Longitudinal association of built environment pattern with DXA-derived body fat in elderly Hong Kong Chinese: a latent profile analysis.

Background: One major limitation of prior studies regarding the associations between built environment (BE) and obesity has been the use of anthropometric indices (e.g., body mass index [BMI]) for assessing obesity status, and there has been limited evidence of associations between BE and body fat. This study aimed to explore the longitudinal association between BE and body fat in a cohort of elderly Hong Kong Chinese and examine whether the BE-body fat associations differed by BMI categories.
Methods: Between 2001 and 2003, 3944 participants aged 65-98 years were recruited and followed for a mean of 6.4 years. BE characteristics were assessed via Geographic Information System. Body fat (%) at whole body and regional areas (trunk, limbs, android, and gynoid) were assessed by dual energy X-ray absorptiometry at baseline and three follow-ups. Latent profile analysis was used to derive BE class, and linear mixed-effects models were used to investigate the associations of BE class with changes in body fat. Stratified analyses by BMI categories were also conducted.
Results: Three BE classes were identified. Participants in Class 2 (characterized by greater open space and proportion of residential land use) had a slower increase in whole body fat (B = -0.403, 95% confidence interval [CI]: -0.780, -0.014) and limbs fat (-0.471, 95% CI: -0.870, -0.071) compared with participants in Class 1 (characterized by high proportion of commercial land use). There were significant interactions of BE class with BMI, and participants in Class 2 had a slower increase in whole body fat and regional fat compared with participants in Class 1 (B ranging from -0.987 [limbs] to -0.523 [gynoid]) among overweight and obese participants only.
Conclusions: We found that those who resided in the areas characterized by greater open space and proportion of residential land use had a slower body fat increase.
(© 2021. The Author(s), under exclusive licence to Springer Nature Limited.)