Key words air leak - persistent bronchopleural fistula - interventional bronchus occlusion -
secondary spontaneous pneumothorax - pneumothorax - Amplatzer devices
Schlüsselwörter Luftleckage - Persistierende bronchopleurale Fistel - interventionelle Bronchusokklusion
- sekundärer Spontanpneumothorax - Pneumothorax - Amplatzer Okkluder
Introduction
Persistent air leak (PAL), defined as an air leak that persists for greater than
5–7 days [1 ], is a rare but potentially
life-threatening complication in children with alveolar-pleural fistulas (APF) or
bronchopleural fistulas (BPF) following traumatic-, iatrogenic-, primary or
secondary spontaneous pneumothorax (SSP). SSP is defined as accumulation of air in
the pleural cavity in patients with pre-existing lung disease [2 ]. Although data on the overall incidence of PAL
in children are not available, necrotizing pneumonia (NP) is likely the most common
cause [3 ]
[4 ].
PAL has been reported in up to 33% of children hospitalized for NP in a
recently published retrospective single center study [5 ]. In addition to the direct impact on respiration and gas exchange, PAL
is associated with secondary complications like immobilization, pain, impaired mucus
clearance and increased risk for pulmonary or pleural infections [6 ]. Furthermore, PAL prevents the maintenance of
positive end-expiratory pressure during mechanical ventilation. Therefore, PALs are
associated with high morbidity, extended duration of chest tube therapy and
prolonged hospital stay, and higher resource utilization [7 ]. Unfortunately, no data on mortality rates in
children exists, but mortality from PAL ranges from 16% -72% in
adult patients [8 ]
[9 ]. Chest tube placement is the current standard of care [10 ]
[11 ] but is
not always sufficient to achieve adequate expansion of the affected lung and
sustained contact of the visceral pleura to the parietal pleura, which is a
precondition for the healing of the defect [12 ].
Furthermore, continuous suction through chest tubes, often applied in patients with
PALs to evacuate the pleural space, may prevent spontaneous healing due to
continuous flow of air through the lesion. Chemical or autologous blood pleurodesis
represent a viable treatment option in patients with PAL [13 ]
[14 ]
[15 ] but can cause significant complications
including tension pneumothorax secondary to obstruction of the chest tube and
empyema [16 ]. Furthermore, pleurodesis can be
ineffective, especially in the absence of continuous contact of the lung surface to
the thoracic cavity. Therefore, current guidelines on the management of spontaneous
pneumothorax recommend surgical treatment in case of PAL when medical treatment
fails [17 ]
[18 ].
Surgical approaches include video-assisted thoracoscopic surgery (VATS) or open
thoracotomy with either mechanical or chemical pleurodesis or pleurectomy, wedge
resection, direct stump closure with intercostal muscle reinforcement,
segmentectomy, lobectomy or even pneumectomy [19 ].
While surgical treatment is highly effective and safe in most cases, it can be
associated with complications such as bleeding, pain, wound infections and loss of
potentially recoverable lung tissue [6 ]
[20 ]. Many promising case reports or case series on
interventional bronchus occlusion in adult patients with PAL have been published
using various devices, primarily one-way endobronchial valves (EBV) [6 ]. To date, reports on successful EBV-placement in
children with PAL are limited to a total number of 6 patients [12 ]
[21 ]
[22 ]. Another viable treatment option for PAL is
interventional bronchial occlusion using Amplatzer devices (ADs). ADs are
self-expanding devices made from a Nitinol wire mesh, designed to provide
transcatheter closure of blood vessels or congenital heart defects. ADs come in
various configurations and sizes, and their deployment characteristics make them an
attractive option. They can be fully deployed, retrieved, and repositioned prior to
detachment which is a critical advantage for precise positioning and occlusion of
the fistula. Although there are several case reports and case series on successful
treatment of PAL with ADs in adult patients [20 ]
[23 ]
[24 ]
[25 ]
[26 ]
[27 ], no pediatric cases have as yet
been published. Here, we describe our first experiences regarding the feasibility,
efficacy and safety of IBO using ADs in children with secondary spontaneous
pneumothorax and PAL.
Methods
Study Design
We performed a retrospective analysis of feasibility, efficacy and safety of IBO
using ADs in patients below 18 years of age with PAL after SSP between January
2016 and December 2020.
Data collection
Demographic data as well as information on clinical symptoms, disease courses,
pre-existing diseases, radiological findings, number of chest tubes, chest tube
duration, respiratory support, surgical interventions and medical treatment
before and after intervention, length and number of interventions, number and
types of devices used, intraprocedural vital signs and radiation dose were
collected from patient’s medical records.
Patient selection
Children with PAL were carefully evaluated by a multidisciplinary team consisting
of pediatric pulmonologists, pediatric surgeons, pediatric intensivists,
pediatric anesthesiologists and pediatric interventional cardiologists. Patients
were considered candidates for IBO after exhaustion of all medical treatment
options and if a surgical procedure was considered to pose a high risk of
serious complications and/or a treatment failure. Parents provided
informed consent following detailed counselling regarding the off-label use of
ADs, possible complications, and alternative treatment options.
Anesthetic management
Standard monitoring consisting of electrocardiography, noninvasive BP measurement
and pulse oximetry was established prior to anesthesia. Intravenous access was
obtained if not already established. Anesthesia was started with intravenous
injection of 5 mg/kg Propofol and maintained by a total
intravenous anesthesia (TIVA) with 8-15 mg/kg/h of
Propofol and 0,3-0,5 μg/kg/min of Remifentanil.
Volatile anesthetics were avoided in order to prevent environmental
contamination of the staff. Depth of sedation was monitored by EEG. Except for
patient 1 who has already been intubated in the intensive care unit, airway
management consisted of a single-use inflatable laryngeal mask as a supraglottic
airway device. The larger intraluminal space provided by the laryngeal mask
offers an advantage over conventional intubation, as this provides more space
for a large bronchoscope and occluder equipment and allows better ventilation
with lower airway pressure. End-tidal CO2 was consistently monitored. A
conductive warming air blanket was used to preserve normothermia during the
procedure. Continuous esophageal temperature measurement was employed. After
intervention, most patients underwent rapid weaning and arrived in the post
anesthesia care unit (PACU) awake and breathing spontaneously.
Interventional bronchus occlusion (IBO)
Procedures were performed by an experienced multidisciplinary team including
pediatric pulmonologists, pediatric cardiologists and pediatric
anesthesiologists in the pediatric cardiac catheterization laboratory, allowing
the use of fluoroscopy when needed. Prior to intervention, chest tubes were
connected to a digital thorax drainage system (Thopaz, MEDELA ®) which
provides quantitative real time data on the air leak. As APFs or BPFs were
localized peripherally in all patients and could therefore not be visualized
directly by the bronchoscope, a balloon catheter (Arndt Airway Catheter, Cook
Medical®) was used to sequentially occlude the bronchial segments to
locate the airway leading to the fistula [28 ]
[29 ]. When bronchoscopes with a
working channel of at least 2,0 mm diameter were used the balloon
catheter was advanced directly through the working channel. In smaller children,
we employed a hybrid technique (passing the guiding catheter outside the working
channel into the bronchi under simultaneous bronchoscopic control). In cases
with multiple fistulas, more than one balloon catheter was used and different
bronchi were blocked simultaneously until the fistula flow stopped. On some
occasions, fistula flow diminished or ceased completely after induction of
anesthesia, making indirect identification of the fistula difficult or even
impossible. In this case, bronchography was applied for direct visualization of
the air leak. After the bronchi leading to the fistula were identified, a
guidewire was passed under fluoroscopic and bronchoscopic control beneath the
bronchoscope into the target bronchus. A guiding catheter or sheath for the
delivery of of the AD was passed over the guidewire which was subsequently
passed over the guidewire. The device was then advanced through the sheath or
guiding catheter and deployed into the bronchus under bronchoscopic and
fluoroscopic surveillance. Device size was calculated using previously obtained
computer tomography images as well as direct measurement during bronchoscopy and
bronchography using devices at least 1-2 mm larger than the bronchial
diameter. Chest tubes were left in place for at least 24 hours after the
procedure and removed when the 24-hour graph of the
Thopaz® -System indicated no fistula flow and the affected
lung was completely expanded on chest-x-ray. In some patients with multiple
fistulas and diminished but persistent air leaks, the procedure was stopped due
to prolonged anesthesia duration. In some cases, residual air leaks then
resolved spontaneously in the days following the intervention. If not, patients
underwent a second procedure. [Fig. 1 ]
illustrates the step-by-step procedure of interventional bronchus occlusion
using patient 3 as an example. In all patients with NP, occluders were removed
bronchoscopically using flexible forceps when closure of the fistula was
suspected based on clinical and radiological findings.
Fig.1 Illustration of sequential IBO in patient
3 a : Block of the lower lobe bronchus using a 7 Fr. Arndt
(Cook Medical®) endobronchial blocker under simultaneous
measurement of fistula flow using a Thopaz MEDELA®. b :
After significant decrease of the fistula flow, the first occluder
(*) is placed under bronchoscopic guidance. c : A second
occluder (**) was placed into segment B6 after total
cessation of fistula flow following occlusion of the segment by an Arndt
blocker as described in A. d : Second intervention after
recurrence of a small air leak. A third occluder
(***) was placed into a subsegment of B6 which
was not completely blocked by the second occluder. e :
Fluoroscopic view directly after insertion of occluder 3
(***). Bronchoscope (a.) and sheath (b.) are
still in place. f : Magnification of implanted devices after
removal of bronchoscope and sheath.
Devices
In all cases Amplatzer devices (ADs) were used. In all but one case,
Amplatzer® Vascular Plugs II (Abbott, Chicago, Illinois,
USA) in different sizes were used. In one patient Amplatzer Duct Occluder II AS
(Abbott, Chicago, Illinois, USA) were additionally applied. Bronchoscopies were
performed with Olympus Europa (SE & Co. KG, Hamburg, Germany) pediatric
flexible fiberoptic bronchoscopes (3C160, Q180, XP160F).
Statistical analysis
All statistical analyses were performed using Microsoft Excel. Descriptive data
were calculated as mean with interquartile range (IQR). Categorical data were
reported as frequencies and percentages.
Ethical approval
Study approval by the institutional ethical review board was waived given its
retrospective observational design. All representatives of pediatric patients
had previously provided written informed consent regarding anonymized use of
personal clinical data for research purposes.
Results
Between January 2016 and December 2020, seven children with PAL were referred to our
institution. One previously healthy, 2-year-old boy with right sided PAL related to
NP was already treated with two chest tubes for 14 days and need of oxygen underwent
IBO of the right upper lobe bronchus using a 4 mm Amplatzer Vascular Plug II
without any complications. Fistula flow decreased significantly after intervention,
allowing prompt removal of one chest tube and cessation of supplemental oxygen
therapy. As PAL persisted (with significantly decreased fistula flow) a second
intervention was planned. One day before the second intervention and 10 days after
the initial IBO, the patient died suddenly and unexpectedly. Post-mortem imaging
showed the occluder in the correct position and an association with the NP or the
intervention and the sudden death could be definitely excluded by subsequent
autopsy. The cause of death remains unknown. For this reason, this patient was
excluded from further analysis.
Mean age at first intervention of the remaining six patients (four males) was 9.8
years (IQR 5.5-14.6). PAL was secondary to NP in four cases. In the other two
patients, PAL occurred as a SSP as a non-infectious complication of preexisting
diffuse lung disease. Three children with NP were previously healthy and in one
patient with cystic fibrosis (CF) NP occurred as a complication of the underlying
disease. All patients were treated with chest tubes for a mean time of 28.2 days
(IQR: 21.0-36.8). Before admission, three patients underwent surgery in the
reffering hospitals (attempt of surgical coverage of a BPF in patient 1; surgical
debridement and pleurectomy complicated by bronchial injury in patient 4;
thoracoscopic bullectomy and apical pleurectomy in patient 5) without success. All
patients required oxygen therapy, one patient required additional noninvasive
ventilator support (patient 5), and one patient required venovenous extracorporeal
membrane oxygenation (ECMO) support because ventilation alone was not sufficient due
to an enormous fistula flow through the PAL (patient 1). Detailed Information on
demographic data, causes of PALs and concomitant chronic diseases are given in [Table 1 ].
Table 1 Demographic data and patients characteristics prior to
interventional bronchial occlusion
Patient no.
1
2
3
4
5
6
Sex
f
f
m
m
m
m
Age (yrs.)
5.4
15.2
5.6
4.4
12.9
15.5
Cause of PAL
NP
NP
NP
NP
SSP
SSP
Preexisting chronic lung disease
/
CF
/
/
Pulmonary GVHD after HSCT
CF
Affected side
right
left
right
right
right
right
Number of chest tubes insertion before first
intervention
5
4
5
5
3
2
Number of days with at least one chest-tube before first
intervention
30
43
21
39
21
15
Respiratory support at time of intervention
VV-ECMO
oxygen only
oxygen only
oxygen only
High-Flow
oxygen only
Surgical treatment attempts before intervention
yes
no
no
yes
yes
no
Abbreviations; yrs.=years; NP=Necrotizing Pneumonia;
SSP=Secondary Spontaneous Pneumothorax; CF=Cystic Fibrosis;
GVHD=Graft Versus Host Reaction; HSCT=Haematopoietic Stem
Cell Transplantation; VV-ECMO=VenoVenous Extracorporeal Membrane
Oxygenation; VATS=Video-assisted thoracoscopic surgery
Overall, 11 procedures with placement of 14 ADs were performed in 6 patients without
any complications. Specifically, repositioning or removal of ADs was not required,
and device migration, iatrogenic fistula formation, and infection was not observed.
Three patients required a second intervention and, of these, two patients underwent
a third procedure. All patients improved significantly after the first intervention
with an observed drop or cessation of fistula flow and a reduced or absent need for
supplemental oxygen. The mean time to removal of all chest tubes was 11.3 days
(IQR:6.5-13.0) after first intervention and the patients were discharged between 6
and 28 days after first intervention in a good clinical condition and without
supplemental oxygen (except patient 5 who was already oxygen dependent before PAL
occurred). Patient 1 represents an especially challenging case as the child had been
on a ventilator, ECMO and full sedated for 30 days prior to the first intervention.
This patient required a total of four ADs, placed sequentially during three
consecutive interventions. Therefore, this patient had a much longer chest tube
duration and hospital stay compared to the others. Impressively, while areas of lung
parenchyma initially seemed totally destroyed, the patient recovered completely and
chest-x-ray was normal on 4 years post intervention ([Fig. 2 ]). In patient 5, a persistent air leak despite previous placement
of 3 occluders in the right upper lobe bronchus was successfully sealed by
autologous blood and fibrin glue during the third procedure. Information on number
of procedures and devices per patient, radiation time, chest tube duration, time
until discharge after first intervention and time until removal of occluders (when
performed) are given in [Table 2 ]. ADs were
removed in all patients with PAL related to NP (n=4) after a mean of 70.0
days (IQR:46.5-94) post first intervention without any complications and without PAL
reoccurrence. At a mean follow up time of 3.3 years (IQR: 3-3.3), three children
were completely asymptomatic without any respiratory complaints and the one patient
with CF was in a stable condition with respect to the underlying disease. In patient
5, occluders were left in place until successful bilateral sequential lung
transplantation at our institution 7 month after IBO. In patient 6, both ADs are in
place at last follow up. It is noteworthy that in this patient with CF and chronic
colonization with mycobacterium abscessus, who is thus at increased risk for
pulmonary infection, no infection related to bronchus occlusion has occurred despite
the long duration of occluders remaining in-situ (5.6 years post intervention).
Furthermore, because of the absence of PAL recurrence after intervention, previous
plans to pursue lung transplantation for this patient were postponed. After the
initiation of triple therapy with Tezacaftor, Elexacaftor and Ivacaftor, lung
function has improved significantly and the patient reported high quality of life
with minimal limitations during everyday activities on last follow up.
Fig.2 Sequential Chest imaging before and after IBO and after removal
of occluders in patient 1 a : Chest x-ray showing a right
sided tension pneumothorax although a chest tube was inserted (1). b :
Chest x-ray showing a persistent pneumothorax despite placement of another
chest tube (1). Additionally there are bilateral diffuse ground glace
opacities, consistent with ARDS. 2: ECMO-cannula coming from the right
femoral vein; 3: ECMO-cannula coming from the right jugular vein; 4:
tracheal tube; 5: central venous line. c : Chest CT scan from the
right lung showing a large seropneumothorax (1), a consolidated lower lobe
with pneumathoceles (2) and a persistent pneumothorax despite 2 chest tubes
(3+4) d : Chest x-ray after placement of multiple ADs
(1–4) showing full expansion of the lung. The patient is still on
ECMO, ventilated via a tracheal cannula and there is still one chest tube
left (5). e : Last chest-x-ray before discharge. The ADs are still in
place, the last chest tube was removed and the patient was decannulated.
f : Last chest-x-ray on follow up 4 years post intervention. ADs
were removed. Both lungs appear unaffected without consolidations,
infiltrates or any other residual defects.
Table 2 Characteristics of interventional bronchial occlusion
and clinical course post intervention
Patient
1
2
3
4
5
6
Affected side
right
left
right
right
right
right
Number and Localization of ADs 1st intervention
2 – B6
1 – B10
2 – LLB+B6
1 - MLB
2 – B1+B2
2 – B2
Number and Localization of ADs 2nd intervention
1 – B9
/
1 – B6
/
1 – B3
/
Number and Localization of ADs 2nd intervention
1 – B10
/
/
/
/*
/
Total radiation time as dose area product cGycm2
701
2113
2348
1145
n. a.
215
Days until removement of the last chest tube after first
intervention
25
8
14
5
10
6
Days of hospitalization until discharge after first
intervention
28
10
9
7
21
6
Days until occluder extraction after first
intervention
97
93
48
42
/
/
Abbreviations; ADs=Amplatzer Devices; B1-10=Segment 1-10;
LLB=Lower Lobe Bronchus; MLB=Middle Lobe Bronchus;
n. a.=not applicable; *Additional insertion of
autologous blood and fibrin glue between AD and bronchus wall in B3
Discussion
We describe a novel interventional treatment approach for children with PAL
complicating SSP. We demonstrated that IBO using ADs is safe even in critical ill
patients.
A major advantage of the described procedure is its minimal-invasive character of
this parenchyma preserving approach. In all six cases reported here, surgery could
be avoided in favor of the IBO procedure. Even in children with chronic pulmonary
lung disease and increased risk of pulmonary infections, we observed no infectious
complications on follow up. Furthermore, no other adverse events such as bronchial
injuries, irritations or occluder migration occurred. In most cases, ADs were
eventually removed, leading to full functional recovery of the previously affected
lung. Especially in children with SSP related to NP it was surprising to observe the
enormous healing capacity of damaged lung tissue. Lung parenchymathat seemed
destroyed on initial chest imaging in these patients, displayed near total recovery
on long term follow up ([Fig. 2 ]). This
observation underlines the efficacy of this intervention and stresses its viability
as an alternative treatment option to surgery. We show that IBO using ADs can be
performed independent of patients’ size or age as instruments do not have to
pass through the working channel of the bronchoscope. Furthermore, in contrast to
surgery, children do not experience pain after intervention and chest tubes can be
removed after a relatively short time leading to early patient mobilization,
avoidance of prolonged pain therapy and sedation and thus to a faster recovery with
a corresponding shortening of hospital stay. Some of the children described here
suffered from severe underlying progressive diffuse lung disease and had already
undergone unsuccessful surgical procedures for their PALs. As result of these
endobronchial interventions, lung transplantation could be postponed or rendered
unnecessary in these patients. Nevertheless, multiple interventions were necessary
in a relevant proportion of our patients.
It is relevant to note that ADs are not licensed for endobronchial occlusion. The
only approved bronchial occluder for children is the endobronchial Watanabe-Spigot.
Watanabe-Spigots are conical, non-self-expanding silicone occluders, available in
different sizes and developed to treat pulmonary hemorrhage and PAL. In the original
trial, which included no pediatric patients, only 23 of 60 adult patients
experienced complete cessation of air leak after placement of the spigot [29 ]. Furthermore, Watanabe-Spigots are very
challenging to handle, especially in children, rendering the risk of dislocation
very high. In addition, due to the lack of self-expansion, they often fail to
adequately seal the bronchi, often resulting in treatment failure. We previously had
disappointing experiences with the use of spigots in children with severe pulmonary
hemorrhage. Case reports in adults have noted successful closure of fistulas with
various adhesives [6 ]. In our experience, however,
gluing a bronchus alone is not sufficient. Moreover, adhesives cannot be accurately
dosed and applied through the working channel of the bronchoscope, which increases
the risk of accidental occlusion of unaffected bronchi. Nevertheless, additional
autologous blood or fibrin glue might still be a treatment option when complete
sealing by ADs fails, as shown in patient 5.
An attractive and promising alternative to ADs are endobronchial valves (EBV). EBV
are one-way valves that are deployed in the lobar, segmental, or subsegmental
airways with a flexible bronchoscope. They are engineered to block air from flowing
through the fistula while allowing distal secretions to drain normally. The two
largest published studies supporting the use of EBV for PAL are retrospective
analyses in adults reporting success rates of over 90% [30 ]
[31 ]. One case
series including four children and two single case reports have demonstrated that
IBO using EBVs is also applicable even in critically ill children [12 ]
[21 ]
[22 ]. An obstacle to the use of EBVs in younger
children is the large diameter of the application system, which has to be passed
through the working channel of the bronchoscope and requires a minimum diameter of
2,1 mm. Therefore, EBVs can only be used in older children and
adolescents.
Our study is limited by its retrospective design, the small number of patients and
the absence of a control group. However, facing the individual character and
severness of SSPs in different age groups, propective studies will potentially not
be achievable. Thus cases should be systematically documented and collected, for
example within the framework of registries, to ensure high quality retrospective
analyses are possible in future.
Conclusion
We describe a technique of IBO using ADs in children as a non-invasive, safe and
valuable treatment option for PAL. Advantages include reduced pain, high efficacy,
a
low rate of complications and the ability to preserve lung parenchyma. Given their
expertise in positioning and deploying these devices, collaboration between
pediatric pulmonologists and interventional cardiologists is recommended.
Contributor's Statement
N.Schwerk, C.Happel, H.Bertram, developed the procedure. N.Schwerk, K.Schütz,
C.Happel, H.Bertram designed the research. K. Schütz, N. Schwerk, O. Keil,
A. Zychlinsky Scharff and C. Happel designed and primarily wrote the manuscript and
compiled tables and figures. All authors made substantial contributions to the
conception, drafting and revising of the manuscript, and final approval of the
submitted version.
