Persistent endotheliopathy in the pathogenesis of long COVID syndrome

Abstract

Background: Persistent symptoms including breathlessness, fatigue, and decreased exercise tolerance have been reported in patients after acute SARS-CoV-2 infection. The biological mechanisms underlying this "long COVID" syndrome remain unknown. However, autopsy studies have highlighted the key roles played by pulmonary endotheliopathy and microvascular immunothrombosis in acute COVID-19.

Objectives: To assess whether endothelial cell activation may be sustained in convalescent COVID-19 patients and contribute to long COVID pathogenesis.

Patients and methods: Fifty patients were reviewed at a median of 68 days following SARS-CoV-2 infection. In addition to clinical workup, acute phase markers, endothelial cell (EC) activation and NETosis parameters and thrombin generation were assessed.

Results: Thrombin generation assays revealed significantly shorter lag times (p < .0001, 95% CI -2.57 to -1.02 min), increased endogenous thrombin potential (p = .04, 95% CI 15-416 nM/min), and peak thrombin (p < .0001, 95% CI 39-93 nM) in convalescent COVID-19 patients. These prothrombotic changes were independent of ongoing acute phase response or active NETosis. Importantly, EC biomarkers including von Willebrand factor antigen (VWF:Ag), VWF propeptide (VWFpp), and factor VIII were significantly elevated in convalescent COVID-19 compared with controls (p = .004, 95% CI 0.09-0.57 IU/ml; p = .009, 95% CI 0.06-0.5 IU/ml; p = .04, 95% CI 0.03-0.44 IU/ml, respectively). In addition, plasma soluble thrombomodulin levels were significantly elevated in convalescent COVID-19 (p = .02, 95% CI 0.01-2.7 ng/ml). Sustained endotheliopathy was more frequent in older, comorbid patients, and those requiring hospitalization. Finally, both plasma VWF:Ag and VWFpp levels correlated inversely with 6-min walk tests.

Conclusions: Collectively, our findings demonstrate that sustained endotheliopathy is common in convalescent COVID-19 and raise the intriguing possibility that this may contribute to long COVID pathogenesis.

Keywords: convalescent COVID-19; endothelial cell (EC) activation; long COVID.

FIGURE 1

Representative thrombin generation curves (A); each curve shows the mean of duplicate reactions from one individual's plasma. The quantitative parameters of (B) lag time, (C) endogenous thrombin potential (ETP), (D) peak thrombin, and (E) time to peak were derived from the thrombin generation curves, comparing convalescent COVID‐19 (n = 50) and healthy controls (n = 17). Comparison of EC activation parameters between convalescent COVID‐19 (n = 50) and healthy controls (n = 17) are shown, including (F) FVIII:C and (H) VWF:Ag. Data are presented as median and the interquartile range. Comparisons between groups were assessed by the Mann‐Whitney U test. Dotted lines represent the upper and lower limits of the local normal range with results in the green shaded areas falling within the normal reference range. Correlations are shown between plasma levels of FVIII:C and (G) Peak thrombin and (I) VWF:Ag. Correlations were evaluated using the Spearman rank correlation test. (*p < .05, **p < .01, ****p < .0001). EC, endothelial cell; FVIII, factor FVIII; VWF:Ag, von Willebrand factor antigen

FIGURE 2

Comparison of EC activation parameters between convalescent COVID‐19 (n = 50) and healthy controls (n = 17) including: (A) VWFpp and (C) sTM. Data are presented as median and the interquartile range. Comparisons between groups were assessed by the Mann‐Whitney U test. Dotted lines represent the upper and lower limits of the local normal range with results in the green shaded areas falling within the normal reference range. Correlations are shown between plasma levels of (B) VWF:Ag vs. VWFpp and (D) VWF:Ag vs. sTM. (E) Heatmap visualization indicating EC marker levels detected in each subject (columns) for each protein (rows). (F) Convalescent VWF:Ag results are grouped according to whether acute infection was managed as an outpatient or inpatient; patients were aged ≥50 or <50 years and whether comorbidity counts were ≥2 or <2, respectively. Correlations are shown between 6‐min walk test distance and plasma levels of (G) VWF:Ag and (H) VWFpp, respectively. Correlations were evaluated using the Spearman rank correlation test. (ns, not significant, *p < .05, **p < .01, ****p < .0001). EC, endothelial cell; VWF:Ag, von Willebrand factor antigen; VWFpp, von Willebrand factor propeptide