Rare and unusual case of hepatic and disseminated tuberculosis in an immunocompetent patient

Abstract

Mycobacterium tuberculosis complex disease (tuberculosis (TB)) of the liver is rare and liver abscesses as a result are even rarer. In an immunocompetent individual, the disease tends to be localised. To the best of our knowledge, we report one of the most severe TB involvements of the liver in an immunocompetent individual. A young woman with a history of previous TB infection, presented in septic shock. Scans showed a liver filled with possible abscesses, one of which was aspirated and confirmed TB. Multiple HIV tests were negative but she remained lymphopaenic. Although she improved substantially with anti-tuberculous treatment, she later developed non-tuberculous central nervous system disease that we were unable to fully explain. Despite a stormy recovery period, she continues to do well.

Background

Tuberculosis (TB) is caused by the Mycobacterium tuberculosis complex. The disease normally runs a chronic indolent course due to the relatively slow-growing nature of the bacteria. It can affect virtually every organ in the body and its clinical manifestation depends on a complex interplay between the causative organism and the immune system of the body over a period of time.¹ In most cases, the immune system is able to keep the bacterium in-check as it does in patients with latent TB. As living standards have improved, the latter part of the twentieth century saw a decline in the global disease burden. However, in recent times, diseases like HIV, and drugs such as chemotherapy that promote suppression of the immune system and have led to a resurgence.¹

Our patient was not only immunocompetent but the burden of her disease was in the liver, creating further diagnostic uncertainty due to the unusual presentation. Hepatic TB comprises only 1% of all TB infections.² Based on imaging, hepatic TB can be classified as micronodular or macronodular. The micronodular form, which normally indicates miliary TB, is more common than the macronodular type which is described as hypodense, calcified or as abscesses on CT scans.¹ This patient had very rare macronodular involvement of the liver with a very rare pattern of severity compared with other cases described in literature. Her subsequent clinical sequelae are also worthy of note.

Case presentation

A university student, who lived most of her life in Zimbabwe, presented to a local emergency department with features of severe sepsis. During her childhood years, she had a positive history of contact with an uncle who had TB and a grandma that died of TB. At the age of 17 years, she was diagnosed with pulmonary TB after she submitted a sputum sample to a health centre in Zimbabwe. Unfortunately, she only took one dose of her anti-tuberculous treatment before discontinuing as she felt better.

Prior to her presentation in the UK, she studied in the USA. She presented to a hospital in Kansas, with a history of right upper quadrant pain, breathlessness and fever. She was informed that a CT scan showed pleural effusion, multiple liver lesions (deemed at that point in the USA to be simple cysts) and some evidence of cholecystitis. A chest drain was inserted; she was treated with piperacillin/tazobactam and levofloxacin but with seemingly limited response (note that no diagnosis of TB was formally made at this point). A working diagnosis of acalculous cholecystitis due to Epstein-Barr virus infection was considered but further tests were pending. On feeling better in herself, she requested discharged against medical advice.

While she was in the UK, her symptoms continued. Her sister noticed that she was talking irrationally during a telephone conversation and asked a friend to check on her. She was subsequently admitted to the local intensive care unit with severe sepsis and after she was stabilised, transferred to our infectious diseases unit.

Investigations

Initially in the UK, CT scan of the abdomen and pelvis showed multiple hepatic lesions, moderate fluid in pelvis, some omental stranding, paraaortic, periportal and mesenteric lymphadenopathy, likely mild terminal ileal thickening and a cyst in spleen (figure 1). The CT thorax findings were: bibasal consolidation, pleural and pericardial effusion, peribronchial ground glass appearance, focal bronchial dilatation but no lymphadenopathy. The only positive finding of a CT head was multiple low attenuating/cystic lesions in the scalp but no intracranial pathology. On examination, her abdomen was mildly distended due to her organomegaly.

Figure 1.

(A and B) Initial and early CT scans showing multiple liver abscess.

Her initial full blood count was as follows: haemoglobin 68 g/L, platelets 303×10⁹/L, mean cell volume 84fL, white cell count (WCC) 3.7×10⁹/L, neuts 3.5×10⁹/L, monocytes 0.1×10⁹/L, lymphocytes 0.1×10⁹/L, eosinophils 0.0×10⁹/L, basophils 0.0×10⁹/L; her C reactive protein was 215 mg/L. Apart from mild derangement in albumin (25 g/L) and a prothrombin time of 22 s, the other indices of liver function and the renal function were normal. A CD4 count of 4 was noted in subsequent blood samples but no other cause for this was found except her concurrent infection; there was no evidence in the notes from her US admission that they found any evidence of immunocompromise either.

Sputum was negative for acid-fast bacilli (AFB). Multiple blood cultures for Mycobacterium and early morning urine samples for TB were negative with a pH noted of 7.5 (although this value was recorded during treatment). Urine microscopy, culture and sensitivity was also negative. While the pleural fluid was negative for TB microscopy and culture, an aspirate of the liver cyst was strongly positive for TB on auramine stains; the cepheid confirmed it as rifampicin sensitive M. tuberculosis. Subsequent results from the reference laboratory also confirmed that the isolate was sensitive for first line TB medications.

Other relevant tests that came back negative or normal include: repeated tests for HIV and HIV viral load, Human T-Lymphotropic Virus (HTLV), faecal microscopy for ova, cysts and parasites, serology for hepatitis B, hepatitis C, hydatid cyst disease, Entamoeba, syphilis, Toxoplasma, filariasis, Strongyloides, Schistosoma, Aspergillus antigen, Cryptococcal antigen, beta-d-glucan. Multiple blood and urine cultures, autoimmune screens, immunoglobulins levels and compliments were negative or insignificant. An echocardiogram and oesophagogastroduodenoscopy (OGD) showed no significant findings. A bone marrow aspirate, taken due to persistent lymphopaenia was unfortunately insufficient.

Differential diagnosis

Given her initial presentation, she was treated for intra-abdominal sepsis. The initial differential was either liver abscesses or hydatid disease. During the early course of her illness, we did also cover with broad spectrum antibiotics in case of super-added alternative bacterial infection. However, in discussion with the gastrointestinal (GI) radiologist and parasitologists in London, this was deemed less likely and a decision was made to aspirate the liver lesions to confirm a diagnosis of presumed TB.

Treatment

Broad spectrum antibiotic therapy was initially started. Six days after starting these antibiotics, anti-tuberculous medications were added given the positive liver aspirate for AFB (figure 2). Broad spectrum antibiotics were continued for 2 weeks to cover for possible multi-drug resistant organisms and intra-abdominal sepsis. With this regime, her inotropic support was weaned and she was stepped down from the intensive care unit.

Figure 2.

Auramine stain of liver aspirate showing multiple acid-fast bacilli.

Her TB was treated with first line TB medications (rifampicin, isoniazid, pyrazinamide and ethambutol) for 2 months and this was subsequently stepped down to rifampicin and isoniazid.

She was transferred to our infectious diseases unit where she continued to spike low-grade temperatures with sinus tachycardia. A repeat CT scan was ordered to review her liver and look for a cause of increasing breathlessness. This revealed bibasal consolidation and reactive effusions, as well as a likely iatrogenic subcapsular haematoma. The effusions were aspirated but negative for TB. There was also concern of an evolving ileocaecal tuberculoma although this was dismissed on subsequent imaging; the hepatic images showed marked improvement with treatment (figure 3).

Figure 3.

CT scan showing a remarkable improvement of the liver after 3 months of treatment.

She continued to be pancytopenic with neutrophils dropping to as low as 0.22×10⁹/L. A bone marrow biopsy was done, but unfortunately the sample was not sufficient enough to make a diagnosis and it was not in her best interests to repeat the test. Given her lymphopaenia, a multi-disciplinary team (MDT) decision was made to start pneumocystis pneumonia prophylaxis with co-trimoxazole. The WCC and platelets recovered subsequently with granulocyte-colony stimulating factor treatment.

During this time her haemoglobin level dropped and she suffered two hypotensive episodes with a systolic blood pressure of 85–90 mm Hg. Blood transfusions were given and an OGD confirmed a normal upper GI tract. She also developed an acute kidney injury, initially thought to be medication related. All nephrotoxic agents, including her anti-tuberculous treatment, were suspended. Renal biopsy confirmed an acute tubular necrosis, likely due to her hypotensive episodes. She was started on high-dose steroids and following discussion with renal physicians, her anti-tuberculous medications were restarted. Her renal injury was fluid responsive.

While on the anti-TB medication, she made steady improvement with the temperatures and the tachycardia starting to settle. Unfortunately, her mood remained low and there was some concern regarding her change in behaviour. Her oral intake was poor and coupled with her poor renal function, necessitated careful fluid management. MRI head was suggestive of central pontine myelinolysis rather than TB disease. She was unable to tolerate lumbar puncture despite attempts by interventional radiology. A decision was made in our MDT to treat with at least 12 months of anti-tuberculous therapy to cover for the possibility of central nervous system disease, but with follow-up by the team we referred to to adjust this in accordance with clinical response.

Outcome and follow-up

Having made significant improvement, our patient was keen to transfer her care to be closer to family in another city. Unfortunately, we later heard from them that her renal function had again declined and she was readmitted to hospital. Thankfully, she informs us that she is now recovering well.

Discussion

This is a rare case of hepatic TB involvement with an unusual severity that is seldom described in literature, and without a background of immunosuppression. In fact, the disseminated TB led to immunosuppression of such an extent that our patient became pancytopenic.

Several articles and case reports in the literature have described interesting cases of isolated hepatic TB or in association with disseminated disease.^3–7 However, none has clearly described hepatic involvement to such extent in an immunocompetent patient like in this case. For instance, Abesekera et al,³ Kayar et al ⁴ and Sharma et al ⁵ described rare cases of isolated hepatic TB in immunocompetent patients. N’goran et al ⁶ also described a case of miliary hepatic TB and Çalışkan et al ⁷ described a case of multiple hypoechoic lesions in a child; none of them was a severe macronodular type/abscesses involving the liver. Hepatic TB is very rare comprising only 1% of TB infections.²

Several publications have described TB reactivation as a consequence of immune-deficiency and thus the re-emergence of TB globally has been linked to the HIV virus.^8–11 However, quite a few have highlighted that TB on its own can cause severe immunosuppression, as it did in our patient. It is thought that one mechanism by which this occurs includes the fact that TB can cause anorexia^{12 13} which might lead to an impaired immune system due to malnutrition. TB can also cause reduced leptin levels in patients with weight loss.¹⁴ Since leptin can promote cell mediated immunity, low levels may lead to more severe disease.

Cases of hepatic or disseminated TB presenting with sepsis or septic shock have been described in literature and our case was no different.^{15 16} Mycobacterial bacteraemia has been described as having high-mortality rates¹⁵ as can bacteraemia caused by other bacteria. There is also a possibility that patients with severe immune compromise due to disseminated TB infections can be predisposed to other bacterial infection and sepsis.¹⁷ Although TB blood cultures were negative, mycobacterial bacteraemia remains a possibility, given that the sensitivities of the tests for culturing Mycobacterium from the blood can be poor and dependent on the volume of blood taken.¹⁸

As in the literature, there is usually good response to treatment with first line TB medications.^{19 20} The findings of the MRI scan of her head, central pontine myelinolysis, could not be explained by her sodium levels which remained stable throughout this admission. Aggressive fluid resuscitation and malnutrition might have contributed.²¹ This was despite dietician input and adequate vitamin replacement even from early on. On the other hand, TB can cause demyelination.^{22 23} Unfortunately, we could not get cerebrospinal fluid analysis for this case and that would probably have given us a better understanding of the MRI head findings.

Her recovery has been complicated by her renal issues. Though the cause of the acute tubular necrosis (ATN) could be explained by the preceding and cumulative hypotensive episodes, it has been shown that TB medications like rifampicin can also cause ATN.²¹ Also, she was put on co-trimoxazole shortly prior to her renal function getting significantly worse. The renal team, however, were happy to restart the necessary suspended medications as they felt the ATN was hypotensive in origin; furthermore, it improved with fluid resuscitation despite restarting her TB medications. The exact cause of her hypotensive episodes remains unclear, but again, may be related to her poor nutritional status and significant burden of disease.

Learning points.

• To our knowledge, this is the most severe case of hepatic involvement of tuberculosis (TB) in the literature but still responded well to routine anti-tuberculous medications.

• Immunosuppression is certainly a risk factor for developing TB but TB alone can cause severe immunosuppression and predispose patients to other infections.

• TB may cause demyelinating disease of the central nervous system without meningeal involvement; the likely mechanism of this is malnutrition.

• The global burden of TB is such that it should remain high on the differential of any significantly unwell patient from countries with a high incidence of the disease.