It was one of few studies exploring the relationships between green space and hypertension and blood pressure in developing countries. The effects of green space may reduce the risk of hypertension. Besides, sex, smoking, and drinking could further modify the associations. The studies of associations between green space and blood pressure are still in the initial stage and lack of systematic studies in developing countries. Given increasing environmental pollution accompanied by the accelerated urbanization process and high prevalence of hypertension in China as well as other developing countries, our findings may be particularly important for public health.
Some studies reported similar results [6, 8, 13, 27–30]. Our study showed that an IQR (0.08 units) increment in NDVI500m was significantly associated with an 8% decrease in odds of hypertension. A study conducted in the urban area of the United States including 249, 405 indicated that a 0.1-units increase in NDVI1000m was related to a 7% decrease in odds of hypertension . Additionally, AM Dzhambov, I Markevych and P Lercher  reported that each IQR (0.16 units) increase in NDVI500m was related to a 36% decrease in odds of hypertension based on a study of 555 adults in an Alpine valley of Austria. However, some studies showed inconsistent findings. A study including 3,063 women in Germany reported null associations . A meta-analysis of 4 studies reported no significant association of greenness with hypertension incidence . Differences in characteristics of participants, study setting, greenness exposure assessment, statistical models, and adjustment for covariates may explain the inconsistency.
We found that higher green space was related to lower blood pressure. Our findings were in line with a cross-sectional study of 3,150 in India  showing an interquartile range increase in NDVI250m related to a decrease in SBP of 4.3 mmHg and a decrease in DBP of 1.2 mmHg, respectively. However, several previous studies showed inconsistent results [7, 8, 18, 21]. BY Yang, I Markevych, MS Bloom, J Heinrich, Y Guo, L Morawska, SC Dharmage, LD Knibbs, B Jalaludin, P Jalava, et al.  reported that higher green space was related to lower SBP, whereas a null association was found between NDVI500m and DBP. A study including 427 newborns in Belgium reported that an IQR (20.3%) increment in green space (5000 m buffer) was related to decrease in DBP of 1.2 mm Hg (95% CI: − 2.4, − 0.0), whereas no association was found for SBP (− 1.2 mm Hg, 95% CI: − 2.5, 0.1) . LD Bloemsma, U Gehring, JO Klompmaker, G Hoek, NAH Janssen, E Lebret, B Brunekreef and AH Wijga  conducted a study of 1,505 children, which found that green space was not significantly related to changes in blood pressure. The study participants listed above (e.g., adults, newborns, and children) have different characteristics, lifestyles, living, and working environments that may modify the effects of greenness. Thus, the relationships between green space and blood pressure remained unclear.
Residing in green space may improve cardiovascular disease outcomes . The mechanisms underlying the association may be explained by immunological and psychological pathways [6, 33, 34]. Specifically, people residing in green space are exposed to more diverse microbes beneficial to the host immune system (improving immune regulation) and less noise and air pollution (reducing inflammatory response), are more likely to increase physical activity (strengthening immune and nervous system), promote the exchange of microbiota (increase social interactions), and promote metabolism (sunlight helps in the synthesis of vitamin D) . Higher green space may benefit mental health such as depression [35, 36].
The subgroup analyses indicated that the effects of green space were more pronounced in males, smokers, and drinkers. The results of sex modification in previous studies have been inconsistent. X Jia, Y Yu, W Xia, S Masri, M Sami, Z Hu, Z Yu and J Wu  suggested that green space was related to lower hypertension prevalence for males (OR = 0.18, 95% CI: 0.09, 0.27) and for females (OR = 0.78, 95% CI: 0.48, 0.99). However, BY Yang, I Markevych, MS Bloom, J Heinrich, Y Guo, L Morawska, SC Dharmage, LD Knibbs, B Jalaludin, P Jalava, et al.  reported that 0.17 units increase in NDVI500m was related to lower hypertension prevalence for females, whereas a null association was found for males. Inconsistent results may due to them that women have a higher frequency of using green space in China, for example, square dancing was the most popular exercise among them . AM Dzhambov, I Markevych and P Lercher  concluded that there was no evidence of significant interactions between sex and greenness-hypertension pathway. There has been limited evidence of how behavioral factors affect the relationships between green space and hypertension and blood pressure. Existing evidence showed that green space could provide diverse microbes, some of which are important inducers of the immunoregulatory pathways, and activation of the immune regulatory system consequently reduced chronic inflammation . C Menni, C Lin, M Cecelja, M Mangino, ML Matey-Hernandez, L Keehn, RP Mohney, CJ Steves, TD Spector, C-F Kuo, et al.  indicated the significant relationships between gut microbial diversity and lower arterial stiffness. Smoking and drinking could trigger inflammatory response, oxidative stress, and metabolic disorders, which further affect systemic vascular resistance [38, 39]. Thus, the effects of green space on inflammatory response and immune system could be more pronounced in smokers and drinkers than normal people.
Several limitations existed in this study. First, as an inherent drawback of cross-sectional design, causal associations should be treated with caution. Second, demographic covariates, socioeconomic covariates, and health behavior covariates were collected using the questionnaire, which may introduce recall bias. Third, some potential confounders including traffic noise, the walkability of a community, psychological status, and indoor greenness exposure were not adjusted in the model because of data unavailability [8, 18].