Observational epidemiological evidence supports a linear and independent association between serum gamma-glutamyltransferase (GGT) concentrations and the risk of Alzheimer's disease (AD). However, the causality of this association has not been previously investigated. We sought to assess the causal nature of this association using a Mendelian randomization (MR) approach. Using inverse-variance weighted MR analysis, we assessed the association between GGT and AD using summary statistics for single nucleotide polymorphism (SNP)-AD associations obtained from the International Genomics of Alzheimer's Project of 17,008 individuals with AD and 37,154 controls. We used 26 SNPs significantly associated with GGT in a previous genome-wide association study on liver enzymes as instruments. Sensitivity analyses to account for potential genetic pleiotropy included MR-Egger and weighted median MR.

The odds ratio of AD was 1.09 (95% confidence interval, 0.98 to 1.22; p  =  0.10) per one standard deviation genetically elevated GGT based on all 26 SNPs. The results were similar in both MR-Egger and weighted median MR methods. Overall, our findings cannot confirm a strong causal effect of GGT on AD risk. Further MR investigations using individual-level data are warranted to confirm or rule out causality.