Tuberculosis in Patients Referred for Liver Transplant: A Case Series and Review of Literature

Case 1

45 year old male patient presented to us with end stage liver disease caused by alcohol. He was referred to our center for evaluation of liver transplant. At presentation he had jaundice, refractory ascites, hyponatremia, muscle wasting, coagulopathy and hypoalbuminemia. He had previously undergone multiple time band ligations for esophageal varies over a period of 18 months. He was abstinent from alcohol for 18 months before presentation. He had bilirubin 8.3 mg/dl, ALT 28 IU/ml, AST 65 IU/ml, Serum Albumin 21 gram/dl, International ratio (INR) 2.89, Serum Creatinine 0.65 mg/dl, Serum Sodium 128 mg/dl. He did not have overt encephalopathy. He was evaluated for liver transplant. During his evaluation his chest x-ray showed bilateral mild pleural effusion. His diagnostic ascitic fluid tap showed protein 2 grams, SAAG > 1.1, 620 cells with lymphocytes 90 percent, ADA 2, MTB Genxpert® [3] negative and negative AFB culture. In view of high protein high SAAG ascites, he underwent a pleural tap which also showed negative AFB culture, negative MTB GenXpert®, negative ADA and protein of 2.6 grams. Our pulmonologist went ahead with bronchoscopy which was normal but broncheo-alveolar lavage was positive for AFB and MTB GenXpert was positive for Rifampicin sensitive mycobacterium tuberculosis. As per guidelines neither of Rifampicin and Isoniazid was given to the patient. He was started on ethambutol, moxifloxacin, clofazamine and cycloserine. His transplant was postponed. On starting anti tuberculosis treatment and naso gastric feeds, he started improving symptomatically. His pleural effusion on chest x-ray improved and even ascites decreased. Four weeks after starting treatment, he developed sudden onset breathlessness and was admitted with severe metabolic acidosis (pH - 6.8). He was resuscitated, intubated and started on renal replacement therapy. He was given carbapenem, azole and teicoplanin. He died 12 hours later. His blood culture reports came positive 3 days later showing extensive resistant Klebsiella Pneumonia showing sensitivity only to colistin/polymyxin B.

Case 2

A 37 year old male with alcohol related cirrhosis of liver diagnosed recently was transferred under our care from an outside hospital in view of worsening dyspnea for considering an urgent liver transplant. At presentation he had history of worsening jaundice for 4 weeks, abdominal distension and pedal edema for 2 weeks and breathlessness for 7 days. He had pallor, icterus, anasarca and absent air entry on left side with an intercostal drain (ICD) placed draining blood. He did not have encephalopathy and was passing good urine. He had undergone ICD insertion outside after requiring daily taps for pleural effusion outside 2 days before presentation. The initial pleural fluid reports were showing transudative picture. However, after he bled on ICD insertion it contained red blood cells > 50,000. His Hb was 6.9 gm/ dl, white cell count 12,000 per cumm, platelets 45,000 per cumm, Bilirubin 5.2 mg/dl, albumin 2.8 gram/dl, INR 2.24, Creatinine 1.45 mg/dl, Sodium 139 mEq/L and Fibrinogen 98 mg/dl. His chest x-ray showed massive left sided effusion with collapse. Local ultrasound revealed hemorrhagic content in the chest and septations. He was started on antibiotics, albumin, nasogastric feeds, multivitamins and coagulopathy correction based on thromboelastography. On third day the ICD got blocked and was removed. The pleural fluid sample was aspirated which showed AFB negative, ADA 18, MTB GenXpert® Negative, 580 cells with lymphocytes predominant, Protein 1.2 grams/L, Albumin 0.4 grams/L. His fluid was sent for AFP culture. His symptoms improved over a week and he was discharged on nasogastric feeds for high calorie high protein diet. He presented again after 8 days with a fistula arising from left side of chest, a site different from the ICD insertion site. The fistula on local ultrasound was communicating with pleural cavity. He denied consent for repeat ICD insertion or pleural tap and so a local fistula closure was done under general anesthesia and sample from fistula was sent for tuberculosis screening. This time it showed AFB negativity but MTB GenXpert® AFB positivity and so was started on Ethambutol and Moxifloxacin. Two months after starting his 2 drug ATT, his pleural effusion and fistula has completely resolved with only minimal pleural thickening remaining on chest x-ray. His ascites and pedal edema has disappeared and liver tests have normalized. His

liver has compensated (Bilirubin 1.2 mg/dl, albumin 3.9 g/ dl, sodium 140 mEq/L, INR 1.01, no ascites on ultrasound) and was started on Rifampicin initially and isoniazid later on. His pleural fluid AFB cultures have turned out to be negative. His transplant evaluation has been on hold and is not on transplant list.

Case 3

A 51 year old male diagnosed to have alcohol related liver disease for 3 years had started worsening over the last 3 months and was referred for evaluation for liver transplant. He had been symptom free and abstinent from the time of diagnosis. He had portal hypertension (HT) manifesting as varies and was started on beta blockers as primary prophylaxis. Over previous 3 months he had development of refractory ascites, edema and significant loss of muscle mass. He had developed loss of appetite but denied any fever, jaundice, decreased urine output or an episode suggestive of encephalopathy. At presentation his reports showed hemoglobin 11 g/dl, platelets 120 per cumm, bilirubin 2.4 mg/dl (conjugated 0.1 mg/dl), albumin 2.6 grams/dl, INR 1.91, creatinine 0.7 mg/dl and sodium 136mEq/L. His ascites was high SAAG and negative for tuberculosis workup. He had isolated minimal right sided effusion and had undergone a tap outside 15 days before presentation which corroborative of hepatic hydrothorax. He was started on liver transplant evaluation and pulmonologist advised a repeat tap of pleural effusion before clearance. He underwent a repeat tap of right sided pleural effusion which showed cell count of 879 with lymphocyte predominant, protein 4 grams, Albumin 2.3 grams, ADA 81, AFB negative and MTB GenXpert® negative. Considering this picture, he was started on anti-tuberculosis therapy (ATT) - Rifampicin, Ethambutol and Moxifloxacin. Within 4 weeks of starting treatment, his ascites improved and edema disappeared. His appetite improved and even liver tests normalized. Isoniazid was added 8 weeks after starting ATT and has completed 12 months of therapy. He is asymptomatic currently. He is not evaluated for transplant.

Case 4

A 48 year old male patient with diagnosis of alcohol related liver disease was referred to us for non-resolving right sided hepatic hydrothorax with refractory ascites. At presentation he was breathless. He had icterus, pallor, anasarca, tense ascites and absent air entry on right side with a respiratory rate of 30 per minute. His last tap was 48 hours before. His previous ascetic and pleural tap reports were negative. His chest x-ray showed hemithorax right sided opacity. With coagulopathy correction based on thromboelastography, he underwent an indwelling pleural catheter insertion and pleural fluid reports were repeated. An HRCT thorax was carried out which showed consolidation in left basal area. The pulmonologist advised a bronchoscopy which revealed MTB GenXpert positive broncheo-alveolar lavage. The patient was started on Ethambutol and Moxifloxacin. He was re admitted 7 days later with decreased urine output. He was diagnosed with urinary tract infection with sepsis. He developed progressive multi organ dysfunction over next 4 days and died. His culture showed Klebsiella Pneumonia sensitive only to polymyxin and colistin.

Case 5

A 61 year old male with ethanol related cirrhosis was referred to our center for requiring repeated ascitic taps. He had history of developing encephalopathy on multiple occasions but had no history of hematemesis. At admission he was sarcopenic with ascites out of proportion to edema and had severe muscle wasting. His reports showed bilirubin 3.1 mg/dl, INR 2.14, serum creatinine 1.28, serum albumin 3.2 and serum sodium 140mEq/L. On admission his ascetic tap was sent which was high SAAG high protein ascites with lymphocyte predominant cellularity. It was negative for MTB GenXpert, ADA and AFB staining. However, the ascetic fluid culture grew MTB at 20 days. The patient was started on Ethambutol and Levofloxacin. The Table 1 mentions the summary of all 5 patients.

Prevalence

Tuberculosis and cirrhosis both are widely prevalent in India and other south Asian countries. Though data of tuberculosis in transplant evaluated patients is scanty, there is data on tuberculosis in cirrhosis in general. In a Taiwanese nationwide longitudinal study including 41,076 cirrhotic patients and 204,244 non-cirrhotic controls from 1998 through 2007, a total of 957 (2.32%) cirrhotic patients developed TB, yielding a rate that was significantly higher than that of the 955 (0.46%) non-cirrhotic patients (P < 0.001) [4]. In a retrospective database published from Nepal [5], tuberculosis prevalence in cirrhosis was 9% while it was 7.38% in a study from western India [6]. In a Danish study cohort of 22675 patients with liver cirrhosis, they identified 151 cases of tuberculosis from 1977 to 1993 with an incidence of 168.6 per 100000 person-years and 30-day case-fatality rate was 27.3% [7].

Prognosis

In our case series 2 patients died with complications of cirrhosis leading to a case fatality rate of 40 percent. In a study which looked at mortality of patients with cirrhosis and tuberculosis, 0-day, 90-day, 1-year and 3-year mortalities were 10.1%, 24.2%, 43.1%, and 63% in the TB group and the hazard ratios (HR) in cirrhotic patients with TB for 30-day, 30 to 90-day, 90-day to 1-year, and 1 to 3-year mortalities were 1.33, 1.91, 1.46, and 1.10 compared to the non-TB group respectively [8]. This study however does not mention about the severity of underlying disease. In a Japanese study, 18/44 patients with cirrhosis and tuberculosis died with liver disease. The severity of cirrhosis was not mentioned even in this study [9].

Risk factors

In our case series all the patients suffered from ethanol related cirrhosis of liver. In previous study, alcoholism and hepatitis C infection were associated with significantly higher TB risk with adjust hazard ratios of 2.184. Alcoholism was also the most common etiology of liver disease (60%) in the study from western India [6]. In a study from Northern India, the etiologies of cirrhosis of liver were: alcohol in 49% cases; the others being post viral infection (23%), cryptogenic (16%) and NASH (12%) [10]. In a Chinese case control study, Hepatitis B was the underlying etiology in 44% patients while ethanol was responsible for 22% cases [11].

Site

In our series, 2/5 (40%) had pulmonary tuberculosis. Even in Nepalese study, extra pulmonary tuberculosis was more common (67%) than pulmonary (33%). In extra pulmonary tuberculosis, distribution was as pleural effusion, abdominal, intestinal, peritoneal, and spine [5]. In the study by Amarapurkar, et al. [6], 40% had pulmonary tuberculosis while other sites involved were pleural in 28%, peritoneal 18%, cervical node 6%, liver 4%, intestine 2% and spine was 2%. As per the Taiwanese database4, 80.0% had pulmonary TB, 0.9% had military TB, 16.6% had extrapulmonary TB and 2.5% had both pulmonary and extrapulmonary TB. In a study from Northern India [10], site of tuberculosis was pulmonary (37%), pleural effusion (16%), peritoneal (29%), lymph nodes (4%), liver (1.5%), intestines (4%), bone (4%), brain (1.5%) and disseminated (3%).primary carried out to look at outcomes of drug induced liver injury with ATT in cirrhosis included 67 patients, of which 16 were belonging to CTP C category.

Diagnosis

In all our patients, tuberculosis diagnosis was not straight forward. The tests were negative at the first instance and only on repeated analysis they turned out to be positive. This is very important as a single negative sample should not exclude tuberculosis in these patients. The clinical presentations of low grade fever, muscle loss, loss of appetite are routinely seen in patients with end stage liver disease and may be symptoms of cirrhosis or tuberculosis [10]. So the diagnosis is mainly based on investigations and high clinical suspicion. Tuberculosis peritonitis can simulate ascites from liver disease or spontaneous bacterial peritonitis. The diagnosis is difficult in these patients because the ascitic fluid may not be of the exudative type as a result of the low albumin level in serum, and lymphocytes do not predominate in all cases. The time required to achieve a correct diagnosis was significantly longer in cirrhotic than in non-cirrhotic patients [12]. In a study to differentiate spontaneous bacterial peritonitis (SBP) from tuberculosis peritonitis (TP), TP in cirrhotic patients was more frequently associated with extra-peritoneal tuberculosis (53.3%), an insidious onset (≥2 weeks in 60%) and advanced liver disease (CTP B in 80%) while, TP showed lower white blood cell count in ascites, higher protein concentration and higher adenosine deaminase activity in ascites [13]. The utility of ADA from peritoneal fluid in cirrhosis has been questioned although. In a US study, the overall sensitivity of the ADA determination in diagnosing tuberculosis peritonitis was only 58.8%, and the specificity was 95.4% and was only 30% sensitive in detecting TP in the setting of cirrhosis. In addition, malignancy-related ascites (13%) and bacterial peritonitis specimens (5.8%) yielded false-positive results [14]. In a Taiwanese study, a 27 U/L as the cut-off value of ADA on the contrary, showed sensitivity and specificity of 100% and 93.3%, respectively to differentiate TP in cirrhosis [15] which is similar to patients without cirrhosis [16, 17]. Limited data are available to support the use of GeneXpert MTB in the diagnosis of abdominal tuberculosis. In our series 3 of 5 patients had GenXpert MTB positive (one in tissue culture from cutaneous fistula and two from Bronchoalveolar Lavage). None of the patients had peritoneal or pleural fluid report positive. Data are also not favoring use of MTB GenXpert in these patients with a study from Pakistan having 0% specificity for the test [18]. The studies from India have found a sensitivity of 8 to 32% for intestinal tuberculosis [19, 20, 21]. A Cochrane review also found that sensitivity of the test varied as per tissue sample, the highest sensitivity in urine, bone and joint while for pleural fluid it was just 31% [22]. Mantoux or interferon gamma release assays (IGRA) were not carried out in our patients. The interpretation of Mantoux test in cirrhosis is difficult. Besides, almost all Indians are vaccinated with tuberculin. Published Indian literature also states that there is no role of Mantoux testing before liver transplant while the utility of IGRA is not established as compared to west [23, 24, 25]. The diagnosis in the study from western India6 was based on histopathology or fluid analysis in 60% cases. Remaining 40% were considered tuberculosis based on clinical and radiological parameters. In the study from Northern India, Mantoux was positive in only 43% patients with CTP C cirrhosis compared to more than 70% with CTP A and CTP B cirrhosis. 59/67 patients had TB confirmed histopathologically [10]. In our series as well, 80% diagnosis was made from tissue sample and only one patient was diagnosed on biochemical and radiological parameters.

Treatment

We treated all patients with Ethambutol and Quinolone. Aminoglycosides were not given considering the risk of acute kidney injury. One patient was given cycloserine. Isoniazid was added once the clinical parameters improved to CTP B state. Isoniazid was initially given at dose of 150 mg and then increased to 300 mg after 15 days. Weekly monitoring was carried out. Rifampicin was added later in one patient in step wise manner. Dhiman, et al. [26] recommends the following management guidelines for tuberculosis in patients of cirrhosis - ATT should include no more than 2 hepatotoxic drugs (Rifampicin (R) and Isoniazid

(H)) in patients with CLD or liver cirrhosis and stable liver function (CTP ≤7), only a single hepatotoxic drug (R or H) in those with advanced liver dysfunction (CTP 8-10) and no hepatotoxic drugs with very advanced liver dysfunction (CTP ≥11). The duration of therapy is for 18-24 months. However, as per study by Sharma et al. they have used R+H for first two months in 2 patients with CTP C cirrhosis along with Ethambutol (E) and Quinolones (Q). They used Rifampicin without isoniazid in 7 patients while used Isoniazid without rifampicin in 2 patients. Of 16 patients treated with CTP C cirrhosis, the mortality was 25% and both patients with R+H use developed liver injury. Drug induced toxicity was seen in 35% patients in patients started with RHEQ, HEQ and REQ. Saigal, et al, in their study comparing Pyrizinamide (Z) with R+H in patients of cirrhosis have shown that R+H is more hepatotoxic than a combination with Q+Z [27]. The previous Indian study published treated 14 patients with E+Q with or without aminoglycosides and clarithromycin6. However, the selective outcome of these patients in CTP C patients was not mentioned, although four patients died of liver related complications from an entire cohort of 50 patients. In a Chinese study [11], only 57% of total patients completed full course of treatment and 7/36 patients had CTP C cirrhosis. Six of patients with CTP C cirrhosis died due to progression of liver disease. None of them received R or H. In an interesting case series published of 3 patients, all 3 patients were given Isoniazid based therapy, one patient getting transplanted 1 month after starting treatment while one patient was started anti tuberculosis treatment (ATT) after transplant. One patient deteriorated 6 months after full dose of ATT [28].

Liver Transplant

In our series, none of the patient was transplanted after the diagnosis. The two patients died within short period of starting therapy. The remaining patients improved with therapy so much as to take them off the transplant list. For patients treated with R +H, it is usually considered safe to transplant after the intensive therapy. However, when patient is just taking E+Q, the timing for transplant is not known. At the same time, delaying transplants in very sick cirrhotic patients can lead to poor outcome. Further research is needed in optimizing the timing of these patients. Although, in a published Chinese case series [11], the patient was transplanted and then ATT was started, from our series we can see that some patients may have dramatic improvement and may no longer require the transplant. The comparative evaluation of ATT by different groups is shown in Table 2. The management of Tuberculosis as proposed by our team is mentioned in Figure 1.

Click to enlarge