Abstract
Identifying modifiable risk factors for Alzheimer’s disease (AD) is a global health priority, yet observational studies are often limited by confounding and reverse causation. This study utilized a two-sample Mendelian randomization (MR) approach to systematically screen 1,054 risk factors and 28 drug classes for causal associations with late-onset AD. Our results provide robust evidence that higher educational attainment, higher intelligence, and increased trunk fat-free mass are protective against AD. Conversely, a genetic predisposition to glioma was associated with an increased risk of AD. Analysis of 28 drug treatments, including antihypertensives, yielded limited evidence of causal associations, suggesting that lifelong genetic exposure to these pathways may not translate directly to late-life pharmacological intervention.
Introduction
Alzheimer’s Disease (AD) remains one of the most significant challenges in modern medicine, characterized by a complex etiology and a lack of curative treatments. While hundreds of observational studies have proposed various risk factors—ranging from dietary habits to cardiovascular health—these findings are frequently confounded by lifestyle variables or reverse causation, where the early stages of neurodegeneration influence the exposure rather than vice versa. Furthermore, the "cherry-picking" of specific traits for study can lead to publication bias.
To address these limitations, this research employs a systematic, hypothesis-free screening of over 1,000 potential risk factors and medical conditions. By using genetic variants as proxies for these traits, we can simulate a randomized controlled trial environment to determine which factors truly drive AD risk.
Methodology
This study utilized Two-sample Mendelian Randomization (MR). MR relies on the principle that genetic variants (Single Nucleotide Polymorphisms, or SNPs) are randomly assigned at conception. Because these variants are generally fixed throughout life and are not influenced by the progression of AD, they can serve as "instrumental variables" to estimate the causal effect of an exposure on an outcome.
Data Sources
- Exposures: Genetic instruments for 1,054 traits and 28 drug classes were sourced from the IEU GWAS database. Each trait required at least three independent, genome-wide significant markers to be included.
- Outcome: Summary statistics for late-onset AD were obtained from the International Genomics of Alzheimer’s Project (IGAP), comprising 17,008 cases and 37,154 controls of European ancestry.
Statistical Analysis
The primary analysis used the Inverse Variance Weighted (IVW) method. To ensure the validity of the findings and account for potential horizontal pleiotropy (where a gene affects the outcome through pathways unrelated to the exposure), several sensitivity analyses were performed, including MR-Egger, Weighted Median MR, and MR-PRESSO. Multiple comparisons were corrected using a False Discovery Rate (FDR) threshold of < 0.05.
Key Results and Discussion
Cognitive Reserve and Education
The most significant protective factors identified were related to cognitive reserve. Higher educational attainment (specifically obtaining a college or university degree) was associated with a 59% reduction in AD risk (OR 0.41). General intelligence also showed a strong protective effect (OR 0.72). Conversely, individuals with no formal qualifications faced a significantly higher risk (OR 3.68). These findings reinforce the hypothesis that early-life brain development and cognitive stimulation provide a buffer against neurodegenerative decline.
Body Composition: Muscle vs. Fat
While obesity is frequently discussed as an AD risk factor, this systematic screen provided a more nuanced view. Trunk fat-free mass (representing lean muscle and bone mass in the torso) was found to be significantly protective (OR 0.78 per standard deviation increase). This suggests that maintaining muscle mass, rather than simply focusing on total body weight or BMI, may be a critical modifiable target for preserving brain health in aging populations.
The Glioma Link
A surprising finding was the causal association between a genetic predisposition to glioma and an increased risk of AD (OR 1.13). While some observational studies previously suggested an inverse relationship between cancer and AD, this MR analysis suggests that shared biological pathways—perhaps involving the neural environment or cell growth regulation—may link these two conditions in a way that warrants further molecular investigation.
Drug Repurposing
The study investigated 28 drug classes, including antihypertensives and anti-inflammatory agents like aspirin, which have been hypothesized to reduce AD risk. However, the MR analysis found limited evidence to support their efficacy. This discrepancy may stem from the difference between lifelong genetic predisposition and the acute, late-life administration of pharmacological treatments.
Limitations and Conclusion
While this study provides a high-level systematic overview of AD risk, several limitations exist. The data is primarily derived from populations of European ancestry, which may limit the generalizability of the findings to other ethnic groups. Additionally, MR estimates the effects of lifelong exposure; therefore, the null results in the drug repurposing analysis do not necessarily rule out the benefits of short-term clinical interventions.
In conclusion, this systematic MR analysis confirms the vital role of education and intelligence in AD prevention and identifies trunk fat-free mass and glioma risk as significant factors. These results help prioritize future research efforts toward muscle maintenance and the shared pathways between oncology and neurodegeneration.
Summary updated