Pancreaticopleural Fistula: A Review of Imaging Diagnosis and Early Endoscopic Intervention

A 49-year-old male with history of chronic alcohol-induced pancreatitis presented with one month of worsening left pleuritic chest pain and shortness of breath. Chest radiograph demonstrated bilateral pleural effusions. Thoracentesis revealed increased amylase in the pleural fluid. Computed tomography (CT) and magnetic resonance cholangiopancreatography (MRCP) showed a fistula tract between the left pleural cavity and pancreas which was confirmed on endoscopic retrograde cholangiopancreatography (ERCP). Patient was treated with placement of a pancreatic stent with complete resolution of the fistula tract approximately in 9 weeks. A systematic literature search was performed on reported cases with pancreaticopleural fistula (PPF) who underwent early therapeutic endoscopy within the last 10 years. Imaging modalities, particularly CT and MRCP, play essential role in prompt preprocedural diagnosis of PPF. Early therapeutic ERCP is an effective and relatively safe treatment option for PPF, so invasive surgery may be avoided.


Introduction
Pancreaticopleural fistula (PPF) is an uncommon complication of chronic pancreatitis, often presenting with recurrent unilateral or bilateral pleural effusions [1,2]. There is usually an abnormal communication to the pleural space from posterior pancreatic duct disruption or pancreatic pseudocyst extension into the pleural cavity [3]. Minimally invasive endoscopic intervention following medical therapy is usually the first choice before invasive surgical management [2,3]. To our knowledge, there is no recent comprehensive data available regarding the outcome of early endoscopic treatment in the patients with PPF. In this study, we present a case of PPF treated successfully by endoscopic retrograde cholangiopancreatography (ERCP) and transpapillary pancreatic duct stent placement. In our case imaging played an important role in the investigation of PPF and followup after ERCP. We performed a systematic literature search of reported PPF cases who underwent early therapeutic endoscopy within the last 10 years, with a review of the utility of diagnostic imaging and the outcome of endoscopic management of PPF.

Methods
A review of the literature was performed using the MEDLINE database with searching of the keywords: pancreaticopleural fistula, pancreatico-pleural fistula, and pancreaticopleural effusion. Inclusion was limited to cases reported in the English or Spanish language between 2007 and 2017 who underwent early ERCP alone or before any operative management. Forty-three patients (including the patient in this study) with PPF were reported during the 2007-2017 period [2,. Data is presented as percentage or mean values ± standard deviation (SD). N throughout the manuscript stands for total number of cases with available data for that variable.

Case Presentation.
A 49-year-old male with history of alcohol-induced pancreatitis presented with 1 month of worsening left pleuritic chest pain and shortness of breath. Initial physical exam was unremarkable, and laboratory tests were only remarkable for increased lactate dehydrogenase and borderline-low albumin. Initial chest radiograph showed moderate left and small right pleural effusions with left retrocardiac opacity ( Figure 1). Computed tomography (CT) pulmonary angiogram was performed and was negative for pulmonary embolism and showed only small bilateral pleural effusions. On admission, patient underwent thoracentesis and approximately 1L of brown fluid from the left pleural space plastic stent was placed in the pancreatic duct and a biliary and pancreatic sphincterotomy was performed to divert the fluid leak. The hospital course was complicated by post-ERCP acute pancreatitis, which was treated with bowel rest and empiric antibiotics. Postprocedure chest radiograph showed worsening pleural effusions and surgical management including distal pancreatectomy was considered. However, surgery was deferred after the patient improved clinically with medical management, and a repeat chest radiograph revealed improvement of the pleural effusions. The patient improved clinically and was discharged with pain medications.
The patient returned to the emergency department within the first week after discharge with complaints of worsening abdominal and chest pain. Another thoracentesis was performed with removal of 400 mL fluid. The study of pleural fluid specimen demonstrated amylase level > 1000 IU/L which was positive for Klebsiella and enterococcus faecalis. The patient was treated with intravenous antibiotics and the pancreatic stent was upsized to 7Fr.
Approximately 9 weeks following the initial ERCP, repeat ERCP demonstrated no extravasation of contrast from the pancreatic tail, and the stent was downsized to 5Fr. Repeat ERCP 15 weeks after the index procedure showed mild diffuse dilatation of the main pancreatic duct with no extravasation of contrast or stenosis ( Figure 4). Therefore, the pancreatic duct stent was removed. Chest radiograph and CT abdomen pelvis showed complete resolution of pleural effusions bilateral as well as the PPF tract (Figures 1 and 2). Table 1 summarizes the demographics and clinical data of previously reported cases of PPF. Seventy-nine percent of the cases were male with a mean age of 50 ± 13 years (N = 41, range:  The majority of the patients (85%, N = 39) received medical therapy before any intervention. The medical therapy included octreotide, somatostatin, total parenteral nutrition, and antibiotics. The time to ERCP after admission was available only in 4 cases and ranged 1-21 days. Post-ERCP acute pancreatitis was reported in 9.5% of cases (N = 21) and superinfection of the pleural fluid occurred in 27.0% of patients (N = 22). Twenty-three out of 43 patients required surgery after the initial ERCP, including distal pancreatectomy (30%), surgical thoracostomy/decortication (30%), pancreaticojejunostomy (18%), exploratory laparoscopy, and external drainage (18%). One patient (4%) underwent surgical sphincterotomy secondary to difficult anatomy for ERCP. ERCP alone and ERCP with surgery were successful in 100% and 96% of patients, respectively. The total duration of the hospital stay was 27 ± 16 days (N = 28). One patient (2%, N = 43) with refractory pleural effusion refused surgical intervention and died.

Demographics and Presentations.
Pancreatic Pleural Fistula (PPF) is a rare entity with an approximate incident rate of 0.4%. Based on this literature review, PPF usually involves males in their late 40s with chronic pancreatitis mainly from excessive alcohol abuse, consistent with previous studies [1]. A smaller percentage of cases may be due to biliary stones, trauma, and idiopathic pancreatitis [1]. Patients with PPF usually present to the emergency room with respiratory symptoms including shortness of breath [37], as seen in our patient, and less commonly with diffuse upper quadrant abdominal pain, nausea, and vomiting without rebound or guarding [1,2,8,27].

Diagnosis.
History and physical examination are nonspecific for PPF and further evaluation with imaging and laboratory work-up is required. In patients with PPF, chest radiograph usually demonstrates unilateral or bilateral pleural effusion, with the left side being more common. Blood laboratory evaluation may show increased levels of gammaglutamyl transferase [37]. Thoracentesis is usually required for symptom relief. Laboratory analysis of the pleural fluid shows an amylase level greater than 1000 IU which may be suggestive for PPF in the absence of malignant cells [1]. In our patient, the initial CT imaging showed the fistula communication between the tail of the pancreas and left pleural space which was confirmed by MRCP and ERCP. From our review, CT was performed on majority of patients with PPF (93%) and was able to identify the fistula in 63% of patients which was comparable with ERCP (70%). Compared to ERCP, CT has the advantages of being noninvasive and relatively less expensive and can be completed in a much shorter time with no sedation required. ERCP has the advantages that it can locate the site of ductal leak and perform simultaneous intervention [2,3,30,37]. In old studies CT was reported to identify fistula tract only in 30% of cases with PPF [1]; application of high resolution CT images with multiplanar reconstruction has likely increased the ability to visualize the fistula tract in newer studies. MRCP showed fistula tract in 83% of patients with PPF in this study, which is similar to reported studies [38]. In addition, CT and MRCP are useful in evaluating additional pancreatic parenchyma and ductal structural abnormalities such as pancreatic mass, pseudocysts, or collections [32,35], which is valuable for surgical or interventional planning [1].

Management and Outcome.
The majority of patients (85%) in the studies we reviewed received medical therapy before any intervention, consistent with previous reports [1]. The medical therapy included octreotide, somatostatin, total parenteral nutrition, and antibiotics. King et al. [1] reported the outcome of early operative management in 63 patients with PPF between 1970 and 2008. In their study, the early operative management had a success rate of 94%. Six patients underwent ERCP which was successful in all cases. ERCP alone and ERCP with surgery was successful in 100% and 96% of patients in our review, which is comparable to king et al. study. The total duration of the hospital stay in this review was 27 ± 16 days which was relatively shorter than reported 40 ± 6 days for patients who received medical therapy and surgery [1]. Postprocedural acute pancreatitis and superinfection of the pleural fluid were the most common complications after ERCP based on our review. Endoscopic pancreatico-or cystogastric plastic stent placement may be helpful to avoid postprocedural acute pancreatitis. Some other less common complications are intra-abdominal infections, wound infections, entrapped lung and in some cases diabetes mellitus after surgery [2]. Only one patient (2%) died after the index endoscopic procedure who had refractory pleural effusion and refused surgical intervention. The mortality rate of patients with PPF who underwent surgical intervention with or without medical therapy has been reported as high as 12-17% [1,3]. In addition, in patients after ERCP ductal stenting there is a faster transition to feeding and reduction of recovery time. ERCP is both a diagnostic and therapeutic tool which has reduced the hospital stay and mortality rate compared to the traditional operative management [2][3][4]37], although further prospective studies are still required to compare the cost-effectiveness and long-term outcome of ERCP with surgery.

Conclusion
Imaging modalities, particularly CT and MRCP, play essential role in prompt preprocedural diagnosis of PPF. Early therapeutic ERCP is an effective and relatively safe treatment option for PPF, so invasive surgery may be avoided.

Conflicts of Interest
The authors declare that they have no conflicts of interest.