Mixed pyogenic and tuberculous liver abscess: clinical suspicion is what matters

Abstract

Tuberculosis (TB) is a rare cause of liver abscess, even in country like India where it is a very common infection. Moreover, tubercular liver abscess (TLA) is the most unusual pattern of hepatic tuberculosis. We report an unusual case of liver abscess in an immune-competent patient presenting only with weight loss. On investigation, initially it appeared pyogenic, but later turn out to be a mixed infection with tuberculosis. He responded well to antibiotic and antitubercular drugs. A mixed pyogenic and TLA is very uncommon. We conclude that, tuberculosis should be suspected in liver abscess, especially in the absence of typical features and failure to respond to antibiotics.

Background

Liver abscess is common in tropical countries like India. Most of them are pyogenic or amoebic abscesses. Uncommonly, it can be a manifestation of hepatic tuberculosis (TB). We report a patient with liver abscess initially thought to be due to Proteus mirabilis. As the clinical presentation of the patient was more consistent with a chronic infection like TB and since the patient did not respond to antibiotics, further workup was carried out which ultimately proved the suspicion. The case has been reported for its unusual presentation and rarity, and to emphasise the fact that clinical suspicion and repeated workup can clinch the diagnosis.

Case presentation

A 60-year-old man presented with weight loss of 8 kg (from 78  to 70 kg) and decreased appetite for 2 months. There was no history of fever, night sweats, cough, haemoptysis, breathlessness, abdominal pain, vomiting or altered bowel habit. He was a smoker (15 pack-years), but not an alcoholic, diabetic or hypertensive. There was no history of TB or high-risk sexual behaviour. On examination, he was a cooperative patient with average build (body mass index 22.3 kg/m²). Blood pressure was 124/74 mm Hg, pulse rate 88/min, respiratory rate 16/min and temperature 97.4 F. General physical examination was within normal limits. In systemic examination, there were diminished breath sounds in right infrascapular region with stony dullness, and tender moderate (3 cm below costal margin) hepatomegaly.

Investigations

On investigation, haemoglobin was 13.2 g% and total lymphocyte count was 8700 cells/mL (P78, L20, E2). Liver function test was normal except for raised alkaline phosphatase (480 IU/L). Rest of biochemistry and urinalysis were normal. In further workup, erythrocyte sedimentation rate was raised to 60 mm in first hour (Westergren); Mantoux test was positive (15×14 mm); sputum was negative for acid-fast bacilli (AFB); hepatitis B surface antigen, antihepatitis C virus and ELISA for HIV were negative. Chest X-ray showed right pleural effusion; however, lung parenchyma was normal. Contrast-enhanced CT scan of the chest and abdomen was carried out, which had a 5.2×4.7 cm² size liver abscess located in segment IV of the liver with right sided pleural effusion (figure 1). Amoebic serology was negative. Pleural fluid aspirated was straw colored, the analysis of which showed 200 cells/mL (90% lymphocyte) and 5.4 g% protein (Light's ratio of protein 0.66). Adenosine deaminase level was not raised (22.7 U/L;<24 U/L) and it was negative for Gram's stain, Ziehl-Neelsen (Z-N) stain, malignant cell and PCR for TB. Liver abscess was aspirated, which had thick foul smelling brownish pus; it was negative on Gram stain and Z-N stain, but grew P mirabilis on culture. According to sensitivity, the patient was started on injection piperacillin-tazobactam. Over the next 2 weeks, there was some improvement in general well-being and no further loss of weight, however, appetite did not improve. Moreover, pleural effusion increased, reaching half of the hemithorax. A repeat pleural tap was carried out and 300 mL of fluid was sent for analysis. Cytology and biochemistry were similar as previously reported, but the centrifuged sample was positive for AFB and PCR for TB. Liver abscess was aspirated again, which also showed AFB on Z-N staining (figure 2).

Figure 1.

Contrast-enhanced CT of the chest and abdomen showing hepatomegaly with abscess in segment 4 of the liver and moderate right-sided effusion.

Figure 2.

Acid-fast bacilli (AFB) staining of liver abscess aspirate showing AFB-positive tubercular bacilli.

Treatment

The patient was started on antitubercular treatment (ATT) regimen with four drugs.

Outcome and follow-up

At the end of 2 months of ATT, weight had increased by 2 kg, appetite had improved and pleural effusion reduced to minimal (figure 3).

Figure 3.

Follow-up chest X-ray at the end of 2 months showing only blunting of right costophrenic angle, with resolution of the effusion.

Discussion

Hepatic involvement is common in disseminated TB. However, isolated hepatic TB is uncommon, probably because low oxygen level in liver is unfavourable for bacilli. It usually has a primary focus in lung or gastrointestinal tract (GIT). Later, it spreads to the liver as a part of haematogenous spread from lung via hepatic artery, and less commonly via portal vein from gastrointestinal focus.^{1 2} Reed et al³ described three morphological pattern of hepatic TB: diffuse involvement of liver with miliary or pulmonary TB; diffuse parenchymal involvement without any evidence of existing TB anywhere (primary miliary TB of liver); and focal lesion in the liver, multiple or solitary, presenting as nodule or abscess. Out of these patterns, isolated tubercular liver abscess (TLA) is the rarest; Essop et al⁴ found TLA in only 0.34% of patients with hepatic TB.

Symptoms of the disease are non-specific: fever, abdominal pain, abdominal mass, anorexia, weight loss, lethargy, weakness and malaise. Hepatomegaly is a common physical finding. Jaundice is an uncommon manifestation. The main differential diagnoses of TLA are hepatoma, pyogenic liver abscess and amoebic liver abscess. The radiological findings of TLA have a low specificity.⁵ It can present as mass on ultrasonography (hypoechoic lesions with hyperechoic rims and complex masses) and CT scan (hypodense lesions). The definitive diagnosis can only be made on histological and bacteriological workup. It is difficult to demonstrate AFB in TLA; pus is positive in only 0–45% of cases.⁶ Cultures can be positive in only 10% of patients, which can increase up to 60% in miliary cases.⁶ PCRis coming as a useful diagnostic tool for hepatic TB. In a study by Diaz et al,⁷ as compare to conventional diagnostic techniques, PCR was positive in 57% of tubercular hepatic granulomas. Another advantage is that it can distinguish Mycobacterium tuberculosis from other Mycobacterium.

Our case had some important peculiarities. Our patient, unlike other reported cases, did not have fever, any symptoms pertaining to lung and GIT or leukocytosis. Similar to our case, Lee et al⁸ reported a case of TLA in a patient who presented only with 3 months history of intermittent chills. Moreover, the spread of infection to liver from lung was unlikely, as there was no parenchymal lesion. It was most probably a case of primary TLA. Also, there was no predisposing factor like diabetes mellitus, HIV, alcoholism and diseases of gut or cancer. A primary localised TLA in an immune-competent patient is quite rare.

Mixed pyogenic and tuberculous infection in liver abscess are rare, but has been reported in the literature. Silverberg et al⁹ reported liver abscess which was co-infected with Escherichia coli, P mirabilis, P morgani, Bacteroides fragilis and M tuberculosis. In another case of a patient with carcinoma colon, Rafailidis et al¹⁰ reported a liver abscess co-infected with Klebsiella pneumoniae, Candida albicans and M tuberculosis. In our case, initially we attributed P mirabilis as the cause of abscess. However, clinical picture was more consistent with tuberculous process and repeated investigation confirmed it. The diagnosis of TB could have been easily missed, had not the index of suspicion for TB been high. Similar observation was made by Singh et al,¹¹ where they picked up TLA when the patient did not respond to antibiotic and amoebicides. The infection with P mirabilis could be a secondary process, due to the spread from the gastrointestinal tract. It could have contributed to abscess formation leading to immediate deterioration before admission.

Management of tuberculous liver abscess is still a subject of discussion. Standard four drug ATT is recommended for 1 year. But, according to Gracey,¹² thick fibrous tissue around the abscesses and large size may prevent antibiotics from reaching the target. Hence, percutaneous drainage, combined with ATT has been used in difficult cases.¹³ Adding further modification, cases has been managed by percutaneous drainage followed with transcatheter infusion of antitubercular drugs.^{14 15} Surgery is indicated if percutaneous aspiration is not successful, or difficult because of site, and multiseptate abscess.¹⁶

Learning points.

• Tuberculosis should be kept in differential diagnosis of liver abscess, especially in the absence of typical features and failure to respond to antibiotics, as also concluded by Lee et al⁸ and Singh et al.¹¹

• Apart from investigation, clinical impression should always be used as the guiding factor of decision-making in patient management.

• Primary tubercular liver abscess in an immune-competent patient is quite uncommon and a co-infection with pyogenic organism is very unusual.