The Epidemiological and Mechanistic Understanding of the Neurological Manifestations of COVID-19: A Comprehensive Meta-Analysis and a Network Medicine Observation

Abstract

The clinical characteristics and biological effects on the nervous system of infection with the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) remain poorly understood. The aim of this study is to advance epidemiological and mechanistic understanding of the neurological manifestations of coronavirus disease 2019 (COVID-19) using stroke as a case study. In this study, we performed a meta-analysis of clinical studies reporting stroke history, intensive inflammatory response, and procoagulant state C-reactive protein (CRP), Procalcitonin (PCT), and coagulation indicator (D-dimer) in patients with COVID-19. Via network-based analysis of SARS-CoV-2 host genes and stroke-associated genes in the human protein-protein interactome, we inspected the underlying inflammatory mechanisms between COVID-19 and stroke. Finally, we further verified the network-based findings using three RNA-sequencing datasets generated from SARS-CoV-2 infected populations. We found that the overall pooled prevalence of stroke history was 2.98% (95% CI, 1.89-4.68; I ²=69.2%) in the COVID-19 population. Notably, the severe group had a higher prevalence of stroke (6.06%; 95% CI 3.80-9.52; I ² = 42.6%) compare to the non-severe group (1.1%, 95% CI 0.72-1.71; I ² = 0.0%). There were increased levels of CRP, PCT, and D-dimer in severe illness, and the pooled mean difference was 40.7 mg/L (95% CI, 24.3-57.1), 0.07 μg/L (95% CI, 0.04-0.10) and 0.63 mg/L (95% CI, 0.28-0.97), respectively. Vascular cell adhesion molecule 1 (VCAM-1), one of the leukocyte adhesion molecules, is suspected to play a vital role of SARS-CoV-2 mediated inflammatory responses. RNA-sequencing data analyses of the SARS-CoV-2 infected patients further revealed the relative importance of inflammatory responses in COVID-19-associated neurological manifestations. In summary, we identified an elevated vulnerability of those with a history of stroke to severe COVID-19 underlying inflammatory responses (i.e., VCAM-1) and procoagulant pathways, suggesting monotonic relationships, thus implicating causality.

Keywords: SARS-CoV-2; cerebrovascular disease; coronavirus disease 2019 (COVID-19); network medicine; protein-protein interactome; stroke; vascular cell adhesion molecule 1 (VCAM-1).

FIGURE 1

Meta-analysis revealing association of stroke with disease severity of COVID-19. Random intercept logistic regression model was used to estimate pooled prevalence and I² was used to show heterogeneity among studies. The red bar denotes the pooled prevalence using random effect model.

FIGURE 2

The elevated inflammatory factors and coagulation indicator are associated with disease severity of COVID-19. The meta-analysis (inverse variance method) was performed to estimate the MD in three groups. I² was used to show heterogeneity among studies. The red bar denoted the pooled MD using random effect model; MD, mean difference.

FIGURE 3

Network-inferred inflammatory endophenotypes shared by stroke and COVID-19. Virus target genes are shown in Blue. The stroke-associated genes were shown in Green. Stroke-associated genes which were also the direct targets of the virus, shown in Orange. The links between genes/proteins denote the physical protein-protein interactions (including SARS-CoV-2 viral protein and human protein interactions, and human protein-protein interactions).

FIGURE 4

The role of VCAM1 in SARS-CoV-2 induced vasculitis and clotting cascade. VCAM1, vascular cell adhesion molecule 1; ACE2, angiotensin-converting enzyme 2; Ang II, angiotensin II; Ang (1–7), angiotensin (1–7); ERK1/2, extracellular signal-regulated kinase 1/2; IL-6, interleukin-6; IL-1β, interleukin-1β; IFN-γ, interferon-γ; MCP-1, monocyte chemoattractant protein-1; MIP, macrophage inflammatory protein.

FIGURE 5

Selected genes involved in COVID-19-associated stroke were differentially expressed between SARS-CoV-2 positive patients and controls. (A) Four SARS-CoV-2 entry host factors were significantly differentially expressed in nasal tissues from COVID-19 positive patients compared to controls. (B) Stroke-related genes were differentially expressed in peripheral blood mononuclear cell (PBMC) samples from COVID-19 positive patients and controls. The data are represented as a boxplot where the middle line is the median, the lower and upper edges of the box are the first and third quartiles, the whiskers represent the interquartile range (IQR) × 1.5. (C) The bar plot shows the stroke-related genes were upregulated in iPSC–cardiomyocytes after SARS-CoV-2 infection. Adjusted p-value (q) were computed by Benjamini-Hochberg method. The cutoff of differentially expressed genes were | log₂-fold change| > 0.5 and adjusted p-value (q) < 0.05 (Supplementary Tables S5–S8).