Coronavirus disease 2019 (COVID-19) is responsible for more than 4,550,00 deaths worldwide at the beginning of September 2021 according to the World Health Organization (WHO). The vast majority of the infected people have subclinical to moderate forms, but some of them develop respiratory failure with acute respiratory distress syndrome (ARDS) (1). Postmortem histological analysis from COVID-19-related ARDS non-survivor patients exhibited vascular microthrombi in lung capillaries that participate in lung damage (2). Moreover, about 20% of critically ill COVID-19 patients develop pulmonary embolism (PE), which can aggravate their pulmonary condition and impair oxygenation because of shunt effect (3). Immunothrombosis is a physiological innate immune response that leads to formation of thrombi inside blood vessels in order to contain and destroy pathogens such as bacteria, fungi, and viruses (4). It involves neutrophils, monocytes, platelets, and activation of hemostasis. Activation of neutrophils by pathogens causes the emission of neutrophil extracellular traps (NETs) that are DNA fragments decorated with proteins of neutrophil origin such as myeloperoxidase (MPO) (5). When uncontrolled, immunothrombosis becomes detrimental to the host. As NETs are procoagulant (6, 7) and cytotoxic for lung vascular endothelial cells (8, 9), increased NETosis has been found to participate in various pathological processes such as arterial and venous thrombosis (6), ARDS (10), and other critical conditions not linked to COVID-19 (11).

It is now well-admitted that circulating markers of NET formation are associated with COVID-19 severity (12–14). Whether measurement of NET biomarkers early after admission for COVID-19 can be of prognostic value is a major question, as it would strengthen the rationale to target NETs and may help in clinical decision-making. The aim of this study is thus to evaluate whether plasma levels of NET markers measured shorty after admission of hospitalized COVID-19 patients are associated with clinical outcomes in terms of clinical worsening, survival, and PE occurrence.

The prediction of disease evolution is a challenge among COVID-19 patients and especially among the most severe ones, i.e., with COVID-19-related ARDS. Our study was designed to assess the involvement of circulating markers of NETs for COVID-19 evolution among inpatients who were admitted at the hospital for moderate and severe COVID-19. We thus included 46 COVID-19 patients with moderate disease and 50 with ARDS. We studied the association between 3 circulating markers of NETs measured shortly after hospital admission and disease evolution in terms of survival, aggravation (for patients with moderate disease only), and PE occurrence. Our findings confirm previous reports showing that circulating markers of NETs correlate with the clinical severity at the time of blood sampling (12–14), but we also report an association between circulating markers of NETs and later survival within the subgroup of patients admitted to the hospital for severe COVID-19 (ARDS patients). The study of Ng et al. (14) previously reported an association between circulating markers of NETs and clinical outcome in a cohort of 106 patients with moderate to severe COVID-19 patients, but this study did not analyze the prognostic value of circulating markers of NETs specifically in each group of patients. Here we did not report any significant association between circulating markers of NETs and disease evolution in the moderate COVID-19 patients in terms of clinical aggravation (i.e., transfer to ICU, PE occurrence, or death). On the contrary, we did find an association between plasma levels of NET markers and survival in ARDS patients.

These results reinforce the hypothesis that immunothrombosis and in particular NETosis are involved in the complications of COVID-19, especially in the most severe forms (13, 25–27). Only experiments in animal models could definitely prove the pathogenic role of NETosis in COVID-19 progression but, at present, several lines of evidence show a link between severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and NET formation. First of all, several studies reported induction of NET release during COVID-19 either by the virus itself (28), plasma and serum from severe COVID-19 patients (probably through the hyperinflammation typical of severe forms of COVID-19), activated platelets from severe patients (12, 13, 29), and anti-phospholipid antibodies (30). Second, a recent study compared lung specimens from four patients who died from COVID-19 and four from a COVID-19-unrelated cause. The authors reported that NETs infiltrated the lung airways and interstitial and vascular compartments only in severe COVID-19 patients but not in controls, supporting the hypothesis that NETs may drive severe pulmonary complications of COVID-19 (31). The third range of evidence comes from the known role of NETs in thrombosis (32) and the observation that severe SARS-CoV-2 infection induces a prothrombotic state manifesting especially with microthrombosis (33). In line with that, a consortium of authors recently proposed that exaggerated immunothrombosis, occurring for the most part within lung microvessels, drives the clinical manifestations of COVID-19 (32), with the atypical ARDS of COVID-19 being summarized as “microvascular COVID-19 lung vessels obstructive thromboinflammatory syndrome” (MicroCLOTS) (34).

It may look contradictory that we observe an association between circulating NET markers and survival in patients who arrive at the hospital for an already severe form of COVID-19 but that we do not find any association with transfer to ICU in the moderate ones. Part of the explanation could be that host response is still regulated in moderate patients but not in ARDS patients, causing the accumulation of NETs. Another plausible explanation is that NETosis generation could occur mostly within the lung tissue (31, 35) and not the circulation. Lung biopsies are not available in these patients for ethical reasons, as their complications are frequent and possibly lethal.

From a therapeutic point of view, as our results show that NETosis is already highly activated in severe patients who will die, we wonder whether targeting NETosis in severe patients is not already too late. In our opinion, our results suggest that NETosis should be targeted to prevent COVID-19 aggravation, before ARDS occurrence. There are ongoing clinical trials that aim to either prevent NET formation or degrade already-formed NETs. The first ones use anti-inflammatory drugs such as the Janus Kinases 1/2 (JAK1/2) inhibitor ruxolitinib (NCT04338958), dipyridamole (NCT 04391179), and ticagrelor (NCT02735707, NCT04518735). We suspect that these drugs should be more efficient in the less severe patients. DNase I (dornase alpha) can degrade already formed NETs and is currently tested by inhalation in patients with COVID-19 (NCT04402944, NCT04355364, NCT04432987, NCT04359654, NCT04445285, NCT04402970). A major issue with the nebulizing administration route in severe COVID-19 is that there is concern that it will not reach the perialveolar vessels due to the high amounts of platelet factor 4 (PF4) (13) that compacts NETs and decrease their susceptibility to DNase degradation (36).

Thrombosis, and especially PE, is a frequent feature in COVID-19 patients and is an independent risk factor for death (37). It should be noted that the embolic origin of the pulmonary vessel occlusions is questionable, and it may be that the so-called PEs are rather pulmonary thrombi that occur directly in pulmonary arteries (38, 39). Given the known role of NETs in venous thrombosis, we looked for differences in circulating markers of NETs. We found that 2 markers were higher in hospitalized COVID-19 patients who subsequently developed PE than those who did not. But when we restricted our analysis within the specific subgroup of COVID-19-related ARDS patients, who are the ones at more risk to develop PE, we did not find any difference between patients who will develop PE or not. This suggests, if confirmed in a larger cohort, that NETosis in itself is not a major driver for PE (venous thromboembolic event in the macrocirculation). Whereas there is an abundant literature to search for biological markers of clinical aggravation, there are only very few studies that report an association between a biological marker or a clinical parameter that is associated with occurrence of thrombosis among severe COVID-19 patients. D-dimers, which are reported to have a significant predictive value for mortality both in non-critical and critical COVID-19 patients (40, 41), have a limited predictive value for venous thromboembolism (VTE) occurrence, with an area under the curve (AUC) of 0.565 (42).

Our study has several limitations. First, we observe, as others in COVID-19 patients (12), dichotomies between the three circulating NET markers. This could be due to the relatively small number of samples we analyzed and also to the lack of standardization for NET marker measurement. We used 3 different plasma markers to measure NETs. Indeed, despite the discovery of the process of NETosis in 2004 (5), there is still no reference test for NET quantification (6). NETs can be visualized and quantified with conventional fluorescence microscopy, but this assay is hardly reproducible and time-consuming and is better when performed right after blood sampling. To overcome this issue and quantify NETosis that occurs in vivo, various plasmatic tests, and mostly ELISAs, have been developed. Given the large amount of tests available and the lack of homogeneity between them, we decided to perform 3 of them: total DNA, MPO–DNA complexes, and H3Cit. Total DNA dosage is not specific for NETs, as it also measures DNA coming from necrotic cells. MPO–DNA complexes are more specific, as it measures DNA together with MPO that specifically comes from neutrophils, but this assay is often not standardized in publications. Here we used a standardized method, with a calibration curve, to allow precise measurement and reproducibility. Lastly, H3Cit measurement can appear to be the most reliable marker, as it directly measures histone H3 that has been citrullinated, a process that is specific from NETosis. But currently, all available ELISAs lack reproducibility and standardization (43). A Scientific and Standardization Subcommittee of the International Society of Thrombosis and Haemostasis is currently running a study aiming at providing recommendation for NETs’ dosage standardization.

Second, computerized tomography pulmonary angiograms were not systematically performed because of in-hospital transport issues regarding these critically ill patients. We only considered here clinically relevant PE. PE occurrence might have been underestimated, as the pretest probability and clinical likelihood of PE could have presumably been lower in patients treated with therapeutic heparin, especially in the ARDS subgroup. Moreover, transportation to CT scan may have been avoided in the more severe patients because of the risks of transferring patients with critical respiratory failure. As catheter-related thrombosis and limb deep-vein thrombosis screening strategy was heterogeneous among centers, we did not analyze those outcomes.

Taken together, our data demonstrate that circulating markers of NETs are linked to survival but not to PE occurrence in patients with COVID-19 and especially among the most severe ones. Even if measuring NET markers could not be easily implemented in clinical practice to become prognostic biomarkers, our findings are important, as they strengthen the fact that NETosis is a proper therapeutic target in COVID-19 disease, but, more specifically, they argue that NETosis should be targeted before COVID-19 aggravation and ARDS occurrence to be the most efficient.

The raw data supporting the conclusions of this article will be made available by the authors without undue reservation.

According to French law and the French Data Protection Authority, the handling of these data for research purposes was declared to the Data Protection Officer of the University Hospital of Bordeaux. Patients or relatives were notified about the anonymized use of their healthcare data via the departments’ booklets, and non-opposition was recorded. This study was approved by the French institutional authority for personal data protection [Commission Nationale de l’Informatique et des Libertés (CNIL), registration number DEC20-086] and ethics committee (ID-CRB 2020-A00763-36) and by the institutional review board of the University Hospital of Bordeaux (declaration number CE-GP-2020-39). Samples from healthy controls were authorized by the Comité de Protection des Personnes Sud Ouest et Outre Mer III DC 2015/94.

RP and CJ designed the study. ADe, ADu, DG, AO, AR, SS, JG, MC, EJ, JP, and GG included the patients and collected the data. SL-C, AD, LG, SS, GG, CJ, and AR collected the samples. SC performed NET measurements. RP, AD, CJ, SC, AD, AR, and SS analyzed the data. RP, AD, and CJ wrote the article. All the authors read and substantially improved the article.

This work was funded by a COVID-19 grant from the University of Bordeaux and an ANR-COVID (CORONET R21025GG).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.