• •

  To adequately identify patients with ARF amenable to the start of noninvasive respiratory support (NIRS).

• •

  To know the processes underlying high risk of failure of NIRS.

Acute respiratory failure is the most common cause of death among patients with influenza and viral infections in general.9 With regard to COVID-19 (the infectious disease caused by SARS-CoV-2), 3.4% of the patients infected in China presented acute respiratory distress syndrome (ARDS), representing 40.3% of the patients with a serious disease.5 It is therefore very important to adopt a therapeutic strategy for ARF secondary to COVID-19 infection.

A fundamental element in the management of ARF in COVID-19 patients in monitoring – preferably of a non-invasive nature. These patients need to be located in individual rooms, ideally with negative pressure. This implies the need for centralized monitoring control (pulsioxymetry, respiratory frequency), preferably with video monitoring.

In the same way as in other processes that cause ARF, in patients with SARS-CoV-2 we can use the respiratory management strategy proposed by Scala and Heunks.10 Conventional oxygen therapy involves the administration of oxygen at different concentrations, and constitutes the basis of the therapeutic pyramid. The next step is high-flow nasal cannula (TAFCN) therapy. This technique involves the use of a gas mixture at high flow rates (up to 60lpm), with variable proportions (FiO2) of air and oxygen administered through a nasal cannula. The administered gas must be warm and with 100% humidification.11 The advantages with respect to conventional oxygen are a constant FiO2, dead space reduction, and the generation of a positive pressure that implies redistribution of intraalveolar fluid and alveolar recruitment.12 The next step is non-invasive mechanical ventilation (NIMV). The main characteristic of NIMV is its use under spontaneous breathing conditions; sedation is therefore null or low.13 The penultimate step is invasive mechanical ventilation (IMV). In this case ventilation in controlled mode is usually used, and tracheal intubation is required.10 The last step is extracorporeal membrane oxygenation (ECMO).14 At present, the respiratory therapies that define NIRS and HFNC and NIMV.

As an orientation and taking the above into account, the criteria for starting respiratory support in ARF secondary to COVID-19 would be the following15:

Clinical criteria:

• •

  Moderate to severe dyspnea with signs of labored breathing and the use of accessory muscles, or paradoxical abdominal motion.

• •

  Tachypnea in excess of 30rpm.

Blood gas criteria:

• •

  PaO2/FiO2 <200 (or the need to administer FiO2 >0.4 to secure SpO2 of at least 92%).

• •

  Acute ventilatory failure (pH <7.35 with PaCO2 >45mmHg).

In patients without any of the above criteria, the initially indicated treatment would be conventional oxygen therapy. In the presence of any of the above criteria, respiratory support would be indicated – whether invasive or non-invasive. The use of NIRS for SARS and other viral pandemics is subject to controversy, with NIMV failure rates of about 30%.16

More recently, NIMV has also been used in patients with ARF due to influenza A H1N1, with failure rates of between 13 and 77%.17–19 Despite the uncertainty of the evidence and the lack of randomized clinical trials, the positive data obtained by most of the observational studies suggest that its use can be considered in carefully selected patients in experienced centers and in a protected environment (ideally rooms with negative pressure). In the present Chinese epidemic, 5.1% of the patients required NIMV, 2.3% IMV and 0.5% ECMO.5 Accordingly, the treatment choice will also depend on the patient background disease (mainly of a respiratory nature), the location of the patient (room with negative pressure, closed individual room with air renewal), and the possibility of an eventual need for aerosol generating procedures. Thus, globally speaking, we may encounter three different clinical scenarios:

• 1.

  Patients without prior disease (de novo ARF) with hypoxemic respiratory failure, and therefore amenable to escalate treatment up to ECMO. The NIRS failure rate in this clinical scenario – mainly in relation to NIMV – is extremely high. Furthermore, there is evidence of increased mortality if the start of IMV is delayed. We therefore do not recommend the use of NIRS in these patients.1,20,21 It could only be contemplated in carefully selected patients, and provided all the following criteria are met19,21–27:

• •

  PaO2/FiO2 >100 despite conventional oxygen therapy.

• •

  Absence of multiorgan failure (APACHE score <20).

• •

  A team with expertise in continuous monitoring is required. It is therefore advisable to perform the technique in special units with a pre-specified nursing staff ratio as in ICUs and IRCUs.

• •

  Early orotracheal intubation (OTI) in the following hour, in the absence of criteria of improvement. In this regard, and in addition to the traditional intubation criteria, we could consider intubation in patients treated with HFNC presenting a respiratory rate-oxygenation (ROX) index ([SpO2/FiO2]/respiratory frequency) of <3, <3.5 and <4 at 2, 6 and 12 hours after the start of HFNC treatment. Likewise, intubation could be considered in patients with a heart rate, acidosis, consciousness, oxygenation, and respiratory rate (HACOR) index of >5 one hour after the start of NIMV therapy.

Extrapolating the evidence referred to de novo ARF, the use of HFNC would be the first choice.21,28 Non-invasive mechanical ventilation is defined as the second option in the event of insufficient patient response and without immediate criteria for IMV.

• 2.

  Patients with hypoxemic failure, instructions against OTI and therapeutic ceiling in NIMV. The start of NIRS would be indicated in these patients, provided the preventive measures are ensured. In this context it is essential to establish the treatment objectives with the patient and family, delimiting the therapeutic ceiling. In general, it is advisable to start treatment with HFNC before resorting to NIMV.21,28 We recommend the following considerations for the use of NIRS29,30:

• •

  Adjust FiO2 to secure a target SpO2 of about 95%.31

• •

  If HFNC is used, it is advisable to administer flows of over 50lpm; if possible, start with 60lpm.

• •

  If NIMV is decided, it is advisable to use positive end-expiratory pressure (PEEP) values and low assist pressures (in order to obtain a high end-tidal volume (VTe) of <9ml/kg ideal body weight).31,32

• 3.

  Patients with severe exacerbation of chronic obstructive pulmonary disease (COPD) with acute or exacerbated hypercapnic respiratory failure: perform a NIRS therapeutic test, especially with NIMV. High-flow nasal cannulas may be useful in these patients in the event of intolerance to NIMV, or for the lowering of NIMV.33–36

To limit transmission of the infection to both the healthcare staff and to other patients.

The use of NIRS is a practice of special risk of SARS-CoV-2 transmission.37–40 The capacity of SARS-CoV-2 to infect healthcare workers has been confirmed, though it is still too soon to establish comparisons with MERS and SARS.40 Thus, the Spanish Ministry of Health, Consumer Affairs and Social Wellbeing has defined a series of recommendations for management of the disease, and this document seeks to disclose the measures that limit transmission of the infection to the healthcare staff.2,39 In order to be able to apply MIRS to a patient with suspected or confirmed SARS-CoV-2 infection, it is crucial to follow the general preventive recommendations. In this regard, the public administrations must be in possession of the materials required for safe NIRS use.

• 1.

  The healthcare staff assisting cases under investigation for COVID-19 infection or with already confirmed infection subjected to NIRS must wear personal protective equipment (PPE) for the prevention of infection2,22,37,39,41 during aerosol-generating procedures that have been associated to an increased risk of air transmission of pathogens. The preventive measures should be targeted to microorganisms that are transmitted via droplets and through contact, and are to include:

• •

  A high-efficacy FFP2 mask or preferably an FFP3 mask (if available).

• •

  Integral-frame protective goggles.

• •

  Retraction of long hair into a knot or tail; a surgical cap can be used in this regard.

• •

  Gloves.

• •

  Long-sleeved, impermeable microbiological protective gowns.

• 2.

  Patient location within the hospital setting is related to the possibility of performing aerosol-generating procedures. In this respect, the available evidence on the use of NIRS devices as invasive procedures and the risk of pathogen transmission to the healthcare professionals is stronger in cases of orotracheal intubation and patients with an artificial airway, as well as in noninvasive ventilation – though these data come from limited studies of very low quality. Interpretation of the data is therefore difficult (Table 1). Therefore, in mild cases it is advisable to keep the patient in a room with negative pressure. If no such rooms are available, the patient can be located in an individual room with bathroom. The door to the room should be kept closed at all times.39 Patients with severe hypoxemia are to be admitted to special units.

  Table 1.

  Summary of aerosol generation procedures and risk of transmission of acute respiratory diseases.

  Aerosol generation procedures	Type of studies	Odds ratio (95%CI)
  Orotracheal intubation	4 cohort studies	3.0 (1.4–6.7)
  		22.8 (3.9–131.1)
  		13.8 (1.2–167.7)
  		5.5 (0.6–49.5)
  	Pooled effect (I2=39.6%)	6.6 (2.3–18.9)
  Orotracheal intubation	4 case–control studies	0.7 (0.1–3.9)
  		9.2 (4.2–20.2)
  		8.0 (3.9–16.6)
  		9.3 (2.9–30.2)
  	Pooled effect (I2=61.4%)	6.6 (4.1–10.6)
  Aspiration before intubation	2 cohort studies	13.8 (1.2–161.7)
  		1.7 (0.7–4.2)
  	Pooled effect (I2=59.2%)	3.5 (0.5–24.6)
  Aspiration after intubation	2 cohort studies	0.6 (0.1–3.0)
  		1.8 (0.8–4.0)
  	Pooled effect (I2=28.8%)	1.3 (0.5–3.4)
  Nebulizations	3 cohort studies	6.6 (0.9–50.5)
  		0.1 (0–1.0)
  		1.2 (0.1–20.7)
  	Pooled effect (I2=73.1%)	0.9 (0.1–13.6)
  Bronchoscopy	2 cohort studies	3.3 (0.2–59.6)
  		1.1 (0.1–18.5)
  	Pooled effect (I2=0%)	1.9 (0.2–14.2)
  Endotracheal aspirations	1 cohort study	1.0 (0.2–5.2)
  High-flow nasal cannula therapy	1 cohort study	0.4 (0.1–1.7)
  Noninvasive mechanical ventilation	2 cohort studies	2.6 (0.2–34.5)
  		3.2 (1.4–7.2)
  	Pooled effect (I2=0%)	3.1 (1.4–6.8)
  Oxygen administration	1 case–control study	1.0 (0.3–2.8)
  Chest compressions	1 case–control study	4.5 (1.5–13.8)
  Chest compressions	2 cohort studies	3.0 (0.4–24.5)
  		0.4 (0.0–7.8)
  	Pooled effect (I2=27.3%)	1.4 (0.2–11.2)
  Defibrillation	2 cohort studies	0.5 (0.0–12.2)
  		7.9 (0.8–79.0)
  	Pooled effect (I2=55.3%)	2.5 (0.1–43.9)
  Tracheotomy	1 case–control study	4.2 (1.5–11.5)
  NIMV mask manipulation	1 cohort studies	4.3 (0.6–27.4)
  Mechanical ventilation	1 cohort studies	0.9 (0.4–2.0)
  Manual ventilation before intubation	1 cohort studies	2.8 (1.3–6.4)
  Sputum sample collection	1 cohort studies	2.7 (0.9–8.2)
  Insertion of nasogastric tube	2 cohort studies	1.7 (0.2–11.5)
  		1.0 (0.2–4.5)
  	Pooled effect (I2=0%)	1.2 (0.4–4.0)

  I2: statistical heterogeneity index; CI: confidence interval.

  Results of studies selected in the systematic review that measured the risk of SARS transmission to healthcare professionals exposed to the cited procedures versus those not exposed to the same procedures.48,49

• 3.

  When in-hospital transfer is required, both the patient and the professional in charge of moving the patient are to wear surgical masks. During transfer, the patient bed is to be covered with a disposable clean sheet that should be eliminated afterwards as group III waste.39