Urban green space landscape patterns and preterm birth: a multicenter cohort study of over one million births in...

While the health benefits of urban green spaces for perinatal outcomes are established, evidence on how landscape configuration relates to preterm birth (PTB) and its subtypes remains limited. We estimated associations between landscape metrics of urban green space and PTB risk, and quantified potential mediating effects of ambient temperature and fine particulate matter (PM_2.5). We conducted a retrospective cohort study of 1,058,772 singleton live births in Chengdu, China (2014-2022). For each participant, we characterized green space within a 2-km residential buffer using three metrics: total area (TA), patch density (PD; fragmentation), and landscape shape index (LSI; morphological complexity). Multilevel logistic regression models were used to estimate associations with overall, early, and late PTB. We also performed stratified and mediation analyses to assess effect modification and indirect effects. Increases in TA and LSI were associated with reduced overall PTB risk (TA: odds ratios [OR] = 0.98, 95% confidence interval [CI]: 0.96-0.99 per 133.75-ha increase; LSI: OR = 0.98, 95% CI: 0.97-0.99 per 17.46-unit increase). These associations were primarily observed for late PTB. Mediation analyses revealed that PM_2.5 mediated 8.08% of the total effect of TA, 1.51% of PD, and 12.15% of LSI. The protective effects of TA and LSI were more pronounced in areas characterized by higher temperatures, lower PM_2.5, and higher ozone levels. Our results suggest that large and more morphologically complex green spaces may reduce PTB risk, partly through alleviating of air pollution and temperature, and highlight the importance of considering spatial configuration in urban planning for maternal health.