Artificial lung as an alternative to mechanical ventilation in COPD exacerbation

To the Editors:

Acute exacerbation of chronic obstructive pulmonary disease (COPD) is commonly treated with different kinds of noninvasive positive pressure devices, ranging from helmet or face-mask continuous positive airway pressure (CPAP) to noninvasive pressure support ventilation (NPPV), or Bi-PAP [1]. The use of positive end-expiratory pressure (PEEP) and NPPV often results in the successful treatment of COPD patients with respiratory distress [1, 2]. If, despite maximal medical management, respiratory distress and gas exchange deteriorate with increasing tachypnoea and acidosis, and with altered level of consciousness, then tracheal intubation and mechanical ventilation (MV) become mandatory [3]. However, tracheal intubation and MV have several detrimental side-effects that may concur to determine the high morbidity and mortality reported in COPD patients requiring them [3, 4]. Indeed, the increase in airway resistance, the prolonged time required for lung emptying and the resulting dynamic hyperinflation, named auto-PEEP [4], are the most important physiological alterations during COPD exacerbation. In this condition, the application of MV could increase lung hyperinflation and lead to barotrauma and circulatory failure [4, 5]. Furthermore, tracheal intubation is usually associated with the need for sedation. The side-effects of intubation, sedation and MV may initiate a vicious circle, often resulting in a very difficult or impossible weaning.

In this report, we describe a patient with severe exacerbation of COPD, in whom after failure of noninvasive ventilation, we decided to treat the respiratory acidosis, tachypnoea and ventilatory fatigue by removing CO₂ with an artificial lung. This avoided the need for tracheal intubation and MV, leaving the patient in spontaneous breathing.

In September 2010, a 72-yr-old, female heavy smoker with a history of COPD (forced expiratory volume in 1 s (FEV₁) 50% predicted, forced vital capacity (FVC) 94% and Tiffeneau index 41%) arrived at the emergency department with an acute COPD exacerbation. She had acute respiratory failure with severe respiratory acidosis (pH 7.27, carbon dioxide tension 79 mmHg) and tachypnoea (40 breaths·min⁻¹), without any other organ dysfunction. A chest radiograph showed lung hyperinflation without signs of pneumonia. The patient was given standard medical treatment (antibiotics and bronchodilators) and steroids. After 96 h of NPPV, her respiratory acidosis further deteriorated (before extracorporeal membrane oxygenation; pre-ECMO in table 1). Tracheal intubation with MV was considered, but, as an alternative, we proposed to the patient and her son the use of an artificial lung, informing them of the possible advantages (and risks) of this approach, including the relief of dyspnoea, the natural lung deflation, the maintenance of consciousness and spontaneous breathing, without tracheal intubation. Thereafter, we applied a veno-venous bypass by a percutaneous cannulation of both femoral veins. The procedure was performed under mild sedation and local anaesthesia. The extracorporeal system consisted of the Permanent Life Support Set, “Bioline Coating”, Maquet BE-PLS 2050® (Rotaflow RF 32 Pump and Quadrox PLS Oxygenator; Maquet GmbH, Rastatt, Germany).

The extracorporeal blood flow was set at 2 L·min⁻¹ and maintained stable, while the gas flow, which is the main determinant of the extracorporeal CO₂ removal, was increased slowly (starting from 1 L·min⁻¹) to avoid huge and sudden alteration of the acid–base status (table 1). The patient was nonintubated and breathing a humidified air–oxygen mixture to maintain normal arterial oxygen tension (P_a,O2; table 1). After 48 h, the patient’s metabolic CO₂ production was almost totally cleared through the membrane lung (4 L·min⁻¹ gas flow) (fig. 1a). With this setting, the artificial CO₂ removal almost abolished the need to clear the CO₂ through the natural lung. In fact, the patient's dyspnoea was relieved shortly after starting the bypass, and, when total extracorporeal CO₂ removal was reached, the respiratory rate decreased from 32 to 8 breaths·min⁻¹ and the pleural pressure, as measured by tidal oesophageal pressure swings, an index of work of breathing, completely normalised (fig. 1b). Radiographic measurements on chest radiographs confirmed the progressive resolution of the dynamic hyperinflation (table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1–

a) Percentage CO₂ clearance, performed by the artificial lung and the estimated CO₂ clearance of the natural lung as a function of the bypass time.b) Respiratory frequency f_R (▪) and tidal oesophageal pressure variations (▴) as a function of bypass time. The above variables were unavailable during the first 24 h because the patient clinical conditions did not allow insertion of invasive monitoring.

We therefore began to slowly reduce the extracorporeal CO₂ removal, allowing the patient to increase her minute ventilation, while monitoring respiratory rate, respiratory pleural pressure variations and breathing coordination. On day 6, we removed the artificial CO₂ removal and the patient showed an acceptable respiratory pattern and gas exchange without clinical and radiological signs of hyperinflation (table 1). After a few hours, the extracorporeal circuit was removed. The patient was discharged from the intensive care unit 1 day later, and at 6 months follow-up she was well and had no need of additional oxygen.

Since 1980, extracorporeal CO₂ removal has been proposed to provide lung rest by reducing mechanical ventilation load in acute respiratory distress syndrome [7]; it has also been used, in combination with tracheal intubation and MV, to resolve acute asthma attacks [8] and dynamic hyperinflation [9]. Recently, this technique has been used in spontaneously breathing patients with the purpose of bridging to lung transplantation [10, 11].

We report herein the successful use of the extracorporeal CO₂ removal in a nonintubated, spontaneously breathing patient to treat acute respiratory failure and dynamic hyperinflation during COPD exacerbation. Our primary aim was to aid recovery from dynamic hyperinflation while avoiding tracheal intubation and MV with its associated side-effects, such as the increase in hyperinflation, barotrauma, ventilatory-acquired pneumonia, critically polyneuropathy illness and difficult weaning [2, 3]. The improvement of the extracorporeal biocompatible materials and circuit characteristics (i.e. new centrifugal pumps, high-performance artificial lungs and heparin-coated circuits) reduces the risks (mainly haemorrhage) of this invasive technique, which could become a possible alternative to conventional mechanical ventilation.

Starting the extracorporeal CO₂ removal in this awake and spontaneously breathing patient almost immediately produced a new clinical scenario. The need of breathing was sharply reduced, proportionally to the CO₂ cleared by the membrane lung, the respiratory muscles were at rest and the lungs deflated naturally. Despite her severe respiratory condition, the patient was able to communicate, drink and eat spontaneously. In this condition, the relationship between the patient and the intensive care unit team (medical doctors and nurses) had to be completely modified, since the patient's needs were totally different from those of sedated mechanically ventilated patients.

In this awake nonintubated patient, the approach to the monitoring of the usual respiratory parameters was particularly difficult. In fact, tidal volume, FEV₁ and FVC measurements were attempted using a mouthpiece spirometer, but were too affected by patient cooperation and ventilatory fatigue to speculate on. Furthermore, auto-PEEP and mixed expired CO₂ could not be directly measured. We recorded the patient oxygen consumption and the extracorporeal oxygen supply using the Fick principle, the extracorporeal CO₂ clearance, measuring the CO₂ concentration in the sweep gas flow, and derived the patient CO₂ clearance assuming a respiratory quotient equal to 1. We also recorded the blood gas parameters, the respiratory mechanical coordination, the modified Borg dyspnoea scale and the intra-tidal oesophageal pressure swings, as an index of the work of breathing. Although some of these parameters were approximate, we believe that their comprehensive evaluation was very useful to understand the mechanisms involved in the vicious circle of dynamic hyperinflation and to guide the appropriate use of the extracorporeal CO₂ removal. Daily chest radiographs helped us to confirm the recovery from dynamic hyperinflation. Indeed, more accurate and precise noninvasive methods to monitor a spontaneous breathing patient during extracorporeal treatment would be highly desirable.

In conclusion, in patients with acute severe COPD exacerbation who fail the noninvasive ventilation approach, the use of veno-venous extracorporeal CO₂ removal during spontaneous breathing could be a successful strategy to reduce dynamic hyperinflation and to avoid the detrimental effects of tracheal intubation and mechanical ventilation.

• ©ERS 2012