Original Article
JOURNAL OF CLINICAL AND EXPERIMENTAL HEPATOLOGY
Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis Praveen Sharma, Pankaj Tyagi, Vikas Singla, Naresh Bansal, Ashish Kumar, Anil Arora
Introduction: There is paucity of data on tuberculosis and antituberculous therapy (ATT) induced hepatotoxicity in patients with chronic liver disease. Aim: To study demographic, clinical characteristics of tuberculosis, pattern of drug induced liver injury and treatment responses to ATT in patients with liver cirrhosis. Material and method: All cases of liver cirrhosis diagnosed with tuberculosis (TB) between January 2010 and June 2013 were enrolled. Drug induced liver injury (DILI) was defined as follows (1) an aspartate aminotransferase (AST)/alanine aminotransferase (ALT) value exceeding 3 times the normal upper limit if the baseline level was normal (2 mg/dl from baseline. Results: Sixty seven patients had confirmed TB with underlying cirrhosis and formed the study group. The mean age was 52 ± 12 years and M:F ratio was 57:10. Mean Child Turcotte Pugh (CTP) score 8.5 ± 1.5 (CTP A:B:C:7:44:16). The sites of TB included: pulmonary (25, 37%); pleural effusion (10, 16%) peritoneal (19, 29%); chest lymph nodes (3, 4%); liver (1, 1.5%); intestines (3, 4%), vertebra (3, 4%), brain (1, 1.5%) and disseminated (2, 3%). Thus, extrapulmonary TB was more common in the cirrhotic patients as compared to pulmonary TB. Patients with Child's status A (n = 7) received 4 drugs (R: rifampicin, H: Isoniazid, E: ethambutol, Z: pyrizinamide) and could tolerate well even during follow up without any drug induced toxicity. In rest of patients commonest regimen followed was combination of drugs (RHEO, n = 32) followed by RHE (n = 11). DILI occurred in 35% started with either RHEO, HEO and REO. Median time of onset of DILI was 12 days (4–34) days. There was no DILI related death during hospital stay or follow up. Conclusions: Extrapulmonary TB is common in patients with cirrhosis and DILI is common in Child B and C with combination of rifampicin and isoniazid regimen. ( J CLIN EXP HEPATOL 2015;-:1–6)
T
uberculosis is major health problem in developing countries. Reported prevalence of all forms of tuberculosis in India is 5/1000.1–3 Standard antitubercular therapy (ATT) which includes rifampicin (R), isoniazid (H) and pyrizinamide (P), have hepatotoxicity as a major side-effect.4,5 The mechanism of hepatotoxicity and the risk factors that predispose to it are still not clearly defined. Liver cirrhosis is considered as an acquired immune deficiency state due to malnutrition or due to advanced liver problem. Drug induced liver injury (DILI) is difficult to diagnose in patients with chronic liver disease because of elevated enzymes and bilirubin at baseline and fluctuations in enzymes level caused by underlying liver diseases may
Keywords: cirrhosis, tuberculosis, treatment Received: 3.3.2014; Accepted: 14.1.2015; Available online: xxx Address for correspondence: Anil Arora, Department of Hepatology and Gastroenterology, Sir Ganga Ram Hospital, New Delhi, India. E-mail: [email protected] Abbreviation: AST/ALT: Aspartate aminotransferase/alanine aminotransferase; ATT: Antituberculous therapy; CTP: Child Turcotte Pugh; DILI: Drug induced liver injury; E: ethambutol; H: Isoniazid; R: Rifampicin; TB: Tuberculosis; Z: pyrizinamide http://dx.doi.org/10.1016/j.jceh.2015.01.003 © 2015, INASL
confound the diagnosis. As these drugs are metabolized in the liver, there is a theoretical risk of increased hepatotoxicity in patients with underlying chronic liver disease (CLD).6,7 ATT constitutes an important acute precipitating event in decompensation of underlying known or unknown liver disease in India.8 There is paucity of data in defining hepatotoxicity in patients with cirrhosis and clear guidelines do not exist regarding the monitoring of such patients, modification of the antitubercular regimen and mode of reintroduction of individual antitubercular therapy. Little is known about the characteristics of TB in patients with liver cirrhosis. The relationship between TB and liver cirrhosis in terms of clinical characteristics, treatment responses and adverse effects is very limited.9–11 The aim of this study was to elucidate demographic, clinical characteristics of tuberculosis, pattern of drug induced liver injury and treatment responses to ATT in patients with liver cirrhosis.
MATERIALS AND METHODS This study is a retrospective analysis of prospective collective data of all cases of liver cirrhosis diagnosed with tuberculosis between January 2010 and June 2013. Patients were
Journal of Clinical and Experimental Hepatology | - 2015 | Vol. - | No. - | 1–6
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
Tuberculosis & Liver Cirrhosis
Department of Gastroenterology, Sir Ganga Ram Hospital, New Delhi, India
PATIENTS WITH LIVER CIRRHOSIS
Tuberculosis & Liver Cirrhosis
either admitted in hospital or were seen in outpatient department. All the patients of cirrhosis were not screened for tuberculosis. Patients with cirrhosis were screened for tuberculosis only if patients presented with following parameters: fever, cough for >2 weeks, hemoptysis, unexplained weight loss, increasing ascites not responding to diuretics, unexplained bowel symptoms (diarrhea, constipation, or subacute intestinal obstruction), radiological lesions suggestive of tuberculosis and past or family history of tuberculosis. Patients were diagnosed to have cirrhosis if they had liver histology suggestive of cirrhosis or based on clinical, biochemical and radiological imaging findings (either on ultrasound, transient elastography and computed scan). Medical records, bacteriological examinations and radiographic findings were recorded for clinical characteristics, responses to treatment and adverse effects of anti-TB treatment. If suspected, investigations including hemogram, erythrocyte sedimentation rate (ESR), Mantoux test, sputum for acid-fast bacillus (AFB) repeated thrice, chest X-ray, ultrasonography, Endoscopic ultrasonography with fine needle aspiration of lymph node, CT scan of thorax and abdomen, and barium studies were done. Mantoux test (tuberculin skin test) was considered positive if there was $5 mm of induration at 48–72 h. A detailed ascitic fluid and pleural fluid examination was done including cytology, protein and albumin content, serum ascites albumin gradient, adenosine deaminase levels, acid fast bacillus culture and PCR for Mycobacterium tuberculosis. Aspiration cytology and tissue biopsy including bone marrow and liver biopsy were done if deemed necessary for the diagnosis of fever of an unknown origin. Tuberculosis was diagnosed based on a good clinical response to chemotherapy and one or more of the following: (i) histological evidence of caseating granulomas; (ii) sputum positivity for AFB; (iii) growth of Mycobacterium tuberculosis on culture; and (iv) positive PCR for Mycobacterium tuberculosis in tissues. The diagnosis of peritoneal and pleural tuberculosis was established if three of the following four criteria were fulfilled: (i) raised ascitic fluid/pleural cell count with lymphocyte predominance; (ii) raised ascitic fluid/pleural fluid albumin (>2.5 g/dL); (iii) raised adenosine deaminase (>33 U/L)12 in ascites fluid and 40 Ul/L in pleural fluid13 and (iv) positive PCR for M. tuberculosis. Treatment regimen was decided by the treating physician and it was his decision to start the antitubercular treatment in combination of his choice considering the condition of the patient. Patients were followed up every 15 days in the initial 2 months after the discharge if admitted and then at month interval. In case of any problem patients were asked to come to the hospital in between the planned visits. All these patients were seen by the authors of these studies and patients who were diagnosed to have cirrhosis and tuberculosis were regularly followed up and their data was collected. Drug induced liver injury 2
SHARMA ET AL
(DILI) was defined as follows (1) an aspartate aminotransferase (AST)/alanine aminotransferase (ALT) value exceeding 3 times the normal upper limit if the baseline level was normal (2 mg/dl in absence of no other alternative identifiable reason for the above changes.14
STATISTICAL ANALYSIS Statistical comparisons were performed using spss (SPSS for Windows, version 12.0; SPSS; Chicago, IL, USA). Values were expressed in median with range Chi-square tests were used to test differences in categorical variables and t-tests were used for continuous variables. Statistical significance was assumed at P < 0.05.
RESULTS Sixty seven (n = 67) patients were diagnosed to had TB with underlying cirrhosis and formed the study group during this study interval. Baseline characteristics of sixty seven patients with TB and liver cirrhosis were shown in Table 1. The median age of patients were 52 (24–85) years and male-to female ratio was 57:10. Median CTP score of patients were 8.5 (5–12) (CTP A:B:C:7:44:16) (Table 1). The etiology of cirrhosis of liver were: alcohol (33, 49%); postviral infection—hepatitis B and C (15, 23%), cryptogenic (11, 16%) and NASH related (8, 12%). Diagnosis of NASH related cirrhosis was labeled in 8 patients based on histological diagnosis of cirrhosis. Four of these patients had liver biopsy done before which showed F3 Fibrosis, steatosis and steatohepatitis and four patient had radiological evidence of steatosis in past (n = 4). Hence diagnosis of NASH related cirrhosis were labeled in these
Table 1 Baseline Characteristics of Patients with Cirrhosis and Tuberculosis (n = 67). Parameter
Median (range) (n = 67)
Age (yr)
52 12
M:F
57:10
CTP score
8.5 1.5
Total bilirubin mg/dl
4.8 6.1
ALT IU/L
61 42
AST IU/L
76 54
International normalized ratio
1.8 0.7
Child's A:B:C
7:44:16
Etiology of cirrhosis ALD:Viral:Cryptogenic:NASH
33:15:11:8
AST: aspartate aminotransferase ALT: alanine aminotransferase CTP: Child-Turcotte-Pugh Score. © 2015, INASL
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
JOURNAL OF CLINICAL AND EXPERIMENTAL HEPATOLOGY
Table 2 Clinical and Radiographic Characteristics of Patients with TB with Cirrhosis.
Table 3 Diagnostic Modalities for Tuberculosis. Pulmonary Tuberculosis (n = 25)
Extrapulmonary TB Pleural effusion (n = 10) Nodal (N = 3)
Chest (n = 3) Peritoneal (n = 19)
Bone (n = 3)
Meningeal (n = 1)
Disseminated (n = 2)
Liver (n = 1) Intestinal (n = 3)
Number (%) Clinical symptoms Fever
45 (67%)
Unexplained weight loss
23 (34%)
Sputum acid fast bacilli positive (n = 16) Fibrocavitatory lesion on chest X-ray and CT scan of thorax (n = 12) Bronchoalveolar lavage with AFB +ve (n = 5) Raised ADA (n = 10) in pleural fluid (detected in CXR with lymphocytic predominance and high protein Chest CT: shows mediastinal lymph node and EUS FNAC showed granuloma with caseation (n = 3) High ADA in ascitic fluid (n = 12) Peritoneal biopsy (n = 2) EUS FNAC intra-abdominal lymph nodes along with high SAAG septated ascites (granuloma with AFB, n = 5) CT scan and MRI scan showing vertebra destruction (n = 3) Positive Mx and high ESR (n = 3) Bone biopsy showing granuloma (n = 1). Cerebrospinal fluid showed lymphocytic predominance with high protein findings, PCR positive for tuberculosis, CT brain suggestive of meningitis Chest X-ray (miliary shadow n = 2) CT abdomen and thorax showed multiple lymphadenopathy (n = 1) EUS guided FNAC from abdominal lymph node (n = 1) and omental biopsy (n = 1) showed granulomatous lesion. Liver biopsy which showed granuloma Double balloon enteroscopy with biopsy suggestive of granuloma (n = 1) CT: abdomen suggestive of stricture and necrotic lymph node (n = 2)
Loss of appetite
55 (82%)
Antitubercular Therapy and Side Effects
Increasing ascites not responding to diuretics
17 (25%)
Patients with Child's Pugh status A (n = 7) received 4 drugs (rifampicin, Isoniazid, Ethambutol and pyrizinamide RHEZ) for the initial two months and then two drugs rifampicin and isoniazid for 7 month without any drug induced toxicity. All the patients had good response and ATT was stopped after nine month of duration and there was no mortality in this group. In Child's B and C, patients were started on combination of drug treatment which included [(Rifampicin R, isoniazid H, ethambutol E and ofloxacin O (RHEO)]. The commonest combination of drugs used was [RHEO, n = 32). This drug combination was started by the treating physician based on patients clinical and biochemical parameters (Table 4). Pyrizinamide was not used instead quinolones was preferred in these patients with advanced liver disease (Child's B and C). Drug induced toxicity was seen in 35% patients in patients started with either RHEO, HEO and REO (Table
Cough Pain abdomen
8 (12%) 14 (21%)
Site of involvement Pulmonary tuberculosis
25 (37%)
Extrapulmonary TB
42 (63%)
Pleural effusion
10 (16%)
Nodal (chest)
3 (4%)
Intestinal
3 (4%)
Peritoneal
19 (29%)
Bone
3 (4%)
Meningeal
1 (1.5%)
Disseminated
2 (3%)
Liver
1 (1.5%)
Journal of Clinical and Experimental Hepatology | - 2015 | Vol. - | No. - | 1–6
3
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
Tuberculosis & Liver Cirrhosis
patients. Cryptogenic cirrhosis was diagnosed based on histological evidence of cirrhosis with no obvious etiology based on histological report of liver biopsy in these patients and negative work up for viral markers (hepatitis B and C), autoimmune markers, Wilson and hemochromatosis disease. None of the patients had HIV co-infection. The TB sites included: pulmonary (25, 37%); pleural effusion (10, 16%) peritoneal (19, 29%); chest lymph nodes (3, 4%); liver (1, 1.5%); intestines (4, 4%); and bone (3, 4%), brain (1, 1.5%) and disseminated (2, 3%). In disseminated tuberculosis histological diagnosis was made based on omental biopsy (n = 1) and EUS guided lymph node FNAC in one case. Thus, extrapulmonary TB was more common in the cirrhotic patients than pulmonary TB (P < 0.05). Most common presenting symptom was loss of appetite (82%) and fever (67%), weight loss (34%) and increasing ascites (25%) (Table 2). Mantoux test (tuberculin skin test) was positive in 43 (64%, Child's status: A:B:C:71%:70%:43%) patients. In 59 patients, TB was confirmed either histopathologically (either biopsy or fine needle aspiration) AFB smear or by fluid analysis. In the remaining 8 patients (4 with pulmonary TB, 2 with intestinal, 2 with bone), indirect indicators such as a history of fever and weight loss, high erythrocyte sedimentation rate (ESR), a positive Mantoux test, typical X-ray and CT thorax and abdomen appearance were taken into account before starting treatment (Tables 2 and 3).
PATIENTS WITH LIVER CIRRHOSIS
SHARMA ET AL
Table 4 Regimens of Anti-TB Treatment and Toxicity. Parameters
Baseline regimen (N)
Follow up regimen after 2 months (N) till completion of therapy
Child's A (n = 7)
RHEZ(7)
RH(7)
Child's B (n = 44)
RHEO(32) + RHE(n = 9) + HEO(3)
RH(28) + HEO(8) + EO(2)
Child's C (n = 16)
REO(7) + RHE(2) + HEO(2) + EO(5)
REO(4) + HEO(4) + EO(4)
Fluroquinolones based regimen
51/67 (76%)
*Rifaximin and isoniazid could be reintroduced in 6 and 4 patients respectively without side effects: 4 patients lost to follow up in Child's B; Ò6 patients died during hospital stay (Child's B, n = 2, Child's C n = 4). R: rifampicin, H: isoniazid, E: ethambutol, Z: pyrizinamide, O: ofloxacin.
Tuberculosis & Liver Cirrhosis
3). Median time of onset of DILI was 12 days (4–34) days and median follow up period of these patients was 13 months (13 days–26 month). Of 43 patients with (R + H) regimen in Child's B and C it could be continued for more than 2 months in 22 patients due to drug related toxicity in these patients. Reintroduction of rifampicin was possible in 6 patients and INH in 4 patients respectively after drug induced toxicity subsided. In these patients ethambutol, ofloxacin with either rifampicin or isoniazid were continued for rest of their treatment. All patients in Child's Pugh status B and C continued quinolones and ethambutol (100%) till end of therapy. Injection streptomycin was used in 3 patients in those who developed drug induced toxicity all in Child B status for 2 months till drug induced toxicity was over. RH combination was not used in Child C patients except in two patients in the first two months. The duration of treatment ranged from 9 to 18 months. Eighteen month treatment was given in patients with disseminated TB, meningeal TB and bone TB patients. Peritoneal TB, tubercular pleural effusion and intestinal TB was treated with 12 months of treatment and all responded clinically and pleural effusion disappeared on follow up. Ascites was controlled and dose of diuretics decreased. Abdominal pain disappeared and patient gained during follow up in intestinal TB. Fever subsided and weight gain was seen in patient with liver tuberculosis on follow up, repeat biopsy was not done in this patient. Pulmonary TB was treated for 9 months and all 19 patients who completed the therapy responded. Four patients with pulmonary TB were lost to follow up after hospital discharge all in Child's B and two patients (Child's C) died during hospitalization. No patient developed drug-induced liver failure. In hospital mortality was 6 (9%) and was due to decompensated liver disease (Child's B, n = 2 and Child's C n = 4) and these are the patients who could not complete the tubercular therapy. Of these 6 patients four patients died due to severe upper gastrointestinal bleed and two developed pneumonia with sepsis and renal failure. There was no DILI induced mortality in our study either during hospitalization or during follows up of these patients. Four patients with pulmonary tuberculosis in Child's B were lost to follow up. 4
DISCUSSION Pulmonary TB was noted in 37% of cases and extrapulmonary TB in 63%. Drug induced liver injury was also higher (35%) in this cohort of patients. RHEO is the commonest combination used in Child B cirrhosis patients, whereas ethambutol and ofloxacin based combination is commonly used in Child C patients. The major antitubercular drugs: isoniazid, rifampicin and pyrazinamide are all metabolized in the liver. The reported incidence of antituberculosis drug-induced hepatotoxicity varies between 2% and 28%.Pooled analysis of Indian trials in non cirrhotic patients found 11.5% incidence of drug induced liver injury whereas a meta analysis of western studies found a pooled risk of 4.28% in patients with no cirrhosis.4,5 Common risk factors are advanced age, female sex, slow acetylator status, malnutrition, HIV and pre-existent liver disease.15 However predictive model for identification of DILI is not available. Patients with cirrhosis have been shown to have a longer half-life of isoniazid and rifampicin, thus leading to their accumulation in the plasma. Isoniazid, pyrazinamide and rifampicin have hepatotoxic potential, and can lead to such reactions during antituberculosis chemotherapy. Most of the hepatotoxic reactions are dose-related; some are, however, caused by drug hypersensitivity.16,17 In a meta-analysis, drug-induced hepatic dysfunction usually occurs within the initial few weeks of the intensive phase of antituberculosis chemotherapy. Vigilant clinical (including patient education on symptoms of hepatitis) and biochemical monitoring are mandatory to improve the outcomes of patients with drug-induced hepatotoxicity during antituberculosis chemotherapy. Some fluoroquinolones like ofloxacin/levofloxacin may have a role in constituting non-hepatotoxic drug regimens for management of tuberculosis (TB) in the presence of hepatic dysfunction.18 Our results support these observations, as 35% of the patients treated with combination treatment developed hepatotoxicity and majority of these patients are from Child B and Child C as these patients have low albumin and poor nutritional status. None of patient from Child © 2015, INASL
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
A developed DILI but these patient constitute only 10% of our cohort. Similarly only 2 patients received RH combination of 16 from Child C and they developed DILI. Thus, a combination of isoniazid and rifampicin in patients with underlying CLD especially in Child's B and C should possibly be avoided. In a similar study by Park et al14 18 (17%) experienced DILI: 11 (24%) among 46 patients with chronic hepatitis and 7 (15%) among 46 patients with compensated liver cirrhosis. We could not predict factors associated with DILI in our study as all patients did not receive the same drugs so predicting factors for DILI would have bias in our study. None of the patients had any sideeffects from fluoroquinolones and ethambutol during the entire course of therapy and follow up which could be a better alternative in patients with advanced liver disease and this is in accordance with Saigal et al10 who concluded that in patients with tuberculosis who have underlying CLD: (i) an ofloxacin-based antitubercular regimen without rifampicin is as effective as a rifampicin-based regimen; and a combination of isoniazid with rifampicin is more hepatotoxic than a combination with ofloxacin and pyrizinamide. Median duration of development of DILI in our study was 12 days with onset as early as day 4 of starting ATT. Our practice in patients with cirrhosis and tuberculosis is to monitor liver function tests every third day till the patient is admitted and then weekly for initial two month of therapy and every fifteen days interval thereafter. Extrapulmonary tuberculosis was seen in 63% of patients in our study. Similar were the results of Baijal et al9 60%, Saigal et al10 65% and Cho et al11 in whom 31% had extrapulmonary TB of 36 patients enrolled. Little is known about the immunopathogenesis of TB in such clinical conditions. Diagnosis of extrapulmonary TB also needs meticulous investigation along with clinical data. With the availability of EUS guided FNAC diagnosis of nodal TB has increased and all measures should be taken to make a tissue diagnosis in patients with cirrhosis, as unnecessary treatment can lead to further deterioration of liver functions in this group of patients. Even with best measures empirical ATT was started in 8 patients in this study based on clinical and biochemical parameters. We found Mantoux test positive in 64% of patients in this study. There is no recommendation of testing Mantoux test in India to screen for tuberculosis in patients with cirrhosis. However study by Çelikbilek M et al19 has shown that in patients with cirrhosis who were waiting for liver transplant and no active tuberculosis, the tuberculin skin test findings were significantly higher in end-stage liver disease caused by viruses than those with a non-viral etiology (P < 0.05). The mean induration of the tuberculin skin test in Child B patients was found to be 7.15 6.4 mm, while in Child C patients it was 12.64 7.5 mm. In Child C patients, the tuberculin skin test scores were significantly higher than in Child B pa-
tients (P < 0.05). So Mantoux test gets affected by status of liver disease and etiology of liver disease. We did it all of our patients and recommend it for all patients as it contributes to clinical decision making along with symptoms if other imaging were not contributory for the diagnosis of tuberculosis. Multi drug induced hepatotoxicity has been well defined in patients with normal liver function tests and no cirrhosis, however its definition in patients who already had impaired liver function tests has not well been documented. Saigal et al10 take arbitrary values for defining DILI as ALT/AST level increased to more than fivefold the baseline level, or to more than 400 IU/L, or if the bilirubin increased by 2.5 mg/dL after exclusion of superimposed acute hepatitis and its definition varies with other published literature. Future studies are needed on this issue to define DILI in patients who are already jaundices and had elevated AST and ALT levels at baseline. Similarly there is no consensus on treatment duration of extrapulmonary TB in patients with cirrhosis and the duration varies from 6 to 18 months. In our study also duration varies in extrapulmonary tuberculosis from 9 to 18months. So future studies should address this issue. However different drugs, different Child Status of patients and different site of involvement might preclude studies to answer this very important clinical problem in day to day practice of Hepatologist. Saito et al6 had shown in a cohort of 44 patients that tuberculous patients with liver cirrhosis are characterized with higher mortality rate and higher frequency of adverse effects of antituberculous chemotherapy. Multi-drug combination regimen could be tolerable under adequate surveillance of side effects even in the situation of preexisting liver dysfunction. We also found 9% in hospital mortality. The main drawbacks of our study was that as patients with cirrhosis were seen by all authors in their respective outpatients department and only few of them were admitted. We did not have record of patients who were seen in outpatient department and had cirrhosis and but not admitted. Hence exact frequency of tuberculosis in patients with cirrhosis cannot be determined in such situation. We did not use fixed combination of anti tubercular drugs as per the Child's status as this study was a collection of data of patients seen by many authors in a real time practice and their follow up till they continued the ATT. A fixed combination of antitubercular therapy were not used as patient had different Child's status and had different course of disease during follow up. However it would be useful to conduct more studies in patients with cirrhosis and tuberculosis and use a standardized treatment in them according to their Child's status and know about its safety so that a particular regimen could be used in future trials according to the response. To conclude extrapulmonary tuberculosis is common in patient with cirrhosis and drug toxicity due to
Journal of Clinical and Experimental Hepatology | - 2015 | Vol. - | No. - | 1–6
5
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
Tuberculosis & Liver Cirrhosis
JOURNAL OF CLINICAL AND EXPERIMENTAL HEPATOLOGY
PATIENTS WITH LIVER CIRRHOSIS
combination of rifampicin and isoniazid is more common in patients with decompensated cirrhosis. Quinolone based patient tailored extended therapy is an alternative in patient with advanced cirrhosis.
CONFLICTS OF INTEREST All authors have none to declare. REFERENCES
Tuberculosis & Liver Cirrhosis
1. Chakraborty AK. Epidemiology of tuberculosis: current status in India. Indian J Med Res. 2004;120:248–256. 2. Pai M, Das J. Management of tuberculosis in India: time for a deeper dive into quality. Natl Med J India. 2013;26:65–68. 3. Agarwal SP. TB across the globe. Tuberculosis in India – the past and prospects for the future. Scott Med J. 2000;45:11–13. 4. Steele MA, Burk RF, Desprez RM. Toxic hepatitis with isoniazid and rifampicin. A meta-analysis. Chest. 1991;19:465–471. 5. Singla R, Sharma SK, Mohan A, et al. Evaluation of risk factors for antituberculosis treatment induced hepatotoxicity. Indian J Med Res. 2010;132:81–86. 6. Saito A, Nagayama N, Yagi O, et al. Tuberculosis complicated with liver cirrhosis. Kekkaku. 2006;81:457–465. 7. Thulstrup AM, Molle I, Svendsen N, Sorenson HT. Incidence and prognosis of tuberculosis in patients with cirrhosis of liver. A Danish nationwide population based study. Epidemiol Infect. 2000;124:221–225. 8. Duseja A, Chawla YK, Dhiman RK, Kumar A, Choudhary N, Taneja S. Non-hepatic insults are common acute precipitants in patients with acute on chronic liver failure (ACLF). Dig Dis Sci. 2010;55:3188– 3192.
6
SHARMA ET AL
9. Baijal R, Praveen kumar HR, Amarapurkar DN, Nagaraj K, Jain M. Prevalence of tuberculosis in patients with cirrhosis of liver in western India. Trop Doct. 2010;40:163–164. 10. Saigal S, Agarwal SR, Nandeesh HP, Sarin SK. Safety of an ofloxacin-based antitubercular regimen for the treatment of tuberculosis in patients with underlying chronic liver disease: a preliminary report. J Gastroenterol Hepatol. 2001;16:1028–1032. 11. Cho YJ, Lee SM, Yoo CG, et al. Clinical characteristics of tuberculosis in patients with liver cirrhosis. Respirology. 2007;12:401– 405. 12. Dwivedi M, Misra SP, Misra V, Kumar R. Value of adenosine deaminase estimation in the diagnosis of tuberculous ascites. Am J Gastroenterol. 1990;85:1123–1125. 13. Porcel JM. Tuberculous pleural effusion. Lung. 2009;187:263– 270. 14. Park WB, Kim W, Lee KL, et al. Antituberculosis drug-induced liver injury in chronic hepatitis and cirrhosis. J Infect. 2010 Oct;61:323– 329. 15. Tostmann A, Boeree MJ, Aarnoutse RE, de Lange WC, van der Ven AJ, Dekhuijzen R. Antituberculosis drug-induced hepatotoxicity: concise up-to-date review. J Gastroenterol Hepatol. 2008;23:192–202. 16. Gurumurthy P, Krishnamurthy MS, Nazareth O, et al. Lack of relationship between hepatic toxicity and acetylator phenotypes in three thousand South Indian patients during treatment with isoniazid for tuberculosis. Am Rev Respir Dis. 1984;129:58–61. 17. Mitchell JR, Thorgiersson UP, Black M. Increased incidence of isoniazid hepatitis in rapid acetylators: possible relations to hydrazine metabolites. Clin Pharmacol Ther. 1975;18:70–79. 18. Yew WW, Leung CC. Antituberculosis drugs and hepatotoxicity. Respirology. 2006;11:699–707. 19. Çelikbilek M, Selçuk H, Yilmaz U. The effect of hepatotropic virus (HBV-HCV) infections on tuberculin skin test in patients with cirrhosis. Turk J Gastroenterol. 2012;23:234–238.
© 2015, INASL
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
JOURNAL OF CLINICAL AND EXPERIMENTAL HEPATOLOGY
Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis Praveen Sharma, Pankaj Tyagi, Vikas Singla, Naresh Bansal, Ashish Kumar, Anil Arora
Introduction: There is paucity of data on tuberculosis and antituberculous therapy (ATT) induced hepatotoxicity in patients with chronic liver disease. Aim: To study demographic, clinical characteristics of tuberculosis, pattern of drug induced liver injury and treatment responses to ATT in patients with liver cirrhosis. Material and method: All cases of liver cirrhosis diagnosed with tuberculosis (TB) between January 2010 and June 2013 were enrolled. Drug induced liver injury (DILI) was defined as follows (1) an aspartate aminotransferase (AST)/alanine aminotransferase (ALT) value exceeding 3 times the normal upper limit if the baseline level was normal (2 mg/dl from baseline. Results: Sixty seven patients had confirmed TB with underlying cirrhosis and formed the study group. The mean age was 52 ± 12 years and M:F ratio was 57:10. Mean Child Turcotte Pugh (CTP) score 8.5 ± 1.5 (CTP A:B:C:7:44:16). The sites of TB included: pulmonary (25, 37%); pleural effusion (10, 16%) peritoneal (19, 29%); chest lymph nodes (3, 4%); liver (1, 1.5%); intestines (3, 4%), vertebra (3, 4%), brain (1, 1.5%) and disseminated (2, 3%). Thus, extrapulmonary TB was more common in the cirrhotic patients as compared to pulmonary TB. Patients with Child's status A (n = 7) received 4 drugs (R: rifampicin, H: Isoniazid, E: ethambutol, Z: pyrizinamide) and could tolerate well even during follow up without any drug induced toxicity. In rest of patients commonest regimen followed was combination of drugs (RHEO, n = 32) followed by RHE (n = 11). DILI occurred in 35% started with either RHEO, HEO and REO. Median time of onset of DILI was 12 days (4–34) days. There was no DILI related death during hospital stay or follow up. Conclusions: Extrapulmonary TB is common in patients with cirrhosis and DILI is common in Child B and C with combination of rifampicin and isoniazid regimen. ( J CLIN EXP HEPATOL 2015;-:1–6)
T
uberculosis is major health problem in developing countries. Reported prevalence of all forms of tuberculosis in India is 5/1000.1–3 Standard antitubercular therapy (ATT) which includes rifampicin (R), isoniazid (H) and pyrizinamide (P), have hepatotoxicity as a major side-effect.4,5 The mechanism of hepatotoxicity and the risk factors that predispose to it are still not clearly defined. Liver cirrhosis is considered as an acquired immune deficiency state due to malnutrition or due to advanced liver problem. Drug induced liver injury (DILI) is difficult to diagnose in patients with chronic liver disease because of elevated enzymes and bilirubin at baseline and fluctuations in enzymes level caused by underlying liver diseases may
Keywords: cirrhosis, tuberculosis, treatment Received: 3.3.2014; Accepted: 14.1.2015; Available online: xxx Address for correspondence: Anil Arora, Department of Hepatology and Gastroenterology, Sir Ganga Ram Hospital, New Delhi, India. E-mail: [email protected] Abbreviation: AST/ALT: Aspartate aminotransferase/alanine aminotransferase; ATT: Antituberculous therapy; CTP: Child Turcotte Pugh; DILI: Drug induced liver injury; E: ethambutol; H: Isoniazid; R: Rifampicin; TB: Tuberculosis; Z: pyrizinamide http://dx.doi.org/10.1016/j.jceh.2015.01.003 © 2015, INASL
confound the diagnosis. As these drugs are metabolized in the liver, there is a theoretical risk of increased hepatotoxicity in patients with underlying chronic liver disease (CLD).6,7 ATT constitutes an important acute precipitating event in decompensation of underlying known or unknown liver disease in India.8 There is paucity of data in defining hepatotoxicity in patients with cirrhosis and clear guidelines do not exist regarding the monitoring of such patients, modification of the antitubercular regimen and mode of reintroduction of individual antitubercular therapy. Little is known about the characteristics of TB in patients with liver cirrhosis. The relationship between TB and liver cirrhosis in terms of clinical characteristics, treatment responses and adverse effects is very limited.9–11 The aim of this study was to elucidate demographic, clinical characteristics of tuberculosis, pattern of drug induced liver injury and treatment responses to ATT in patients with liver cirrhosis.
MATERIALS AND METHODS This study is a retrospective analysis of prospective collective data of all cases of liver cirrhosis diagnosed with tuberculosis between January 2010 and June 2013. Patients were
Journal of Clinical and Experimental Hepatology | - 2015 | Vol. - | No. - | 1–6
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
Tuberculosis & Liver Cirrhosis
Department of Gastroenterology, Sir Ganga Ram Hospital, New Delhi, India
PATIENTS WITH LIVER CIRRHOSIS
Tuberculosis & Liver Cirrhosis
either admitted in hospital or were seen in outpatient department. All the patients of cirrhosis were not screened for tuberculosis. Patients with cirrhosis were screened for tuberculosis only if patients presented with following parameters: fever, cough for >2 weeks, hemoptysis, unexplained weight loss, increasing ascites not responding to diuretics, unexplained bowel symptoms (diarrhea, constipation, or subacute intestinal obstruction), radiological lesions suggestive of tuberculosis and past or family history of tuberculosis. Patients were diagnosed to have cirrhosis if they had liver histology suggestive of cirrhosis or based on clinical, biochemical and radiological imaging findings (either on ultrasound, transient elastography and computed scan). Medical records, bacteriological examinations and radiographic findings were recorded for clinical characteristics, responses to treatment and adverse effects of anti-TB treatment. If suspected, investigations including hemogram, erythrocyte sedimentation rate (ESR), Mantoux test, sputum for acid-fast bacillus (AFB) repeated thrice, chest X-ray, ultrasonography, Endoscopic ultrasonography with fine needle aspiration of lymph node, CT scan of thorax and abdomen, and barium studies were done. Mantoux test (tuberculin skin test) was considered positive if there was $5 mm of induration at 48–72 h. A detailed ascitic fluid and pleural fluid examination was done including cytology, protein and albumin content, serum ascites albumin gradient, adenosine deaminase levels, acid fast bacillus culture and PCR for Mycobacterium tuberculosis. Aspiration cytology and tissue biopsy including bone marrow and liver biopsy were done if deemed necessary for the diagnosis of fever of an unknown origin. Tuberculosis was diagnosed based on a good clinical response to chemotherapy and one or more of the following: (i) histological evidence of caseating granulomas; (ii) sputum positivity for AFB; (iii) growth of Mycobacterium tuberculosis on culture; and (iv) positive PCR for Mycobacterium tuberculosis in tissues. The diagnosis of peritoneal and pleural tuberculosis was established if three of the following four criteria were fulfilled: (i) raised ascitic fluid/pleural cell count with lymphocyte predominance; (ii) raised ascitic fluid/pleural fluid albumin (>2.5 g/dL); (iii) raised adenosine deaminase (>33 U/L)12 in ascites fluid and 40 Ul/L in pleural fluid13 and (iv) positive PCR for M. tuberculosis. Treatment regimen was decided by the treating physician and it was his decision to start the antitubercular treatment in combination of his choice considering the condition of the patient. Patients were followed up every 15 days in the initial 2 months after the discharge if admitted and then at month interval. In case of any problem patients were asked to come to the hospital in between the planned visits. All these patients were seen by the authors of these studies and patients who were diagnosed to have cirrhosis and tuberculosis were regularly followed up and their data was collected. Drug induced liver injury 2
SHARMA ET AL
(DILI) was defined as follows (1) an aspartate aminotransferase (AST)/alanine aminotransferase (ALT) value exceeding 3 times the normal upper limit if the baseline level was normal (2 mg/dl in absence of no other alternative identifiable reason for the above changes.14
STATISTICAL ANALYSIS Statistical comparisons were performed using spss (SPSS for Windows, version 12.0; SPSS; Chicago, IL, USA). Values were expressed in median with range Chi-square tests were used to test differences in categorical variables and t-tests were used for continuous variables. Statistical significance was assumed at P < 0.05.
RESULTS Sixty seven (n = 67) patients were diagnosed to had TB with underlying cirrhosis and formed the study group during this study interval. Baseline characteristics of sixty seven patients with TB and liver cirrhosis were shown in Table 1. The median age of patients were 52 (24–85) years and male-to female ratio was 57:10. Median CTP score of patients were 8.5 (5–12) (CTP A:B:C:7:44:16) (Table 1). The etiology of cirrhosis of liver were: alcohol (33, 49%); postviral infection—hepatitis B and C (15, 23%), cryptogenic (11, 16%) and NASH related (8, 12%). Diagnosis of NASH related cirrhosis was labeled in 8 patients based on histological diagnosis of cirrhosis. Four of these patients had liver biopsy done before which showed F3 Fibrosis, steatosis and steatohepatitis and four patient had radiological evidence of steatosis in past (n = 4). Hence diagnosis of NASH related cirrhosis were labeled in these
Table 1 Baseline Characteristics of Patients with Cirrhosis and Tuberculosis (n = 67). Parameter
Median (range) (n = 67)
Age (yr)
52 12
M:F
57:10
CTP score
8.5 1.5
Total bilirubin mg/dl
4.8 6.1
ALT IU/L
61 42
AST IU/L
76 54
International normalized ratio
1.8 0.7
Child's A:B:C
7:44:16
Etiology of cirrhosis ALD:Viral:Cryptogenic:NASH
33:15:11:8
AST: aspartate aminotransferase ALT: alanine aminotransferase CTP: Child-Turcotte-Pugh Score. © 2015, INASL
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
JOURNAL OF CLINICAL AND EXPERIMENTAL HEPATOLOGY
Table 2 Clinical and Radiographic Characteristics of Patients with TB with Cirrhosis.
Table 3 Diagnostic Modalities for Tuberculosis. Pulmonary Tuberculosis (n = 25)
Extrapulmonary TB Pleural effusion (n = 10) Nodal (N = 3)
Chest (n = 3) Peritoneal (n = 19)
Bone (n = 3)
Meningeal (n = 1)
Disseminated (n = 2)
Liver (n = 1) Intestinal (n = 3)
Number (%) Clinical symptoms Fever
45 (67%)
Unexplained weight loss
23 (34%)
Sputum acid fast bacilli positive (n = 16) Fibrocavitatory lesion on chest X-ray and CT scan of thorax (n = 12) Bronchoalveolar lavage with AFB +ve (n = 5) Raised ADA (n = 10) in pleural fluid (detected in CXR with lymphocytic predominance and high protein Chest CT: shows mediastinal lymph node and EUS FNAC showed granuloma with caseation (n = 3) High ADA in ascitic fluid (n = 12) Peritoneal biopsy (n = 2) EUS FNAC intra-abdominal lymph nodes along with high SAAG septated ascites (granuloma with AFB, n = 5) CT scan and MRI scan showing vertebra destruction (n = 3) Positive Mx and high ESR (n = 3) Bone biopsy showing granuloma (n = 1). Cerebrospinal fluid showed lymphocytic predominance with high protein findings, PCR positive for tuberculosis, CT brain suggestive of meningitis Chest X-ray (miliary shadow n = 2) CT abdomen and thorax showed multiple lymphadenopathy (n = 1) EUS guided FNAC from abdominal lymph node (n = 1) and omental biopsy (n = 1) showed granulomatous lesion. Liver biopsy which showed granuloma Double balloon enteroscopy with biopsy suggestive of granuloma (n = 1) CT: abdomen suggestive of stricture and necrotic lymph node (n = 2)
Loss of appetite
55 (82%)
Antitubercular Therapy and Side Effects
Increasing ascites not responding to diuretics
17 (25%)
Patients with Child's Pugh status A (n = 7) received 4 drugs (rifampicin, Isoniazid, Ethambutol and pyrizinamide RHEZ) for the initial two months and then two drugs rifampicin and isoniazid for 7 month without any drug induced toxicity. All the patients had good response and ATT was stopped after nine month of duration and there was no mortality in this group. In Child's B and C, patients were started on combination of drug treatment which included [(Rifampicin R, isoniazid H, ethambutol E and ofloxacin O (RHEO)]. The commonest combination of drugs used was [RHEO, n = 32). This drug combination was started by the treating physician based on patients clinical and biochemical parameters (Table 4). Pyrizinamide was not used instead quinolones was preferred in these patients with advanced liver disease (Child's B and C). Drug induced toxicity was seen in 35% patients in patients started with either RHEO, HEO and REO (Table
Cough Pain abdomen
8 (12%) 14 (21%)
Site of involvement Pulmonary tuberculosis
25 (37%)
Extrapulmonary TB
42 (63%)
Pleural effusion
10 (16%)
Nodal (chest)
3 (4%)
Intestinal
3 (4%)
Peritoneal
19 (29%)
Bone
3 (4%)
Meningeal
1 (1.5%)
Disseminated
2 (3%)
Liver
1 (1.5%)
Journal of Clinical and Experimental Hepatology | - 2015 | Vol. - | No. - | 1–6
3
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
Tuberculosis & Liver Cirrhosis
patients. Cryptogenic cirrhosis was diagnosed based on histological evidence of cirrhosis with no obvious etiology based on histological report of liver biopsy in these patients and negative work up for viral markers (hepatitis B and C), autoimmune markers, Wilson and hemochromatosis disease. None of the patients had HIV co-infection. The TB sites included: pulmonary (25, 37%); pleural effusion (10, 16%) peritoneal (19, 29%); chest lymph nodes (3, 4%); liver (1, 1.5%); intestines (4, 4%); and bone (3, 4%), brain (1, 1.5%) and disseminated (2, 3%). In disseminated tuberculosis histological diagnosis was made based on omental biopsy (n = 1) and EUS guided lymph node FNAC in one case. Thus, extrapulmonary TB was more common in the cirrhotic patients than pulmonary TB (P < 0.05). Most common presenting symptom was loss of appetite (82%) and fever (67%), weight loss (34%) and increasing ascites (25%) (Table 2). Mantoux test (tuberculin skin test) was positive in 43 (64%, Child's status: A:B:C:71%:70%:43%) patients. In 59 patients, TB was confirmed either histopathologically (either biopsy or fine needle aspiration) AFB smear or by fluid analysis. In the remaining 8 patients (4 with pulmonary TB, 2 with intestinal, 2 with bone), indirect indicators such as a history of fever and weight loss, high erythrocyte sedimentation rate (ESR), a positive Mantoux test, typical X-ray and CT thorax and abdomen appearance were taken into account before starting treatment (Tables 2 and 3).
PATIENTS WITH LIVER CIRRHOSIS
SHARMA ET AL
Table 4 Regimens of Anti-TB Treatment and Toxicity. Parameters
Baseline regimen (N)
Follow up regimen after 2 months (N) till completion of therapy
Child's A (n = 7)
RHEZ(7)
RH(7)
Child's B (n = 44)
RHEO(32) + RHE(n = 9) + HEO(3)
RH(28) + HEO(8) + EO(2)
Child's C (n = 16)
REO(7) + RHE(2) + HEO(2) + EO(5)
REO(4) + HEO(4) + EO(4)
Fluroquinolones based regimen
51/67 (76%)
*Rifaximin and isoniazid could be reintroduced in 6 and 4 patients respectively without side effects: 4 patients lost to follow up in Child's B; Ò6 patients died during hospital stay (Child's B, n = 2, Child's C n = 4). R: rifampicin, H: isoniazid, E: ethambutol, Z: pyrizinamide, O: ofloxacin.
Tuberculosis & Liver Cirrhosis
3). Median time of onset of DILI was 12 days (4–34) days and median follow up period of these patients was 13 months (13 days–26 month). Of 43 patients with (R + H) regimen in Child's B and C it could be continued for more than 2 months in 22 patients due to drug related toxicity in these patients. Reintroduction of rifampicin was possible in 6 patients and INH in 4 patients respectively after drug induced toxicity subsided. In these patients ethambutol, ofloxacin with either rifampicin or isoniazid were continued for rest of their treatment. All patients in Child's Pugh status B and C continued quinolones and ethambutol (100%) till end of therapy. Injection streptomycin was used in 3 patients in those who developed drug induced toxicity all in Child B status for 2 months till drug induced toxicity was over. RH combination was not used in Child C patients except in two patients in the first two months. The duration of treatment ranged from 9 to 18 months. Eighteen month treatment was given in patients with disseminated TB, meningeal TB and bone TB patients. Peritoneal TB, tubercular pleural effusion and intestinal TB was treated with 12 months of treatment and all responded clinically and pleural effusion disappeared on follow up. Ascites was controlled and dose of diuretics decreased. Abdominal pain disappeared and patient gained during follow up in intestinal TB. Fever subsided and weight gain was seen in patient with liver tuberculosis on follow up, repeat biopsy was not done in this patient. Pulmonary TB was treated for 9 months and all 19 patients who completed the therapy responded. Four patients with pulmonary TB were lost to follow up after hospital discharge all in Child's B and two patients (Child's C) died during hospitalization. No patient developed drug-induced liver failure. In hospital mortality was 6 (9%) and was due to decompensated liver disease (Child's B, n = 2 and Child's C n = 4) and these are the patients who could not complete the tubercular therapy. Of these 6 patients four patients died due to severe upper gastrointestinal bleed and two developed pneumonia with sepsis and renal failure. There was no DILI induced mortality in our study either during hospitalization or during follows up of these patients. Four patients with pulmonary tuberculosis in Child's B were lost to follow up. 4
DISCUSSION Pulmonary TB was noted in 37% of cases and extrapulmonary TB in 63%. Drug induced liver injury was also higher (35%) in this cohort of patients. RHEO is the commonest combination used in Child B cirrhosis patients, whereas ethambutol and ofloxacin based combination is commonly used in Child C patients. The major antitubercular drugs: isoniazid, rifampicin and pyrazinamide are all metabolized in the liver. The reported incidence of antituberculosis drug-induced hepatotoxicity varies between 2% and 28%.Pooled analysis of Indian trials in non cirrhotic patients found 11.5% incidence of drug induced liver injury whereas a meta analysis of western studies found a pooled risk of 4.28% in patients with no cirrhosis.4,5 Common risk factors are advanced age, female sex, slow acetylator status, malnutrition, HIV and pre-existent liver disease.15 However predictive model for identification of DILI is not available. Patients with cirrhosis have been shown to have a longer half-life of isoniazid and rifampicin, thus leading to their accumulation in the plasma. Isoniazid, pyrazinamide and rifampicin have hepatotoxic potential, and can lead to such reactions during antituberculosis chemotherapy. Most of the hepatotoxic reactions are dose-related; some are, however, caused by drug hypersensitivity.16,17 In a meta-analysis, drug-induced hepatic dysfunction usually occurs within the initial few weeks of the intensive phase of antituberculosis chemotherapy. Vigilant clinical (including patient education on symptoms of hepatitis) and biochemical monitoring are mandatory to improve the outcomes of patients with drug-induced hepatotoxicity during antituberculosis chemotherapy. Some fluoroquinolones like ofloxacin/levofloxacin may have a role in constituting non-hepatotoxic drug regimens for management of tuberculosis (TB) in the presence of hepatic dysfunction.18 Our results support these observations, as 35% of the patients treated with combination treatment developed hepatotoxicity and majority of these patients are from Child B and Child C as these patients have low albumin and poor nutritional status. None of patient from Child © 2015, INASL
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
A developed DILI but these patient constitute only 10% of our cohort. Similarly only 2 patients received RH combination of 16 from Child C and they developed DILI. Thus, a combination of isoniazid and rifampicin in patients with underlying CLD especially in Child's B and C should possibly be avoided. In a similar study by Park et al14 18 (17%) experienced DILI: 11 (24%) among 46 patients with chronic hepatitis and 7 (15%) among 46 patients with compensated liver cirrhosis. We could not predict factors associated with DILI in our study as all patients did not receive the same drugs so predicting factors for DILI would have bias in our study. None of the patients had any sideeffects from fluoroquinolones and ethambutol during the entire course of therapy and follow up which could be a better alternative in patients with advanced liver disease and this is in accordance with Saigal et al10 who concluded that in patients with tuberculosis who have underlying CLD: (i) an ofloxacin-based antitubercular regimen without rifampicin is as effective as a rifampicin-based regimen; and a combination of isoniazid with rifampicin is more hepatotoxic than a combination with ofloxacin and pyrizinamide. Median duration of development of DILI in our study was 12 days with onset as early as day 4 of starting ATT. Our practice in patients with cirrhosis and tuberculosis is to monitor liver function tests every third day till the patient is admitted and then weekly for initial two month of therapy and every fifteen days interval thereafter. Extrapulmonary tuberculosis was seen in 63% of patients in our study. Similar were the results of Baijal et al9 60%, Saigal et al10 65% and Cho et al11 in whom 31% had extrapulmonary TB of 36 patients enrolled. Little is known about the immunopathogenesis of TB in such clinical conditions. Diagnosis of extrapulmonary TB also needs meticulous investigation along with clinical data. With the availability of EUS guided FNAC diagnosis of nodal TB has increased and all measures should be taken to make a tissue diagnosis in patients with cirrhosis, as unnecessary treatment can lead to further deterioration of liver functions in this group of patients. Even with best measures empirical ATT was started in 8 patients in this study based on clinical and biochemical parameters. We found Mantoux test positive in 64% of patients in this study. There is no recommendation of testing Mantoux test in India to screen for tuberculosis in patients with cirrhosis. However study by Çelikbilek M et al19 has shown that in patients with cirrhosis who were waiting for liver transplant and no active tuberculosis, the tuberculin skin test findings were significantly higher in end-stage liver disease caused by viruses than those with a non-viral etiology (P < 0.05). The mean induration of the tuberculin skin test in Child B patients was found to be 7.15 6.4 mm, while in Child C patients it was 12.64 7.5 mm. In Child C patients, the tuberculin skin test scores were significantly higher than in Child B pa-
tients (P < 0.05). So Mantoux test gets affected by status of liver disease and etiology of liver disease. We did it all of our patients and recommend it for all patients as it contributes to clinical decision making along with symptoms if other imaging were not contributory for the diagnosis of tuberculosis. Multi drug induced hepatotoxicity has been well defined in patients with normal liver function tests and no cirrhosis, however its definition in patients who already had impaired liver function tests has not well been documented. Saigal et al10 take arbitrary values for defining DILI as ALT/AST level increased to more than fivefold the baseline level, or to more than 400 IU/L, or if the bilirubin increased by 2.5 mg/dL after exclusion of superimposed acute hepatitis and its definition varies with other published literature. Future studies are needed on this issue to define DILI in patients who are already jaundices and had elevated AST and ALT levels at baseline. Similarly there is no consensus on treatment duration of extrapulmonary TB in patients with cirrhosis and the duration varies from 6 to 18 months. In our study also duration varies in extrapulmonary tuberculosis from 9 to 18months. So future studies should address this issue. However different drugs, different Child Status of patients and different site of involvement might preclude studies to answer this very important clinical problem in day to day practice of Hepatologist. Saito et al6 had shown in a cohort of 44 patients that tuberculous patients with liver cirrhosis are characterized with higher mortality rate and higher frequency of adverse effects of antituberculous chemotherapy. Multi-drug combination regimen could be tolerable under adequate surveillance of side effects even in the situation of preexisting liver dysfunction. We also found 9% in hospital mortality. The main drawbacks of our study was that as patients with cirrhosis were seen by all authors in their respective outpatients department and only few of them were admitted. We did not have record of patients who were seen in outpatient department and had cirrhosis and but not admitted. Hence exact frequency of tuberculosis in patients with cirrhosis cannot be determined in such situation. We did not use fixed combination of anti tubercular drugs as per the Child's status as this study was a collection of data of patients seen by many authors in a real time practice and their follow up till they continued the ATT. A fixed combination of antitubercular therapy were not used as patient had different Child's status and had different course of disease during follow up. However it would be useful to conduct more studies in patients with cirrhosis and tuberculosis and use a standardized treatment in them according to their Child's status and know about its safety so that a particular regimen could be used in future trials according to the response. To conclude extrapulmonary tuberculosis is common in patient with cirrhosis and drug toxicity due to
Journal of Clinical and Experimental Hepatology | - 2015 | Vol. - | No. - | 1–6
5
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
Tuberculosis & Liver Cirrhosis
JOURNAL OF CLINICAL AND EXPERIMENTAL HEPATOLOGY
PATIENTS WITH LIVER CIRRHOSIS
combination of rifampicin and isoniazid is more common in patients with decompensated cirrhosis. Quinolone based patient tailored extended therapy is an alternative in patient with advanced cirrhosis.
CONFLICTS OF INTEREST All authors have none to declare. REFERENCES
Tuberculosis & Liver Cirrhosis
1. Chakraborty AK. Epidemiology of tuberculosis: current status in India. Indian J Med Res. 2004;120:248–256. 2. Pai M, Das J. Management of tuberculosis in India: time for a deeper dive into quality. Natl Med J India. 2013;26:65–68. 3. Agarwal SP. TB across the globe. Tuberculosis in India – the past and prospects for the future. Scott Med J. 2000;45:11–13. 4. Steele MA, Burk RF, Desprez RM. Toxic hepatitis with isoniazid and rifampicin. A meta-analysis. Chest. 1991;19:465–471. 5. Singla R, Sharma SK, Mohan A, et al. Evaluation of risk factors for antituberculosis treatment induced hepatotoxicity. Indian J Med Res. 2010;132:81–86. 6. Saito A, Nagayama N, Yagi O, et al. Tuberculosis complicated with liver cirrhosis. Kekkaku. 2006;81:457–465. 7. Thulstrup AM, Molle I, Svendsen N, Sorenson HT. Incidence and prognosis of tuberculosis in patients with cirrhosis of liver. A Danish nationwide population based study. Epidemiol Infect. 2000;124:221–225. 8. Duseja A, Chawla YK, Dhiman RK, Kumar A, Choudhary N, Taneja S. Non-hepatic insults are common acute precipitants in patients with acute on chronic liver failure (ACLF). Dig Dis Sci. 2010;55:3188– 3192.
6
SHARMA ET AL
9. Baijal R, Praveen kumar HR, Amarapurkar DN, Nagaraj K, Jain M. Prevalence of tuberculosis in patients with cirrhosis of liver in western India. Trop Doct. 2010;40:163–164. 10. Saigal S, Agarwal SR, Nandeesh HP, Sarin SK. Safety of an ofloxacin-based antitubercular regimen for the treatment of tuberculosis in patients with underlying chronic liver disease: a preliminary report. J Gastroenterol Hepatol. 2001;16:1028–1032. 11. Cho YJ, Lee SM, Yoo CG, et al. Clinical characteristics of tuberculosis in patients with liver cirrhosis. Respirology. 2007;12:401– 405. 12. Dwivedi M, Misra SP, Misra V, Kumar R. Value of adenosine deaminase estimation in the diagnosis of tuberculous ascites. Am J Gastroenterol. 1990;85:1123–1125. 13. Porcel JM. Tuberculous pleural effusion. Lung. 2009;187:263– 270. 14. Park WB, Kim W, Lee KL, et al. Antituberculosis drug-induced liver injury in chronic hepatitis and cirrhosis. J Infect. 2010 Oct;61:323– 329. 15. Tostmann A, Boeree MJ, Aarnoutse RE, de Lange WC, van der Ven AJ, Dekhuijzen R. Antituberculosis drug-induced hepatotoxicity: concise up-to-date review. J Gastroenterol Hepatol. 2008;23:192–202. 16. Gurumurthy P, Krishnamurthy MS, Nazareth O, et al. Lack of relationship between hepatic toxicity and acetylator phenotypes in three thousand South Indian patients during treatment with isoniazid for tuberculosis. Am Rev Respir Dis. 1984;129:58–61. 17. Mitchell JR, Thorgiersson UP, Black M. Increased incidence of isoniazid hepatitis in rapid acetylators: possible relations to hydrazine metabolites. Clin Pharmacol Ther. 1975;18:70–79. 18. Yew WW, Leung CC. Antituberculosis drugs and hepatotoxicity. Respirology. 2006;11:699–707. 19. Çelikbilek M, Selçuk H, Yilmaz U. The effect of hepatotropic virus (HBV-HCV) infections on tuberculin skin test in patients with cirrhosis. Turk J Gastroenterol. 2012;23:234–238.
© 2015, INASL
Please cite this article in press as: Sharma et al., Clinical and Biochemical Profile of Tuberculosis in Patients with Liver Cirrhosis, Journal of Clinical and Experimental Hepatology (2015), http://dx.doi.org/10.1016/j.jceh.2015.01.003
