atopy, education etc? What are the critical characteristics?

WHO states:

"Answer

In general, individuals with asthma are expected to be more responsive to short-term exposure to inhaled agents, when compared to individuals without asthma. Controlled human exposure studies of short-term responses of persons with and without asthma to NO₂ have not been carried out. There is limited evidence from epidemiological studies that individuals with asthma show steeper concentration-response relationships. Small-scale human exposure studies have not shown consistent effects of NO₂ exposure on airways reactivity in persons with asthma, even at exposure levels higher than typical ambient concentrations. As for other pollutants, children can reasonably be considered to be at increased risk. There is limited evidence for influence of the other listed factors on the effects of NO₂.

Rationale:

In healthy adults changes in lung function in experiments with controlled human exposure to NO₂ occur only at concentrations in excess of those normally encountered in ambient air. However, asthmatic subjects are characterized by having airways that are hyper-responsive to a wide variety of inhalation stimuli and, as a consequence, might be expected to exhibit a greater airways response to NO₂ than in normal subjects. Small scale human exposure studies in adult asthmatics with mild to moderate disease have failed to demonstrate consistent effects of NO₂ on either baseline airway calibre or on direct (e.g., methacholine) or indirect (e.g., SO₂, cold air) airway hyper-reactivity even at concentrations higher than those typically achieved in outside air (397, 398, 399, 400). It is noteworthy that there are no studies that have included both normal and asthmatic subjects in the same study, nor patients with severe disease.

Some cross sectional studies in adults and children have shown associations between ambient NO₂ concentrations and impaired lung function (93, 236, 392, 401, 402) but it is not possible to determine whether this is due to NO₂ itself or to the secondary pollutants that are derived from it. Several new longitudinal cohort studies have also shown associations between outdoor NO₂ concentrations and impaired growth in lung function (22, 23, 24, 235) but this effect is mostly weakened when the pollutant models take account of the effects of other outdoor pollutants such as ozone, particles or acid aerosols and indoor exposures. In the study of Horak et al. (235)), a seasonal difference was found with NO₂ enhancing the effect of PM₁₀ on lung function in the summer and vice versa in the winter.

Cross sectional studies using symptoms, lung function and hospital admissions have provided some evidence for an increased association between NO₂ exposure and asthma but the effects are not consistent (79, 94, 231, 233, 236, 393, 394, 402, 403, 404, 405). As with the cross section and cohort studies, NO₂ effects on asthma appear to be more prominent in children (231, 233, 236, 393, 402, 403, 405) than in adults (94, 394, 403) as observed for the aggravating effects of other air pollutants on asthma (79). As might be predicted, there are also greater associations between outdoor NO₂ exposure and respiratory outcome measures in children who spend more time outdoors (22, 24). Some epidemiological studies have reported gender effects of NO₂ on asthma or lung function changes but these are inconsistent.

There is also limited evidence that lower educational attainment is a risk factor for NO₂ with risk estimates that are independent of smoking, diet or alcohol, but less than observed for particulate matter and could be explained by increased exposure to air pollutants (12, 13, 392)."

Source & © : WHO Regional Office for Europe
"Health Aspects of Air Pollution" (2003), Chapter 7 Nitrogen dioxide,
Section 7.2 Answers and rationales, Question 4

2.5 Is there a threshold below which nobody’s health is affected by NO₂?

WHO states:

"Answer:

The evidence is not adequate to establish a threshold for either short or long-term exposure. While a number of epidemiological studies have described concentration-response relationships between ambient NO₂ and a range of health outcomes, there is no evidence for a threshold for NO₂.

Rationale:

As noted in the introduction, threshold points in dose-response relationships are not readily established, based on either experimental or observational data. For the acute effects of NO₂, the evidence comes primarily from human exposure studies at concentrations in the range of several hundred µg/m³ and higher. In general, current exposures in Europe are below this range. Thus, human exposure studies carried out with relatively small numbers of health[y] volunteers do not provide any evidence for the existence of thresholds in a range relevant to current standard setting.

For the effects of chronic exposure, the detection of possible thresholds is complicated by the relationship between NO₂ and the formation of O₃ and of secondary particulate matter. For these secondary pollutants, thresholds have not been demonstrated and adverse effects are observed at currently prevalent levels, even towards the lower end of the concentration range. For NO₂ itself, the relevant evidence comes primarily from a few selected studies of outdoor air pollution and also from the studies of indoor air pollution (395). Although the evidence needs to be interpreted with the above limitations in mind, several studies show associations of NO₂ with adverse health outcomes at current levels of exposure without strong evidence for a threshold. In a meta- analysis of time-series studies carried out recently (391), the overall effect estimate for NO₂ and all-cause mortality was positive. The SAPALDIA Study found evidence of adverse respiratory effects in adults in Switzerland in both cross-sectional and longitudinal data (94, 392, 396). The effects of NO₂ occurred at a mean concentration in the range of 30 µg/m³. The authors noted the difficulty of separating the effects of NO₂ from the effects of other pollutants. The indoor studies provide mixed evidence for effect of NO₂ and are inadequate for the purpose of determining a threshold for adverse effects."

Source & © : WHO Regional Office for Europe
"Health Aspects of Air Pollution" (2003), Chapter 7 Nitrogen dioxide,
Section 7.2 Answers and rationales, Question 3

See also: General Issues and Recommendations on Air Pollutants:

• question 1.3 on uncertainties in defining thresholds
• question 3.1 recommendations regarding the concept of threshold