COVID
COVID
The Berlin definition of ARDS has limited applicability in settings where blood gas analyzers, chest
radiographs, and/or positive pressure ventilation are not reliably available. The Kigali Modification to
the Berlin criteria has been developed to address this gap (Riviello et al).
The Kigal modification requires the following four criteria to diagnose ARDS:
Key differences in the Kigali Modification include allowing ultrasound as a method of identifying
pulmonary opacities, the use of SpO2 instead of PaO2, and the lack of requirement for PEEP. Several
studies have validated the use of SpO2 in place of arterial blood gases (Chen et al; Sanz et al; Brown
et al) and ultrasound in place of chest xray for diagnosing ARDS (Lichtenstein). One single center
study has validated the modification in ventilated patients (Vercesi et al). However, a large validation
study comparing the Kigali Modification to the Berlin Criteria is needed.
When patients are not mechanically ventilated, the FiO2 will need to be estimated.
See relevant sections in Inpatient Management (listed below) for options in patients that do not
require intubation and mechanical ventilation. Some of these can be performed in the ICU or outside
of the ICU depending on the institution:
Candidacy
There are several potential indications for intubating a patient with respiratory failure. These
include:
Ultimately, the decision to intubate is based on multiple factors including the patient’s complete
clinical scenario (including goals of care) as well as the availability of local resources to safely manage
the airway and provide mechanical ventilation. Of note, early in the pandemic there were anecdotal
reports advocating for earlier intubation of COVID patients as compared to other patients with
respiratory failure from other causes. However, there are no data to support this practice. Intubation
criteria for COVID patients with respiratory failure are generally the same as for non-COVID patients
with respiratory failure.
1. Prioritize CXR and vent settings over procedures (such as central venous catheter
placement) if possible.
2. If portable chest x-ray is not available, confirm endotracheal tube placement with
bilateral breath sounds and CO2 detection
1. Typical starting rates will be 16-24 titrated to goal minute ventilation of 6-8 L/min
2. Consider starting rates of 24-28 titrated to goal minute ventilation of 8-12 L/min in
setting of acidosis (pH < 7.25) pre-intubation. If blood gas is unavailable, higher initial
minute ventilation should be targeted for patients with a pre-intubation respiratory
rate above 35.
1. Calculate P/F ratio from initial post-intubation ABG. Adjust oxygenation as described
in ARDS Oxygenation.
2. Goal pH 7.20 to 7.45. Adjust ventilation as described in ARDS Ventilation.
Mechanical Ventilation
This section addresses the management of mechanical ventilation for COVID ARDS specifically, not
mechanical ventilation for other indications. It discusses the use of AC/VC as a ventilatory mode.
Some settings may prefer APRV, which is not discussed in detail. Pressure support ventilation (PSV)
mode is often used as the patient is recovering and preparing for extubation.
This section does not discuss managing COVID ARDS with concurrent obstructive lung disease
(asthma, COPD), which ideally should be done by experienced clinicians only.
Patients with ARDS receiving mechanical ventilation are at risk of lung damage, often referred to as
ventilator induced lung injury (VILI). However, steps can be taken to reduce the risk of VILI and
decrease mortality for patients with ARDS. This is referred to as lung protective ventilation (LPV).
Minute Ventilation (respiratory rate x tidal volume) helps control pH and PCO2.
Tidal volumes should always be within the 4-8 cc/kg range, ideally 4-6cc/kg based
on Ideal Body Weight [IBW]. See ARDSNet table to look up tidal volume by gender and
height in cm or inches.
The respiratory rate often has to be high to accommodate low tidal volumes; typical RR is
20-35 breaths/minute. In patients with obstructive lung disease these are lower.
PEEP and Fi02 drive oxygenation. The goal is to deliver a partial pressure of oxygen to perfuse tissues
(PaO2 ≥ 65, Sp02 ≥ 92%) while limiting lung injury from high distending pressures (with plateau
pressures ≤ 30) and oxygen toxicity (with FiO2 ≤ 60%, SpO2 ≤ 96%).
Lower limit goals for PaO2 / SpO2 are widely debated; PaO2 > 55 and SpO2 >88% are also commonly
used. Many clinicians use PaO2/FiO2 (called the P/F ratio) to guide oxygenation as it is a shorthand
way of assessing the A/a gradient for the patient and seeing if their oxygenation is
improving. SpO2/FiO2 ratio can be used if arterial blood gases are unavailable
If FiO2 >60%; patient requires ventilator optimization (ask a specialist). If persistent, see
the Refractory Hypoxemia pathway.
1. Within half an hour of initial ventilation settings (typically PEEP 5 for BMI <40 and
PEEP 10 for BMI >40) reset PEEP and FiO2 to target oxygenation SpO2 92-96% using
the following tables:
FiO2 0.3 0.4 0.5 0.5 0.6 0.7 0.7 0.7 0.8 0.9 0. 0.9 1
9
PEEP 5 5 8 8 10 10 12 14 14 14 16 28 18-24
FiO2 0.3 0.3 0.3 0.3 0.3 0.4 0.4 0.5 0.5 0.5-0.8 0.8 0.9 1 1
PEEP 5 8 10 12 14 14 16 16 18 20 22 22 22 24
Higher levels of PEEP can cause hypotension. It is important to monitor blood pressure when
increasing PEEP.
2. Readjust Frequently
1. SpO2 <92% or >96% (do not use more oxygen or PEEP than is needed)
2. PaO2 <65 or >100
3. pPlat >30 (see this pathway)
Plateau Pressure:
It is important to avoid elevated plateau pressures (with goal ≤ 30) which can indicate relative lung
overdistention (Slutsky et al).
1. Check plateau pressure with every change in tidal volume, PEEP, or clinical deterioration
(worsening oxygenation) but not as part of routine practice. In order to accurately
measure plateau pressure, the patient must be passive (i.e. not actively breathing) on
AC/VC mode with a constant flow delivery (as opposed to decelerating flow delivery).
2. If plateau pressure is >30 cm H20, then decrease tidal volume by 1 mL/kgIBW (minimum
4 mL/kgIBW)
3. If plateau pressure is < 25 cm H20 and tidal volume < 6 mL/kgIBW, then increase tidal
volume by 1 mL/kgIBW until plateau pressure is > 25 cm H2O or tidal volume = 6
mL/kgIBW
4. If plateau pressure is < 30 cm H20 and patient is breath stacking or dyssynchronous, then
increase tidal volume in mL/kgIBW increments to 7 mL/kgIBW or 8 mL/kgIBW while
plateau pressure is < 30 cm H20
Compliance:
Compliance measures can give an indication about whether a patient’s lung stiffness is improving or
declining over time, and can help with prognostication and management.
Troubleshooting
Resistance: Troubleshooting increased Peak Inspiratory Pressure due to high resistance: Work
outside (machine) to inside (alveoli); circuit problem, ETT kink/occlusion/biting, ETT
obstructed/malpositioned, large airway obstruction (mucous plug), small/ medium airway
obstruction (bronchospasm); auscultation & passing a suction catheter can quickly eliminate many of
these.
PSV
APRV
There exists significant practice variation around the use of bilevel ventilatory modes. APRV should
only be used by providers with experience and familiarity with this mode. At this time there are no
data to support the superiority of APRV in ARDS patients, including those with COVID-19. One recent
small study in Australia found that APRV was associated with decreased survival (Zorbas et. al).
Prone Ventilation
Early proning in COVID-associated ARDS requires intensive management but can significantly
improve oxygenation. Pronation is one of the only interventions shown to improve mortality in ARDS
(PROSEVA) (Guérin et al). In one representative study in 62 COVID patients on ventilators who
underwent prone positioning, as compared to 199 similar controls who met criteria for prone
positioning but did not receive the intervention, showed a multivariate-adjusted hazards ratio for
mortality of 0.57 (0.42-0.76) in the proned patients (Shelhamer et al).
Timing: We recommend early proning in severe ARDS (<36 hrs) and prefer to initiate proning prior to
use of continuous paralytics (or inhaled pulmonary vasodilators), despite the fact that in the
PROSEVA trial over 95% of patients in both the intervention and the control arm were on continuous
paralytics.
For patients with a tracheostomy, we recommend that patients have their tracheostomy
replaced by oral endotracheal intubation (ETT) when possible, while recognizing that
decannulating a tracheostomy and placing an ETT poses an infectious risk to staff.
Renal replacement therapy can be performed while prone, typically via a femoral line.
Monitor for complications: Prone ventilation can lead to increased incidence of brachial
plexopathy in the context of increased pressure to anterior portions of the arm and shoulder
(Scholten; Goettler). Prone positioning for surgery has been associated with abdominal or
limb compartment syndromes, or Rhabdomyolysis (Kwee).
1. Have one person hold ET tube and lines to assure they are not dislodged and
continuous medications are not disrupted
1. Prone ≥16 hrs per 24 hrs. Supine ≥ 4 hrs per 24 hrs. Repeat every day. There is no day
limitation for maintaining prone ventilation and it should be repeated every day while
beneficial.
2. Discontinue neuromuscular blockade if initiated for dyssynchrony and re-assess.
1. Resupinate. Consider trying again the next day, as sometimes the recruitability of lung
tissue will change over time.
1. Unscheduled extubation
2. Endotracheal tube obstruction
3. Severe or significantly worsening hypoxemia, e.g. Sp02 <85% and consideration of
ECMO if available
4. Hemodynamic instability
Refractory Hypoxemia
If patient is hypoxic (PaO2 <75) despite PEEP optimization as above); and FiO2 > 0.6 or PaO2/FiO2
ratio < 150 then consider trying each of the following
1. Assess for improvement in oxygenation (stable oxygenation metrics while being able
to reduce FiO2 by 0.1)
2. Try stopping neuromuscular blockade daily if possible, and discontinue completely if
the patient maintains PaO2>75 with FiO2<0.75 without it.
A recruitment maneuver is the deliberate administration of a high airway pressure for protocolized
periods of time to open collapsed alveoli There are multiple protocols for performing recruitment
maneuvers. Studies have shown that recruitment maneuvers are associated with a temporary
increase in oxygen levels but do not impact clinical outcomes. (Brower et al; Meade et al; Oczenski et
al). A more recent study of recruitment maneuvers in combination with best PEEP trials found an
association with increased mortality (Calvacanti et al). Recruitment maneuvers can increase
intrathoracic pressure enough to affect blood return to the heart and thus hemodynamics. We do
not recommend regular use of recruitment maneuvers in the management of refractory hypoxemia.
Oxygen Care
Initiation
Maintenance
If SpO2 is not available, one may consider empiric escalation of oxygen therapy for
tachypnea or increased work of breathing. However, these signs are weak surrogates for
SpO2 when titrating oxygen therapy.
1. Proning can be used with all types of oxygen delivery systems, or for patients on no
oxygen. For all oxygen systems, and non-invasive in particular, it is important to be
able to monitor the patient closely enough to ensure the risks of proning do not
outweigh the benefits.
3. If oxygen goals are not met by nasal cannula at <6 L/min then consider one of the
following:
1. Unlike simple and non-rebreather facemasks where you set the oxygen flow rate,
with Venturi masks you set the percent of oxygen (e.g. 40%). The percent of
oxygen is controlled using a valve attached to either the mask or the flowmeter.
First, select and attach the valve that corresponds to the correct FiO2 (or setting
the percent of oxygen if the valve is adjustable). The markings on the valve will
instruct you what flow rate to set. Because the valve blends pure oxygen with
room air, the actual flow delivered to the patient will be higher than the flow set
on the flowmeter.
4. If oxygen goals are still not met with the above options, consider escalation to the
following:
1. Non-rebreather facemask (at 10-15L/min, do not go below 10L or carbon dioxide can
be retained in the mask)
1. For patients with severe hypoxemia, some clinicians will place a non-rebreather
facemask on top of a nasal cannula. This is used when more intensive oxygen
delivery systems (e.g. high flow nasal cannula, non-invasive, and intubation) are
unavailable.
5. If oxygen goals are still not met, consider one of the options in the table below. The
clinical situation, availability of options at your institution, and goals of care discussions
should guide selection.
Institutional Factors:
1. Wean oxygen completely to off while monitoring at bedside with pulse oximetry, for at
least 5 minutes (unless the patient rapidly desaturates)
1. If oxygen saturation falls below SpO2 target (92% if no target specified), restart the
oxygen at the lowest flow rate necessary to meet the patient’s clinical (SpO2) goal.
2. If a patient maintains saturations above the clinical target without oxygen, oxygen
therapy may be discontinued.
2. Check oxygen saturation 30 minutes later and then again at 1 hour to ensure saturation
remains adequate without oxygen therapy.
For stable patients on simple, Venturi, or non-rebreather facemasks: attempt weaning at least once
a day by decreasing oxygen flow until goal oxygen saturation is met.
1. Minimum oxygen flow rates are required for non-rebreather face masks and Venturi face
masks to function properly, so do not decrease below the manufacturer recommended
flow. Switch to a lower intensity oxygen delivery device once a patient is stable on the
minimum flow rate for their current oxygen delivery device.
1. Simple facemask: Minimum flow rate is often 4 to 5L/min. At this setting, the next
step in oxygen weaning is to switch to nasal cannula at 5 to 6L/min.
2. Venturi facemask: Minimum flow rate depends on the oxygen concentration setting
(FiO2). In general, once a patient is stable on 40%, they are ready to attempt
switching to nasal cannula at 5 to 6 L/min.
3. Oxymizer: There is no minimum flow rate, but once a patient is stable on 4 to 5 L/min
they can be switched to nasal cannula at 5 to 6 L/min.
4. Non-rebreather Facemask: Minimum is often 10L per minute. At this setting, the next
step is simple facemask at 10 L/min , or, if a simple facemask is not available, nasal
cannula at 5 to 6 L/min.
The degree to which different oxygen delivery devices are thought to cause aerosolization remains
an area of active research, and the exact amount of aerosolization in each situation is not known.
Patient factors like coughing (which produces aerosols) and viral load, as well as the dynamics of
droplet particle size and dispersion, make this quite complex (Klompas et al). Meaningful distinction
between “safe” and “unsafe” levels of aerosols at this point is not possible. See Aerosols, Droplets,
and Fomites.
Flow rate: Lower oxygen flow rates hypothetically should reduce aerosols. However, a preprint
study in healthy volunteers showed that there was no variation in aerosol level between room air,
6L/min nasal cannula, 15 L/min non-rebreather, 30L/min high-flow nasal cannula and 60 L/min high-
flow nasal cannula regardless of coughing (Iwashyna et al).
This chart provides an overview, but is subject to change. Please follow your institution’s IPC
Practices regarding droplet or aerosol precautions. Aerosol Generating Procedures (AGPs) are
discussed here.
It is important to know the oxygen supply capability at your facility as well as the consumption rates
for different delivery devices. Depending on a facility’s oxygen supply type (liquid oxygen versus
cylinders versus an oxygen generating pressure swing adsorption plant) some oxygen delivery
devices may not be practical. Even relatively well-resourced facilities with liquid oxygen can exhaust
supplies when ramping up use of devices like high-flow nasal cannula during a surge census. The
tools above can help you determine which delivery device to use and the flow rate needed.
The values represent estimates of FiO2. Actual FiO2 delivered is dependent on multiple factors
including oxygen supply quality and patients minute ventilation. One general estimation rule is using
oxygen flow rate: FiO2 =0.21 + 0.03 x oxygen flow rate in L/min (Frat et al).
When available, high-flow nasal cannula is one option for selected patients for whom non-
rebreather or Venturi mask is not adequate to maintain goal oxygenation. While standard cannulas
and masks can provide flow rates of up to 15 liters per minute, an HFNC system delivers oxygen flow
rates as high as 80 L/min with variable concentrations of oxygen up to 100%.
In general, HFNC has been demonstrated as an effective intervention for management of acute
hypoxemic respiratory failure, improving survival (Frat et al) and reducing the need for mechanical
ventilation (Ferreyro et al).
Several small studies show COVID-19 patients might avoid intubation using high-flow nasal
cannula (HFNC) (Demoule et al). This is especially true with concomitant proning. (Tu et
al; Despres et al; Xu et al).
In one study of 293 COVID patients in South Africa, 47% percent were weaned off of HFNC
and did not require intubation (Calligaro et al).
A patient who is not meeting oxygenation goals on escalating therapies (e.g. facemask,
venturi mask, or non-rebreather) and does not meet criteria for intubation
A patient who is not meeting oxygenation goals despite maximal oxygen therapy AND
intubation is either not available or not within goals of care,
The patient does not need significant assistance with work of breathing or hypercapnia
(HFNC helps oxygenation but does not tend to help ventilation)
Contraindications:
High flow nasal cannula is often labeled as an aerosol generating procedure (perhaps better
stated as an aerosol enhancing device); However, data to support this notion or quantify risk
to healthcare workers remain evolving; Nonetheless, all patients on HFNC should be
required to wear a surgical mask over the cannula (Leung et al; Ferioli et al).
Heated and humidified oxygen must be used to avoid drying of mucous membranes and
secretions to prevent ciliary damage.
Start with lower flow rates if possible to minimize potential aerosols
Transport on HFNC is often not logistically possible, so conversion to non-rebreather is
recommended. In addition, non-rebreather may generate less aerosol.
Standard HFNC systems usually consist of a high capacity flow meter, an air-oxygen blender
(typically connected to wall air and oxygen sources), tubing, cannula, and a heater-
humidifier.
Some HFNC systems require a connection to high-pressure air in addition to high-pressure
oxygen sources. There are also several systems which do not require wall air and entrain
room air instead (either by Venturi effect or turbine)
Humidification: For optimal patient comfort and adherence, HFNC systems should deliver
gas to the patient at 44 mg H2O/L or 100% relative humidity (Spoletini et al; Restrepo et al).
HFNC consumes significant amounts of oxygen. For example, a patient receiving HFNC at 50
L/min and 0.8 FiO2 will consume approximately 37 L/min of oxygen and 13 L/min of air. A J-
type oxygen cylinder (1.45m height) contains 6800 liters of oxygen. At this rate, the cylinder
would last less than 3 hours.
Non-invasive Positive Pressure Ventilation (NIPPV)
When available, non-invasive positive pressure ventilation (e.g. CPAP, BiPAP) can be considered for
patients with the indications for which it would normally be used (e.g. OSA, COPD flare) whether or
not they have COVID.
In some institutions NIPPV is not a preferred method of delivering oxygen in worsening COVID-19-
related pneumonia/ARDS, though this is an area of active research and recommendations are often
changing. To see a summary of different guidance institutions recommendations, see our dashboard.
Some early studies indicate it may help avoid intubation, though mortality statistics remain
unknown: In one study of 47 patients about a third of patients treated with CPAP were able
to avoid intubation (Alviset et al). In another study of 53 patients, 83% were successfully
treated with NIPPV (Brusasco et al). Careful patient selection is likely important in
determining candidacy and ultimately success.
NIPPV may also be a way to help avoid ICU capacity overload and manage some patients on
the floor (Lawton et al).
Patients on NIPPV need to be closely monitored as high tidal volumes or work of breathing
may risk patient-induced lung injury in ARDS (Brochard).
Helmet NIV has been shown to be equivalent to high flow nasal cannula in moderate to
severe hypoxemia. Greico et al showed no significant difference in the number of days free
of respiratory support within 28 days, but did show decrease rate of endotracheal intubation
and number of days free of invasive ventilation in the helmet NIV group.
A patient has increased work of breathing or increasing pCO2 despite maximal oxygen
therapy (including HFNC if available) AND intubation is either not available, not within goals
of care, or not advised by the clinician caring for the patient.
Similar indications as in non-COVID-19 patients:
Contraindications:
NIPPV should be generally avoided in the same situations that NIPPV is avoided in COVID-19
negative patients (e.g., severe ARDS without short-term reversibility; the presence of
relative contraindications such as altered mental status, aspiration risk, secretions).
Some institutions require that NIPPV be done with aerosol precautions, others do not. This is
an area of active research.
Other Considerations:
Prolonged use of NIPPV has been linked to malnutrition and should be monitored (Turner et
al).
Prone Positioning
Benefits of Proning
Proning is thought to provide physiologic benefits for patients with COVID-19: It improves
recruitment of alveoli in dependent areas of the lungs and may improve perfusion to ventilated
areas, improving ventilation-perfusion mismatching. Typically proning is used in ventilated ICU
patients, however the same benefits may be found in non-ventilated patients.
Intubated Proning: Proning is one of the mainstays of ARDS therapy for intubated patients,
showing both 28 day and 90 day mortality benefit in the PROSEVA 2013 trial (Guerin). See
also Proning of Intubated Patients.
Self-proning (non-intubated) in non-COVID patients: Small non-COVID-19 patient cohorts
(ARDS, post-lung transplant, and post-surgery) showed association with lab, radiographic, or
clinical improvement.
Self-Proning in COVID-19. Multiple trials have shown benefit in oxygen with proning (Coppo
et al; Weatherald et al). A randomized trial of 1121 patients showed that awake proning
improved outcomes: the hazard ratio for intubation was 0.75 (0.62−0.91), and the HR for
mortality was 0.87 (0.68−1.11) compared with standard care. (Ehrmann et al).
Risks of Proning
Patient Selection
Self-proning can be used on stable patients (on room air or supplemental oxygen) and as a “rescue”
for those who have escalating supplemental O2 requirements.
Self-proning can be done with any type of oxygen delivery system with careful consideration
and ideally multidisciplinary discussion for safety. At higher levels of oxygen, the patient may
require more frequent monitoring (see protocol below).
The patient should ideally be able to move independently and have the cognitive and
physical status to supinate themselves if they become uncomfortable. This includes the
ability to safely manage their supplemental oxygen, IV tubing, SpO2 monitor and other leads
and attachments (within reason).
In certain situations, patients who are unable to position themselves may be candidates
for assisted proning as a “rescue” therapy. For example, assisted proning can be considered
if a patient has a low SpO2 (below 92%) on non-rebreather facemask, and HFNC, NIPPV, and
intubation are all unavailable or contraindicated (see below for additional considerations).
Contraindications:
Absolute:
Relative Contraindications:
1. Monitoring
2. Telemetry: If telemetry is indicated, EKG leads can remain on anterior chest wall for
continuous monitoring, avoiding pressure points.
2. Prior to Proning
1. Make plans in advance for toileting, contacting nurses, and cellular phone if patient
has one.
2. If possible, place the bed in reverse Trendelenburg (head above feet, 10 degrees) to
help reduce intraocular pressure.
3. Have patient empty bladder.
4. Educate the patient.
5. Arrange tubing to travel towards the top of the bed, not across the patient, to
minimize risk of dislodging. Ensure support devices are well-secured to the patient.
(Eg. sleeve over IV access site, position urinary catheter)
6. Assess pressure areas to avoid skin breakdown and dress any wounds and use skin
protective devices as needed.
7. If the patient will require assistance, assess the patient’s size and weight to determine
adequacy of the bed frame and the mattress in addition to the number of staff
required to safely turn the patient.
3. Prone Position
1. The patient should lay on their abdomen (arms at sides or in “swimmer” position). If
this is not a tolerable position, they can try laying on their side, though this may not
work as well (Bentley et al).
1. Show the patient how to choose which side to roll to so that they avoid any IV
tubing and how to adjust oxygen delivery device and pillows as needed
2. If a patient is unable to tolerate, they may rotate to lateral decubitus or partially prop
to the side (in between proning and lateral decubitus) using pillows or waffle
cushioning as needed. Ideally the patient should be fully proned rather than on the
side as there is currently no data about whether side positioning is beneficial.
1. Patient should try proning every 4 hrs and overnight, and stay proned as long as
tolerated (ideally at least 30 minutes). Our ideal goal is 16 hrs per 24 hours (e.g. 4
times for 4 hours each session) based on common interpretations of the PROSEVA
trial (Guerin). However, we realize that few (if any) patients will tolerate 16 hrs of
proning per 24 hrs.
1. We recommend continuing daily cycles of proning until the patient is on nasal cannula
(<4 L/min) with SpO2>92%. The patient may choose to stop self-proning at any time
2. Stop proning if any of the following occur: