THE ASSOCIATION BETWEEN BLOOD PRESSURE VARIABILITY AND DEMENTIA: A TWO-SAMPLE MENDELIAN RANDOMIZATION STUDY

Abstract

Objective: This study aims to investigate the causal inference by applying a mendelian randomization (MR) study using single-nucleotide polymorphisms (SNPs) from the genome-wide association study (GWAS) based on the UK biobank. Design and method: The genetic instruments were chosen from BPV GWAS based on UK Biobank data. Four GWA studies were used for the genetic evaluation of dementia, including (1) The International Genomic of Alzheimer's Project (IGAP) GWAS for dementia; (2) and (3) UK Biobank GWAS for proxy phenotypes of Alzheimer's dementia: the maternal and paternal family history (MFH and PFH-UKBB); (4) A combined GWAS meta-analysis from the abovementioned three GWAS results. The following exclusion criteria were applied for SNPs: (1) confounding SNPs by searching hypertension and artery stiffness GWAS; (2) significant SNPs that are associated with dementia in the datasets; (3) palindromic SNPs with an effect allele frequency between 0.4 and 0.7. Proxy SNPs were manually selected for some SNPs. (4) only SNPs with linkage disequilibrium r2 < 0.01 were selected as genetic instruments, also known as independent SNPs. The effect/reference alleles were checked to ensure the consistency of the base pair between BPV and dementia GWAS, which is also known as data harmonization. After data harmonization, four MR methods were employed to measure the causal effect: IVW-MR with random effect, IVW-MR with fixed effect, MR-Egger, and weighted median MR. Results: Finally, six independent SNPs were chosen as instrument variants for SBPV, and five for DBPV. Significant causal effects of SBPV on dementia were found in the PFH-UKBB dataset, and the odds ratio of dementia per 10-unit increase in SBPV was 1.028, 1.015, and 1.015 for MR-Egger, weighted median, and IVW-MR, respectively (Table 1). In contrast, only one significant result was found for DBPV in MFH-UKBB. No significant results were found on other datasets. Conclusions: Along with the evidence from different observational studies, this GWAS study demonstrates that systolic blood pressure variability is a potential causal risk factor for Alzheimer's dementia, while the evidence from diastolic blood pressure variability is still uncertain.