key: cord-0782059-i15i6cjr authors: Chosidow, Samuel; Contou, Damien; Fraissé, Megan; Pajot, Olivier; Mentec, Hervé; Cally, Radj; Plantefève, Gaëtan title: “Early” and “delayed” intubation of COVID-19 patients: different definitions, different populations. date: 2022-02-25 journal: Respir Med Res DOI: 10.1016/j.resmer.2022.100897 sha: c31960e90c845c33b595c886991656d0bbc3bc8d doc_id: 782059 cord_uid: i15i6cjr nan The Coronavirus Disease 2019 (COVID-19) pandemic has raised the question of the timing of initiation of invasive mechanical ventilation (IMV). Some authors have advocated early intubation, arguing that it would lower the risk of unprepared urgent intubation and lessen patient self-inflicted lung injury as well as droplet aerosolization for healthcare workers [1] . Others have recommended using more conventional criteria for initiation of IMV [2] . Several studies have compared "early" to "delayed" intubation strategies in COVID-19 patients but could not evidence definite messages [3] . "Early" and "delayed" intubation have mainly been defined according to the duration between Intensive Care Unit (ICU) admission and intubation. In clinical practice however, the decision to initiate IMV usually relies on the patient's clinical features, including respiratory rate (RR) and oxyhemoglobin saturation (SpO 2 ) [4, 5] . An "early" approach to IMV can therefore be construed as intubating patients showing less severe or partial signs of respiratory distress [6] . We aimed to assess the discrepancies in the categorization of patients using two definitions of "early" and "delayed" Table 1 . Using the clinical definition, 54 patients were classified in the "early" intubation group and 60 in the "delayed" intubation group. On the day of intubation, RR (36 vs. 45 breaths/min, p<0.001) and SpO 2 (91 vs. 85%, p<0.001) significantly differed between groups. There was no difference between these groups regarding the delay between baseline and intubation (36 vs. 35 hours, p=0.75) ( Table 1 and Supplementary Material Figure 1 ). With the temporal definition that uses the median delay between baseline and intubation (median 36 hours), 59 patients were classified in the "early" intubation group and 55 in the "delayed" intubation group. Fifty-seven patients (50%) were mismatched when comparing these classifications. Precisely, 48% of the "early" intubation group and 52% of the "delayed" intubation group as determined by the clinical definition were classified differently with the temporal definition. There was no difference in ICU mortality between the "early" and "delayed" intubation groups using the clinical definition (52% vs. 65%, p=0.22) or the temporal definition (54% vs. 63%, p=0.41). No cardiac arrest occurred on induction of anesthesia. In this study, using a clinical definition to define "early" and "delayed" intubation yielded a different categorization of patients than using a temporal definition. Using solely a temporal criterion as it has been done in recent studies [3] might constitute highly heterogeneous groups in regards to clinical features prior to intubation. Moreover, it is difficult to determine at what point one should "start the clock", since rates of disease progression seem to vary drastically between patients. The clinical definition, however, does not depend on this variability of disease progression but still relates to the potential risks of clinically "delayed" intubation, such as patient self-inflicted lung injury or hypoxic cardiac arrest. Using both definitions, there was a trend towards higher ICU mortality in the "delayed" intubation groups which did not reach statistical significance, possibly due to a lack of power. These results must be interpreted with caution, considering that the patients were not randomized and that the groups were significantly different in regards to several characteristics (such as period of admission, dexamethasone therapy, oxygenation device). Nonetheless, one interesting finding is the higher proportion of patients treated with HFNC in the "delayed" intubation group as categorized with the temporal definition. Although the retrospective design of this study impedes definite conclusions, an explanatory hypothesis is that HFNC, by improving respiratory mechanics and oxygenation, might lengthen the time spent breathing spontaneously before a possible intubation. There is conceivably an implication of the oxygenation device on the time-course of the respiratory distress in spontaneously breathing patients. In conclusion, investigating the appropriate timing of intubation of COVID-19 patients is paramount since "early" and "delayed" intubations are both associated with still unclear risk-benefit ratios. In this study, we showed that a categorization based on RR and SpO 2 provides a new approach to explore this issue by classifying patients differently, and we believe in a more suited way, than using a temporal definition. Intubation and Ventilation amid the COVID-19 Outbreak Caution about early intubation and mechanical ventilation in COVID-19 Effect of timing of intubation on clinical outcomes of critically ill patients with COVID-19: a systematic review and meta-analysis of non-randomized cohort studies High-Flow Oxygen through Nasal Cannula in Acute Hypoxemic Respiratory Failure International variation in the management of severe COVID-19 patients High-flow nasal oxygen in patients with COVID-19-associated acute respiratory failure Simplified Acute Physiology Score; HFNC, High-Flow Nasal Cannula; ICU, Intensive Care Unit Oxyhemoglobin saturation We warmly acknowledge Dr Elsa Logre, Dr Jo-Anna Tirolien, Dr Olivia Picq, Dr Florence Sarfati, Dr Paul Desaint, and all the residents who cared for the patients as well as Dr Clara Finck for her writing assistance. SC, GP, RC and DC are responsible for the conception and design. All the authors took care of the patients. SC is responsible for data acquisition. DC performed the statistical analysis. All the authors were responsible for analysis and interpretation of data. All authors read, critically reviewed and approved the final manuscript. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. None