Thoracobiliary fistulas: literature review and a case report of fistula closure with omentum majus

Background Thoracobiliary fistulas are pathological communications between the biliary tract and the bronchial tree (bronchobiliary fistulas) or the biliary tract and the pleural space (pleurobiliary fistulas). Review of the literature We have reviewed aetiology, pathogenesis, predilection formation points, the clinical picture, diagnostic possibilities, and therapeutic options for thoracobiliary fistulas. Case report A patient with an iatrogenic bronchobiliary fistula which developed after radiofrequency ablation of a colorectal carcinoma metastasis of the liver is present. We also describe the closure of the bronchobiliary fistula with the greater omentum as a possible manner of fistula closure, which was not reported previously according to the knowledge of the authors. Conclusions Newer papers report of successful non-surgical therapy, although the bulk of the literature advocates surgical therapy. Fistula closure with the greater omentum is a possible method of the thoracobiliary fistula treatment.


Introduction
In the case of bronchobiliary (BBF) and pleurobiliary fistulas (PBF), there is an existing pathological communication between the biliary tract and bronchial tree in the first case and pleural space in the second.
The literature indicates several possible causes for the conditions, but all potential causes can be summarized in five groups: [1][2][3][4][5] 1. Congenital bronchobiliary or pleurobiliary fistulas; fistulas that are the result of 2. hepatic hydatid disease or liver abscess (echinococcic, amoebic, pyogenic), 3. biliary tract obstruction (iatrogenic cause or trauma excluded), 4. injury (blunt or penetrant) and 5. iatrogenic fistulas (liver resection, radiofrequency ablation -RFA, bile duct stricture, irradiation, thoracic drainage, etc…). Pathogenesis may be, except in the case of the congenital form of the disease, explained by two mechanisms. 2,4,6 In the first case biliary tract obstruction is the primary reason for fistula formation. Causes may be scars (trauma, surgery, after radiation, etc…), inflammatory diseases, foreign bodies, primary tumours, metastases or granulomas of different aetiologies, which obstruct the bile ducts. The result is the retention of bile proximal to the barrier, the formation of a liver biloma and subsequently the abscess formation. By increasing, the abscess gradually erodes the diaphragm. In case of adhesions between the lower lung lobe and the diaphragm (due to previous pleural or lung pathology), the abscess erodes directly into the lung parenchyma until it reaches the nearest bronchus and a BBF is formed. When no previous pleural pathology is present, the abscess gradually erodes into the pleural space and a PBF with pleural empyema is formed.
In the second case, the formation of a thoracobiliary fistula (TBF) takes place without biliary tract obstruction. In this case a hydatid cyst or a liver abscess is the primary reason for the fistula formation. The abscess can be echinococcic (most often), amoebic, or pyogenic in origin. As described above, the cyst or the abscess gradually enlarge and erode the diaphragm. Depending on the previous state of the pleural space, a BBF or a PBF are formed. The predilection point for the fistula formation is the posteromedial part of the right hemidiaphragm-i.e. the part which is in a direct contact with the area nuda hepatis. 7 Because BBF and PBF are rare phenomena, larger studies on their frequency and the most com-mon causes do not exist. In the literature, liver abscess of echinococcic origin or hydatid disease of the liver are stated as the most frequent causes in the developing countries and also globally. 2,3,5,6,8,9 Information about the most common cause in the developed world is contradictory (Table 1). 1,2,4,5,7,[10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25] The clinical picture of patients with BBF and PBF may be present either in the acute or chronic form. 2,7 Symptoms result from the underlying disease of the liver, biliary tract, and lung pathology. Bilioptysis, whenever encountered, is pathognomonic for bronchobiliary fistula. 2,3,7,8,10,21 In the case of the acute form 2,7 , the patient is in distress, with elevated body temperature, and complains of pain in the lower right part of the chest. In the case of BBF, bilioptysis is present and during physical examination, inspiratory crackles over the lowerright parts of the lung are noticed. The fulminant disease presents in the form of acute respiratory distress syndrome (ARDS). When the patient has a PBF, the cough is dry and irritating, with the absence of respiratory phenomena above the right bottom of the lungs. A chronic TBF 2,7 presents with dry, irritating cough, occasional yellowish sputum, intermittent fever with malaise, and weight loss. The clinical picture mimics recurrent pneumonia.
Beside the pulmonary symptoms there are also symptoms due to the underlying disease of the liver and biliary system. 2,7 There might be upper right quadrant abdominal pain that irradiates to the right shoulder. Jaundice in the case of bile duct obstruction. Sometimes a biliocutaneus fistula can also be encountered.
When examining the laboratory results elevated white cell blood count and C-reactive protein (CRP) is noticed. Anaemia and hypoalbuminaemia are signs of chronic TBF. In the case of bile duct obstruction, elevated levels of direct and total bilirubin are present. 2 The microbiological analysis of sputum reveals the presence of the following microbes: E. coli, Klebsiella spp., Pseudomonas aerug., Enterococcus spp. and Enterobacter cloacae. 3,8,17 On chest x-ray a raised right hemidiaphragm, shadowing in the area of the lower and middle right lobus, atelectasis and pleural effusion might be seen. 6,9 Occasionally, gas-fluid levels are present on abdominal x-ray images. 4,8 In cases when a clinical suspicion for BBF or PBF exists, the first dilemma is which diagnostic method should be used to either confirm or refute our suspicion. The possible diagnostic methods listed in the literature are bronchoscopy 2,21 , bronchogram 2,4 , CT 3,21 ( Figure 1) and MRI 21 that are routinely used  Regarding the therapy there are three possible approaches to treat TBF: surgical, conservative, and the combined approach. The bulk of the literature advocates the surgical approa ch. 1,2,4,5,7,[10][11][12][13][14]17,18,20,21,23,24,29 In 1967, Ferguson and Burford 1 published an article on TBF, summarizing the basic steps necessary for a successful TBF treatment: early aggressive treatment by thoracotomy, adequate subcostal drainage of the hepatic bed under direct vision, secure closure of the diaphragmatic perforation by non-absorbable sutures, decortication for pleurobilia, if necessary lobectomy for bronchobiliary fistula and the awareness of the need for prophylactic decompression of the biliary tree.
In the next decades other authors have published their experience in treating TBF. The studies reviewed included from 2 to 123 cases. 1,2,4,5,7,10-14,16,18-25 By comparing these studies, one can notice that non-traumatic TBF requires more operations for the successful treatment than traumatic TBF. This statement is best illustrated by the publications of Warren et al. 2 and Gugenheim et al. 4 In the first case 15 patients have been cumulatively operated 63 times and in the second case 16 patients have been operated 42 times. In contrast, Oprah and Mandal 12 report of only 17 operative interventions needed for the successful treatment of 14 patients with traumatic TBF. The common denominator in TBF treatment is to remove or solve the abdominal pa-thology that caused TBF formation. Additionally, adequate drainage of the subphrenic space and a secure closure of the diaphragmatic defect are of importance. Regarding the need for decortication or lobectomy, the authors do not share the same opinion. A lot of them think that lobectomy is necessary only occasionally, if the lung tissue is chronically inflamed or damaged 2,4,7 to the point that it could cause problems in the future.
According to the literature the possible ways of TBF closure are: non-absorbable sutures 1 , the use of a mesh 30,31 , AlloDerm® 3 , pericardial fat tissue flap 17 , pleural flap 5 , and the use of a vascularized pedicel of intercostal muscle. 17 In the last decade of the 20 th century reports of conservative ways of TBF management started to appear. In these cases the treatment was comprised of biliary drainage using PTC (percutaneous transhepatic biliary drainage -PTBD) or ERCP and percutaneous drainage of the subphrenic, subhepatic or intrahepatic abscess if it existed. When performing ERCP there is the additional option of endoscopic sphincterotomy (EST) and stent placement or nasobiliary drainage (NBD). 32 The goal of these interventions is to minimize the pressure in the biliary tree, drain an abscess if present, prevent the flow of bile through the TBF and enable the healing of the TBF. In 1996 Yilmaz et al. 15 published a series of 11 cases of BBF after hydatid cyst operation. In all 11 patients an ERCP with nasobiliary drainage was performed, which lead to fistula closure. In some articles histoacyrl embolization of the TBF is stated as an additional option to the procedures listed above. Richter et al. 33 report of successful histoacyrl occlusion of a BBF after a selective duct cannulation during ERCP. Kim et al. 34 report of a successful histoacyrl embolization of BBF under bronchoscopic guidance. According to some authors, there is also the option of applying octreotide -a somatostatin analogue which is thought to reduce the secretion through enteral and biliary fistulas and thus promote their healing. 19,22,35 But it should be kept in mind that there were reports of an increased number of septic and thrombotic complications while using octreotide. 36 The combined approach uses biliary drainage (via ERCP or PTC) and abscess drainage (US or CT guided) in the first step, followed by a delayed surgical intervention. The reason for such a course of action are patients who are initially not candidates for surgery (ARDS, sepsis, comorbidity, etc…) 3 and patients where non-surgical interventions have failed. 8,17,19,22,28

Case report
The patient was referred to our medical centre in September 2009. At the time he was 73 years old. In the past an elective cholecystectomy for cholecystolithiasis was performed. He had essential arterial hypertension (regulated), suffered from a myocardial infarction, and in his youth he also had pulmonary TBC. In July 2009 a right hemicolectomy for colon carcinoma (stage T3N0M0) was performed at a regional general hospital. Preoperatively, pleural thickening in the left apical and right basal part of the lung was described on chest x-ray (due to TBC during his youth).
On follow-up examinations, the progression of the disease (using US and CT) with two metastases in the right hemiliver was diagnosed. One was located in the deep of the right anterior section (border of 5 th and 8 th segment) and was 4 cm in size. The other was located in the 7 th segment and was 2 cm in size. There were also two metastases found in the lungs; one in the 2 nd left and the other in the 6 th right segment. At the multidisciplinary team meeting it was decided that liver surgery should be attempted first. A formal anatomical right hemihepatectomy was planned. In the beginning of October 2009 the patient was operated. With intraoperative inspection and US we found the two tumours in the right hemiliver as depicted by CT. Unexpectedly we also found two metastases in the left hemiliver, located in the 3 th and 4 th segment. Both were 1 cm in diameter and were not visible on the preoperative CT scan. A metastasectomy in the 3 rd and 4 th liver segment was performed. The remaining two metastases in the 7 th segment and right anterior liver section were treated with radiofrequency ablation (RFA). A few days after the surgery the patient reported malaise and increasing pain in the right upper part of the abdomen. His body temperature and inflammatory parameters were elevated. A control US showed a hypoechogenic fluid collection in the area of the liver where RFA had been performed. ERCP demonstrated a biliary leak from the right anterior section into the cavity after RFA, endoscopic papillotomy (EPT) was performed and US-guided puncture and drainage was done, which revealed the fluid to be an infected biloma. Despite drainage and the broad spectrum antimicrobial therapy, the patient's state deteriorated. Pneumonia of the right lower and middle lung lobes and a biliocutaneus fistula developed. A reoperation was performed, which included the evacuation and lavage of the abscess cavity, sutures at the area of biliary leakage and the placement of two drains. After the inter-  vention the patient's state promptly improved. At the end of November 2009 he was discharged from our hospital, with the intention of the further oncological and surgical therapy.
In January 2010 the patient was again urgently admitted to our department. He was complaining of pain in the upper right part of the abdomen and pleuritic pain in the lower right part of the thorax. His body temperature was 39°C and he had a productive cough with yellowish sputum. White cell blood count and CRP were elevated and shadowing was identified in the lower right part of the lungs on chest x-ray. A CT scan of the thorax and abdomen was performed, showing a subphrenic abscess with a BBF. Again an operative evacuation, lavage and drainage of the abscess were performed. Sutures were placed at the site of biliary leakage and at the abdominal ostium of the BBF. A right hepatectomy was not performed, because the remaining liver volume would have been too small. Instead, a right portal vein embolization (PVE) was planned. We speculated that this would lead to atrophy of the right hemiliver, cessation or limitation of biliary leakage and compensatory left liver hypertrophy. The patient was discharged with a drain. The right PVE mentioned followed and his case was presented at an oncological council. Because the number of pulmonary metastases increased, the conclusion of the council was that he should receive palliative oncological therapy.
In August 2010 he was again admitted with malaise, intermittent fever, and cough with yellowish sputum. On CT scan the abscess in the remaining right liver was again present. As it was speculated, the right hemiliver was atrophic and the left liver was hypertrophic as a result of right PVE. The BBF seen on the previous CT scan was still present. Despite the pulmonary progression, we decided to perform a right hepatectomy with the intention to eliminate the inflammatory focus which was the reason for the BBF. Decisive for this step was the left liver hypertrophy taking place after the right PVE. After the right hepatectomy, we intraoperatively confirmed the presence of the BBF by placing a small diameter catheter through the abdominal ostium ( Figure 2) and visualizing the tip of the catheter with a bronchoscope (Figure 3). Because of the pulmonary metastatic progression we did not decide to excise the fistula and the right lower lung lobe. Instead we closed the abdominal ostium of the BBF with a part of the omentum majus ( Figure 4).The omentum was sutured over the abdominal opening of the BBF and was also used to cover the resection surface of the remaining left hemiliver, but was not sutured to it. After lavage a drain was placed. Postoperatively, a prompt decrease of the white cell blood count and CRP, cessation of bilioptysis and the improvement of the patient's general status was recorded. On followup examinations there was no sign of BBF recurrence. The patient expired in March 2011 due to the pulmonary metastatic progression of the disease.

Discussion
By comparing our case to the literature reviewed, one can notice several parallels. Our case describes an iatrogenic BBF. In fact more cases of TBF in the developed world are probably due to trauma and iatrogenic causes than due to liver infection (echinococcic amoebic or pyogenic). Pathogenesis and the formation point of BBF described, coincides with the reviewed papers. The clinical suspicion was in our case confirmed with contrast enhanced CT of the thorax and upper abdomen. According to the literature this is not the most sensitive method. ERCP offers better results and has the additional advantage of biliary tree decompression as shown above. Intraoperative, TBF can be confirmed with cholangoigraphy. In the case of the liver resection, as it was shown in our clinical case, it can also be confirmed by visualizing a small diameter catheter put through the abdominal ostium with a bronchoscope.
The biggest dilemma regarding TBF is the proper treatment. Although the bulk of the literature advocates surgery, non-surgical or conservative therapy is a good alternative in some cases. As shown by Yilmaz et al. 15 , Singh et al. 19 and Ertugrul et al. 37 , the conservative therapy with biliary and abscess drainage can be effective in cases where TBF is the result of liver abscess, complicated liver hydatidosis after surgical cyst removal, biliary tract obstruction due to stones or short strictures and in selected cases of posttraumatic TBF.
For the rest, surgery is probably the best solution. The basic steps outlined by Ferguson and Burford 1 are a good basis but need some comment. Thoracotomy is necessary if the preoperative investigation indicates the need for the lung resection 4,14 or in the case of posttraumatic TBF. 4 Laparotomy is mandatory if biliary tract obstruction which cannot be managed conservative is the cause of TBF. 2,4 In cases of extensive disease two separate incisions (approaches) are probably superior to thoracophrenolaparotomy. 14 Regarding the diphragmal defect closure, we can add our report of BBF closure with omentum majus to those listed above. In our opinion preoperative biliary decompression would act beneficial on TBF therapy outcome in all TBF cases not just those caused by the biliary tract obstruction.
In our case two operations were needed for a successful BBF treatment. Despite the palliative nature of the treatment a surgical approach was chosen. The reason was the persistence of biliopytisis and recurrent septic bursts, although the conservative treatment with ERCP, EPT and US guided abscess drainage had been done first. We think that biliary leakage into the cavity produced by RFA was the reason for failure of the conservative measures. The only possible solution was the right hepatectomy, which removed the inflammatory focus and closure of the fistula as described above. Though this might seem an extensive treatment for a patient with a poor prognosis, it was the only way of improving his quality of life.

Conclusions
Thoracobiliary fistulas are pathological communications between the biliary tract and the bronchial tree or the biliary tract and the pleural space. The first are termed bronchobiliary and the second pleurobiliary fistula. Etiologically, they can be divided into congenital TBF, TBF resulting of liver hydatid disease or liver abscess, biliary tract obstruction, traumatic TBF, and iatrogenic TBF. The article summarizes the characteristics of the clinical picture and laboratory findings of BBF and PBF. The most sensitive methods for TBF confirmation are ERCP, PTC, and fistulography of a biliocutaneus fistula, if it exists. Opinions regarding the therapy differ. Although the bulk of the literature advocates surgical therapy, newer papers report of successful non-surgical therapy. These interventions are comprised of a biliary drainage via ERCP or PTBD, percutaneous US or CT guided drainage of a subdiaphragmatic, subhepatic or intrahepatic abscess if it exists and ERCP or bronchoscopic guided fistula embolization. The authors presented the fistula closure with the greater omentum as a possible method of TBF closure which was according to the authors' knowledge not reported on previously.