The successful use of extracorporeal carbon dioxide removal as a rescue therapy in a patient with severe COVID‐19 pneumonitis

Summary

We present a patient with severe COVID‐19 pneumonitis; poor respiratory compliance; dangerously high ventilator pressures; and hypercapnia refractory to conventional treatment including low tidal volume ventilation, neuromuscular blockade and prone position ventilation. Extracorporeal carbon dioxide removal was used as a rescue therapy to facilitate safer ventilator pressures and arterial partial pressures of carbon dioxide. After 6 days of treatment, the patient had improved to the extent that the extracorporeal support was able to be weaned and the patient was decannulated from the device. Following a prolonged respiratory wean, the patient was subsequently discharged from the intensive care unit and then from the hospital to home with no adverse events related to the therapy.

Introduction

Coronavirus disease 2019 (COVID‐19) was first identified in Wuhan, Hubei province, China from a cluster of 27 cases of pneumonia of unknown aetiology in patients connected to the open seafood market. Despite lockdown measures it rapidly spread and a global pandemic was announced by the World Health Organization on 11 March 2019 [1]. Although the majority of patients suffer a mild illness or are asymptomatic, around 5% require intensive care and 2.3% go on to develop severe respiratory failure from acute respiratory distress syndrome (ARDS) requiring ventilatory support [2]. Acute respiratory distress syndrome is defined by the Berlin criteria with the presence of bilateral pulmonary infiltrates, not wholly explained by left ventricular failure, with impaired oxygenation and a PaO₂/FiO₂ ratio of < 40 KPa [3]. Although novel treatments for COVID‐19 are being investigated, a reasonable approach in the interim is to adopt strategies previously found efficacious in ARDS. The mainstay of treatment of ARDS is to prevent further damage or ventilator‐induced lung injury by using lung protective ventilation with tidal volumes of 6 ml.kg⁻¹, plateau pressure < 30 cmH₂O, driving pressure < 15 cmH₂O. Other measures recommended include higher positive end‐expiratory pressure; prone position ventilation; neuromuscular blockade; and extracorporeal membrane oxygenation (ECMO) for refractory cases [4]. Although there has been discussion as to whether COVID‐19 pneumonitis represents truly ‘classical’ ARDS with poor compliance or whether highly compliant phenotypes exist [5], it is evident from the initial ARDSnet paper that compliance was variable in their cohort and a quarter of their patients had plateau pressure between 10 and 20 cmH₂O which would be considered compliant [6]. With limited evidence from COVID‐specific cohorts, it therefore seems reasonable to use those strategies that are already recommended in ARDS. Indeed, the treatment strategy we used at our institution during the pandemic was to draw on our previous experience of treating ARDS and use these evidence‐based methods.

Report

A 48‐year‐old man with a history of asthma and hypercholesterolaemia was admitted to our institution with 7 days of productive cough, fever and shortness of breath. He had a chest x‐ray demonstrating bilateral infiltrates and a respiratory viral swab was subsequently positive for COVID‐19. He was initially managed with oxygen via facemask and oral doxycycline to treat possible co‐existent bacterial pneumonia. Due to evolving respiratory failure, he was trialled on continuous positive airways pressure (CPAP) and despite escalating treatment to 100% FiO₂ with a pressure of 15 cmH₂O he did not improve. On day four, he was taken to the intensive care unit (ICU) and his trachea was intubated to facilitate mechanical ventilation. The ICU at our facility offers standard level 3 care and advanced therapies such as ECMO are only available at regional centres. Over the subsequent week, the patient's condition progressively worsened and was consistent with ARDS from COVID‐19–pneumonitis. Various strategies for treatment of severe ARDS were trailed including 6 ml.kg⁻¹ tidal volumes; multiple episodes of prone position ventilation; neuromuscular blockade; high positive end‐expiratory pressure; recruitment manoeuvres; and continuous inhaled prostacycline analogue. The treating team felt despite maximal treatment, the patient was deteriorating as evidenced by progressive chest x‐ray infiltrates, ongoing hypercapnic respiratory failure with PaO₂/FiO₂ ratio of 8, PCO₂ of 12.4 kPa and poor compliance with peak airway pressure of 35 cmH₂O and a driving pressure of 23 cmH₂O. He was therefore referred to the local ECMO centre for refractory ARDS, but the referral was declined on the basis of futility. The treating team felt that mechanical ventilation with dangerously high airway pressure was contributing to the progression of his ARDS and therefore considered extracorporeal carbon dioxide removal as a rescue therapy. This was initiated uneventfully using the Hemolung RAS device (ALung Technologies Inc., Pittsburgh, USA) with a 15.5‐Fr catheter inserted in the right jugular vein and heparin infusion. The sweep gas flow was titrated up to 10 l.min⁻¹ and blood flow to 510 ml.min⁻¹ resulting in a peak carbon dioxide removal of 136 ml.min⁻¹. The patient's blood gases this improved the PCO₂ from 12.4 to 8.8 kPa and allowed both tidal volume and respiratory rate to be reduced, resulting in the peak airway pressure falling from 35 to 28 cmH₂O and the driving pressure from 23 to 16 cmH₂O. Given the severity and trajectory of the patient's condition, a course of methylprednisolone was also commenced. The patient spent 6 days on the Hemolung without bleeding events or vasopressor requirement and could continue to receive prone position ventilation without complication. On the sixth day, the patient improved to a point where spontaneous ventilation was restored and the Hemolung was weaned off. A slow respiratory wean followed and a percutaneous tracheostomy was performed. The patient was eventually discharged from intensive care following decannulation. His intensive care stay was 37 days and he was discharged home 4 days later with mobility and cognition intact.

Discussion

This report demonstrates the use of extracorporeal carbon dioxide removal as a rescue therapy for patients with hypercapnic respiratory failure as a result of ARDS in COVID‐19 pneumonitis. The Hemolung device allowed airway pressure to be reduced from dangerously high levels to facilitate lung protective ventilation. Further ventilation at such elevated airway pressures would likely have been deleterious to the patient and propagated the inflammatory process contributing to ARDS. Had standard ventilator strategies alone been continued, the patient's respiratory failure might have progressed and likely been non‐survivable, especially considering that ECMO had been ruled out.

Previous work on extracorporeal carbon dioxide removal has been regarding its use to facilitate ultra‐lung protective ventilation with tidal volumes <6 ml.kg⁻¹. The rationale behind this is that even with lung protective ventilation according to ARDSnet protocol where plateau pressure is < 30 cm H₂O, tidal volumes are 6–8 ml.kg⁻¹ and driving pressure is < 15 cmH₂O, the processes by which ventilator‐induced lung injury occur are still ongoing. Extracorporeal carbon dioxide removal allows a reduction in all of these parameters by reducing the amount of carbon dioxide needing to be removed by ventilation and therefore reducing the mechanical forces to which the lung is submitted. Terragni et al. showed in a single‐centre study that it was possible to reduce tidal volumes below 6 ml.kg⁻¹ and remove the resulting excess carbon dioxide with a modified dialysis circuit while reducing markers of lung inflammation in the process [7]. Combes et al. conducted a multicentre trial of ARDS patients that built on the experience of Terragni et al. and showed the feasibility of ultra‐low tidal volume ventilation facilitated by extracorporeal carbon dioxide removal with novel commercially available devices. The investigators noted however that the serious adverse event rate (6/95 patients) approached the rate of reduction in mortality from the original ARDSnet paper and that a randomised controlled trial was warranted [8]. The results of a study by McNamee et al. [9], which is a randomised control trial examining the use of extracorporeal carbon dioxide removal to facilitate ultra‐lung protective ventilation, are awaited. Pending this, evidence for routine use of extracorporeal carbon dioxide removal to facilitate ultra‐lung protective ventilation is lacking.

This report highlights a different indication for extracorporeal carbon dioxide removal than in the previously outlined and that is as a rescue therapy where the treating clinicians felt that the severity of the respiratory acidaemia and dangerously high ventilator pressures meant that patient survival without intervention was unlikely. Ideally, this indication would be studied in a randomised control trial. Despite limited evidence for this indication, refractory hypercapnia is already an accepted indication for ECMO in ARDS [10]. It therefore seems reasonable for clinicians to consider the use of extracorporeal carbon dioxide removal as a rescue therapy in severe ARDS with high ventilator pressures and refractory hypercapnia, either in preference to ECMO or if ECMO is not deemed appropriate or cannot be delivered due to resource constraints particularly during times of increased resource uptake such as during a viral pandemic. Potential advantages of extracorporeal carbon dioxide removal over ECMO include lack of requirement for transfer to an ECMO centre; smaller catheter size; and lower blood flow rate which may reduce the likelihood of complications. Extracorporeal carbon dioxide removal is now considered in the authors' institution if the arterial pH is unable to be maintained at 7.15 or above due to hypercapnia using lung protective tidal volumes and pressures.