Background: Observational studies have reported structural changes in the brains of patients with COVID-19; it remains unclear whether these associations are causal.
Aim: We evaluated the causal effects of COVID-19 susceptibility, hospitalization, and severity on cortical structures.
Design: Mendelian randomization (MR) study.
Methods: Data on the different COVID-19 phenotypes were obtained from the latest large-scale genome-wide association study (GWAS) (R7) of the COVID-19 Host Genetics Initiative. Brain structure data, including cortical thickness (TH) and surface area (SA), were obtained from the ENIGMA Consortium. Additionally, we employed the round 5 dataset released in January 2021 as the validation cohort. The inverse-variance weighted (IVW) method was used as the primary analysis in MR. Sensitivity analyses were conducted to evaluate heterogeneity and pleiotropy. We performed enrichment analysis on the MR analyses that passed the sensitivity analysis filtering.
Results: After IVW and sensitivity analyses, we observed causal associations between COVID-19 susceptibility and rostral middle frontal SAw (P = 0.0308, β=-39.1236), cuneus THw (P = 0.0170, β=-0.0121), medial orbitofrontal THw (P = 0.0002, β = 0.0225), postcentral THw (P = 0.0217, β= -0.0106), temporal pole THw (P = 0.0077, β = 0.0359), medial orbitofrontal SAnw (P = 0.0106, β= -24.0397), medial orbitofrontal THnw (P = 0.0007, β = 0.0232), paracentral SAnw (P = 0.0483, β= -20.1442), rostral middle frontal SAnw (P = 0.0368, β= -81.9719), and temporal pole THnw (P = 0.0429, β = 0.0353). COVID-19 hospitalization had causal effects on medial orbitofrontal THw (P = 0.0053, β = 0.0063), postcentral THw (P = 0.0143, β= -0.0042), entorhinal THnw (P = 0.0142, β = 0.0142), medial orbitofrontal THnw (P = 0.0147, β = 0.0065), and paracentral SAnw (P = 0.0119, β= -7.9970). COVID-19 severity had causal effects on rostral middle frontal SAw (P = 0.0122, β=-11.8296), medial orbitofrontal THw (P = 0.0155, β = 0.0038), superior parietal THw (P = 0.0291, β= -0.0021), lingual SAnw (P = 0.0202, β=-11.5270), medial orbitofrontal THnw (P = 0.0290, β = 0.0039), paracentral SAnw (P = 0.0180, β= -5.7744), and pars triangularis SAnw (P = 0.0151, β=-5.4520).
Conclusion: Our MR results demonstrate a causal relationship between different COVID-19 phenotypes and cortical structures.
© The Author(s) 2024. Published by Oxford University Press on behalf of the Association of Physicians. All rights reserved. For permissions, please email: journals.permissions@oup.com.