The present study compared patients with CARDS to patients with ARDS. The main findings of our study are: i) the pulmonary hemodynamic is not profoundly altered in CARDS compared to usual ARDS and ii) auto-regulation of lung vessels seems to be largely intact in patients with SARS-CoV2-related lung failure, iii) mechanical ventilation intensity was statistically significant higher in CARDS compared to usual ARDS.
Usual ARDS is defined by decreased diffusion capacity and ventilation deficit as a consequence of inflammation and low compliance7, resulting in a “baby lung”. In addition, there is often pulmonary hypertension due to increased vascular resistance with the risk of consecutive right heart failure4. In COVID-19 patients admitted to our ICU we could not completely affirm these pathophysiological findings.
Further paCO2 in COVID-19 patients was rather similar compared to patients with ARDS at 52 mmHg and 41 mmHg, respectively. Our data shows that all COVID-19 patients had tidal volumes and a compliance resulting in minute ventilation of 6 L/min to 20 L/min. These findings might not be typical for patients with ARDS. In these patients tidal volume is low for the compliance of the lung is decreased13. As CO2 has a higher diffusion coefficient than O2, hypercapnia is normally a problem of ventilation rather than diffusion14. In usual ARDS, patients hypercapnia can be explained by low tidal volume and the following lack of alveolar ventilation. But in CARDS patients minute ventilation was not decreased. Our group of CARDS patients had significantly higher tidal volume and a higher compliance than ARDS patients. Matching these two groups by PaO2 and oxygenation index, these parameters were quite similar in both groups. The question that arises is the pathophysiology behind the hypercapnia despite high minute ventilation and preserved compliance. V/Q mismatch could be an explanation for the ineffective gas exchange in CARDS patients. Also, non-effective regulation of hypoxic vasoconstriction10 could be another reason why blood volume passes non-ventilated areas in the lungs and is not able to participate in gas exchange. However, we could not find differences between CARDS and ARDS with regard to calculated physiological dead-space and shunt fraction. Calculated dead-space in CARDS and ARDS patients was in the range expected according to PaO2/FiO2 ratio15. However, we found a significant difference in the estimated CO2 production, which might explain at least some of the CARDS features. Indeed, the intensity of ventilation was significantly higher in CARDS patients than ARDS patients, but the intensity in CARDS patients was not surprisingly high and comparable to previously reported data in patients with ARDS16.
In ARDS patients there is often pulmonary hypertension as consequence of hypoxia or hypercapnia. Further, there is a higher rate of right heart failure and lower cardiac output in ARDS patients because of a higher right ventricular afterload17,18. In CARDS patients we did not see signs of right heart failure or decreased cardiac output. The cardiac output in CARDS patients was 7.8 L/min in median compared to 6.1 L/min in ARDS patients (cardiac index 3.9 L/min/m² vs. 3.1 L/min/m²). Pulmonary vascular resistance was significantly lower in CARDS patients than in ARDS patients (133 dyn × sec × cm− 5 in COVID-19 vs. 248 dyn × sec × cm− 5 in ARDS patients). One reason for increased PVR is higher PEEP. It has been shown previously, that increasing the PEEP increases PVR in mechanically ventilated patients19. In our study, PEEP was similar in both groups and also pulmonary capillary wedge-pressure did not differ and there was no correlation between PEEP and PVR. The difference in PVR may thus simply result from increased cardiac output of CARDS patients. CARDS patients probably had an increased cardiac output in comparison to ARDS patients due to inflammation and low systemic vascular resistance.
Mean pulmonary arterial pressure was similar in COVID-19 patients and in patients with ARDS, nevertheless both groups showed pulmonary arterial hypertension to some degree. Both groups had post-capillary forms of pulmonary hypertension with PCWP > 15 mmHg, but the diastolic pressure gradient was 7 mmHg in median in CARDS patients vs. 8 mmHg in ARDS patients. With a DPG less than 7 mmHg there is no precapillary component in pulmonary hypertension according to the actual guidelines12. In our study, the median DPG was 7 mmHg and PVR 1.7 WU × m² (vs. DPG 8 mmHg, PVR 3.1 WU × m² in ARDS) meaning the pulmonary arterial hypertension in CARDS may be more a result of a post-capillary component, than a precapillary. Hence, there was only a slight difference and a defect in pulmonary vascular regulation in the presence of hypoxemia and hypercapnia unable to avoid a pathologically high right-left shunt20 can be ruled out. We did not calculate an unexpected high shunt in CARDS patients. Rather, an elevated CO2 production might explain differences between CARDS and ARDS. Patients presented with prolonged fever and unexpected long periods of elevated inflammatory parameters, both might contribute to increased VCO2.
Early recommendations for COVID-19 patients with respiratory failure consisted in early invasive ventilation strategies to avoid contagious aerosols and the use of high PEEP analogue to usual ARDS21. These therapies may be most effective in patients with lower diffusion capacity and compliance but the benefit in ventilation/perfusion mismatch needs to be questioned. These recommendations may be effective in ARDS patients, however in CARDS patients an important cornerstone according to our data might be to limit excessive CO2 production e.g. by rigorous temperature control.
Moreover veno-venous extracorporeal life support (vvECMO) should be considered to treat these patients. Six patients in the CARDS group recieved vvECMO therapy later on, because of increasing hypercapnia and pH < 7.1 at exhausted mechanical ventilation.