The main finding of our study is that, among intubated patients with severe respiratory failure and COVID-19, intubation after 2 days of hospital admission was associated with increased mortality. Furthermore, patients who were intubated 6 days after hospital admission had a much higher mortality rate than those patients who were intubated early (i.e., within 3 days of admission). In general, survivors had a shorter time to intubation than non-survivors. In addition, age, a high SOFA score, a high WBC number at ICU admission, a low PaO2/FiO2 ratio after intubation, and low static compliance of the respiratory system were also associated with increased mortality.
A valid reason for this clinical course could be the spontaneous, prolonged ventilatory efforts before intubation, inducing the progression of patients’ lung damage, also known as P-SILI1,2,23. Both disease progression and superimposed P-SILI could result in failure of the CPAP and/or HFNC support therapies, and a need for intubation. Non-invasive respiratory support has been considered a very effective therapy for overcoming gas exchange impairment and potentially averting the need for intubation in ARDS patients24. However, patients failing non-invasive ventilation have been shown to have a particularly poor prognosis25. Intubation and MV yield protective effects by decreasing inspiratory effort and tidal volumes, thus limiting the extent of P-SILI. There was an early hypothesis-driven advisory that COVID-19 patients should be ventilated early in the disease progression to prevent lung injury26. However, new studies on COVID-19 management and outcomes have called this paradigm into question27,28. As such, early management of COVID-19-induced hypoxemia employs non-invasive forms of oxygenation to forestall the need for intubation and MV.
In our study, reduction in respiratory system compliance, especially in late and very late intubated patients, is probably associated with a significant increase in non-aerated lung tissue caused by alveolar and interstitial oedema, consolidation and /or fibrosis. We do not know the impact of late intubation and prolonged HFNC and CPAP administration on the development of these lung lesions, but our findings are similar to those of other clinical trials29–31.
In COVID-19 patients, the time to intubation is still the subject of intense debate32. Our findings are similar to those of other studies, suggesting that delayed intubation in patients with severe hypoxemia worsens their prognosis, especially after a prolonged CPAP trial6–9. However, many observational studies10–13 and one meta-analysis of non-randomised cohort studies, spanning approximately 9,000 patients,14 report non-statistically significant differences in mortality between patients intubated early or late during the course of the disease.
The majority of these studies are retrospective, with a small number of patients treated with MV. In a study by Hernandez-Romieu et al., of the 231 patients admitted to the ICU, only 97 (47.2%) were eventually intubated, while the remaining were treated with HFNC11. The short median period between hospital and ICU admission (1 day) may have limited the appearance of different phenotypes of lung damage and disease progression33,34. Concerning the meta-analysis,14 there was a significant variability in the definition of early and late intubation, which was a major limitation of the study.
In our study, a high SOFA score and a high WBC number were associated with higher mortality. Perhaps this finding means that markers of the disease progression or worsening and computer tomography scans should also be considered in our final decision to intubate the patients35,36.
Our study has some critical limitations. First, it is a single-centre, observational, cohort study, and our results do not necessarily reflect the reality of other hospitals, even in our own country. In our study, however, ICU data were prospectively collected, and we included a considerable number of intubated patients. Second, we included only intubated patients, most of whom had failed the HFNC/CPAP treatment, while there were many patients with severe hypoxemia in our hospital treated with non-invasive ventilation who survived without intubation. Therefore, determining the optimal time to intubation remains a challenge.
In conclusion, our study findings indicate that among critically ill intubated COVID-19 patients, late intubation is associated with poor outcomes. During patients’ hospitalisation, additional risk factors such as age, a high SOFA score and a high WBC number may increase the mortality risk associated with late intubation. A lower PaO2/FiO2 ratio following intubation and low static compliance of the respiratory system are also significant risk factors. Further prospective studies are required to establish the best time to intubation for COVID-19 patients suffering from severe ARDS.