Respiratory Physiology of COVID-19-Induced Respiratory Failure Compared to ARDS of Other Etiologies

Domenico Luca Grieco; Filippo Bongiovanni; Lu Chen; Luca S. Menga; Salvatore Lucio Cutuli; Gabriele Pintaudi; Simone Carelli; Teresa Michi; Flava Torrini; Gianmarco Lombardi; Gian Marco Anzellotti; Gennaro De Pascale; Andrea Urbani; Maria Grazia Bocci; Eloisa S. Tanzarella; Giuseppe Bello; Antonio M. Dell'Anna; Salvatore M. Maggiore; Laurent Brochard; Massimo Antonelli


Crit Care. 2020;24(529) 

In This Article


The result of this matched-cohort study can be summarized as follows:

  • In COVID-19 patients, the severity of hypoxemia was related to respiratory system compliance reduction. This suggests that aeration loss is a relevant mechanism of hypoxemia.

  • Similarly to ARDS from other causes, respiratory mechanics of COVID-19 patients was highly heterogeneous.

  • Average respiratory system compliance and ventilatory ratio of COVID-19 patients were slightly higher than those of ARDS from other etiologies. Although statistically significant, differences appear clinically small.

  • The potential for PEEP-induced lung recruitment was variable. Average recruitability was similar to ARDS from non-COVID etiology. COVID-19 patients showed frank oxygenation response to PEEP, independently from recruitability.

Respiratory Mechanics

Few data are available about respiratory mechanics and response to PEEP COVID-19 patients with acute respiratory failure.[18–20,34] Our study compared respiratory mechanics and response to PEEP of patients with COVID-19 with those of matched ARDS from other etiologies, who have undergone the same procedures.

In our study, respiratory mechanics was highly heterogenous both in COVID-19 patients and in ARDS of other etiologies. Average values of respiratory system compliance were slightly higher in COVID-19 patients. Albeit statistically significant, mean difference (7 ml/cmH2O) may not be clinically relevant. This may depend on the small differences in body mass index between the two cohorts, and heterogeneous ARDS causes in the control group. In COVID-19 patients, compliance reduction was linearly related to oxygenation impairment: this indicates that aeration loss is a causative mechanism of hypoxemia, which is the hallmark of ARDS pathophysiology (i.e., the baby lung).[35–38]

Gattinoni et al. have hypothesized that the acute respiratory failure caused by COVID-19 is a time-related disease spectrum within different phenotypes.[18] Our results indicate that, soon after intubation, heterogeneity and average values of respiratory mechanics are similar to ARDS of other etiologies. Our data come from a limited sample. However, results appear consistent with those of a recent large study on 742 patients[39] and with the recently published physiologic data by Haudebourg and coworkers.[19] They found no major differences in respiratory mechanics between patients with ARDS from COVID-19 and other etiologies. They did not exactly match their patients as done in our study, but could not detect relevant differences between ARDS from COVID-19 and other etiologies. Also, other authors have reported high heterogeneity in the respiratory mechanics and response to PEEP of COVID-19 patients.[22,40,41] These considerations strengthen the hypothesis that, from a ventilatory standpoint, clinicians should approach COVID-19 patients who fulfill ARDS criteria with our current evidence-based practices, informed by bedside physiology.[15,42–44]

Whether the microvascular involvement represents a disease-specific feature of COVID-19 disease is debated.[27,45] In our study, ventilatory ratio was slightly higher in COVID-19 patients than in ARDS of other causes. The ventilatory ratio is correlated with dead space and can reflect microvascular thrombosis, which yields ventilation-perfusion mismatch.[22] However, microcirculatory involvement and increased dead space are hallmarks of ARDS as well.[46,47] Larger cohorts will be needed to subtle differences on this specific aspect.

Response to PEEP—Gas Exchange

More than 95% of patients improved oxygenation with high PEEP, independently from recruitability. The oxygenation improvement achieved with high PEEP was greater in COVID-19 patients than in patients affected by ARDS of other causes, although the potential for lung recruitment was not different. PEEP-induced improvement in oxygenation without alveolar recruitment could be caused by decreased cardiac output, with redistribution of lung perfusion towards the normally aerated compartment.[22,34] This indicates that, similarly to ARDS from other causes, the oxygenation response to PEEP is not informative about alveolar recruitment in COVID-19 as well.

Interestingly, PEEP-induced alveolar recruitment was correlated with PaCO2 changes, and all patients with low potential for lung recruitment developed increases in PaCO2 with high PEEP. Changes in PaCO2 due to PEEP reflect dead space modifications. In case of poorly recruitable lungs, alveolar dead space increases due to compression of pulmonary vessels,[48] and airway dead space augments due to gas compression in the respiratory circuit and airways.[32] With alveolar recruitment, overdistension by tidal volume is mitigated, and this reduces ventilation-perfusion mismatch.[49,50]

Response to PEEP—Recruitability

In our study, recruitability of COVID-19 patients was variable, with an average value similar to ARDS from other etiologies. Our results are consistent with most recent data indicating great heterogeneity in the response to PEEP in COVID-19 patients.[19,22,23] This has relevant clinical implications, as PEEP setting should balance between its capability to recruit new alveoli and the unavoidable overdistension in already open tissue.[51,52] As such, a high PEEP should be beneficial only in patients having greater potential for lung recruitment, in whom PEEP increases the size of the aerated lung available for tidal ventilation. Conversely, in non-recruitable patients, PEEP only enhances lung injury by increased static stress and strain.[50] Recruitability could not be predicted by changes in oxygenation, compliance, or driving pressure in response to PEEP, which represent popular proposed PEEP-setting strategies.[53] This suggests that bedside assessment of the potential for lung recruitment appears warranted in COVID-19 patients. The recently developed recruitment-to-inflation ratio (which represents recruited volume normalized to aerated lung size) offers a simple, timely, and reproducible assessment of gas recruitment.[29] This may help distinguish patients showing high vs. low recruitability at the bedside, possibly supporting PEEP setting.

Clinical Outcome

In our study, 28-day mortality of COVID-19 patients was higher than that reported for ARDS of other causes in a large multicenter observational study (35% for moderate and 43% for severe ARDS).[14] We report high mortality rate despite tidal volume, plateau, and driving pressure were within a relatively safe range and prone position was applied in most of the patients. The majority of our patients were intubated after receiving noninvasive respiratory support, which could have selected the most severe population at higher risk for mortality.[14,54–56] However, the small size of our cohort precludes from further speculation on the reasons for this high mortality.


This study has limitations.

First, our COVID-19 patients were studied within 24 h after endotracheal intubation: it is possible that respiratory physiology varies over time, as suggested by other investigator showing low respiratory system compliance and minimal recruitability at a later stage of COVID-19 ARDS.[57] This reinforces the clinical message of our study, which addresses individualized interventions based on bedside assessment of physiology.

Second, while the matched comparison with non-COVID-19 historical cohort has several strengths, we cannot exclude that uncontrolled individual characteristics of studied patients affected some of study results. In particular, duration of intubation prior to measurements could not be matched, and patients in the control group showed heterogeneous causes of ARDS: both these features may have affected study results.