Association of residential greenness and incident depression: Investigating the mediation and interaction effects of particulate matter.

BACKGROUND

Growing evidence has linked residential greenness to depression, the results from prospective cohort study are still limited. And it remains unclear whether particulate matter (PM) modify, mediate, or interact the greenness-depression relationship.

METHODS

We collected data from Yinzhou Cohort(N = 47,516) which was recruited between June 2015 and December 2017. Depression cases before April 2020 were ascertained from local Health Information System covered all residents' health care records. Residential greenness (the Normalized Difference Vegetation Index, NDVI, and the Enhanced Vegetation Index, EVI) and PM (particulate matters with diameters≤2.5 μm, PM and particulate matters with diameters≤10 μm, PM) were estimated based on participants' residential coordinates. We conducted Cox models employing age as timescale to estimate the association between residential greenness within different buffers and incident depression. Furthermore, we explored the potential confounding, mediation and interaction relationship between greenness and PM.

RESULTS

During the 99,556 person-years of follow-up, 1043 incident depression cases occurred. In single exposure models, residential greenness was inversely associated with depression incidence (e.g. Hazard Ratio (HR) = 0.86, 95% confidence interval (CI): 0.79, 0.94 for per interquartile range (IQR) increase NDVI 250 m). The protective association between greenness was attenuated after introducing PM and PM into the models. We identified multiplicative interactions between greenness and PM exposure for depression (e.g. HR  = 0.91, 95%CI: 0.85, 0.98 for per IQR decrease NDVI 250 m and per IQR increase PM). Besides, we found the protective association of greenness was partly mediated by PM (e.g. mediation proportion = 52.9% between NDVI 250 m and PM).

CONCLUSIONS

In this longitudinal cohort study, residents living in greener neighborhoods had a lower risk of depression incidence and the benefits were interacted and partly mediated by PM. Improvement in residential greenness could be an actionable and planning intervention to prevent depression.