Abstract
The Centers for Disease Control and Prevention (CDC) has previously reported surveillance data of severe liver injury in patients treated for latent tuberculosis infection (LTBI) with a daily and twice-weekly 2-month regimen of rifampin with pyrazinamide (RZ). On the basis of these initial reports, the CDC cautioned clinicians in the use of this therapy with advised additional monitoring. To estimate the incidence of RZ-associated severe liver injury and provide more precise data to guide treatment for LTBI, the CDC collected data from cohorts of patients in the US who received RZ for the treatment of LTBI during January 2000–June 2002 and for whom data were reported to the CDC through June 6, 2003. This report summarizes the analysis, which found high rates of hospitalization and death from liver injury associated with the use of RZ. On the basis of these findings, the American Thoracic Society (ATS) and the CDC now recommend that this regimen should generally not be offered to persons with LTBI. The revised ATS/CDC recommendations described in this report have been endorsed by the Infectious Diseases Society of America (IDSA). Clinicians are advised to use the recommended alternative regimens for the treatment of LTBI. Rifampin and pyrazinamide should continue to be administered in multidrug regimens for the treatment of persons with active tuberculosis (TB) disease.
A 2-phase retrospective survey was conducted to estimate the incidence of severe liver injury among persons receiving RZ for treatment of LTBI. In December 2001 (phase I), the CDC sent a questionnaire by E-mail to TB control programs in 12 large cities and all 50 states asking them to identify programs and healthcare providers prescribing RZ for treatment of LTBI. All controllers responded and, in February 2002, the CDC staff called the programs and healthcare providers identified to confirm its use. In September 2002 (phase II), the CDC mailed a second questionnaire to the 150 healthcare providers identified during the first phase, requesting aggregate cohort data for January 2000–June 2002; 109 (78%) healthcare providers responded by June 6, 2003.
Of 7737 patients who were reported to have started RZ for treatment of LTBI during the survey period, 5980 received daily doses and 1757 received twice-weekly doses. A total of 204 patients discontinued RZ because of aspartate aminotransferase concentrations >5 times the upper limit of normal (26.4/1000 treatment initiations). An additional 146 patients discontinued RZ because of symptoms of hepatitis (18.9/1000 treatment initiations).
Of the 48 cases of severe liver injury reported to the CDC through passive surveillance, 30 also were detected in the second phase of the survey. Based on these 30 cases, the estimated rates of hospitalization and death during the survey period were 3.0 per 1000 treatment initiations.
The CDC cohort analysis found that the rates of severe liver injury and death related to the use of RZ are higher than the rates for isoniazid (INH)-associated liver injury in the treatment of LTBI. Although initial studies attributed hospitalization rates as high as 5.0 per 1000 treatment initiations and mortality rates as high as 1.0 per 1000 to INH, studies conducted since 1991 involving more than 1 million persons treated with INH have reported hospitalization rates of 0.1–0.2 and mortality rates of 0–0.3 per 1000. This decrease from earlier studies might reflect careful selection of patients and active monitoring for early signs of adverse events. In addition to the survey on the use of RZ described in this report, recent studies have reported episodes of liver injury and hospitalization associated with RZ for treatment of LTBI, including the need for transplantation in 1 patient. Among first-line agents in the treatment of active TB disease, pyrazinamide might be the most hepatotoxic.
The ATS and CDC now recommend that this regimen should generally not be offered to persons with LTBI who are HIV-negative or HIV-in-fected. Based on of the investigation of potential cofactors in the 48 patients with serious liver injury, this regimen should never be offered to patients who (1) are concurrently taking other medications associated with liver injury, (2) drink excessive amounts of alcohol, even if alcohol use is discontinued during treatment, (3) have underlying liver disease, or (4) have a history of INH-associated liver injury.
If the potential benefits of this regimen outweigh the risk for severe liver injury and death associated with it, use of RZ might be considered in carefully selected patients, but only if (1) the preferred or alternative regimens (i.e., 9 mo of daily or biweekly INH, 6 mo of daily or biweekly INH, or 4 mo of daily rifampin) are judged not likely to be completed and (2) oversight by a clinician with expertise in the treatment of LTBI can be provided. A TB/LTBI expert should be consulted before RZ is offered. In addition, patients should be asked whether they have had liver disease or adverse effects from taking INH or other drugs, informed of potential hepatotoxicity of the RZ regimen, and advised against the concurrent use of potentially hepatotoxic drugs, including over-the-counter drugs such as acetaminophen.
To facilitate periodic clinical assessments of persons taking an RZ regimen, clinicians should dispense no more than a 2-week supply, with a daily pyrazinamide dose of <20.0 mg/kg/d (maximum 2.0 g) and a twice-weekly dose of <50.0 mg/kg/d (maximum 4.0 g). Patients should be reassessed by a healthcare provider after 2, 4, 6, and 8 weeks of treatment for adherence, tolerance, and adverse effects. The 8-week assessment also should be used to document treatment completion.
For persons taking this regimen, serum aminotransaminases (AT) and bilirubin should be measured at baseline and after 2, 4, 6, and 8 weeks of treatment. Because the majority of these patients had onset of symptoms of liver injury after the fourth week of therapy, patients should be monitored throughout the entire course of treatment. Use of RZ should be discontinued immediately and not resumed for any of the following findings: (1) AT >5 times the upper limit of normal range in an asymptomatic person, (2) AT greater than normal range when accompanied by symptoms of hepatitis, or (3) serum bilirubin concentration greater than the normal range, whether or not symptoms are present.
The recommendations against the use of RZ for treatment of LTBI described in this report do not apply to the appropriate use of rifampin and pyrazinamide in multidrug regimens for the treatment of persons with active TB disease. In these circumstances, the risk for morbidity and mortality from TB disease is substantially greater than with LTBI. Rifampin and pyrazinamide are essential components of recommended ATS/CDC/IDSA regimens that render patients noninfectious rapidly and are effective in curing patients with drug-susceptible Mycobacterium tuberculosis strains within 6 months. (MMWR 2003;52:735-9)