Recent advances in the treatment of air leaks

Robert James Cerfolio

Purpose of review
Alveolar-pleural fistulas (air leaks) are an extremely common
clinical problem and remain the most common complication
after elective pulmonary resection and video-assisted
procedures. The decision making process used to manage air
leaks and chest tubes that control them has been,
until very recently, based on opinions and training preferences
as opposed to facts derived from randomized clinical trials.
Recent findings
Recently, several prospective randomized trials have
studied air leaks. An objective, reproducible classification
system has also been designed and clinically validated to
help study air leaks. This system and these studies have
shown that water seal is superior to wall suction to help
stop most leaks. Even in patients with a pneumothorax and
an air leak, water seal is safe and best; however, if a patient
has a large leak (greater than an expiratory 3 on the
classification system) or experiences subcutaneous
emphysema or an expanding pneumothorax that causes
hypoxia, then some suction ( 10 cm of water) should
be applied to the chest tubes.
Summary
Air leaks were a poorly understood yet extremely common
clinical problem that had never been scientifically studied.
Over the past 5 years, prospective randomized studies
have shown that water seal is the best setting for chest
tubes and that a pneumothorax is not a contraindication
to leaving tubes on seal. Further studies are needed to
investigate the ideal management of alveolar-pleural
fistulas (air leaks) in different clinical scenarios besides
those that occur postoperatively.
Keywords
air leaks, chest tube management, water seal
Curr Opin Pulm Med 11:319323. 2005 Lippincott Williams & Wilkins.
Section of Thoracic Surgery, University of Alabama at Birmingham, and Division of
Cardio-Thoracic Surgery, Department of Surgery, Birmingham Veterans
Administration Hospital, Birmingham, Alabama, USA
Correspondence to Robert J. Cerfolio, Cardiothoracic Surgery, University of
Alabama at Birmingham, 1900 University Blvd., THT 712, Birmingham,
AL 35294, USA
Tel: 205 934 5937; fax: 205 934 6218, e-mail: Robert.cerfolio@ccc.uab.edu
Current Opinion in Pulmonary Medicine 2005, 11:319323
2005 Lippincott Williams & Wilkins.
1070-5287
Introduction
Alveolar-pleural fistulas, or air leaks, remain the most
common complication after pulmonary resection. Until re-
cently there has been no consensus on how best to mange
them. Air leaks have many different causes; they include
spontaneous, iatrogenic, traumatic, and postresectional air
leaks, among others. Clinically, most air leaks occur after
pulmonary resection. Because of the homogenicity of this
type of air leak, it has been the only type that has been
truly scientifically studied in clinical prospective trials
and thus will be the focus of this review. The principles
learned from these studies, however, are applicable to
any patient with an air leak from any cause. Until very re-
cently, few if any scientific studies were designed to eval-
uate the best treatment of this problem. Surgeons and
pulmonologists have been called upon for years to manage
chest tubes in patients with an air leak. Most decisions,
unfortunately, were based on where and by whom these
specialists were trained rather than on objective facts
gleaned from review of the most recent scientific litera-
ture. Recent studies, however, have provided important
data to help guide these decisions.
Definition
One of the biggest problems with air leaks is confusion of
the terms [1]. Very frequently, physicians pervert the
terms and call alveolar-pleural fistulas bronchopleural fis-
tulas. An alveolar-pleural fistula is a communication be-
tween the pulmonary parenchyma distal to a segmental
bronchus and the pleural space. A bronchopleural fistula
is a communication between a mainstem, lobar, or seg-
mental bronchus and the pleural space. The two terms
refer to completely different clinical problems. This dis-
tinction is not merely academic. Most air leaks after elec-
tive pulmonary resection are alveolar-pleural fistulas, not
bronchopleural fistulas. The latter cannot occur unless the
patient has had a pneumonectomy, lobectomy, or segmen-
tectomy (unless there has been an iatrogenic injury or
trauma to the airway from a suction catheter or a double-
lumen tube, or a deceleration or missile injury, all of which
are exceedingly rare). The proper classification of an air
leak as one or the other is important because the treat-
ments and natural histories of the two types of fistulas
are drastically different. Bronchopleural fistulas al-
most always require re-operation or some type of surgical
intervention. They usually require a muscle or omental
flap or occasionally can be treated, if very small, with glue
injection into the sidewall of the bronchus. These types of
fistulas have significant morbidity. By contrast, an alveolar
pleural fistula almost never, if ever, requires re-operation.
319
Time and patience are usually all that is needed. With
patience almost all of these leaks will seal; even if they
do not, the tubes can be removed within a few weeks,
and the pleural space adhesions prevent a tension
pneumothorax [2].
Incidence
Alveolar-pleural fistulas are extremely common and are
reported to occur in most large series in about 33% of
patients after elective pulmonary resection, depending
on how and when they are defined [25]. Several risk fac-
tors have been shown to increase the chance that an air
leak will occur. These include wound-healing problems
such as steroid use, emphysematous qualities such as a
low forced expiratory volume in 1 second, and large resec-
tions that leave a large pleural space deficit. These con-
cepts are all further expanded upon in the next section.
Many of these patients characteristics are also found in
those who have a spontaneous pneumothorax. The princi-
ples presented below, obtained from randomized trials on
patients who undergo pulmonary resection, are applicable
to these patients as well. Both the air leak and the pleural
space need to be controlled, but the latter need not be
overtreated, as it has been in the past.
Risk factors for air leaks
Brunelli et al. [4] performed a retrospective analysis of a
prospective database on 588 patients who underwent lo-
bectomy and identified predictors of prolonged air leaks.
In that article, the authors defined prolonged air leak as an
air leak that persisted for more than 7 days after resection.
A low predicted postoperative forced expiratory volume
in 1 second, the presence of pleural adhesions, an upper
lobectomy, and a bilobectomy were all identified as factors
that increased the risk of a prolonged leak. Bilobectomy
(the removal of two lobes on the right side) removes a large
amount of pulmonary parenchyma and leaves only one
lobe in a large pleural space, with little chance of parietal
pleural-to-visceral pleural apposition. Thus, we performed
a prospective randomized trail on these patients and showed
that the creation of a pneumoperitoneum is safe and de-
creases the incidence of air leaks in these patients [6].
This procedure can be performed on patients with pro-
longed leaks who have not undergone surgery but who
have a leak and a basilar pneumothorax. Similarly, Brunelli
et al. [7] in a separate article also showed in a prospective
trial that the creation of a pleural tent, a technique that
brings the parietal pleural of the upper chest wall down
to the remaining pulmonary parenchyma after upper lo-
bectomy, is another surgical technique that helps reduce
the duration of air leaks. This technique addresses upper
hemithorax pleural space problems. Both these techni-
ques show the importance of pleuralpleural apposition
as a critical component in the sealing of air leaks. It is
not a necessary component, however. Some patients have
a fixed pleural space deficit. This is defined as a nonre-
solving pneumothorax in a patient with a fully expanded
lung and patent chest tubes that are on suction and in the
pneumothorax. This space is best left alone and not over-
treated. If the patient does not have an air leak, the tubes
should be removed and the space will fill with fluid.
The largest series of predictors of air leaks was published
by our group. It was a retrospective review of an electronic
prospective database on 688 patients [2]. We found that
steroids, male gender, a leak with a pneumothorax, and
a lobectomy were all risks for the occurrence of a prolonged
leak. In that series we defined a prolonged leak as one that
is present on postoperative day 4.
Initial evaluation of an air leak: is it real?
If confronted with an alveolar-pleural fistula (air leak), the
clinician at the bedside must ensure that the leak is real
and is not a systemleak. All connections between the chest
tube and the drainage system should be checked. When
the leak is confirmed as coming from inside the patients
chest and not the system, it should be classified. Careful
observation at the bedside reveals that the natural history
of air leaks is based on two main features: the type of air
leak (the qualitative aspect of the system; determined by
when the air leak occurs during the respiratory cycle) and
the size of the air leak (the quantitative aspect of the clas-
sification system). We have developed and refined a classi-
fication system for alveolar-pleural fistulas (air leaks) [8,9].
It is formally referred to as the Robert David Cerfolio
(RDC) classification system for air leaks (named in honor
of my father, a practicing surgeon). Given that it helps
guide treatment, it will be briefly presented below.
Qualitative aspect of the classification
system for air leaks
There are only four types of air leaks. The largest and most
uncommon type of leak is a continuous (C) air leak. These
leaks are present throughout the entire respiratory cycle.
When the physician asks the patient to breathe in and out,
there is continuous bubbling in the air leak chamber.
These types of leaks are rare and are usually seen only
in the patient who is using a ventilator or who has a bron-
chopleural fistula. If the patient is using a ventilator, the
bubbling occurs continuously during the inspiratory and
expiratory phases of the mechanically delivered breath.
The second largest type of air leak, which is also uncom-
mon, is an inspiratory (I) air leak. These leaks are present
only during inspiration. They too are seen almost exclu-
sively in the patient who is using a ventilator and has a
sizable alveolarpleural fistula or a small bronchopleural
fistula. These leaks are most commonly seen in patients
with severe emphysema who have experienced a spontane-
ous pneumothorax from a ruptured bleb.
The third and fourth types of air leaks are much more
common than C or I leaks. The third largest leak is called
320 Diseases of the pleura
an expiratory (E) air leak. An E leak is present only during
expiration (it is also evident during forced expiration).
When the physician at the bedside asks the patient to
take deep breaths in and out, bubbling is seen in the leak
chamber (or in the air leak meter) only during expiration.
This type of leak is commonly seen after pulmonary sur-
gery, and it suggests a parenchymal air leak (alveolar-
pleural fistula). Finally, if a patient is asked to take deep
breaths in and out and no air leak is seen in the leak meter
chamber, the patient should then be asked to cough. If
a leak is present only with coughing, it is referred to as
forced expiration. This type of air leak is termed a forced
expiratory (FE) leak. FE leaks are also a very common type
of air leak. More than 98% of air leaks after elective pul-
monary surgery in patients not using ventilators are E or
FE leaks. As leaks begin to resolve or heal they usually
go from an E leak to an FE leak.
Quantitative aspect of the classification
system for air leaks
The other feature of air leaks that was critical to the de-
velopment of the classification system is the size of the air
leak. A commercially available air leak meter is contained
within the Sahara S1100a Pleur-evac Chest Drainage Sys-
tem, as shown in Figure 1. The air leak meter features a
chamber in which the leak is measured on a scale from 1 to
7 (Fig. 2). A number 1 leak is the smallest leak; a number 7
leak is the largest. Each of these chambers has a different
size, hence a different resistance. The air leak meter is
a reliable and accurate way of quantifying the size of air
leaks.
The RDC system is simple to use and teach. Air leaks are
scored according to the qualitative and quantitative crite-
ria. For example, a leak may be called an expiratory 2 air
leak or a forced expiratory 3 air leak, referred to as an
E2 or an FE3. Medical students, residents, and fellows
learn this system in 10 to 15 minutes. It facilitates com-
munication among physicians about patients air leaks
without their having to be at the patients bedsides. This
information has allowed more efficient chest tube man-
agement and contributes to fast-tracking of patients after
pulmonary resection, which enables patients to be dis-
charged routinely with a high degree of satisfaction by
postoperative day 3 or 4. Armed with this system, we are
able to review the most recent literature.
Recent literature
Over the past several years, we and others have studied
the problem of alveolar-pleural fistulas (air leaks) using
prospective randomized trials or predetermined algo-
rithms in an attempt to bring some science to what has
always been a subjective art form. The first prospective
study, which was from our group at the University of
Alabama at Birmingham, found that most air leaks oc-
curred during expiration [8]. We also reported in that first
study that pulmonary function testing results consistent
with emphysema increased the risk that an air leak would
occur after pulmonary resection. This study showed that
placing chest tubes on water seal not only was safe for air
leaks but also seemed to be superior to suction at stopping
leaks. It provided safety data for the performance of a pro-
spective randomized study.
The second study on air leaks was also from our institu-
tion, the University of Alabama at Birmingham. It was a
prospective randomized trial on 140 patients, of whom
33 had air leaks [9]. This study showed that patients
Figure 1. The commercially available Pleur-evac systemused for
the Robert David Cerfolio classification system for air leaks
Figure 2. The air leak meter contained within the Pleur-evac
system, which measures the leak on a scale from 1 to 7
Air leaks Cerfolio 321
who had their tubes placed on water seal instead of wall
suction were more likely to have their leak stop sooner.
Air leaks were also smaller when tubes were on water seal,
but water seal did not stop large expiratory leaks. The clas-
sification system for air leaks was further refined and
validated between blinded observers. The classification
system has become a critical component for the manage-
ment of tubes. It helps guide treatment. For example, if
a patient has an E5 leak, the tube is best left on suction
and not placed to water seal because an enlarging pneu-
mothorax is probable.
Marshall et al. [10] from the University of Pennsylvania
reported another prospective randomized study and found
that placing chest tubes on water seal after pulmonary re-
section shortens the duration of air leaks and decreases
the time chest tubes remain in place. By contrast, Brunelli
et al. [11

] recently published a series of selected patients,

many of whom had undergone pleural tenting. As de-
scribed above, this technique decreases the incidence and
duration of air leaks. In that series the authors did not con-
clude that water seal was a better chest tube setting than
suction; however, the advantages of water seal can be seen
in their data. Interestingly, the authors reported that the
patients using water seal had more complications than did
those who were treated with suction. This finding needs
to be further explored.
Other studies have shown that if patients have large E6 or
E7 air leaks on postoperative day 1, they will continue
to have an air leak by postoperative day 4 irrespective
of the chest tube management [2]. These patients are dis-
charged home (if otherwise ready for discharge) using a
Heimlich valve. Because of the accuracy and reliability of
the classification system, these patients can be informed
about the need for discharge with an indwelling tube early
in their hospital course. This allows the patient, the fam-
ily, nurses, and physicians to prepare both mentally and
physically for the patients discharge home with
a Heimlich valve. Moreover, this information has helped
us care for patients with spontaneous pneumothoraces.
If a patient has experienced a first spontaneous pneumo-
thorax, we usually place a chest tube only and observe the
patient. But if the air leak is large an E4 or greater we
now know that the natural history of that leak is pro-
longed. Thus, we now offer video-assisted thorascopic sur-
gery early in the hospital stay as opposed to waiting day
after day for the leak to resolve. In our most recent man-
uscript on leaks we have shown that water seal is safe for
patients with an air leak and a pneumothorax [12]. If the
leak is large (greater than an E4) or the pneumothorax is
large (greater than 8 cm on a measurement scale defined
in that paper), water seal is not safe.
Many studies have evaluated the efficacy of using pul-
monary sealants to prevent leaks [3,1315]. The only pul-
monary sealant approved by the US Food and Drug
Administration is no longer commercially available, how-
ever. Several companies are currently in phase II and III
studies trying to develop the ideal sealant.
Treatment of persistent air leaks
We define a persistent air leak as one that prolongs hospi-
talization. If on postoperative day 3 the leak is larger than
an FE3, it will not seal overnight. For that reason, the
patients chest tubes are connected to a Heimlich valve,
and the other end of the valve is connected to a urinary leg
bag or a portable drainage system. Two holes are cut in the
top of the bag. This system allows air to escape, but the
bag collects fluid, so the patients leg or clothes stay dry.
A chest radiograph is obtained after 24 hours of Heimlich
valve use, and if no new subcutaneous emphysema or no
new or enlarging pneumothorax is seen, the patient is dis-
charged home on postoperative day 4 or 5. If the chest
radiograph identifies a problem, the patient must be
returned to water seal or 10 cm H
2
O of suction, which-
ever is needed to alleviate the pneumothorax. This pro-
cess is repeated again in 2 days. If a second pneumothorax
occurs, the options are to perform a bedside chemical
pleurodesis. If a bedside pleurodesis using doxycycline is
performed, the tubing cannot be clamped. Tubing should
be hung about 6 feet off the ground. An extra length of
rubber tubing is often needed to accomplish this height.
This technique allows the sclerotic agent to stay in the
chest but allows air to escape.
Provocative chest tube clamping
When the patient is home with a Heimlich valve, a chest
radiograph is obtained every week. If the air leak resolves,
the tube can be removed. This can be determined by plac-
ing the end of the Heimlich valve in a cup of water and
asking the patient to take deep breaths and cough. If
the air leak has resolved, there will be no bubbles. If the
air leak is still present after the patient has been home
for 2 weeks with a Heimlich valve, the tubes can still
be removed safely. The safety of this method has been
demonstrated [2]. It uses a technique called provoca-
tive chest tube clamping, which was first described by
Kirschner [16]. The reason it is safe to remove a chest
tube despite the presence of a leak is probably that the
pleural space develops adhesions from the tube. These
adhesions prevent a pneumothorax from developing be-
cause the rest of the lung is stuck, even though part of
the lung is still leaking.
Conclusion
In conclusion, air leaks are a very common clinical prob-
lem. The management of tubes and drains and air leaks
can be studied with randomized trials and objective data.
A validated, objective classification system is now avail-
able and helps guide treatment. This system and random-
ized studies have shown that placing chest tubes to water
322 Diseases of the pleura
seal is superior to suction and better helps stop air leaks.
Large leaks (i.e., greater than E4), however, will probably
not respond to water seal, and patients may experience a
pneumothorax or an enlarging subcutaneous emphysema.
In these patients, some suction is best. Prolonged air leaks
are more common in patients with emphysematous lungs
and with pulmonary resections that removed large amounts
of lung. A pneumothorax is not an indication for suction.
Finally, patients can safely go home with an air leak and
with chest tubes. The tubes can be managed on an out-
patient basis and then removed, even if the patient still
has an air leak, as long as there is no subcutaneous emphy-
sema or a symptomatic pneumothorax. Further random-
ized studies are needed.
References and recommended reading
Papers of particular interest, published within the annual period of review, have
been highlighted as:

of special interest
of outstanding interest
1 Cerfolio RJ. Advances in thoracostomy tube management. Surg Clin North
Am 2002; 82:833 848.
2 Cerfolio RJ, Bass CS, Pask AH, et al. Predictors and treatment of persistent
air leaks. Ann Thorac Surg 2002; 73:1727 1731.
3 Fabian T, Federico JA, Ponn RB. Fibrin glue in pulmonary resection: a prospec-
tive, randomized, blinded study. Ann Thorac Surg 2003; 75:1587 1592.
4 Brunelli A, Monteverde M, Borri A, et al. Predictors of prolonged air leak after
pulmonary lobectomy. Ann Thorac Surg 2004; 77:1205 1210.
5 Rice TW, Kirby TTJ. Prolonged air leak. Chest Surg Clin N Am 1992; 2:803
811.
6 Cerfolio RJ, Holman WL, Katholi CR. Pneumoperitoneum after concomitant
resection of the right middle and lower lobes (bilobectomy). Ann Thorac Surg
2000; 70:942 947.
7 Brunelli A, Refai MA, Monteverde M, et al. Pleural tent after upper lobectomy:
a randomized study of efficacy and duration of effect. Ann Thorac Surg 2002;
74:1958 1962.
8 Cerfolio RJ, Tummala RP, Holman WL, et al. A prospective algorithm for the
management of air leaks after pulmonary resection. Ann Thorac Surg 1998;
66:1726 1731.
9 Cerfolio RJ, Bass CS, Katholi CR, et al. Prospective randomized trial com-
pares suction versus water seal for air leaks. Ann Thorac Surg 2001; 71:
1613 1617.
10 Marshall MB, Deeb ME, Bleier JI, et al. Suction vs water seal after pulmo-
nary resection. a randomized prospective study. Chest 2002; 121:831
835.

11 Brunelli A, Monteverde M, Borri A, et al. A. Comparison of water seal and suc-

tion after pulmonary lobectomy: a prospective, randomized trial. Ann Thorac
Surg 2004; 77:1932 1937.
In this prospective randomized trial, the authors studied patients who underwent
lobectomy and found no difference between suction and water seal, but a careful
review of their paper shows that when the patients who underwent pleural tenting
were eliminated from the analysis, there was an advantage for those with a seal,
but it did not achieve statistical significance.
12 Cerfolio RJ, Bryant AS, Singh S, et al. The management of chest tubes in pa-
tients with a pneumothorax and an air leak after pulmonary resection. Chest
2005. (in press).
13 Lang G, Csekeo A, Stamatis G, et al. Efficacy and safety of topical application
of human fibrinogen/thrombin-coated collagen patch for treatment of air leak-
age after standard lobectomy. Eur J Cardiothorac Surg 2004; 25:160
166.
14 Porte HL, Jany T, Akkad R, et al. Randomized controlled trial of a synthetic
sealant for preventing air leaks after lobectomy. Ann Thorac Surg 2001;
71:1618 1622.
15 Wain JC, Kaiser LR, Johnstone DW, et al. Trial of a novel synthetic sealant in
preventing air leaks after lung resection. Ann Thorac Surg 2001; 71:1623
1629.
16 Kirschner PA. Provocative clamping and removal of chest tubes despite air
leak [letter]. Ann Thorac Surg 1992; 53:740 741.
Air leaks Cerfolio 323