Percutaneous drainage of tubercular liver abscess

Authors


To the Editor

We would like to report our experience of managing tubercular liver abscess. Between 2002 and 2005 we have treated five cases of tubercular liver abscess. All patients were male with mean age of 53 years (34–63 years). All patients presented with clinical features suggestive of amoebic liver abscess. Two patients had received antitubercular treatment in the past; one patient for Pott's spine 12 years ago and the other for tubercular cervical lymphadenopathy 3 years ago.

The laboratory investigations showed leukocytosis with elevated ESR in all patients. Liver function tests showed elevated transaminases (upto three times normal) and minimal elevation of serum bilirubin and alkaline phosphatase. Ultrasonography revealed large right lobe liver abscess (mean diameter 7.5 cm; range 6.5–9 cm) in four patients and a 4.5 cm diameter abscess in the left lobe in the fifth patient. The rest of the liver and other viscera were normal and there was no ascites or retroperitoneal lymphadenopathy in any patient. Plain film of the chest showed elevated right hemidiaphragm in three cases whereas it was normal in the other two patients. All patients were started on metronidazole but did not improve after 48 h. Percutaneous aspiration was carried out in four patients with right lobe abscess and broad spectrum antibiotics were added. Gram stain of the aspirate was negative and cultures were sterile in all four. As there was no improvement after another 48 h of antibiotics and metronidazole, percutaneous drain was inserted in these four patients. After drainage, pus was evaluated for Mycobacterium tuberculosis by acid fast stain and polymerase chain reaction (PCR). Stain for acid fast bacilli (AFB) was positive in two patients and PCR-assay based test for M. tuberculosis was positive in all four.

PCR for tuberculosis, Gram stain, AFB stain and culture for pyogenic bacteria was done on the diagnostic aspirate in the patient with left lobe abscess who did not respond to 48 h of metronidazole. Gram stain and AFB stain were negative and culture for pyogenic organisms was sterile whereas PCR for M. tuberculosis was positive.

All patients were started on four drug antitubercular therapy (ATT). Four patients with percutaneously placed drains showed gradual improvement in symptoms. Drain output was high in all patients and persisted beyond 2 weeks in all the four patients. The mean duration of drainage was 26 days (19–33 days). ATT in these patients was continued for 12 months. The fifth patient was directly started on ATT after diagnostic aspiration but showed minimal clinical improvement after 3 weeks of treatment. Repeat USG showed 3 cm heteroechoic lesion in the left lobe of liver. This patient underwent laparotomy and left lateral segmentectomy. The histological examination showed features of organized abscess with multiple tubercular granulomas in the wall. He had an uneventful postoperative course and was discharged on ATT for 12 months.

On mean follow-up of 17 months (8–21 months) all patient are asymptomatic with normal ESR and liver functions.

Primary tubercular liver abscess are uncommon. Most cases are diagnosed late, either after percutaneous drainage or aspiration when the abscess do not respond to antibiotics. Tubercular bacilli may reach the liver via hematogenous route, lymphatics and adjacent abdominal viscera (1, 2). Alvarez has classified tuberculosis of the liver into three types; military, granulomatous hepatitis and localized type (abscess or nodular mass like) (1). Ultrasonographic and computerized tomography features of tubercular liver abscess are nonspecific (3, 4). AFB are sometimes demonstrable in the pus or in the biopsy from the abscess wall. (3, 5). PCR-based test on liver biopsy or aspirated pus may give better yield compared with AFB staining (6, 7).

High index of suspicion is required for diagnosis. In patients with liver abscess who do not show typical features of pyogenic/amoebic abscess and who fail to respond to antibiotics/amoebicidal drugs, the possibility of tubercular abscess should be considered. Radiologically guided diagnostic aspiration with staining for AFB and PCR help in confirming the diagnosis. Antitubercular drugs alone are sufficient for treating most such abscesses. Many patients, however, undergo percutaneous drainage as the diagnosis is not suspected before percutaneous tube drainage. These patients require long duration percutaneous drainage along with multidrug anti-tubercular treatment (3). Liver resection is required in few cases which do not respond to medical management or when there is suspicion of a mass (3, 5).

We conclude that tubercular liver abscess should be an important differential diagnosis in patients with liver abscess not responding to amoebicidal drugs and broad spectrum antibiotics, specially in regions with high prevalence of M. tuberculosis infection. However, in most cases the diagnosis is made on PCR or staining the pus for AFB when patients fail to improve even after percutaneous drainage. Multidrug antitubercular treatment is required along with long-term percutaneous drainage for managing these patients.

Ancillary