The association between neighborhood greenness and cardiovascular disease: an observational study

Abstract

Background: Previous studies have demonstrated links between cardiovascular disease and physical inactivity and poor air quality, which are both associated with neighborhood greenness. However, no studies have directly investigated neighborhood greenness in relation to coronary heart disease risk. We investigated the effect of neighborhood greenness on both self-reported and hospital admissions of coronary heart disease or stroke, accounting for ambient air quality, socio-demographic, behavioral and biological factors.

Method: Cross-sectional study of 11,404 adults obtained from a population representative sample for the period 2003-2009 in Perth, Western Australia. Neighborhood greenness was ascertained for a 1600 m service area surrounding the residential address using the mean and standard deviation of the Normalized Difference Vegetation Index (NDVI) obtained from remote sensing. Logistic regression was used to assess associations with medically diagnosed and hospitalization for coronary heart disease or stroke.

Results: The odds of hospitalization for heart disease or stroke was 37% (95% CI: 8%, 57%) lower among adults in neighborhoods with highly variable greenness (highest tertile) compared to those in predominantly green, or predominantly non-green neighborhoods (lowest tertile). This effect was independent of the absolute levels of neighborhood greenness. There was weaker evidence for associations with the mean level of neighborhood greenness.

Conclusion: Variability in neighborhood greenness is a single metric that encapsulates two potential promoters of physical activity - an aesthetically pleasing natural environment and access to urban destinations. Variability in greenness within a neighborhood was negatively associated with coronary heart disease and stroke.

Figure 1

Four simulated service areas. Area A has low mean but high variability (standard deviation) in greenness. Area B has both high mean and variability in greenness. Area C has both low mean and variability in greenness. Area D has high mean but low variability in greenness. In general, the prevalence of high NDVI pixels illustrates mean greenness, and the prevalence of contrasting high and low NDVI pixels illustrates variability in greenness.

Figure 2

Illustration of a service area with a high variability in greenness (NDVI Mean 0.04, SD 0.13). The lighter areas represent lakes, roads and buildings. Darker areas represent parks, ovals and bush land. The white section with east–west alignment at the north side of the service area and the white section with north–south alignment in the middle of the service area is commercial land. The remaining area is predominantly residential. A large section of bush land lies outside and south-east of the service area.

Figure 3

Illustration of a service area with low variability in greenness (NDVI Mean 0.10, SD 0.08). The lighter areas represent lakes, roads and buildings. Darker areas represent parks, ovals and bush land. The figure is more uniform in terms of NDVI. The Indian Ocean lies outside and to the west of the service area. This service area has a lower spatial variability (i.e. predominantly high levels and few low levels) of greenness than Figure 1.