© 2004 American Thoracic Society
Central Airway ObstructionPulmonary and Critical Care Division, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts; Department of Interdisciplinary Endoscopy, Thoraxklinik, Heidelberg, Germany; Cleveland Clinic Foundation, Cleveland, Ohio Correspondence and requests for reprints should be addressed to Armin Ernst, M.D., Director, Interventional Pulmonology, Beth Israel Deaconess Medical Center, Harvard Medical School, 330 Brookline Avenue, Boston, MA 02215. E-mail: Aernst{at}caregroup.harvard.edu
Central airway obstruction is a problem facing all medical and surgical subspecialists caring for patients with chest diseases. The incidence of this disorder appears to be rising because of the epidemic of lung cancer; however, benign causes of central airway obstruction are being seen more frequently as well. The morbidity is significant and if left untreated, death from suffocation is a frequent outcome. Management of these patients is difficult, but therapeutic and diagnostic tools are now available that are beneficial to most patients and almost all airway obstruction can be relieved expeditiously. This review examines current approaches in the workup and treatment of patients suffering from airway impairment. Although large, randomized, comparative studies are not available, data show significant improvement in patient outcomes and quality of life with treatment of central airway obstruction. Clearly, more studies assessing the relative utility of specific airway interventions and their impact on morbidity and mortality are needed. Currently, the most comprehensive approach can be offered at centers with expertise in the management of complex airway disorders and availability of all endoscopic and surgical options.
Key Words: bronchoscopy • central airway obstruction • stenting • tumor destruction
Etiology, Presentation, and Diagnostic Approach Securing the Airway Anesthesia for Central Airway Obstruction Therapeutic Approaches Dilation of the Airways (Bronchoplasty) Electrocautery Argon Plasma Coagulation Laser Therapy Photodynamic Therapy Cryotherapy External Beam Radiation and Brachytherapy Airway Stents Surgical Resection Airway Complications after Lung Transplantation Postintervention Care Summary: The Future of Central Airway Obstruction Obstruction of the central airways, the trachea and mainstem bronchi, can result from a variety of disease processes and is the cause of significant morbidity and mortality. Although the actual incidence and prevalence of central airway obstruction (CAO) are unknown, the epidemiology of lung cancer would suggest an increasing number of patients develop complications of proximal endobronchial disease (1). An estimated 20–30% of patients with lung cancer will develop complications associated with airway obstruction (atelectasis, pneumonia, dyspnea, etc.) (2), and up to 40% of lung cancer deaths may be attributed to locoregional disease (3). In addition, with increased use of artificial airways such as endotracheal intubation in an enlarging and aging population, the incidence of both benign and iatrogenic complications is also likely increasing. With the growth of subspecialties such as interventional pulmonology, there has been more attention and research directed toward treating patients with complex airway pathology resulting from both benign and malignant disease. In 1898, Gustav Killian was the first to report using a bronchoscope to remove a foreign body (a piece of pork bone) from a farmer's trachea, introducing the era of modern bronchoscopy (4). Since that time, and especially since the development of the flexible bronchoscope in 1967 by Dr. Shigeto Ikeda, many modalities have become available to diagnose and alleviate central airway obstruction. The choice of interventions is generally dictated by the patient's stability, the nature of the underlying problem, the overall prognosis, the patient's quality of life, the particular expertise of the physician, and the available technology. The evaluation and management of patients with CAO require a thorough knowledge of the etiology, physiology, diagnostic, and treatment options, as well as a multidisciplinary team approach including chest radiologists, anesthesiologists, medical oncologists, thoracic surgeons, and interventional pulmonologists. Designing large randomized trials to determine the best strategies for evaluation and treatment in this patient population is extremely difficult. Limiting factors include selecting patients with comparable disease and comorbidities, as well as the fact that these patients are often critically ill. Many investigators believe that it would be unethical to randomize patients with obviously symptomatic disease, who may experience immediate benefit from therapy aimed at opening their airway, to a nonintervention arm (5). Double-blinding is also impossible in this population for obvious reasons. The literature, therefore, is largely made up of case series and retrospective analyses. Nonetheless, the impact of therapy on quality of life and survival is impressive, and as new technology becomes available the primary question will not be "Is therapy helpful?" but "Which therapy is best for a particular patient?" As many of the therapies used to treat CAO are quite expensive and require specialized training, their use tends to be concentrated in "centers of excellence." For example, over the course of a 6-year period, Noppen's group in Belgium experienced a fivefold increase in interventional bronchoscopy activity, with the majority of cases being referred from outside institutions (3). For this article, Medline was searched with keywords "airway" AND "obstruction," as well as "airway" and each of the therapeutic modalities, such as "cryotherapy." The references of the articles were also searched. In this review, we discuss the diagnosis of and general approach to central airway obstruction in adults. Specific management modalities are discussed in relevant detail, with a focus on the most relevant literature.
Central airway obstruction may be caused by a variety of malignant and nonmalignant processes, as listed in Table 1 . We use the term "nonmalignant" instead of benign, as the signs and symptoms of CAO are typically extremely disturbing and just as much a threat to life as malignant causes of CAO, and are therefore not benign. The most common cause of malignant CAO is direct extension from an adjacent tumor—most commonly bronchogenic carcinoma, followed by esophageal and thyroid carcinoma. Primary tumors of the airway are relatively uncommon, with an estimated incidence of 600–700 cases per year. The majority (70–80%) of primary tracheal tumors are squamous cell carcinoma or adenoid cystic carcinoma. Carcinoid tumors, however, account for the majority of primary airway tumors distal to the carina (6). Distant tumors may also metastasize to the airway, with the most common causes including renal cell, breast, and thyroid carcinoma. Although there are few epidemiologic data, the most commonly encountered nonmalignant causes of CAO include granulation tissue resulting from prior endotracheal/tracheostomy tubes, airway foreign bodies, and likely tracheal or bronchomalacia.
A particular patient's presentation will not only depend on the underlying disease, but on the location and rate of progression in airway obstruction as well as their underlying health status and other associated symptoms such as postobstructive sequelae. Patients may develop symptoms suddenly, as can happen with foreign body aspiration, or more gradually, as is often the case with slowly growing malignant obstructions. If airway obstruction is mild, it may have little effect on airflow and therefore patients may be asymptomatic. The inflammation associated with even mild respiratory tract infections can cause mucosal swelling and mucus production, which may further occlude the lumen. For this reason, patients may be misdiagnosed with exacerbations of chronic obstructive pulmonary disease or asthma, especially because symptoms such as wheezing and dyspnea may improve with therapy aimed at treating the superimposed infection. Signs and symptoms develop when the CAO impairs airflow to the point of increasing the work of breathing or altering cardiopulmonary interactions. Alvin Barach, in 1935, postulated that the primary sensation of air hunger in patients with CAO was not related to hypoxia or hypercapnia, but rather to "the increased effort required to obtain the normal velocity of air delivered to and from the lungs" (7). Although wheezing indicates airflow through a narrowed orifice, its location does not always conform to the site of airflow obstruction. Wheezing heard best over the trachea does not necessarily indicate that the source of the obstruction is the trachea (8). Unilateral wheezing, however, often suggests airway obstruction distal to the carina (8). The presence of persistent unilateral wheezing should always prompt the investigation of focal airway obstruction. Stridor is a sign of severe laryngeal or tracheal obstruction. Patients may also present with other nonspecific symptoms such as exertional dyspnea and positional wheezing. With an anatomically fixed obstruction, shortness of breath and wheezing are typically unresponsive to bronchodilators, and failure of a patient to improve with these measures should prompt the physician to consider the presence of CAO. It is important to note that the trachea is typically significantly narrowed, to less than 8 mm, before exertional dyspnea is present. Once the lumen is less than 5 mm, symptoms present at rest (8, 9). Because of the dramatic loss of airway diameter before the development of symptoms, up to 54% of patients with tracheal stenosis can present in respiratory distress (10). When evaluating patients with suspected CAO, it is crucial to examine the shape of the flow–volume loop and not only the FEV1, FVC, and FEV1/FVC ratio. The characteristic blunting of the flow–volume loop that is indicative of central airway obstruction is typically seen before spirometry yields abnormal results, yet may not be appreciated until the airway is already narrowed to about 8–10 mm (11–13). As one would expect, it is the initial, effort-dependent portion of the flow–volume loop that is affected by central airway obstruction. At lower lung volumes, the more peripheral airways normally and progressively limit flow (11, 14). One should also note that airflow obstruction at multiple anatomic sites, as can be seen in patients with emphysema and CAO, may produce atypical flow–volume loops (15). Because of the possibility of inducing respiratory failure, spirometry should not be undertaken in patients with respiratory distress or advanced CAO. Conventional chest radiographs are rarely diagnostic, yet are often obtained as the initial radiologic test. Obvious pathology, such as tracheal deviation (Figure 1) , can be identified; however, the chest X-ray is unable to determine airway invasion or aid in procedure planning. Standard computed tomography (CT) scans provide much more information, including the ability to document dynamic airway collapse, and help predict response to treatment such as photodynamic therapy (16, 17). Advances in airway imaging, however, now allow multiplanar and three-dimensional reconstruction with internal (virtual bronchoscopy) and external rendering, and excellent image quality is achievable by low-dose techniques (18–23). These new imaging protocols give better characterization as to whether the lesion is intraluminal, extrinsic to the airway, or has features of both types of lesions (Figure 2) and whether the airway distal to the obstruction is patent. In addition, the length and diameter of the lesions, and the relationship to other structures such as vessels, are assessed to a much higher degree of accuracy. All these features are invaluable in helping the physician determine the appropriate therapy.
Bronchoscopy (either rigid or flexible) is always necessary in assessing airway obstructions. There is debate, however, as to whether the treating physician should routinely perform (and even reperform if initially done by the referring physician) flexible bronchoscopy, or to defer endoscopy to the time of definitive treatment (24, 25). Direct visualization allows the nature and extent of the obstruction to be determined, and provides useful treatment planning information such as the relative amount of intraluminal and extraluminal disease. Most importantly, bronchoscopy allows a tissue diagnosis to be made. The addition of endobronchial ultrasound (EBUS) has been shown to be extremely sensitive for determining the degree of tracheal invasion as well as aiding in planning therapeutic interventions (26–28). In a study by Miyazu and coworkers, EBUS demonstrated extracartilaginous disease in patients initially thought to be appropriate candidates for photodynamic therapy, and therefore other therapies were selected (28). EBUS has also been used to help identify the distal end of an obstructing lesion that would have otherwise been inaccessible with a bronchoscope to facilitate stent placement and plan brachytherapy (29). The largest series describing the use of EBUS in therapeutic bronchoscopy was published by Herth and colleagues (30). EBUS was utilized in 1,174 of 2,446 cases over a 3-year period, including mechanical tumor debridement, stent placement, neodymium:yttrium–aluminum–garnet (Nd:YAG) laser, argon plasma coagulation (APC), brachytherapy, foreign body removal, and the endoscopic drainage of abscesses. EBUS was found to guide or change management in 43%, and changes included selecting proper stent size, guiding tumor debridement, and selecting patients for endoscopic therapy versus surgical therapy. Flexible bronchoscopy may be difficult, and potentially dangerous, when obstruction is severe, as the instrument will further obstruct the remaining lumen and does not allow for ventilatory support. In addition, the use of conscious sedation may depress ventilation and relax the respiratory muscles enough so that a relatively stable airway becomes unstable. As such, access to a team equipped for advanced airway management is essential when undertaking flexible bronchoscopy in these patients.
In the unstable patient, the airway must be immediately stabilized. Endotracheal intubation is indicated to ensure adequate ventilation and oxygenation. This can be accomplished with either a standard endotracheal tube or the rigid bronchoscope. For patients with severe proximal upper airway obstruction, tracheotomy may be the stabilizing procedure of choice (31). Obviously, however, tracheotomy will not alleviate more distal airway obstruction. It is essential that the most experienced physician available manage the airway in a patient with respiratory distress due to CAO. When there is any doubt as to the stability of a patient's airway, we recommend rigid endoscopy as the procedure of choice. The rigid bronchoscope not only provides a secure airway, allowing excellent control of oxygenation and ventilation, but is also a therapeutic tool that allows the passage of various instruments. An obstructing lesion can be "cored out" (32), a stenosis can be dilated with the barrel of the bronchoscope, and therapeutic modalities such as laser, electrocautery, and stent placement can be employed. Certain modalities, such as placement of most silicone stents, can be performed only through a rigid scope. In addition, the barrel of the bronchoscope can be used to tamponade a bleeding central lesion, and large forceps can be used to remove foreign bodies or for mechanical debridement (1, 33, 34). A study by Colt and Harrell reviewed the records of 32 patients with CAO requiring admission to an intensive care unit before therapeutic intervention with a rigid scope (35). Twenty of these patients had an immediate reduction in the level of care following therapeutic rigid endoscopy. Of the 19 patients requiring mechanical ventilation, 10 (53%) had immediate discontinuation of mechanical ventilation, including 7 of the 8 patients (87%) with benign disease (35). The use of heliox, a mixture of 60–80% helium and 20–40% oxygen, should be considered as a bridge in patients with CAO and respiratory distress. Helium has a density one-third that of nitrogen. By reducing the Reynolds number, heliox decreases the tendency for turbulent flow to develop and allows laminar flow to become established more quickly after changes in airway diameter (7, 36, 37). This results in a lower driving pressure required to achieve a given flow, or an increase in flow at the same driving pressure. The reduction in work of breathing may be enough to allow for a more stable intubation, either with an endotracheal tube or the rigid bronchoscope. After the initial stabilization, or in patients who do not need emergent intervention, detailed and careful bronchoscopy and imaging studies are performed to plan additional measures. Secretions are suctioned, and diagnostic tissue is obtained. If a dedicated airway team is not available, patient transfer to a specialized center should be considered at that time. As the number and scope of therapeutic options have increased dramatically, the appropriate measures must be chosen carefully in the context of each patient's situation and the expertise of the team. Expert opinion supports the use of multimodality and multidisciplinary approaches featuring a combination of several interventions to produce effective long-term success (10, 38–40). Figure 4 describes a suggested algorithm in the treatment of tracheobronchial obstructions.
The anesthetic approach to patients with CAO is a critical part of their care. In addition to CAO, coexisting disease, such as pulmonary sepsis and superior vena cava obstruction, contribute to the fact that the majority of these patients are American Society of Anesthesiologists Grade III or IV (41). Patients in severe respiratory distress may not be able to lie supine, and the use of neuromuscular blocking drugs may eliminate the only muscular tone that is keeping the airway patent. There has been debate in the anesthesia literature concerning the benefits of intravenous versus inhalational induction in patients with respiratory distress from CAO. Loss of airway control has been reported using both types of induction. Arguments favoring intravenous induction include a more rapid and smoother induction, as well as less airway irritation (42); however, if induction is too rapid, the airway may be lost. Halothane has been the most widely used anesthetic for inhalation, and is favored by Mathisen and Grillo (32). Sevoflurane, a volatile agent with a low blood gas solubility coefficient (0.68), a pleasant smell, and less airway irritation, has been increasingly suggested as the agent of choice for inhalational induction of anesthesia (41, 43, 44). In the stable patient, the combination of a short-acting intravenous anesthetic agent such as propofol, along with midazolam, fentanyl, and vecuronium, provides effective and safe anesthesia, amnesia, pain control, and muscle relaxation (1). It is necessary that the endoscopist be stationed at the head of the bed throughout induction of anesthesia in case urgent rigid bronchoscopy is required to control the airway. Ventilation can be achieved via a closed system, although we prefer using jet ventilation. Both the endoscopist and anesthesiologist must be aware of the possibility of dynamic hyperinflation (41). This may result from central airway resistance that is more pronounced during exhalation as compared with inhalation, and can be seen with either technique. Biro and coworkers, however, found that in patients with otherwise normal lung parenchyma, near total obstruction of the airway during high-frequency jet ventilation does not lead to dynamic hyperinflation (45). The team should also be alert for the redevelopment of CAO after the procedure, which is typically due to the mobilization of secretions from a previously obstructed bronchus; however, sloughed tissue, blood clot, or a migrated stent may also be the cause (41). Airway fires are another concern in this group of patients. Fires can result from either the use of a flammable anesthetic, or, more commonly, from the use of high concentrations of oxygen in the presence of lasers or electrocautery. In addition, endotracheal tubes and stents can ignite with the use of laser or electrocautery (46, 47). The FIO2 should be less than 0.4 whenever these techniques are used, and it is therefore essential to have constant communication with the anesthesiologist. This is true especially when using jet ventilation through the rigid bronchoscope, as the jet entrains room air with 100% oxygen. The anesthesiologist must therefore hold ventilation for several seconds while the endoscopist uses suction to remove the excess oxygen. Other precautions relating to the choice of endotracheal tubes and treatment of airway fires are discussed in the literature (46–49).
Dilation of the Airways (Bronchoplasty) In emergencies, airways may be dilated with the barrel of the rigid bronchoscope. In less urgent cases sequential dilation, for example, with either balloons or semirigid (Jackson) dilators, is preferred. The major advantage of sequential balloon dilation as compared with dilation with the rigid bronchoscope is that it produces less mucosal trauma and subsequent granulation tissue formation. This technique has been used successfully for patients with airway stenosis following lung transplantation and surgical resection of the airway, in patients with postintubation tracheal stenosis, as well as in patients with malignant airway obstruction (50–53). Before 1991, balloon bronchoplasty was performed with the aid of rigid bronchoscopy or without bronchoscopy under fluoroscopic guidance (54, 55). Several authors, however, have shown it to be safe, effective, and well tolerated in awake patients using flexible bronchoscopy and conscious sedation (50, 51, 56). In Hautmann and colleagues analysis of 78 consecutive patients undergoing balloon bronchoplasty for malignant airway obstruction, 79% of patients had an immediate improvement in airway diameter (56). They suggest balloon bronchoplasty is particularly effective in preparing stenotic airways for stent placement, in expanding stents after insertion, and in the placement of brachytherapy catheters that would otherwise be impeded by high-grade stenoses. Dilation is immediately effective for intrinsic and extrinsic compression, but results are typically not sustained. In fact, mucosal trauma may lead to granulation and accelerate restenosis (39). For this reason, dilation is commonly followed by laser or stenting procedures (52). The main side effect of balloon bronchoplasty is airway rupture resulting in pneumothorax or pneumomediastinum, mediastinitis, and bleeding (50). We have also used the microdebrider, a tool borrowed from our otorhinolaryngology colleagues, to provide mechanical tumor excision in the trachea and mainstem bronchi (57, 58). The microdebrider utilizes a spinning blade contained within a rigid suction catheter to cut while providing suction to remove blood and tumor/granulation tissue. See the online supplement for a brief video demonstrating the use of the microdebrider.
Electrocautery As the majority of commercially available bronchoscopes are not electrically grounded, the bronchoscopist risks becoming the grounding electrode should the unipolar probe tip touch the scope while the current is on (62). Newer bipolar probes have been developed that eliminate this risk, as the current completes the arc through the probe. Electrocautery using the snare device is especially suited to the removal of pedunculated lesions. By cauterizing the stalk, the majority of tissue can be removed without destruction, and therefore becomes available for pathologic review (61). It has been used with curative intent in patients with early stage and intraluminal squamous cell lung cancer (64) as well as in advanced malignancies, in combination with other modalities (59, 65, 66). Side effects of electrocautery include bleeding, airway perforation, endobronchial fire, as well as damage to the bronchoscope. Coulter and Mehta prospectively evaluated the influence of electrocautery on the need for laser photoresection (63). Forty percent of the lesions evaluated were believed to be amenable to electrocautery, and the procedure was successful in 89% of these cases. van Boxem and colleagues documented that the amount of mucosal damage visualized after electrocautery has good correlation with histologic tissue damage (67). They suggest this as a potential benefit over other therapies, such as laser, where the actual histologic damage may be more severe than what is visualized after the procedure. Disadvantages include a loss of effectiveness with bleeding due to the diffusion of the current across a larger surface area, and the fact that more debridement and probe cleaning are required as compared with laser therapy (62, 63). See the online supplement for a brief video demonstrating the use of endobronchial electrocautery.
Argon Plasma Coagulation Although generally used for malignant CAO, there are case reports of using APC to treat airway obstruction resulting from the formation of granulation tissue at the site of surgical anastomosis (73, 74), as well as respiratory papillomatosis (75). APC has also been used successfully in patients with airway stents without apparent damage to the endoprosthesis (71, 74). See the online supplement for a brief video demonstrating the use of endobronchial APC.
Laser Therapy Dumon's group published the first large series using Nd:YAG laser photoresection for both benign and malignant airway disease in 111 patients (80). Their best results were obtained in patients with malignant disease. For patients with benign tracheal stenosis they advocated a combination of laser therapy and "gentle" rigid dilation. This combined approach for tracheal stenosis has also been supported by Shapshay and coworkers (38) and Mehta and coworkers (39). Cavaliere's group has published the largest series to date. In their initial publication, they reviewed their results in treating 1,000 patients (78). Of the 649 patients with malignant CAO, 70% had squamous cell carcinoma, 7% had adenocarcinoma, 5% had small cell carcinoma, and 4% had large cell carcinoma; 15 patients had metastatic disease from a nonlung primary tumor. Their success rate depended more on the location of the tumor than on the pathologic tumor type. Tumors in the trachea, right mainstem bronchus, and bronchus intermedius were treated successfully 97, 94, and 90% of the time, respectively. This compares with a success rate of 86% for tumors in the left mainstem bronchus, and 58% for tumors in the left upper lobe (78). In a follow-up article, encompassing 2,253 treatments in 1,585 patients, radiographic improvement was noted in 93% of patients with bronchogenic carcinoma, and their overall complication rate was 2.3% (81). Table 2 summarizes the major publications of laser therapy for benign and malignant CAO. It can be seen that this modality is extremely effective for restoring airway diameter and improving symptoms such as dyspnea and hemoptysis.
As compared with historical controls, laser resection for malignant airway obstruction is associated with improved survival rates (60% of patients alive at 7 months in one study) (82–84). Shea and colleagues found that the combination of Nd:YAG laser therapy with endobronchial brachytherapy was more effective than Nd:YAG therapy alone in patients with squamous cell carcinoma (85). As these studies were not randomized trials, it is impossible to conclude that the laser therapy was responsible for the improved survival. It is more likely that laser therapy provided a rapid and safe means of relieving CAO, which is associated with a high mortality rate. This is supported by data from the Cleveland Clinic (Cleveland, OH), where survival for patients receiving emergent Nd:YAG laser therapy was significantly longer than that of historic control subjects who received only emergent external beam radiation therapy (267 vs. 150 days, p = 0.04) (5). Lesions most amenable to laser therapy are central, intrinsic, and short (< 4 cm), with a visible distal endobronchial lumen. When lesions meet these criteria, patency can be reestablished in more than 90% of cases (78, 83, 86). With appropriate precautions, the safety record of laser therapy is excellent; significant complications develop in fewer than 5% of cases, and in a summary of close to 7,000 laser treatments, the overall complication rate was 0.99% (49, 76–79, 82, 83, 86, 87). Local disease recurrence with airway obstruction is typical unless tumor debulking is followed by radiation therapy, photodynamic therapy, or stenting. It has been shown that the application of 10 W at a distance of 10 mm is safe for treating granulation tissue around uncovered metal stents and silicone stents; however, covered metal stents can ignite and fracture at this delivered energy (87). Many authors perform laser photoresection under general anesthesia and use rigid equipment for control of ventilation and bleeding, to aid in airway debridement, and to dilate stenotic segments (49, 71, 77, 79, 86, 88, 89). Flexible scopes can be inserted through the rigid scope to apply the laser therapy, especially in more distal lesions. See the online supplement for a brief video demonstrating the use of endobronchial laser therapy.
Photodynamic Therapy The amount of energy delivered depends on the duration of light treatment (93). Many authors use about 200 J/cm treated (400 mW/cm length of diffuser for 500 seconds) during the initial treatment session (94, 95). This typically takes about 8 minutes, and can therefore be easily applied via outpatient flexible bronchoscopy with conscious sedation and local anesthesia. Cell death is achieved via a Type II photooxidation reaction. This reaction describes the formation of singlet oxygen, as opposed to a Type I reaction, which generates free radicals (91). The singlet oxygen then causes vascular stasis followed by edema and hemorrhage. As cytotoxicity is delayed, follow-up bronchoscopies are necessary to remove secretions and cellular debris from the airways. PDT can be curative for early-stage lung cancer of the airways, and if used for carcinoma in situ, the complete remission rate may be as high as 83% (96). This is therefore an extremely attractive option for treating patients who are not surgical candidates. PDT has been shown to obviate the need for surgical resection in up to 22% of patients with early-stage non–small cell lung cancer (97). In a prospective 14-year study, McCaughan and coworkers treated 175 patients with endobronchial tumors with PDT (98). The median survival for all patients was 7 months. For the 16 patients with Stage 1 disease, median survival was not reached. As compared with historical data indicating a 5-year survival of 52–62% for patients with Stage 1 disease, the patients in this series had a 5-year survival of 93%. In addition to achieving a length of palliation equal to or better than that of historical control subjects, the authors suggest PDT as an alternative treatment for Stage 1 non–small cell lung cancer, especially for patients who are considered to have a high surgical risk (98). These data obviously need to be interpreted understanding that at this time, surgery remains the "gold standard" for patients with Stage 1 non–small cell carcinoma; however, in the patient who is not a good surgical candidate, PDT remains an effective therapeutic option. Tables 2 and 3 summarize the PDT literature concerning patients with malignant CAO, as well as CAO due to typical carcinoids (99) and laryngeal papillomas (100).
McCaughan's group compared outcomes of 102 patients with malignant airway disease who were treated with CO2 laser, YAG laser, and PDT (60). The only seven patients to achieve "no evidence of disease" status had received PDT. Similarly, Diaz-Jiménez and coworkers randomized 31 patients with malignant CAO to treatment with the Nd:YAG laser or to PDT (101). Although there was no difference in symptom improvement between the groups, patients treated with PDT had significantly longer time until treatment failure (50 vs. 38 days, p = 0.03), and median survival (265 vs. 95 days, p = 0.007). These favorable results may have been due to differences in tumor staging at the time of randomization, with the PDT group containing fewer patients with advanced disease (101). As with most other therapies for CAO, PDT is most effective in patients with intraluminal disease. Mean duration of complete response is 22 weeks if more than 50% of the obstruction is due to mucosal disease, as compared with only 7 weeks if the tumor is submucosal or external (91). The major benefit of PDT is that it can easily be applied in patients who are not candidates for surgery or general anesthesia. In addition, there is generally complete healing of the normal tracheobronchial tree, with less risk of perforation and minimal effects on cartilage as compared with other therapies (102). As PDT uses a nonthermal laser, there is no risk of endobronchial fire or electrical exposure. Side effects are few and include bleeding and airway obstruction from necrotic tumor and blood. The main adverse effect is skin photosensitivity, which, as stated above, may last for 6 weeks. It is therefore crucial to educate the patient before initiation of therapy. Newer drugs are being developed with the hopes of increasing tumor selectivity and reducing the duration of skin phototoxicity (103). The use of EBUS, as suggested by Miyazu and colleagues (28), to help determine the extent of disease before injecting the patient with the photosensitizer might be beneficial in the precise delivery of the laser light. See the online supplement for a brief video demonstrating the use of endobronchial PDT.
Cryotherapy There is experimental evidence that cryotherapy also produces beneficial effects on the immune system, although the mechanisms remain largely speculative. Maiwand and Mathur support the notion that cryotherapy can induce proliferation of a population of natural killer cells (105). Other studies, in patients with prostate cancer and melanoma, have found a reduction in distant disease with local cryotherapy, possibly due to an increase in lymphocyte activation in the peripheral circulation (106, 107). The Joule–Thomson effect describes the decrease in temperature that is observed during the expansion of gas from a high-pressure to a low-pressure environment. Although CO2 can be used, it produces a "snow" when it expands at atmospheric temperatures, which may clog cryoprobes. Nitrous oxide (N2O), which is stored at room temperature under high pressure, is the most commonly used gas for cryotherapy. When N2O is released at the tip of the cryoprobe, the temperature falls to –89°C within several seconds. Liquid nitrogen has also been used; however, maximal negative temperatures are achieved after 1–2 minutes, thus creating less cellular injury as compared with N2O (104). Cryotherapy has been used to successfully treat both benign and malignant CAO (Table 4) . It is effective in reducing or eliminating hemoptysis due to malignant disease in up to 93% of patients (108), and Maiwand and Homasson recommend cryotherapy as a first-line treatment in patients with posttransplant anastomotic strictures (104).
Cryotherapy may have synergistic effects with chemotherapy or radiation (109, 110). In one of the only studies using survival as a key end point, Vergnon and colleagues treated 38 patients with unresectable non–small cell lung cancer with a combination of cryotherapy and external beam radiation therapy (110). The 26 patients (65%) who had favorable results had a median survival of 397 days. Seventeen patients in this group had no evidence of residual tumor after irradiation, and experienced a median survival of 560 days. The historical control subjects treated solely with radiotherapy, on the other hand, had only a 35% local control rate (110). Again, when comparisons to historical control subjects are made, it is important to understand the selected nature of the patients under investigation. Cryotherapy has also been used to remove foreign bodies and blood clots in the airways (111). Freezing the object to the probe tip allows retrieval of the foreign body along with removal of the cryoprobe and bronchoscope unit from the airways. It is especially useful for objects with high water content, such as grapes, and friable objects that may otherwise disintegrate when grasped with forceps (104). The primary advantage of cryotherapy lies in its relative safety. As freezing and recrystallization depend on cellular water content, cartilage and fibrous tissue are relatively cryoresistant (104, 108). The incidence of airway perforation is therefore markedly reduced. Bleeding also tends to be less common because of the hemostatic effects of cryotherapy. In addition, cryotherapy is not associated with the risk of airway fires, electrical accidents, or radiation exposure. The major disadvantage of cryotherapy is that its maximal effects are delayed, and it should therefore not be used to treat patients with acute, severe airway obstruction (108, 112). Some authors feel that although patients may experience significant improvement after the first treatment, additional benefit is frequently observed after a second or third treatment (113). It is also necessary to perform a follow-up bronchoscopy in 2–4 days to remove sloughed tissue, and therefore the patient is exposed to several procedures (112, 114). Cryotherapy can be used via both the rigid and flexible bronchoscopes, and rigid, semirigid, and flexible probes are commercially available. As the size of the probe tip is proportional to tissue injury, many authors recommend rigid bronchoscopy (104); however, flexible bronchoscopy is certainly more accessible to the majority of pulmonologists, and good results can also be obtained with the flexible cryoprobe (112). When using the flexible bronchoscope, it is crucial to have the probe protrude several millimeters from the distal tip of the scope, so as not to freeze the video chip. Approximately three 60-second freeze–thaw cycles are performed in each area. Some authors recommend more frequent applications of shorter (30 second) freeze–thaw cycles (105). After the probe is thawed and easily detached from the tissue, it is then moved several millimeters, for another three cycles, such that there will be a slight area of overlap (105, 112).
External Beam Radiation and Brachytherapy The technical aspects of brachytherapy have been extensively reviewed (123–126), and most authors currently recommend using the afterloading technique. In this technique, a blind-tipped catheter is placed at the desired position and the radiation source is then loaded afterward. A major benefit of this method is the ability to use higher intensity isotopes without exposing the endoscopy staff to radiation (118). Brachytherapy may be delivered by either low-dose rate (LDR), intermediate-dose rate (IDR), or high-dose rate (HDR) methods (85, 117, 119). There is no consensus regarding dose rate and cumulative dose in distinguishing LDR, IDR, and HDR brachytherapy. LDR therapy is arbitrarily defined as delivering 75–200 cGy/hour (117, 119, 123), and involves placing the radiation source adjacent to the lesion for 20–60 hours. A typical, although arbitrary, cumulative dose of 3,000 cGy at a radius of 10 mm in the trachea, and 5 mm in the bronchi, is commonly applied (123). LDR brachytherapy, therefore, requires hospitalization, and the typical treatment is one session. IDR uses fractions of 200–1,200 cGy/hour, with each session lasting 1–4 hours, and cumulative total doses similar to LDR (119, 127, 128). HDR delivers more than 1,000–1,200 cGy/hour (119, 128). A typical regimen delivers 500 cGy at 10 mm, with an average of three fractions at weekly intervals, and a total dose of 1,500 cGy (124), although fractions of up to 1,000 cGy have also been used (129). Each fraction lasts between 3 and 30 minutes, and can therefore be given on an outpatient basis. Multiple bronchoscopies, however, are required for repeated catheter placement (85). Table 5 summarizes the results of studies investigating the use of brachytherapy for malignant CAO. There have been no head-to-head, comparison trials of LDR versus IDR versus HDR brachytherapy. As all forms offer significant palliation (128), some authors suggest LDR for inpatients with poor performance status, or for those who travel a great distance to the treatment hospital, and HDR for outpatients who live relatively close to the hospital (127).
A distinct advantage of brachytherapy is that the catheters can be placed in the upper lobe bronchi, as well as in segmental bronchi, areas typically inaccessible to laser therapy. Endobronchial radiotherapy has also been used successfully in patients with peribronchial disease, and patients often require less retreatment for disease recurrence (122). Disadvantages to brachytherapy include intolerance of the catheter; excessive radiation-induced bronchitis; cough; fistula formation between the esophagus, pleura, or great vessels; hemorrhage; and infection, including abscess formation (123, 124, 130, 131). Although the risk of fatal hemorrhage was 21% in Macha's study, they believed this was not secondary to the HDR brachytherapy itself, but more likely was due to the underlying histology of the tumor (squamous in the majority of patients) and disease progression. The incidence of hemoptysis appears to be associated with the location of the tumor/treatment. Treatment of tumors in the right and left upper lobes has the highest incidence of hemoptysis, as these are located near the great vessels (125, 130, 132). Treatment planning, with CT imaging, should therefore be used before the initiation of brachytherapy. Potential future applications may include treatment for carcinoma in situ and nonmalignant airway obstruction from excessive granulation tissue and tracheobronchial amyloid (133–136). There are also studies underway examining the use of drugs such as paclitaxel as radiosensitizing agents (132).
Airway Stents The development of airway stents began when Montgomery introduced a silicone T-tube in 1965, for use in patients with tracheal stenosis (139). In 1982, Westaby and coworkers modified the Montgomery T-tube and designed the T-Y tube with the intent of providing airway stability in patients with tracheomalacia and granulomatous obstruction of the distal trachea and mainstem bronchi (140). The primary disadvantage of both of these tubes is that they require a tracheotomy to anchor the horizontal limb of the stent. The first dedicated, completely endoluminal airway stent, however, was introduced by Jean-François Dumon in 1990 (141). His team placed 118 studded silicone prostheses in 66 patients with airway obstruction due to both benign and malignant disease. Immediate relief of respiratory symptoms and significant quality survival were achieved in all but two patients, and none of the patients who died did so from airway obstruction. Since that time, there have been numerous different designs, each of which exhibits various advantages and disadvantages (142). The ideal stent should be: (1) easy to insert and remove, yet not migrate; (2) of sufficient strength to support the airway, yet flexible enough to mimic normal airway physiology and promote secretion clearance; (3) biologically inert to minimize the formation of granulation tissue; and (4) available in a variety of sizes. There are currently two main types of stents: metal and silicone. Although metal stents are easily placed, they can be extremely difficult to extract (Figure 3) . Metal stents are available in covered (typically with Silastic or polyurethane) and uncovered varieties. For malignant airway obstruction, the only appropriate metal stents are covered models, which prevent tumor ingrowth. Some authors believe that there is no indication for an uncovered metal stent (143). A distinct advantage of metal stents is their larger internal:external diameter ratio as compared with silicone stents. In addition, many centers routinely place metal stents via flexible bronchoscopy. Silicone stents, on the other hand, require rigid bronchoscopy for placement, but they are more easily removed and are significantly less expensive. The rate of stent migration, however, tends to be higher with silicone stents than metal stents. Current research aims to combine advantages of both stent designs (144, 145). The Polyflex stent utilizes a polyester mesh embedded in silicone to achieve the luminal patency of metal stents with the benefits of easy removal of silicone stents. Stent migration, however, can be a problem, and a studded version has been shown to decrease this complication (146). An important note to those who place stents is that the Dumon stents are packaged with a reading for the outer diameter (not including the studs), whereas Polyflex and Ultraflex stents are packaged with a reading for the inner diameter.
Some researchers are investigating the efficacy and tissue compatibility of bioabsorbable stents, made either of Vicryl filaments (147) or poly-L-lactic acid (148–150). These stents may be particularly useful in the pediatric patient, as the internal diameter of silicone stents can be too narrow, and metal stents are difficult to remove once epithelialized (149). The Gianturco, Palmaz, and Strecker stents (16, 151–153) are now rarely used in airway stenting. The most commonly used metal stents are made from nitinol (143). Nitinol is a super-elastic biomaterial that has the ability to undergo large deformations in size and shape (154). In addition, nitinol has "shape memory," that is, at cold temperatures the stent is easily deformable, and at higher temperatures (i.e., body temperature) it regains its original shape (152). The risk of airway perforation seems to be lower with nitinol stents because they do not change length once expanded and are flexible enough to change shape with a cough, yet have excellent radial strength during constant compression by tumor or stenoses (155). Nonmetallic stents are generally made from molded silicone, and are shaped to prevent migration (141) or contain polyester wire mesh embedded in silicone (143). Dynamic stents contain metal struts embedded in silicone and are Y-shaped. Silicone stents are commonly placed with the aid of a specially designed stent introducer system in which the stents are preloaded into the introducer and inserted into the stricture with the aid of a stent pusher. The Dumon stent is currently the most widely used stent throughout the world, and some believe that it is the "gold standard" to which future stents will have to be compared (143). In a well designed study using ex vivo tracheas, Freitag's group evaluated the mechanical properties of several stents (156). Excised human and minipig tracheas were subjected to area-versus-pressure, area-versus-load, and force-versus-compression measurements, and the stress–strain relationship of various stents was measured. The Dumon stent was found to be the most collapsible, and the authors recommend using either a metallic or hybrid stent for very rigid stenoses. However, both the Dumon and dynamic stents had excellent force–compression relationships resulting in a smooth distribution of pressure to the airway surface. The authors recommend that, as the ends of stents tend to collapse more readily than do the central portions, the stent exceed the stenosis by 5 mm on each end (156). Table 6 summarizes the outcomes of several clinical trials examining the efficacy and complications of airway stenting.
It is not clear whether stenting may be beneficial for some or all cases of tracheobronchial malacia (TBM) with symptoms of airway obstruction (157, 158). The dynamic characteristics of TBM are quite different from static causes of CAO and therefore the forces placed on the stents are also different. The shape of the airway in patients with TBM is different from that of the normal trachea, and also different from the typical cylindrical shape of most stents, thus altering the surface contact dynamics between stent and airway. We recommend performing dynamic flexible bronchoscopy, that is, bronchoscopy with active inhalation and exhalation, on all patients with suspected TBM. Newer CT scanning techniques are proving extremely helpful in noninvasively diagnosing patients with TBM (16, 20, 21). The radiographic appearance of TBM during dynamic breathing has been shown to correlate extremely well with the bronchoscopic appearance (20, 21). One of the problems in stenting patients with TBM is in identifying the site of flow limitation. Placing a stent in one segment that appears to collapse maximally on exhalation may just result in the site of flow limitation moving more distally (16), and studies are currently underway to help predict physiologically who will benefit most from stent placement. A problem with stent placement in patients with TBM is the forces placed on the stent by the sometimes dramatic changes in airway size and shape. This often results in an excessive incidence of stent fracture of metal endoprosthesis and possible airway perforation (157, 159, 160). As such, some authors recommend a more conservative approach for patients with TBM and use continuous positive airway pressure to act as a pneumatic stent (161, 162). The obvious disadvantage of this approach is the loss of airway support without the continuous positive airway pressure. Details of stent placement techniques have been reviewed (152). Metal stents are easily placed with the flexible bronchoscope (152, 155, 163–165). Although most authors recommend using fluoroscopic guidance, studies suggest this is not necessary (166). Silicone stents can also be inserted via flexible bronchoscopy (167). Many authors, however, prefer to place both silicone and metal during rigid bronchoscopy as it provides control of the airway, allows the use of other modalities, facilitates removal or repositioning of the stent, and avoids the problems encountered during prolonged procedures when topical anesthesia wears off (3, 137, 168, 169). Miyazawa's group used spirometry and ultrathin bronchoscopy to identify the choke point in 64 patients with malignant CAO (170). They also used EBUS to identify the precise size of stent to be placed, as well as the cause of dynamic collapse and choke point migration. All patients had significant improvements in dyspnea and spirometry after stenting, and the authors suggest that the correct positioning of the stent at the choke point leads to maximal symptomatic benefit in these patients. In addition, advanced techniques, such as EBUS and ultrathin bronchoscopy, can identify choke point migration after initial stent placement, and prompt placement of a second stent (170). It is crucial that the indications for stent placement be clear, that the appropriate stent be selected, that an endoscopist with significant experience inserts the stent, and that the patient be provided with appropriate education and follow-up. Especially in the case of benign CAO, placement of a metal stent should only be undertaken if no other therapeutic options, including surgical correction, exist. See the online supplement for a brief video demonstrating the use of endobronchial stenting.
Surgical Resection For malignant CAO, it has been shown that initial endoscopic relief can be followed by better preoperative preparation of the patient, a more appropriate surgical technique, as well as improved surgical results (6, 177, 178). For example, infections can be treated, the tumor staging may be downgraded because of resolution of atelectasis, and if neoadjuvant chemotherapy is indicated it can be give more safely once the atelectasis has resolved. Although in development, there are currently no prosthetic tracheal substitutes. The location of the pathology is extremely important in determining the ability to offer surgical resection. Up to half of the cervical or intrathoracic trachea can be resected, as compared with only about 4 cm of the airway at the level of the carina (6). The operative morbidity and mortality are also higher for carinal surgery (172). Radial extension into adjacent structures typically precludes resection, and therefore careful preoperative planning with techniques such as EBUS is crucial. Recent developments in reabsorbable stent technology may be used in conjunction with surgery for the treatment of TBM; however, their exact role remains to be defined (150).
The refinement in airway surgery over the last 50 years, combined with advances in immunosuppression, has allowed an increasing number of patients with end-stage lung disease to benefit from lung transplantation. The initial attempts at transplantation, however, were limited by anastomotic failure in up to 80% of cases (179). It is thought that bronchial ischemia is the underlying cause of airway complications. In the first 1–4 weeks after transplantation, blood supply to the donor bronchus is dependent on retrograde perfusion via the pulmonary arteries to the bronchial circulation. Despite significant advances in surgical technique, airway complications after lung transplantation continue to pose a significant problem. Detection of airway complications generally requires bronchoscopy—symptoms and spirometry are generally nonspecific. Although advanced airway imaging with virtual bronchoscopy has been studied, the overall accuracy approached only 80% (180). In an analysis of 348 anastomoses, the St. Louis group found airway complications necessitating intervention in 9.5%. Five of the 229 patients studied died as a result of their airway complication. Features associated with the development of these complications included single (versus bilateral) transplantation, the use of simple interrupted suturing, and a longer duration of postoperative mechanical ventilation. Over the 6-year course of their series, the airway complication rate fell from 14.3 to 4%, and the authors suggest better surgical technique, surveillance of rejection, and improved maintenance immunosuppression as the factors leading to this improvement (181). In other series, airway complications remain significantly higher. Herrera and colleagues, in a study of 147 bronchial anastomoses, found an overall complication rate of 23.8% per anastomosis (182). Complications included granulation tissue, fibrotic stricture, bronchomalacia, mucosal slough, and infection. Risk factors, both for the donor and recipient, could not be identified, aside from the isolation of Aspergillus fumigatus during the first 30 days. It is impossible to know, however, whether Aspergillus is the cause or a consequence of airway complications. In a similar review of 127 anastomoses by Kshettry and coworkers, the overall airway complication rate was 15% (179). Interestingly, they found an almost 50% reduction in anastomotic complications after changing from a technique using an omental wrap, to the telescoping technique (p = 0.15, potentially due to a small sample size) (179). Almost all the techniques described above, including mechanical debridement, balloon dilation, Nd:YAG laser, and brachytherapy, have been used to treat airway complications in lung transplant recipients (136, 179, 182–186). Anastomotic dehiscence has also been treated successfully with cyanoacrylate glue (187). As with almost all studies in this field, there have been no randomized trials comparing different treatment modalities, and treatment primarily depends on the experience of the treating physician/team, and available resources. The endoscopist must be aware of the likelihood of concomitant disease, such as infection and stenosis, before embarking on a treatment plan. Clearly, underlying infection should be adequately treated before the insertion of a stent. In a retrospective analysis by Burns and coworkers, infectious complications actually decreased after stent placement, possibly because of improved mucociliary clearance (186). In addition, although airway perforation can be a serious complication of laser therapy in a normal host, it may be devastating in the immunocompromised patient. Although initially associated with an increased mortality rate, the study by Burns's group suggests a mortality equivalent to patients without airway complications (186).
All patients with a history of airway obstruction should carry a card or bracelet identifying them as patients with complicated airways or indwelling airway stents. The presence of a complicated airway, however, does not preclude intubation, if needed. On completion of the procedure used to relieve the CAO, the large majority of patients can be extubated. Some patients, however, may have been in respiratory failure for some time before the intervention, and may have limited pulmonary reserve. These patients, despite relief of the CAO, may benefit from a brief period of positive pressure ventilation (25). Close follow-up |
TOP
ABSTRACT
CONTENTS
