Greenness, trees, allergenic trees and allergic outcomes: results from sensitivity analyses
Effect estimates were similar when we adjusted the three models with tree exposure variables additionally for NDVI (Table 3). Similar associations were observed for the 100 m (Table E3) and 1,000 m (Table E4) buffers, with some exceptions. In case of the 100 m buffer, associations between NDVI and asthma, the association between allergenic trees definition 1 and food sensitization, and the associations of total number of trees and NDVI with aeroallergen sensitization were not present. For the 1,000 m buffer, no associations with aeroallergen sensitization were observed and the association between medium NDVI and asthma was lost.
Effect estimates from crude and additionally adjusted models were very similar to those from the main models (Table E5). The same was true when children with partially missing outcome variables were excluded (data not shown). The associations between trees and asthma, allergenic trees definition 2 and asthma, and NDVI and aeroallergen sensitization got stronger with age. Effect modification by moving between birth and 2 years did not reveal a distinct pattern (Figure E3). Similarly, a mixed pattern was observed considering effect modification by residing close to structured green spaces (Figure E4).