Diagnosis and Management of a Pancreaticopleural Fistula in a Patient with AIDS and a Large Pleural Effusion

Abstract

Pleural effusions typically present with nonspecific pulmonary complaints in the setting of either acute or chronic diseases. In the general population, these illnesses include congestive heart failure, infection, and malignancy. However, in people living with HIV/AIDS (PLWHA), pleural effusions often result from opportunistic infections and AIDS-defining malignancies, such as Kaposi sarcoma and non-Hodgkin lymphoma. Since the introduction of highly active antiretroviral therapy, there has been a decline in the frequency of AIDS-defining opportunistic infections and AIDS-defining cancers and an increase in certain non-AIDS-defining malignancies including lung cancer. Throughout this period, longer life expectancy in PLWHA has contributed to an increased risk of those chronic diseases that can result in pleural effusions. This case describes an HIV-infected man who was an active cigarette smoker and alcoholic and who presented with a large pleural effusion of uncertain etiology. The authors review several important noncardiac risk factors associated with pleural effusions in PLWHA. The authors also emphasize the importance of obtaining a detailed medical history and the use of appropriate imaging and laboratory tests in order to identify an underlying cause and to provide optimal treatment.

Case Presentation

A 42-year-old man with AIDS and slowly progressive multifocal leukoencephalopathy of 8 years’ duration presented to medical attention with several days of cough and a blood-tinged sputum. He had no associated weight loss, fever, or night sweats. Despite a history of variable compliance with highly active antiretroviral therapy (HAART) and multidrug-resistant HIV infection, on his current salvage regimen of etravirine, maraviroc, and atazanavir-boosted with ritonavir (ATV/r), he had achieved a stable CD4 count >300 cells/mm³ and an HIV viral load <40 copies/mL. His other medications included levetiracetam for seizure prophylaxis and a calcium–vitamin D over-the-counter supplement. He had a >20 pack-year smoking history and reported drinking at least 2 large glasses of vodka daily for the past 10 years. On physical examination, he was normotensive with a heart rate of 90 beats/min and a respiratory rate of 16/min. Breath sounds were diminished throughout the left lung. Complete blood count, chemistry, and liver function tests were all within a broad range of normal, whereas a chest radiograph (CXR) demonstrated a large, free-flowing, left-sided pleural effusion. Following a short course of empiric antibiotics, the cough and hemoptysis resolved, yet his pleural effusion persisted. Two months later, he reluctantly agreed to an ultrasound-guided thoracentesis and 1000 cm³ of clear and yellow fluid was drained from the left pleural cavity with the following characteristics—protein 3.9 g/dL, glucose 95 mg/dL, lactate dehydrogenase 230 U/L, and white blood cell (WBC) 318 × 10⁶/L (segmented neutrophils 1%, large mononuclear lymphocytes 60%, and small mononuclear lymphocytes 30%). Seven days later, a repeat CXR showed reaccumulation of the left-sided pleural effusion.

For the next 2 months, the patient remained asymptomatic and during this interval, he refused further investigations. Shortly thereafter, he presented with acutely worse dyspnea on exertion and hemoptysis. He continued to deny any fevers, chills, nausea, vomiting, chest pain, or productive cough. Due to a leukocytosis of 18 × 10⁹/L, he was hospitalized and started on broad-spectrum parenteral antibiotics for presumed pneumonia. A CXR showed a stable, free-flowing, left-sided pleural effusion with interval development of a left-sided hydropneumothorax for which a chest tube was placed. A contrast-enhancing computed tomography (CT) of the thorax highlighted the known large left hydropneumothorax (Figure 1A and B) as well as a mixed fluid collection in the tail of the pancreas (Figure 1A). Repeat thoracentesis and fluid analysis showed an uncomplicated, exudative effusion, with a pH 7.3, normal protein and glucose, WBC 356 × 10⁶/L (segmented neutrophils 14%, large mononuclear lymphocytes 85%, and small mononuclear lymphocytes 1%). Cytospin preparations showed reactive changes with no malignancy; special studies for fungus and mycobacterium species were unremarkable. Pleural fluid analysis showed a pleural amylase level of 22 000 U/L and pleura lipase >12 000 U/L. This corresponded to an elevated serum amylase of 524 U/L and lipase of 395 U/L.

Figure 1.

A, Contrast-enhanced coronal computed tomography (CT) obtained the day of admission demonstrates a large left pleural effusion (P). A 4.3-cm hemorrhagic pseudocyst (arrow) was incidentally noted adjacent to the tail of the pancreas. A left-sided pneumothorax was also present (not shown). B, Sagittal reformation shows continuity of the peripancreatic inflammatory process with the left hemidiaphragm and pleural space (arrow).

A diagnosis of chronic pancreatitis was further supported by endoscopic retrograde cholangiopancreatography (ERCP) findings of a high-grade pancreatic duct stricture and a mixed solid and cystic lesion toward the tail of the pancreas (Figure 2). Fine needle aspiration cytology showed no malignancy. After ERCP with stent placement across the pancreatic duct stricture, there was a decrease in the size of the fluid collection on subsequent magnetic resonance cholangiopancreatography (MRCP). Magnetic resonance cholangiopancreatography showed no evidence of cholelithiasis or pancreatic malignancy. Because the patient’s hydropneumothorax and trapped left lung failed to resolve with nonoperative management, he underwent a left decortication with pleurectomy. Three months later, a repeat ERCP was performed. His stent was removed and a pancreatogram showed multiple irregular side branches at the tail of the pancreas; but there was resolution of the fluid collection and pancreatic duct leak. At last follow-up, he denied any residual respiratory or abdominal symptoms but reported ongoing tobacco and alcohol use. A follow-up CT scan showed no residual pneumothorax or pleural effusion and resolution of peripancreatic fluid collection.

Figure 2.

Fluoroscopic image acquired during endoscopic retrograde cholangiopancreatography (ERCP) shows irregularity in the pancreatic tail duct (arrow). Contrast extravasation is seen outside the pancreatic duct (arrowhead). The common bile duct is labeled “B” and the pancreatic duct is labeled “P.”

Discussion

In the general population, the most common causes of pleural effusions are congestive heart failure (CHF; 37%), pneumonia (22%), and malignancy (11%). ¹ In contrast, in people living with HIV/AIDS (PLWHA), infections and malignancies are more common causes of pleural effusions (62.5% and 15%, respectively). ² Despite a decline in the prevalence of infection and malignancy since the introduction of HAART, the initial workup of pleural effusions in HIV-positive individuals should focus on these etiologies. ^{3 –5} As this case illustrates, diseases typically associated with aging become more relevant to consider in PLWHA and these diseases may manifest 10 or 15 years earlier than those seen in otherwise healthy non-HIV infected cohorts.

Infection remains the leading cause of pleural effusions among PLWHA, most often due to bacterial community–acquired pneumonia (BCAP). ² Streptococcus pneumonia infections are the most common, followed by Haemophilus influenza, Pseudomonas aeruginosa, and Legionella pneumophila. ⁶ Bacterial community–acquired pneumonia in PLWHA may present with acute onset of fever, chills, and a productive cough or may present atypically with a subacute low-grade fever, nonproductive cough, and dyspnea. Risk factors include current cigarette smoking, alcohol abuse, intravenous drug use, cirrhosis, CD4 count <200 cells/mm³, and uncontrolled HIV-viral replication. ⁶ Estimates of the incidence of pleural effusions among PLWHA with BCAP vary widely from as low as 12% to as high as 47%. ² A definitive diagnosis is made by sputum, blood, or pleural fluid culture and pleural fluid analysis. Early empiric antibiotic treatment and hospitalization are recommended for all HIV-positive patients with BCAP who have a CD4 count <200 cells/mm³. Further treatment algorithms are based on pneumonia severity index. ⁷

Although less common than BCAP in developed countries, worldwide an estimated 8% to 12% of pleural effusions in PLWHA are due to Mycobacterium tuberculosis infection. ² Diagnosis is based on history, imaging findings, and laboratory evaluation with purified protein-derivative skin testing, sputum acid-fast bacillus stain and culture, and pleural fluid culture. Tubercular pleural effusions are more commonly seen in HIV-positive patients with CD4 counts >200 cells/mm³. ⁸ Less common infectious causes of pleural effusions include Pneumocystis jiroveci (4%) and cryptococcus. ²

Malignancy remains the second leading cause of pleural effusions among PLWHA, ² but the relative frequencies and typical presentations of HIV-related malignancies have changed since the advent of HAART. ^9,10 Kaposi sarcoma (KS), once the most common malignancy among PLWHA, can present with pulmonary involvement. In the pre-HAART era, the incidence of KS was estimated to be 12.5 per 1000 person-years, with pulmonary involvement in approximately 15% of patients with KS, but with HAART, the incidence of KS has fallen to 4.9 per 1000 person-years and pulmonary involvement is seen far less frequently than it once was and is now largely restricted to those patients with CD4 counts less than 50 cells/mm³ . ^2,5 Among patients having KS with pulmonary involvement, the prevalence of pleural effusions varies widely from 15% to 95%. ² They typically present with nonspecific findings of progressive shortness of breath, a nonproductive cough, and fever ¹¹ ; up to 25% will have no history of cutaneous involvement. ¹² Pleural fluid in patients with KS having effusions is typically exudative and blood tinged with monocellular predominance, but there are no specific pleural fluid findings. ² After an otherwise negative workup, a diagnosis of KS-associated pleural effusion was made clinically based on the evidence of endobronchial KS lesions by imaging or bronchoscopy. Without treatment, pleural involvement of KS is associated with short survival measured in months; although regression can occur with HAART alone, more commonly such patients are treated with cytotoxic chemotherapy. ¹³

Non-Hodgkin lymphoma (NHL) accounts for approximately 3% of pleural effusions in PLWHA. ² Both diffuse large B-cell lymphoma and Burkitt lymphoma can present with secondary effusions characterized as exudative effusions with elevated WBC. Diagnosis is confirmed by cellular morphology and detection of specific clonal B-cell marker expression. An estimated 4% of PLWHA with NHL will have primary effusion lymphoma (PEL). ¹⁴ This rare, aggressive large B-cell neoplasm is associated with human herpesvirus type 8 coinfection. Primary effusion lymphoma should be suspected in HIV-positive patients with low CD4 counts; a malignant pleural, pericardial, or abdominal effusion; and no detectable tumor mass. Up to half of patients with PEL will also have a diagnosis of KS, whereas 75% will have evidence of Epstein-Barr virus infection. ^15,16 Primary effusion lymphoma carries a poor prognosis, and current treatment strategies involve both chemotherapy and antiretroviral therapy. ¹⁴

Less frequent causes of pleural effusions among PLWHA include CHF (1.7% of cases), renal failure (2.8%), hypoalbuminemia (3.9%), as well as pancreatitis, hepatic cirrhosis, and pericarditis (approximately 1% each). ² In the last 20 years, the life expectancy of PLWHA has increased along with the prevalence and burden of chronic diseases, in particular CHF due to coronary arterial disease and cardiomyopathy as well as cirrhosis due to hepatitis C coinfection. ¹⁷

Additionally, PLWHA are at risk of pancreatitis due to common risk factors such as gallstones, alcohol abuse, or a greater than 20 pack-year smoking history. ¹⁸ Pancreatitis rates also may be higher due to the increased prevalence of smoking among HIV-infected patients as well as HIV treatment–specific risk factors such as hypertriglyceridemia due to protease inhibitors or mitochondrial toxicity associated with elevated intracellular levels of didanosine (ddI). ⁹ Among all patients with chronic pancreatitis, copleural fistulae (PPFs) occur in only 0.4% of patients when disruptions in the pancreatic duct or rupture of a peripancreatic fluid collection lead to extravasation of pancreatic fluid. ^{19 –22} Pancreatic fluid can fistulalize through the esophageal or the aortic hiatus or directly through the diagram, resulting in a left-sided, right-sided, or bilateral pleural effusion. ^{22 –24} Pancreaticopleural fistulae are characterized as exudative effusions with high protein and amylase levels. A diagnosis is supported by visualization of the pancreatic duct disruption and fistula tract by MRCP or ERCP. Magnetic resonance cholangiopancreatography is the preferred imaging modality as it is noninvasive, does not require contrast, and has a higher sensitivity than ERCP (80% versus 48%-78%). ²¹ In contrast, ERCP is operator dependent, invasive, and can be limited by severe strictures of the pancreatic duct. ²⁵ However, ERCP often provides definitive treatment for PPF and can spare the patient a partial pancreatectomy. ²¹ In cases where endoscopic treatment is not successful, surgical interventions are required for definitive treatment. ²⁶

Conclusion

This case illustrates the extensive workup, early empiric treatment, and tertiary-level care often needed to manage complications of HIV infection and pancreatitis. After stabilization of the patient with placement of a chest tube and empiric antibiotic coverage, efforts are focused on identifying the underlying etiology of the effusion. Our patient’s HIV serostatus alone was sufficiently compelling to immediately raise the concern of infection, despite his well-preserved CD4 count and nondetectable viral load. Early identification of opportunistic infection is critical, since even after the survival gains seen with the introduction of HAART, the 5-year mortality after an AIDS-opportunistic illness is still only 65%. ²⁷ Our patient’s initial history of pulmonary complaints and subsequent constitutional symptoms was consistent with infection, but his lack of improvement after empiric antibiotics, his significant smoking history, and long-term immunosuppressed status raised the concern for malignancy, leading us to repeat a thoracentesis for further cytological examination. It was not until a CT scan was performed that an incidental finding suggested pancreatitis as the underlying cause of the pleural effusion. This lead to a reevaluation of his history, including risk factors for pancreatitis such as alcohol and tobacco abuse. However, these same risk factors also placed him at increased risk of cardiovascular and hepatic disease. Although PLWHA who present with a pulmonary effusion should be evaluated for both malignant and infectious causes, the improvement in survival and immune reconstitution associated with HAART therapy means that PLWHA are now at increased risk of developing many of the same chronic diseases that afflict age-matched HIV-negative groups.

This rise in the prevalence and burden of chronic diseases in PLWHA increases the importance of engaging with HIV-positive patients to actively modify their risk factors for ill health. In this case, it was the patient’s long-term alcohol and tobacco abuse, not directly his HIV infection, which lead to the chronic fibrosis and pancreatic duct stricturing that resulted in a large hydropneumothorax and trapped lung requiring a prolonged hospital course with surgery and multiple procedures. Because HIV-positive patients are now exposed to decades of chronic inflammation and the potentially toxic effects of medications, it is becoming increasingly important to address how to modify those risk factors which can reduce chronic disease and their associated morbidity and mortality.