|| Checking for direct PDF access through Ovid
Evidence of the impact of green spaces on pregnancy outcomes is still limited. We analysed green spaces and socio-economic position (SEP) as effect modifiers of the effect of high temperatures and air pollutants (PM10, NO2 and O3) on the risk of preterm birth (PB).A cohort of new-borns in Rome, from April to October, 2001–2013, were analysed. Pre-terms, identified through the Certificate of Delivery Care Registry, were defined as births between the 22nd and the 36th week of gestation. A time series approach was used, with maximum apparent temperature (MAT), PM10, NO2 and O3 as exposure variables. We used a lag of 0–2 days for all exposure when analysing preterm births, except for PM10 (lag of 12–22 days). As green indicators we considered both the distance between mothers’ residence address and green spaces and the Normalised Difference Vegetation Index (NDVI) within a 100 m buffer centred on woman’s residence address. Women were also classified according to age, SEP and education level.We enrolled 56 576 total births, 5.1% of which PB. In our cohort 24% of women were younger than 30 years, 36% had a low SEP and 33% completed primary school. We observed a 2.0% (95% CI: 0.7 to 3.2) increase in the daily number of PB per 1°C increase in MAT, adjusting by PM10. Among pollutants only PM10 was associated to a significant increase in PB (+0.7%; 95% CI: 0.1 to 1.3) per 1 mg/m3 increase in PM10 (adjusted by MAT). SEP was an effect modifier for both MAT/PB and PM10/PB relationship; MAT increased the risk of PB only among women of medium or low SEP while PM10 among those of high SEP. Green was an effect modifier of MAT/PB relationship, with the highest effect of MAT on PB among women living very close to green spaces (within 100 m).Socio-economic position resulted to be an important effect modifier for both MAT/PB and PM10/PB relationship. In particular we found the highest effect of temperature on preterm birth risk in women with low socio-economic position and living very close to green areas. How green acts in modifying this association should be further investigated.