Anaesthetic management for endobronchial valve insertion: lessons learned from a single centre retrospective series and a literature review.

BACKGROUND
Endoscopic lung volume reduction using one or more endobronchial valves is a treatment option for a select group of patients with severe emphysema. Patients presenting for this procedure pose various challenges to the anaesthetist; in addition to their lung condition, they are often elderly with multiple comorbidities. The procedure is usually performed outside the operating room. Monitored anaesthesia care with intravenous sedation, and general anaesthesia with an endotracheal tube have both been described for these procedures, aiming for adequate ventilation and haemodynamic stability.


METHODS
We present our experience on 20 of these procedures in relation to the anaesthetic techniques employed and discuss the perioperative challenges involved in managing these cases.


RESULTS
Twenty one planned endobronchial valve insertion procedures were identified on 18 patients. There were ten cases of monitored anaesthesia care with sedation and 10 cases which used general anaesthesia with an endotracheal tube. Two have been excluded; one had features of anaphylaxis and the procedure was abandoned, and the other required conversion from monitored anaesthesia care to general anaesthesia with endotracheal tube.


CONCLUSIONS
Both monitored anaesthesia care with sedation and general anaesthesia with endotracheal tube were well tolerated during endobronchial valve insertion procedures. General anaesthesia with endotracheal tube may offer better interventional conditions, patient comfort and reduced anaesthetic time.


Background
Endoscopic lung volume reduction (ELVR) using an endobronchial valve (EBV) is a treatment option for selected patients with severe emphysema [1,2]. EBV uses a fibreoptic bronchoscope to implant one or more one-way valves into the airways of the most emphysematous part of the lung [2,3]. The valve blocks the inspired air from entering the target lobe but allows air and secretions to leave during expiration, leading to lobar collapse and hence diminishing gas trapping (Figs. 1, 2 and 3) [2,3].
Decreasing the volume of the affected region may increase overall lung elasticity, and promote healthier regions to expand and function more effectively.
The advanced age, and concurrent comorbidities of the patients, and the frequent use of a "remote location" such as a bronchoscopy suite, provide many challenges for anesthetists managing these cases. The anaesthetic requirements are: to provide good interventional bronchoscopic conditions; to facilitate patient comfort and minimise movement (during sedation); to maintain adequate ventilation whilst avoiding hypercapnia, and preventing air trapping (during general anaesthesia); to maintain haemodynamic stability; to diagnose and manage acute perioperative complications; and (ideally) allow an early recovery. General anaesthesia with endotracheal intubation (GA via ETT), and monitored anaesthesia care (MAC) using sedation together with airway topicalisation have both been described for these procedures [3][4][5][6]. Nonetheless, EBV is still a novel technique, and literature discussing the anaesthetic implications of this procedure is limited. We describe the anaesthetic management of 20 procedures of EBV implantation in 17 patients as a retrospective review, with the objective of analysing the anaesthetic techniques and complications encountered during and after this procedure.

Methods
After ethics committee approval from the Central Adelaide Local Health Network (Approval No: Q20170604), we performed a retrospective chart review (manual and electronic records) of patients who underwent EBV implantation procedures between April 2015 and December 2017 at The Queen Elizabeth Hospital, South Australia. Perioperative medical records and discharge summaries were analysed. Patients' demographic profile including age, gender, weight, American Society of Anesthesiologists (ASA) physical status were recorded. Preoperative pulmonary function tests and arterial blood gas analyses during the procedure were also reviewed.
The following were also recorded: anaesthetic data during the procedure, including the type of anaesthesia administered (MAC or GA via ETT); details of airway topicalisation; medications used for sedation and general anaesthesia; methods to secure the airway; duration of the procedure; and total anaesthetic time. In addition, perioperative and postoperative complications relating to anaesthesia or the procedure, such as haemodynamic and respiratory instability (eg. pneumothorax), and length of hospital stay were also documented.

Anaesthetic techniques
This is a new procedure and not frequently performed. It is not surprising that each anaesthetist employs their own favourite technique (and even this may change with the experience of previous cases). Broadly the techniques used in our institution fall into two groups; general anaesthesia with a tracheally intubated patient, and 'monitored anaesthesia care' with a sedated unintubated patient.
Dexmedetomidine using a bolus followed by infusion, was used in eight cases of MAC along with midazolam and fentanyl. In nine cases, glycopyrrolate was administered prior to commencing dexmedetomidine and six had airway topicalisation. Total intravenous anaesthesia (TIVA) with propofol was used in all GA via ETT cases. In 8 out of 10 of the GA via ETT cases, combinations of propofol, remifentanil and rocuronium were used. The   details of the cases analysed presented in Fig. 4. Sugammadex was used as the reversal agent in 9 of the 10 GA cases. While the GA cases were positioned supine, MAC cases were positioned 15-20 degrees head up.

EBV placement procedure
The technical aspects of this procedure from a pulmonology perspective have been extensively reviewed recently by two groups [1,2]. The patients are selected according to generally agreed spirometry criteria, namely hyperinflation with a residual volume > 175% of predicted, and forced expiratory volume < 50% of predicted [1]. The efficacy of this technique depends on the absence of collateral ventilation between the lobe with implanted valves and the ipsilateral lobe [1]. Precise measurement of the amount of collateral ventilation is performed using endobronchial flow catheters. The Chartis Diagnostic System (Pulmonx, Redwood City, California, USA) is one such system [1]. This is an endobronchial flow catheter and associated computer software (Fig. 5). It is generally recommended that both the Chartis measurement and EBV implantation be performed as a single-stage procedure whenever possible, to minimize the risk of bronchoscopy induced COPD exacerbations [1]. The Chartis system comprises a balloon catheter attached to a sensor remotely housed in a console. The catheter is passed through the working channel of the bronchoscope; the main bronchus of the target lobe is blocked by inflating the balloon; and collateral ventilation is assessed by measuring the expiratory flow over time (Fig. 6). If no collateral flow is observed, a flexible delivery catheter is then used to deploy the EBV into the targeted bronchial lumen. An inbuilt measurement gauge in the catheter helps in choosing the appropriate size of the valve for the target lumen [7]. In our review, all the procedures were done by the same interventional pulmonologist, while different consultant anaesthetists participated in the anaesthetic management. An arterial line was inserted in all patients and serial blood gas analyses were performed.

Results
A total of 22 EBV planned insertion procedures were identified on 19 patients. Three patients underwent a second procedure as a valve replacement technique.
The procedure had to be abandoned in one patient who manifested severe haemodynamic instability mimicking an anaphylaxis. In one instance, there was a conversion from MAC to general anaesthesia. This patient had a refractory cough and the procedure could not be performed using MAC. (The case was excluded from the analysis). The mean preoperative FEV1 was 0.78 ± 0.26 (range: 0.43-1.62) litres and the mean residual volume was 178% of predicted ±23.49 (range: 143-212) ( Table 1). All were ASA status III with multiple co-morbidities. Emphysema was the most common indication. Fifteen patients had the Chartis endobronchial evaluation for collateral ventilation performed immediately prior to the valve insertion. A Zephyr valve (Pulmonx) was used in all cases and the procedures were performed in a dedicated bronchoscopy suite located outside the main operating room complex. Although the average procedure duration was similar between the MAC (24.90 min) and GA via ETT (20.70 min) group, the total anaesthetic time was much longer for the MAC group (61 min vs 37 min). Of the analysed 20 patients,  10 patients had MAC and 10 had GA with ETT. All GA via ETT cases were extubated at the end of the procedure (Fig. 4).
All values of PaCO 2 which were recorded in the patients' case notes are shown in Fig. 7. The mean intraoperative rise in PaCO 2 was 11 mmHg. By 2 h, the PaCO 2 returned to near baseline values.

Intraoperative complications
Mild hypotension was observed in 5 patients. The hypotension was treated with a single bolus of metaraminol in 4 cases and a bolus of ephedrine in 1 case. Intravenous infusions of inotropes or vasopressors were not required. One case of profound hypotension -thought likely to be anaphylaxis -occurred in the GA via ETT group, and the procedure was abandoned. There was one case of inadequate topicalisation. All patients were discharged to the respiratory high dependency unit.

Postoperative complications
Postoperative complications were encountered in seven patients. One episode of rapid AF occurred on extubation. There were exacerbations of chronic obstructive pulmonary disease in three instances, and two cases of haemoptysis (one being described as massive) and one valve migration (Table 2). There were no pneumothoraces encountered and the mean hospital length of stay was 5 days. As an aid to comprehending the distribution of these complications, they are also included in Fig. 4.

Discussion
Our review demonstrates that EBLV was well tolerated by this high-risk cohort of patients, with either MAC or GA via ETT and the peri-procedural complications were  low. The PaCO 2 trends were similar using both techniques. As a minimally invasive approach, endoscopic lung volume reduction (ELVR) has been shown to improve pulmonary function, exercise capacity, and quality of life in patients with severe emphysema [1]. Two types of valves are commercially available, the Zephyr valve (Pulmonx) and the Spiration valve (Olympus, Japan) [2]. The Zephyr valve is a second generation device that has replaced the first generation Emphasys valve, but the recent literature has been focused predominantly around the Zephyr valve (Fig. 1). Compared to the first generation valves, the Zephyr valves offer less flow resistance and promote better air flow especially with the larger valves. Also, the deployment manoeuvre is much simpler allowing them to be placed under lighter levels of anaesthesia [7].
While, an earlier case series on the first generation valve insertion described the application of GA via ETT [4] both MAC and GA via ETT have been described for the deployment of Zephyr valves as well for the Chartis measurement [3][4][5][6]. A recent recommendation by an expert panel on ELVR is in favour of advocating a GA via ETT technique. The arguments in its support are: a more secure airway; less distortion to the Chartis measurement -especially with low ventilation rates of 8-10/min and prolonged expiration (I/E ratio 1:3-1:4) [8]. Further, ventilation helps in reducing CO 2 retention [1]. General anaesthesia also reduces the duration of the Chartis measurement [1].
Air trapping (increased residual volume and increased ratio of residual volume to total lung capacity) and development of intrinsic positive end-expiratory pressure (PEEPi) is a constant threat in patients with advanced COPD. There is elevated expiratory airway resistance, along with expiratory airflow limitation, reduced elastic recoil, and diminished expiratory time due to increased ventilatory rates. This impedes the pulmonary system in reaching the ideal elastic equilibration volume [9][10][11]. This phenomenon, referred to as dynamic hyperinflation can potentiate V/Q (ventilation/perfusion) mismatch, hypoxia, hypercapnia, barotrauma, and haemodynamic instability due to impaired venous return [12]. During EBLV procedures, ventilation is further compromised, as the respiratory passage is shared with bronchoscopy instruments, further limiting expiratory airflow. Ventilatory management should be targeted at reducing the dynamic hyperinflation, PEEPi, and air trapping [13]. Allowing a longer expiratory time (I/E ratio 1:3-1:4) is a major consideration to reducing air trapping and thereby decreasing the occurrence of 'breath stacking'. Keeping a vigil on dynamic hyperinflation is paramount. One recommended way to differentiate dynamic hyperinflation from tension pneumothorax during GA is to detach the ventilator circuit from the endotracheal tube and observing for the blood pressure to return to baseline (usually within a minute). In our series, information as to the exact ventilatory pressures and I:E ratios employed was poorly documented. However, it is likely that appropriate and optimal ventilator settings would have been utilized.

GA techniques
A TIVA technique may be preferable compared to a volatile based anaesthetic regime. Leakage of volatile agents is a concern due to repeated airway manipulation. Additionally, propofol may have a favourable effect of maintaining hypoxic pulmonary vasoconstriction, an effect beneficial in this scenario [14]. Remifentanil may be a good narcotic choice considering its potency and rapid onset and offset, allowing immediate return of respiratory function post procedure. Moreover, by continuing a low dose infusion during extubation, cough and hypertensive responses could be lessened. Sugammadex may be preferable to prevent residual neuromuscular blockade in the presence of compromised baseline respiratory function.

Monitored Anaesthesia care
Monitored anaesthesia care with sedation along with airway topicalisation are guided by Institutional practices. Adequate airway topicalisation may help in preventing cough and movement. Mucosal atomizing devices such as the DeVilbiss (DeVilbiss health care, Devilbiss International), can generate local anaesthetic droplets between 6 and 12 μm, which are delivered to the distal tracheobronchial tree [15]. Excessive suctioning to clear secretions can result in airway oedema, mucosal bleeding and inflammation, and affect the valve sizing judgements [1]. Apart from reducing the secretions, an antisialagogue might also help in reducing the dilution of the delivered local anaesthetic into the tracheobronchial tree during airway topicalisation. However, antisialagogue benefits should be weighed against the risk of their chronotropic effects [1]. Achieving optimal Data are presented as absolute numbers AF Atrial fibrillation, COPD Chronic obstructive pulmonary disease, SVT Supraventricular tachycardia sedation is always challenging, and never more so than during the Chartis Procedure. The patients' should display adequate tidal volume to validate the detection of a collateral-free target lobe. This should be balanced against minimizing coughing and/or secretion production [1]. Avoiding respiratory failure and CO 2 retention is a major concern during sedation techniques. Deep sedation should be attempted only if appropriate rescue resources are readily available [1]. Titrated doses of a short acting benzodiazepine and a narcotic may be an attractive option, as the effects can be readily reversed. Dexmedetomidine may be a promising selection in view of its sympatholytic and respiratory stability effects. The sympatholytic effect needs careful monitoring in these patients. For MAC technique, co-induction using reversible agents which reduce the requirements for dexmedetomidine, and thereby avoiding large drops in arterial pressure seemed popular with our anaesthetists. An increasing trend towards increasing GA utilization was noted in our review. It became evident that GA via ETT enabled quicker Chartis measurements. The absence of cough and movement made the technique easier. A GA technique also meant that any prolonged procedure could be easily tolerated. An increase in the anaesthetic times of about 24 min were noted for the MAC group compared with the GA via ETT group (61 vs 37 min). Reassuring the patients, and gaining their cooperation takes time; likewise, obtaining effective topical anaesthesia of the airway further prolongs the anaesthetic time required. Currently, GA via ETT is the preferred choice at our institution. A supraglottic airway together with neuromuscular blockade may also offer similar bronchoscopic conditions without the need for intubation. However, the bronchoscope may be inserted and removed on more than one occasion, and trauma to the vocal cords and surrounding structures from repeated passage of the bronchoscope; along with the risk of displacement of a supraglottic device, means that an ETT may prove to be the preferred choice. Making such decisions must involve the pulmonologist, and close cooperation between the proceduralist and the anaesthetist is absolutely essential.
A wide range of complications can occur during the procedure, such as respiratory and also haemodynamic instability. None of the patients own medications (such as beta blockers and calcium channel blockers) were omitted preoperatively and there is a possibility of these agents worsening dexmedetomidine induced hypotension. This does not seem to have been a significant problem in our patients. The liberal use of glycopyrrolate prior to commencing dexmedetomidine may have averted significant bradycardia. The hypotension observed in the GA via ETT cases was in line with the expected drop after any general anaesthetic. It responded to boluses of vasopressors namely ephedrine and metaraminol, and no infusions (either vasopressors or inotropes) were required.
Pneumothorax is a major complication of EBV procedures, attributed to the sudden expansion of the lobe adjacent to the target lobe after the collapse of the treated lobe, and patients showing significant volume reduction on post procedure Chest X-ray may be at an increased risk [1]. The incidence varies between 15 and 25% [16], and the median onset is 2 days post procedure [16,17]. It can also occur within hours after valve insertion and anytime during the procedure [2]. Two studies have looked into the factors responsible for pneumothorax occurrence after EBV insertion [17,18]. One found that a strategy of post procedure bed rest and cough reduction helped decrease the incidence [18].
Presently, it is unclear whether the type of anaesthetic technique has any role in precipitating pneumothorax. Although, there is a theoretical possibility of increased pneumothorax with positive pressure ventilation in these high risk patients especially with bullous lung disease, the risk may be reduced if a specific individually optimised ventilatory approach is adhered to. If High Frequency Jet Ventilation is available this may be an attractive option, however, we do not have this device in our institution. If a small pneumothorax does occur, − in the absence of tension -careful monitoring should suffice. A chest drain is indicated for larger pneumothoraces. Infections, COPD exacerbations, pneumonia and temporary haemoptysis are other possible postoperative complications [2].

Limitations of this review
The small numbers, the retrospective nature and missing data are some of the limitations of this review.

Conclusion
To conclude, both MAC and GA via ETT were well tolerated during ELVR in a high risk group of patients with multiple comorbidities. On one hand, hypercapnia is a major threat during sedation, on the other, positive pressure ventilation may impose a pneumothorax risk. Nonetheless, GA via ETT strategy may offer superior interventional conditions, patient comfort and reduced anaesthetic time both during the Chartis measurement as well as during valve deployment. Close communication between the interventional pulmonologist and the anaesthetist is crucial in identifying and managing serious peri-procedural complications.