Alveolar-pleural fistula, causing spontaneous pneumothorax in a dog- surgical management

vladi prAuthor:

Dr. Vladislav Zlatinov

Central Veterinary Clinic

Chavdar Mutafov str, 25 B, Sofia, Bulgaria



Referring vet: Dr. Jordan Jordanov



Most commonly the pneumothorax in small animals is caused by trauma 1. It could be open or closed; closed traumatic pneumothorax is often the result of blunt trauma (HRS, automobile accidents, etc.). The mechanism includes a chest compressed against the closed glottis, the airway or lung parenchyma can rupture with resultant air leakage.

When there is no evidence of trauma,  air leakage from the lung parenchyma is termed spontaneous pneumothorax. This is relatively rare but potentially lethal condition in small animals. The most common underlying causes are pulmonary bulla, subpleural blebs/emphysema (68% of dogs); neoplasia (11% of dogs); migrating plant seeds; pulmonary abscesses; feline chronic allergic bronchitis; chronic pneumonia; heartworm disease. Siberian Huskies and large chested breeds are overrepresented for the condition2.

In internal pneumothorax, the trachea, bronchi, alveolar ducts could be the source of leakage. The alveolar-pleural fistula (APF) is a communication between the pulmonary parenchyma distal to a segmental bronchus and the pleural space, while a broncho- pleural fistula (BPF) is a communication between a main stem, lobar, or segmental bronchus and the pleural space3.This distinction is important because the treatment for the two types could vary. Indeed in the veterinary literature there are not too many publications, regarding incidence, therapy, etc. of APF.



Case report


Six years old, large (40kg) mix breed dog was presented to us with clinical signs of respiratory distress. The difficult breathing had started suddenly 48 hours ago. The referring veterinarian had done primary diagnostics- chest X -rays and blood work. The radiographs had revealed bilateral pneumothorax; thoracocentesis has been accomplished several times, every time evacuating more than one liter of air. The owner didn’t report any primary trauma. The day before the onset, the dog was treated with parasite prevention drug-Ivermectin 300 ug/kg , s.c.


Our clinical examination revealed anxiety, tachypnea- frequent (120/min) shallow breathing, distant breath sounds bilaterally, mild fever- 39.5. The patient was over- all hemodynamically stable- normal mucous membranes, strong femoral pulse.

Emergency care


After sedation with Butorphanol + Midazolam (i.v.), thoracocentesis was done at the right 10-th intercostal space. Almost two liters of air were removed.




Orthogonal chest radiographs (just after the centesis), revealed residual pneumothorax, left lung lobe collapse (atelectasis), right shifted cardiac silhouette (Fig.1). Pleural fissures were detectable, but no significant pleural effusion was visible; the lung pattern showed mild signs of diffuse alveolo-interstitial pattern, more pronounced in the right hilus area.unnamed


A serological Diroffilaria test (Anigen, Bionote, Seul, Korea), done by the referral veterinarian went out positive. Nevertheless, the disease was not confirmed by us- the SNAP 4Dx Test (Idexx comp.) and microfilaria blood smear tests came out negative. Larvo/ovoscopic test (done 6 days later) was negative.

Working algorithm


Working algorithm Within the next 8 hours, several thoracocentesis and air evacuations were accomplished, demonstrating the significant air leakage present. The underlying cause was not evident- a spontaneous rupture of a lung lesion (bulla) was supposed. A bronchoscopic or thoracoscopic diagnostic options were not available. A CT study could be used but needed transport to another facility, which was evaluated as too risky. Because of the lack of any signs of lung sealing within 56 hours, an open chest surgical exploration was planned and accomplished.


Anesthetic protocol


The patient was pre medicated with Acepromazine/ Butorphanol combination- i.m., low range doses. After 10 minutes of hyper oxygenation the patient was induced in anesthesia rapidly (Diazepam/ Propofol) and intubated as fast as possible. The maintenance was achieved by Isoflurane gas (2-3%) and Ketamine drop. Positive end-expiratory pressure (PEEP) of 15 cm H20 ventilation was applied to the patient, using anesthetic ventilator (Midmark Matrix). The hemodynamic parameters- saturation, pulse rate, blood pressure was closely monitored during the whole surgery; no anesthesiological incident was met. The PEEP was adjusted ( up to 25 cm H20), accordingly to the desired lung hyperinflation after chest opening.


Surgical protocol


A standard ventral median sternotomy approach was used. The sternum was osteotomised (with oscillating saw) in the mid line – from the 3-th to 7-th sternebra; later the approach was extended further caudally , leaving intact the end part of the xiphoid process. Two automatic wound retractors were placed cranially and caudally, achieving excellent approach to all chest structures.

A thorough cavity inspection was accomplished. No free fluid or obvious air leakage was found. Chest walls and heart had a normal surface and structure. The lungs had a normal surface, structure and inflation except one collapsed lobe area.  The pathology was found in the caudal part of the left cranial lung lobe-  atelectasis , with lack of parenchyma inflation. On the dorsal surface there were three small (1 mm) confined lesions areas with serosal  surface defects,filled with granulation tissue (Video 1)


A bubble test was accomplished- flooding the chest with warm sterile saline and searching for bubbles during positive-pressure ventilation (up to 40 cm H20). Surprisingly no source of air leakage, including the lesion area, was found.

A second careful  and thorough lung exploration was done, but did not reveal any other areas in question. Finding the visible pathology , we proceeded towards partial lobectomy of the distal collapsed caudal part of the cranial left lobe. One relatively big bronchus and blood vessel were encountered proximally and ligated individually. The resected area was first sutured with  continuous overlapping suture (3-0 PDS). The edge of the incision is over sewn in continuous pattern (4-0 PDS). The sutured sites were tested for leakage again. The lung tissue was separated for histology examination.

Single chest tube was placed before thoracic cavity closure. The thoracotomy was closed routinely, using several full cerclage wires, compressing tightly the osteotomised stenebrae

(Fig.2).vl 5



Postoperative care and follow up


Peri operative analgesia was provided using multimodal approach:


NSAID (Meloxicam s.c.) – pre and post op

Transdermal Fentanyl patch 100 μg (applied 4 h pre op),

MLK (Morphine, Lidocaine, Ketamin)  i.v. infusion for 10 hours post op

Local pleural infiltration- 20 ml Levobupivacaine (2,5 mg/ml) in the chest tube every 6 hours.


Continuous air suction was not applied but the chest tube was tested and air evacuated every 2 hours.


The patient showed fast anesthetic recovery with good pain control after the procedure (comfortable laying in sternal recumbence).


The air presence was monitored carefully. After 2 hours of negative finding, an abundant air accumulation started, just as the preoperative status. With the frequent air evacuation, the next 24 h the patient was respiratory stable with occasion onsets of tachypnea. The chest tube was closely examined, and found adequately airtight.


Additional diagnostics


Significant air leakage (>500 ml/h) was appreciated in the next 24h after the surgery, with no tendency of rate deceleration.  This let us look for a major broncho-pleural fistula with hidden /complex localization. We used an uncommon diagnostic technique- contrast bronchography. After induction in short light anesthesia, the dog was intubated and 3,5 ml Omnipaque (Iohexol 350 mg/ml) were injected through the endotracheal tube, with the head in elevated position. The procedure went uneventfully; the X rays, following 2 minutes of Ambu bag hyperventilation revealed normal bronchial tree, without noticeable pathology(Fig.3).vl 8

Second explorative surgery


Considering, that we didn’t adequately addressed the air leakage, and after a fair discussion with the owner we reluctantly decided to go for a second explorative surgery.


The same anesthetic protocol was used; we approached the cavity through the recent thoracotomy wound. The lobectomy site was explored, showing perfect sealing and early signs of fibrin formation on the edge. The chest tube intercostal passage looked smooth and nice without soft tissue laceration.

This time, all the lung lobes showed normal inflation and again a frustrating lack of any leakage signs. Special attention was payed on to explore the obscure hilus lung areas. Two saline/bubble test were accomplished without result. At the end (after 30 minutes exploration) an air leakage was demonstrated by an accidental left lung lobe maneuvering, while the chest still filled with saline. We found a single, small (0,5 mm) , barely visible, smooth margins (obviously epithelised) opening on the dorsal surface of the cranial left lung lobe (Fig.4 ). Based on the macroscopic findings we diagnosed alveolar-pleural fistula, having stealthy dynamic characteristics- emissive only in specific lung lobe positions, and dorsal body recumbence not facilitating it (look at the discussion section)


A single purse- string  suture (4-0 PDS) was laced around the fistula opening. After the sealing confirmed and final chest lavage the thoracotomy was closed routinely, leaving a chest tube (Fig.5 ).vl 4


Follow up and result


Immediate cease of the air accumulation was evident after the second surgery (Fig.6). Despite the fast recovery, the dog was closely monitored in clinic for 3 days, afterward the chest tube was removed and the dog discharged for home care.vl 6

The histological report that came later was suggesting not specific granulomatous lung parenchyma changes, with no causative agent (parasites/larva) present in the tested tissue.


The dog made full clinical recovery. Its condition was followed 4 years after the thoracotomy surgeries (Fig.7 ).vl 3




In the presented case report multiple discussions could be initiated. They may regard the possible underlying cause of the pneumothorax, the best diagnostic tools for APF, and the best treatment – conservative vs. surgical, sternal vs. intercostal approach.

In our case we were able to confirm the mechanism of air leakage but we couldn’t reveal the exact primary cause of the lesions. Considering the anamnesis (deteriorating after ivermectin treatment) and the morphology of the lesions- small localized, discreet , we supposed parasitic) migration that had caused alveolar-visceral pleural fistulation. The adult or larva forms of some nematodes may have caused this.

There are numerous evidences that  confirm that D. immitis could cause spontaneous pneumothorax (sometimes lethal) 4,5. There are case reports of a Angiostrongylus vasorum causing massive lung lesion and spontaneous pneumothorax6. Ascaris larvae invasion was reported as a cause of spontaneous pneumothorax in humans7.

Despite we suggest parasitic etiology, it is beyond our knowledge and the scope of the report  to evaluate the morbidity potential of the mentioned or other (Capillaria, Oslerus) parasites. Nevertheless the cause, our retrospective interpretation of the lesions showed these: the fistula canals in the caudal part of the cranial lobe (found at first surgery) had collapsed, causing no serious  air leakage, whereas the cranial fistula (found at the second surgery) was covered with epithelium, preventing it’s successful sealing and creating an air valve. It was interesting that despite its small size the APF was capable of causing serious leakage, with no tendency for healing. Indeed, the small size and the normal local serosal surface around made it difficult to distinguish during both surgeries. Another misleading factor was it’s dorsal localization. With the patient laying in dorsal recumbence, a leakage intermittently blocked by compression of the fistula against the thoracic wall was suggested by us (Fig.8-9 ). This could have been the reason for negative bubble tests- no air leakage, unless lobe manipulation and lifting it from the thoracic wall.Untitled-1

In the presented case, we didn’t had the chance to use advanced imaging diagnostics, so we relied on surgical exploration as diagnostic and vl 9therapeutic option, as recommended in many veterinary sources. We were able to manage the case successfully, but met some difficulties and a revision surgery was needed. It is interesting to suggest how useful could be the advanced imaging tools in the diagnose of APF. Increased accuracy (compared to chest radiography) in finding pathologic lung bullas/ blebs have been proven by studies8.9. One study comparing the accuracy of radiography and CT for bulla and bleb identification in dogs with spontaneous pneumothorax found: radiography to be accurate in 16% of cases and CT to be accurate in 80% of cases. In our case, such circumscribed , air filled structure was missing, so it is questionable if the small air leakage source (APF) could be precisely identified. Possibly, just suggestive signs of gross hypoventilation of the cranial left lung could have been found.

Our suggestion is that a bronchoscopy could also be inconclusive, dealing with a fistulation of small alveolar duct, not accessible for examination from the segmental bronchi. It is interesting if a thoracoscopy study could find such small APF. In a 2003 study,10,11 spontaneous pneumothorax, caused by bullas was successfully diagnosed and treated in three dogs using thoracoscopy,

An argument for a more consistent conservative approach- constant air evacuation system for longer period (>4 days), may be raised. Knowledge  about potential causes of spontaneous pneumotorax is important factor when considering surgical versus nonsurgical management. We based our clinical decision on several reports that show the advantages of the more aggressive surgical  management in canine patients . In one study, recurrence rates and mortality rates for dogs with spontaneous pneumothorax treated surgically (3% and 12%, respectively) were lower than for dogs treated conservatively (50% and 53%, respectively) 1,12.

The feline patients look more prone to conservative management (first choice treatment) of spontaneous pneumothorax, because of prevailing inflammatory ethnologies 13.

It is interesting to mention that in human patients there are publications14,15 for successfully treatment of APF by use of blood patch, synthetic hydrogel and valves, delivered endobronchialy. Watanabe spigots (IBV® valves -Olympus Corp., Japan) are specifically designed for reducing air leaks by means of total occlusion of the affected bronchus. For now, no such interventional option has been reported in the veterinary sources.

The surgical approach that we used -a median sternotomy, may be considered as more aggressive and painful than the alternative- intercostal thoracotomy. The median sternotomy is our preferred choice in such cases because it gives a vast access to all lung lobes and other thoracic structures. There are evidences that in dogs, pain, the degree of cardiopulmonary impairment, and complication rates with between two approaches do not differ 16,17, 18. In humans, median sternotomy causes less postoperative discomfort than intercostal thoracotomy19.

We recognize  that a good pain control is a critical factor , when dealing with thoracotomy patients. Post operatively pain may prevent full thoracic wall excursion, reducing the ventilation and causing hypoxemia20. Pain results in catecholamine release, which contributes to vasoconstriction, decreased tissue perfusion and arrhythmias. A multimodal medical approach was used in the case providing haemodynamic stability of the patient within two open chest surgeries.



We would like to stress out the importance of following the medical algorithms in managing such spontaneous pneumothorax cases. In rare cases, the “culprit” / the underlying cause, could not be easily identified even if we know it is there. Still a persistence and systematic approach could be rewarded even in perplexed, frustrating  scenarios




  1. Puerto DA, Brockman DJ, Lindquist C, et al: Surgical and nonsurgical management of and selected risk factors for spontaneous pneumothorax in dogs: 64 cases (1986– 1999). J Am Vet Med Assoc 220:1670, 2002.