Abstract
The effects of a high carbon dioxide on cerebral perfusion and intracranial pressure are well known. We report the case of a man who presented after with a severe traumatic brain injury including intracranial and extradural haemorrhage. Neuroprotective ventilation was impossible without supramaximal tidal volumes due to a combination of chest trauma and severe bronchospasm. A pump driven Novalung iLA active® system was inserted to achieve both ARDSnet ventilation and a lowering of intracranial pressure. To our knowledge, this is the first time this system has been used to this effect. The patient went on to make a good recovery.
Keywords
Case report
A 21-year-old male with a medical history of asthma was found unconscious with evidence of trauma to the head and face. On a primary survey at scene, paramedics recorded a Glasgow Coma Score (GCS) of 6 (motor 4, verbal 1 and eye opening 1) with a core temperature of 30℃. His nose was deformed and there was airway contamination with blood. He was transferred directly to the local major trauma centre.
On arrival at the major trauma centre, a repeat primary survey confirmed the presence of blood in the airway and demonstrated equal bilateral air entry with no evidence of chest wall injury. He had a respiratory rate of 12 with oxygen saturations of 92% on 15 L/min of inspired oxygen. His blood pressure was 140/100 with a heart rate of 100 bpm. He had a soft abdomen with no evidence of pelvic or long bone injury. He had full spinal immobilisation. Two intraosseous needles had been inserted on scene. In the emergency department, intravenous access was established by inserting a large bore pulmonary artery catheter introducer into the right subclavian vein.
Due to his GCS of 7 (motor 4, verbal 2, eye 1), a decision was made to intubate for airway protection, neuroprotection and to facilitate safe transfer to the computerized tomography (CT) scanner. A modified rapid sequence induction was performed with 200 mcg Fentanyl, 120 mg Ketamine and 80 mg Rocuronium. In line immobilisation was maintained during the procedure. Oxygen saturations fell rapidly to 80%. The C-MAC intubating video laryngoscope was used to visualise the airway. Due to copious, ongoing bleeding the view was difficult but the endotracheal tube was seen to pass through the vocal cords. After intubation no end tidal carbon dioxide (CO2) was seen on the monitor and the team were unable to ventilate the patient. An oesophageal intubation was suspected so the endotracheal tube was removed, and bag valve mask ventilation with a guedel airway was attempted. This was ineffective, producing no air entry or chest movement and his oxygen saturations fell significantly. He was reintubated using the C-MAC and the endotracheal tube was again seen to pass through the cords, however, ventilation was still impossible. The oxygen saturations fell quickly to 30%. Bilateral needle decompressions were performed followed by bilateral thoracostomies. Ventilation was still impossible. A size 6.0 endotracheal tube was inserted through the cricothyroid membrane and resulted in some air entry but ventilation remained very difficult due to high airway pressures. Five milligrams of 1:10,000 adrenaline was injected via the cricothroidostomy tube which did result in an apparent reduction in airway pressures, improved ventilation and oxygen saturations. At this point the difficulty with ventilation was thought to be due to severe bronchospasm secondary to blood and gastric contents. The team were unaware at this point that the patient was an asthmatic.
He was subsequently transferred to CT. CT scans (Figures 1–3) showed fractures of the frontal and parietal bones, extending inferiorly to involve the right temporal bone and petrous bone. A longitudinal fracture of the right petrous temporal bone was present extending into the anterior wall of the external auditory canal, crossing the right carotid canal and sphenoid sinus. There was a left sided subdural haematoma involving the left temporal, frontal and parietal regions with a maximum depth of 6 mm. There was also subdural blood bilaterally over the tentorium cerebelli. There was mass effect with effacement of the frontal and occipital horns of the left ventricle and midline shift of 6 mm. There was almost complete effacement of the cerebral sulci. Subarachnoid blood was seen within both Sylvian fissures and the sulci of the right frontal and parietal lobes. Multiple contusions were present involving the left temporal and left frontal lobes. The basal cisterns were effaced with the cerebral tonsils protruding into the foramen magnum. Chest CT demonstrated bilateral contusions in the left upper lobe, lingual, apical segment of right lower lobe and posterior segment of basal left lower lobe. Abdominal and pelvic CT were unremarkable.
Head CT demonstrating subarachnoid blood and effacement of the ventricles. Protrusion of the cerebellar tonsils. Demonstrating presence of pulmonary contusions, pneumothorax and chest drain.
Following the scan, he was moved to the general intensive care unit where invasive haemodynamic monitoring was established.
On the general intensive care unit a right frontal Camino intracranial pressure (ICP) bolt was inserted. The initial ICP reading was 35 mmHg despite deep sedation using morphine and midazolam combined with therapeutic paralysis. The high ICP was thought to be related to the high partial pressure of arterial carbon dioxide (PaCO2) of 6 kPa and presumed raised intra thoracic pressure. To achieve temporary stability, a bolus of 30 g of mannitol was given with good effect resulting in reduction of ICP to 22 mmHg. A bronchoscopy was performed. A mucus plug was suctioned from right upper lobe and gastric fluid suctioned from the left upper lobe and left lower lobe. ICP during the procedure increased to 25 mmHg but cerebral perfusion pressure was maintained above 60 mmHg with the aid of noradrenaline. At this stage, the PaCO2 had risen to 6.44 kPa causing a pH level of 7.19 despite tidal volumes of 10 mL/kg and peak airway pressures of 36 H2O. A decision was taken to use an extracorporeal carbon dioxide removal device to reduce his CO2 levels and thus, control the ICP.
A 22 Fr Novaport twin® cannula was inserted into the right femoral vein under direct ultrasound guidance. The cannula was connected to the pump driven Novalung iLA active® system platform. No anticoagulation was used for circuit priming or systemically due to concerns of worsening his intracranial injuries, however all blood contact surfaces of the device are pre-coated with heparin. Initial blood flow thorough the iLA active® was 1 L/min and the oxygen sweep gas was increased gradually from 1 to 10 L/min over the next 3 h. The PaCO2 dropped from 6.44 to 5.06 kPa with a corresponding fall in ICP from 26 to 7 mmHg. We simultaneously adopted a lung protective ventilation. His peak airway pressures fell from 36 to 30 cm H2O while reducing tidal volume (TV) from 10 to 6 mL/kg.
Change in titratable variables with time.
After a further 24 h, he was transferred to the neuro-intensive care unit for ongoing care. The ICP bolt was removed 19 days after insertion and he was transferred to his local district general hospital. After a short stay there he was discharged to his own home. Four months after admission he has made a good recovery and is returning to work.
Discussion
This case demonstrates the challenge of controlling the ventilation to manage severe head injury in the context of a significant lung injury. In this case, the situation was made more challenging by the underlying severe asthma aggravated by aspiration. The need to control ICP and maintain cerebral perfusion was made even more critical after the period of marked hypoxaemia around the time of securing the airway with potential for secondary injury to the brain.
It is now well established that invasive ventilation in patients with acute respiratory distress syndrome (ARDS) with lower (4–8 mL/kg) rather than a traditional (10–15 mL/kg) tidal volumes reduces mortality and increases ventilator free days. 1 The primary focus of the ARDSnet ventilation is to reduce the peak ventilatory pressures and reduce the risk of baro and volu-trauma to the lung whilst tolerating permissive hypercapnia. In this case, we were unable to tolerate hypercapnia. Elevated PaCO2 causes vasodilation, increased blood flow (a near doubling from physiological values to that at 11 kPa) and concomitant rise in intracranial pressure. Whilst controversy exists over the lack of a mortality benefit shown from insertion of an ICP bolt, there is also data that failing to control the ICP results in a greater in-hospital mortality. A systemic review of four studies in 2007 showed an ICP of between 20 and 40 mmHg was associated with a 3.5 odds ratio of death, increasing to 6.9 for an ICP of over 40 mmHg. 2
In this case, PaCO2 (and hence ICP) could not be controlled with standard ventilatory measures putting him at vastly increased risk of lung injury and ongoing inflammation.
It has been hypothesised that there exists a ‘double hit’ picture where a systemic inflammatory response syndrome exists as a result of a traumatic brain injury, increasing the likelihood of an acute lung injury even where lung protective ventilation measures were used. 3 An observational study of 137 patients over a 4-year period at a level one trauma centre in the United States found a mortality of 38% in patients with ALI compared to 15% in the non-ALI group (p = 0.004). 4
The solution we embarked upon, the Novalung iLA Active® was a therapy with which we had some experience, having used it for hypercapnic respiratory failure patients previously. There are several case reports involving the use of an extracorporeal membrane oxygenation (ECMO) circuit to treat respiratory failure in the context of a traumatic brain injury, some without the need for heparin anticoagulation.5–7 There are also reports of pumpless (i.e. an arterio-venous system reliant on the patient’s cardiac output for blood flow) extracorporeal circuits. A 2013 study in the Journal of Neurosurgical Anaesthesiology looked into the effect of these systems on brain perfusion in a series of pig models. They found that in these models, removal of CO2 with the pumpless system increased cardiac output by 44.9%. 8 None of these models had any simulated trauma and were observational only. As cardiac injuries are not uncommon in trauma patients it would appear unsuitable to use this device. As our primary problem was control of CO2 and ICP rather than oxygenation, ECMO was not required.
The Novalung iLA active® pumped extracorporeal circuit is a veno-venous system with a 22F cannula. The use of this therapy in the context of severe traumatic brain injury is to the best of our knowledge the first time a pumped extracorporeal carbon dioxide removal circuit has been used. The system is primed with a volume of 175 mL which in our experience causes minimal cardiovascular changes on commencement of therapy.
Commencement of the Novalung without anticoagulation was undertaken due to the nature of his injuries. The design of the Novalung is such that if blood clots form, they can be seen within the membrane, and the pressures across the membrane are measured allowing early cessation of therapy or replacement of the circuit if necessary. Whilst there has been a suggestion of the use of citrate anticoagulation for the extracorporeal anticoagulation 9 which makes use of the collective experience from renal replacement therapy, this was not something that we had experience of and retrospectively was not necessary.
Pumped veno-venous extracorporeal CO2 removal is a relatively new therapy but with the advent of a single venous canula and removal of the necessity of large bore arterial cannulation, the risks of this procedure approach those of haemofiltration. We expect to see a greater use of this therapy in the future. Increased familiarisation with the technology is being achieved; we are about to see the start of the REST trial 10 which is introducing a similar system into ‘at least 40 intensive care units’. We would suggest early employment of extracorporeal decarboxylation devices to reduce the risk of ventilator associated lung injury (VALI) whilst maintaining physiological CO2 in similar clinical circumstances.
Footnotes
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.