Frontiers in Nutrition · Nov 18, 2026
Community modern-market change and subsequent abdominal adiposity in Chinese adults: a longitudinal analysis of the China Health and Nutrition Survey
By Lin Guo, Yanjie Jiang, Jianing Yue et al.
Background: Community food environments are changing rapidly in China, yet evidence remains limited on whether change in a heterogeneous modern-market environment precedes later abdominal adiposity and whether any association is explained by macronutrient composition. Methods: We linked the 1997 and 2004 community modern-markets score in the China Health and Nutrition Survey to waist circumference measured in 2011 and 2015. The exposure-eligible clinical cohort included 9,604 adults aged 20–80 years; fully adjusted models included 3,649 participants in 2011 and 2,949 in 2015 and used community-clustered standard errors. A separate 1997–2006 dietary panel was analyzed using strict individual and survey-wave fixed effects after outcome-specific complete-case selection. Additional analyses evaluated score-change categories, score comparability, attrition weighting, and missing-outcome sensitivity. Results: Each two-point greater 1997–2004 increase in the modern-markets score was associated with a positive but statistically imprecise baseline-waist-adjusted difference in subsequent waist circumference in 2011 (0.449 cm; 95% CI, −0.068 to 0.966; P=0.089) and a positive estimate whose confidence interval excluded zero in 2015 (0.555 cm; 0.106 to 1.004; P=0.015). Holm-adjusted P values were 0.089 and 0.031, respectively. In corrected two-way fixed-effects models, the corresponding per-point differences in energy share were 0.057 percentage points for fat (−0.104 to 0.217), 0.048 for protein (0.007 to 0.088; nominal P=0.022; BH q=0.066), and −0.121 for carbohydrate (−0.283 to 0.041). No macronutrient estimate survived Benjamini–Hochberg correction, and no secondary biomarker survived false-discovery-rate correction. Conclusions: Greater community modern-market score change was associated with a positive but imprecise baseline-waist-adjusted difference in subsequent waist circumference in 2011 and a more precise positive difference in 2015. The separate dietary panel did not identify precise fat or carbohydrate energy-share shifts and cannot confirm or exclude dietary mediation. Changes in score content around 2004, incomplete follow-up, and residual confounding limit causal interpretation.