Shigeto Ikeda introduced the flexible bronchoscopy (FB) for clinical use in 1967,1 and since then it has become the most important technique in the practice of modern pulmonology. Particularly for patients hospitalized in intensive care units (ICUs), FB has changed current diagnostic and therapeutic approaches. The versatility of the flexible bronchoscope, combined with its portability, allows one to perform the technique at the bedside, and this is of major importance in the unstable patient, who is often unable to be transported safely to the bronchoscopy suite.
FB performed in the ICU has its own specificities, either with regard to the patient or the environment that surrounds him or her. The critical patient often has one or more organ failures, which makes him or her a high-risk patient for the procedure. Moreover, as respiratory system involvement is a common feature, regardless of the precipitating factor of critical illness, the risk of respiratory failure is not negligible. Finally, the unstable character of such patients' diseases imposes time constraints and puts pressure on the bronchoscopist: the procedure must be rapidly performed and in certain circumstances absolute safety cannot be guaranteed throughout the examination, yet the urgency of it makes it imperative. In relation to the surroundings, the bronchoscopist is faced with a reduced amount of space, because of crowding of the monitoring equipment and therapeutic devices, quite different from the comfortable and controlled environment of a bronchoscopy suite. These aspects, taken together, imply an additional technical preparation. FB can be performed safely, as long as certain principles are met, but one must keep in mind potentially serious complications that can arise from this procedure. Because of its efficiency and safety in trained hands, it has been used increasingly in the unstable patient, particularly the patient under mechanical ventilation.
FB in critical patients should always be performed by skilled pulmonologists or intensivists. There is no clear evidence in the literature suggesting the minimum number of procedures one needs to perform to be considered well trained, and most probably that number varies on an individual basis, depending on one's dexterity.2,3 The lead assistant to the bronchoscopist, usually a specialized nurse, is essential for the performance of all aspects of the examination and must have training and expertise in all bronchoscopic procedures, equipment, and accessories.2–4
In some instances, FB must be performed outside the ICU for specific procedures (such as for transbronchial lung biopsies under fluoroscopic guidance). The transport of critically ill patients is associated with increased risk of morbidity and mortality and thus must be carefully planned. Guidelines for secure transport of critical patients are published elsewhere.5 However, basic principles include pretransport coordination and communication, accompanying trained personnel and resuscitation equipment, and monitoring of the patient during transport.
Although one can generally group the indications for bronchoscopy into diagnostic and therapeutic, there are circumstances in which the examination serves both purposes. Examples include hemoptysis and bronchial obstruction (by a foreign body, endobronchial tumors, or mucus plugs).
Between 65% and 79% of FB performed in ICUs are conducted in patients on mechanical ventilation, and 47% to 75% have therapeutic indications.6
Olopade and Prakash7 studied the use of FB in 198 critically ill patients. In their study, 45% were performed for the removal of retained bronchial secretions, 35% for collecting samples from the lower respiratory tract, 7% for assessing the airway, 2% for hemoptysis, 0.5% for assisting tracheal intubation, and 0.5% for the removal of foreign bodies.
Among the primarily diagnostic indications,1,2,6,8,9 one can include:
- Thoracic trauma
- Tracheal obstructive pseudomembrane
- Inhalation airway injury
- Upper airway and vocal cord function assessment
Radiographic abnormalities are a common finding in critically ill patients, particularly those on mechanical ventilation.1 In such patients, pneumonia is the most frequent infection.6 In fact, nosocomial pneumonias occur in 9% to 25% of patients on mechanical ventilation and in 70% of patients with acute respiratory distress syndrome. The impact of this disease is not negligible, as ventilator-associated pneumonia (VAP) has a mortality rate of 35% to 90%.6
Successful treatment depends on the initial antibiotic choice; inappropriate antibiotic coverage is an independent predictor of increased mortality in the patient with pneumonia.6 In addition, the delay in diagnosis implies a delayed start of therapy, which is also associated with worse clinical outcome.
The role of bronchoscopy in patients with suspected pneumonia is to identify the infectious agent, thereby allowing one to narrow the antibiotic spectrum; it also avoids treating patients without infection, favoring the emergence of resistant strains.6,10
FB is particularly useful in immunocompromised patients with pulmonary infiltrates, as it is a technique with a high diagnostic yield in these patients, especially in the identification of Pneumocystis jirovecci, mycobacteria, and fungi.6,8
Among the several bronchoscopic procedures available for the diagnosis of pneumonia, protected specimen brush (PSB) and bronchoalveolar lavage (BAL) are the most valuable.6,10,11 Transbronchial lung biopsy (TBLB) is a risky procedure in patients under mechanical ventilation, and its use should be restricted to specific circumstances, such as on suspicion of noninfectious lung disease, assuming that TBLB is an alternative to more invasive diagnostic procedures. For example, in lung transplant recipients, it is invaluable in establishing the differential diagnosis between infection and graft rejection.6,11
The technique of BAL fluid collection11,12 implies wedging of the tip of the flexible bronchoscope into an airway lumen, isolating that airway from the rest of the bronchial tree. Then, at least 120 mL of isotonic saline is instilled in several aliquots (3 to 6) through the working channel of the bronchoscope, and gentle hand suction is applied to retrieve the fluid. The amount of fluid returned, usually 40% to 70%, can vary significantly in specific circumstances, affecting the results (eg, a very small return may result in false-negative results). No more than 30 minutes should elapse between BAL collection and processing for microbiological analysis.
The methodology for PSB11 implies the use of a double-lumen catheter brush system with a distal occluding plug to prevent contamination from airway secretions during the passage of the catheter through the flexible bronchoscope channel. As for BAL, rapid processing of the samples is desirable.
Differential cell counts and quantitative cultures are of paramount importance in distinguishing between contamination and true infection.11 Both BAL and PSB samples should be rejected, if >1% of total cellularity are squamous or bronchial epithelial cells. In the case of BAL fluid, the diagnostic threshold for infection is 104 CFU/mL. For PSB samples, the proposed cutoff is 103 CFU/mL. The sensitivity of BAL ranges from 60% to 90% for bacterial infections; 70% to 80% for mycobacterial, fungal, and most viral infections; and 90% to 95% for P. jirovecci pneumonia.12
Mini-BAL sampling13,14 is a nonbronchoscopic and thus blinded technique that uses a telescopic plugging catheter through which 20 to 150 mL of fluid is instilled. Along with otherblinded nonbronchoscopic sampling techniques, it has been studied essentially for the diagnosis of VAP, to obviate the need to perform FB (reducing associated costs, potential complications, and the need for trained personnel). The cutoff for distinguishing pneumonia from colonization is ≥103 to 104 CFU/mL. The reported, fairly high, sensitivity and specificity values (63% to 100% and 66% to 96%, respectively) have been attributed to the more diffuse nature of VAP, thus increasing the yield of such sampling methods. Mini-BAL sampling methodology is not standardized and should be restricted to units where FB is not readily available.
Apart from the limitations specifically related to each one of the procedures mentioned, there are other factors that may affect the diagnostic yield of BAL and PSB samples. On the one hand, false-positive results by upper airway contamination must be minimized by using an aseptic technique and avoiding tracheal and main bronchi aspiration. In contrast, the use of lidocaine should be restricted, as it can inhibit bacterial growth.6,11 If the patient is already under antibiotic coverage, the diagnostic yield of BAL and PSB will be very low (unless very specific pathogens are involved, as previously mentioned).11
Apart from nosocomial pneumonia, FB plays an important role in the diagnosis of lower airway aspiration of stomach contents.1 In addition to confirming aspiration, it allows evaluation of the degree of tracheobronchial mucosal damage. Aspiration injury is usually visualized as areas of erythema and mucosal friability.
FB plays an important role in patients with hemoptysis. Endoscopic evaluation within the first 12 to 18 hours usually allows the identification of the site of bleeding and the guidance of subsequent therapeutic interventions. If the source of bleeding is not visible, segmental lavages can be performed in search of fresh blood in recovered fluid.1
The approach of mild to moderate hemoptysis requires the instillation of cold saline, serum-diluted epinephrine, and fibrin precursors.1
In massive hemoptysis, which accounts for 10% of cases, it is not clear whether it is preferable to use the rigid bronchoscope or the flexible bronchoscope.1,10 The former has the advantage of allowing control of the airway: proper ventilation during the procedure, better visualization, and effective aspiration of blood clots. The flexible bronchoscope despite providing access to more distal areas of the bronchial tree, has a limited suction capacity, but allows some basic procedures for airway maintenance and immediate control of the bleeding. In addition, a cryoprobe can be used to remove large clots from the airway.
After locating the source of hemoptysis, a 200-cm-long Fogarty balloon-tipped catheter can be introduced through the flexible bronchoscope working channel to tamponade the bleeding bronchial subsegment. This is achieved by inflating the balloon to occlude the bleeding zone. The balloon is deflated after 24 to 48 hours.15 In cases of unilateral massive bleeding, selective endobronchial intubation of the nonbleeding lung can be a life-saving measure.1,16,17
If an endobronchial lesion is detected, electrocautery, cryosurgery, and laser photocoagulation through the flexible bronchoscope are useful therapeutic tools; however, they may be difficult to accomplish with massive bleeding. Rigid bronchoscopy is the preferred approach in such a scenario. The main disadvantage of rigid bronchoscopy is the need to move the patient to the operating room and the requirement for sedation and muscle relaxation.1,8,10,16,17
Thus, for massive hemoptysis, FB is an essentially therapeutic technique, allowing balloon tamponade, selective endobronchial intubation of the nonbleeding lung, and other measures to control the bleeding, buying time for a definitive intervention.1,8,16,17
Tracheobronchial lesions affect 2.8% of patients with severe closed chest trauma. They may arise in the form of fractures or lacerations of the tracheobronchial tree. FB is the fastest and safest way to diagnose such injuries.6
The clinical picture depends on the type and location of the injury, but may be subtle and go unnoticed during early stages.6 Indications1 to perform FB in a patient with thoracic trauma include fracture of upper ribs, clavicle, or sternum; pulmonary or chest wall contusion; hemoptysis, dyspnea, or cough; and evidence of pneumothorax, pneumomediastinum, atelectasis, or subcutaneous emphysema. In particular, a pneumothorax associated with a persistent large air leak after tube thoracostomy is an indication for urgent FB.1 The radiologic evidence of lung collapse in the most dependent area of the lung field (falling lung sign) is rare but pathognomonic of total rupture of a mainstem bronchus.6,18
Tracheal Obstructive Pseudomembrane
Tracheal obstructive pseudomembrane is an infrequent complication of tracheal intubation.8 It usually appears in the cuff's previous location and is an excessive fibrotic response to compression of the tracheal mucosa. The endoscopic appearance is that of a white, thick membrane, tubular-shaped, and tightly adherent to the tracheal wall, similar to the membranes observed in staphylococcal infections in children and invasive aspergillosis in the immunocompromised, but with no infectious cause. It seems to be the first stage of a process leading to postintubation tracheal stenosis.
It manifests clinically as an acute upper airway obstruction shortly after extubation. FB establishes the diagnosis and allows for planning of future endoscopic therapeutic procedures using rigid bronchoscopy. After removal of the pseudomembrane, the appearance is one of a superficial hemorrhagic abrasion of the mucosa without cartilage damage, which usually heals without progressing to stenosis.8
Airway Inhalation Injury
Airway inhalation injury is common in fire victims, especially when plastic-derived or other synthetic material combustion fumes are inhaled. It can be divided into chemical or thermal injury, and predominates in the upper airways or lower airways, depending on the type and amount of inhaled toxin, its water solubility, and duration of exposure.19 Inhalation injury can occur in the absence of skin lesions and may be asymptomatic during the first 72 hours, even in patients with the most serious injuries, and this is the reason why FB must be performed early in suspected cases. The indications for performing FB are facial or nasal cilia burns, carbonaceous sputum, suspected acute obstruction of the airway, and inhalation of toxic vapors or fumes.8 In these particular patients, laryngeal edema develops quickly and can compromise the airway. Early FB can accurately identify those situations and, if necessary, allow tracheal intubation under endoscopic guidance.1
In lower airway burns, the endoscopic appearance of the mucosa can be almost normal at an early stage, with slight hyperemia and edema, which can go unnoticed, especially in the absence of carbon particles. However, even minor changes must be valued. Hours later, the mucosa may showscaly and necrotic areas, with carbon particles and focal areas of ulceration, alternating with areas of normal mucosa, creating a “mosaic” or “leopard skin” appearance. The appearance ofa thin, dry mucosa and the absence of edema are severity signs, as they reflect a deep mucosal lesion or complete epithelial abrasion.19 The cough reflex is typically absent, and should bepromptly investigated in nonanesthetized bronchi.8,20
Accidental injuries resulting from steam inhalation, either in the workplace or at home, as a consequence of cooking accidents and microwave superheated foodstuffs, can also be disastrous, especially when associated respiratory diseases are present.21 Steam inhalation induces direct thermal damage, with airway inflammation and edema, eventually followed by blistering and sloughing of bronchial mucosa. Once again, FB is the standard procedure for recognizing and evaluating such inhalation injuries.
Upper Airway and Vocal Cord Function Assessment
In the nonintubated patient, FB can be very useful when upper airway obstruction is suspected, identifying infections such as epiglotitis, laryngeal tumors, vocal cord paralysis, and other diseases presenting with stridor.1,22 In the ICU setting, however, the primary reasons for performing FB are upper airway evaluation after prolonged intubation, management of patients in whom stridor develops after extubation, and extubation over FB in suspected glottic or subglottic obstruction.1,6,9
The following are major therapeutic indications for performing FB in critically ill patients1,6,8,9:
- Endotracheal intubation
- Tracheobronchial obstruction (foreign body or endoluminal lesion)
- Percutaneous tracheostomy
Tracheal intubation under endoscopic visualization represents 0.07% to 3.4% of all intubations in ICU.1 FB plays a key role in 4 major groups of situations in which airway management is not simple: evaluation of the airways before intubation, intubation of the nonsedated patient, intubation in cases where neck extension is prohibited, and intubation in cases of suspected airway disease.10
To perform intubation under endoscopic visualization, the flexible bronchoscope of choice is the one with an outer diameter of 5.7 mm. The pediatric flexible bronchoscope should not be used because it is more flexible and is, therefore, more likely to bend into the esophagus.6 One should always try the orotracheal route, for several reasons: to avoid damaging the nasal mucosa, to prevent the development of purulent sinusitis and otitis media, and to allow intubation with a large caliber endotracheal tube (ETT), thereby decreasing the work of breathing and facilitating the clearance of bronchial secretions. If orotracheal intubation is to be performed, it is recommended to use a mouthguard to prevent biting of the flexible bronchoscope. In adults, one should try to pass an 8.0 mm inner diameter ETT. With nasotracheal intubation, it may be possible to pass a 7.5 mm ETT, but in patients with a narrow nasal cavity the risk of injury to the turbinates is not infrequent.
The lubricated ETT is passed over the flexible bronchoscope and the latter is used as a guide stent; the scope is advanced through the nose or mouth until it reaches the middle third of the trachea. The next step is to slide the ETT through the flexible bronchoscope into the airway. The advantages are obvious: there is visual control throughout the intubation process, and therefore it can be done safely. Moreover, continued observation allows for immediate endoscopic intervention if necessary, and, after the procedure, the bronchoscopist can ensure that the ETT is left in the correct position.8
The need for a skilled bronchoscopist is the major drawback of the technique. The attempt at intubation by an untrained endoscopist may prove to be disastrous by extending the period of hypoxia and damaging the upper airway mucosa, making subsequent attempts more difficult.
Intubation under endoscopic visualization must be planned and anticipated in cases of suspected difficult airway, which cannot be easily intubated or properly ventilated (cannot ventilate, cannot intubate scenarios).23,24 Indeed, intubation with flexible bronchoscope in an unplanned scenario is more risky, given the presence of secretions and inadequate topical anesthesia and the fact that it places the patient under increased risk of severe hypoxemia. Still, approximately 25% of flexible bronchoscope-guided intubations are performed in an emergency setting, after several failed attempts with direct laryngoscopy. This causes damage of pharyngeal and laryngeal structures, with bleeding and reduction in airway caliber by edema, which can compromise endoscopic visualization and further complicate the procedure.
Intubation under endoscopic control is very useful in the placement of double-lumen tracheal tubes (Robert Shaw) and also in the intubation of patients carrying airway stents, as blind tracheal intubation carries the risk of migration of the stent.8
In the study by Hasegawa et al,25 27% of emergency ICU FB were performed because of atelectasis and retention of secretions. The advantage of bronchial aspiration over aggressive chest physiotherapy for the elimination of secretions is a very controversial issue.1,6,26 Marini et al27 showed, in a prospective study, no advantage of bronchial aspiration over aggressive chest physiotherapy in patients with lobar atelectasis, whether under mechanical ventilation or not.
On the one hand, there are specific subsets of patients in whom endoscopic removal of secretions seems to have an added benefit: patients with spinal cord or brain injury and patients with neuromuscular disorders.1,8,10 In these patients, local directed suctioning, combined with acetylcysteine instillation, can be very effective.26 In contrast, the presence of an air bronchogram on chest radiography (indicating the presence of a patent airway) and persistent left lower lobe collapse after abdominal or thoracic surgery are features associated with lower effectiveness of FB in resolving atelectasis.10 Urgent endoscopic bronchial aspiration is sometimes necessary in patients with tenacious bronchial secretions, which form thick mucus plugs that are extremely difficult to aspirate even with flexible bronchoscope suction. In those circumstances, FB should not be delayed because of hypoxemia; respiratory failure in these patients is the clinical indication for performing the procedure.1
Use of mucolytic agents during bronchoscopy is controversial, and there are no reports directly evaluating the efficacy of this approach. There have been some isolated studies relating to endobronchial instillation of mucolytics to help in removing mucus plugs in ventilated patients with status asthmaticus, allowing weaning from mechanical ventilation.10 Apart from the apparent benefit in patients with neuromuscular disorders, there is no clear indication for the routine use of endobronchial mucolytics in patients with atelectasis in the literature.
In conclusion, although FB superiority cannot be documented in every patient with atelectasis, indications for endoscopic removal of secretions are lobar collapse unresponsive to aggressive chest physiotherapy and total lung collapse.1,10 Except for total lung collapse, radiologic improvement is delayed, appearing only 6 to 24 hours after endoscopic intervention, long after improvement in auscultatory findings and blood gas exchange.1
The removal of tracheobronchial foreign bodies may be attempted by FB, using a biopsy forceps or a Dormia basket.1,6 However, there are situations when it is preferable to use the rigid bronchoscope, because of the foreign body's size or for safety reasons (risk of aspiration and fragmentation).10
In relation to airway obstruction by endobronchial lesions or extrinsic compression, the recommended approach is rigid bronchoscopy, by use of procedures ranging from mechanical dilation with the bronchoscope itself or a balloon, photocoagulation with laser, and, if indicated, placement of a tracheobronchial stent.
Percutaneous dilational tracheostomy23,28,29 is a safe and effective alternative to standard surgical tracheostomy for patients dependent on mechanical ventilation. Percutaneous approaches were introduced in 1955 by Shelden et al and since then, the procedure has undergone several modifications. Of the 3 techniques available, Ciaglia's percutaneous dilational technique is the most common approach. One can use straight or curved dilators, although the latter have been more extensively studied in the literature.
Percutaneous tracheostomy can be performed at the bedside in ICU patients and a flexible bronchoscope can be very useful during the procedure.28–30 The bronchoscope is introduced through the ETT with its tip kept inside the tube to avoid damage by the introducer needle. The advantages are obvious: the needle is introduced under direct visualization along with the tracheostomy tube, and proper positioning of the ETT during the procedure is guaranteed. In this way, one can prevent paratracheal insertion of the tracheostomy tube and damage to the posterior tracheal wall during initial insertion of the needle. In addition, bronchoscopic visualization avoids impalement of the ETT by the needle. However, the presence of the flexible bronchoscope inside the ETT is associated with hypoventilation—if using FB, one should try to limit the bronchoscopy time. After successful placement of the tracheostomy tube, the bronchoscope is again inserted into the ETT, the cuff is deflated, and the tube gently pulled back from its suprastomal position, while carefully observing the proximal trachea, subglottic space, posterior commissure, and supraglottic structures.
The introduction of the flexible bronchoscope into the airway decreases the cross-sectional area available for airflow and induces changes at various levels, including respiratory mechanics, gas exchange, and hemodynamics.
A flexible bronchoscope with an outer diameter of 5.7 mm occupies approximately 10% of the cross-sectional area of the trachea6,10 and 15% of the cross-sectional area at the cricoid ring in an adult patient.6,31 The immediate consequence is some degree of airway obstruction. In a conscious, spontaneously breathing patient, the obstruction it creates is mild to moderate, perfectly tolerated, and does not induce significant intratracheal pressure variations. However, in the ventilated patient, the obstructive effect of the flexible bronchoscope is added to that of the ETT. Indeed, a 5.7 mm flexible bronchoscope occupies 40% of the cross-sectional area of a 9.0-mm inner diameter ETT, 51% of 8.0-mm inner diameter ETT, and 66% of 7.0-mm inner diameter ETT.31 This obstruction leads to a significant increase in airway resistance, which in turn generates significant intratracheal pressure variations during the respiratory cycle.
Lindholm et al19,31 studied the effect of FB on respiratory mechanics in 55 patients and found that in 38 patients who were under mechanical ventilation, high tracheal expiratory pressures rapidly induced a positive end-expiratory pressure (PEEP) effect (10.4±9.3 cmH2O), with a tendency to rise with the smaller ETT caliber and reduce the real expiratory time (which, in turn, is generated by higher respiratory rates and lower peak inspiratory flows). The importance of the ETT caliber is vital; with an 8.0 mm ETT, auto-PEEP usually remains below 20 cmH2O, but PEEP values of 35 cmH2O were reported in a patient carrying a 7.0 mm ETT, making otherwise healthy patients prone to develop pneumomediastinum and pneumothorax.6,10,31
The auto-PEEP induced by the presence ofthe flexible bronchoscope in a ventilated patient's airway induces a 30% increase in functional residual capacity a 40% decrease in forced expiratory volume in one second16 as well as a reduced vital capacity.32 As a consequence, expiratory tidal volume (VT) is significantly reduced.
During the inspiratory phase of the respiratory cycle, pre-existing lung inflation prevents the input of an acceptable VT given the significant increase of peak pressure. The resulting hypoventilation may induce significant blood gas changes. Thus, there are some investigators who suggest a minimum difference of 2.0 mm between the inner diameter of the ETT and the outer diameter of the flexible bronchoscope, to ensure the maintenance of an adequate VT.33
In conclusion, the airway obstruction limits not only expiration but also inspiration, causing lung inflation and alveolar hypoventilation (Fig. 1).
Suction applied during FB, although beneficial in patients with atelectasis in whom FB is performed to aspirate mucus plugs, can also have a negative impact on respiratory mechanics. Continuous and prolonged suction periods may reduce VT and functional residual capacity significantly, leading to small airway collapse and serious ventilation-perfusion mismatch, inducing severe hypoxemia.6,31 In this sense, suction should be quick and intermittent.
There are several determinants of changes occurring in the partial pressure of oxygen in arterial blood (PaO2) during FB6 (Fig. 2).
The presence of the flexible bronchoscope in the airway is associated with 10 a to 20 mm Hg reductions in PaO2 in an uncomplicated examination.34,35 When suction is applied, however, PaO2 fall can be more pronounced. Indeed, each suction can induce a 200 to 300 mL fall in VT, which, along with CRF reduction, can induce a 40% decline in PaO2.6 If, in addition, there is some VT loss through the swivel adaptor in ventilated patients, desaturation can be even more pronounced. Another feature contributing to the observed hypoxemia during FB is reflex bronchoconstriction, mediated by subepithelial parasympathetic nervous system receptors located in the large airways.6 An adequate topical anesthesia can reduce this effect.
The performance of BAL has a known deleterious effect on oxygenation. The PaO2 decline shortly after BAL can be explained by 2 phenomena: epithelial surface changes induced by the instilled fluid and local proinflammatory mediators release. After the procedure, there is a gradual return to baseline PaO2 levels, which can take approximately 15 minutes in the normal individual to several hours in the presence of severe pulmonary parenchymal disease.6
In some patients, there may be a rise in PaO2. Such cases are associated, almost invariably, with the resolution of atelectasis by suction of tracheobronchial secretions or blood clots. Another phenomenon that may contribute to better oxygenation is the auto-PEEP generated during the procedure, which, by recruiting collapsed alveoli, improves ventilation-perfusion relation.6
With regard to changes in ventilation, a slight rise of approximately 8.2 cmH2O in the partial pressure of carbon dioxide in arterial blood is common during the procedure.6 This increase reflects alveolar hypoventilation induced by the cumulative effect of the reduction in VT and lung inflation. Prolonged periods of suction, particularly in the absence of secretions, can exacerbate this phenomenon.
The combined effects of hypoxemia, hypercapnia, mechanical irritation of the airways, and the patient's own anxiety (less important feature in sedated and ventilated patients) cause adrenergic stimulation, with consequent increase in mean arterial pressure, heart rate, and pulmonary artery pressure.6,31 Lindholm et al31 reported a 50% increase in cardiac output during FB, with a return to baseline levels 15 minutes after the procedure.
The performance of FB in patients with severe brain injury raises safety concerns related to changes in intracranial pressure (ICP). Peerless et al36 showed an average increase of 114% in ICP (13.5±8.8 mm Hg) in 15 patients who underwent FB despite severe brain injury. The simultaneous increase in mean arterial pressure during the procedure allowed an adequate cerebral perfusion. Both variables returned to normal a few minutes after the removal of the bronchoscope from the airway and restoration of normal ventilation.
There are several factors that may contribute to increased ICP. In the spontaneously breathing patient, cough (generating a PEEP effect) is the main determinant. In ventilated patients, cough in response to stimulation of the airways generates a cumulative effect in relation to the auto-PEEP generated by the flexible bronchoscope itself. Proper sedation and muscle relaxation before performing FB can suppress the elevation of ICP in these patients. In addition, one should use the largest ETT possible, to lessen the PEEP effect.36
Hypoxemia may also induce a rise in ICP, but this can be prevented through close monitoring of peripheral oximetry. Hypercapnia, by causing cerebral vasodilation, can also increase ICP.
In conclusion, although there are many factors that can potentially induce intracranial hypertension, the performance of FB in patients with severe brain injury is a safe procedure that does not adversely affect the neurological status of patients with cerebral lesions, as long as it is carried out respecting basic rules.36
If a spontaneously breathing patient is on supplemental oxygen by breathing mask, a small hole is fashioned in the mask, opposite the nostril through which the flexible bronchoscope will be introduced.6 After proper local anesthesia with lidocaine and lubrication of the flexible bronchoscope, the same is passed nasally. The nasotracheal route should be attempted whenever possible, as it induces less gag reflex and does not require a mouthpiece.
In patients under mechanical ventilation, there are a number of practical recommendations that must be met according to the pathophysiological considerations outlined above. They relate mainly to the adaptation of the ventilatory circuit and the adjustment of ventilator parameters1,6,26,31,33:
- Anesthetize the tracheobronchial tree properly;
- Sedate the patient (add muscle relaxation in selected cases);
- The ETT should have an inner diameter of at least 8.0 mm; the difference between the inner diameter of the ETT and the outer diameter of the flexible bronchoscope should be ≥2 mm. If the patient is intubated with an ETT <8.0 mm, and tube change is not a viable option, then a pediatric or an ultrathin bronchoscope can be used;
- Use a swivel adapter to minimize VT loss through the circuit;
- Ventilate on volume control mode. Triggered modes are not generally recommended because ventilation will not be guaranteed unless the peak pressure limit is increased to ensure that the desired VT is delivered. Pressure-control ventilation will also result in a reduced VT, unless the inspiratory pressure level is increased to compensate for high airway resistance during the procedure. Exhaled VT must be continuously monitored (ensure that it is similar to prebronchoscopic levels during the procedure). In addition to the increase in VT, increase the ventilator rate, if necessary, to maintain adequate ventilation;
- Set PEEP to 0 cmH2O (if this is not feasible, reduce the PEEP level by 50%). Despite concern over derecruitment, this measure is proposed by most investigators to avoid dangerous hyperinflation and barotrauma, as generation of auto-PEEP during the examination is well documented;
- Increase FiO2 to 1.0, starting 5 to 15 minutes before the procedure, for adequate preoxygenation. Maintain FiO2 at 1.0 during and up to 1 hour after FB, with the purpose of keeping SaO2 as close to 100% as possible;
- Set the peak inspiratory flow level to ≤60 L/min;
- Visually monitor the magnitude of thoracic excursions during the procedure;
- Apply suction only for short periods (3 s or less).
During the procedure, continuous electrocardiogram, blood pressure, and peripheral O2 saturation monitoring is mandatory. Cardiopulmonary resuscitation equipment should always be at hand. The team conducting the FB must have the necessary technical skills to perform the procedure and to act promptly should complications arise. After the procedure, the patient should receive a chest radiograph to exclude pneumothorax or pneumomediastinum.
Sedation and Analgesia
Sedation1,37,38 is used to achieve patient comfort, safety, and cooperation. Adequate sedation must provide anxiolysis, anterograde amnesia, and analgesia. Some analgesics, such as opiates, provide additional suppression of cough and gag reflexes, which are important effects when performing FB. Patients under mechanical ventilation need to be properly sedated during the performance of FB, to increase comfort and minimize respiratory and cardiovascular derangements induced by the procedure.
Most frequently, a combination of benzodiazepines and opiates is used.1,37 Midazolam is the preferred benzodiazepine, given its rapid onset of action (<5 min) and short half-life. Sedation can be prolonged in elderly patients, and dose reduction is necessary for hepatic failure. Fentanyl39 has an onset of action of <90 s and is the preferred analgesic in patients with hemodynamic compromise, because cardiovascular effects are minimal. Respiratory depression and hypoxemia are potential adverse effects of these drugs, and the combination of both can induce greater hypoventilation than midazolam alone.26,37
Propofol is an anesthetic that can also be used for sedation, either by bolus administration or continuous infusion. Its rapid onset of action (<1 min) and recovery time are advantageous. Clearance is not affected by renal or hepatic failure. However, respiratory and cardiovascular depression are more likely to occur with this drug, as deep sedation and general anesthesia and thus propofol use requires some degree of experience and expertise. It is a very good agent for sedation of mechanically ventilated patients, although it must be managed with caution in hemodynamically unstable patients.26,37,40
Ketamine elicits sedative and analgesic effects without cardiovascular depression. Respiratory depression is also minimal, unless the drug is infused too rapidly or inappropriately high doses are used. However, its dissociative properties may induce a state of emergence delirium, which is the reason why it must be used with caution in adult patient sedation. Laryngospasm is another troublesome side effect.37,39
Thiopental is a barbiturate used for short-term sedation. Hypotension is its most prominent adverse effect, and advanced age and critical illness potentiate this effect.41 It is not routinely used for FB.38 Both propofol and thiopental are purely sedative/amnestic drugs and must be used in association with an analgesic drug.41
The individual response to medication can vary. The clinician must be prepared to act appropriately in the case of inappropriately deep sedation or general anesthesia, especially when using propofol, ketamine, or thiopental. Oversedation with benzodiazepines and opiates can be reversed with flumazenil and naloxone, respectively. The other drugs have no specific reversal agents.16,37
The Patient With Respiratory Failure
As noted earlier, FB is associated with a mild to severe decline in oxygenation. The performance of such a procedure in patients with respiratory insufficiency poses serious risks of severe respiratory failure, hemodynamic instability, and arrythmias.34,35
In patients with severe hypoxemia [need for continuous positive airway pressure (CPAP) or a FiO2 of at least 0.5 to obtain a PaO2 of at least 75 mm Hg],34 FB should be deferred. However, clinical judgment must prevail, as performing the FB may be the only way to improve gas exchange in selected cases. In addition, it is of paramount importance in establishing the diagnosis in patients with pulmonary infiltrates, often with significant blood gas changes. Thus, clinicians have sought solutions that can avoid intubation and mechanical ventilation of high-risk patients to perform FB safely.
Maitre et al35 studied CPAP in candidates for diagnostic FB with moderate respiratory failure (defined as PaO2/FiO2<300), and concluded that the use of positive pressure during the procedure allowed the retention of adequate gas exchange, preventing subsequent respiratory failure. Antonelli et al,34 in turn, conducted a comparative study in patients with moderate to severe partial respiratory failure (defined as PaO2/FiO2<200) using noninvasive ventilation (NIV), as opposed to the use of supplemental O2 by Venturi mask. Thirteen patients underwent FB under NIV (full-face mask, PEEP 5 cmH2O, pressure support of 15 to 17 cmH2O, and FiO2 of 0.9), and 13 patients underwent the procedure with FiO2 of 0.9 by Venturi mask. In patients under NIV, as opposed to the conventional group, the PaO2/FiO2 ratio remained significantly higher during and after the procedure, mean arterial pressure remained stable, and heart rates were lower after the procedure.
In general, an oronasal interface is adjusted to the patient and the ventilator set either in CPAP mode or for NIV support, with supplemental oxygen as required. The bronchoscope is inserted through a hole in the front of the mask.42
There are, to date, no studies showing the superiority of NIV compared with CPAP. However, according to the literature,34,35,43–45 ventilatory support, either in the form of continuous positive airway pressure or with the addition of pressure support, seems to be well tolerated and effective in stabilizing gas exchange during FB in patients with severe hypoxemia.
FB is considered a safe procedure, as long as the bronchoscopist is aware of indications and contraindications and basic safety precautions are taken.1,6,26 Major complications arise in 0.08% to 0.15%, and minor complications occur in approximately 6.5% of the cases.1 There are studies reporting a mortality rate of 0.01% to 0.04%.6 A review conducted by Olopade and Prakash, which included 6 studies and a total of 804 FB performed in the ICU, did not record any procedure-related deaths.6 However, in a large retrospective analysis of diagnostic and therapeutic bronchoscopies performed in 23,682 patients, the mortality rate was 0.013% and the severe complication rate (massive hemoptysis, laryngospasm, bronchospasm, tracheospasm, arrhythmias, pneumothorax, subcutaneous emphysema, esophagotracheal fistula, tracheal perforation, and airway obstruction) was 0.637%.46 The authors believe that the implementation of therapeutic bronchoscopic techniques, such as electrocautery, argon-plasma coagulation, laser, balloon dilation, and stenting, is in close relation to the increase in severe complication rate. Prebronchoscopic evaluation is still the most important step in guaranteeing a complication-free procedure.46
The following are considered major, potentially life-threatening complications2,26: respiratory depression, pneumonia, pneumothorax, hemorrhage, airway obstruction, cardiorespiratory arrest, arrhythmias, and pulmonary edema.
Complications can be classified under the following 3 groups1:
- Associated with sedation and anesthesia
- Associated with FB itself
- Associated with ancillary techniques
Complications Associated With Sedation and Anesthesia
The first group includes respiratory depression and respiratory arrest, syncope, tachycardia, hypotension, laryngospasm, nausea, and vomiting, among others. One should always be aware of complications associated with the use of lidocaine for local anesthesia, ranging from laryngospasm and bronchospasm to arrhythmias, seizures, and methemoglobinemia, which is the reason why bronchoscopists should not forget that lidocaine is a drug with documented absorption by the airway mucosa and that its maximal dose must therefore not be exceeded (8.2 mg/kg). Careful attention should be paid to patients with hepatic failure, in whom lidocaine metabolism is impaired.2,16,26
Complications Associated With FB
In the second group, epistaxis, infection, fever, laryngospasm, bronchospasm, arrhythmias, and hypoxemia are among the most significant complications.
The risk for bacteremia after FB, although it was once thought to be very small, can be as high as 6.5%, according to some published reports (especially in immunocompromised patients47). Still, endocarditis prophylaxis is not generally recommended, except in asplenic patients, patients with prosthetic valves, or patients with previous endocarditis history. Pneumonia is a rare complication.1,26
Fever after FB is not common, although reported frequencies of this complication range from 1% to 20%.1,47 It is usually self-limited, subsiding in the first 24 hours. Transient bacteremia and the release of proinflammatory cytokines as a consequence of FB have been thought to play a role in postbronchoscopy fever.47 However, bacteremia is a rare finding in immunocompetent individuals and cannot explain, at least inthese patients, the occurrence of fever, as Umetal47 have shown. Reported predisposing factors1,47 are advanced age, abnormal bronchoscopic findings, endobronchial obstruction, bronchoscopic intervention for malignancy, severity of bleeding during FB, endotoxin contamination, instillation of topical lidocaine through the bronchoscope, abnormal differential cell counts and bacterial growth in BAL fluid, and performance of bronchial brushing and BAL. In patients undergoing BAL, fever can occur in up to 30% of cases.1
Arrhythmias result from the combined effects of hypoxemia and increased sympathetic tone during the examination, with associated tachycardia and myocardial ischemia. However, of all potentially arrhythmogenic factors, hypoxemia is the most important and should be strongly avoided by administration of supplemental oxygen, by performing the procedure as quickly as possible, and, if necessary, by intermittent removal of the flexible bronchoscope from the patient's airway to allow ventilation (especially in patients under mechanical ventilation).1
Hypoxemia is one of the most common complications of FB.1 The decline in PaO2, which may reach 30 to 60 mm Hg in critically ill patients, can persist for up to 2 hours after the procedure.1,6 Among the several mechanisms associated with desaturation during FB, respiratory mechanics changes induced by the presence of the flexible bronchoscope in the airway and continuous suction are probably the most important.
Complications Associated With Ancillary Techniques
The third group includes the most significant complications of bronchoscopy: bleeding and pneumothorax.
Bleeding during bronchoscopy is considered significant if it exceeds 50 mL.6 The risk of bleeding is highest with TBLB, followed by bronchial biopsy, bronchial brushing, and finally BAL, although it is rare if one respects all contraindications.6 The risk of TBLB-associated bleeding is 9% to 11% and, in most cases, is self-limited or can be stopped with the local instillation of cold saline or adrenaline. However, there have been situations of uncontrolled bleeding that must be managed with endobronchial tamponade, selective intubation of the nonbleeding lung, and, in extreme cases, surgery.
Factors that increase the risk of bleeding include coagulopathies, thrombocytopenia, platelet dysfunction, severe uremia, hepatic failure, pulmonary hypertension (PH), superior vena cava syndrome, and malabsorption. The bleeding risk is, likewise, increased in immunocompromised patients and patients with severe malnutrition.2,6,26
Pneumothorax is described as a complication of TBLB in approximately 1% to 5%,26 and risk is especially high in patients under mechanical ventilation, in which case it may reach 7% to 15% (rates of 23% have been described).1,6,22,26,48,49 Nevertheless, TBLB can still be performed with an acceptable risk in mechanically ventilated patients, as long as the benefit clearly outweighs the risks in the individual patient (ie, if one is to expect a change in patient management).48
TBLB can be performed blindly or under fluoroscopic guidance. The primary reason for using fluoroscopy is to reduce the risk of pneumothorax. Moreover, when performing biopsies for localized peripheral lesions, fluoroscopic guidance allows for the tracking and directing of the tip of the flexible bronchoscope toward the lesion, increasing the diagnostic yield. However, TBLB can be performed blindly in specific situations, especially diffuse lung diseases, infiltrates localized to a lung apex, or radiologically well-defined infiltrates, involving an entire segment of a lobe.22 There is no general consensus with regard to using fluoroscopic guidance to increase the safety of TBLB in mechanically ventilated patients, although it seems the risk of pneumothorax after TBLB is higher without fluoroscopic control. It seems to be a reasonable option, when available.
Although much less frequent, pneumothorax can also arise in patients under mechanical ventilation by barotrauma (particularly if FB is performed through an ETT with <8.0 mm inner diameter or if ventilatory parameters are not properly adjusted). In a ventilated patient, an iatrogenic pneumothorax is an indication for placement of a pleural drain, although some investigators refer to the need for pleural drainage in only 50% of patients.1,26 In patients under mechanical ventilation one should never perform bilateral TBLB simultaneously, because of the small risk for bilateral pneumothorax.1 Signs and symptoms may not be immediate; however, it is very uncommon for a pneumothorasc to develop more than one hour after the procedure.26
There are very few contraindications for performing FB.1,2,6,10,22,26
The only absolute contraindications are the noncooperation or refusal by the patient, an inexperienced bronchoscopist, lack of suitable facilities or equipment, and inability to maintain adequate oxygenation during the procedure.
The remaining are considered to be relative contraindications as they place the patient at risk of certain complications. The performance of FB in such particular settings should be carefully considered, and the potential benefit for the patient weighed against the risks involved (Table 1).
Hypoxemia has been associated with increased risk of arrhythmias and is considered a relative contraindication. There are several proposed definitions of severe hypoxemia. In nonintubated patients, a requirement of CPAP or FiO2 of at least 0.5 to achieve a PaO2 ≥75 mm Hg is considered a contraindication to FB.34 For other investigators, a PaO2 level <70 mm Hg while on FiO2 of 0.7 is considered a high-risk situation.1 In these patients, the procedure probably should be performed under CPAP or NIV or with elective intubation. As a general rule, a safety SaO2 threshold of 90% must be guaranteed, and it seems reasonable to postpone FB if one is unable to keep SaO2 above 90% while breathing 100% oxygen, except in life-saving situations.1,6,26
Acute coronary syndromes are generally considered to increase the risk of complications during FB. In a retrospective analysis of patients who underwent diagnostic FB (including airway examination, BAL, endobronchial biopsies, endobronchial brushing, and TBLB) within 30 days (as early as 24 h) of an acute myocardial infarction, the procedure was safe and no adverse events resulted in its interruption. One must say, however, that all patients had previously undergone a revascularization procedure.50 These findings are consistent with those of Dunagan et al,51 who conducted a retrospective review of 40 patients undergoing FB while in a coronary care unit. The practical application of such findings is that, in the face of a clinical indication for performing FB in a patient with a recent acute myocardial infarction, the decision to postpone the procedure must be individualized.
There has been some discussion with regard to PH and the risk of complications during diagnostic FB, namely bleeding. Some authors suggest that PH is a contraindication to TBLB, given the risk of uncontrollable bleeding as a result of increased perfusion of the pulmonary vascular bed.1 The rationale behind this statement is that patients with PH are at risk for adverse hemodynamic events during FB. Bronchial suctioning is associated with an increase in pulmonary capillary wedge pressure, and some mechanically ventilated patients undergoing FB have an acute pulmonary hypertensive response during the procedure.52 Moreover, patients with chronic venous PH may have dilation of submucosal bronchial veins.52 In the face of these findings, one can understand the general concern in performing FB in PH patients, particularly if TBLB is indicated. However, a retrospective study conducted by Diaz-Guzman and Vadi52 found no difference in several recognizable FB complications, namely bleeding, between patients with and without mild to moderate PH. Thus, it appears that bronchoscopic diagnostic procedures (including TBLB) can be performed safely in mild to moderate PH. However, data are limited for severe PH, and therefore, caution must be used in such patients.52
Thrombocytopenia is a relative contraindication for FB. When performing bronchoscopy only for simple airway examination, there is a proposed cutoff of 20,000 to 50,000/μL.1,49 Most authors propose platelet counts above 50,000/μL for biopsies (Tai6 proposes counts above 75,000/μL for TBLB). BAL collection is a less risky procedure, and thus, it is contraindicated only below 20,000 platelets/μL.2,22 Uremia is associated with platelet dysfunction and carries some bleeding risk; thus, it has been suggested that serum creatinine levels of 3 mg/dL or greater and blood urea nitrogen levels of 30 mg/dL or greater should be relative contraindications to performing TBLB.22,53 Desmopressin can temporarily reverse the platelet dysfunction associated with uremia and can thus be useful for short procedures in patients with renal failure.1,49 Although it is generally agreed that changes in prothrombin time (PT) and activated partial thromboplastin time (APTT) should be corrected before FB, and some authors state that PT or APTT values greater than 1.5 times control increase bleeding risk, there is no clearly defined cutoff above which the procedure is contraindicated.1,26 PT >50 s has been proposed as a contraindication for BAL.2 For biopsies, it is generally agreed that clotting disorders should be properly corrected, but there is no defined cutoff for PT or APTT.2 In patients receiving oral anticoagulation, published guidelines suggest stopping anticoagulants 3 days before the procedure or the administration of low-dose vitamin K.26 In patients with high thromboembolic risk, in whom anticoagulation cannot be stopped, international normalized ratio should be kept under 2.5 and heparin started.26
In some instances, patients have neither recognizable bleeding diathesis nor are under oral anticoagulation. They are, however, on antiplatelet drugs, such as acetylsalicylic acid (ASA) or clopidogrel, for cardiovascular disease. In a prospective study performed in 1217 patients, ASA alone did not pose significant bleeding risk after TBLB.54 However, in a randomized trial55 of 12,562 patients presenting with acute coronary syndromes without ST-segment elevation, the rate of major bleeding complications was higher with clopidogrel than that with ASA alone. These findings led Wahidi et al56 to analyze the risk of bleeding during TBLB in pigs, and the study showed no increase in bleeding complications on clopidogrel-treated animals (with or without ASA). Later on, Ernst et al57 conducted a study on 604 patients undergoing TBLB, with or without clopidogrel. The study had to be stopped prematurely because of the excessive bleeding rates in the clopidogrel group (with or without ASA). There was no difference in bleeding rates or severity among groups with different indications for TBLB. In the face of such findings, the authors concluded that clopidogrel greatly increases the risk of bleeding after TBLB and recommend stopping the drug 5 to 7 days before the procedure. Immune suppression and malnutrition do not pose as much bleeding risk as once thought; however, the decision to perform FB in these patients should be carefully considered.1
Since its introduction by Ikeda, the role of FB has been expanding and is nowadays a key diagnostic and therapeutic procedure in the ICU. It is a safe technique with widespread use in all sorts of critically ill patients, as long as the benefits outweigh the risks in such particular candidates. However, performing FB in a safe manner in unstable and ventilated patients requires sound knowledge of the specificities of such a setting. It should, therefore, be conducted only by skilled bronchoscopists or pulmonologists with specific training in this area.