Physiological and quantitative CT-scan characterization of COVID-19 and typical ARDS: a matched cohort study



To investigate whether COVID-19-ARDS differs from all-cause ARDS.


Thirty-two consecutive, mechanically ventilated COVID-19-ARDS patients were compared to two historical ARDS sub-populations 1:1 matched for PaO2/FiO2 or for compliance of the respiratory system. Gas exchange, hemodynamics and respiratory mechanics were recorded at 5 and 15 cmH2O PEEP. CT scan variables were measured at 5 cmH2O PEEP.


Anthropometric characteristics were similar in COVID-19-ARDS, PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS. The PaO2/FiO2-matched-ARDS and COVID-19-ARDS populations (both with PaO2/FiO2 106 ± 59 mmHg) had different respiratory system compliances (Crs) (39 ± 11 vs 49.9 ± 15.4 ml/cmH2O, p = 0.03). The Compliance-matched-ARDS and COVID-19-ARDS had similar Crs (50.1 ± 15.7 and 49.9 ± 15.4 ml/cmH2O, respectively) but significantly lower PaO2/FiO2 for the same Crs (160 ± 62 vs 106.5 ± 59.6 mmHg, p < 0.001). The three populations had similar lung weights but COVID-19-ARDS had significantly higher lung gas volume (PaO2/FiO2-matched-ARDS 930 ± 644 ml, COVID-19-ARDS 1670 ± 791 ml and Compliance-matched-ARDS 1301 ± 627 ml, p < 0.05). The venous admixture was significantly related to the non-aerated tissue in PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS (p < 0.001) but unrelated in COVID-19-ARDS (p = 0.75), suggesting that hypoxemia was not only due to the extent of non-aerated tissue. Increasing PEEP from 5 to 15 cmH2O improved oxygenation in all groups. However, while lung mechanics and dead space improved in PaO2/FiO2-matched-ARDS, suggesting recruitment as primary mechanism, they remained unmodified or worsened in COVID-19-ARDS and Compliance-matched-ARDS, suggesting lower recruitment potential and/or blood flow redistribution.


COVID-19-ARDS is a subset of ARDS characterized overall by higher compliance and lung gas volume for a given PaO2/FiO2, at least when considered within the timeframe of our study.


Acute respiratory distress syndrome (ARDS)—as currently defined—is a syndrome which broadly includes diverse conditions grouped on the basis of an oxygenation deficit of acute onset and bilateral radiographic infiltrates that cannot be attributed solely to a cardiovascular cause [1]. The severity of ARDS is classified by a single criterion only: oxygenation deficit, expressed as PaO2/FiO2 ratio. By this broad definition, hypoxemic patients with coronavirus disease 2019 (COVID-19) and bilateral chest X-ray infiltrates clearly satisfy the definition of ARDS. We have reported [2, 3] that severe hypoxemia with relatively well-preserved respiratory system compliance (Crs) measured under standard conditions is characteristic of COVID-19-ARDS, and it differs from ARDS of other causes (typical ARDS), while others did not recognize consistent differences [4,5,6,7]. However, the increased frequency of higher Crs in COVID-19-ARDS was noted by some of the same authors [4, 5], and a significantly higher Crs associated with severe hypoxemia was recently documented [8]. The heterogeneous nature of ARDS allows that—at the population level—there may be wide overlap between COVID-19-ARDS and typical ARDS, and these observations continue to drive a debate [6, 9]. It is worth remembering that during the Berlin conference, the experts’ panel initially agreed on using an upper threshold of 40 ml/cmH2O of respiratory system compliance (Crs) to qualify as severe ARDS. This variable was not implemented, however, as it did not add further prognostic value to bilateral infiltrates and PaO2/FiO2 ratio [10]. In other words, the PaO2/FiO2 ratio and the Crs deteriorated together in typical ARDS. This pairing may not occur in COVID-19-ARDS, where a discrepancy between the severities of hypoxemia and respiratory mechanics may be the key issue, rather than their individual absolute values per se.

Indeed, this discrepancy could be due to the underlying pathogenesis of COVID-19-ARDS, which is highly atypical and quite distinct from most other forms of typical ARDS that are routinely encountered [2, 11, 12]. In typical ARDS, the primary site ‘hit’ is the alveolar space, particularly in pulmonary ARDS. In contrast, in COVID-19-ARDS, the prevalent pathophysiological mechanism is initiated on the vascular side of the pulmonary unit. In addition, the endothelialitis, typical of COVID-19 patients, results in a powerful activation of the coagulation cascade, with micro and macro thrombosis occurring in pulmonary tissues and throughout the body [13,14,15]. Undoubtedly, microthromboses are recognized autopsy findings [16], and pulmonary artery filling defects (vascular occlusion or compression) have been described in typical ARDS for decades [17, 18]. A striking difference between typical ARDS and COVID-19-ARDS, however, is the remarkable frequency and extent to which pulmonary [15, 19] and extrapulmonary thrombosis [20] occur in the latter. While regional atelectasis, edema and fibrosis may coexist, disrupted vasoregulation strikingly alters the matching of perfusion to ventilation, a pathophysiologic mechanism which may be the predominant contributor to hypoxemia in the early phase of this evolving disease [12, 21].

To investigate whether and to what degree, COVID-19-ARDS differs from typical ARDS, we compared the physio-anatomical characteristics of COVID-19-ARDS patients with two historically matched cohorts of typical ARDS. Quantitative CT scan analysis, and measurements of respiratory system mechanics and gas exchange were performed under standardized and identical conditions, both in COVID-19-ARDS and typical ARDS, thus avoiding the biases of acquiring CT scans and physiological variables under highly heterogeneous “clinical” conditions.


Study population

Thirty-two COVID-19-ARDS patients, consecutively admitted to the ICU of ASST Santi Paolo e Carlo Hospital, Milan over the period between February 21st, 2020 and May 7th, 2020 were prospectively enrolled. This study was conducted in accordance with the pre-existing Ethics Committee approval that allows physiological and CT scan studies for all patients with severe respiratory failure admitted to our critical care unit (ethics committee numbers: 42937/2016 and 9890/2017). All had documented COVID-19 positive RT-PCR o nasal or pharyngeal swab and bilateral infiltrates documented by chest X-ray. This COVID-19-ARDS population was matched with cohorts from two separate non-COVID ARDS populations: one matched 1:1 for PaO2/FiO2 (PaO2/FiO2-matched-ARDS) and a second one matched 1:1 for respiratory system compliance (Compliance-matched-ARDS). The values of PaO2/FiO2 ratio and Crs used for matching these cohorts were the ones measured in COVID-19-ARDS at 5 cmH2O of PEEP during mechanical ventilation, immediately before the CT scan. The PaO2/FiO2 ratio and Crs of the two historical non-COVID ARDS cohorts were measured under exactly the same conditions.

The two matched, entirely independent population samples were extracted from our ARDS dataset which includes 232 patients studied between 2003 and 2018. These patients had previously been screened and included in clinical physiopathologic studies performed by our group over the same time span. Therefore, they met all criteria that define ARDS and underwent a common and standardized intervention (e.g., CT scan, PEEP trials, measurement of respiratory mechanics and gas exchange) and a standardized data collection protocol.


In every studied patient (of both COVID-19-ARDS and the matched non-COVID ARDS populations), gas exchange, respiratory mechanics, hemodynamics and CT scan variables were recorded under standardized conditions (Volume Controlled ventilation, tidal volume 7–8 ml/kg of Ideal Body Weight (IBW), muscle relaxation, 5 cmH2O of Positive End-Expiratory Pressure, PEEP). Both COVID-19-ARDS and non-COVID ARDS population cohorts underwent CT scanning and PEEP testing within a median of 3 [IQR 1—4] days after the admission to ICU.

Gas exchange

We measured FiO2, PO2, PCO2, hemoglobin saturation and derived variables (using arterial and central venous blood) and end-tidal PCO2 (PETCO2). Venous admixture was computed using central venous blood values as surrogates for the mixed venous ones [22].

Respiratory system mechanics

We measured plateau pressure, PEEP, driving pressure and respiratory system compliance at the standardized value of 5 cmH2O.

CT-quantitative anatomical variables

In each patient, the whole lung CT was performed under static conditions during an end-expiratory hold at 5 cmH2O of PEEP. Lung profiles of each CT scan slice were manually contoured, excluding hilar structures. Then, quantitative analysis was performed with dedicated software (Maluna [23]). We estimated lung weight, gas volume, amount of over-inflated tissue (voxel density − 1000 to − 900 Hounsfield Units, HU), well-aerated tissue (− 899 to − 500 HU), poorly aerated tissue (− 499 to − 100 HU) and non-aerated tissue (− 100 to + 100 HU). Analyses were performed on each whole slice as well as on ten equally spaced segments along the sterno-vertebral axis.

PEEP response

All patients underwent a “PEEP-test” in which PEEP was raised from 5 to 15 cmH2O while keeping constant respiratory rate, tidal volume and FiO2. Gas exchange, hemodynamics and mechanical variables were re-measured at 15 cmH2O of PEEP after a 15-min equilibration period (See Supplement for details.)

Statistical analysis

The one-to-one matching procedure was performed with the nearest-neighboring method using the optimal algorithm, without replacement, with the MatchIt package for R (R Foundation for Statistical Computing version 4.0.2) [24]. Data are presented as mean ± standard deviation. Student’s t test assessed the statistical significance of the difference between group means when data were distributed normally; otherwise, the Wilcoxon test was used. Chi square test or Fisher’s exact test was used to construct the contingency tables. Linear regression tested the relationship between continuous variables. Two-way analysis of variance allowing interaction was used to evaluate the gas volume distribution along the segments of sterno-vertebral axis. These statistical analyses were performed with R (R Foundation for Statistical Computing version 4.0.2) and its package Tidyverse.


Study population

COVID-19-ARDS (32 consecutive patients) and both non-COVID ARDS population cohorts (PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS, 32 patients each) had similar baseline characteristics regarding age, sex, Ideal Body Weight and Body Mass Index. The Simplified Acute Physiology Score II (SAPSII), although lower in COVID-19-ARDS patients, was not statistically different from the comparison cohorts. However, it is likely that overall clinical severity in the PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS groups was greater, as indicated by longer ICU length of stay (Table 1). PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS had comparable distributions of prevalence regarding etiology of lung injury (p = 0.86). The majority (68.7%) of both COVID-19-ARDS and PaO2/FiO2-matched-ARDS patients had PaO2/FiO2 ratios consistent with severe ARDS, based on the Berlin definition of ARDS severity. In contrast, severe ARDS represented only 18.7% of Compliance-matched-ARDS patients; (p < 0.001) (Table 1). Outcome measures for the three populations are reported in Table 1.

Table 1 Baseline clinical characteristics of the three cohorts

Oxygenation and respiratory mechanics

When COVID-19-ARDS was compared to the PaO2/FiO2-matched-ARDS cohort, i.e., at similar oxygenation, its respiratory system compliance was significantly higher (49.9 ± 15.4 vs 39.9 ± 11.1 ml/cmH2O; p = 0.003, Fig. 1a) and plateau and driving pressures were significantly lower (Table 2). When COVID-19-ARDS was compared to the Compliance-matched-ARDS population, i.e., at similar respiratory system mechanics, the PaO2/FiO2 ratio was significantly lower in COVID-19-ARDS (106.5 ± 59 vs 160 ± 62 mmHg; p < 0.001, Fig. 1b), as were the other oxygenation variables. In PaO2/FiO2-matched-ARDS patients, the PaO2/FiO2 ratio was linearly related with the respiratory system mechanics (p = 0.036), whereas no significant correlation was found neither in Compliance-matched-ARDS (p = 0.9), nor in COVID-19-ARDS (p = 0.81, Figure E1).

Fig. 1

a Respiratory system compliance in COVID-19-ARDS (orange) and in PaO2/FiO2-matched-ARDS populations (blue). The matched PaO2/FiO2 ratios were similar (COVID-19-ARDS = 106 ± 59 mmHg, PaO2/FiO2-matched-ARDS = 106 ± 60 mmHg). Note that, for the same PaO2/FiO2 ratio, the Crs in COVID-19-ARDS is significantly higher (~ 11 ml/cmH2O) than in PaO2/FiO2-matched-ARDS (median values 49.5 and 38.4 ml/cmH2O, respectively). b PaO2/FiO2 ratio in COVID-19-ARDS (orange) and in Compliance-matched-ARDS populations (blue). The matched respiratory system compliance values were similar (COVID-19-ARDS = 49.9 ± 15.4 ml/cmH2O, Compliance-matched-ARDS  = 50.1 ± 15.7 ml/cmH2O). Note that, for the same Crs, the PaO2/FiO2 ratio in Compliance-matched-ARDS is significantly higher (~ 70 mmHg) than in COVID-19-ARDS (median values 155.0 and 85.4 mmHg, respectively)

Table 2 Gas exchange, respiratory mechanics, hemodynamics and CT variables of the three cohorts

CO 2 clearance and dead space

With regard to the ventilation parameters, COVID-19-ARDS and its two matched populations (PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS) had comparable values for tidal volume, alveolar dead space ventilation and ventilatory ratio. Minute ventilation was significantly higher and PaCO2 lower in COVID-19-ARDS, compared to PaO2/FiO2-matched-ARDS, due to a higher respiratory rate (Table 2).

CT scan variables

Despite similar total lung weights, patients with COVID-19-ARDS, compared to PaO2/FiO2-matched-ARDS, had significantly higher lung gas volume (1670 ± 791 vs 930 ± 644 mL; p < 0.001), a greater amount of normally aerated tissue (475 ± 185 vs 287 ± 154 g; p < 0.001) and less non-aerated tissue (591 ± 293 vs 960 ± 567 g; p = 0.002) (Table 2). The weights of the normally aerated and non-aerated lung tissues were similar in COVID-19-ARDS and in Compliance-matched-ARDS populations, but total lung gas volume was higher in COVID-19-ARDS compared to Compliance-matched-ARDS. Notably, the distribution of gas volume was remarkably different for the three populations: patients with COVID-19-ARDS had the highest gas volumes in each lung segment, whereas the lowest gas volumes were measured in the corresponding segments of the PaO2/FiO2-matched ARDS population (Fig. 2).

Fig. 2

Lung gas volume measured in the 10 equally spaced lung segments along the sterno-vertebral axis (level 1 = closest to the sternum, level 10 = closest to the vertebra). The gas volume of both the PaO2/FiO2-matched-ARDS (dark blue) and Compliance-matched-ARDS (light blue) was significantly different from COVID-19-ARDS (p < 0.001 and p = 0.043, respectively). Note that the gas volume was higher in COVID-19-ARDS, even compared to the Compliance-matched-ARDS. The extent of the differences is particularly evident in the most dependent lung regions, where the gas volume at each level was even more than double in COVID-19-ARDS than in PaO2/FiO2-matched-ARDS

Hemodynamics, venous admixture and non-aerated tissue

Hemodynamic values are presented in Table 2. In Fig. 3, we present venous admixture as a function of the fraction of non-aerated tissue. In each of the non-COVID ARDS cohorts, venous admixture increased with the fraction of non-aerated tissue (both p = 0.004). In contrast, in COVID-19-ARDS, the venous admixture remained approximately constant and independent from large variations in the observed fraction of non-aerated tissue (p = 0.75). The regression model also shows that in the COVID-19-ARDS population the constant term of the model equates to a venous admixture of 0.50 (95% CI 0.33–0.67), suggesting that significant venous admixture is theoretically present, even for an assumed zero fraction of non-aerated tissue. PaO2/FiO2 ratio and A-aPO2 as a function of the fraction of non-aerated tissue (Figures E2 and E3) showed the same behavior of venous admixture.

Fig. 3

Venous admixture as a function of the fraction of non-aerated tissue, in PaO2/FiO2-matched-ARDS (PF-ARDS, left panel), COVID-19-ARDS (CARDS, middle panel) and Compliance-matched-ARDS (Crs-ARDS, right panel). As shown, in PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS, the venous admixture increases proportionally with similar slopes (0.83 and 0.89, respectively) with the increase fraction of non-aerated tissue, implying a coupling between the shunt fraction and the fraction of non-aerated tissue. In contrast, in COVID-19-ARDS, the two variables were uncoupled. The relationships followed the regression equations: PaO2/FiO2-matched-ARDS, venous admixture = 0.83 × fraction of non-aerated tissue + 0.14, p = 0.003, R2 = 0.32 (22 observations). CARDS, venous admixture = − 0.07 × fraction of non-aerated tissue + 0.5, p = 0.75, R2 = − 0.03 (29 observations). Crs-ARDS, venous admixture = 0.89 × fraction of non-aerated tissue + 0.13, p = 0.004, R2 = 0.35 (19 observations). Missing data were due to the lack of central venous blood samples

Response to PEEP test

The responses of the physiological variables to the PEEP test, i.e., increasing PEEP from 5 to 15 cmH2O are summarized in Table 3. As shown, despite a similar increase in oxygenation in all three populations, the respiratory system mechanics and dead space all improved in the PaO2/FiO2-matched-ARDS cohort but did not change or deteriorated in patients with COVID-19-ARDS and those with Compliance-matched-ARDS.

Table 3 Gas exchange, respiratory mechanics and hemodynamic response to the PEEP increase (5–15 cmH2O)


In this study, which compares COVID-19-ARDS patients with two different non-COVID-19 ARDS populations, we found the following: (1) COVID-19-ARDS patients, compared to PaO2/FiO2-matched ARDS (i.e., similar oxygenation), had consistently better respiratory system compliance and nearly double the end-expiratory gas volume as their counterparts in the comparison groups; (2) COVID-19-ARDS patients, compared to a separate population of non-COVID-19 ARDS patients matched on Crs (i.e., with similar respiratory system mechanics) had consistently worse oxygenation variables; (3) COVID-19-ARDS, PaO2/FiO2-matched-ARDS, and Compliance-matched-ARDS experienced similar oxygenation improvement when raising PEEP from 5 to 15 cmH2O. Importantly, however, while that oxygenation improvement in the PaO2/FiO2-matched-ARDS population was associated with significantly improved CO2 clearance and respiratory mechanics, these variables did not change or deteriorated in both COVID-19-ARDS patients and Compliance-matched-ARDS patients.

Patient populations

A single matching variable was used for each matching procedure. No other variables were included, due to the limited size of our ARDS dataset. The anthropometric characteristics of the three populations were not statistically different. Bilateral pneumonia was the only cause of lung injury in COVID-19-ARDS patients. By comparison, pneumonia accounted for 53.1% and 43.8% in PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS cohorts, respectively, incidence frequencies similar to the LUNG-SAFE study (59.4%) of 3022 patients (p = 0.16) [25]. The prevalence of sepsis was also similar among LUNG-SAFE, PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS (16%, 18.7% and 25% in, respectively; p = 0.36). Therefore, our sample of matched ARDS patients appears representative of the ARDS populations enrolled in pre-COVID ARDS clinical trials. The distribution of mild, moderate and severe ARDS (as measured at 5 cmH2O of PEEP [26]) in COVID-19-ARDS and in our PaO2/FiO2-matched-ARDS subgroup was identical (Table 2) [1]. In contrast, the overall severity of Compliance-matched-ARDS patients was lower, as the prevalence of severe ARDS category was only 18.8% vs 68.7% in COVID-19-ARDS. The general clinical severity, as indicated by SAPSII, tended to be lower in COVID-19-ARDS patients, compared to the two non-Covid ARDS populations, perhaps accounting for their shorter length of stay in the ICU.

Oxygenation, lung mechanics and the mechanism of hypoxemia

Differently from typical ARDS, where the decrease of PaO2/FiO2 ratio is associated with a decrease in Crs, in our COVID-19-ARDS population PaO2/FiO2 ratio and Crs were unrelated. This has also been found in a recent larger study comparing typical ARDS with COVID-19-ARDS [8]. This contrasts with the decision taken in Berlin to exclude Crs from the ARDS definition as unnecessary, as it added no prognostic value to the PaO2/FiO2 ratio alone [10]. It is then possible that the mechanisms leading to hypoxemia are somehow different between COVID-19-ARDS and typical ARDS. Hypoxemia due to venous admixture [27] originates from two potential mechanisms: true right to left shunt (i.e., perfusion of non-aerated tissue) and/or low ventilation–perfusion (VA/Q) ratio (perfusion of poorly ventilated lung regions). In typical ARDS, the primary component of venous admixture is right-to-left shunt. Accordingly, the greater the fraction of non-aerated tissue, the greater the venous admixture [28, 29]. In COVID-19-ARDS the venous admixture was unrelated to the non-aerated tissue fraction; indeed, it was very high even when the fraction of non-aerated tissue was very low (Fig. 3). This observation strongly suggests that the major component of the venous admixture in COVID-19-ARDS is ventilation–perfusion mismatch, rather than true right-to-left shunt. The important role of VA/Q mismatch in COVID-19-ARDS is consistent with (but not entirely explained by) the reported high incidence of micro and macro thrombosis in this disease [11, 14, 30, 31] and with the importance of markers of immune-thrombosis (e.g., d-dimers) in the outcome of COVID-19-ARDS [8].

Respiratory system compliance and lung gas volume

The relative importance of VA/Q mismatching as opposed to right-to-left shunt in COVID-19-ARDS is consistent with its relatively higher lung gas volume, which correlates with the respiratory system compliance (see Figure E4). Moreover, the gas volume was remarkably higher in COVID-19-ARDS compared to PaO2/FiO2-matched-ARDS for each lung section along the gravitational axis, including the most dependent ones, which are almost gasless in typical ARDS (Fig. 2). Again, this difference, likely due to the vasocentric nature of COVID-19-ARDS (as compared to ‘gas space-centered’ nature of typical ARDS), is not entirely surprising. Unexpectedly, however, we found that Compliance-matched-ARDS patients had lower gas volume than did COVID-19-ARDS patients, despite having similar values of respiratory system mechanics. The interpretation of these findings is currently only speculative. However, it is tempting to hypothesize that the increased gas volume in COVID-19-ARDS is caused by newly formed emphysema-like functional regions that may develop as a consequence of the ischemic changes and diffuse micro thromboses described in autopsy findings [11, 14].

Response to PEEP

The improved oxygenation in our PaO2/FiO2-matched-ARDS patient cohort in response to the PEEP test was likely due recruitment. Indeed, these patients were more recruitable, as indicated by higher baseline non-aerated tissue mass in conjunction with a significant decrease of plateau pressure and an improvement of Crs when PEEP was raised. In contrast, Compliance-matched-ARDS and COVID-19-ARDS patients, with lower baseline non-aerated tissue mass, showed unaltered or worsened respiratory system mechanics and PaCO2 in response to the PEEP test (Table 3). These findings suggest—in line with previous observations—[32, 33] that the primary mechanism of oxygenation improvement was a decrease/redistribution of blood flow away from airless zones rather than recruitment.

Atypical features of COVID-19-ARDS

Our data suggest that COVID-19-ARDS is an atypical subset of ARDS. We may then wonder why, for a given severity of hypoxemia, the Crs values of our COVID-19-ARDS patients appear higher than those reported by other authors [6, 34]. As the virus is the same worldwide, its manifestations everywhere should be more or less consistent. The differences observed among various reports may depend on two main factors: the timing of the observations and the conditions of measurement. Indeed, COVID-19-ARDS evolves rather rapidly with time, as reflected by a CT scan appearance that shifts progressively from bilateral ground glass opacities to overt consolidations/collapse [35]. Crs changes accordingly [36]. It is not surprising that, with passing time, Crs may decrease to impressively low values. The conditions of measurement are also important. Most studies, such as the largest one yet published on COVID-19 pathophysiology [8], report Crs values measured under the prevailing “clinical conditions”. In all cohorts of our COVID-19-ARDS, PaO2/FiO2-matched-ARDS and Compliance-matched-ARDS populations, all measurements were performed at a standard PEEP of 5 cmH2O. It is obvious that Crs measured at 10–15 cmH2O of PEEP may lead to different values than those we report here.

Clinical implications

Our COVID-19-ARDS patients were studied 9.6 ± 4 days after the onset of symptoms and were compared with “early” historical ARDS patients (within 1 week from admission). Within this initial timeframe, the sharp physio-anatomic distinctions between non-COVID-ARDS and COVID-19-ARDS suggest the need to modify our standard practice of ARDS management for COVID-19 patients. Specifically, the dramatically greater gas volume and better compliance of COVID-19 lungs, when present, discourage interventions intended to further inflate the lungs. Indeed, for a similar marginal improvement of oxygenation in response to a PEEP increment, signs of overdistension became manifest in our COVID-19 patients. In contrast, respiratory mechanics improved and PaCO2 decreased in the PaO2/FiO2-matched-ARDS cohort. Attempts to aggressively recruit the lung to improve O2 exchange by applying higher than customary levels of mean airway pressure seem ill-advised during this early disease phase. We must stress, however, that COVID-19 pneumonia rapidly evolves with time. Consequently, the safest ventilatory strategy could well be different at different stages which range from initial modest ground-glass opacities with preserved Crs to an intermediate stage (as described in the present study), to a final stage characterized by extensive opacities, prevalent fibrosis, and very low Crs.


These data are unique in documenting physiologic measurements and quantitative images under identical conditions in closely matched COVID-19-ARDS and non-COVID-ARDS patients. However, our study has several limitations: first, the limited size of our historical ARDS dataset. Second, patients were enrolled in a single center, within a limited time frame of their illnesses. Earlier or later stages may present sharply different behaviors. In addition, we did not perform a second CT scan at 15 cmH2O. Finally, a comprehensive set of hemodynamic data were not acquired, preventing full characterization of the mechanisms underlying the gas exchange variations we observed.


COVID-19-ARDS and non-COVID ARDS patients differ significantly in their radiological and physiological features, both in terms of the relationship between oxygenation and lung mechanics and their responses to PEEP. The different stages of the disease call for a rethinking of the traditional lung protective ventilation targets which take into account the peculiarities of this novel ARDS variant.


  1. 1.

    Ranieri VM, Rubenfeld GD, Thompson BT, Ferguson ND et al (2012) Acute respiratory distress syndrome: the Berlin Definition. JAMA 307(23):2526–2533

    Google Scholar 

  2. 2.

    Gattinoni L, Coppola S, Cressoni M, Busana M et al (2020) COVID-19 does not lead to a “Typical” acute respiratory distress syndrome. Am J Respir Crit Care Med 201(10):1299–1300

    CAS  Article  Google Scholar 

  3. 3.

    Gattinoni L, Chiumello D, Caironi P, Busana M et al (2020) COVID-19 pneumonia: different respiratory treatments for different phenotypes? Intensive Care Med 46(6):1099–1102

    CAS  Article  Google Scholar 

  4. 4.

    Schenck EJ, Hoffman K, Goyal P, Choi J et al (2020) Respiratory mechanics and gas exchange in COVID-19-associated respiratory failure. Ann Am Thorac Soc 17(9):1158–1161

    Article  Google Scholar 

  5. 5.

    Ferrando C, Suarez-Sipmann F, Mellado-Artigas R, Hernandez M et al (2020) Clinical features, ventilatory management, and outcome of ARDS caused by COVID-19 are similar to other causes of ARDS. Intensive Care Med.

    Article  PubMed  PubMed Central  Google Scholar 

  6. 6.

    Ziehr DR, Alladina J, Petri CR, Maley JH, Moskowitz A, Medoff BD, Hibbert KA, Thompson BT, Hardin CC (2020) Respiratory pathophysiology of mechanically ventilated patients with COVID-19: a cohort study. Am J Respir Crit Care Med 201(12):1560–1564

    CAS  Article  Google Scholar 

  7. 7.

    Cummings MJ, Baldwin MR, Abrams D, Jacobson SD et al (2020) Epidemiology, clinical course, and outcomes of critically ill adults with COVID-19 in New York City: a prospective cohort study. Lancet 395(10239):1763–1770

    CAS  Article  Google Scholar 

  8. 8.

    Grasselli G, Tonetti T, Protti A, Langer T et al (2020) Pathophysiology of COVID-19-associated acute respiratory distress syndrome: a multicentre prospective observational study. Lancet Respir Med.

    Article  PubMed  Google Scholar 

  9. 9.

    Fan E, Beitler JR, Brochard L, Calfee CS et al (2020) COVID-19-associated acute respiratory distress syndrome: is a different approach to management warranted? Lancet Respir Med 8(8):816–821

    CAS  Article  Google Scholar 

  10. 10.

    Ferguson ND, Fan E, Camporota L, Antonelli M et al (2012) The Berlin definition of ARDS: an expanded rationale, justification, and supplementary material. Intensive Care Med 38(10):1573–1582

    Article  Google Scholar 

  11. 11.

    Ackermann M, Verleden SE, Kuehnel M, Haverich A et al (2020) Pulmonary Vascular Endothelialitis, Thrombosis, and Angiogenesis in Covid-19. N Engl J Med 383(2):120–128

    CAS  Article  Google Scholar 

  12. 12.

    Santamarina MG, Boisier D, Contreras R, Baque M et al (2020) COVID-19: a hypothesis regarding the ventilation-perfusion mismatch. Crit Care 24(1):395

    Article  Google Scholar 

  13. 13.

    Teuwen LA, Geldhof V, Pasut A, Carmeliet P (2020) COVID-19: the vasculature unleashed. Nat Rev Immunol 20(7):389–391

    CAS  Article  Google Scholar 

  14. 14.

    Wichmann D, Sperhake JP, Lutgehetmann M, Steurer S et al (2020) Autopsy findings and venous thromboembolism in patients with covid-19: a prospective cohort study. Ann Intern Med 173(4):268–277

    Article  Google Scholar 

  15. 15.

    Patel BV, Arachchillage DJ, Ridge CA, Bianchi P et al (2020) Pulmonary Angiopathy in Severe COVID-19: physiologic, imaging, and hematologic observations. Am J Respir Crit Care Med 202(5):690–699

    Article  Google Scholar 

  16. 16.

    Tomashefski JF Jr, Davies P, Boggis C, Greene R et al (1983) The pulmonary vascular lesions of the adult respiratory distress syndrome. Am J Pathol 112(1):112–126

    PubMed  PubMed Central  Google Scholar 

  17. 17.

    Vesconi S, Rossi GP, Pesenti A, Fumagalli R et al (1988) Pulmonary microthrombosis in severe adult respiratory distress syndrome. Crit Care Med 16(2):111–113

    CAS  Article  Google Scholar 

  18. 18.

    Greene R, Zapol WM, Snider MT, Reid L et al (1981) Early bedside detection of pulmonary vascular occlusion during acute respiratory failure. Am Rev Respir Dis 124(5):593–601

    CAS  PubMed  Google Scholar 

  19. 19.

    Shi H, Han X, Jiang N, Cao Y et al (2020) Radiological findings from 81 patients with COVID-19 pneumonia in Wuhan, China: a descriptive study. Lancet Infect Dis 20(4):425–434

    CAS  Article  Google Scholar 

  20. 20.

    Zhang L, Feng X, Zhang D, Jiang C et al (2020) Deep vein thrombosis in hospitalized patients with COVID-19 in Wuhan, China: prevalence, risk factors, and outcome. Circulation 142(2):114–128

    Article  Google Scholar 

  21. 21.

    Lang M, Som A, Mendoza DP, Flores EJ, Reid N, Carey D, Li MD, Witkin A, Rodriguez-Lopez JM, Shepard JO, Little BP (2020) Hypoxaemia related to COVID-19: vascular and perfusion abnormalities on dual-energy CT. Lancet Infect Dis.

    Article  PubMed  PubMed Central  Google Scholar 

  22. 22.

    Ladakis C, Myrianthefs P, Karabinis A, Karatzas G et al (2001) Central venous and mixed venous oxygen saturation in critically ill patients. Respiration 68(3):279–285

    CAS  Article  Google Scholar 

  23. 23.

    Herrmann P, Nguyen XP, Luecke T, Quintel M (2002) MALUNA 1.03 ein Softwaretool zur analyse computertomographischer Schnittbilder del Lunge. In: Jamal R, Jaschinski H (eds) Virtuelle instrumente in der praxis. Huethig Verlag, Heidelberg, pp 389–395

    Google Scholar 

  24. 24.

    Stuart EA (2010) Matching methods for causal inference: A review and a look forward. Stat Sci 25(1):1–21

    Article  Google Scholar 

  25. 25.

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

    CAS  Article  Google Scholar 

  26. 26.

    Caironi P, Carlesso E, Cressoni M, Chiumello D et al (2015) Lung recruitability is better estimated according to the Berlin definition of acute respiratory distress syndrome at standard 5 cm H2O rather than higher positive end-expiratory pressure: a retrospective cohort study. Crit Care Med 43(4):781–790

    Article  Google Scholar 

  27. 27.

    Riley RL, Cournand A (1949) Ideal alveolar air and the analysis of ventilation-perfusion relationships in the lungs. J Appl Physiol 1(12):825–847

    CAS  Article  Google Scholar 

  28. 28.

    Reske AW, Costa EL, Reske AP, Rau A et al (2013) Bedside estimation of nonaerated lung tissue using blood gas analysis. Crit Care Med 41(3):732–743

    Article  Google Scholar 

  29. 29.

    Gattinoni L, Caironi P, Cressoni M, Chiumello D et al (2006) Lung recruitment in patients with the acute respiratory distress syndrome. N Engl J Med 354(17):1775–1786

    CAS  Article  Google Scholar 

  30. 30.

    Henderson WR, Chen L, Amato MBP, Brochard LJ (2017) Fifty Years of Research in ARDS. Respiratory Mechanics in Acute Respiratory Distress Syndrome. Am J Respir Crit Care Med 196(7):822–833

    Article  Google Scholar 

  31. 31.

    Diehl JL, Mercat A, Pesenti A (2019) Understanding hypoxemia on ECCO2R: back to the alveolar gas equation. Intensive Care Med 45(2):255–256

    Article  Google Scholar 

  32. 32.

    Pan C, Chen L, Lu C, Zhang W et al (2020) Lung recruitability in COVID-19-associated acute respiratory distress syndrome: a single-center observational study. Am J Respir Crit Care Med 201(10):1294–1297

    CAS  Article  Google Scholar 

  33. 33.

    Roesthuis L, van den Berg M, van der Hoeven H (2020) Advanced respiratory monitoring in COVID-19 patients: use less PEEP! Crit Care 24(1):230

    Article  Google Scholar 

  34. 34.

    Bos LDJ, Paulus F, Vlaar APJ, Beenen LFM et al (2020) Subphenotyping acute respiratory distress syndrome in patients with COVID-19: consequences for ventilator management. Ann Am Thorac Soc 17(9):1161–1163

    Article  Google Scholar 

  35. 35.

    Shi H, Han X, Jiang N, Cao Y, Alwalid O, Gu J, Fan Y, Zheng C (2020) Radiological findings from 81 patients with COVID-19 pneumonia in Wuhan, China: a descriptive study. Lancet Infect Dis 20:425–434

    CAS  Article  Google Scholar 

  36. 36.

    Gattinoni L, Coppola S, Cressoni M, Busana M et al (2020) Reply by Gattinoni et al. to Hedenstierna et al., to Maley et al., to Fowler et al., to Bhatia and Mohammed, to Bos, to Koumbourlis and Motoyama, and to Haouzi et al. Am J Respir Crit Care Med 202(4):628–630

    CAS  Article  Google Scholar 

Download references


We would like to thank Sartorius AG, Göttingen, Germany for an unrestricted grant for lung injury-related research towards the Department of Anesthesiology of Göttingen University Medical Center.


Open Access funding enabled and organized by Projekt DEAL. Institutional.

Author information




Conception and design: DC, SC, PF, MC. Data collection: SC, FR, PF, MB, TP, MMP, MC, PH. Analysis and interpretation: MB, KM, LC, MQ, LG. Drafting the manuscript for important intellectual content: DC, LC, JJM, LG.

Corresponding author

Correspondence to Luciano Gattinoni.

Ethics declarations

Conflicts of interest

The authors have no interests to disclose.

Ethics approval

Ethics Committee approval that allows physiological and CT study for all patients with severe respiratory failure admitted to our critical care unit (ethical committee numbers: 42937/2016 and 9890/2017).

Consent to participate


Consent for publication

All authors have approved the manuscript.

Availability of data and material

Dataset available upon reasonable request.

Additional information

Publisher's Note

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

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary file1 (DOCX 2727 kb)

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, which permits any non-commercial 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

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Chiumello, D., Busana, M., Coppola, S. et al. Physiological and quantitative CT-scan characterization of COVID-19 and typical ARDS: a matched cohort study. Intensive Care Med (2020).

Download citation


  • COVID-19
  • ARDS
  • Respiratory system mechanics
  • Mechanical ventilation
  • CT scan