Background: Antituberculosis drug-induced hepatotoxicity (anti-TB-DIH) is a worldwide serious medical problem among TB patients. Oxidative stress has also been proposed as one of the possible mechanisms involved in anti-TB-DIH. Aims: The main aim of this study was to evaluate the risk factors for anti-TB-DIH and further, to investigate the possible association of anti-TB-DIH with oxidative stress. Materials and Methods: This prospective cohort study was conducted in 244 TB patients receiving anti-TB treatment. Liver function (ALT, AST, Bilirubin) and oxidative stress parameters (Malondialdehyde, Glutathione, Superoxide dismutase) were estimated before and during the treatment period, along with clinical observation. Results: 14.3% patients developed Anti-TB-DIH. Patients with female gender, low body mass index, extra-pulmonary TB and positive HIV were significant risk factors for anti-TB-DIH. During the treatment period, patients with DIH showed a significant (P<0.0001) higher level of malondialdehyde, superoxide dismutase and a significant (P<0.0001) lower level of glutathione as compared to patients without DIH. Conclusion: In summary, we found that TB patients with female gender, low body mass index, extrapulmonary and positive HIV were at higher risk for anti-TB-DIH. Our finding further conclude that higher level of MDA with altered level of antioxidants in patients with DIH, may be due to oxidative stress-resulting from anti-TB drugs.
Keywords: Antituberculosis drug-induced hepatotoxicity, oxidative stress, risk factors, tuberculosis patients
Tuberculosis (TB) is one of the major causes of illness and death worldwide. Globally, 8.9 million new cases of TB occur annually, of which 1.8 million (20%) occur in India.  Directly observed treatment short-course (DOTS) strategy using first-line drugs (isoniazid, rifampicin and pyrazinamide) is very effective treatment for TB and cure 95% cases.  Despite the high efficacy of DOTS, these anti-TB drugs may cause hepatotoxicity, skin reactions, gastrointestinal and neurological disorders. Among these, the most commonly encountered and serious side effect is hepatotoxicity, which limits the clinical use of drugs, contributes to treatment failure and thereby leads to drug resistance. , Some patients taking anti-TB drugs develop severe hepatitis that may progress to liver failure and finally death. 
Early recognition of anti-TB drug-induced hepatotoxicity (anti-TB-DIH) and its risk factors hold great significance to arrest severe liver injury and prevent treatment failure.  The most common risk factors that have been found to be associated with anti-TB-DIH, are advanced age, female sex, poor nutritional status, liver disease, inappropriate use of drugs, infection with hepatitis B virus (HBV), hepatitis C virus (HCV) and human immunodeficiency virus (HIV), and high alcohol intake.  However, the exact factors associated with anti-TB-DIH are still obscure.
The anti-TB-DIH is often suggested to be mediated through oxidative stress, leading to generation of lipid peroxidation and alteration in antioxidants. , Lipid peroxidation occurs as a chain reaction initiated by excess production of free radicals. In response to deleterious effects of free radical induced lipid peroxidation, cells activate antioxidant defense mechanisms in which superoxide dismutase and reduced glutathione act synergistically to detoxify the effects of lipid peroxidation.  Therefore, the evaluation of lipid peroxidation and antioxidants in patients with DIH may be of importance to further evaluate the possible involvement of oxidative stress in the pathogenesis of anti-TB-DIH.
The present study was undertaken to identify the risk factors for anti-TB-DIH and to further explore the possible association of anti-TB-DIH with oxidative stress in North Indian population. The findings of this study may be imperative in screening among individuals at high risk for anti-TB-DIH and ultimately can refine DIH prevention efforts.
This prospective cohort study was conducted in TB patients with age ranging 15-65 years, taking anti-TB drugs under DOTS, visiting Department of Pulmonary Medicine, C.S.M. Medical University, Lucknow, India. The study was carried out from March 2009 to February 2011. The patients fulfilled following criteria- (i) they were going to start anti-TB drugs, (ii) not receiving any other hepatotoxic drugs parallel with anti-TB treatment, (iii) no history of chronic alcohol intake, (iv) normal findings of liver function parameters at the beginning of the treatment, and (v) negative for hepatitis B/C. The study was approved by the Ethical Committee and informed consent was taken from all the patients included in the study. The questionnaire was intended to elicit information of the patients’ demographic characteristics, full details of all anti-TB drugs with doses, concomitant use of other drugs, diseases characteristics and any clinical manifestation of anti-TB-DIH.
Diagnosis of anti-TB-DIH
Presence of at least one of the following criteria was used to define anti-TB-DIH: 
The patients were followed up both clinically (response to therapy, any adverse effects) and biochemically with special reference to liver function and oxidative stress parameters. Liver function parameters included ALT, AST and bilirubin. Oxidative stress parameters included malondialdehyde, glutathione and superoxide dismutase. All these parameters were performed before the initiation of the treatment. After the initiation of drug treatment, follow-up were performed every two weeks during the first month and then monthly until completion of treatment. These biochemical parameters were repeated later, whenever symptoms suggestive of hepatotoxicity like nausea, anorexia, vomiting occurred. Proper monitoring was done in these patients and instructed to report any unusual signs and symptoms they will come across during their course. During the treatment period, only peak values of liver function parameters were recorded and at the same time level of oxidative stress parameters were also recorded.
Estimation of liver function parameters
Liver function parameters (ALT, AST and bilirubin) were measured in blood serum by utilizing an auto-analyzer in the pathology lab, C.S.M. Medical University. Normal range in the laboratory was 9-43 U/L for ALT, 9-43 U/L for AST and 0.4-1.2 mg/dl for bilirubin.
Estimation of oxidative stress parameters
The rate of lipid peroxidation was estimated by measuring the blood plasma concentration of malondialdehyde (a stable aldehydic products of lipid peroxidation), according to the method of Ohkawa et al. (1979).  200 μl of blood plasma was mixed with 1ml of 20% acetic acid, subsequently added 200 μl of 8% SDS (pH adjusted to 4). Following that, 1.5 ml of 0.8% thiobarbituric acid (TBA) and 1.1ml of distilled water (DW) were added. Reaction mixture was incubated in boiling water bath for one hour. After cooling, 3ml of n-butanol was added. A clear butanol fraction thus obtained was used for measuring the absorbance at 532 nm and expressed as nmol/ml.
Blood GSH was measured according to the method of Beutler et al. (1963)  with slight modification. 200 μl of whole blood was mixed with 1.8 ml of cold distilled water and incubated for 10 min at 37 0 C. Six hundred microliter of sulphosalicylic acid was added to the reaction mixture and centrifuged at 2000 rpm for 15 min. To 200 μl of supernatant, 400 μl of phosphate buffer and 80 μl of dithiobis nitrobenzoic acid (DTNB) were added. The absorbance of the yellow color developed was measured at 412 nm and expressed as μg/ml.
Superoxide dismutase (SOD) activity was measured according to the method of McCord and Fridovich (1969) 13] with slight modification. Two hundred microliter of whole blood was washed thrice with ice cold normal saline by centrifuging three times at 10,000 rpm. RBCs collected were hemolyzed by adding 1.5 volume of cold water. Cold ethanol (300 μl) and chloroform (180 μl) were then added and centrifuged. The resulting supernatant containing SOD was taken for the measurement of its activity. Two reactions setup were run in parallel for SOD estimation. The first setup (experimental) received 1.2 ml (0.052 M) of sodium pyrophosphate, 0.3 ml (186 μM) of phenazine methosulphate, 0.3 ml (300 μM) of nitroblue tetrazolium and 0.2 ml of enzyme source. The second setup (reference) received the entire above reagents except the enzyme source. 0.8 ml and l ml DW was added respectively in both sets and finally reactions were started simultaneously by the addition of 0.2 ml (780 μM) of NADH. After an interval of 90 seconds, 1 ml of glacial acetic acid was then added to each reaction tube and absorbance was read at 560 nm against a blank on spectrophotometer. The SOD activity was expressed as Unit/g Hb.
The data collected was entered in Microsoft Excel and checked for any inconsistency. The descriptive statistics such as mean, standard deviation and percentages were calculated. The chi-square test was used to compare categorical/dichotomous variables. The unpaired t-test was used to compare independent continuous variables. The relative risk and its 95% confidence interval were calculated to find out the risk factors for anti-TB-DIH in univariate analysis. Further, multiple logistic regression analysis was also used to determine the adjusted risk factors for anti-TB-DIH. The P-value <0.05 was considered significant. All the analysis was carried out by using SPSS 15.0 version.
Baseline demographic and disease characteristics of total patients
The baseline characteristics of the patients are given in the [Table 1]. A total of 244 cases were included in the study. Patients with advanced age group (≤35 years) were in majority (73.4%) and nearly half of the patients (50.4%) were males. More than half (58.2%) of the patients were married. Similarly, more than half of the patients (69.3%) patients belonged to lower socio-economic status. About one fifth (20.5%) had body mass index (BMI) <18.5 (considered as low BMI). About half (52.9%) of the patients had pulmonary TB and remaining (47.1%) extra-pulmonary TB included cervical lymphadenitis (39.1%), pleural effusion (24.3%), bone (7.8%), meningitis (7.0%), ocular (6.1%), genitourinary (8.7%) and gastrointestinal (7.0%). There were only 7.4% HIV positive patients.
Prevalence of anti-TB-DIH and its risk factors
During the treatment period, 14.3% (35/244) patients developed anti-TB-DIH, which was detected by clinical examination and confirmed by liver function parameters. As shown in [Table 2] pretreatment level of ALT and AST were almost similar in patients with and without DIH (ALT=27.17±4.02 versus 26.60±3.82, p=0.42; AST=28.83±3.40 versus 29.33±2.64, P=0.32). Whereas, during the treatment period, all patients developing DIH showed significant higher (P<0.0001) level of ALT, AST and bilirubin as compared to patients without DIH (ALT=144.03±24.35 versus 30.02±4.46; AST=141.63±21.02 versus 32.53±3.18; bilirubin=1.27±0.36 versus 0.79±0.14). However, we also found that pretreatment bilirubin level was already significantly higher (P<0.02) in patients without DIH as compared to patients with DIH (0.66±0.18 versus 0.59±0.09).
The time interval from the initiation and the onset of hepatotoxicity was 15-50 days. The clinically observed symptoms of anti-TB-DIH were vomiting in 80% patients, nausea in 65.7% patients, anorexia in 31.4% patients, abdominal pain in 17.1% patients and jaundice in 2.8% patients.
The univariate analysis indicated that the prevalence of anti-TB-DIH was significantly (P=0.03) higher in younger age group (≤35 years) as compared to older (RR=2.81, 95% CI=1.03-7.66). Females were more (p=0.005) affected with anti-TB-DIH as compared to males (RR=0.39, 95%CI=0.20-0.78). There was no statistically significance (P=0.82) difference in the incidence of anti-TB-DIH between married and unmarried groups. Anti-TB-DIH was insignificantly higher in lower socio-economic status (17.8%) as compared to middle (7.1%) and upper middle status (5.3%). The prevalence of anti-TB-DIH was significantly (P=0.002) higher in low BMI (<18.5) than BMI ≥18.5 (RR=2.59, 95% CI=1.42-4.72). Also, pulmonary patients were found to be at 59% lower risk (P=0.006) for anti-TB-DIH as compared to extra pulmonary patients (RR=0.41, 95s%CI=0.21-0.80). Similarly, the prevalence of anti-TB-DIH was significantly (P=0.02) higher in HIV positive patients than HIV negative patients (RR=2.60, 95%CI=1.24-5.43) [Table 3].
The multivariate logistic regression analysis indicated that age was not a significant (P=0.058) risk factor for anti-TB-DIH. Further, the analysis showed that females had 58% of higher risk (P=0.04) of being hepatotoxic as compared to males (adjusted OR=0.42, 95% CI=0.18-0.98). Patients who had BMI <18.5 were at higher risk for the development of anti-TB-DIH (adjusted OR=3.49, 95%CI=1.52-8.01). The analysis also indicated that pulmonary patients had 68% lower risk for anti-TB-DIH than extra pulmonary patients (adjusted OR=0.32, 95%CI=0.14-0.74). Further, HIV positive patients had about 5 times higher risk of being hepatotoxic as compared to HIV negative patients (adjusted OR=5.13, 95%CI=1.51-17.39) [Table 4].
Level of oxidative stress parameters in patients with and without DIH
In our study, we estimated the level of oxidative stress parameters in both patients with and without DIH at different stages i.e. before and during the treatment period. We observed that there was no significant difference (P=0.80) in the pretreatment level of MDA in patients with and without DIH (6.88±1.36 versus 6.95±1.54). However, during the treatment period, the level of MDA was significantly higher (P<0.0001) in patients with DIH as compared to patients without DIH (10.68±1.0 versus 7.79±1.28) [Table 5].
Likewise, before the treatment period, no significant difference (P=0.33) was observed in the level of GSH, between patients with and without DIH (63.69±9.79 versus 65.55±10.46). However, during the treatment period, the GSH level was found to be significantly lower (P<0.0001) in patients with DIH than patients without DIH (50.40±3.21 versus 61.63±5.08) [Table 5].
Before the treatment period, SOD level was also nearly similar (P=0.37) in both patients with and without DIH (84.44±14.02 versus 82.14±13.91). Whereas, during the treatment period, the level of SOD was significantly higher (P<0.0001) in patients with DIH as compared to patients without DIH (103.15±9.70 versus 85.26±13.78) [Table 5].
Due to wide spread application of DOTS strategy in TB treatment, it is imperative to evaluate their risk on human health. Hepatotoxicity is one of the potential risks related to anti-TB drugs. It is estimated that India has the highest burden of TB of any county in the world. Also, Indian populations are highly exposed to the risk of a potentially serious hepatotoxic effect of anti-TB drugs. Therefore, anti-TB-DIH is very important issue in India and the knowledge of risk factors for anti-TB-DIH, prior to medication, may greatly enhance TB prevention programs.
In our study, the prevalence of anti-TB-DIH was observed in 14.3% (35/244) patients, which was higher as compared to other developed countries ranging from 3%-4%.  The time interval from the initiation and the onset of hepatotoxicity was 15-50 days in which, patients manifested as anorexia, nausea, vomiting, abdominal pain and jaundice. Makhlouf et al. (2008)  also observed the similar clinical manifestations of anti-TB-DIH in patients and reported time interval of 15-60 days.
In this study, age was not a significant risk factor for anti-TB-DIH. Similarly, other studies, , in agreement with ours, showed no increased risk for the development of anti-TB-DIH regarding age. Our study clearly showed that hepatotoxicity induced by anti-TB drugs was more frequent in females as compared to males. Several studies reported that the prevalence of anti-TB-DIH was higher in female gender. , It has been suggested that slow acetylator are more prone to hepatotoxicity compared to rapid acetylator and females being a slow acetylator, are at higher risk for anti-TB-DIH.  We observed that malnutrition (detected by BMI <18.5 kg/m 2 ) was a significant risk factor for anti-TB-DIH. In a study done in India, incidence of Anti-TB-DIH was found to be three times higher in malnourished patients.  In our study, patients with extrapulmonary cases were found to be more susceptible for anti-TB-DIH. This was also reported in another study,  in which extrapulmonary TB patients were more affected with anti-TB-DIH compared with pulmonary TB patients. Similar to the result of one previous study,  we found that HIV positive patients were at increased risk for anti-TB-DIH, as compared to HIV negative patients. The decrement in the immune status of the HIV positive patients may be one of the reasons for higher risk of anti-TB-DIH. 
Oxidative stress plays an influential role in the pathogenesis of anti-TB-DIH, which is characterized by increased level of MDA with altered level of antioxidants.  Among the first line anti-TB drugs, metabolic intermediates of isoniazid metabolism are mainly responsible for the occurrence of anti-TB-DIH. These metabolic intermediates cause hepatocyte damage or necrosis through lipid peroxidation. In response to detoxify the toxic effects of these hepatotoxic metabolites, a large amount of glutathione is consumed, resulting in compromised anti-oxidative capacity. 
In this study, before the initiation of drug treatment, there was no significant alteration in the level of MDA, between patients with and without DIH. However, during the treatment period, we observed that MDA level was significantly higher in patients with DIH as compared to patients without DIH. Higher level of MDA (Lipid peroxidation product) indicates increased rate of oxidative stress in patients developing DIH. Mandai and Das (2005)  also suggested that oxidative stress with increased lipid peroxidation products are involved in toxic and drug-induced liver injury. Similarly, before the initiation of drug therapy, no significant alteration was to be observed in the level of glutathione between patients with and without DIH. Whereas, significant lower level of glutathione was to be observed after the initiation of anti-TB drug therapy in patients with DIH as compared to patients without DIH. Others studies, in agreement to ours, suggested that the level of glutathione level was to be decrease in rats after the administration of isoniazid or hydrazine. , Similarly, Chowdhury et al. (2006)  observed that the isoniazid and rifampicin induced hepatotoxicity was associated with decreased level of glutathione and increased rate of MDA production. Thus, the decreased level of GSH can be as a consequence of anti-TB-DIH. SOD is an antioxidant enzyme, extensively used as a biochemical indicator of oxidative stress. In this present study, before the treatment period, when we compared between patients with and without DIH, with special reference to SOD level, no significant difference was observed. However, during the treatment period, the activity of SOD was significantly higher in patients with DIH as compared to patients without DIH. In other previous studies, , reduced concentration of SOD was observed after the administration of isoniazid or hydrazine in rats. SOD catalyzes the dismutation of highly toxic superoxide radicals to H 2 O 2 . Increased generation of superoxide radicals lead to increased rate of lipid peroxidation. Thus, SOD has a protective role against free radial induced damage, and their induction can be understood as an adaptive response to oxidative stress.
In conclusion, our study indicates that TB patients with female gender, low body mass index, extrapulmonary and positive HIV are at higher risk for anti-TB-DIH. Further, increased lipid peroxidation products with altered profile of antioxidants, in patients with DIH, indicated that anti-TB-DIH appeared to be mediated through oxidative stress. Thus, there is a need for a regular biochemical and clinical follow-up for those patients who are at higher risk. Close monitoring through liver function and oxidative stress parameters might be helpful in the early diagnosis of anti-TB-DIH, which is crucial to prevent progression of severe liver injury.
Source of Support: None, Conflict of Interest: None
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]