LAUSR

BACKGROUND Previous studies have indicated that residential greenness can affect human health, but limited studies have examined the association between residential greenness and glucose homeostasis during pregnancy. We aimed to investigate the associations of residential greenness with plasma glucose, glycosylated hemoglobin (HbA1c) and gestational diabetes mellitus (GDM). METHODS We recruited a total of 587 pregnant women aged 20-45 years in Guangzhou, China. We extracted normalized difference vegetation index with different buffers (NDVI-250m, 500m and 1000m) from remote satellite data based on maternal residential addresses. We measured plasma glucose levels and HbA1c during 20-28 weeks' gestation, and GDM was diagnosed with a 75-g oral glucose tolerance test. We collected the information of relevant covariates by face-to-face interviews and questionnaires. We used generalized linear regression to examine the associations of residential greenness with quantitative and categorized outcomes. RESULTS In the final analyses, 123 (21.0%) of the pregnant women were diagnosed as GDM at 20-28 weeks gestation. With a 0.1 unit increase in NDVI-250m, the percent of HbA1c changed by -0.05% [95% confidence interval (CI): -0.08, -0.02]. The results of HbA1c were consistent when using different resolution of NDVI [NDVI-500m: 0.03, 95%CI: -0.05, -0.01; NDVI-1000m: 0.05, 95%CI: -0.08, -0.02]. We observed non-significant associations of glucose levels and the risk of GDM in relation to NDVI with different resolutions when adjusted for confounding. The results remained robust in sensitivity analyses. CONCLUSION The present study in Guangzhou, China was the first to identify negative association of NDVI with HbA1c in pregnant women, but we did not observe its association with plasma glucose levels or the risk of GDM. The results support that building sufficient green infrastructure could be considered in urban design and planning to promote maternal health.