We describe three Covid-19 cases presenting with critical respiratory failure managed in the pre-hospital setting. The cause of hypoxemia in these patients is a matter of speculation, and several mechanisms have been suggested. Two recent autopsy studies have reported interstitial thickening and congested capillaries, making both diffusion failure and a low V/Q ratio secondary to dead space ventilation plausible mechanisms4, 5. However, CT imaging studies also report consolidations in the majority of patients with Covid-19 pneumonia6. Consolidations may accordingly contribute to a substantial shunt fraction, which explains the failure of oxygen therapy in some of these patients. Prone positioning has been reported to improve oxygenation in COVID-19 patients7, lending support to the notion that pulmonary shunting is a contributing factor to hypoxemia. Furthermore, Gatinoni et al. postulated dysregulation of pulmonary vascular tone with loss of hypoxic vasoconstriction and potentially increased pulmonary shunting8. Covid-19 patients are reported to have near normal lung compliance and consequently they can maintain large minute volumes for a prolonged period of time without succumbing to exhaustion.
Carbon dioxide (CO2) diffuses through tissues about twenty times more rapidly compared to oxygen (O2), and these properties likely underlie the disproportional pulmonary exchange of CO2 and O2 in these patients. The increased minute ventilation in these patients might be a sign of progressive respiratory failure with limited diffusion and increased dead space ventilation. Normal to increased end-tidal CO2 in hypoxemic patients, might therefore be a sign of exhaustion or imminent respiratory collapse.
Despite being severely hypoxemic all three patients described in this case series were responsive and cooperative upon presentation. All cases presented with low pulse oximetry values in the 35% – 60% range. However, how well this reflects arterial oxygen saturation (SaO2) is uncertain since pulse oximetry is not validated for SaO2 values below 70% and tends to overestimate hypoxemia below this threshold9.
Early intubation of Covid-19 patients have been recommended by some institutions, however this strategy has recently been challenged10. Anecdotal evidence suggest that favorable outcome can be achieved with NIV support even in patients with severe respiratory failure and persistent hypoxemia11, and that NIV may reduce the need for intubation12. Several factors can contribute to a favorable outcome with NIV. First, as these patients seem to have near normal lung compliance, they do not necessarily have increased work of breathing implying that they can endure long periods with NIV support, without the need for heavy sedation. Second, prone positioning has been reported to be effective in Covid-19 and this can be easier to accomplish in awake patients on NIV compared to intubated patients which demands heavy sedation7. Further, profound hypoxemia, hypotension and even cardiac arrest have been reported in a significant number of Covid-19 patients during induction of anesthesia and tracheal intubation13. We argue that the decision to intubate Covid-19 patients in the pre-hospital setting should rely on a careful risk-benefit analysis, and one should consider NIV in patients with severe hypoxemia. We speculate that pre-hospital intubation should only be sought in hypoxemic Covid-19 patients with clear signs of decompensation, i.e. altered level of consciousness, increased work of breathing with apparent use of accessory respiratory muscles, or circulatory compromise.
The Covid-19 patients reported in this paper responded poorly to oxygen therapy and due to their high minute ventilation high oxygen flow rates were required. With the delivery systems applied in the pre-hospital context, oxygen flow rates are often limited to 15 liters/minute. Non-rebreather masks and CPAP systems will dilute oxygen with a highly variable volume of ambient air, depending on the patient´s minute ventilation, oxygen flow rate and degree of mask seal. In a study in healthy volunteers the FiO2 obtained with a non-rebreather mask with oxygen flow of 15 liters/minute was approximately 0.6 and fell rapidly with increasing respiratory rate and tidal volume14. The FiO2 delivered via a BVM is variable depending on the manufacturer. One study reported a FiO2 in the range of 0.43–0.54 at flow rates of 10 and 15 liters/minute, but at higher O2 flow rates an FiO2 above 0.9 was achieved15. We would like to highlight two factors which must be taken into consideration in order to deliver a high FiO2 with the BMV. First, a tight seal between the patients face and the mask is necessary in order to avoid ambient air being drawn in during inspiration. Second, the oxygen flow rate must exceed the patient´s minute ventilation. If not, the reservoir bag collapses, and ambient air will be pulled into the bag diluting the oxygen and reducing the FiO2. This is an important point in the context of COVID-19 patients as they often have a high minute ventilation and thus require a high oxygen flow rate to achieve a high FiO2. A PEEP valve may be connected to the BVM, however we would like to stress that a BVM with a PEEP-valve is not able to generate a continuous positive airway pressure. The PEEP-valve only applies positive airway pressure during expiration and not continuously throughout the respiratory cycle. Lastly, attaching a viral filter to the BVM will cause increased airway resistance during inspiration compared to spontaneous breathing with a non-rebreather. This increased airway resistance will increase the work of breathing and may cause atelectasis. Assisting the patient’s breathing by pressing the bag during inspiration may compensate for this and may even facilitate lung recruitment. However, doing this synchronously with the patient’s spontaneous effort requires practice and focus and is especially challenging in patients with a high respiratory rate.
Hypoxemic patients may fail to respond adequately to supplemental oxygen despite a high FiO2, reflecting a considerable shunt fraction or compromised diffusion. Reducing the shunt fraction can be achieved by recruiting non-ventilated lung volume or by facilitating redistribution of perfusion to areas of the lung with better ventilation. Several non-invasive interventions can be applied in the pre-hospital context in order to reduce shunt fraction.
In an in-hospital setting CPAP has shown to be effective in improving hypoxemia in Covid-19 patients11, however the time required from onset of CPAP-treatment to an increase in oxygen saturation has not been specifically studied in these patients. Unfortunately, disposable CPAP systems generally fail to deliver a high FiO2, and we speculate that this contributes to the ineffective use of CPAP in one of our patients. Adding extra oxygen via a nasal cannula underneath the CPAP may improve the FiO2, however this set up requires two oxygen sources. A transport ventilator in CPAP mode could permit a FiO2 close to 1.0 as these ventilators draw oxygen to match the patients inspiratory flow rate without the need for ambient air. Additionally, a mechanical ventilator may sustain a more stable positive airway pressure. This could also be used as a bridge to intubation in patients with critical hypoxemia requiring optimal preoxygenation.
Prone position has been reported to improve outcome in intubated patients with acute respiratory distress syndrome (ARDS)16. In awake patients the evidence is scarce, possibly reflecting the less severe nature of ARDS in non-intubated patients. Prone position can theoretically improve oxygenation via several mechanisms. Perhaps the most immediate effect is due to redistribution of blood from atelectatic dorsal areas of the lungs to less atelectatic ventral lung tissue. Proning may also help in recruitment of dorsal atelectasis and help in mobilization of secretions and thereby preventing further atelectasis. Several studies have reported that Covid-19 patients have dorsal consolidations whereas the ventral parts of the lungs are less affected6, 17. This may explain why many Covid-19 patients respond favourably to prone positioning. Caputo et al reported the effects of awake prone position in 50 patients with confirmed Covid-19 presenting to the emergency department7. A majority showed improvement in oxygenation within 5 minutes. Awake prone positioning appears to be a safe and effective intervention in cooperative hypoxemic Covid-19 patients when done in an in-hospital setting11. It requires little or no specialized equipment and little extra resources. Awake prone positioning for hypoxemic Covid-19 patients has already been implemented in some EMS protocols in the United States18.
A pilot study from 2004 reported that inhaled nitric oxide had an immediate effect on oxygenation in spontaneously breathing SARS patients with hypoxemic respiratory failure19. A randomized controlled trial investigating the effects of inhaled nitric oxide in mechanically ventilated patients with Covid-19 is underway20. Administering pulmonary vasodilators like inhaled nitric oxide drugs is logistically challenging in a pre-hospital setting. In situations with critical hypoxemia were the appropriate equipment is readily available this could be an option.
The risks of a pre-hospital intubation may outweigh the benefits in cooperative and alert patients despite severe hypoxemia. Although early intubation was initially recommended in Covid-19 patients that were hypoxemic despite supplemental oxygen, WHO guidelines now suggest a trial of NIV before proceeding to invasive mechanical ventilation in selected patients21. There are many possible explanations why intubating these patients is considered high risk. Initially, several guidelines advised against mask ventilating these patients through the apnoeic period prior to laryngoscopy due to the risk of aerosolization. This may however lead to an unnecessary long apnea time potentially causing a critical drop in oxygen saturation. In line with this, some international guidelines now recommend mask ventilation through the apneic period22. Another possible mechanism by which apnea leads to a rapid deterioration in oxygen saturation is the effect of increased PaCO2 on haemoglobin oxygen affinity. In Covid-19 patients with a high minute ventilation, hypocapnic hypoxia causes a left shift in the oxygen dissociation curve implying that at a given PaO2 the hypocapnic patient has a higher oxygen saturation than the normocapnic patient. If the PaCO2 suddenly rises due to apnea this may cause a reduction in haemoglobin oxygen affinity at the level of the lung thereby leading to a sudden drop in SaO223. A sudden drop in SaO2 to a critically low level accompanied by a simultaneous drop in coronary perfusion pressure secondary to the vasodilatory and cardiodepressive effects of the anesthetic drugs, may contribute to hemodynamic collapse.