Is COVID-19 different from other causes of acute respiratory distress syndrome?

Abstract

Coronavirus disease 2019 (COVID-19) pneumonia can lead to acute hypoxemic respiratory failure. When mechanical ventilation is needed, almost all patients with COVID-19 pneumonia meet the criteria for acute respiratory distress syndrome (ARDS). The question of the specificities of COVID-19-associated ARDS compared to other causes of ARDS is of utmost importance, as it may justify changes in ventilatory strategies. This review aims to describe the pathophysiology of COVID-19-associated ARDS and discusses whether specific ventilatory strategies are required in these patients.

Introduction

The coronavirus disease 2019 (COVID-19) pandemic was associated with a massive influx of patients suffering from acute respiratory failure, leading to important clinical and healthcare organizational questions. The severity of the disease varies according to the patient's characteristics and the severe acute respiratory syndrome coronavirus 2 (SARS-COV-2) strain; however, when mechanical ventilation is needed, almost all patients with COVID-19 pneumonia meet the criteria for acute respiratory distress syndrome (ARDS).[1], [2] ARDS is defined as a heterogeneous syndrome and includes patients with various etiologies of acute lung injury.^[3] From a physiological point of view, “classical” ARDS is characterized by lung edema leading to a decrease in aerated lung volume and an increase in shunt fraction responsible for hypoxemia. The decreased lung volume is associated with decreased respiratory system compliance that can be easily measured at the bedside.^[4] Although some pathophysiological mechanisms may vary according to ARDS risk factors and some sub-phenotypes with specific features have been described, the global ventilatory strategies proposed for patients with ARDS are similar regardless of the suspected cause.[5], [6] Very early during the pandemic crisis, questions regarding the specificities of this new disease emerged. Several hypotheses emerged from the literature. In particular, some authors described that this viral infection may lead to specific vascular injuries either not observed or rarely observed in other causes of ARDS.^[7] Moreover, significant differences in respiratory mechanics including, in particular, a high respiratory system compliance and an increased dead space fraction in COVID-19-associated ARDS patients compared to patients with “classical forms” of ARDS have been suggested by some research groups.[8], [9] From a clinical point of view, this question regarding COVID-19-associated ARDS specificities is of utmost importance, because it may lead to changes in the proposed ventilatory strategies to manage these patients with ARDS.^[10] In addition, the huge number of critically ill patients has led intensivists to adapt intensive care units’ structures and strategies to attempt to take care of as many patients as possible. Differences in patient management between COVID-19 and non-COVID-19-associated ARDS may therefore explain some differences in the physiological measurements observed between these two patient populations. This review aims to describe the pathophysiology of COVID-19-associated ARDS and discuss whether it is substantially different from ARDS unrelated to COVID-19.

Compliances and Recruitability in Patients with COVID-19-associated ARDS

Although it was not included in the criteria of the Berlin definition, decreased respiratory system compliance may be considered a key feature of ARDS pathophysiology.[2], [11] This decreased compliance because of a decrease in the volume of aerated lung tissue (and not to a “stiff lung”) observed in computed tomography-based studies led to the description of the concept of “baby lung” by Gattinoni et al.^[4] The presence of non-aerated lung tissue (due to lung edema) is associated with intra-pulmonary shunt, responsible for most of the hypoxemia observed in patients with ARDS.^[12] Based on clinical observations and very small descriptive series, some authors suggested very early in the COVID-19 crisis that COVID-19-associated ARDS may be characterized by severe hypoxemia despite relatively normal respiratory system compliance.[9], [10] In addition, according to these first reports, some of the patients with COVID-19-associated ARDS were supposed to present with lower levels of recruitability than those with “classical” ARDS.[9], [10] A high level of recruitability is defined by a high volume of lung tissue in which aeration can be restored with an increase in applied positive end-expiratory pressure (PEEP).^[13]

Studies published later on aiming to assess the respiratory mechanics of COVID-19-associated ARDS showed heterogeneous results.[8], [14], [15], [16], [17], [18], [19] Some studies actually showed a slightly higher respiratory system compliance in patients with COVID-19-associated ARDS than in those with non-COVID-19-associated ARDS when the partial pressure of oxygen (PaO₂) /fraction of inspired oxygen (FiO₂) ratios were similar.[8], [14], [15], [18] However, the clinical relevance of such small differences may be questioned. Some other studies did not find any difference in the respiratory mechanics between patients with and without COVID-19-associated ARDS.[16], [17], [20] In a study comparing the evolution over a 7-day period of respiratory mechanics in COVID-19 and non-COVID-19 pulmonary ARDS matched for age, Simplified Acute Physiology Score (SAPS) II, PaO₂/FiO₂ ratio, and PEEP level, patients with COVID-19-associated ARDS exhibited a slightly but significantly higher respiratory system compliance at day 1 after ARDS diagnosis than those with pulmonary non-COVID-19-associated ARDS.^[15] On day 7, the respiratory system compliance did not differ between COVID-19 and non-COVID-19 patients; however, the PaO₂/FiO₂ ratio was lower in COVID-19 than in non-COVID-19 patients. Multivariate analyses showed that the difference in respiratory system compliance observed on day 1 no longer existed on day 7 after adjustment of the PEEP level and PaO₂/FiO₂ ratio. These data suggested that the natural course of evolution after intubation tends to erase the slight differences in compliance initially observed between COVID-19 and non-COVID-19-associated ARDS. Furthermore, the patient's phenotype subgroup presented relatively high compliance and a low PaO₂/FiO₂ ratio which has been suggested to be frequent in COVID-19-associated ARDS and has been shown to be frequent in the very large international lung-safe study published in 2016 including 2377 patients with non-COVID-19-associated ARDS from 459 intensive care units (ICUs) and 50 countries.[21], [22] Furthermore, a recently published meta-analysis showed that respiratory system compliance measured early after intubation was normally distributed in patients with COVID-19-associated ARDS, and no specific clinical phenotypes could be distinguished in these patients based on respiratory mechanics features.^[23]

Recruitability was not different between patients with and without COVID-19-associated ARDS in two physiological studies that used recruitment-to-inflation (R/I) ratio measurements.[14], [17], [21], [22] The R/I ratio is a recently described bedside tool that uses a drop in PEEP over a single breath maneuver to estimate the recruited volume over a given range of PEEP.^[24] A high R/I ratio is associated with a high potential for lung recruitment with PEEP, but it is important to note that the recruitability can significantly vary over time. In another physiological study that only included patients with COVID-19-associated ARDS, most patients considered poorly recruitable on day 1 after intubation was considered highly recruitable on day 5.^[25] Of note, neither oxygenation nor respiratory system compliance measured either at a low or high level of PEEP nor changes to these parameters with PEEP level changes could be used to identify highly recruitable or poorly recruitable patients.^[25]

To sum up, patients with COVID-19-associated ARDS may exhibit early after intubation slightly higher respiratory system compliance than those with non-COVID-19-associated ARDS when the PaO₂/FiO₂ ratios were similar (Table 1). However, this difference may be related to differences in ventilatory management between COVID-19 and non-COVID-19 patients and tends to disappear after several days of mechanical ventilation.

Table 1.

Specificities and common features of COVID-19-associated ARDS compared to “classical forms” of ARDS.

Items	Specificities of COVID-19-associated ARDS	Common features
Anatomopathological findings	Specific vascular features: much more prevalent microangiopathy with alveolar capillary microthrombi intussusceptive angiogenesis	Diffuse alveolar damage
Respiratory system compliance	May be slightly higher early after intubation	Similar after several days of mechanical ventilation
Recruitability	No evidence of difference (can significantly vary over time)
Dead space fraction	Might be higher (might increase with duration of mechanical ventilation)
Mechanisms of hypoxemia	Higher ventilation-perfusion mismatch due to marked alteration of hypoxic pulmonary vasocontraction?	Shunt due to non-aerated lung volume (lung edema)
Outcomes	Longer duration of mechanical ventilation	No difference in mortality when adjusted for main risk factors

COVID-19: Coronavirus disease 2019; ARDS: Acute respiratory distress syndrome.

Ventilation-perfusion Mismatch, Dead Space Fraction, and Specific Vascular Injuries in COVID-19-associated ARDS

The possible dissociation between marked hypoxemia and non-severely altered respiratory mechanics (i.e., the small amount of non-aerated lung tissue) that may be observed in patients with COVID-19-associated ARDS is likely explained by greater perfusion of regions with low ventilation to perfusion ratio related to a marked alteration of hypoxic pulmonary vasocontraction.^[15] Hypoxic pulmonary vasoconstriction is a reflex contraction of pulmonary circulation vascular smooth muscle in response to a low regional oxygen partial pressure.^[26] It represents a very important physiological response in hypoxemic respiratory failure, minimizing arterial hypoxemia owing to shunt or low ventilation-perfusion area. It is possible that alteration in COVID-19 pneumonia may be related to multiple mechanisms such as the reduction in alveolar partial pressure in carbon dioxide (CO₂), increased nitric oxide production via cytokine-mediated upregulation of inducible nitric oxide synthase activity, or changes in mitochondrial proteins and transduction pathways involved in oxygen sensing. In addition, pulmonary artery smooth muscle cells are known to express angiotensin-converting enzyme 2 (ACE-2), which is targeted by SARS-COV-2 and may thus be altered and lose their hypoxic sensitivity in case of COVID-19.^[26] Of note, the possibly altered hypoxic pulmonary vasoconstriction has been suggested to contribute to the so-called “silent hypoxemia” (defined as the presence of severe hypoxemia without marked symptoms of respiratory distress) observed in patients with acute respiratory failure due to COVID-19 pneumonia.[26], [27], [28]

In a physiological study including 10 patients and using electrical impedance tomography to assess ventilation and perfusion, Mauri et al.^[29] found that elevated ventilation-perfusion mismatch in COVID-19-associated ARDS was more related to an increased dead space (ventilated non-perfused lung units) than to an increased shunt (perfused non-ventilated units).

In addition, some authors suggested that an increased dead space fraction may characterize COVID-19-associated ARDS. The ventilatory ratio (VR), a simple bedside index of impaired efficiency of ventilation that is considered a good surrogate for dead space fraction has thus been found to be increased in some studies including patients with COVID-19 pneumonia.[30], [31], [32] VR has been described as the ratio of the product of measured minute ventilation and measured partial pressure of carbon dioxide (PaCO₂) to the product of predicted minute ventilation and ideal PaCO₂. In practice, it is actually calculated using the following formula: VR = minute ventilation (mL/min) × PaCO₂ (mmHg)/predicted body weight (kg) × 100 × 37.5.^[31] In a previously described study, VR was found to be lower at day 1 in patients with COVID-19-associated ARDS than in matched patients with non-COVID-19 pulmonary ARDS, but an increase over time was observed only in COVID-19 patients. Of note, additional analyses showed that VR was significantly impacted by the humidification device and the applied PEEP level. The use of heat and moisture exchangers (which may have been more frequent during the COVID-19 pandemic because of a shortage of heated humidifiers and initial concerns about viral transmission) is actually associated with an increased dead space fraction compared to the use of heated humidifiers.^[33] High PEEP levels may be associated with alveolar overdistension also leading to an increased dead space fraction. Changes in VR may therefore not only reflect pathophysiological alterations (as vascular injuries leading to an increase in dead space fraction) but also may be related to changes in ventilatory strategies (as changes in humidification devices or PEEP titration strategies).

These alterations including high dead space fraction and more generally increased ventilation to perfusion mismatch as described above may be related to significant vascular injuries reported on anatomopathological studies performed in patients who died from COVID-19 pneumonia. In a study comparing morphologic and molecular changes in lungs obtained during autopsy from 7 patients who died from COVID-19, 7 patients who died from ARDS secondary to influenza A (H1N1) infection and 10 age-matched, uninfected control subjects, the “classical” morphologic pattern including diffuse alveolar damage and infiltrating perivascular lymphocytes was found in COVID-19 is similar to influenza A ARDS lungs.^[7] However, distinctive vascular features involving severe endothelial injuries with the presence of intracellular viruses and disrupted cell membranes were found in the lungs of patients who died from COVID-19.^[7] Microangiopathy with alveolar capillary microthrombi was much more prevalent in COVID-19 than in influenza A lungs. New vessels were also more frequently found in COVID-19.^[7] More importantly, their development was predominantly related to a mechanism of intussusceptive angiogenesis.^[7]

Using radiologic computed tomography pulmonary angiography and dual-energy computed tomography, Patel et al.^[34] showed a markedly impaired pulmonary perfusion likely due to pulmonary angiopathy and thrombosis in 39 patients with COVID-19-associated ARDS. Furthermore, numerous data found a higher risk of thrombotic complications in patients with COVID-19-associated ARDS than in patients with other causes of ARDS.[35], [36], [37]

Thus, some of the very important data seem to suggest that patients with COVID-19-associated ARDS may exhibit specific vascular injuries that may be associated with an altered hypoxic pulmonary vasoconstriction (Table 1). These alterations may explain the severe hypoxemia observed in some patients with non-severely altered respiratory mechanics.

Natural Course of COVID-19 and Outcomes

The mean delay from the onset of the injury (or viral inoculation in COVID-19 pneumonia) and the onset of acute respiratory failure (or ARDS diagnosis) is significantly longer in COVID-19 than in other causes of ARDS (including other viruses such as influenza).^[38] It is also notable that COVID-19-associated ARDS is associated with longer dependence on mechanical ventilation than non-COVID-19 ARDS.[15], [39] This point is of major importance when considering ventilatory strategies or assessing the risk of ventilator-associated pneumonia.[40], [41]

When the patients are matched on SAPS II, hypoxemia, and age, no difference in overall mortality at day 28 was found between those with COVID-19-associated ARDS and those with pulmonary non-COVID-19-associated ARDS.^[15]

Specificities of Ventilatory Strategies for Patients with COVID-19 associated ARDS

The question of the specificities of COVID-19-associated ARDS pathophysiology naturally leads to the question of the potential optimization of ventilatory strategies in these patients. In other words, should ventilatory support strategies of patients with COVID-19 be in line with guidelines for ARDS unrelated to COVID-19 pneumonia?

Non-invasive ventilatory support in patients with COVID-19 acute hypoxemic respiratory failure

Before the COVID-19 pandemic, oxygenation using a high-flow nasal cannula (HFNC) has been shown to reduce mortality in patients with non-hypercapnic acute respiratory failure (i.e., in patients with a PaO₂/FiO₂ ratio <300 mmHg but without hydrostatic pulmonary edema or hypercapnia) compared to conventional oxygen therapy or non-invasive ventilation.^[42] HFNC was initially sparsely used in COVID-19 patients in numerous centers because of the absence of data reporting the efficacy of this technique in this new population of patients and because of the potential risk of environmental contamination and caregivers’ infections.

This risk for caregivers was not evidenced by later data.[43], [44] Some preliminary works then suggested that HFNC may reduce the intubation rate in patients with acute hypoxemic respiratory failure due to COVID-19 compared to standard low-flow oxygen therapy.^[45] Frat et al.^[46] showed, in a large randomized clinical trial including 711 patients with COVID-19, that HFNC did not significantly reduce the 28-day mortality compared with standard oxygen therapy (10% and 11%, respectively, P=0.60). The intubation rate was however significantly lower in patients treated with HFNC (45% vs. 53% respectively, P=0.04). Of note, the number of ventilator-free days at day 28 and the rate of ventilator-associated pneumonia were not different between the two groups. Another trial performed in three hospitals in Colombia and including 220 patients with COVID-19 pneumonia (PaO₂/FiO₂ ratio: <200 mmHg) found that, compared with conventional oxygen therapy, HFNC decreased the need for mechanical ventilation support and time to clinical recovery assessed at day 28 using a 7-category ordinal scale.^[47] Beyond the absence of clearly demonstrated individual benefits on mortality, HFNC may decrease the number of patients needing a ventilator and other ICU resources during major health crises. Interestingly, Ranieri et al.^[48] showed that the oxygenation criterion of ARDS based on the PaO₂/FiO₂ ratio can be applied to COVID-19 patients treated with HFNC to predict the presence of ARDS immediately after intubation. However, as the patient outcome is very different between the patients who are intubated after HFNC failure and those who do not need to be intubated, oxygenation in HFNC cannot be used alone as a predictive factor.

The use of continuous positive airway pressure (CPAP) has also been studied in patients with COVID-19 pneumonia. Data published before the COVID-19 pandemic showed that CPAP improves oxygenation and decreases the work of breathing in patients with acute respiratory failure.[49], [50] No benefit in terms of intubation or mortality reduction has however been evidenced so far with this non-invasive support except for hydrostatic pulmonary edema.^[49] In COVID-19-associated acute hypoxemic respiratory failure, a British parallel group, adaptive, randomized clinical trial including 1273 patients found that CPAP significantly reduced the risk of tracheal intubation or mortality at day 30 compared with conventional oxygen therapy.^[51] No difference in this composite criterion was found between HFNC and conventional oxygen therapy in this trial.^[51] In another large randomized controlled trial using a 2 × 3 factorial design to assess the benefit of high-dose dexamethasone compared with standard-of-care dexamethasone and to assess the benefit of HFNC or CPAP compared with conventional oxygen therapy in patients with COVID-19 pneumonia, HFNC or CPAP did not reduce the risk of intubation at day 28.^[52] Of note, the mortality observed in this study was significantly higher than in Frat's study (>20% vs. around 10% at day 28) but was close to the one observed in Perkins’ study (mortality from 17% to 20%).[46], [51], [52]

The use of non-invasive ventilation has also been proposed in COVID-19-associated acute respiratory failure. Grieco et al.^[53] found in a randomized trial including 110 patients that the use of helmet non-invasive ventilation did not reduce the number of days free of respiratory support within 28 days but reduced the risk of intubation compared with HFNC. In a large randomized clinical trial including 322 patients with COVID-19-associated acute hypoxemic respiratory failure in Saudi Arabia and Kuwait, helmet non-invasive ventilation did not significantly reduce mortality at day 28 compared with usual respiratory support including mask non-invasive ventilation, HFNC, or conventional oxygen therapy (27% vs. 26.1%, P=0.85).^[54] Furthermore, no difference was found between the two groups in terms of risk of intubation or barotrauma in this study.

The question of non-invasive support of oxygenation in acute hypoxemic respiratory failure is closely related to the question of optimal timing for intubation in these patients. There was an intensive debate on this matter in patients with COVID-19 pneumonia.[55], [56], [57], [58] As mentioned above, some authors reported a high prevalence of “silent hypoxemia” in patients with COVID-19-associated hypoxemic respiratory failure.^[28] In addition, some data suggested that patients with COVID-19 treated with non-invasive support may present with high tidal volume and transpulmonary pressure and thus develop potential patient self-induced lung injury (P-SILI).^[59] The concept of P-SILI, was proposed by Brochard et al.^[60] in 2017, defines effort-dependent lung injury by analogy with the concept of ventilator-induced lung injury (VILI) due to positive pressure ventilation.^[61] Other authors argued that hypoxemia in the absence of increased work of breathing should not be per se considered as an indication for invasive mechanical ventilation when oxygenation support allows the oxygenation target to be achieved without needing to apply PEEP.^[57] For these authors, intubation for the acute respiratory failure of patients who do not present with severe symptoms of respiratory distress may expose them to an unjustified risk of complications related to invasive mechanical ventilation.

Ventilator settings in intubated patients with COVID-19-associated ARDS

The small amplitude of the potential differences in respiratory system compliances between COVID-19 and non-COVID-19-associated ARDS does not support the use of different tidal volumes. A tidal volume of 6 mL/kg predicted body weight should thus be considered a safe target for COVID-19 patients as for patients with “classical forms” of ARDS.^[62] A more individualized approach to the tidal volume setting may repose on the respiratory system driving pressure measurement. The respiratory system driving pressure or distending pressure is a very important concept representing how much pressure is transmitted to the respiratory system during tidal ventilation.^[63] It is very easily calculated at the bedside as the difference between the plateau pressure and the final end-expiratory pressure. As the compliance of the respiratory system is equal to the tidal volume divided by the driving pressure, it is equal to the tidal volume divided by the respiratory system compliance. Therefore, using the driving pressure to adjust the tidal volume is a simple way to titrate tidal volume for respiratory system compliance (i.e., to normalize the tidal volume to the functional lung size and not to the predicted lung size in healthy persons). In a study including 51 patients with moderate-to-severe ARDS (37 related to COVID-19 pneumonia), a strategy targeting a driving pressure between 12 cm H₂O and 14 cm H₂O was required to change the tidal volume in 90% of the patients as compared to a conventional strategy targeting a tidal volume of 6 mL/kg predicted body weight.^[64] Such driving pressure-guided strategy targeting was associated with a significant decrease in mechanical power. Mechanical power corresponds to the amount of energy generated by mechanical ventilation and released on the respiratory system per unit of time and can be considered a determinant of VILI.^[65]

The optimal PEEP titration strategy is still a matter of debate in patients with the “classical forms” of ARDS. Meta-analyses of large randomized clinical trials suggested that high PEEP levels were associated with improved survival in patients with what is now called “moderate-to-severe ARDS”.[66], [67] From a physiological point of view, high levels of PEEP should be applied only in patients with a high potential for lung recruitment. Thus, the formal guidelines of the French Intensive Care Society recommend in non-COVID-19-associated ARDS that “High PEEP should probably be used in patients with moderate or severe ARDS, but not in patients with mild ARDS” but should be “reserved for patients in whom it improves oxygenation without marked deterioration of respiratory system compliance or hemodynamic status” and mentioned that “PEEP settings should be individualized”.^[68] Numerous studies seeking to assess the optimal PEEP titration strategy in patients with COVID-19-associated ARDS were published.[69], [70], [71] However, there is no evidence that PEEP should be titrated differently in COVID-19 and non-COVID-19-associated ARDS. The previously described data reporting various and evolutive recruitability in patients with COVID-19-associated ARDS actually suggest that an individual assessment of the potential for lung recruitment in response to a high level of PEEP may be needed to optimize PEEP settings in these patients as in patients with “classical forms” of ARDS.[14], [24], [25]

Prone positioning in intubated or non-intubated patients with COVID-19 pneumonia

Prone positioning has been shown to reduce the mortality of patients with a “classical form” of ARDS and PaO₂/FiO₂ ratio <150 mmHg.^[72] This treatment must thus be considered as a key element in the ventilatory strategies of all the patients meeting these criteria, including those for whom proning is not associated with an increase in oxygenation.^[73] No evidence has been found until now to propose different recommendations for patients with COVID-19-associated ARDS. In a study describing the lung protective effects of prone position in patients with COVID-19-associated ARDS, some authors showed that prone positioning was associated with extensive alveolar recruitment in the dorsal regions despite relatively preserved lung aeration in the supine position and that dorsal recruitment reduced the risk of regional atelectrauma compared to the supine position.^[74] In addition, this study found that ventral lung regions were characterized by a decreased fraction of ventilated non-perfused units and reduced dead space/shunt ratio after pronation.

Prone positioning has also been proposed in non-intubated patients. In the context of COVID-19 patients treated with HFNC, proning was associated with an increase in oxygenation and may reduce the need for intubation without a significant impact on the mortality rate.[75], [76], [77], [78] In a post hoc analysis of data collected for a randomized controlled trial including 125 patients with COVID-19-associated acute hypoxemic respiratory failure, Kaur et al.^[79] showed that the patients who received early awake prone positioning (i.e., within 24 h after the beginning of treatment with HFNC) had lower mortality than patients who received prone positioning later (i.e., after more than 24 h of HFNC).

Inhaled nitric oxide in intubated patients with COVID-19-associated ARDS

Pathophysiological hypotheses concerning vascular injuries in COVID-19 led researchers to specifically assess the effect of inhaled nitric oxide in these patients. As in classical forms of ARDS, improved oxygenation was observed with this treatment but no impact on the mortality rate was found.[80], [81], [82] In addition, a physiological study including 10 patients with COVID-19-associated ARDS found that the association of inhaled nitric oxide and almitrine was associated with better oxygenation than inhaled nitric oxide alone.^[83]

Steroids

No specific pharmacological treatment has been proven to benefit patients with ARDS. In particular, although numerous molecules and dosing regimens have been proposed, the efficacy of corticosteroids remains controversial in classical forms of ARDS.[84], [85], [86] In addition, some data suggest that their use may be harmful in patients with influenza-associated ARDS.[87], [88] On the contrary, dexamethasone has been shown to improve the outcomes of patients with COVID-19. Although its use may be associated with an increased risk of ventilator-associated pneumonia, dexamethasone was associated with an increase in the number of ventilator-free days over 28 days in patients with moderate-to-severe COVID-19-associated ARDS and with a decrease in 28-day mortality in those receiving invasive mechanical ventilation or oxygen alone at randomization.[89], [90], [91], [92]

Extracorporeal membrane oxygenation (ECMO) in patients with COVID-19-associated ARDS

The Extracorporeal Life Support Organization (ELSO) guidelines recommend the use of the same criteria to consider ECMO in patients with COVID-19-associated ARDS and in those with “classical forms” of ARDS.^[93] Mortality of non-COVID-19 patients treated with venovenous ECMO has been shown to increase with patients’ age and with prolonged exposure to mechanical ventilation before ECMO.^[94] Strict age criteria and a maximal delay between the beginning of mechanical ventilation and the cannulation of 7 days or 10 days were thus usually recommended in patients with COVID-19-associated ARDS. More stringent contraindications were often applied when resources became more constrained during the pandemic.

Encouraging results were reported in selected patients with COVID-19-associated ARDS with mortality rates similar to those observed in non-COVID-19 patients.[93], [95] Of note, interesting results were also found in a less selective population of COVID-19 patients not using strict criteria such as limitation of age or duration of exposure to mechanical ventilation to consider ECMO (hospital mortality: 48%).^[96] In this cohort, a late ECMO initiation was not associated with increased mortality.

Conclusions

Most available data suggest that the pathophysiology of COVID-19-associated ARDS may be considered substantially similar to the pathophysiology of ARDS not related to COVID-19. Some important research studies have however shown frequent and specific vascular injuries that may be associated with an altered hypoxic pulmonary vasoconstriction in patients with COVID-19. Published studies actually do not support the use of specific ventilatory strategies in these patients. A strategy based on current evidence concerning “classical forms” of ARDS and individualized according to information provided by bedside physiology should thus be recommended in COVID-19 as in non-COVID-19 patients.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of Interest

The author declares that he has no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Managing Editor: Jingling Bao