• asthma epidemiology
• paediatric lung disaese

For many, it seems self-evident that a greener city improves the quality of life and health for its citizens. Epidemiological studies are providing increasing evidence that the presence and amount of vegetation around locations where one spends a lot of time (home, work and school) have numerous beneficial effects on physical and mental health, including increased longevity.1 Interestingly, the evidence supporting a positive role of vegetation on allergic and respiratory health is much weaker, possibly because of the complex role played by air pollution. Beneficial effects attributed to vegetation may be due to less air pollution, as levels tend to be lower in vegetated areas. Some vegetation types may also actively reduce air pollution levels, although the scientific consensus is that this occurs only to a limited extent.2 Air pollutants may also interact with vegetation, such as pollen,3 to influence associations with health. These mechanisms are relevant in urban settings where air pollution levels can be high.

Various metrics are used to estimate ‘vegetation exposure’ in epidemiological studies. The Normalised Difference Vegetation Index (NDVI) is common and assesses surrounding levels of green vegetation, often termed ‘greenness’. The NDVI is calculated from satellite data giving the ratio of visible and near-infrared light reflection and ranges from −1 (water), 0 (barren rock) to +1 (dense green vegetation). It provides a measure of the average level of green vegetation present but no information on the vegetation species (eg, grasses, shrubs, trees), whether the greenness is accessible or frequently used (eg, public space, private gardens) or whether one might expect particular patterns of ambient pollen exposures. Other vegetation metrics include other indices (eg, Soil-Adjusted Vegetation Index, ‘naturalness’ index), the presence of and distance to structured green spaces (eg, urban parks, agricultural land, forests), per cent tree cover and the quantity and species of allergenic trees. Overall, there is a lack of widespread use and standardisation of these methods, which inhibits comparisons across studies.

In their Thorax study, Zhou et al 4 used data on 6740 schoolchildren (mean age 11.6 years) attending 48 elementary and middle schools in seven cities in northeastern China to determine whether those attending schools with more greenness (primarily assessed using the NDVI) within 500 metres had better lung function, and whether this was the same in areas with low or high air pollution levels. The authors found that an IQR increase in greenness was associated with higher FEV₁ (+57 mL, 95% CI 44 to 70) and FVC (+58 mL, 95% CI 43 to 73) and a corresponding reduced odds of spirometric restriction (OR: 0.63, 95% CI 0.53 to 0.76). These differences remained after adjustment for air pollution levels, except for PM₁. The authors went on to show that in areas with high PM₁ levels, children attending schools surrounded by more greenness had worse lung function compared with those attending schools with less greenness, and the observed benefits of greenness on lung function were limited to children attending schools in areas with lower levels of PM₁ (p interaction <0.05). A similar interaction was observed for SO₂ (for FVC) but not PM_2.5 or NO₂. The authors suggested that greening urban areas may promote lung health only in areas with low-moderate air pollution levels.

To our knowledge, only one other study has examined how associations between greenness and lung function in children differ across air pollution levels. In a paper published after the submission of Zhou et al,4 participants of the English ALSPAC birth cohort with higher greenness levels close to their home and nearby urban green spaces (ie, public green areas used predominantly for recreation) throughout their life had higher FEV₁ and FVC up to 24 years of age.5 Associations between repeated measures of the presence and proportion or urban green space were greater among those living in cities and in areas of high PM₁₀ levels. This latter result is in direct contrast to the results by Zhou et al as it suggests a greater beneficial importance of vegetation on lung function in urban and more polluted settings. One explanation for these conflicting conclusions may be that air pollution levels are much higher in the regions studied by Zhou et al 4 and their low/medium air pollution categories at which beneficial effects are observed likely correspond to the high category of Fuertes et al.5 This is difficult to confirm as different pollutants were assessed in the two studies. Nonetheless, it highlights one of the main strengths of Zhou et al—no study had yet examined how vegetation may affect respiratory health in areas with very high air pollution levels. Other potential reasons for the differing conclusions include geographical variation in the vegetation species present as well as the pollutants assessed and their sources, and that lung function was only measured once and at a younger age in Zhou et al.

To better understand their results, Zhou et al conducted mediation analyses, a technique increasingly used in epidemiological work to calculate how much of an observed association can be assumed to be explained by another factor. The authors found that less than 2% of the association between greenness and lung function was due to increased physical activity and lower body mass index in children attending schools in greener areas. This suggests that the health benefits of greenness are related to its impact on other factors such as early life events (eg, birth weight) or perhaps on asthma and allergic diseases (eg, wild vs urban vegetation, allergenic vegetation such as grasses). Associations between vegetation and asthma, allergic diseases and atopy are inconsistent,6 7 and the role of air pollution in these associations is complex (Markevych et al in press). Presumably, the authors did not know the vegetation types in their study areas, or whether these differed between low and high air pollution settings, thus they are unable to provide further explanations for their results.

Urban planners are seeking guidance from epidemiologists, health professionals, ecologists and many others as to how best to green our cities to maximise health, including respiratory health.8 Studies with detailed and informative vegetation exposure assessments are required for us to answer this call.