Association of green space with bone mineral density change and incident fracture in elderly Hong Kong Chinese: Mr. OS and Ms. OS study

Abstract

Background: A large body of literature has reported positive effects of green space (GS) on various aspects of health and well-being, while no studies explore the role of GS in bone health.

Objectives: The present study aimed to investigate the associations of GS with bone mineral density (BMD) change and incident fracture in a prospective cohort of elderly Hong Kong Chinese.

Methods: Between 2001 and 2003, 3944 participants aged 65 years and older at baseline were recruited. GS (%) within 300-m and 500-m buffers were calculated for each participant based on the Normalized Difference Vegetation Index. BMD at whole body, lumbar spine, total hip, and femoral neck were assessed by dual energy X-ray absorptiometry at baseline and 3 follow-ups. Incident fracture cases were ascertained from the electronic database of Hospital Authority of Hong Kong. Linear mixed-effects models and Cox proportional hazards models were used to investigate the associations of GS with changes in BMD and incident fracture, respectively.

Results: Greater GS within 300-m and 500-m buffers were associated with a slower increase in lumbar spine BMD over 14 years. After adjustment for potential confounders, β and 95% confidence intervals (CIs) of change in BMD across Q2-Q4 (quartiles of GS measured in a 300-m, compared with Q1) were -6.42 (-12.3, -0.59), -7.78 (-13.6, -1.97), and -7.83 (-13.7, -2.00) mg/cm³, respectively. GS was also positively associated with non-spinal fracture and major osteoporotic fracture incidence risks. Multivariable-adjusted hazard ratios (95%CIs) were 1.40 (1.09, 1.79; P-trend = 0.036) for non-spinal fracture and 1.53 (1.13, 2.07; P-trend = 0.010) for major osteoporotic fracture (Q4 compared with Q1 of GS measured in a 300-m buffer). Positive GS-fracture associations were also found for GS within a 500-m buffer.

Conclusions: We found that those who lived near higher GS levels had a slower increase in lumbar spine BMD and had higher incident fracture risk.

Keywords: Bone mineral density; Cohort study; Dual energy X-ray absorptiometry; Fracture; Green space.

Figure 1.. Flowchart of study participants over study period.

* Number of participants whose bone mineral density was assessed by dual energy X-ray absorptiometry (DXA).

Figure 2.. Trajectories of bone mineral density over age by Quartiles (Q) of green space (300-m buffer).

Trajectories derived from linear mixed-effects models, including age, green space (quartiles), and their interaction term in the models, and adjusted for sex, marital status, education level, socioeconomic status, alcohol drinking, smoking, body mass index, physical activity, calcium supplement, number of chronic diseases, any incident fracture, and baseline bone mineral density (whole body, lumbar spine, total hip, and femoral neck, respectively). In comparison with Q1, Q4 had a slower decrease in whole body bone mineral density (Panel a), while a slower increase in lumbar spine bone mineral density (Panel b).

Figure 3.. Shapes of the associations of green space with incident non-spinal fracture.

Restricted cubic spline models were used to estimate the associations of green space with incident non-spinal fracture in the Cox proportional hazards models, adjusted for baseline age, sex, marital status, education level, socioeconomic status, alcohol drinking, smoking, body mass index, physical activity, calcium supplement, number of chronic diseases, previous history of fracture, family history of fracture, and whole body bone mineral density. None of the associations showed significant nonlinearity (P-nonlinearity >0.05). In addition, to make the dose-response association comparable with the quartile (Q) results, we also plotted the hazard ratio (HR) and 95% confidence interval (95% CI) of incident fracture across Q2 to Q4 of green space (compared with Q1, location of each plot corresponds to the median levels within each quartile).