| |
Isoniazid (TB treatment) Hepatotoxicity among Drug Users: The Role of Hepatitis C
|
| |
| |
Treatment of latent tuberculosis infection (LTI) in high-risk patients, such as those in contact with tuberculosis, HIV-seropositive patients, convicts, the homeless, or drug users, is an important part of the strategies required to eliminate tuberculosis from developed countries. Drug users deserve special attention because they are often included in several of the aforementioned groups. The treatment of LTI with isoniazid among these patients while they are enrolled in methadone maintenance programs has proved to be a cost-effective approach. However, one of the main problems with both isoniazid and the combination of rifampin and pyrazinamide is the risk of hepatotoxicity.
The efficacy of LTI treatment as a strategy in special-risk groups depends on the risk of developing active tuberculosis disease, the efficacy of the drugs used, and adherence to the treatment regimen. On an individual level, it depends on the relation between the gains and the risks associated with administering a drug to a patient without the disease.
Drug users have a higher risk of tuberculosis than does the general population, although their risk is lower than that for other groups, such as close contacts of persons with tuberculosis, HIV-infected persons, with silicotic patients, or patients with radiographic findings consistent with prior tuberculosis. Although there is consensus about the convenience of treating the LTI in drug users undergoing disintoxication programs, the risk of isoniazid-associated hepatotoxicity in this group has not been completely studied. The factors that have been related to isoniazid-associated hepatotoxicity are age, alcohol abuse, malnutrition, and liver diseases, including chronic viral hepatitis.
Drug users frequently have multiple risk factors for hepatotoxicity, including chronic infections with hepatitis B or C viruses and excessive alcohol intake, in addition to drug use. The safety of isoniazid use in patients with these risk factors has not been completely studied, and, although several studies showed an incidence of hepatotoxicity higher than that described for the general population, the predisposing factors have not been well defined. The objective of this study was to evaluate the safety of the treatment of LTI with isoniazid in a cohort of former drug users and to study which factors may predict the development of liver complications.
Summary: The incidence of and risk factors associated with hepatotoxicity in patients with chronic hepatitis have not been systematically studied. Therefore, we conducted a prospective study that included former drug users who were treated with isoniazid for latent tuberculosis infection. Of 415 patients, 20 (4.8%; 95% confidence interval [CI], 37.4) had hepatotoxicity diagnosed, and 6 (1.4%; 95% CI, 0.53.2) developed clinical hepatitis, none of whom had serious symptomsÉ. In this study, isoniazid had to be withdrawn from 4.8% of the patients as a result of liver toxicity...Of the factors studied only 2 factors were independently associated with isoniazid hepatotoxicity were excessive alcohol consumption (odds ratio [OR]; 4.2, 95% CI, 1.610.8; P = .002) and a high baseline alanine transaminase level (OR, 4.3; 95% CI, 1.611.4; P = .002). The presence of hepatitis C virus antibodies was associated with hepatotoxicity only on univariate analysis. Treatment with isoniazid in drug users appears to be safe and well tolerated, although frequent asymptomatic elevations in transaminase levels were observed.
"...The toxicity rate was significantly higher among patients who were antibody-HCV positive and who had increased baseline ALT levels compared with the other groups of patients (4 patient groups by whether they had increased baseline ALT and/or HCV-positivity). No cases of hepatotoxicity were observed among anti-HCV negative patients with elevated baseline ALT levels. The percentage of excessive alcohol drinkers in these 4 groups did not differ significantly (15%-20.4%)"... The transaminase elevations usually occur during the first 3 months of treatment, and the role of isoniazid as a causative agent was not at all clear...
... In the present study, among the 214 anti-HCV positive patients, 46% had normal baseline ALT levels, and, in 54%, ALT levels were elevated. The toxicity frequency was significantly higher in the latter (3% vs. 11.4%; P = .02). A normal transaminase value generally implies milder liver disease. This higher risk of hepatotoxicity could be explained by an increase in toxicity among patients with more active hepatitis. For this reason, anti-HCV positive patients with normal transaminase levels do not require closer observation than do anti-HCV negative patients.
Editorial note from Jules Levin: normal ALT does not necessarily imply milder liver disease. ItÕs estimated that 5-12% of HCV infected patients have normal ALT and more advanced liver disease, and in HIV/HCV coinfected a higher percentage of individuals with normal ALT can have fibrosis and more advanced liver disease.
Clinical Infectious Diseases 2003; 36: 293-298
Alberto Fernandez-Villar,1, 2 Bernardo Sopena,3 Rafael Vazquez,1,2 Fernando Ulloa,5 Enrique Fluiters,4 Mar Mosteiro,2 Cesar Mart’nez-Vazquez,3 and Luis Pineiro2
1Unidad de Tuberculosis, Servicios de 2Neumolog’a y 3Medicina Interna, 4Unidad de Investigacion, Hospital Xeral-C’es, and 5Servicio de Analisis Cl’nicos, Hospital Nicolas Pena, Vigo, Spain
A prospective study was performed for the period of November 1997 through December 2001. The study population was composed of former drug users observed at the Drug Users Attention Units of our city (Vigo, Spain). The patients were included either in a methadone or naltrexone maintenance program or in a drug-free program. They were referred systematically to our tuberculosis consult, where a clinical history was obtained and the tuberculin skin test and hematological and biochemical analysis were performed. Serological tests for detection of hepatitis C virus (HCV), hepatitis B virus, and HIV, were performed with use of a third-generation ELISA method. In accordance with Spanish guidelines, chest radiography was performed and sputum samples were obtained from patients who had a 5-mm tuberculin test result.
After these procedures were performed, patients with LTI who completed 7 days of isoniazid therapy were included. Exclusion criteria were as follows: HIV positivity, evidence of active tuberculosis, a history of isoniazid-associated hepatotoxicity, previous receipt of correct treatment of LTI or tuberculosis disease, and elevated baseline aspartate aminotransferase (AST) or alanine aminotransferase (ALT) levels of >3 times the upper limit of normal (ULN; normal level, 40 IU/L).
With the patient's consent, isoniazid (300 mg) and pyridoxine (50 mg) were administered daily for 6 months. The medications were administered under direct supervision along with methadone. For patients who were not in a methadone maintenance program, compliance was ensured by ensuring family supervision, counting the pills, and detecting isoniazid metabolites in urine specimens. Excessive alcohol intake was defined as a mean weekly intake of >280 g of ethanol for men and >140 g of ethanol for women for 1 year. The ALT level was used to correlate the baseline liver function test values with the toxicity frequency. The patients visited the clinic monthly until the end of treatment (month 6), and analytical controls (hemograms and liver function tests) were performed at the end of the first, second, and fourth month. Consults and analysis were scheduled more frequently when hepatotoxicity was suspected.
Criteria for the diagnosis of hepatotoxicity and the withdrawal of isoniazid therapy were as follows. (1) Isoniazid was withdrawn when symptoms of hepatitis with AST and/or ALT value elevations were >5 times the ULN. In asymptomatic patients, AST and/or ALT value elevations of >5 times the ULN in 2 consecutive weekly determinations or a unique elevation of >250 IU/L were required. (2) After the drug's withdrawal, transaminase values should return to the baseline level in <3 weeks. (3) Exclusion of other causes was ensured by performing new serological testing for hepatitis A, B, and C viruses (if previously negative) or ingestion of other drugs. Informed consent was obtained from all participants before enrollment in this study.
RESULTS
Description of the study population. Among the 541 patients with a positive tuberculin skin test result, 126 individuals were excluded. Therefore, 415 patients fulfilled the inclusion criteria and were included in the study. The duration of treatment was 154.1 ± 51.4 days (range, 10180 days). Three hundred sixty-three patients (87.5%) were men, and the mean age was 31.3 ± 5.5 years (range, 1749 years). Three hundred thirteen subjects (75.4%) were included in a methadone maintenance program, 74 (17.8%) were in a drug-free program, and 28 (6.7%) were receiving naltrexone treatment (50100 mg q.d.).
The mean ALT level before the commencement of treatment was 36 ± 27 IU/L (range, 6119 IU/L), and the mean AST level was 33 ± 18 IU/L (range, 8105 IU/L). One hundred thirty-three patients (32%) had an ALT level of >40 IU/L. Sixty-five (16.1%) of 403 patients were taking other psychiatric medication daily, mainly benzodiazepines. In 403 patients (97.1%), the ingestion of alcohol was calculated, and it was considered excessive for 73 patients (17.6%). The mean baseline ALT and AST values were 37 ± 25 IU/L and 36 ± 21 IU/L, respectively, in alcohol drinkers and 36 ± 27 IU/L and 32 ± 17 IU/L, respectively, in patients who did not drink alcohol (P = .3). In 262 patients, the body mass index was known; for 26 patients (9.9%), it was 20 kg/m2.
Two hundred fourteen (51.6%) of 415 patients had HCV antibodies detected, and 8 (1.9%) were hepatitis B surface antibody positive. The mean baseline ALT and AST levels were 48 ± 29 IU/L and 41 ± 21 IU/L, respectively, in anti-HCV positive patients, and 30 ± 14 IU/L and 24 ± 8 IU/L, respectively, in anti-HCV negative patients (P < .0001).
Transaminase elevations and isoniazid hepatotoxicity. In 34 patients (8.2%), elevations in the ALT and/or AST values to >3 times ULN were detected. None of these patients met the hepatotoxicity criteria or had symptoms of hepatitis, and the transaminase values returned to baseline values even though isoniazid therapy was maintained. Eleven patients (32.3%) were heavy drinkers.
Twenty (4.8%) of 415 patients (95% CI, 37.4) developed isoniazid-associated hepatotoxicity and isoniazid was withdrawn. The ALT, AST, and bilirubin peak values reached in these subjects were 339 ± 201 IU/L (range, 207948 IU/L), 184 ± 114 IU/L (range, 73482 IU/L), and 0.8 ± 0.3 mg/dL (range, 0.11.6 mg/dL), respectively. Nineteen (95%) of these 20 patients were men, with a mean age of 30.3 ± 6.6 years. Only 6 patients (1.4%; 95% CI, 0.53.2) had symptoms of hepatitis, such as asthenia, nausea, or mild abdominal pain. None of them presented with jaundice or other manifestations that would imply serious hepatitis. The time until the diagnosis of toxicity was 61.3 ± 28 days (range, 21120 days). In 17 (85%) of 20 patients, the diagnosis was made in the first 3 months, and in only 3 patients was the diagnosis made thereafter (at the end of the fourth month).
On univariate analysis, the factors significantly associated with a higher risk of hepatotoxicity were the elevated ingestion of alcohol, the presence of anti-HCV antibodies in serum, and elevated baseline transaminase values. On multivariate analysis, only the excessive ingestion of alcohol (OR, 4.2; 95% CI, 1.610.8; P = .002) and an abnormal baseline ALT level (OR, 4.3; 95% CI, 1.611.4; P = .002) proved to be independent risk factors for the development of hepatotoxicity. The only factor significantly associated with symptomatic isoniazid hepatotoxicity was excessive alcohol intake (OR, 9.5; 95% CI, 1.752; P = .01).
The toxicity rate was significantly higher among patients who were antibody-HCV positive and who had increased baseline ALT levels compared with the other groups of patients (4 patient groups by whether they had increased baseline ALT and/or HCV-positivity). No cases of hepatotoxicity were observed among anti-HCV negative patients with elevated baseline ALT levels. The percentage of excessive alcohol drinkers in these 4 groups did not differ significantly (15%-20.4%).
DISCUSSION
The efficacy of LTI treatment as a strategy in special-risk groups depends on the risk of developing active tuberculosis disease, the efficacy of the drugs used, and adherence to the treatment regimen. On an individual level, it depends on the relation between the gains and the risks associated with administering a drug to a patient without the disease. Drug users have a higher risk of tuberculosis than does the general population, although their risk is lower than that for other groups, such as close contacts of persons with tuberculosis, HIV-infected persons, with silicotic patients, or patients with radiographic findings consistent with prior tuberculosis. Although there is consensus about the convenience of treating the LTI in drug users undergoing disintoxication programs, the risk of isoniazid-associated hepatotoxicity in this group has not been completely studied. The factors that have been related to isoniazid-associated hepatotoxicity are age, alcohol abuse, malnutrition, and liver diseases, including chronic viral hepatitis.
In this study, isoniazid had to be withdrawn from 4.8% of the patients as a result of liver toxicity. Nearly one-third of our patients had elevated serum transaminase levels before they began receiving treatment, 17.6% of them were excessive alcohol drinkers, and 51.6% had detectable HCV antibodies. This prevalence of HCV infection is lower than has previously been reported, because only 44% of our patients were parenteral drug users (data not shown).
The definition of hepatotoxicity in patients with previous hepatic diseases is much disputed, because it is difficult to establish the influence of the natural evolution of the underlying disease. Although it is generally recommended that the isoniazid therapy be interrupted when transaminase levels increase to >35 times the ULN, this limit has not been defined in patients whose values were already elevated before they began receiving treatment. Schenker et al. reported that elevations in the ALT and/or AST levels to >50100 UI/L more than the baseline levels might define toxicity. If these interruption limits had been applied in our study, 64 (15.4%) of 415 patients would have had hepatotoxicity diagnosed. Nevertheless, only 6 patients (1.4%) had symptoms of liver toxicity, and the peak transaminase level that was reached was much lower than that reported in a recent study that involved the general population, in which the monitoring of the disease was clinical. In a previous study, it has been shown that "biochemical hepatitis" is 6 times more common than "clinical hepatitis". Performance of periodic analytical controls, which, at present, are only recommended in patients with risk factors, can contribute to overestimating the actual frequency of hepatotoxicity when it is diagnosed in asymptomatic patients with temporary transaminase value elevations produced by use of isoniazid itself, the ingestion of alcohol, the patient's underlying disease, or, more likely, the synergistic action of them all. It is reasonable to think that, in patients with a previous liver alteration, the early detection of the drug's toxicity might lessen the severity of the liver damage. From a practical point of view, when dealing with the treatment of an LTI, the approach should be similar, leading to the drug's withdrawal and careful evaluation as to whether it should be readministered.
In this study, as in those that have involved the general population, most of the patients received diagnoses in the first 3 months of the treatment, when the possibility of this complication must be especially considered. Although our patients were observed until the end of the treatment, late biochemical hepatotoxicity reactions could not be excluded, because the last analytical control was performed at the fourth month.
Of the factors studied, only excessive ingestion of alcohol and previous alteration of ALT values were independently associated with a higher risk of hepatotoxicity. The presence of HCV antibodies was only associated with higher risk on univariate analysis. Ungo et al. reported that HCV-associated chronic infection increases the risk of toxicity 5 times in patients treated for tuberculosis disease with isoniazid, rifampin, and pyrazinamide. Recently, in a cohort of drug users treated with isoniazid for LTI, 95% of whom were HCV seropositive, Sadaphal et al. found hepatotoxicity (ALT level of >5 times ULN) in 9% of patients. Because only 6 patients were anti-HCV negative, and because none of them had hepatotoxicity, a causative role of HCV in isoniazid hepatotoxicity could not be proved. To our knowledge, the present report is the first to study the safety of isoniazid in a large number of anti-HCVpositive patients compared with patients with similar ages and habits but who were anti-HCV negative.
Chronic hepatitis C has a variable clinical spectrum, and the rate of chronic hepatitis C is elevated in parenteral drug users. The natural history of this infection comprises an oscillating course with phases of diverse clinical activity, often remaining in an asymptomatic phase and showing normal transaminase values. Inglesby et al. showed that, in a cohort of HCV-positive drug users, 42% presented with persistently normal ALT values, 15% had permanently elevated values, and 43% experienced intermittent elevations. In the present study, among the 214 anti-HCVpositive patients, 46% had normal baseline ALT levels, and, in 54%, ALT levels were elevated. The toxicity frequency was significantly higher in the latter (3% vs. 11.4%; P = .02). A normal transaminase value generally implies milder liver disease. This higher risk of hepatotoxicity could be explained by an increase in toxicity among patients with more active hepatitis. For this reason, anti-HCVpositive patients with normal transaminase levels do not require closer observation than do anti-HCVnegative patients. We have not performed confirmatory tests with determination of the HCV RNA level, so it is possible that some of the anti-HCVpositive patients, particularly those with normal baseline transaminase values, do not have active chronic liver disease.
The other factor associated with the development of hepatotoxicity was the excessive ingestion of alcohol. In a previous study conducted among the general population, the hepatic toxicity of isoniazid in habitual drinkers was 4-fold the hepatic toxicity in others. The only factor associated with a higher risk of hepatotoxicity in the study of Sadaphal et al. was the excessive ingestion of alcohol. However, in another study, which involved patients receiving isoniazid and rifampin, there were no significant differences in the toxicity between alcoholic patients and patients who were not alcoholics. In the present study, transaminase values and anti-HCV status were similar among drinkers and nondrinkers. It has been described that the ingestion of alcohol may increase the severity of the HCV-associated infection; therefore, the use of isoniazid in these patients might have an additional effect. However, these results must be read cautiously, because the information about drinking habits may be potentially biased, especially in a group comprising drug users. None of the other studied factors were associated with the presence of hepatotoxicity. Age, which is one of the most accepted risk factors, was not correlated because of the narrow age range of the studied population.
In summary, the treatment of LTI with isoniazid in HIV-negative drug users seems to be safe and well tolerated. In this study, no cases of serious hepatotoxicity were observed, although frequent transaminase value elevations, generally asymptomatic or oligosymptomatic, were found. The transaminase elevations usually occur during the first 3 months of treatment, and the role of isoniazid as a causative agent was not at all clear. These findings are especially important among patients who drink alcohol excessively and those who had transaminase value alterations before they began to receive treatment, especially if they have HCV antibodies.
|
|
| |
| |
|
|
|
