Results

Of 8583 Tasmanian school children enrolled in 1968 at age 7 years, 5729 were included in the 2002 proband study (figure 1), of whom 338 were excluded because the participant did not provide a residential address allowing ambient air pollution exposure to be estimated. Of the remaining 5391 participants, 2238 were excluded because they did not participate in the 2012 proband study. Of the 3153 with data from both the 2002 and 2012 proband studies, 2369 had a valid SPT result. There were no important differences between those followed and those lost to follow-up, except that those lost were more likely to be smokers or exposed to smoke, less likely to report hay fever and being from lower socioeconomic status (Supplementary table 2).
The mean age at the 2012 follow-up was 53 years and 50.3% were males (Table 1). At 53 years 281 participants (8.96%) of the participants had prevalent eczema, 115 participants (3.67%) had incident current eczema and 201 participants (6.38%) had persistent eczema (supplementary table 3). When comparing baseline and follow-up concentrations, ambient air pollution markers decreased slightly over time (table 1).

Association between ambient air pollution and prevalent eczema at age 53.

There was evidence that sex modified the association (P for interaction<0.1) between NO2 and prevalent eczema (table 2). Thus, in males, baseline exposure to NO2 was associated with increased risk of having prevalent eczema at follow-up (adjusted odds ratio (aOR): 1.15 [95%CI 0.98-1.36] per IQR [2.27ppb] NO2 increase), while higher exposure in females was associated with reduced risk of prevalent eczema (aOR: 0.83 [95%CI 0.67-1.03 per 2.27 ppb NO2increase).
Likewise, associations also differed by sex when prevalent eczema was classified by atopy status (table 3). In males, baseline NO2 exposure was associated with increased risk of having both NAE (aOR 1.39 [95%CI 1.02-1.90]) and AE (aOR 1.26 [1.00-1.59]) per 2.27 ppb NO2 increase, while baseline PM2.5 was associated with increased risk of having AE (aOR 1.47 [1.04-2.06]) per 1.56 µg/m3PM2.5 increase. By contrast, in females, higher NO2 exposure at baseline was associated with a reduced risk of having AE (aOR 0.65 [95% CI 0.43-0.99]) per 2.27 ppb NO2 increase. A similar trend was seen at follow-up, NO2 was associated with increased risk of prevalent NAE (aOR 1.37 [95%CI 1.03-1.83] per 2.27 ppb) in males while in females it was a protective of NAE (aOR 0.67 [0.44-1.01]) per 2.21 ppb increase (Table 3). When the same associations were assessed restricting only to those who did not changed their address (non-movers), only negligible variation of the estimated effects were shown (supplementary table 4).

Association between ambient air pollution at baseline and incident current eczema.

There was weak evidence of associations, nor effect modification by sex, between baseline ambient air pollution markers at baseline and incident current eczema (Table 2). Similarly, when incident current eczema was classified by atopy groups (Table 4), there was weak evidence of association.

Association between ambient air pollution at baseline and persistent current eczema.

There was evidence that sex modified the association (P for interaction<0.1) between NO2 exposure and as a result, increased risk of persistent eczema was stronger in males than females (Table 2). Likewise, when persistent eczema was classified by atopy status (Table 5), there was evidence of interaction by sex (P for interaction<0.1). As such, NO2 exposure in males was found to increase persistent AE risk (aOR 1.25 [95% CI 0.95-1.65] per 2.27 ppb NO2 increase, while in females, the association tended towards reduced risk of persistent AE (aOR 0.57 [0.36-0.91] per 2.27 ppb NO2 increase). Similarly, PM2.5 exposure in males increased the risk of persistent AE (aOR 1.53 [1.04-2.25] per 1.56 µg/m3PM2.5 increase) and in females the risk of persistent AE was reduced (aOR 0.71 [0.49-1.02] per 1.56 µg/m3PM2.5 increase).

Restricted definition of non-atopic eczema

When using the same analytic approach as the SALIA cohort study6 there were no significant associations between ambient air pollution markers at baseline and incident NAE with increasing the strictness (i.e. those participants without hay fever and SPT negative) of the definition (supplementary table 5). However, a non-significant trend between DMR at follow-up and increased risk of prevalent NAE was seen, and this effect became stronger when the NAE definition became stricter (supplementary table 5). Furthermore, when the analyses were restricted to women, there was an association between DMR at follow-up and increased risk of prevalent eczema (supplementary table 6).
Association between ambient air pollution at baseline and aeroallergen sensitisation at age 53 years.
At baseline, increased exposure to PM2.5 was associated with increased odds of being sensitized to aeroallergens (aOR 1.15 [95% CI 1.03-1.30]) per 1.56 µg/m3PM2.5 increase) at age 53 (supplementary table 7). Furthermore, there were associations between ambient air pollution markers and specific aeroallergen sensitisations that are described in more detail in supplementary table 8.