Systematic analysis of associations between obesity and memory decline

Despite the development of numerous obesity indicators, the optimal measure associated with memory remains to be elucidated. We examined and compared these associations whilst also exploring potential interactions with education. We analysed the associations between 20 obesity indicators and memory (measured by Delayed Word Recall Test (DWRT)) in middle-aged and older participants from Guangzhou Biobank Cohort Study by linear regression and stratification analyses. We used two-sample Mendelian randomisation (2SMR) to analyse the associations of obesity with cognitive performance. Of 20 obesity indicators, 6 and 7 were associated with poorer memory function measured by baseline and follow-up DWRT; 7 and 10 with faster memory decline measured by mean annual change (MAC) and MAC rate of DWRT. Central obesity measured by waist-to-hip-to-height ratio (WHHR) z-score showed the greatest effect size, with the βs (95% CIs) being - 0.09 (- 0.11 to - 0.07), - 0.07 (- 0.09 to - 0.04), - 0.02 (- 0.03 to - 0.01) and - 0.49 (- 0.69 to - 0.29), respectively. The associations were weaker in those with higher education (P for education-interaction from 0.008 to 0.049). In 2SMR, body mass index (BMI), waist circumference and BMI-adjusted waist-to-hip ratio were negatively associated with cognitive performance using inverse-variance weighted method, with βs (95% CIs) being - 0.11 (- 0.15 to - 0.07), - 0.07 (- 0.12 to - 0.02) and - 0.06 (- 0.09 to - 0.02), respectively. In conclusion, obesity, especially central obesity measured by WHHR, was associated with poorer memory function and faster memory decline in middle-aged and older people, with a weaker association observed amongst those with higher education.