Abstract: BACKGROUND: Alzheimer’s disease has been associated with sarcopenia, but a causal relationship has not been established. Exploring the causal relationship between the two most common disability-burdening diseases in the aging population - Alzheimer’s disease and sarcopenia - and their potential mediating factors holds certain implications for further alleviating the healthcare costs and socioeconomic burden for older adults in China.
OBJECTIVE: To explore the potential causal relationship between Alzheimer’s disease and sarcopenia in the general population using a Mendelian randomization study and to explore the role of body mass index in this context.
METHODS: Two-sample Mendelian randomization analysis based on published genome-wide association studies (GWAS) were used to infer causality, and univariate Mendelian randomization and mediation analyses were used in the study design. Through the Integrative Epidemiology Unit (IEU) database, ieu-b-2 was selected as the Alzheimer’s disease dataset (sample size: 63 926), ieu-b-4816 as the body mass index dataset (99 998), ebi-a-GCST90000027 as the appendicular lean mass dataset (244 730), ukb-b-7478 as the left hand grip strength dataset (461 026), ukb-b-10215 as the right hand grip strength dataset (461 089) and ukb-b-4711 as the walking pace dataset (459 915). Inverse-variance weighting was used as the primary analysis method, and the results were validated by pleiotropy and heterogeneity analysis. The Steiger Directionality Test was performed to validate the reasonableness of the causal direction. 
RESULTS AND CONCLUSION: (1) The Mendelian randomization analyses provided evidence that Alzheimer’s disease predicted the risk of appendicular lean mass [odds ratio (OR)=1.009; 95% confidence interval (CI), 1.001-1.017; P=0.023), and walking pace (OR=1.010; 95% CI, 1.003-1.017; P=0.008). No correlation with hand grip strength was observed. (2) Alzheimer’s disease was negatively correlated with body mass index (OR=0.893; 95% CI, 0.811-0.984; P=0.022); body mass index was positively correlated with appendicular lean mass (OR=1.084; 95% CI, 1.031-1.141; P=0.002) and negatively correlated with walking pace (OR=0.975; 95% CI, 0.969-0.980; P < 0.001). (3) Mediation analyses showed that the causal relationship between Alzheimer’s disease and appendicular lean mass and walking pace was partially mediated by body mass index, with the proportion of mediations being 50.25% and 32.11%, respectively. (4) The results of this study suggest that based on large-scale population studies, genetic prediction of Alzheimer’s disease is a potential risk factor for sarcopenia, in which body mass index plays an important mediating role. This suggests that in clinical practice, attention should be paid to the muscle condition of patients with Alzheimer’s disease, and weight management should be implemented, as maintaining a body mass index within the normal high range may have a preventive effect on the occurrence of sarcopenia in patients with Alzheimer’s disease. However, further research is needed to verify the applicability of this conclusion to other ethnic groups. This study utilized an international public database for analysis, providing a reference for research on the correlation between Alzheimer’s disease and sarcopenia in the Chinese population. It also highlights the significant mediating role of body mass index, offering insights for further prevention and treatment of sarcopenia among Chinese individuals.

Key words: Alzheimer’s disease, body mass index, sarcopenia, two-step Mendelian randomization, mediation analysis, causal association