Keywords
experimental - lung - air leak - sealant
Introduction
Alveolar air leaks (AAL) are common sequelae of lung surgery, particularly in pleural
decortication, dissection of firm pleural adhesion, and division of incomplete fissures.
They are associated with increased postoperative morbidities, delayed removal of chest
tubes, and prolonged hospital stay.[1]
[2]
[3] In the surgical practice, the conventional closure techniques including suturing
and stapling might not always be feasible, especially in video-assisted thoracic surgery
(VATS) procedures. In the last two decades, various surgical sealants have been developed
and widely practiced in lung surgery as adjuncts to conventional repairs.[4] The recent survey reveals that many surgeons choose surgical sealants based on individual
experience and preference.[4] This is believed to, at least partly, be attributed to the limited data comparing
the sealing efficacy of different sealants.
The sealing efficacy indicated by producers is usually determined by means of the
modified industrial standard test method for burst pressure strength of surgical sealants
(e.g., American Society for Testing and Maerials F 2392–04), which does not conform
to the surgical reality. Databases on animal experiments are lacking, perhaps due
to the ethical concerns.[3] The prospective, randomized clinical trials assessing sealants in treating AAL demonstrated
very inconsistent results.[1]
[2] More importantly, the primary end points usually used in these trials (viz., the
length of hospital stay and the duration of chest tube) are dependent on various factors
besides significant AAL, limiting their reliability.
Recently, our group has developed an in vitro lung model assessing the sealing efficacy
of various sealants.[5] Two widely used sealants—a human thrombin-fibrinogen sponge (TachoSil; Takeda Pharmaceutical
Company Limited, Osaka, Japan) and an albumin-glutaraldehyde glue (BioGlue; CryoLife
Europa Ltd., Surrey, United Kingdom)—were tested by means of this model. The results
are hereafter reported.
Materials and Methods
Experimental Protocol
Lungs were freshly excised from German Landrace pigs (around 80 kg) without preference
in gender, which were euthanized in a local slaughterhouse. Within 2 hours after harvest,
the lungs were dissected along the trachea until the tracheal bifurcation was reached.
The lower lobe was selectively intubated, followed by inflation through manual ventilation.
The fully bloated lower lobe was subsequently immersed in warm water to ensure its
impermeability. After being connected to the ventilation machine (Evita, Dräger, Lübeck,
Germany), the lower lobe was ventilated in a volume-controlled mode with a positive
end-expiratory pressure of 5 cm H2O, I:E ratio of 1:2 and a ventilation frequency of 12/min. The lower lobe was fully
inflated when inspired tidal volume (TVi) ≥ 400 mL. Over-inflation was observed, if
TVi ≥ 800 mL. To create a standardized superficial parenchymal lesion, a rectangle
measuring 40 × 25 mm was first marked with a marker pen on the fully inflated lung
lobe. Using a small, conic headed drill, the defect was created by carefully applying
pressure to the marked site, starting at the edges and advancing toward the middle
of the designated area ([Fig. 1]). Surgical knots were then tied on the cranial and caudal edge of the lesion. Starting
ventilation at TVi = 300 mL, TVi was increased by 100 mL in steps until a maximal
inspiratory pressure (Pmax) of 40 cm H2O was reached. Following each increase in TVi, the expiratory tidal volume (TVe),
resistance, compliance, as well as Pmax, mean inspiratory pressure (Pmean), and plateau
inspiratory pressure (Pplat) were recorded after five cycles. AAL was calculated as
the difference between TVi and TVe.
Fig. 1 A superficial lung defect was created in a previously marked area of 40 × 25 mm on
the inflated lower lobe.
Thereafter, TachoSil and BioGlue were applied in a randomized order (n = 20 in each group) according to the usage guide, respecting a safety margin of 1
cm to all sides. TachoSil sponge (9.5 × 4.8 cm) was shortly humidified in isotone
saline and subsequently placed on the pleural defect of inflated lung lobe. Using
a moist bandage, mild pressure was applied for a period of 5 minutes to the sealant
sponge to ensure full adhesion according to the usage guide of the producer. As to
BioGlue, pleural defect was dried before the sealant application. In a slow and steady
manner, 2 mL liquid sealant was then applied on the pleural defect through the primed
applicator tip, resulting in an even coating. Before the tests, a full 2 minutes was
given to allow BioGlue to achieve the full strength according to the usage guide.
The lower lobe was then ventilated again with TVi rising slowly from 100 mL. Commencing
at 400 mL TVi, the same parameters as before sealant application were measured after
every increase in TVi by 100 mL. The distance between the two sutures was measured
to evaluate the elasticity of sealants.
Air leak at the site of pleural defect was assessed in submersion tests and graded
following the Macchiarini scale as grade 0 (no leak), grade 1 (countable bubbles),
grade 2 (stream of bubbles), and grade 3 (coalesced bubbles).[6] Sealing was considered, if no bubble was visible after five cycles of ventilation
(grade 0). Sealing failure was determined once air bubbles were observed (grade 1
or higher). If sealing failure occurred, Pmax of the last ventilation cycle was determined
as the maximally tolerated pressure (MTP). The sealing failure was furthermore categorized
into adhesive or cohesive failure. Adhesive failure was considered, if the failure
occurred at the interface between sealant and parenchymal defect. Cohesive failure
was defined as the failure within the sealant. Application failure was considered,
when cohesive or adhesive failure of sealant was observed before starting the test
at 400 mL TVi. The results were adjusted by two investigators independently. Any disagreement
would be arbitrated by a third investigator.
Finally, the lung specimens containing the attached sealant were resected and fixed
with 10% formalin. The specimens were further embedded in paraffin, and processed
to obtain sections for hematoxylin-eosin staining. The tests were performed under
room temperature and moderate air humidity. The experimental procedures have been
established in the previous study.[5]
Statistical Analysis
The normality of variables was tested by the Kolmogorov–Smirnov one-sample test. Descriptive
statistics are presented as mean ± standard deviation in case of normal distribution.
Categorical variables are expressed as percentages. Continuous data were compared
using Student t-test. Multiple linear regression analysis was used to determine the ventilation parameters
associated with AAL. Statistical significance was assumed if p < 0.05. All statistical evaluation was performed using SPSS (version 16.0 for Windows;
SPSS, Inc., Chicago, Illinois, United States).
Results
A total of 40 consecutive tests were performed in the present experiment. In the assessment
before sealant applications, AAL increased linearly with ascending TVi. Among the
recorded ventilation parameters, Pmax was the only parameter predictive of AAL in
the multiple linear regression analysis (p < 0.001).
Sealant application failure occurred once in each group. There was no disagreement
on the adjustment of air tightness in the submersion tests between the investigators.
At TVi = 400, 500, 600, and 700 mL, BioGlue achieved sealing in 19 (100%), 19 (100%),
16 (84.2%), and 14 (73.7%) tests, while TachoSil sealed in 19 (100%), 14 (73.7%),
4 (21.1%), and no (0%) tests, respectively (p = NS, p = 0.04, p < 0.001, p < 0.001). [Fig. 2] demonstrates the sealing proportion in the two groups. The MTP of BioGlue was 40.3 ± 3.0
cm H2O, significantly higher than that of TachoSil (36.0 ± 4.9 cm H2O, p = 0.003). Cohesive and adhesive failures were found in 10 and 1 tests in the BioGlue
group respectively, while all burst failures of TachoSil were adhesive. Concerning
elasticity, TachoSil allowed more expansion of the covered lung defect than BioGlue
(6.3 ± 3.9 vs. 1.4 ± 1.0 mm, p < 0.001, [Fig. 3]). Hematoxylin-eosin staining of the sealed lung specimen showed the TachoSil sponge
and the BioGlue foam attaching densely to the underlying lung surface ([Fig. 4]).
Fig. 2 Sealing proportion of TachoSil and BioGlue with ascending TVi.
Fig. 3 Expansion of the lung defects covered by TachoSil and BioGlue.
Fig. 4 Photomicrograph of histological sections showed the sealants layer attaching densely
the underlying lung tissue (A: TachoSil, B: BioGlue).
Discussion
In treating AAL, surgical sealants are an important element in the armamentarium of
thoracic surgeons. However, the multitude of clinical trials assessing the efficacy
of various sealants demonstrated inconsistent results and presented often limited
validity due to small sample size and inherent flaws of study design.[1]
[5] In addition, there are very few comparative clinical trials in this regard. As more
and more sealants with different properties are available, questions have arisen concerning
the purported overall clinical benefits. In this respect, it is pivotal to compare
the sealing efficacy of different sealants and find the most effective one.
In our study, both TachoSil and BioGlue presented strong sealing efficacy in treating
AAL. The MTP of these two sealants are close to the upper limit of the commonly applied
ventilation pressure (40 cm H2O). When compared with TachoSil, BioGlue demonstrated significantly higher MTP, thus
the superior sealing efficacy. Moreover, adhesive burst failure was found only in
1 out of 19 tests with BioGlue, while all burst failures of TachoSil sponges were
adhesive. BioGlue consists of the bovine serum albumin (BSA) as the active chemistry
and the glutaraldehyde as the connector molecule. The aldehyde groups of glutaraldehyde
react with the amine groups of BSA and those present in the extracellular matrix and
cell surface, resulting in strong cross-linking after 2 to 3 minutes.[7] In contrast, TachoSil is a thin collagen patch sponge coated with dried human fibrinogen
and human thrombin.[8] Upon application, fibrinogen is activated by thrombin and converted into fibrin
monomers, which subsequently form a highly concentrated and stable fibrin clot.[9] According to our results, the adhesive strength of fibrin clot, which holds the
TachoSil sponge tightly to the wound surface, appears lower than that of the cross-links
in BioGlue. This finding is consistent with the results of the previously published
in vitro experiments. Pedersen et al created pleural defect with a diameter of 10 mm
in harvested porcine lungs and assessed the sealing efficacy of diverse surgical sealants
including TachoSil and BioGlue in submersion tests. The recorded burst pressure of
TachoSil (35 cm H2O) was lower than that of BioGlue (55 cm H2O).[10] Of interest is also the in vitro experiment of Carbon et al, in which a piece of
porcine lung tissue containing a standardized defect with a diameter of 10 mm was
clamped in the flange over the top of a pressure chamber.[11] Pressure was applied in the chamber till the sealant material was lifted off or
destroyed. In this lung model, TachoSil achieved a burst pressure of 51.2 cm H2O (50.2 hPa).
Another interesting finding of our study is the marked elasticity of TachoSil. This
result is consistent with other clinical and experimental studies.[12] In a multicenter, prospective, and randomized study of 189 patients undergoing lobectomy,
Lang et al tested the efficacy and safety of TachoComb (Takeda Pharmaceutical Company)
as the processor of TachoSil for air leak treatment. They reported that the elastic
properties of TachoComb was especially suitable for sealing of dynamically expanding
systems like lung tissue.[8] In addition, Izbicki et al used TachoComb for treating diffused parenchymal bleeding
and/or air leaks in 52 patients and parenchymal lesions in 12 procines.[12] They found that “the high elasticity of the fleece does not disturb lung extension
movement.” In contrast, BioGlue presented a very rigid nature. Good elasticity allows
surgical sealants to adapt the movement of the lung surface after application. This
biomechanical property is particularly important, if the lung is trapped due to chronic
inflammatory process of the lung and pleura. In decortication surgery for re-expansion
of the residual lung parenchyma, removal of the peel from the underlying parenchyma
results almost inevitably in superficial tears and AAL.[13] A rigid sealant like BioGlue might prevent the re-expansion and worsen the stretchability
of trapped lung. This disadvantage would be more apparent, if the long-standing resorption
time of BioGlue (2 years) is taken into account.[14] Comparatively, TachoSil will be reabsorbed 4 to 5 months after application. In this
patient subpopulation, TachoSil with its high sealing efficacy and marked elasticity
appears to be the sealant of choice. In view of the higher MTP, BioGlue might be considered
in favor over TachoSil for treating AAL in patients with underlying lung disease,
which requires high inspiratory pressure. For instance, in patients with cystic fibrosis,
surgical pleurodesis as an effective treatment for large and recurrent pneumothorax
presents a relative contraindication for lung transplantation later.[15] Successful air leak sealing by topical application of BioGlue has been reported
in cystic fibrosis patient.[16]
In our in vitro model, porcine lungs were used due to the similarity in design and
biomechanical properties to human lungs.[11] The experimental procedures were constructed to simulate the real surgical and clinical
setting of lung surgery. The MTP of TachoSil determined in our tests is comparable
to that in the previously published animal experiment, suggesting the reliability
of the present model.[3] Recently, Pedersen et al compared sealing efficacy of six different surgical sealants
including TachoSil and BioGlue by means of another in vitro lung model.[10] They fixed harvested porcine lungs in a Plexiglas chamber filled with isotonic saline.
After sealant application on deflated lungs, the lungs were ventilated with incremental
peak airway pressure and air leaks were assessed by submersion tests. Their results
demonstrated a median burst pressure of 35 cm H2O in TachoSil group. Compared with our model, some issues may have complicated their
study to the disadvantage of TachoSil. First, TachoSil was applied to deflated lungs,
whereas the manufacturer recommends that lung should be inflated during application.
The subsequent lung re-inflation and stretch may have impaired the bonding between
fibrin sponge and lung surface. It is also worth noting that the time from lung harvest
to experimentation averaged as long as 24 hours. The resulted degradation of proteins,
especially the factor XIII being essential for the cross-linking of fibrin strands,
may also have weakened the sealing efficacy of TachoSil.[9]
As VATS has been widely adopted and practiced in lung surgery, air leak sealing by
means of topical sealant application through trocars has become a feasible approach.[16] On the contrary, prolonged air leaks are still one of the major complications after
VATS major lung resections and result in prolonged hospital stay.[17] This illustrates the need for further improvement of air leak management during
VATS procedures. Surgical sealants permitting a precise application to target sites
thorugh trocars provide further advantage in these aspects. In the surgical practice,
TachoSil sponge can be rolled and delivered through trocar easily. As liquid sealant,
BioGlue can also be applicated thoracoscopically using a delivery tip extention.
Despite the strong sealing efficacy of TachoSil and BioGlue demonstrated in our in
vitro tests, caution should be taken for potential side effects of these materials.
While concern has been arised about the potential risk for transmission of blood-borne
diseases by TachoSil and BioGlue derived from human or bovine blood plasma,[2]
[3]
[4]
[14] there have been report on the cytotoxic effects of BioGlue in in vitro and in vivo
examinations.[18] Moreover, animal experiment has demonstrated that BioGlue reinforcement of aortic
anastomoses impairs vascular growth.[19] In addition, Klimo et al found a strong association between the use of BioGlue and
postoperative wound complications in their pediatric neurosurgical practice due to
triggered acute pyogenic and chronic granulomatous inflammatory response.[20]
As one of the limitations of the present experiment, a certain variation in the size
of ventilated procine lower lobes could not be totally avoided. To minimize this confounding
feature, the lungs were harvested from the pigs in almost the same weight (around
80 kg). We did not observe marked difference in this regard during the tests. The
lower lobes were fully inflated when TVi was 400 mL or higher in all tests. In addition,
the randomization of the applied sealants might also contribute in reducing this bias.
The authors recognize that the sealant applications were not blinded for the assessment
of air tightness and the measurement of elasticity in the present experiment. It may
have resulted in information bias, which was certainly minimized by randomization
of the sealants. Finally, the observation bias might have arisen due to the inevitable
subjectiveness in the judgment of air bubbles, even though it was performed by two
investigators independently. Nevertheless, the statistic analysis revealed significant
results in various aspects. The present experiment could make a positive contribution
for the evidence-based sealant use in lung surgery.
In conclusion, our in vitro tests demonstrated the high sealing efficacy of TachoSil
and BioGlue in treating AAL. While BioGlue was superior in resisting higher ventilation
pressure, TachoSil possessed better elasticity. It is recommended to perform adequately
powered and well-designed prospective clinical trials before a systematic sealant
use.
