Skip to main content

CXCL10 could drive longer duration of mechanical ventilation during COVID-19 ARDS



COVID-19-related ARDS has unique features when compared with ARDS from other origins, suggesting a distinctive inflammatory pathogenesis. Data regarding the host response within the lung are sparse. The objective is to compare alveolar and systemic inflammation response patterns, mitochondrial alarmin release, and outcomes according to ARDS etiology (i.e., COVID-19 vs. non-COVID-19).


Bronchoalveolar lavage fluid and plasma were obtained from 7 control, 7 non-COVID-19 ARDS, and 14 COVID-19 ARDS patients. Clinical data, plasma, and epithelial lining fluid (ELF) concentrations of 45 inflammatory mediators and cell-free mitochondrial DNA were measured and compared.


COVID-19 ARDS patients required mechanical ventilation (MV) for significantly longer, even after adjustment for potential confounders. There was a trend toward higher concentrations of plasma CCL5, CXCL2, CXCL10, CD40 ligand, IL-10, and GM-CSF, and ELF concentrations of CXCL1, CXCL10, granzyme B, TRAIL, and EGF in the COVID-19 ARDS group compared with the non-COVID-19 ARDS group. Plasma and ELF CXCL10 concentrations were independently associated with the number of ventilator-free days, without correlation between ELF CXCL-10 and viral load. Mitochondrial DNA plasma and ELF concentrations were elevated in all ARDS patients, with no differences between the two groups. ELF concentrations of mitochondrial DNA were correlated with alveolar cell counts, as well as IL-8 and IL-1β concentrations.


CXCL10 could be one key mediator involved in the dysregulated immune response. It should be evaluated as a candidate biomarker that may predict the duration of MV in COVID-19 ARDS patients. Targeting the CXCL10-CXCR3 axis could also be considered as a new therapeutic approach.

Trial registration, NCT03955887


Since December 2019, the world is experiencing an outbreak of coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Clinical, radiological, and biological differences have been found between COVID-19-related acute respiratory distress syndrome (ARDS) and ARDS from another origin. Indeed, the sudden clinical deterioration observed 1 week after symptom onset, together with deep hypoxemia contrasting with “normal” (> 40 mL/cmH2O) lung compliance, and the higher incidence of thromboembolic events suggest that SARS-CoV-2 is driven by a unique pathogenesis resulting in an atypical form of ARDS [1, 2]. Notably, mechanical ventilation appears to be required for twice as long in COVID-19 than in non-COVID-19 ARDS patients [3, 4].

It has been established that the virus invades type 2 alveolar cells and ciliated epithelial cells that express ACE2 [5]. Subsequently, as they die, infected cells release virus particles and intracellular components including molecules likely to act as alarmins (i.e., danger signals), as reflected by rising lactate dehydrogenase (LDH) levels in the plasma [6]. The ensuing recruitment and activation of immune cells lead to lung damage [7]. For now, several studies have established that the hyperinflammatory response (namely the cytokine storm) induced by SARS-CoV-2 is associated with disease severity and could contribute to the development of ARDS [8, 9]. In addition, since most patients need to undergo mechanical ventilation in this context, ventilator-induced lung injury (VILI) could exacerbate tissue damage as well as local and systemic inflammation, thus acting as a “second hit.” Our team has previously shown that mitochondrial alarmins (i.e., mitochondrial DNA) are released by human epithelial cells submitted to cyclic stretch, and these alarmins are also recovered from bronchoalveolar lavage (BAL) fluid obtained from either ventilated rabbits or ARDS patients. By promoting chemotaxis and activation of polymorphonuclear neutrophils (PMNs), mitochondrial alarmins are thought to represent proximal endogenous mediators of VILI and ARDS when they are released by injured alveolar cells [10, 11].

The aim of our study was to compare cytokine response patterns, in both alveolar and systemic compartments, between COVID-19-related ARDS and non-COVID-19-related ARDS (i.e., ARDS complicating pneumonia from another origin). In addition, we sought to establish the extent to which the immune signature could be associated with clinical evolution according to ARDS etiology.


Study design and patient selection

This is an ancillary study of the ongoing prospective PNEUMOCHONDRIE study ( NCT03955887) which started in June 2019 in three intensive care units and the department of pneumology at the University Hospital of Dijon (France). Patients were eligible if a BAL was considered necessary by the attending physician and if they had (1) pneumonia (≥ 2 acute signs including cough, purulent sputum, dyspnea, chest pain, temperature < 35 °C or ≥ 38.5 °C, and novel radiological pulmonary infiltrate); (2) ARDS (according to the Berlin Definition) [12]; (3) requiring mechanical ventilation (MV); and (4) BAL was performed within 72 h of the start of MV. COVID-19 ARDS patients were eligible if they tested positive for SARS-CoV-2 by reverse transcriptase polymerase chain reaction (RT-PCR) for at least one respiratory sample. The control population consisted of outpatients without fever during the last 15 days, not under mechanical ventilation, and undergoing a BAL for a condition not related to acute infection (evaluation of a pulmonary nodule, chronic interstitial syndrome, or unexplained chronic pulmonary sign). Seven control patients and 7 bacterial ARDS patients were prospectively included between June 11, 2019, and January 20, 2020 (5 weeks before the pandemic COVID-19 started in Burgundy, France). Fourteen COVID-19 ARDS patients were included between March 30 and April 09, 2020. Oral consent was obtained from the patients or their legal representatives. Approval was obtained from the ethics committee (Comité de Protection des Personnes Sud-Est III; 2018-035 B), and an amendment was obtained to include supplementary patients with COVID-19-related ARDS.

Variables of interest, clinical outcomes, and data collection

Demographic data, comorbidities, clinical and biological parameters, and severity scores (Sequential Organ Failure Assessment (SOFA) [13] and the new simplified acute physiology score (SAPS II) [14]) were calculated on the first day of ARDS. Septic shock was defined as persistent hypotension requiring vasopressors and a serum lactate level > 2 mmol/L despite adequate volume resuscitation. Clinical outcomes were recorded up to 30 days after the start of ARDS: 30-day mortality; number of hospital-, ICU-, and ventilator-free days; and hospital-acquired complications (ventilatory-acquired pneumonia (VAP), thrombo-embolic disease). The “ventilator-free days” outcome was defined as the number of days alive from day 1 of ARDS to day 30 during which the patient was breathing without MV. Dedicated clinical research assistants collected all data using a standardized electronic case report form. Automatic checks were generated for missing or incoherent data. All abnormal data were controlled by a physician-scientist.

The concentration of inflammatory cytokines and mitochondrial alarmins (cell-free mitochondrial DNA concentrations) was measured within both the systemic and alveolar compartments.

Sample collection and analysis

Bronchoalveolar lavage was performed by fiberoptic bronchoscopy as part of standard care with sterile saline at 37 °C. The ten first milliliters of aspirated fluid, reflecting a bronchial sample, was not considered for the study. Ten milliliters of BAL fluid (BALF) was dedicated to the study, transported at 4 °C, and used within 2 h of collection. Fifty microliters of whole BALF was homogenized with 200 μL of Thrombo-Plus (Sarstedt) before cell count by light microscopy. BAL was filtered through a 70-μm cell strainer (Fisher) and centrifuged at 500×g for 10 min at 4 °C to remove mucus and cells. The supernatant was then centrifuged again at 3200×g for 5 min at 4 °C to remove the remaining debris and stored at − 80 °C until use. In addition, three additional blood tubes (EDTA) were collected just before the BAL procedure (with a maximum delay of 3 h), as part of standard care, and then centrifuged at 2000×g for 10 min at 4 °C and stored at − 80 °C until use.

Measurement of cytokines

Forty-five analytes were quantified in the plasma and BALF using Human XL Cytokine Magnetic 45-plex Luminex® assay (R&D Systems, USA) according to the manufacturer’s instructions: C-C motif chemokine ligand (CCL)2, CCL3, CCL4, CCL5, CCL11, CCL19, CCL20, CD40 ligand, fractalkine, C-X-C motif chemokine ligand (CXCL) 1, CXCL2, CXCL10, epidermal growth factor (EGF), fibroblast growth factor (FGF), FMS-like tyrosine kinase 3 ligand (FLT3L), granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), granzyme B, interferon (IFN)-alpha, IFN-beta, IFN-gamma, interleukin (IL)-1α, IL-1β, IL-1RA, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12, IL-13, IL-15, IL-17A, IL-17E, IL-33, programmed death-ligand 1 (PDL1), platelet-derived growth factor (PDGF)-AA, PDGF-AB/BB, transforming growth factor (TGF)-α, tumor necrosis factor (TNF)-α, TNF-related apoptosis inducing ligand (TRAIL), and vascular endothelial growth factor (VEGF).

Cell-free mitochondrial DNA

For mitochondrial DNA isolation, collected conditioned media from the plasma and BALF were first centrifuged at 3200×g for 5 min to remove cell debris, followed by DNA extraction using Qiagen DNEasy kit (Qiagen, Valencia, CA, USA). Quantitative PCR was performed on a StepOnePlus™ Real-Time PCR (Applied Biosystem), using SYBR green reagent (PowerUp®, Thermo) with one-tenth and one-fiftieth dilutions of the final product of plasma and BALF respectively, using mitochondrial-specific PCR primers for cytochrome C oxidase subunit III (COXIII) and NADH dehydrogenase subunit I [15], in reference to a standard curve of human mitochondrial DNA to quantify the amount of mitochondrial DNA, amplified as previously reported [10, 11, 15].

SARS-CoV-2 quantification in BALF

RNA extraction was performed on a NucliSENS® easyMAG® platform (Biomerieux, Marcy-l’Étoile, France), from 200 μL of cell-free BALF, and according to the manufacturer’s instructions. Two target genes were amplified and tested simultaneously, namely the RNA-dependent RNA polymerase (RDRP) IP2 and IP4 regions [16]. Amplification was performed using QuantStudio 5 rtPCR Systems (Thermo Fisher Scientific, Waltham, MA, USA). Quantification of the number of RNA copies was done according to a scale ranging from 103 to 106 copies/mL. All positive BALF samples were quantified and expressed as the number of RNA copies/μL.

Epithelial lining fluid

To correct for dilution of BALF, the ELF concentration of all analytes or SARS-CoV-2 was calculated by multiplying the BALF concentration with a dilution factor and using urea, according to the formula [analyte]ELF = [analyte]BALF × [urea]plasma/[urea]BALF, as described by Rennard et al. [17]. Determination of the urea concentration in BALF and plasma was performed using the liquid chromatography tandem mass spectrometry (LC-MSMS) method, adapted from Han et al. [18] with modifications. Plasma (10 μL) or BALF (20 μL) were extracted with 1 mL of cold ethanol containing 1000 ng or 50 ng of 13C and 15N2 urea used as internal standard, respectively. High-performance liquid chromatography (HPLC) was performed on an Agilent 1260 LC system coupled to a 6460 QqQ mass spectrometer. The analysis was conducted in a positive selected reaction monitoring mode. Calibration curves using authentic urea standards (Santa Cruz, Dallas, USA) dissolved in water were prepared. Linear regression was used for calculations.

Statistical analysis

Collected data were described according to the COVID-19 status of the ARDS (i.e., non-COVID-19 or COVID-19) and for the control group. Continuous variables were expressed as median and inter-quartile ranges (IQR) and categorical variables as number and percentages. The univariate analysis consisted of comparisons between variables, according to the COVID-19 status, performed using the chi-square test (Fisher’s exact test if conditions were not met) for percentages and Wilcoxon Mann-Whitney test for medians and IQRs. Subsequently, the association between COVID-19 s non-COVID-19 ARDS and the number of ventilator-free days was assessed using two multivariate median regression models including some other clinically relevant explicative variables: (i) age, septic shock (yes/no) and PaO2/FiO2 (model 1) and (ii) age and baseline SOFA score (model 2). Then, cytokines were presented by boxplots to visualize potential associations between cytokines and COVID-19 status (those for which differences were observed according to the COVID-19 status or according to physio-pathological considerations). Finally, multivariate median regression models were built to test the association with the most pertinent clinical outcome associated with COVID-19 status in univariate analysis comparison, including the five most relevant covariates (i.e., COVID-19 status and the cytokines that differed the most between COVID-19 and non-COVID-19 ARDS patient, either clinically or with a p value below 0.3) and treatments that differ between both groups. In order to avoid overfitting, the choice of variables was not only made on the p value but also considering statistical (i.e., collinearity) and physio-pathological considerations. Model variability explicative power was quantified using the R2 coefficient. Then, mitochondrial DNA levels were compared between controls, non-COVID-19, and COVID-19 ARDS patients using a Kruskal-Wallis analysis of variance test. To account for multiple comparisons, the p value was adjusted for a false discovery rate (FDR) using the Benjamini and Hochberg method. Correlations were sought using the Spearman test between the concentration of cytokines, mitochondrial DNA, and outcome, and also ELF concentrations of SARS-CoV-2, and depicted with a heatmap representation and scatter plot. All tests were two-tailed. A p value lower than 0.05 was considered statistically significant. All analyses were performed using the Stata (13.1, Stata Corporation, College Station, TX, USA) or Prism software (GraphPad Prism®, version 8.0, San Diego, CA, USA).


Characteristics of the study population

Twenty-eight patients were enrolled in this study (7 in the control group, 7 in the non-COVID-19 ARDS group between June 2019 and January 2020, and 14 in the COVID-19 ARDS group between March and April 2020). The control patients were outpatients suffered from sarcoidosis (n = 2), pulmonary Langerhans histiocytosis (n = 1), pulmonary carcinoma (n = 1), scleromyositis (n = 1), focal bronchiectasis (n = 1), or unknown cause (n = 1), and none had any evidence of current infection or acute disease. Bacterial pneumonia was proven in five patients from the non-COVID-19 ARDS group (Pseudomonas aeruginosa (n = 1), Legionella pneumophila (n = 3), Mycoplasma pneumonia (n = 1)), while co-infection occurred in only one COVID-19 ARDS patient (Staphylococcus aureus). While demographic and comorbidity data were not statistically different, baseline severity according to the SOFA score in particular was found to be lower in COVID-19 ARDS patients (p = 0.045), with a trend toward less frequent septic shock (p = 0.120) and a marginally lower baseline arterial lactate levels (p = 0.079) at the onset of ARDS (Table 1). According to the Berlin criteria, ARDS severity was marginally significantly different between non-COVID-19 and COVID-19 patients, with a median baseline PaO2:FiO2 of as low as 68.5 (IQR = 60.9–90.7) and 88.4 (IQR = 79.2–116.6) mmHg, respectively (p = 0.067). Moreover, we found no difference in terms of ventilator settings or lung mechanics. Procalcitonin tended to be only slightly higher in COVID-19 ARDS patients than in the other ARDS patients (0.3 [0.2–0.5] vs. 21.6 [3.8–38.7] μg/L, respectively; p = 0.167) (Table 1).

Table 1 Patient characteristics and outcomes


The 30-day mortality rate reached 14% in the non-COVID-19 group and 21% in the COVID-19 group (Table 1). However, the number of ventilator-free days was significantly lower in COVID-19 patients, compared to non-COVID-19 patients (8 [0–15] vs. 18 [17–21]; p = 0.034), along with a higher rate of ventilator-acquired pneumonia (10 (71%) vs. 0 (0%); p = 0.004). COVID-19 etiology remained associated with a fewer ventilator-free days within the 30 days following admission, even after adjustment on age and baseline severity (i.e., septic shock and PaO2:FiO2 (model 1; p = 0.046) or baseline SOFA score (model 2; p = 0.076)) (Supplementary Tables 1 and 2).

Comparison of systemic and pulmonary concentrations of inflammatory mediators between non-COVID-19 and COVID-19 ARDS patients

Plasma cytokine levels are shown in Fig. 1 and Additional Table 3. COVID-19 ARDS patients had significantly higher plasma concentrations of CCL5 (p = 0.025) and non-significantly higher plasma concentrations of CXCL2 (p = 0.094), CXCL10 (p = 0.287), CD40 ligand (p = 0.125), IL-10 (p = 0.232), and GM-CSF (p = 0.332) compared with non-COVID-19 ARDS patients. At the same time, we observed lower concentrations of plasma IL-2 (p = 0.047), TRAIL (p = 0.059), and G-CSF (p = 0.067). Plasma CXCL10 concentration was independently associated with a greater number of ventilator-free days, after adjustment for the COVID-19 etiology, submission to non-invasive ventilation (NIV) prior to intubation, exposure to multiple antibiotics, CXCL2, CCL5, and CD40 ligand plasma concentrations (p = 0.049) (Table 2). Interestingly, CXCL10 levels were highly correlated with those of GM-CSF (r = 0.991; p < 0.0001) and IL-10 (r = 0.958; p < 0.0001).

Fig. 1

Boxplot graph depicting plasma concentration of cytokines. Plasma cytokines were measured in COVID-19 ARDS (n = 14), non-COVID-19 ARDS (n = 7), and control patients (n = 7). COVID-19 ARDS patients had significantly higher plasma concentrations of CCL5 and non-significantly higher plasma concentrations of CXCL2 (p = 0.09), CXCL10 (p = 0.29), CD40 ligand (p = 0.14), IL-10 (p = 0.23), and GM-CSF (p = 0.33) compared with non-COVID-19 ARDS patients. We observed lower concentrations of plasma IL-2 (p = 0.04), TRAIL (p = 0.055), and G-CSF (p = 0.06). The Mann-Whitney U test was used for comparison between non-COVID-19 and COVID-19 ARDS patients (Pneumochondrie study, 2019–2020)

Table 2 Plasma cytokines associated with the number of ventilator-free days in the 21 patients with ARDS (multivariate median logistic regression; pseudo-R2 = 0.432; n = 21)

Epithelial lining fluid cytokine levels are shown in Fig. 2 and Additional Table 4. A trend toward higher ELF concentrations of CXCL1 (p = 0.287), CXCL10 (p = 0.287), granzyme B (p = 0.110), TRAIL (p = 0.094), and EGF (p = 0.094) is observed in COVID-19 patients. We also observed significantly lower ELF concentrations of IL-2 (p = 0.001), G-CSF (p = 0.046), and IL-17A (p = 0.042) and a trend toward lower concentrations of CCL3 (p = 0.079), CCL4 (p = 0.079), CCL20 (p = 0.172), IL-6 (p = 0.110), INF-γ (p = 0.067), and TNF-α (p = 0.110). CXCL10 concentration in the epithelial lining fluid was independently associated with a greater number of ventilator-free days after adjusting for the COVID-19 etiology, submission to NIV prior to intubation, exposure to multiple antibiotics, TRAIL, TNF-α, and granzyme B ELF concentrations (p = 0.030) (Table 3).

Fig. 2

Boxplot graph depicting epithelial lining fluid (ELF) concentration of cytokines. Epithelial lining fluid concentration of the 16 main cytokines was measured in COVID-19 ARDS (n = 14), non-COVID-19 ARDS (n = 7), and control patients (n = 7). A trend toward higher ELF concentrations of CXCL1 (p = 0.29), CXCL10 (p = 0.29), granzyme B (p = 0.11), TRAIL (p = 0.09), and EGF (p = 0.09) is observed in COVID-19 patients. We also observed significantly lower ELF concentrations of IL-2 (p = 0.001), G-CSF (p = 0.046), and IL-17A (p = 0.037) and a trend toward lower concentrations of CCL3 (p = 0.08), CCL4 (p = 0.08), CCL20 (p = 0.17), IL-6 (p = 0.11), INF-γ (p = 0.06), and TNF-α (p = 0.11). The Mann-Whitney U test was used for comparison between non-COVID-19 and COVID-19 ARDS patients (Pneumochondrie study, 2019–2020)

Table 3 Epithelial lining fluid cytokines associated with the number of ventilator-free days in the 21 patients with ARDS (multivariate median logistic regression; pseudo-R2 = 0.449; n = 21)

Systemic and pulmonary levels of cell-free mitochondrial DNA

We also addressed the magnitude and impact of mitochondrial DNA released into both the systemic and alveolar compartments during non-COVID-19 and COVID-19 ARDS. Plasma and ELF concentrations of mitochondrial DNA were significantly higher in ARDS patients, regardless of COVID-19 involvement (Fig. 3a–d). Interestingly, ELF concentrations of mtDNA were positively correlated with BALF cell count (r = 0.526; p = 0.014), as well as with concentrations of IL-8 (r = 0.647; p = 0.0015), IL-1β (r = 0.637; p = 0.0019), TGF-α (r = 0.582; p = 0.006), EGF (r = 0.532; p = 0.013), and CCL4 (r = 0.508; p = 0.019) (Fig. 3e–f; Additional Figure 1). Conversely, mtDNA concentrations were negatively correlated with CXCL10 (r = − 0.579; p = 0.006) and GM-CSF (r = − 0.519; p = 0.016) (Additional Figure 1). The same results were observed when only COVID-19 ARDS patients were considered (Additional Figure 2). However, neither plasma nor ELF mitochondrial DNA concentrations were correlated with the number of ventilator-free days.

Fig. 3

Boxplot graph depicting the concentration of cell-free mitochondrial DNA in the plasma and epithelial lining fluid (ELF). The concentration of cell-free mitochondrial DNA was measured in the plasma (NADH I (a) and cytochrome c (b) genes) and epithelial lining fluid (NADH I (c) and cytochrome c (d) genes) for COVID-19 ARDS (n = 14), non-COVID-19 ARDS (n = 7), and control patients (n = 7). The Kruskal-Wallis test with a false discovery rate post hoc multiple comparisons correction was used to adjust p values. Correlation between ELF concentrations of NADH I DNA and IL-8 (e) and IL-1β (f). Spearman correlations: *p < 0.05, **p < 0.01, and ***p < 0.01 between each cytokine and ELF concentration of NADH I mitochondrial DNA (Pneumochondrie study, 2019–2020)

Epithelial lining fluid SARS-CoV-2 concentrations

Finally, SARS-CoV-2 concentrations were measured, and correlations with cytokine levels were sought in the alveolar compartment. First, the SARS-Cov-2 genome was detected in 13 COVID-19 ARDS patients with a median of 20,449 (IQR = 893–2,092,104) copies per microliter of ELF. There was a strong negative correlation with ELF IL-6 (r = − 0.719; p = 0.005) and CCL20 (r = − 0.723; p = 0.005), while no correlation was observed with ELF CXCL10 (r = − 0.165, p = 0.573) (Fig. 4a, b).

Fig. 4

Correlations between epithelial lining fluid (ELF) SARS-CoV-2 viral load and ELF concentration of cytokines. Heatmap of the Spearman correlation (r) between epithelial lining fluid (ELF) concentrations of SARS-CoV-2 and ELF concentration of the main correlated cytokines for the 14 COVID-19 ARDS patients (a). Inverse correlation between the ELF concentrations of SARS-CoV-2 and IL-6 (b). Spearman correlations: *p < 0.05 and **p < 0.01 between each cytokine (our outcome) and ELF concentration of SARS-CoV-2 (Pneumochondrie study, 2019–2020)


Our findings suggest that COVID-19 ARDS is associated with prolonged mechanical ventilation in comparison with non-COVID-19 ARDS patients, regardless of the baseline characteristics. This comprehensive evaluation of systemic and pulmonary immune response showed that the higher CXCL10 concentrations in both the systemic and alveolar compartments of patients with COVID-19 ARDS were associated with a longer duration of mechanical ventilation. Interestingly, alveolar concentrations of SARS-CoV-2 were not correlated with alveolar CXCL10 concentration. Finally, in both COVID-19 and non-COVID-19 patients, higher mitochondrial DNA concentrations in the plasma and ELF compartment were highly correlated with alveolar inflammation, as assessed by BALF cell count and ELF IL-8 and IL-1β concentrations. This result highlights the key role of mitochondrial alarmins in ARDS and VILI.

ARDS is the leading cause of death during COVID-19. Given the increasing number of cases, and especially the protracted duration of ARDS, attention was rapidly focused on the number of required ICU beds, ventilators, and intensivists [19]. However, COVID-19-associated ARDS, despite meeting the Berlin Definition, is now considered “atypical” with peculiar features such as the discrepancy between the relatively well-preserved lung mechanics and the depth of hypoxemia [20]. Our findings support these observations since we show that patients with COVID-19 ARDS received prolonged mechanical ventilation when compared to non-COVID-19 ARDS patients, regardless of baseline severity.

In the most severe forms of SARS-CoV-2, common features including systemic cytokine storm (including macrophage activation syndrome), organizing diffuse alveolar damage (acute fibrinous and organizing pneumonia (AFOP)) with excessive immune cells infiltration into the lung (in particular T cell infiltration), and thrombosis have been reported, but these features have also been observed in patients with SARS-CoV-1 infection and Middle East respiratory syndrome (MERS) [9, 21,22,23]. We assume that this atypical dysregulated immune response may cause the lung immunopathology that leads to the protracted and inflammatory nature of COVID-19 ARDS. Since there is no available effective direct anti-viral treatment so far, mitigating the “cytokine storm” through immune modulation is a promising therapeutic avenue. However, it is worth noting that most of the currently tested immunomodulatory agents (especially anti-IL-6, anti-IL-1β, and anti-TNFα) have been put forward despite gaps in our understanding of the immune response behind COVID-19 ARDS, especially within the pulmonary compartment. However, assessing the alveolar compartment is challenging in severe ARDS patients. To the best of our knowledge, there is currently no published study comparing cytokine concentrations in bronchoalveolar lavage fluid obtained from either COVID-19 or non-COVID-19 ARDS patients, and basically, data regarding inflammation within the alveolar space of the former patients are sparse [24]. We report herein new insights into this issue.

We show that plasma CXCL2, CXCL10, CCL5, CD40 ligand, IL-10, and GM-CSF concentrations and ELF CXCL1, CXCL10, granzyme B, TRAIL, and EGF concentrations were found in greater concentrations in COVID-19 than in non-COVID-19 ARDS patients. Interestingly, almost all of these mediators are chemokines or cytokines involved in either recruitment or activation of T lymphocytes (CCL5, CXCL10) and/or monocytes/macrophages (GM-CSF). Others are associated with anti-inflammation (IL-10 and TRAIL) or endothelial dysfunction (CD40 ligand, EGF). Among them, we found that CXCL10 was the most likely to account for the protracted nature of COVID-19 ARDS in both the systemic and the alveolar compartments. CXCL10 (or INF-γ-induced protein (IP-10)) is secreted upon INF-γ stimulation by various cell types (endothelial cells, fibroblasts, monocytes/macrophages, and T lymphocytes) and promotes chemoattraction for activated T lymphocytes, natural killer cells, and monocytes through CXCR3 [25]. Interestingly, plasma concentrations of CXCL10 have been previously reported at high levels in SARS-CoV-1 [26], respiratory syncytial [27], and influenza infections [28], and significantly associated with a higher risk of death in ARDS associated with A/H1N1 influenza infection [29]. Yang et al. recently reported that among 44 inflammatory mediators measured in COVID-19 patients, the plasma concentration of CXCL10 was highly associated with disease severity, especially the Murray score, and could independently predict COVID-19 disease progression [9]. Interestingly, Ichikawa et al. showed that the CXCL10-CXCR3 signaling pathway was critical in viral and non-viral ARDS pathogenesis. Moreover, they showed that lung injury could be prevented, and therefore, survival improved, when the CXCL10-CXCR3 axis was blocked [30]. Since our results emphasize the involvement of the CXCL10-CXCR3 signaling axis in the pathogenesis of the most severe forms of COVID-19 infection, this axis represents a potential therapeutic target. Corticosteroids that showed beneficial effects in the most severe forms of COVID-19 [31] may act upstream from CXCL10 through inhibition of the Th1 pathway. However, specifically blocking CXCL10 (e.g., eldelumab/MDX-1100) or CXCR3 may be a promising therapeutic approach. Interestingly, Lev et al. reported that upon the administration of corticosteroid, a significant decrease of CXCL10 levels was observed in COVID-19 patients [32]. In addition, if validated in larger cohorts, plasma CXCL10 measurement could be helpful in predicting the risk of prolonged MV.

Since plasma CXCL-10 and GM-CSF were strongly correlated (r = 0.991, p < 0.0001), the latter was also highly associated with a longer duration of mechanical ventilation. The most severe forms of COVID-19 were associated with macrophage activation syndrome, characterized by a fulminant hypercytokinemia with multiorgan failure [23]. GM-CSF has recently emerged as a potential therapeutic target, since it plays a pivotal role in initiation (monocyte activation and macrophage transformation) and perpetuation of the immune response and could represent the link between T cell-driven acute pulmonary inflammation and self-amplifying cytokine loop leading to monocyte/macrophage activation [33, 34]. Interestingly, an anti-GM-CSF therapy (lenzilumab) has been previously evaluated in severe COVID-19 and associated with improved clinical outcomes, oxygen requirement, and cytokine release [35].

Still, a cautious approach is required since any immune modulation is likely to alter the body’s antimicrobial defenses. This point is particularly relevant, since deep dysfunctions of the myeloid and lymphoid responses have also been described in COVID-19 patients, with a decrease of HLA-DR expression on monocytes [23] and T cell exhaustion [36]. Interestingly, no correlation was found between the alveolar viral load of SARS-CoV-2 and ELF CXCL10, suggesting that activation of the CXCL10-CXCR3 axis may play a role in the dysregulated immune response independently from viral clearance. Thus, its blockage might not compromise the host’s ability to bring the virus under control. In contrast, we observed a strong negative correlation between ELF concentrations of SARS-CoV-2 and IL-6, suggesting that IL-6 production within the alveolar compartment is associated with effective viral clearance. In addition, dampening inflammation in a context of high immune suppression level is not always a hazardous route. In the setting of chimeric antigen receptor T (CAR-T) cell therapy, GM-CSF inhibition reduces cytokine release syndrome and neuro-inflammation but enhances antitumoral CART-T cell function [37]. Moreover, the IL-6 blocker partially rescues immune dysregulation caused by SARS-CoV-2 [23]. Nevertheless, as advocated for sepsis [38], immune profiling (including CXCL10 measurement) may serve for the selection of patients that could be eligible for immunotherapy.

Finally, it is worth noting that higher viral alveolar loads were associated with more severe ARDS in terms of blood oxygenation and remote organ failure [39], highlighting the need for drugs likely to prevent COVID-19 replication in addition to therapies targeting host response.

In addition to our findings regarding the CXCL10-CXCR3 axis, we observed that the released amount of mediators involved in chemotaxis and/or activation of PMNs (i.e., IL-8, IL-1β, IL-6, and TNF-α) was similar when ARDS was caused by SARS-CoV-2. Moreover, alveolar cell counts and ELF concentrations of IL-8 and Il-1β were highly correlated with the release of mitochondrial DNA, whose levels were significantly increased in both the systemic and pulmonary compartments of ARDS patients, regardless of etiology. It is worth noting that cell-free mitochondrial DNA release can elicit neutrophil-mediated lung injury through the promotion of IL-8 and Il-1β secretion by activation of the TLR-9 (Toll-like receptor 9) and the NLRP-3 (NOD-like receptor pyrin domain 3) inflammasome pathways, respectively [15, 40]. This finding suggests that systemic and pulmonary immune response could be triggered in part by the release of endogenous mediators originating from injured alveolar cells. This reaction may be subsequent to a two-hit lung injury (i.e., infection of alveolar epithelial cells and ventilator-induced lung injury), as suggested by previous clinical and experimental findings from our group [10, 11].

Surprisingly, we observed a trend toward lower ELF concentrations of IL-6 in COVID-19 patients compared with non-COVID-19 ARDS patients (p = 0.11), thus challenging the interest of anti-IL-6 therapies in severe COVID-19 pneumonia. These results are in line with retrospective observations by Sinha et al., showing that plasma IL-6 concentrations were lower in the severe form of COVID-19 as compared to those reported in ARDS from another origin and arguing that the term “cytokine storm” may be misleading in COVID-19 ARDS [41].

Finally, it is worth noting that early bacterial coinfection was unlikely in COVID-19 ARDS patients from our cohort, in accordance with previously published data [42, 43]. One could argue that previous exposure to antibiotics has led to false-negative results. More interestingly, it ascertains the safety of introducing new therapies (i.e., including those targeting CXCL-10) likely to dampen the host inflammatory response at this stage of the viral disease. This statement has to be mitigated by the fact that VAP complicated ARDS in 71% of the COVID-19 patients, as compared to 0% in patients with ARDS from another origin. Although the longer duration of MV could account at least in part for such an obvious difference, one cannot exclude that VAP occurrence reflected exhaustion of the lung immune defense following the strong activation of the CXCL10-CXCR3 axis. Moreover, the impact of blocking this critical pathway remains uncertain regarding the risk of subsequent bacterial infection.

Our study has some limitations. First, the origin of non-COVID-19 ARDS was mainly Gram-negative bacteria, which is not representative of all ARDS of pulmonary origin. As a result of the difficulty in obtaining BALF, the small sample size prevented us from performing a mortality analysis and resulted in a lack of power for some comparisons. As a result, our findings should be taken cautiously, and new studies are required in order to ascertain our hypothesis. However, pulmonary compartment assessment remains challenging, and there is currently no published data comparing BALF in COVID-19 and non-COVID-19 ARDS patients. Moreover, the criteria of inclusion of our control group remain questionable, and non-infected ICU patients under MV would have been preferred but not possible given ethical concerns related to the fiberoptic procedure, which could not be considered as part of standard care. One can also argue that in the context of such an emerging disease, some treatments likely to influence the release of cytokines could have been given to the COVID-19 patients prior to the study inclusion. However, such treatments (i.e., hydrocortisone and hydroxychloroquine) were administrated prior to blood and BALF collection in a small number of patients. Moreover, the time elapsed between treatment administration and sampling was quite short, making unlikely any impact on cytokine concentrations (Additional Table 5). Finally, the specific impact of mechanical ventilation on pulmonary immune response cannot be evaluated since patients with less severe disease who did not require MV were not included.


This study provides new insights into the peculiar pathogenesis of COVID-19 ARDS. First, CXCL10 may represent the dysregulated immune response that drives the duration of MV in COVID-19 ARDS patients. CXCL10 appears to be a potential biomarker for the duration of MV, and the CXCL10-CXCR3 signaling axis may be a potential therapeutic target in COVID-19 ARDS. This target seems all the more interesting since no correlation was found between alveolar concentrations of CXCL10 and SARS-CoV-2 clearance. Apart from those particular features, COVID-19 and non-COVID-19 ARDS share the well-known involvement of both IL-8 and IL-1β within the airway, potentially driven by the release of endogenous mediators originating from injured alveolar cells (e.g., mitochondrial alarmins).

Availability of data and materials

All data are available on demand.



Acute respiratory distress syndrome


Bronchoalveolar lavage


Bronchoalveolar lavage fluid


C-C motif chemokine ligand


Coronavirus disease 2019


C-X-C motif chemokine ligand


CXC chemokine receptor 3


Deoxyribonucleic acid


Extracorporeal membrane oxygenation


Epithelial lining fluid


Epidermal growth factor


Fibroblast growth factor


FMS-like tyrosine kinase 3 ligand


Granulocyte colony-stimulating factor


Granulocyte-macrophage colony-stimulating factor


Intensive care medicine






INF-γ-induced protein


Interquartile range


Mechanical ventilation


Nicotinamide adenine dinucleotide I


NOD-like receptor pyrin domain 3

PaO2:FiO2 :

Arterial pressure of oxygen/oxygen inspiratory fraction


Programmed death-ligand 1


Platelet-derived growth factor


RNA-dependent RNA polymerase


Reverse transcriptase polymerase chain reaction


Simplified Acute Physiology Score II


Severe acute respiratory syndrome coronavirus 2


Sequential Organ Failure Assessment


Transforming growth factor


Toll-like receptor 9


Tumor necrosis factor


TNF-related apoptosis-inducing ligand


Ventilator-acquired pneumonia


Vascular endothelial growth factor


  1. 1.

    Helms J, Tacquard C, Severac F, Leonard-Lorant I, Ohana M, Delabranche X, et al. High risk of thrombosis in patients with severe SARS-CoV-2 infection: a multicenter prospective cohort study. Intensive Care Med. 2020;46(6):1089–98.

    CAS  Article  PubMed  Google Scholar 

  2. 2.

    Marini JJ, Gattinoni L. Management of COVID-19 respiratory distress. JAMA. 2020;323(22):2329–30.

  3. 3.

    Bellani G, Laffey JG, Pham T, Fan E, Brochard L, Esteban A, et al. Epidemiology, patterns of care, and mortality for patients with acute respiratory distress syndrome in intensive care units in 50 countries. JAMA. 2016;315(8):788–800.

    CAS  Article  PubMed  Google Scholar 

  4. 4.

    Cummings MJ, Baldwin MR, Abrams D, Jacobson SD, Meyer BJ, Balough EM, et al. Epidemiology, clinical course, and outcomes of critically ill adults with COVID-19 in New York City: a prospective cohort study. Lancet. 2020;395(10239):1763–70.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  5. 5.

    Sungnak W, Huang N, Bécavin C, Berg M, Queen R, Litvinukova M, et al. SARS-CoV-2 entry factors are highly expressed in nasal epithelial cells together with innate immune genes. Nat Med. 2020;26(5):681–7.

    CAS  Article  PubMed  Google Scholar 

  6. 6.

    Mo P, Xing Y, Xiao Y, Deng L, Zhao Q, Wang H, et al. Clinical characteristics of refractory COVID-19 pneumonia in Wuhan, China. Clin Infect Dis. Clin Infect Dis. 2020;ciaa270.

  7. 7.

    Felsenstein S, Herbert JA, McNamara PS, Hedrich CM. COVID-19: immunology and treatment options. Clin Immunol. 2020;215:108448.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  8. 8.

    Huang C, Wang Y, Li X, Ren L, Zhao J, Hu Y, et al. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet. 2020;395(10223):497–506.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  9. 9.

    Yang Y, Shen C, Li J, Yuan J, Wei J, Huang F, et al. Plasma IP-10 and MCP-3 levels are highly associated with disease severity and predict the progression of COVID-19. J Allergy Clin Immunol. 2020;146(1):119–127.e4.

  10. 10.

    Grazioli S, Dunn-Siegrist I, Pauchard L-A, Blot M, Charles P-E, Pugin J. Mitochondrial alarmins are tissue mediators of ventilator-induced lung injury and ARDS. PLoS One. 2019;14(11):e0225468.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  11. 11.

    Blot M, Pauchard L-A, Dunn I, Donze J, Malnuit S, Rebaud C, et al. Mechanical ventilation and Streptococcus pneumoniae pneumonia alter mitochondrial homeostasis. Sci Rep. 2018;8(1):11718.

    Article  PubMed  PubMed Central  Google Scholar 

  12. 12.

    Definition Task Force ARDS, Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND, Caldwell E, et al. Acute respiratory distress syndrome: the Berlin Definition. JAMA. 2012;307(23):2526–33.

    Google Scholar 

  13. 13.

    Vincent JL, Moreno R, Takala J, Willatts S, De Mendonça A, Bruining H, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med. 1996;22(7):707–10.

    CAS  Article  Google Scholar 

  14. 14.

    Le Gall JR, Lemeshow S, Saulnier F. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA. 1993;270(24):2957–63.

    Article  Google Scholar 

  15. 15.

    Zhang Q, Raoof M, Chen Y, Sumi Y, Sursal T, Junger W, et al. Circulating mitochondrial DAMPs cause inflammatory responses to injury. Nature. 2010;464(7285):104–7.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  16. 16.

    Corman VM, Landt O, Kaiser M, Molenkamp R, Meijer A, Chu DK, et al. Detection of 2019 novel coronavirus (2019-nCoV) by real-time RT-PCR. Euro Surveill. 2020;25(3)..

  17. 17.

    Rennard SI, Basset G, Lecossier D, O’Donnell KM, Pinkston P, Martin PG, et al. Estimation of volume of epithelial lining fluid recovered by lavage using urea as marker of dilution. J Appl Physiol. 1986;60(2):532–8.

    CAS  Article  Google Scholar 

  18. 18.

    Han L, Huang X, Zhang L, Zhang Q, Wang J, Lin H, et al. Candidate reference measurement procedure for determination of urea in serum by liquid chromatography-tandem mass spectrometry. J Pharm Biomed Anal. 2019;162:124–9.

    CAS  Article  Google Scholar 

  19. 19.

    Bhatraju PK, Ghassemieh BJ, Nichols M, Kim R, Jerome KR, Nalla AK, et al. Covid-19 in critically ill patients in the Seattle region - case series. N Engl J Med. 2020;382(21):2012–22.

    CAS  Article  Google Scholar 

  20. 20.

    Gattinoni L, Coppola S, Cressoni M, Busana M, Rossi S, Chiumello D. COVID-19 does not lead to a “typical” acute respiratory distress syndrome. Am J Respir Crit Care Med. 2020;201(10):1299–300.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  21. 21.

    Schaller T, Hirschbühl K, Burkhardt K, Braun G, Trepel M, Märkl B, et al. Postmortem examination of patients with COVID-19. JAMA. 2020;323(24):2518–20.

  22. 22.

    Channappanavar R, Perlman S. Pathogenic human coronavirus infections: causes and consequences of cytokine storm and immunopathology. Semin Immunopathol. 2017;39(5):529–39.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  23. 23.

    Giamarellos-Bourboulis EJ, Netea MG, Rovina N, Akinosoglou K, Antoniadou A, Antonakos N, et al. Complex immune dysregulation in COVID-19 patients with severe respiratory failure. Cell Host Microbe. 2020;27(6):992–1000.e3.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  24. 24.

    Carvelli J, et al. Association of COVID-19 inflammation with activation of the C5a-C5aR1 axis. Nature. 2020.

  25. 25.

    Jiang Y, Xu J, Zhou C, Wu Z, Zhong S, Liu J, et al. Characterization of cytokine/chemokine profiles of severe acute respiratory syndrome. Am J Respir Crit Care Med. 2005;171(8):850–7.

    Article  PubMed  Google Scholar 

  26. 26.

    Wong CK, Lam CWK, Wu AKL, Ip WK, Lee NLS, Chan IHS, et al. Plasma inflammatory cytokines and chemokines in severe acute respiratory syndrome. Clin Exp Immunol. 2004;136(1):95–103.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  27. 27.

    Miller AL, Bowlin TL, Lukacs NW. Respiratory syncytial virus-induced chemokine production: linking viral replication to chemokine production in vitro and in vivo. J Infect Dis. 2004;189(8):1419–30.

    CAS  Article  PubMed  Google Scholar 

  28. 28.

    Guo J, Huang F, Liu J, Chen Y, Wang W, Cao B, et al. The serum profile of hypercytokinemia factors identified in H7N9-infected patients can predict fatal outcomes. Sci Rep. 2015;5:10942.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  29. 29.

    Bautista E, Arcos M, Jimenez-Alvarez L, García-Sancho MC, Vázquez ME, Peña E, et al. Angiogenic and inflammatory markers in acute respiratory distress syndrome and renal injury associated to A/H1N1 virus infection. Exp Mol Pathol. 2013;94(3):486–92.

    CAS  Article  PubMed  Google Scholar 

  30. 30.

    Ichikawa A, Kuba K, Morita M, Chida S, Tezuka H, Hara H, et al. CXCL10-CXCR3 enhances the development of neutrophil-mediated fulminant lung injury of viral and nonviral origin. Am J Respir Crit Care Med. 2013;187(1):65–77.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  31. 31.

    RECOVERY Collaborative Group, Horby P, Lim WS, Emberson JR, Mafham M, Bell JL, et al. Dexamethasone in hospitalized patients with Covid-19 - preliminary report. N Engl J Med. 2020;2021436.

  32. 32.

    Lev S, Gottesman T, Levin GS, Lederfein D, Berkov E, Diker D, et al. Real-time IP-10 measurements as a new tool for inflammation regulation within a clinical decision support protocol for managing severe COVID-19 patients. medRxiv. 2020;20158782.

  33. 33.

    Lang FM, Lee KM-C, Teijaro JR, Becher B, Hamilton JA. GM-CSF-based treatments in COVID-19: reconciling opposing therapeutic approaches. Nat Rev Immunol. 2020;20(8):507–14.

  34. 34.

    Mehta P, Porter JC, Manson JJ, Isaacs JD, Openshaw PJM, McInnes IB, et al. Therapeutic blockade of granulocyte macrophage colony-stimulating factor in COVID-19-associated hyperinflammation: challenges and opportunities. Lancet Respir Med. 2020;8(8):822–30.

  35. 35.

    Temesgen Z, Assi M, Vergidis P, Rizza SA, Bauer PR, Pickering BW, et al. First clinical use of lenzilumab to neutralize GM-CSF in patients with severe COVID-19 pneumonia. medRxiv.2020;20125369.

  36. 36.

    Remy KE, Mazer M, Striker DA, Ellebedy AH, Walton AH, Unsinger J, et al. Severe immunosuppression and not a cytokine storm characterize COVID-19 infections. JCI Insight. 2020;5(17):e140329.

  37. 37.

    Sterner RM, Sakemura R, Cox MJ, Yang N, Khadka RH, Forsman CL, et al. GM-CSF inhibition reduces cytokine release syndrome and neuroinflammation but enhances CAR-T cell function in xenografts. Blood. 2019;133(7):697–709.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  38. 38.

    van der Poll T, van de Veerdonk FL, Scicluna BP, Netea MG. The immunopathology of sepsis and potential therapeutic targets. Nat Rev Immunol. 2017;17(7):407–20.

    Article  Google Scholar 

  39. 39.

    Blot M, Jacquier M, Manoha C, Piroth L, Charles P-E. Pneumochondrie study group. Alveolar SARS-CoV-2 viral load is tightly correlated with severity in COVID-19 ARDS. Clin Infect Dis. 2020;ciaa1172.

  40. 40.

    Shimada K, Crother TR, Karlin J, Dagvadorj J, Chiba N, Chen S, et al. Oxidized mitochondrial DNA activates the NLRP3 inflammasome during apoptosis. Immunity. 2012;36(3):401–14.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  41. 41.

    Sinha P, Matthay MA, Calfee CS. Is a “cytokine storm” relevant to COVID-19? JAMA Intern Med. 2020.

  42. 42.

    Rawson TM, Moore LSP, Zhu N, Ranganathan N, Skolimowska K, Gilchrist M, et al. Bacterial and fungal co-infection in individuals with coronavirus: a rapid review to support COVID-19 antimicrobial prescribing. Clin Infect Dis. Clin Infect Dis. 2020;ciaa530.

  43. 43.

    Hughes S, Troise O, Donaldson H, Mughal N, Moore LSP. Bacterial and fungal coinfection among hospitalized patients with COVID-19: a retrospective cohort study in a UK secondary-care setting. Clin Microbiol Infect. 2020;26(10):1395–99.

Download references


The authors acknowledge all the members of the PNEUMOCHONDRIE study group (François Aptel, Auguste Dargent, Marjolaine Georges, Marie Labruyère, Laurent Lagrost, Audrey Large, Serge Monier, Jean-Baptiste Roudaut, Charles Thomas). We thank Hélène Choubley from the Lipidomic Analytical Platform of the University of Burgundy and Serge Monier and Anabelle Sequeira from the Cytometry Core Facility which are supported by the following institutions: Conseil Régional de Bourgogne Franche-Comté and FEDER. We thank Alexandra Lamotte-Felin, Lydie Rossie, and Aurélie Degot from the CIC-EC (CIC-EC1432) and Solenne Villot and Mathilde Audry from the Intensive Care Medicine Research team. We thank Suzanne Rankin for the editing assistance, Maud Carpentier from the DRCI (Direction de la Recherche Clinique et de l’Innovation), Shaliha Bechoua (Centre de Ressources Biologiques), and Cécile Pitoiset and Magali Darnio for the technical assistance (virology). We thank the patients.


This work was supported by grants: the Appel à Projet Jeunes Chercheurs 2018 (MB) from the Groupement Interrégional de Recherche Clinique et d’Innovation Est (GIRCI), the INSERM (Centre de Recherche UMR 1231), and the ANR Investissements d’Avenir Grant (ANR-11 LABX-0021-01, Labex Lipstic), the FEDER (Fonds européens de développement regional), and the Regional Council of Bourgogne.

Author information





Designing the research study: MB, MJ, LSA, CB, and PEC. Conducting the experiment: MB, MJ, JPP, and DM. Acquiring the data: MB, MJ, LSA, JPP, DM, CB, and PEC. Analyzing the data: MB, MJ, CB, LSA, LP, and PEC. Main clinical investigators: GB, PB, MN, BB, JPQ, PEC, SP, and PA. Writing the manuscript: MB, MJ, LSA, JPP, LP, and PEC. The authors read and approved the final manuscript.

Corresponding author

Correspondence to Mathieu Blot.

Ethics declarations

Ethics approval and consent to participate

Approval was obtained from the ethics committee (Comité de Protection des Personnes Sud-Est III; 2018-035 B), and an amendment was obtained to include supplementary patients with COVID-19. Oral consent was obtained from the patient or their legal representatives.

Consent for publication

All authors consented for the publication.

Competing interests

We declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Additional file 1: Table 1.

Model 1: factors associated with the number of ventilator-free days in the 21 patients with ARDS (multivariate median logistic regression; pseudo R2 = 0.237; n=21, Pneumochondrie study, 2019-2020). Table 2. Model 2: factors associated with the number of ventilator-free days in the 21 patients with ARDS (multivariate median logistic regression; pseudo R2 = 0.223; n=21, Pneumochondrie study, 2019-2020). Table 3. Plasma concentrations of cytokines. Table 4. Epithelial lining fluid concentrations of cytokines. Figure 1. Heatmap of the Spearman correlation (r) between epithelial lining fluid (ELF) concentrations of mitochondrial DNA (NADH I and Cytochrome C), ELF concentration of cytokines, and outcomes for the 21 ARDS patients. Spearman correlations: p<0.05 *; p<0.01 ** between each cytokine and the ELF concentration of NADH I mitochondrial DNA. Figure 2. Heatmap of the Spearman correlation (r) between the epithelial lining fluid (ELF) concentrations of mitochondrial DNA (NADH I and Cytochrome C), the ELF concentration of cytokines and outcomes for the 14 COVID-19 ARDS patients (A). Spearman correlations: p<0.05 *; p<0.01 ** between each cytokine and the ELF concentration of NADH I mitochondrial DNA.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Blot, M., Jacquier, M., Aho Glele, LS. et al. CXCL10 could drive longer duration of mechanical ventilation during COVID-19 ARDS. Crit Care 24, 632 (2020).

Download citation


  • Acute respiratory distress syndrome
  • COVID-19
  • SARS-CoV-2
  • Mechanical ventilation
  • Immune response
  • Bronchoalveolar lavage
  • CXCL10
  • Mitochondrial DNA
  • Biomarker