A 61-year-old Cambodian woman with compensated cirrhosis secondary to chronic hepatitis B virus infection presented with abdominal swelling associated with fatigue and anorexia. Physical examination revealed fever, tachycardia, and scleral icterus. Her abdomen was distended and tense with flank dullness. Laboratory testing showed mild elevations in alanine aminotransferase, aspartate aminotransferase, and alkaline phosphatase values with a total bilirubin level of 4.7 mg/dL (normal = 0.1-1.0 mg/dL), a direct bilirubin level of 2.0 mg/dL (normal = 0.0-0.3 mg/dL), and an international normalized ratio of 1.7. A computed tomography scan of the abdomen showed a cirrhotic liver with splenomegaly and a large amount of ascites (panel A). Analysis of the ascitic fluid showed that it was serous in nature with a nucleated cell level of 487/μL (69% lymphocytes, 27% monocytes, and 3% neutrophils) and a total protein level of 3.4 g/dL. The serum-ascites-albumin gradient (SAAG) was 0.9 g/dL.
The ascitic fluid parameters suggested an infectious etiology; testing for viral, fungal, parasitic, and autoimmune etiologies was unrevealing. There was a history of latent tuberculosis, and the patient had been treated with isoniazid 20 years previously. QuantiFERON and human immunodeficiency virus testing were negative. In addition, an acid-fast stain and culture of the ascitic fluid were negative for tuberculosis. Despite the negative evaluation, tuberculous peritonitis (TBP) was strongly suspected. Laparoscopy with peritoneal biopsy was performed and showed a thickened and inflamed bowel with adhesions to the anterior wall (panel B) and multiple, small, white peritoneal implants (panel C). A histopathological examination revealed multiple granulomas with giant cells (panel D). A culture of the peritoneal biopsy sample was positive for Mycobacterium tuberculosis. She was given a diagnosis of TBP and was treated with a quadruple-drug therapy consisting of isoniazid, rifampin, ethambutol, and pyrazinamide. The ascites resolved, and she continued to do well at follow-up.
The peritoneum is a common extrapulmonary site of tuberculous infection.1 TBP is likely increasing in incidence in developed countries because of population migrations and the increased use of immunosuppressive therapies. Like our patient, patients with cirrhosis are at higher risk of developing TBP. Infections of the peritoneum are most likely due to secondary hematogenous spread from latent tuberculous foci. The disease can present in three different forms: a wet-ascitic form, a fibrotic-fixed form, or a dry-plastic form. Our patient had the wet-ascitic form, which is the most common type of TBP. Regardless of the type, the symptoms and signs of TBP are nonspecific and include abdominal pain, ascites, and weight loss.2
Ascitic fluid analyses of patients with TBP reveal high protein levels (>2.5 mg/dL) and high leukocyte counts with a predominance of lymphocytes. The SAAG is less than 1.1 g/dL, except in patients with cirrhosis and portal hypertensive ascites, for whom the SAAG remains greater than 1.1 g/dL.3 Other conditions causing lymphocyte-predominant, high-protein, low-SAAG ascitic fluid include fungal peritonitis (Histoplasma and Cryptococcus), peritoneal carcinomatosis, and lupus serositis; making the diagnosis of TBP a challenge. Ziehl-Neelsen staining of ascitic fluid is unreliable and is positive in only 3% of TBP cases.2 A mycobacterial culture requires prolonged incubation and is positive in only a third of patients.2 Tuberculin skin testing and interferon-γ assays (QuantiFERON) are useful in the diagnosis of latent infections; TBP is an active infection, and positive results from these tests offer only supportive, not diagnostic information.2 Chest radiography may show evidence of active and old pulmonary tuberculosis infections in 14% and 30% of patients, respectively. Classic findings on abdominal computed tomography scans, which were absent in this case, include lymphadenopathy, high-attenuation ascites, peritoneal/mesenteric thickening, and omental caking.2 Adjunctive tests such as ascitic fluid adenosine deaminase and polymerase chain reaction for Mycobacterium may help in the diagnosis of TBP; however, their yield is typically low, and their role in diagnosing TBP is undefined.2
The diagnostic test of choice for TBP is laparoscopy with peritoneal biopsy. As in this case, a classic finding on a visual inspection of the peritoneum is whitish, miliary nodules that are less than 5 mm in size and are scattered over the parietal peritoneum. Other well-described findings include turbid ascites with fibrinous strands between the bowel and the peritoneum.
In conclusion, this case illustrates the challenge of diagnosing TBP because noninvasive tests such as acid-fast staining and culturing of the ascitic fluid are usually insufficient. It is often necessary to perform laparoscopy and peritoneal biopsy to confirm this diagnosis. Laparoscopy and peritoneal biopsy remain the most reliable and expedient methods for diagnosing TBP.