Introduction Previous studies on the environment and obesity have estimated mutually adjusted effects of different environment factors that are strongly correlated with each other, without carefully assessing whether it is possible to disentangle these effects.

Methods We investigated whether correlated neighbourhood characteristics related to the sociodemographic, physical, service-related, and social-interactional environments measured within ego-centered areas were associated with BMI and waist circumference, and assessed whether or not these associations could be disentangled using an original neighbourhood characteristic-matching technique (analysis of each environmental effect within pairs of individuals similarly exposed to another environmental variable). We conducted cross-sectional analyses of 7230 adults from the RECORD Cohort Study (Paris region, France).

Results After adjustment for individual/neighbourhood socioeconomic variables, both outcomes were negatively associated with characteristics of the physical/service environments reflecting higher densities (eg, built surface area, street network connectivity, and densities of fruit/vegetables selling shops, fast-food restaurants, and healthcare resources). Multiple adjustment models were unable to disentangle the effects of these correlated densities. Analyses by pairs of participants similarly exposed to another environmental variable only identified a few associations, primarily with the density of fruit/vegetables selling shops.

Conclusion Overall, beyond influences of the socioeconomic environment, certain characteristics of the physical/service environments may be associated with weight status, but it may be difficult to disentangle the effects of various environmental dimensions because of the strong correlation between the variables (even if they imply different causal mechanisms and interventions).