Association between neighbourhood greenness and body mass index in a population-based study

Abstract

Worldwide overweight and obesity increased tremendously. Nowadays e.g., more than half of the German population is overweight and quarter of the adults are obese. This rapid increase warrants further investigations of factors involved in weight gain. Studies in developed countries identify the built environment impact on people’s weight status. Greenness is a well-known health resource and seen as a stimulator of physical activity. However, information is still lacking about the impact of neighbourhood greenness on bodyweight in urban areas.

Main objective is to study the relationship between neighbourhood greenness and Body Mass Index (BMI) of urban dwellers. We used the data of the cohort population-based Heinz Nixdorf Recall (HNR) study, which examined 4814 randomly selected citizens, aged 45-75 years, in three cities at the metropolitan Ruhr area since early 2000. Two follow-up examinations took place five and ten years after the baseline.

BMI and covariates were assessed by physical examinations and standardized interviews at the HNR study centre. Neighbourhood greenness was determined by using satellite-based Normalized Difference Vegetation Index (NDVI), calculated for different buffer sizes around participant’s residence address. We predefined confounder using the Directed Acyclic Graph. We conducted two separate cross-sectional analyses using (i) the HNR baseline data n=4780 (ii) the 10-years follow-up data n=3050. We applied multiple linear regression to estimate the effect sizes (ß) per 0.1 unite increase in NDVI and corresponding 95% confidence intervals (95%CI). In sensitivity analysis, we stratified by sex. Here we report results for the 1000 m buffer size.

Overall, baseline prevalence of obesity was 28% increasing to 31% in the 10-years follow-up. At baseline mean NDVI was 0.35 (SD=0.06) reduced to 0.31 (SD=0.05) after 10 years. Results of the adjusted multiple linear regression analyses at both baseline and 10-year follow-up indicate a distinct decrease in BMI with increasing NDVI (resp. ß=-0.35; 95%CI: -0.59, -0.12; ß=-0.39; 95%CI: -0.76, -0.01). However, sex stratified analyses revealed a stronger effect in women (-0.52; 95%CI: -0.90, -0.15) compared to men (-0.19; 95%CI: -0.48, 0.09). After 10 years, this association became even more pronounced among women (-0.82; 95%CI: -1.42, -0.22) while completely disappearing in men (0.03; 95%CI: -0.42, 0.47).

Our results suggest a substantial negative association between neighbourhood greenness within 1000 m and the bodyweight of middle-aged to older women. Interestingly, this effect was not observed for the men in our cohort. Forthcoming, we aim to analyse the mediation effects of physical activity and sports in longitudinal analyses to examine whether the mediators might elucidate our sex-specific outcomes. Overall, our findings support the role of green space in urban areas and its positive impact on public health, which is particularly important given the challenges of climate change.