Systemic inflammatory markers in relation to cognitive function and measures of brain atrophy: a Mendelian randomization study

Abstract

Observational studies have implied associations between multiple cytokines and cognitive decline, anti-inflammatory drugs however did not yield any protective effects on cognitive decline. We aimed to assess the associations of systemic inflammation, as measured by multiple cytokine and growth factor, with cognitive performance and brain atrophy using two-sample Mendelian randomization (MR). Independent genetic instruments (p < 5e − 8 and p < 5e − 6) for 41 systemic inflammatory markers were retrieved from a genome-wide association study conducted in 8293 Finnish participants. Summary statistics for gene-outcome associations were obtained for cognitive performance (N = 257,841) and for brain atrophy measures of cerebral cortical surface area and thickness (N = 51,665) and hippocampal volume (N = 33,536). To rule out the heterogeneity in the cognitive performance, we additionally included three domains: the fluid intelligence score (N = 108,818), prospective memory result (N = 111,099), and reaction time (N = 330,069). Main results were computed by inverse-variance weighting; sensitivity analyses taking pleiotropy and invalid instruments into account were performed by using weighted-median estimator, MR-Egger, and MR PRESSO. After correcting for multiple testing using false discovery rate, only genetically predicted (with p < 5e − 6 threshold) per-SD (standard deviation) higher IL-8 was associated with − 0.103 (− 0.155, − 0.051, padjusted = 0.004) mm3 smaller hippocampal volume and higher intelligence fluid score [β: 0.103 SD (95% CI: 0.042, 0.165), padjusted = 0.041]. Sensitivity analyses generally showed similar results, and no pleiotropic effect, heterogeneity, or possible reverse causation was detected. Our results suggested a possible causal association of high IL-8 levels with better cognitive performance but smaller hippocampal volume among the general healthy population, highlighting the complex role of inflammation in dementia-related phenotypes. Further research is needed to elucidate mechanisms underlying these associations.

Association between cytokine cycling levels and Sjogren's syndrome: genetic correlation and bidirectional Mendelian randomization study

“…Additionally, the data being derived from a single racial group limits the generalizability of our ndings to other populations. Consistent with other studies (51)(52)(53), the use of a P-value threshold of < 5 × 10^-6 due to a limited number of SNPs may have resulted in false positives. However, the robustness of our results is supported by high F-statistics.…”

Associations between abdominal adipose tissue, reproductive span, and brain characteristics in post-menopausal women

“…Also yet to be elucidated are the mechanisms underlying the associations between adipose tissue, oestrogen exposure, and brain characteristics, which are likely multifactorial and interactive. Factors such as elevated inflammatory markers have been associated with increased adipose tissue levels ( Aguilar-Valles et al, 2015 , Miller and Spencer, 2014 ), even specifically during the menopause transition ( Lee et al, 2009 ), as well as decreasing oestrogens ( McCarthy and Raval, 2020 ), brain atrophy ( Luo et al, 2022 ), and dementia risk ( Heneka et al, 2015 , Ransohoff, 2016 ). Biological markers of obesity, such as lipid profile ( Anstey et al, 2017 , Reitz, 2012 ), glucose ( Crane et al, 2013 ), HbA1c ( Ramirez et al, 2015 ), leptin ( Zeki Al Hazzouri et al, 2013 ), and Vitamin B12 ( Lauer et al, 2022 ), may also influence associations between adipose tissue and brain health, and contribute to risk of comorbidities such as type II diabetes and hypertension, which are known to impact neural and cardiometabolic health ( Cole, 2020 , Fuchs and Whelton, 2020 , Newby et al, 2022 , Peters et al, 2014 ).…”

Circulating Inflammatory Biomarkers and Risk of Intracranial Aneurysm: A Mendelian Randomization Study

“…However, a more stringent threshold (P < 5 × 10 − 8 ) results in less available instrumental variables, leading to less statistical power and insu cient sensitivity analyses. Still, multiple prior studies had employed the relaxed threshold to investigate the relationship between cytokines and diseases [51,52]. Second, although we performed pleiotropy test and utilized PhenoScanner, the potential bias from confounders could not be statistically ruled out.…”