0
Original Research: Pulmonary Procedures |

Complications Following Therapeutic Bronchoscopy for Malignant Central Airway ObstructionComplications of Therapeutic Bronchoscopy: Results of the AQuIRE Registry FREE TO VIEW

David E. Ost, MD, MPH, FCCP; Armin Ernst, MD, FCCP; Horiana B. Grosu, MD; Xiudong Lei, PhD; Javier Diaz-Mendoza, MD, FCCP; Mark Slade, MBBS, FCCP; Thomas R. Gildea, MD, FCCP; Michael Machuzak, MD, FCCP; Carlos A. Jimenez, MD, FCCP; Jennifer Toth, MD; Kevin L. Kovitz, MD, FCCP; Cynthia Ray, MD; Sara Greenhill, MD, FCCP; Roberto F. Casal, MD; Francisco A. Almeida, MD, FCCP; Momen Wahidi, MD, FCCP; George A. Eapen, MD, FCCP; Lonny B. Yarmus, DO, FCCP; Rodolfo C. Morice, MD, FCCP; Sadia Benzaquen, MD; Alain Tremblay, MDCM, FCCP; Michael Simoff, MD, FCCP; on behalf of the AQuIRE Bronchoscopy Registry
Author and Funding Information

From the Pulmonary Department (Drs Ost, Grosu, Jimenez, Eapen, and Morice) and the Department of Biostatistics (Dr Lei), The University of Texas MD Anderson Cancer Center, Houston, TX; the Reliant Medical Group (Dr Ernst), Boston, MA; the Department of Pulmonary and Critical Care Medicine (Drs Diaz-Mendoza and Simoff) and the Department of Internal Medicine (Dr Ray), Henry Ford Hospital, Detroit, MI; the Department of Thoracic Oncology (Dr Slade), Papworth Hospital, Cambridge, England; the Department of Pulmonary, Allergy, and Critical Care (Drs Gildea and Almeida), Cleveland Clinic Foundation, Cleveland, OH; the Department of Pulmonary Medicine (Dr Machuzak), Cleveland Clinic, Cleveland, OH; the Penn State Cancer Institute (Dr Toth), Hershey, PA; the University of Illinois Hospital & Health Sciences Center (Dr Kovitz), Chicago, IL; Chicago Chest Center Interventional Pulmonology (Dr Greenhill), Elk Grove Village, IL; the Department of Pulmonary, Critical Care, and Sleep Medicine (Dr Casal), Baylor College of Medicine, Houston, TX; the Department of Internal Medicine (Dr Wahidi), Division of Pulmonary, Allergy and Critical Care Medicine, Duke University Medical Center, Raleigh, NC; the Division of Pulmonary and Critical Care Medicine (Dr Yarmus), Johns Hopkins University, Baltimore, MD; the Division of Pulmonary, Critical Care and Sleep Medicine (Dr Benzaquen), University of Cincinnati, Cincinnati, OH; and the Department of Medicine (Dr Tremblay), University of Calgary, Calgary, AB, Canada.

CORRESPONDENCE TO: David E. Ost, MD, MPH, FCCP, Pulmonary Department, The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Unit 1462, Houston, TX 77030; e-mail: dost@mdanderson.org


FUNDING/SUPPORT: The American College of Chest Physicians (CHEST) funded the database construction for the AQuIRE program. This research was supported in part by the National Institutes of Health through a Cancer Center Support Grant [Grant P30CA016672], Biostatistics Core, MD Anderson Cancer Center.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details.


Chest. 2015;148(2):450-471. doi:10.1378/chest.14-1530
Text Size: A A A
Published online

BACKGROUND:  There are significant variations in how therapeutic bronchoscopy for malignant airway obstruction is performed. Relatively few studies have compared how these approaches affect the incidence of complications.

METHODS:  We used the American College of Chest Physicians (CHEST) Quality Improvement Registry, Evaluation, and Education (AQuIRE) program registry to conduct a multicenter study of patients undergoing therapeutic bronchoscopy for malignant central airway obstruction. The primary outcome was the incidence of complications. Secondary outcomes were incidence of bleeding, hypoxemia, respiratory failure, adverse events, escalation in level of care, and 30-day mortality.

RESULTS:  Fifteen centers performed 1,115 procedures on 947 patients. There were significant differences among centers in the type of anesthesia (moderate vs deep or general anesthesia, P < .001), use of rigid bronchoscopy (P < .001), type of ventilation (jet vs volume cycled, P < .001), and frequency of stent use (P < .001). The overall complication rate was 3.9%, but significant variation was found among centers (range, 0.9%-11.7%; P = .002). Risk factors for complications were urgent and emergent procedures, American Society of Anesthesiologists (ASA) score > 3, redo therapeutic bronchoscopy, and moderate sedation. The 30-day mortality was 14.8%; mortality varied among centers (range, 7.7%-20.2%, P = .02). Risk factors for 30-day mortality included Zubrod score > 1, ASA score > 3, intrinsic or mixed obstruction, and stent placement.

CONCLUSIONS:  Use of moderate sedation and stents varies significantly among centers. These factors are associated with increased complications and 30-day mortality, respectively.

Malignant airway obstruction is a serious complication of lung cancer, resulting in dyspnea, decreased functional status, and asphyxiation risk. In addition, pulmonary metastases from other malignancies, including breast, colon, and renal cell cancer, commonly result in malignant airway obstruction.1 There are three main types of malignant airway obstruction: endobronchial obstruction, extrinsic compression, and mixed pattern. For endobronchial obstruction, ablative techniques that destroy tissue are indicated, including lasers, electrocautery, argon plasma coagulation (APC), photodynamic therapy, microdebriders, and cryotherapy. For extrinsic compression, stents are used to strengthen the bronchial wall and keep the airway open. For mixed patterns, ablation followed by stenting is usually required. Treatment strategies often are multimodal, and variations exist in how physicians perform therapeutic bronchoscopy.

Prior studies of therapeutic bronchoscopy for central airway obstruction212 have included both malignant and benign cases, and most were done retrospectively, although some have focused on malignant disease.1318 Reported complication rates are low, but complications and outcomes differ significantly depending on the indication for the procedure (ie, malignant vs benign disease, isolated hemoptysis vs central airway obstruction), and in most studies, significant heterogeneity existed in terms of patient population and indications.1,4 Many of these studies focused on individual technologies, such as stents, microdebriders, or APC, and most were performed at centers of excellence as part of ongoing research programs. Whether variations in practice patterns affect complication rates is unknown and cannot be answered by single-center studies. In addition, because many previous studies had relatively small sample sizes, formal analysis of rare events like complications has been limited. Whether these results can be generalized to everyday clinical practice is unknown. Additional outcomes data on therapeutic bronchoscopy for malignant central airway obstruction in everyday clinical practice is, therefore, needed to establish benchmarks for quality improvement and clinical effectiveness.

Registries are well suited for this purpose because they provide a more generalizable picture of outcomes and clinical effectiveness. We used the American College of Chest Physicians (CHEST) Quality Improvement Registry, Evaluation, and Education (AQuIRE) program to evaluate therapeutic bronchoscopy for malignant central airway obstruction, focusing on complications and their clinical consequences and 30-day outcomes. The primary objective was to quantify the incidence of and risk factors for complications. The secondary objective was to quantify the incidence and risk factors for bleeding, hypoxemia, respiratory failure, and 30-day mortality and to evaluate the consequences of complications as measured by escalation in level of care and associated adverse events. Data regarding the success rate of therapeutic bronchoscopy and its impact on dyspnea and quality-adjusted survival have been presented separately.19

Data on patients undergoing therapeutic bronchoscopy from January 2009 to February 2013 were entered into the AQuIRE program.20 Not all centers started participating at the same time; some centers participated for the entire duration, whereas others participated for ≥ 1 year. However, participating centers agreed to enter all consecutive patients for the duration of their participation. Institutional review board approval was governed by each site (see e-Appendix 1 for details). The principal investigator for each site was primarily responsible for data quality for that site. Informed consent or a waiver of consent was obtained in accordance with institutional guidelines. Data were entered through the AQuIRE web-based interface using standardized definitions, quality control checks, and protocols as previously described.2123

Patients undergoing therapeutic flexible or rigid bronchoscopy for malignant central airway obstruction were included. Central airway obstruction was defined as occlusion of ≥ 50% of the trachea, mainstem bronchi, bronchus intermedius, or lobar bronchus. Because a registry was used, all clinical decisions, including type of intervention, were left to the discretion of the attending bronchoscopist.

Information extracted from the AQuIRE program included patient demographics, clinical characteristics, physician and hospital information, sedation information, procedural information, complications, outcomes of complications, adverse events, and 30-day survival. See e-Appendix 1 contains for additional details on definitions. The average number of cases per month was used as a center-level variable in hierarchical models to assess the relationship between center volume and outcomes. Adverse events were defined as events resulting from the complication that extended hospitalization, required intervention to avoid permanent impairment, were life threatening, or caused disability or death.

The primary outcome was procedural morbidity, which was defined as any of the following events within 24 h: bleeding requiring intervention, pneumothorax, refractory hypoxemia defined as oxygen saturation < 90% for > 1 min,1,22 clinically significant airway injury, hypotension, cardiac arrest or arrhythmia, or unexpected respiratory failure requiring positive pressure ventilation. Secondary outcomes included subsets of complications (ie, bleeding, hypoxemia, respiratory failure), clinical consequences of complications as measured by complications associated with adverse events, need for an escalation in level of care, and death within 30 days. Escalation in level of care, as a measure of resource utilization, was defined as admission to the hospital if the patient was an outpatient, transfer to the ICU if the patient was an inpatient, or death irrespective of whether there was a complication or adverse event (see e-Appendix 1 for additional details). For the outcome of death within 30 days, we analyzed only first therapeutic bronchoscopy cases. The reason is that the probability of death within 30 days of the procedure for patients having more than one procedure would not be independent of each other.

Statistical Analysis

For binary outcomes, the association of the outcome with each covariate was checked by χ2 or Fisher exact test (for categorical variables) or Wilcoxon-Mann-Whitney two-sample test (for continuous variables) as appropriate. For continuous outcomes, the association of the outcome with each covariate was checked by the F test in analysis of variance models (for categorical variables) or t test in linear regression models (for continuous variables) as appropriate. Variables with P < .20 on univariate analyses were candidates for multivariate models. A backward model selection was applied to retain only variables where P < .05. We used hierarchical models to evaluate the impact of center-level variation on homogeneity. Logistic regression using the maximum likelihood approach was used when there were sufficient numbers of events for the outcome; otherwise, logistic regression with the Firth penalized likelihood approach was applied if there were rare events.24,25 For continuous outcomes, an analysis of covariance model was used. P < .05 was considered significant; all tests were two-sided. All statistical analyses were performed with SAS version 9.3 software (SAS Institute Inc).

Fifteen centers with 26 physicians enrolled 947 patients who underwent 1,115 procedures. Patients with lung cancer either had advanced stage disease to begin with or had a recurrence after initial treatment that was unresectable at the time of intervention. Baseline patient and clinical characteristics are shown in Table 1.

Table Graphic Jump Location
TABLE 1 ]  Patient and Clinical Characteristics

Data are presented as mean ± SD or No. (%). APC = argon plasma coagulation; ASA = American Society of Anesthesiologists; GERD = gastroesophageal reflux disease; HD = hemodialysis.

a 

Patients could have multiple disease locations, multiple types of obstruction, multiple ablative techniques, and multiple types of stents; therefore, these are not mutually exclusive.

b 

Patients having a second primary cancer present other than the one causing obstruction.

c 

Spontaneous-assist ventilation was classified as volume-cycled ventilation.

d 

If a patient had any non-Y-shaped stent placed, whether metal or silicone, it was considered a tube stent. See e-Appendix 1 for interpretation of ORs related to stent type.

e 

Eighty-three Y stents were silicone, and two were metal.

Practice Patterns

Significant differences were found in the type of anesthesia used among centers (P < .001). Four centers used moderate sedation predominantly (57%-100% of cases). The remaining 11 centers used moderate sedation rarely (range, 0%-15%). There were also significant differences in the use of rigid bronchoscopy (P < .001) and type of ventilation used with rigid bronchoscopy (P < .001). Four centers predominantly used flexible bronchoscopy without rigid bronchoscopy (86%-100% of cases), five centers predominantly used rigid bronchoscopy (79%-96% of cases), and six centers often used rigid bronchoscopy moderately (48%-69% of cases). When performing rigid bronchoscopy, some centers used jet ventilation, whereas others used volume-cycled ventilation exclusively (P < .001). Use of stents varied significantly among centers as well. Among the eight centers with data on ≥ 25 cases (n = 1,052), the proportion of cases in which a stent was placed ranged from 13% to 69% (P < .001).

Any Complication

Forty-four patients (3.9%) had complications (Table 2). An escalation in level of care was required in 27 (61%) of these patients. In the four centers that used moderate sedation predominantly, the complication rate was 12.2%; in the 11 centers that used predominantly general anesthesia, the complication rate was 3.0% (P < .001). Among the eight centers with data on ≥ 25 cases (n = 1,052), complication rates ranged from 0.9% to 11.7% (P = .002). On multivariate analysis, urgent and emergent procedures, American Society of Anesthesiologists (ASA) score > 3, redo therapeutic bronchoscopy, and moderate sedation were associated with increased complication rates.

Table Graphic Jump Location
TABLE 2 ]  Patient and Clinical Characteristics for Any Complications

Data are presented as mean ± SD or No. (%) unless otherwise indicated. See Table 1 legend for expansion of abbreviations.

a 

Firth penalized likelihood approach used for rare events.

b 

Wilcoxon two-sample test.

c 

Fisher exact test.

d 

Spontaneous-assist ventilation was classified as volume-cycled ventilation.

Complications That Had an Adverse Event

Six patients (0.5%) died secondary to procedural complications (e-Table 1). Four patients had a complication and died within 24 h. Two patients had a complication and died > 24 h after the procedure, but the complication was believed to be contributory. Two other patients died within 24 h but did not have complications. In both cases, death was due to underlying disease progression refractory to treatment. On multivariate analysis, only urgent and emergent procedures and never smoking were associated with procedural complications leading to death (Table 3).

Table Graphic Jump Location
TABLE 3 ]  Patient and Clinical Characteristics by Complication Resulting in Death

Data are presented as No. (%) unless otherwise indicated. See Table 1 legend for expansion of abbreviations.

a 

Firth penalized likelihood approach used for rare events.

b 

Fisher exact test.

c 

Two-sample t test.

d 

Spontaneous-assist ventilation was classified as volume-cycled ventilation.

Twenty-four patients (2.2%) had complications as well as an adverse event (Table 4). An escalation in level of care was required in 17 (71%) of these patients. On multivariate analysis, Zubrod score > 1, ASA score > 3, and redo therapeutic bronchoscopy were associated with greater risks of having a complication and adverse event.

Table Graphic Jump Location
TABLE 4 ]  Patient and Clinical Characteristics by Any Complication That Also Had an AE

Data are presented as No. (%) unless otherwise indicated. AE = adverse event. See Table 1 legend for expansion of other abbreviations.

a 

Firth penalized likelihood approach used for rare events.

b 

Wilcoxon two-sample test; two-sample t test.

c 

Fisher exact test.

d 

Spontaneous-assist ventilation was classified as volume-cycled ventilation.

Death Within 30 Days of Procedure

There were 800 first therapeutic bronchoscopy procedures (Table 5). In the 30 days following bronchoscopy, 119 patients (14.8%) died. Among the seven centers with data on ≥ 25 cases (n = 735), 30-day mortality ranged from 7.7% to 20.2% (P = .02). On multivariate analysis, Zubrod score > 1, ASA score > 3, any intrinsic or mixed obstruction, and stent placement were associated with greater risks of death within 30 days of the procedure. Y stents had a higher risk of death (OR, 4.92) than tube stents (OR, 1.72). See e-Appendix 1 for interpretation of the ORs for stents.

Table Graphic Jump Location
TABLE 5 ]  Patient and Clinical Characteristics by Death Within 30 Days Among Patients Having Their First Therapeutic Bronchoscopy

Data are presented as No. (%) unless otherwise indicated. See Table 1 legend for expansion of abbreviations.

a 

Two-sample t test.

b 

Fisher exact test.

c 

Spontaneous-assist ventilation was classified as volume-cycled ventilation.

d 

See e-Appendix 1 for interpretation of ORs for stents.

Unexpected Escalation in Level of Care

Forty-nine patients (4.4%) had an unexpected escalation in level of care required (e-Table 2). Among the eight centers with data on ≥ 25 cases (n = 1,052), the proportion of patients requiring an unexpected increase in the level of care ranged from 0.5% to 4.2% (P = .10). On multivariate analysis, ASA score > 3, gastroesophageal reflux disease, electrocautery use, and dilation were associated with a higher likelihood of requiring escalation in the level of care.

Bleeding Requiring Intervention

Six patients (0.5%) had bleeding requiring intervention (e-Table 3). On multivariate analysis, urgent and emergent procedures, APC use, redo therapeutic bronchoscopy, and never smoking were associated with greater risks of bleeding requiring intervention.

Refractory Hypoxemia

Twenty-two patients (2.0%) had refractory hypoxemia (e-Table 4). On multivariate analysis, redo therapeutic bronchoscopy, tracheoesophageal fistula, and moderate sedation were associated with greater risk of refractory hypoxemia.

Unexpected Respiratory Failure Within 24 Hours

Fifteen patients (1.3%) had unexpected respiratory failure (e-Table 5). On multivariate analysis, nonwhite race, ASA score > 3, diabetes, having a hematologic malignancy, left mainstem disease, and electrocautery use were associated with a greater risk of unexpected respiratory failure (see e-Appendix 1 for a detailed discussion).

Hospital-Level Variation and Complications

We evaluated the impact of center and center-level volume on homogeneity. We did not find evidence of a relationship between average number of cases per month and complication rates, but the analysis was limited because the model did not converge when smaller centers were included. When we used only larger centers, we failed to find evidence of heterogeneity caused by center-level variables. This is not to say that centers had the same outcomes but rather that after evaluating for other covariates, the between-center variance was negligible.

In this study, we quantified the risks associated with therapeutic bronchoscopy for malignant central airway obstruction. We found that therapeutic bronchoscopy has a complication rate of 3.9%, with 2.2% of patients experiencing an adverse event and 0.5% of patients dying of complications. We found that moderate sedation, urgent or emergent bronchoscopy, ASA score > 3, and redo bronchoscopy were associated with increased risk of complications. Not all complications resulted in adverse events, but in general, adverse events were more common in patients who had a lower functional status, higher ASA scores, and redo bronchoscopies. We found significant differences among centers in terms of practice patterns, with some centers using predominantly moderate sedation and flexible bronchoscopy and others using general anesthesia and rigid bronchoscopy. Significant differences were also found in stent usage rates. Complication rates and 30-day mortality also varied significantly among centers.

The present study is consistent with and adds to the existing body of evidence regarding therapeutic bronchoscopy complications.1,2,4,5,16,18,2630 We focused solely on patients with malignant central airway obstruction because prior work has shown that the magnitude of the effect of risk factors varies based on whether the patient has malignant or benign disease.1 Many previous studies reporting risk factors for therapeutic bronchoscopy failed to make this distinction. In addition, because of the relatively rare nature of complications and the specialized nature of the procedures involved, many prior studies lacked the statistical power necessary to identify risk factors for rare complications. The present study is among the largest multicenter cohort studies to focus on malignant airway obstruction, and because of this, it allowed us to identify risk factors for specific types of complications that have not been previously reported. We found that redo bronchoscopy and moderate sedation are associated with higher complication rates. Because moderate sedation use varied significantly among centers, this provides an opportunity for improving outcomes in the future. Of note, although the use of neuromuscular paralysis was associated with lower complication rates on univariate analysis (3% vs 6.7%, P = .006), after adjustment for the type of anesthesia, the use of paralysis had no impact on complication rates. Similarly, although spontaneous ventilation was associated with higher complication rates than jet or volume-cycled ventilation, after adjustment for the type of anesthesia, the type of ventilation did not affect complication rates. Importantly, the type of anesthesia used is a patient-level variable at some centers (ie, something decided by the bronchoscopist on a case-by-case basis) and a center-level variable at others (eg, in centers where general anesthesia is not readily available for bronchoscopy). The current results suggest that it is important for hospitals to invest the necessary resources to make general anesthesia readily available for therapeutic bronchoscopy.

We also found significant differences among centers in terms of stent use. This may be important because 30-day mortality was higher in patients who had stents placed. However, the association between stents and 30-day mortality may be the result of residual confounding or confounding by indication. Specifically, patients requiring stents may have a higher disease burden, and therefore, the observed decrease in survival might be due to these unaccounted-for factors. This might also explain why Y stents were associated with higher mortality than tube stents. Similarly, physicians may have been more likely to place stents in patients who had few other treatment options and limited life expectancy. However, prior studies have shown that infections are more common in patients undergoing airway stenting and that these infections are associated with a high case fatality rate.13,14,31 We believe that selectively stenting only patients who really require it is prudent. Although the wide variation in stent use observed among centers (range, 13%-69% of cases; P < .001) could be due to differences in the patient populations being treated, it is probable that the present findings reflect at least some degree of variation among centers in terms of stent strategy.

Stent strategy may differ depending on many subtle variations in the underlying indication for airway stenting. In some cases, a stent may be used due to significant residual luminal stenosis not amenable to ablation. In rare cases, stenting may be used to close a fistula. Finally, stents may be used for their barrier effect following successful reopening of the airway with ablative techniques, even though the lumen postablation is > 50% open. In these instances, the physician must make a judgment about whether the risks of recurrent obstruction warrant stenting. It is likely that most of the observed practice variation arises from this last category of patients. Although all patients had at least 50% stenosis prior to bronchoscopy, we did not record the luminal obstruction present at the moment of stenting, only the percentage obstruction, so this outcome analysis can only provide limited insight in this regard. This study cannot answer the question of best stent strategy given the limitations inherent in its design, but the study does highlight the need for future work in this area given the wide variations in practice patterns and the associated increase in 30-day mortality observed with stenting.

To our knowledge, this study is the first to evaluate the impact of hospital-level variables on therapeutic bronchoscopy outcomes. Prior studies of diagnostic bronchoscopy with endobronchial ultrasound-transbronchial needle aspiration found that although little difference exists among hospitals in terms of complication rates, there are significant differences in diagnostic yield, with higher-volume hospitals having higher diagnostic yields.21,22 In the present study of therapeutic bronchoscopy, the average number of cases per month did not have a significant impact on complication rates. However, the number of centers and the number of patients enrolled from low-volume centers were low relative to the frequency of the events being studied (ie, complications). In addition, the center volumes per month reflected only the number of patients with malignant central airway obstruction. Complex therapeutic bronchoscopies for other indications, such as benign disease, were not considered, but the expertise developed from those procedures would probably be transferrable to malignant cases. Therefore, these findings should be viewed as exploratory because this study lacks the power to reliably detect a relationship between center volume and incidence of complications.

Consistent with prior studies, we found that low performance status was associated with increased 30-day mortality1,13 and that high ASA scores were associated with more complications and increased health-care resource use.1,32 Of note, not all prior studies of therapeutic bronchoscopy found a relationship between ASA scores and complications, but this may have been due to sample size limitations and the frequency of complications being relatively low.12,33,34 In terms of health-care resource use, higher ASA score was associated with a higher probability of escalation in the level of care postbronchoscopy. Escalation in level of care in the present study does not necessarily reflect harm per se because it did not require a clinical complication to be present. For example, observing a patient overnight after an outpatient procedure would have been considered an escalation in level of care. Rather, escalation in level of care in this context is a measure of resource use. We found that 4.4% of patients required increased health-care resources in the short term following therapeutic bronchoscopy. This is similar to previous reports on Nd:YAG laser in which up to 21% of patients could not be immediately extubated and went to the ICU.34 This is in contrast to other studies showing that therapeutic bronchoscopy allows deescalation in the level of care.35 However, those studies focused solely on patients with airway obstruction and acute respiratory failure requiring admission to the ICU prior to bronchoscopy. In such instances, therapeutic bronchoscopy often results in an immediate clinical improvement and deescalation in level of care, albeit perhaps not in the first 24 h; however, that is a very small subset of all patients receiving therapeutic bronchoscopy. The present study provides additional information by assessing the broader population of all patients undergoing therapeutic bronchoscopy for malignant airway obstruction. The data suggest that over the first 24 h, therapeutic bronchoscopy occasionally results in an increase in level of care required, although probably over the long term, quality of life and level of care required improve, consistent with previous studies.12,35

Other limitations inherent to the study design are the possibility of residual confounding and confounding by indication. For example, the association between APC use and bleeding is probably not causal, but rather it is a reflection of the fact that when bleeding occurs, an attempt to cauterize the lesion with APC is often made. As with all observational studies, the associations observed may not necessarily be causal. Another limitation is the use of composite outcomes, such as any complication, rather than looking at specific types of complications, such as hypoxemia or bleeding. This can affect the magnitude of the ORs observed and the inferences drawn because a risk factor for one type of complication (eg, hypoxemia) may not be a risk factor for another (eg, bleeding). To address this, we analyzed the data for each major type of complication and reported them separately. However, because complications are rare, it is possible that we failed to identify significant risk factors (ie, β-error) despite the relatively large sample size.

In conclusion, this multicenter registry study of therapeutic bronchoscopy for malignant central airway obstruction is the first to evaluate patient and hospital predictors of complications and adverse events. Complications were relatively rare, occurring in 3.9% of patients, with death occurring in 0.5%. There were significant variations in how therapeutic bronchoscopy was performed. Of the risk factors identified, moderate sedation and stent use were variables most amenable to physician control. Physicians performing therapeutic bronchoscopy for malignant central airway obstruction should consider using general anesthesia rather than moderate sedation or should have general anesthesia readily available during the procedure. Stent placement was associated with increased 30-day mortality, and significant variation in stent use existed among centers. Although this could be due to residual confounding or confounding by indication, careful consideration about whether a stent is truly necessary if the airway lumen can be reopened to ≥ 50% without it is prudent. Future studies should verify these findings and explore the impact of variations in stent strategy on patient outcomes.

Author contributions: D. E. O. served as the principal investigator, had full access to all of the data in the study, and takes responsibility for the integrity of the data and the accuracy of the data analysis. D. E. O. contributed to the project oversight and organization, data collection and auditing, statistical analysis, and writing of the manuscript; A. E., H. B. G., J. D.-M., M. Slade, T. R. G., M. M., C. A. J., J. T., K. L. K., C. R., S. G., R. F. C., F. A. A., M. W., G. A. E., L. B. Y., R. C. M., S. B., and A. T. contributed to the data collection and writing of the manuscript; X. L. contributed as the primary biostatistician for the project and to the construction of the multilevel models and analyses and writing of the manuscript; and M. Simoff contributed to the registry design and organization, data collection and auditing, and writing of the manuscript writing.

Financial/nonfinancial disclosures: The authors have reported to CHEST the following conflicts of interest: Dr Eapen is a consultant to Olympus Corporation. Dr Tremblay receives royalties from CareFusion Corporation regarding intellectual property licensed through the University of Calgary and is a paid consultant to CareFusion Corporation and Olympus America Inc. Drs Ost, Ernst, Grosu, Lei, Diaz-Mendoza, Slade, Gildea, Machuzak, Jimenez, Toth, Kovitz, Ray, Greenhill, Casal, Almeida, Wahidi, Yarmus, Morice, Benzaquen, and Simoff have reported that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Role of sponsors: The data used for this publication were provided through AQuIRE. Although CHEST has reviewed and approved the proposal for this project, the researcher(s) are solely responsible for the analysis and any conclusions drawn from the data presented in this article. The sponsors had no role in the design of the study, the collection and analysis of the data, or the preparation of the manuscript.

Collaborators: Kevin L. Kovitz, MD, FCCP (principal investigator [PI]); Sara Greenhill, MD, FCCP; Thomas R. Gildea, MD, FCCP (PI); Michael Machuzak, MD, FCCP; Francisco A. Almeida, MD, FCCP; Joseph Cicenia, MD; Momen Wahidi, MD, FCCP (PI); Kamran Mahmood, MD; Paul MacEachern, MDCM (PI); Alain Tremblay, MDCM, FCCP; Michael Simoff, MD, FCCP (PI); Javier Diaz-Mendoza, MD, FCCP; Cynthia Ray, MD; David Feller-Kopman, MD (PI); Lonny B. Yarmus, DO, FCCP; Rosa Estrada-Y-Martin, MD (PI); Roberto F. Casal, MD (PI); Jennifer Toth, MD (PI); Raj Karunakara, MD (PI); Mark Slade, MD, FCCP (PI); Armin Ernst, MD, FCCP (PI); Samaan Rafeq, MD; David E. Ost, MD, MPH, FCCP (PI); George A. Eapen, MD, FCCP; Carlos A. Jimenez, MD, FCCP; Rodolfo C. Morice, MD, FCCP; Sadia Benzaquen, MD (PI); Jonathan Puchalski, MD (PI).

Other contributions: Each site PI was responsible for data collection and auditing at their institution. The authors thank the following AQuIRE staff: Joyce Bruno, MBA, MIPH; Jeff Maitland, BS; and Danielle Jungst, BS.

Additional information: The e-Appendix and e-Tables can be found in the Supplemental Materials section of the online article.

APC

argon plasma coagulation

AQuIRE

American College of Chest Physicians (CHEST) Quality Improvement Registry, Evaluation, and Education Program

ASA

American Society of Anesthesiologists

Ernst A, Simoff M, Ost D, Goldman Y, Herth FJ. Prospective risk-adjusted morbidity and mortality outcome analysis after therapeutic bronchoscopic procedures: results of a multi-institutional outcomes database. Chest. 2008;134(3):514-519. [CrossRef] [PubMed]
 
Ernst A, Feller-Kopman D, Becker HD, Mehta AC. Central airway obstruction. Am J Respir Crit Care Med. 2004;169(12):1278-1297. [CrossRef] [PubMed]
 
Casal RF, Iribarren J, Eapen G, et al. Safety and effectiveness of microdebrider bronchoscopy for the management of central airway obstruction. Respirology. 2013;18(6):1011-1015. [CrossRef] [PubMed]
 
Chung FT, Chen HC, Chou CL, et al. An outcome analysis of self-expandable metallic stents in central airway obstruction: a cohort study. J Cardiothorac Surg. 2011;6:46. [CrossRef] [PubMed]
 
Husain SA, Finch D, Ahmed M, Morgan A, Hetzel MR. Long-term follow-up of ultraflex metallic stents in benign and malignant central airway obstruction. Ann Thorac Surg. 2007;83(4):1251-1256. [CrossRef] [PubMed]
 
Saad CP, Murthy S, Krizmanich G, Mehta AC. Self-expandable metallic airway stents and flexible bronchoscopy: long-term outcomes analysis. Chest. 2003;124(5):1993-1999. [CrossRef] [PubMed]
 
Lin SM, Lin TY, Chou CL, et al. Metallic stent and flexible bronchoscopy without fluoroscopy for acute respiratory failure. Eur Respir J. 2008;31(5):1019-1023. [CrossRef] [PubMed]
 
Madden BP, Datta S, Charokopos N. Experience with Ultraflex expandable metallic stents in the management of endobronchial pathology. Ann Thorac Surg. 2002;73(3):938-944. [CrossRef] [PubMed]
 
Dasgupta A, Dolmatch BL, Abi-Saleh WJ, Mathur PN, Mehta AC. Self-expandable metallic airway stent insertion employing flexible bronchoscopy: preliminary results. Chest. 1998;114(1):106-109. [CrossRef] [PubMed]
 
Sasano S, Onuki T, Adachi T, et al. Nitinol stent for the treatment of tracheobronchial stenosis. Jpn J Thorac Cardiovasc Surg. 2001;49(5):311-316. [CrossRef] [PubMed]
 
Shin JH, Hong SJ, Song HY, et al. Placement of covered retrievable expandable metallic stents for pediatric tracheobronchial obstruction. J Vasc Interv Radiol. 2006;17(2 pt 1):309-317. [CrossRef] [PubMed]
 
Cavaliere S, Foccoli P, Farina PL. Nd:YAG laser bronchoscopy. A five-year experience with 1,396 applications in 1,000 patients. Chest. 1988;94(1):15-21. [CrossRef] [PubMed]
 
Grosu HB, Eapen GA, Morice RC, et al. Stents are associated with increased risk of respiratory infections in patients undergoing airway interventions for malignant airways disease. Chest. 2013;144(2):441-449. [CrossRef] [PubMed]
 
Ost DE, Shah AM, Lei X, et al. Respiratory infections increase the risk of granulation tissue formation following airway stenting in patients with malignant airway obstruction. Chest. 2012;141(6):1473-1481. [CrossRef] [PubMed]
 
Breitenbücher A, Chhajed PN, Brutsche MH, Mordasini C, Schilter D, Tamm M. Long-term follow-up and survival after Ultraflex stent insertion in the management of complex malignant airway stenoses. Respiration. 2008;75(4):443-449. [CrossRef] [PubMed]
 
Furukawa K, Ishida J, Yamaguchi G, et al. The role of airway stent placement in the management of tracheobronchial stenosis caused by inoperable advanced lung cancer. Surg Today. 2010;40(4):315-320. [CrossRef] [PubMed]
 
Saji H, Furukawa K, Tsutsui H, et al. Outcomes of airway stenting for advanced lung cancer with central airway obstruction. Interact Cardiovasc Thorac Surg. 2010;11(4):425-428. [CrossRef] [PubMed]
 
Amjadi K, Voduc N, Cruysberghs Y, et al. Impact of interventional bronchoscopy on quality of life in malignant airway obstruction. Respiration. 2008;76(4):421-428. [CrossRef] [PubMed]
 
Ost DE, Ernst A, Grosu HB, et al. Therapeutic bronchoscopy for malignant central airway obstruction: success rates and impact on dyspnea and quality of life. Chest. 2015;147(5):1282-1298. [CrossRef] [PubMed]
 
ACCP Quality Improvement Registry Education and Evaluation (AQuIRE). CHEST website. http://chestnet.org/guidelines-and-resources/payment-practice-and-quality/practice-resources/AQuIRE. Accessed April 30, 2013.
 
Ost DE, Ernst A, Lei X, et al; AQuIRE Bronchoscopy Registry. Diagnostic yield of endobronchial ultrasound-guided transbronchial needle aspiration: results of the AQuIRE Bronchoscopy Registry. Chest. 2011;140(6):1557-1566. [CrossRef] [PubMed]
 
Eapen GA, Shah AM, Lei X, et al; American College of Chest Physicians Quality Improvement Registry, Education, and Evaluation (AQuIRE) Participants. Complications, consequences, and practice patterns of endobronchial ultrasound-guided transbronchial needle aspiration: results of the AQuIRE registry. Chest. 2013;143(4):1044-1053. [CrossRef] [PubMed]
 
Yarmus LB, Akulian J, Lechtzin N, et al; American College of Chest Physicians Quality Improvement Registry, Education, and Evaluation (AQuIRE) Participants. Comparison of 21-gauge and 22-gauge aspiration needle in endobronchial ultrasound-guided transbronchial needle aspiration: results of the American College of Chest Physicians Quality Improvement Registry, Education, and Evaluation Registry. Chest. 2013;143(4):1036-1043. [CrossRef] [PubMed]
 
Firth D. Bias reduction of maximum likelihood estimates. Biometrika. 1993;80(1):27-38. [CrossRef]
 
King G, Zeng L. Logistic regression in rare events data. Polit Anal. 2001;9(2):137-163. [CrossRef]
 
Razi SS, Lebovics RS, Schwartz G, et al. Timely airway stenting improves survival in patients with malignant central airway obstruction. Ann Thorac Surg. 2010;90(4):1088-1093. [CrossRef] [PubMed]
 
Oviatt PL, Stather DR, Michaud G, MacEachern P, Tremblay A. Exercise capacity, lung function, and quality of life after interventional bronchoscopy. J Thorac Oncol. 2011;6(1):38-42. [CrossRef] [PubMed]
 
Bolliger CT, Sutedja TG, Strausz J, Freitag L. Therapeutic bronchoscopy with immediate effect: laser, electrocautery, argon plasma coagulation and stents. Eur Respir J. 2006;27(6):1258-1271. [CrossRef] [PubMed]
 
Bolliger CT, Breitenbuecher A, Brutsche M, Heitz M, Stanzel F. Use of studded Polyflex stents in patients with neoplastic obstructions of the central airways. Respiration. 2004;71(1):83-87. [CrossRef] [PubMed]
 
Wood DE, Liu YH, Vallieres E, Karmy-Jones R, Mulligan MS. Airway stenting for malignant and benign tracheobronchial stenosis. Ann Thorac Surg. 2003;76(1):167-172. [CrossRef] [PubMed]
 
Agrafiotis M, Siempos II, Falagas ME. Infections related to airway stenting: a systematic review. Respiration. 2009;78(1):69-74. [CrossRef] [PubMed]
 
Fernandez-Bustamante A, Ibañez V, Alfaro JJ, et al. High-frequency jet ventilation in interventional bronchoscopy: factors with predictive value on high-frequency jet ventilation complications. J Clin Anesth. 2006;18(5):349-356. [CrossRef] [PubMed]
 
Perin B, Zaric B, Jovanovic S, et al. Patient-related independent clinical risk factors for early complications following Nd: YAG laser resection of lung cancer. Ann Thorac Med. 2012;7(4):233-237. [CrossRef] [PubMed]
 
Hanowell LH, Martin WR, Savelle JE, Foppiano LE. Complications of general anesthesia for Nd:YAG laser resection of endobronchial tumors. Chest. 1991;99(1):72-76. [CrossRef] [PubMed]
 
Colt HG, Harrell JH. Therapeutic rigid bronchoscopy allows level of care changes in patients with acute respiratory failure from central airways obstruction. Chest. 1997;112(1):202-206. [CrossRef] [PubMed]
 

Figures

Tables

Table Graphic Jump Location
TABLE 1 ]  Patient and Clinical Characteristics

Data are presented as mean ± SD or No. (%). APC = argon plasma coagulation; ASA = American Society of Anesthesiologists; GERD = gastroesophageal reflux disease; HD = hemodialysis.

a 

Patients could have multiple disease locations, multiple types of obstruction, multiple ablative techniques, and multiple types of stents; therefore, these are not mutually exclusive.

b 

Patients having a second primary cancer present other than the one causing obstruction.

c 

Spontaneous-assist ventilation was classified as volume-cycled ventilation.

d 

If a patient had any non-Y-shaped stent placed, whether metal or silicone, it was considered a tube stent. See e-Appendix 1 for interpretation of ORs related to stent type.

e 

Eighty-three Y stents were silicone, and two were metal.

Table Graphic Jump Location
TABLE 2 ]  Patient and Clinical Characteristics for Any Complications