1. CHEST Original Research
INTERVENTIONAL PULMONOLOGY
Treatment of Persistent Pulmonary Air
Leaks Using Endobronchial Valves
John M. Travaline, MD, FCCP; Robert J. McKenna, Jr, MD, FCCP;
Tiziano De Giacomo, MD; Federico Venuta, MD, FCCP;
Steven R. Hazelrigg, MD, FCCP; Mark Boomer, MD;
and Gerard J. Criner, MD, FCCP; for the Endobronchial Valve for Persistent Air
Leak Group*
Background: Prolonged pulmonary air leaks are a significant source of frustration for patients
and physicians. When conventional therapy fails, an alternative to prolonged chest tube drainage
or surgery is needed. Bronchoscopic blockage of a bronchus can be performed with the hope of
accelerating closure of the air leak by reducing the flow of air through the leak. To our
knowledge, this article presents the largest series of patients with prolonged air leaks treated with
an endobronchial valve.
Methods: With Internal Review Board approval, endobronchial valves were compassionately
placed using flexible bronchoscopy in patients with prolonged air leaks at 17 international sites.
Results: Between December 2002 and January 2007, 40 patients (15 women; mean age ؎ SD,
60 ؎ 14 years) were treated with one to nine endobronchial valves per patient. The air leaks had
recurrent spontaneous pneumothorax (n ‫ ,)12 ؍‬postoperative (n ‫ ,)7 ؍‬iatrogenic (n ‫ ,)6 ؍‬first-
time spontaneous pneumothorax (n ‫ ,)4 ؍‬bronchoscopic lung volume reduction (n ‫ ,)1 ؍‬and
trauma (n ‫ )1 ؍‬etiologies. Nineteen patients (47.5%) had a complete resolution of the air leak, 18
(45%) had a reduction, 2 had no change, and 1 had no reported outcome. The mean time from
valve insertion to chest tube removal was 21 days (median, 7.5 days; interquartile range [IQR], 3
to 29 days) and from valve procedure to hospital discharge was 19 ؎ 28 days (median, 11 days;
IQR, 4 to 27 days).
Conclusions: Use of endobronchial valves is an effective, nonsurgical, minimally invasive
intervention for patients with prolonged pulmonary air leaks. (CHEST 2009; 136:355–360)
Abbreviation: IQR ϭ interquartile range
P may cause considerableleaks are common and
rolonged pulmonary air
morbidity, prolonged
eases: bullous emphysema, advanced pulmonary sar-
coidosis, radiation fibrosis, and interstitial lung dis-
hospital stay, and increased health-care costs. When ease. The treatment of patients with air leaks due to
an air leak is present on the fourth postoperative day, underlying pulmonary disease is often very challeng-
the chance of air leak on postoperative day 7 is 83%.1 ing due to the poor ability of the diseased lung to
In addition to occurring after 15% of thoracic oper- heal.
ations,2 other causes of prolonged air leaks may Because the poor performance status caused by
include the following underlying pulmonary dis- pulmonary disease may increase the risk for surgical
intervention, alternative therapies are needed. Treat-
Manuscript received October 6, 2008; revision accepted
February 10, 2009.
Affiliations: From the Temple University School of Medicine © 2009 American College of Chest Physicians. Reproduction
(Drs. Travaline and Criner), Philadelphia, PA; Cedars-Sinai of this article is prohibited without written permission from the
Medical Center (Dr. McKenna), Los Angeles, CA; the University American College of Chest Physicians (www.chestjournal.org/site/
of Rome (Drs. De Giacomo and Venuta), Rome, Italy; Southern misc/reprints.xhtml).
Illinois University School of Medicine (Dr. Hazelrigg), Spring- Correspondence to: John M. Travaline, MD, FCCP, Professor
field, IL; and Saint Francis Hospital (Dr. Boomer), Tulsa, OK. of Medicine, Temple Lung Center, 3401 North Broad St,
*Members and affiliations for the Endobronchial Valve for Philadelphia, PA 19140; e-mail: trav@temple.edu
Persistent Air Leak Group are listed in the Appendix. DOI: 10.1378/chest.08-2389
www.chestjournal.org CHEST / 136 / 2 / AUGUST, 2009 355
Downloaded From: http://journal.publications.chestnet.org/ on 10/23/2012

2. ment of prolonged air leaks by sclerosing agents3 has
shown little efficacy and variable patient tolerance.
Various endobronchial approaches include the appli-
cation of gelfoam, the use of fibrin glue in conjunc-
tion with endovascular metallic ring-shaped coil
placement,4 tracheobronchial stent deployment,5
and endobronchial valves.6 –13 The results of these
nonsurgical approaches appear promising and, for
some patients, may be the only treatment option
after all conventional treatments have failed or are
associated with high risk. To our knowledge, this
article reports the largest series of patients with
prolonged pulmonary air leaks, mostly related to
secondary pneumothoraces, treated with an endo-
bronchial valve (Zephyr EBV; Emphasys Medical;
Redwood City, CA) in order to assess the results of
this bronchoscopic approach for the treatment of
prolonged air leak.
Materials and Methods
In the United States, patients were treated under compassionate-
use regulatory provisions in which individual Food and Drug
Administration and Internal Review Board approvals were ob-
tained for each patient. Outside the United States, the device was
commercially available, with an approved indication for treat-
ment of air leak. Individual case studies for six patients have been
previously reported7–12 in the medical literature.
The endobronchial valve system used in this series (Fig 1)
consists of a silicone-based, one-way valve mounted in a self-
expanding nitinol retainer. The self-expanding retainer stabilizes
the device in the airway and provides an airtight seal against the
bronchial wall. The one-way valve is positioned in the center of
Figure 1. Transcopic endobronchial valve (Zephyr EBV;
the device to allow it to function independently of the anchoring
Emphasys Medical, Redwood City, CA).
portion of the valve. The valve is intended to block air from
flowing through the air leak while allowing distal secretions to
drain normally. The device is designed to function as a perma-
nent implant but can, at physician discretion, be removed with
graspers subsequent to resolution of the air leak. Figure 2 shows Following valve implantation, patients were allowed to recover
a transcopic valve in the right middle lobe bronchus. from anesthesia according to standard hospital practice. Vital
Anesthesia and access approaches varied across centers, de- signs and arterial blood gas levels were closely monitored. Chest
pending on patient status and physician preference. The proce- radiography often was used to assess target lung inflation status.
dure was performed with IV sedation, spontaneous breathing, Between December 2002 and January 2007, 15 centers used the
and flexible bronchoscopy. Alternative approaches involved gen- endobronchial valves to treat patients with prolonged pulmonary
eral anesthesia, rigid bronchoscopy, and mechanical ventilation. air leaks.
Prior to valve implantation, the chest drainage system was The following data were retrospectively abstracted from pa-
observed to assess the air leak. A balloon-tipped catheter was tients’ medical records: demographic information, the cause of
inserted into the lobar airway suspected of supplying the air leak. the air leak, associated diseases, and whether hospitalization was
To provide selective bronchial occlusion, the balloon was inflated medical or surgical. In addition, several end points were chosen
to block the airflow to that region of the lung. The air leak rate to assess endobronchial valve efficacy in treating the underlying
through the chest tube was then assessed qualitatively for bronchopleural fistula, including the lung involved, the number
reduction. If the leak rate was reduced, the balloon catheter was and duration of chest tubes used, the need for other interventions
deflated and repositioned into a more distal airway. The process to treat the air leak, and length of hospital stay. These variables
was repeated to identify the segmental or subsegmental airway or were analyzed using simple descriptive statistics.
airways that, when occluded, offered the greatest reduction in air
leak rate. These airways then were targeted for valve placement.
The endobronchial valves were delivered to the target airway
Results
using a flexible catheter. The valves were compressed into the
distal tip of the delivery catheter using a valve loader supplied Fifteen women and 25 men (mean age Ϯ SD,
with the system. The delivery catheter was then passed through
the working channel of a standard adult bronchoscope (Ն 2.8 mm 60 Ϯ 14 years) had at least one endobronchial valve
inner diameter) and guided to the target airway. Once in place, (range, one to nine valves) placed for the manage-
the valve was deployed. ment of prolonged pulmonary air leak. Primary
356 Original Research
Downloaded From: http://journal.publications.chestnet.org/ on 10/23/2012

3. Table 1—Etiology of Air Leak Grouped by
Classification (n ‫)04 ؍‬
Classification No.
Continuous
Secondary spontaneous 14
Postsurgical 3
Iatrogenesis 3
Primary spontaneous 3
Postbronchoscopic lung volume reduction 1
Trauma 1
Expiratory
Secondary spontaneous 6
Postsurgical 4
Iatrogenesis 3
Primary spontaneous 1
Unidentified
Secondary spontaneous 1
(Fig 3A) but was reduced following the placement of
four endobronchial valves (Fig 3B).
Valve procedures were performed in one session
in 37 patients, two sessions in 2 patients, and four
sessions in 1 patient, for a total of 116 valves inserted.
The classic-type valve was used in 5 patients, and
transcopic valves were used in 35 patients. The mean
number of valves placed per patient was 2.9 Ϯ 1.9
overall, 3.19 Ϯ 2.2 for secondary pneumothorax,
Figure 2. Bronchoscopic view of valve following deployment in
right middle lobe bronchus. 2.83 Ϯ 2.3 for iatrogenesis, and 2.75 Ϯ 1.5 for pri-
mary and 2.28 Ϯ 1.1 for postsurgical air leak. The
valves were placed in the right upper lobe (n ϭ 11),
right middle lobe (n ϭ 3), right lower lobe (n ϭ 3),
comorbidities in these patients were cancer (30%); left upper lobe (n ϭ 11), left lower lobe (n ϭ 5),
COPD (30%); pneumonia (7.5%); and other (22.5%), right upper and right lower (n ϭ 1), right lower and
which were rheumatoid arthritis (n ϭ 2), tubercu- middle lobe (n ϭ 1), and right upper and left upper
losis (n ϭ 1), trauma (n ϭ 1), aspergilloma lobe (n ϭ 1). In four patients, the target lobe was not
(n ϭ 1), bronchiectasis (n ϭ 1), cor pulmonale reported. The valve procedures were performed
(n ϭ 1), lung transplantation (n ϭ 1), and multiple using conscious sedation in 16 patients, general
comorbidities (n ϭ 1). anesthesia in 16 patients, and deep sedation in 8
Table 1 shows the etiology and classification of the patients.
air leaks prior to the endobronchial valve procedure. Following valve placement, the air leaks resolved
Prior to the procedure, 39 patients had at least one or decreased in 37 patients (92.5%); 19 patients
chest tube (range, one to five tubes per patient). One (47.5%) had complete resolution of acute air leak, 18
patient had an Eloesser flap. The mean duration of patients (45.0%) had reduction, and 2 patients
air leak prior to valve treatment was 119 days (5.0%) had no change in air leak status. For one
(median, 20 days; interquartile range [IQR], 15 to 45 patient (2.5%), the immediate change in air leak was
days). Thirty-five patients were treated with an en- not reported. No relationship was found between the
dobronchial valve only. Prior to placement of an resolution or reduction in air leak and the location of
endobronchial valve, five patients had other treat- the valve or the etiology of the air leak (data not
ments, which included blood patch (n ϭ 3), wedge shown). Overall, the mean time from valve place-
resection (n ϭ 1), and pleurodesis (n ϭ 1). ment to chest tube removal for the 28 patients with
Figure 3 shows representative chest radiographs of complete information was 21 days (median, 7.5 days;
a 66-year-old woman with a history of COPD and IQR, 3 to 29 days). The time from insertion of the
coronary artery disease who had a loculated right valve to hospital discharge (35 patients were dis-
pneumothorax and bronchopleural fistula following charged alive; 1 patient had missing information) was
video-assisted thoracoscopy and talc pleurodesis. A 19 Ϯ 28 days (median, 11 days; IQR, 4 to 27 days).
bronchopleural fistula remained despite chest tubes Thirty-four patients had no adverse event related to
www.chestjournal.org CHEST / 136 / 2 / AUGUST, 2009 357
Downloaded From: http://journal.publications.chestnet.org/ on 10/23/2012

4. Figure 3. A: Representative chest radiograph showing patient with right-loculated basilar pneumo-
thorax and extensive soft-tissue air (lower arrow and upper arrow, respectively). B: One week following
an endobronchial valve placement procedure (four valves) [valves encircled; arrows noting absence of
abnormalities noted in A], showing lessened pneumothorax and soft-tissue air.
the valve placement; adverse events for the remain- Prolonged pulmonary air leak leads to consider-
ing six patients were valve expectoration, moderate able morbidity for patients, and complications have
oxygen desaturation; initial malpositioning of the prompted the need for a nonsurgical, minimally
valve that required redeployment, pneumonia, me- invasive approach to treat patients with this condi-
thicillin-resistant Staphylococcus aureus colonization tion. One such approach is an endobronchial valve
following a second valve procedure, and an unspec- system (Zephyr EBV), a device initially developed as
ified event. a nonsurgical method of lung volume reduction for
Of the 40 patients, 8 had the valves removed after emphysema. This bronchoscopically deployed, one-
cessation of the air leak. In two of the eight patients, the way device blocks airflow into targeted areas of the
valves were removed over two sessions. Overall, the lung that are the origin of prolonged air leaks. The
valves were in place for an average of 66 Ϯ 53 days valves are placed in the segmental or subsegmental
(range, 7 to 143 days) prior to removal. In 32 patients,
airways proximal to the area of the air leak. The air is
the clinician elected to leave the valves in place.
prevented from entering the pleural space, thereby
Follow-up ranged from 5 to 1,109 days after valve
enabling the lung to possibly reexpand and heal.
implantation. At last follow-up, 24 patients were
reported to be alive. The causes of death for the The management of persistent air leaks remains a
remaining 16 patients were underlying disease challenge for pulmonologists and thoracic surgeons.
(n ϭ 8), cancer (n ϭ 4), bronchiectasis (n ϭ 1), em- Whether they arise from primary or secondary pneu-
physema (n ϭ 1), respiratory failure (n ϭ 1), and mothoraces or complicate lung resection surgery,
sepsis (n ϭ 1). No deaths were attributed to the they are associated with increased morbidity and
valve or implantation of the valve. frustrate patients who want to be free of chest tubes
and not undergo additional surgical procedures. In
many patients, surgical options for the treatment of
Discussion air leaks do not exist, and effective alternative inter-
ventions are necessary.
We show that the implantation of endobronchial Predictors of persistent air leaks in patients in-
valves is effective for a large number of patients with clude poor wound-healing characteristics, such as
prolonged air leaks following primary and secondary preoperative use of steroids, a low FEV1 percentage,
causes of pneumothorax. Complete cessation of air and low maximum voluntary ventilation percent-
leak following endobronchial valve implantation was age.14 Intraoperative techniques for surgeons to de-
achieved in 48% of patients. An additional 45% crease the likelihood of persistent air leak following
experienced diminution in the magnitude of air leak. lung resection include bovine pericardium buttress-
Combined, 93% of patients had improvement in air ing of staple lines,15 use of pleural tents for upper
leak status following valve treatment. The valve lobectomies,16 use of pneumoperitoneum after lower
implantation procedure and the valve itself also were lobectomies,17 and the use of fibrin glue.18 When
well tolerated, with few adverse events. persistent air leaks develop despite these measures,
358 Original Research
Downloaded From: http://journal.publications.chestnet.org/ on 10/23/2012

5. they are associated with prolonged hospitalization Dr. Boomer is a consultant for Pneum Rx. The remaining authors
and increased cost. Patients with prolonged air leak have reported to the ACCP that no significant conflicts of interest
exist with any companies/organizations whose products or ser-
also have an increased risk for postoperative morbid- vices may be discussed in this article.
ities, such as empyema, fever, and pneumonia.19
Several investigators7–13 have reported the use of
endobronchial valves in the management of persis- Appendix
tent air leaks, and in all the cases, the patient’s
underlying condition made the choice of a nonsur- Members of the Endobronchial Valve for Persistent Air Leak
Group (alphabetically arranged by their institutional affiliations)
gical intervention most appropriate. For example,
are as follows: The Alfred Hospital and Monash University,
Toma and colleagues13 reported the use of endo- Melbourne, Australia: Gregory I. Snell and Trevor J. Williams;
bronchial valves in two complex medical conditions. Cedars-Sinai Medical Center, Los Angeles, CA: Robert J.
One woman with lymphangioleiomyomatosis shortly McKenna, Jr.; Cleveland Clinic, Cleveland, OH: Atul C. Mehta,
following single-lung transplantation required a Thomas R. Gildea, and Michael S. Machuzak; Henry Ford
Hospital, Detroit, MI: Michael J. Simoff; Medical University of
modified Monaldi decompressing procedure on her
South Carolina, Charleston, SC: Charlie Strange; New York
native lung complicated by a persistent air leak. The Presbyterian Medical Center, New York, NY: Roger A. Maxfield;
other patient had bilateral pneumothoraces and a Remington Davis Clinical Research Group, Columbus, OH:
persistent air leak following a complicated course of Edward M. Cordasco, Jr.; Saint Francis Hospital, Tulsa, OK:
severe pneumonia and ARDS. Similarly, the patients Mark Boomer; Southern Illinois University School of Medicine,
Springfield, IL: Steven R. Hazelrigg; Temple University, Phila-
reported here all had significant comorbidities, lim-
delphia, PA: John M. Travaline and Gerard J. Criner; University
iting the safe application of surgical intervention and of California Davis Medical Center, Sacramento, CA: Andrew
necessitating the use of less-invasive therapy. Chan; University of Iowa, Iowa City, IA: J. Scott Ferguson;
The major limitation of this report is its retrospec- University of Kentucky, Lexington, KY: Rolando Berger; Univer-
tive, nonrandomized, and uncontrolled methodol- sity of Rome “La Sapienza,” Rome, Italy: Tiziano De Giacomo
and Federico Venuta; and University of Tennessee Medical
ogy. This case series, however, contributes additional
Center, Knoxville, TN: Thomas E. Gaines.
information to support the application of this non-
surgical approach to the management of persistent
pulmonary air leaks. Further, it serves as a basis to References
study endobronchial valves in a controlled, random-
1 Cerfolio RJ, Tummala RP, Holman WL, et al. A prospective
ized, and prospective fashion. algorithm for the management of air leaks after pulmonary
In summary, this report on a large series of resection. Ann Thorac Surg 1998; 66:1726 –1730
patients confirms what has been previously demon- 2 Abolhoda A, Liu D, Brooks A, et al. Prolonged air leak
strated in case reports about the use of an endobron- following radical upper lobectomy: an analysis of incidence
chial valve for the treatment of prolonged pulmonary and possible risk factors. Chest 1998; 113:1507–1510
3 West D, Togo A, Kirk AJB. Are bronchoscopic approaches to
air leak. Further, based on the efficacy of this valve post-pneumonectomy bronchopleural fistula an effective al-
in the treatment of prolonged air leak, it extends the ternative to repeat thoracotomy? Interact Cardiovasc Thorac
indication for the use of the valve in patients with Surg 2007; 6:547–550
this condition. It is a reasonable, nonsurgical, mini- 4 Sivrikoz CM, Kaya T, Tulay CM, et al. Effective approach for
mally invasive intervention that may be appropriate the treatment of bronchopleural fistula: application of endo-
vascular metallic ring-shaped coil in combination with fibrin
for the treatment of prolonged air leaks in patients glue. Ann Thorac Surg 2007; 83:2199 –2201
who are either operable or inoperable. Prospective 5 Singh SS, Pyragius MD, Shah PJ, et al. Management of a
trials are needed, however, to further characterize large bronchopleural fistula using a tracheobronchial stent.
the clinical circumstances in which this intervention Heart Lung Circ 2007; 16:57–59
may be most appropriate. 6 Fann JI, Berry GJ, Burdon TA. The use of endobronchial
valve device to eliminate air leak. Respir Med 2006; 100:
1402–1406
7 Mitchell KM, Boley TM, Hazelrigg SR. Endobronchial valves
Acknowledgments for treatment of bronchopleural fistula. Ann Thorac Surg
2006; 81:1129 –1131
Author contributions: Dr. Travaline contributed to the concept 8 Ferguson JS, Sprenger K, Van Natta T. Closure of a bron-
and organization for the manuscript, data collection and organi- chopleural fistula using bronchoscopic placement of an en-
zation, statistical analysis, and the primary preparation and dobronchial valve designed for the treatment of emphysema.
writing of the manuscript. Dr. Criner contributed to the concept Chest 2006; 129:479 – 481
and organization of the manuscript. Dr. McKenna contributed to 9 Feller-Kopman D, Bechara R, Garland R, et al. Use of a
the primary preparation and writing of the manuscript. All removable endobronchial valve for the treatment of broncho-
authors contributed to data collection and additional review, pleural fistula. Chest 2006; 130:273–275
writing, and editing of the manuscript. 10 De Giacomo T, Venuta F, Diso D, et al. Successful treatment
Financial/nonfinancial disclosures: Dr. Travaline was a coin- with one-way endobronchial valve of large air-leakage com-
vestigator for the VENT Study-Endobronchial Valve for Emphy- plicating narrow-bore enteral feeding tube malposition. Eur
sema Palliation Trial Emphasys Medical, Inc, from 2003 to 2008. J Cardiothorac Surg 2006; 30:811– 812
www.chestjournal.org CHEST / 136 / 2 / AUGUST, 2009 359
Downloaded From: http://journal.publications.chestnet.org/ on 10/23/2012

6. 11 Snell GI, Holsworth L, Fowler S, et al. Occlusion of a broncho- 16 Robinson LA, Preksto D. Pleural tenting during upper lobec-
cutaneous fistula with endobronchial one-way valves. Ann Tho- tomy decreases chest tube time and total hospitalization days.
rac Surg 2005; 80:1930 –1932 J Thorac Cardiovasc Surg 1998; 115:319 –326
12 Anile M, Venuta F, De Giacomo T, et al. Treatment of 17 Toker A, Dilege S, Tanju S, et al. Perioperative pneumo-
persistent air leakage with endobronchial one-way valves. peritoneum after lobectomy-bilobectomy operations for
J Thorac Cardiovasc Surg 2006; 132:711–712 lung cancer: a prospective study. Thorac Cardiovasc Surg
13 Toma TP, Kon OM, Oldfield W, et al. Reduction of persistent air 2003: 51;93–96
leak with endoscopic valve implants. Thorax 2007; 62:830–833 18 Fabian T, Federico JA, Ponn RB. Fibrin glue in pulmonary
14 Cerfolio RJ, Bass CS, Pask AH, et al. Predictors and treatment resection: a prospective, randomized, blinded study. Ann
of persistent air leaks. Ann Thorac Surg 2002; 73:1727–1731 Thorac Surg 2003; 75:1587–1592
15 Hazelrigg SR, Boley TM, Naunheim KS, et al. Effect of 19 Brunelli A, Monteverde M, Borri A, et al. Predictors of
bovine pericardial strips on air leak after stapled pulmonary prolonged air leak after pulmonary lobectomy. Ann Thorac
resection. Ann Thorac Surg 1997; 63:1573–1575 Surg 2004; 77:1205–1210
360 Original Research
Downloaded From: http://journal.publications.chestnet.org/ on 10/23/2012