© 2002 American Thoracic Society
Evaluation of Clinical and Immunogenetic Risk Factors for the Development of Hepatotoxicity during Antituberculosis TreatmentDepartments of Medicine, Histocompatibility and Immunogenetics, and Biostatistics, All India Institute of Medical Sciences, New Delhi, India Correspondence and requests for reprints should be addressed to S. K Sharma, Professor, Department of Medicine, All India Institute of Medical Sciences, New Delhi-110029, India. E-mail: surensk{at}hotmail.com
Though several risk factors for the development of hepatotoxicity due to antituberculosis drugs have been suggested, involvement of genetic factors is not fully established. We have studied the major histocompatibility complex (MHC) class II alleles and clinical risk factors for the development of hepatotoxicity in 346 North Indian patients with tuberculosis undergoing antituberculosis treatment. Of these, 56 patients developed drug-induced hepatotoxicity (DIH group), whereas the remaining 290 patients did not (non-DIH group). The DIH group was comparatively older, had lower pretreatment serum albumin, and a higher frequency of moderately/far advanced disease radiographically than the latter. Further, patients with high alcohol intake had threefold higher odds of developing hepatotoxicity. In multivariate logistic regression analysis, older age (odds ratio [OR] 1.2), moderately/far advanced disease (OR 2.0), serum albumin < 3.5 g/dl (OR 2.3), absence of HLA-DQA1*0102 (OR 4.0), and presence of HLA-DQB1*0201 (OR 1.9) were independent risk factors for DIH. Our results suggest that the risk of hepatotoxicity from antituberculosis drugs is influenced by clinical and genetic factors.
Key Words: antituberculosis drugs • hepatotoxicity • human leukocyte antigens
It is estimated that one third of the world's population is infected with Mycobacterium tuberculosis, resulting in 8.4 million new tuberculosis (TB) cases in 1999 (1). Antituberculosis therapy with rifampicin, isoniazid, pyrazinamide, and ethambutol/streptomycin is very effective, but the first three drugs are hepatotoxic. A higher risk of hepatotoxicity has been reported in Indian patients (2–5) than in their Western counterparts (6–8). The risk of hepatotoxicity, based on data from four prospective Indian studies, was 11.5% (95% confidence interval [CI]: 9.5–13.5), compared with 4.3% (95% CI: 3.4–5.3) in 14 published studies from the West (9). The reasons for this higher rate of hepatotoxicity in Indian patients are unclear. Reported risk factors for hepatotoxicity include: older age (10), female sex (11), poor nutritional status (12), high alcohol intake (11), pre-existing liver disease (13), hepatitis B carriage (14), increased prevalence of viral hepatitis in developing countries (2, 15), hypoalbuminaemia and advanced tuberculosis (16), and inappropriate use of drugs and acetylator status (17, 18). Convincing evidence has been presented for the genetic association of human leukocyte antigen (HLA)-DR2 with pulmonary tuberculosis in various populations (19–22) and an association of NRAMP1 gene variants (23) with susceptibility to tuberculosis. However, there are no studies on the role of immunogenetic factors in the development of hepatotoxicity due to antituberculosis drugs. The present study was therefore designed to determine the role of various risk factors, including major histocompatibility complex (MHC) class II alleles, in the development of hepatotoxicity in Indian patients receiving short-course antituberculosis chemotherapy. Identification of patients with risk factors will facilitate monitoring for hepatotoxicity, which results in mortality of 6–12% if the drugs are continued even after the onset of symptoms (24).
Study Population We evaluated 361 patients who presented to the Outpatient Department or admitted at the All India Institute of Medical Sciences hospital, New Delhi, between 1996 and 2000. Fifteen patients with chronic illnesses such as cirrhosis of liver, chronic hepatitis, acute viral hepatitis, and/or gastrointestinal, renal, or cardiac diseases, were excluded from the study, leaving a study population of 346 patients. All patients were treated for the full duration of antituberculosis chemotherapy with regular follow-up. Of these patients, 56 developed drug-induced hepatotoxicity (DIH group) whereas the remaining 290 patients had no clinical or biochemical evidence of hepatotoxicity (non-DIH group). The HLA data were compared with those of 275 healthy control subjects of the same ethnic background and similar socioeconomic status as the patients. The diagnosis of pulmonary TB was based on the presence of acid-fast bacilli on sputum smear or M. tuberculosis on sputum culture. In patients with smear-positive TB, culture was not done unless multidrug-resistant TB was strongly suspected. Sputum cultures were done in all patients with smear-negative pulmonary TB. In patients with negative smears and cultures, the diagnosis of TB was based on symptoms, chest radiographic infiltrates in the upper lobes, and clinical and radiographic response to antituberculosis drugs. Focal disseminated TB was diagnosed on the basis of histopathologic and/or microbiologic evidence of TB from two noncontiguous sites. A group of patients with disseminated TB exhibiting classic miliary mottling on chest radiographs was included as "miliary TB" as described earlier (25). Lymph node TB was diagnosed by the demonstration of M. tuberculosis in the smear and/or culture of lymph node aspirates or biopsy specimens. The DIH group included 13 patients with sputum smear-positive pulmonary TB, 10 with sputum smear-negative pulmonary TB (2 with positive cultures), 8 with miliary TB, 19 with focal disseminated TB, and 6 with lymph node TB. The non-DIH group included 126 with sputum smear-positive pulmonary TB, 39 with sputum smear-negative pulmonary TB (8 with positive cultures), 41 with miliary TB, 42 with focal disseminated TB, and 42 with lymph node TB (see online data supplement for drug regimens and dosages).
Diagnosis of Drug-induced Hepatotoxicity
Laboratory Monitoring
DNA Typing for HLA Class II Alleles
Statistical Analysis
All patients included in the study had negative serologic tests for hepatitis A, B, C, and E, and for HIV. Patients in the DIH group were significantly older than those in the non-DIH group, but other parameters, including the male/female ratio, measures of nutritional status and erythrocyte sedimentation rate were similar in both groups (Table 1) . None of the DIH group patients had positive tests for rheumatoid factors or other autoantibodies. Except serum albumin levels, pretreatment liver function test results were similar in both groups (Table 2) . The maximally abnormal values of serum bilirubin, AST, ALT, alkaline phosphatase, total protein, and albumin levels were significantly higher in the DIH than in the non-DIH group. The median time from the initiation of antituberculosis chemotherapy until the development of hepatotoxicity (latency period) was 4 weeks (range, 1–72). None of the patients had recurrence of hepatotoxicity with the reintroduction of antituberculosis drug therapy.
Patients in the DIH group were more likely to have a high alcohol intake (5% versus 2% in the non-DIH group) and moderately/far advanced disease (50% versus 33% in the non-DIH group), but these differences were not statistically significant. A pretreatment serum albumin of less than 3.5 g/dl was present in 32% of DIH patients, compared with 16% of non-DIH group (p < 0.01). No mortality was observed in either group.
HLA Analysis Haplotype analysis revealed that DQB1*0201-associated DRB1*0301 and DRB1*0701 haplotypes occurred more frequently in the DIH patients than the non-DIH group (see Table E3 in the online data supplement). Of these, the latter haplotype, DRB1*0701-DQA1*0201-DQB1*0201, was present in 17% of the DIH group but in only 8% of the non-DIH group (p = 0.01). The other DR7 haplotype (DRB1*0701-DQA1* 0201-DQB1*0303) was distributed equally in the two groups. In the bivariate analysis, DIH patients with advancing age and presence of specific HLA alleles like DRB1*01, DRB1*03, and DRB1*07 were positively associated with development of DIH. However, in multivariate logistic regression analysis, when the pretreatment variables and HLA alleles were simultaneously considered in the model, advancing age, moderately/far advanced disease, hypoalbuminaemia, presence of HLA-DQB1*0201, and absence of HLA-DQA1*0102 were found to be significant risk factors for the development of DIH (Table 3) .
Identification of patients at increased risk for DIH is important because hepatotoxicity causes significant morbidity and mortality and modification of the drug regimen may be required. The incidence of DIH has been reported to be higher in developing countries, and factors such as acute or chronic liver disease, indiscriminate use of drugs, malnutrition, and more advanced tuberculosis have been implicated (5, 16, 26). However, it is unclear if these are independent risk factors for hepatotoxicity. A high incidence of viral hepatitis has been reported in TB patients in developing countries (15, 27), resulting in misdiagnosis of DIH if serologic tests are not performed. All patients with positive serologic tests for hepatitis A, B, C, and E were excluded from the current study. All of our 56 patients with DIH recovered from hepatotoxicity, and antituberculosis medications were reintroduced one by one in gradually increasing doses, with careful monitoring of liver function. No patient developed recurrent hepatotoxicity, a finding that is at variance with experience of others (28). The reasons for this difference are uncertain. We found that advancing age was an independent predictor of DIH, consistent with previous reports (10), but did not confirm previous studies suggesting that females are more likely to have DIH (11, 14). Malnourished children have threefold increased incidence of DIH (5) and we found that patients with pretreatment hypoalbuminaemia had a twofold higher risk of developing DIH. Other measures of malnutrition, such as BMI and triceps skin fold thickness, were not predictors of DIH. We found that high alcohol intake and advanced tuberculosis were associated with DIH, confirming previous reports (11, 16). Of these, moderately/far advanced pulmonary TB was an independent predictor of DIH. Specific HLA alleles have been associated with development of pulmonary tuberculosis in various populations (19–22) and with pulmonary Mycobacterium avium complex infection among the Japanese (29). We therefore investigated the role of these genes in the development of DIH in TB patients. PCR-SSO hybridization techniques were employed because they are more sensitive than serology and define alleles with high resolution. Multivariate analysis revealed that the two HLA-DQ alleles were associated with development of DIH.
The HLA-DQ molecules are unique among class II MHC molecules because both the It is interesting to note that although DRB1*15/16 (DR2) is associated with increased susceptibility to the development and progression of pulmonary TB, these alleles occur with a reduced frequency in patients developing hepatotoxicity due to short-course chemotherapy. A similar corollary exists in rheumatoid arthritis, which is associated with expression of DR4. In contrast, patients with rheumatoid arthritis who developed D-pencillamine–induced polymyositis and nephropathy were DR4-negative and DR2- and DR3-positive (30, 31). Our results suggest that a genetic influence associated with the MHC class II region, particularly the DQ locus, plays an important role in the development of DIH. In summary, we found that tuberculosis patients with advancing age, moderately/far advanced disease, pretreatment hypoalbuminaemia, and the presence/absence of specific HLA-DQ alleles are associated with development of hepatotoxicity. It will be important to evaluate those risk factors in other populations.
The authors wish to thank Ms. Yogita Dixit and Mr. Mukesh Singh for their help in the conduct of the study. Supported by the Department of Biotechnology (DBT), Ministry of Science and Technology, Government of India (Grant No. BT/MB/05/005/HG/96).
This article has an online data supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form August 21, 2001; accepted in final form May 9, 2002
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
- and ß-chains are highly polymorphic and most of the variable amino acid residues are located on the 
