Impaired skin microvascular endothelial reactivity in critically ill COVID-19 patients

The COVID-19 outbreak, caused by the Severe Acute Respiratory Syndrome coronavirus 2 (SARS-CoV-2), has affected unprecedently all regions of the world. About 5% of infected patients require intensive care unit (ICU) admission because of acute respiratory failure but also extra-respiratory disorders, including acute kidney injury or myocarditis [1, 2]. Frequent (arterial and) venous thrombosis affecting large vessels and microcirculation have been observed in COVID-19 patients suggesting that the SARS-CoV-2 virus may impact endothelial cell function and/or survival. Such a hypothesis was supported by 1/experimental studies showing that the SARS-CoV-2 bind to the angiotensin-converting enzyme 2 receptors, widely expressed on lung epithelial cells and vascular endothelial cells [3] 2/ histological analysis showing endothelium damage, named endothelitis, in several organs of COVID-19 non-survivors [4] 3/ sublingual videomicroscopy showing the accumulation of leukocytes and red blood cell microaggregates in the microcirculation of COVID-19 patients [5] despite normal red blood cells deformability [6]. However, the impact of SARS-CoV-2 infection on endothelial function in severely ill COVID-19 patients remained unknown.

This prospective pilot observational study compared skin microvascular endothelial reactivity in critically ill COVID-19 and patients admitted in the ICU for non-COVID-19 bacterial pneumonia (NCBP).

Our study prospectively assessed skin endothelial-dependent microvascular reactivity in SARS-CoV-2 pneumonia and non-COVID-19 bacterial pneumonia. We first observed that the skin microvascular basal blood flow was significantly lower in COVID-19 patients compared to NCBP. Moreover, the endothelium-dependent microvascular reactivity was significantly lower in critically ill COVID-19 patients than NCBP despite higher disease severity.

As both a target organ and an effector, the endothelium has become of paramount interest in COVID-19 patients [10]. Here we explored the skin endothelium-dependent vasodilatation function in the forearm using iontophoresis of acetylcholine. We found altered microvascular reactivity in COVID-19 patients. This result is in line with previous experimental studies. Using either local thermal hyperemia [11] or iontophoresis of acetylcholine [12], moderate-to-severe patients with COVID-19 had an impaired endothelial reactivity. In these two studies, the COVID-19 population was heterogeneous with mechanically ventilated and non-mechanically ventilated patients, and compared to healthy volunteers. Similarly, Mesquida et al. reported impairment in microvascular reactivity using near infrared spectroscopy in COVID-19 patients compared to healthy controls, which was associated with acute respiratory distress syndrome (ARDS) severity [13]. In this study, the control group is constituted of septic patients with bacterial pneumonia non-mechanically ventilated without vasopressor or other organ failure. While it is well known that septic patients have endothelial dysfunction [14], we documented that COVID-19 patients have a more pronounced impaired endothelial function than NCBP patients despite lower severity. One remarkable finding of our study is that, although in the NCBP group microvascular reactivity was heterogeneous, all COVID-19 patients had a consistent impaired skin endothelium-dependent vasodilatation. In addition, we assessed the endothelial reactivity within the first 24 h following ICU admission, only in patients who did not require invasive ventilation. Despite that, we observed a substantial endothelial-mediated vasoreactivity impairment, suggesting an endothelial involvement in moderate to severe SARS-CoV-2 ARDS without secondary bacterial infection and prior to a possible multiorgan failure.

Using Ach iontophoresis, we indirectly explored the Nitric Oxide (NO) pathway as acetylcholine promotes endothelial production of NO. NO plays a crucial role in vascular homeostasis with its vasodilatory and anti-thrombotic actions [15]. In severe SARS-CoV-2 infection, besides the inflammatory and pro-oxidative environment, NO bioavailability is impaired [16]. In addition, SARS-CoV-2 infects host cells binding and downregulating of ACE-2, also reducing NO production [17]. These data suggest that NO insufficiency plays a substantial role in severe COVID-19. Indeed the damaged endothelium and reduced NO bioavailability contribute to inflammatory response and coagulopathy, thus amplifying the microvascular impairment in a vicious circle [18, 19]. Microvascular vasodilatation is ultimately carried out by the vascular smooth muscle cells (SMC). In this study, performing only Ach iontophoresis, we speculate that we indirectly explore the endothelial production of NO, but an impaired response could alternatively involve SMC dysfunction. Indeed, infected SMC [3] might become non-responsive to the EC-produced mediators in COVID-19 patients such as NO but also endothelium-derived hyperpolarizing factor or eicosanoids [20].

In our study, the basal skin blood flow was significantly lower in COVID-19 patients than in NCBP. This is in line with previous studies of sublingual microcirculation in which a reduced microvascular density in severe COVID-19 patients [21], in association with biomarkers of coagulopathy [22]. However, we did not find any significant differences between the COVID-19 group and the bacterial pneumonia group in terms of tissular perfusion parameters (mottling score, skin temperature). Of note, peripheral tissue perfusion was not severely impaired in included patients, probably because we did not include septic shock patients with vasopressors. Another possible explanation for this discrepancy is that microvascular blood flow and tissue perfusion parameters were not assessed in the same area, reflecting the tissue perfusion heterogeneity.

Microvascular alterations are strongly associated with mortality and organ failure in sepsis and septic shock [14, 23]. In our COVID-19 cohort, we did not observe a significant correlation between endothelial reactivity and organ failure (i.e. the coagulopathy [Fibrinogen, D-dimers] nor the severity of respiratory failure [PaO2/FiO2]). These results may be explained by a lack of power in a small cohort of non-invasive mechanically ventilated patients. In previous studies, endothelial dysfunction has been associated with disease severity and outcome [11]. For example, Rovas et al. reported that sublingual glycocalyx thickness, very difficult to assess in vivo, could predict the 60-day mortality [21]. Although these results need to be confirmed in larger clinical studies, we believe that microvascular endothelial dysfunction contributes to organ failure and death in severe COVID-19 and could be a potential target in further studies.

We acknowledge some limitations to this pilot translational observational study. First, this is a single-center study with a limited number of patients. Therefore, we could not identify an association between impaired endothelial reactivity and prognosis. In the same line, we did not find any significant relationship between time from first symptoms to ICU admission and endothelial reactivity in COVID-19 patients. In this observational study, we used a standardized and validated technique to explore the endothelium-dependent vasodilation in response to acetylcholine [24] performed for years in our ICU [14, 25]. However, we cannot infer the underlying mechanisms nor other endothelial functions impairment. We can speculate that endothelial dysfunction pre-existed in patients with COVID-19. Indeed, hypertension, diabetes and cardiovascular disease are common conditions predisposing to severe COVID-19 [26] and are known to be associated with endothelial dysfunction [27, 28]. Nevertheless, comorbidities were not significantly different between groups. Corticosteroids were more frequently used in COVID-19 patients which may be a potential confounder. Moreover, we chose to focus only on patients with an isolated respiratory failure and no need for invasive mechanical ventilation at inclusion. In multiorgan failure patients having support organ therapy, many factors could potentially impair the microvascular reactivity (vasopressors, acidosis, neuromuscular blockade [29], renal replacement therapy [30, 31] and make it difficult to attribute causation of endothelial dysfunction to the type of infection. In addition, using NCBP as a control group for COVID-19 is an arguable choice but severity comparison remains challenging because classical disease severity scores used in ICU such as SOFA or SAPS2 may not be necessarily appropriate in patients with SARS-CoV-2 infection. Acknowledging that there is no “ideal” control group for such explorative study and despite some significant differences in treatment at the time of admission, we assume that NCBP with isolated respiratory failure is conceptually appropriate. Finally, we measured skin microvascular endothelial reactivity at one single early timepoint and changes during ICU stay and more specifically time to recovery were not evaluated.

In critically ill COVID-19 patients without invasive ventilation, we evidenced both a reduced baseline skin microvascular blood flow and a drastically impaired skin microvascular endothelium-dependent vasoreactivity compared to NCBP patients. These results support the hypothesis of a singular and clinically relevant SARS-CoV-2-associated endotheliopathy.