Viral infection and dementia risk: A systematic bidirectional Mendelian randomization analysis

Viral infections have attracted increasing interest as potentially alterable risk factors for dementia; nonetheless, the results from observational research are variable and prone to confounding bias and reverse causation. This work seeks to systematically explore the potential causal associations between several viral infections and various forms of dementia by a Mendelian randomization methodology utilizing summary statistics and extensive genomic data from the Finnish FinnGen R12 database.

This study initially included 14 viral infections, including anogenital herpes, infectious mononucleosis, acute hepatitis B, herpesviral infections, acute poliomyelitis, rubella, varicella, viral warts, herpes zoster, COVID-19, measles, mumps, acute hepatitis A, and HIV infection. The outcome measurements included 15 forms of dementia, including Alzheimer's dementia, vascular dementia and its subtypes, and frontotemporal dementia. The principal analysis utilized the inverse variance weighted (IVW) methodology. This report provides the number of SNPs in each analysis, the mean F statistic, the causal effect estimates, and their 95% confidence ranges. All the MR-Egger regression, weighted median, simple mode, and weighted mode approaches were used in the sensitivity analyses.

The final analysis omitted acute hepatitis A, HIV infection, and measles because of inadequate instrumental variables (only 1-2 SNPs). The IVW primary analysis revealed 6 associations with uncorrected P < 0.05 among the remaining 11 viral infections. Rubella infection was associated with a decreased risk of vascular dementia (subcortical type) (OR = 0.8304; 95% CI: 0.7154-0.9638; P = 0.0145; SNPs = 7; mean F statistic = 25.56) and showed a negative association with the outcome of the composite dementia, including that of Alzheimer's disease (OR = 0.9553; 95% CI: 0.9128-0.9998; P = 0.0488). Viral wart infection was associated with a reduced risk of Alzheimer's dementia (OR = 0.9095; 95% CI: 0.8332-0.9928; P = 0.0338; SNPs = 11; mean F-statistic = 56.01), any dementia (strict exclusion criteria) (OR = 0.9278; 95% CI: 0.8668-0.9932; P = 0.0310), and Parkinson's-related dementia (OR = 0.8182; 95% CI: 0.6796-0.9852; P = 0.0342). There was a suggestive link between mumps infection and a higher risk of vascular dementia (mixed type) (OR = 1.5579; 95% CI: 1.2219-1.9864; P = 0.0003; SNPs = 5; mean F-statistic = 22.63). This link was equally strong in the weighted median analysis (OR = 1.4131, P = 0.0277). In fact, statistically significant connections were not found between any types of viral infections and dementias (after multiple testing correction). Furthermore, for the other viral infections, no associations were observed even at the uncorrected P < 0.05 level. All findings with uncorrected P < 0.05 were successfully tested for heterogeneity and pleiotropy, and sensitivity analyses revealed consistent effect directions.

This study offers suggestive genetic evidence for potential associations between certain viral infections and specific forms of dementia. Infections with rubella and viral warts may correlate with a diminished risk of dementia; however, mumps infection may increase the risk of vascular dementia (mixed type). It was not possible to reliably evaluate acute hepatitis A and HIV infection since there were not enough instrumental variables. These results provide preliminary insights into the influence of infectious agents on neurodegenerative disorders, serving as a hypothesis-generating framework for future epidemiological and mechanistic research.