Hepatotoxicity from pyrazinamide based anti-tuberculous therapy: protective effect from isoniazid and HIV infection?

In 1999 the American Thoracic Society (ATS), the Centers for Disease Control and Prevention (CDC), and the Infectious Diseases Society of America (IDSA) endorsed a 2 month daily regimen of rifampin and pyrazinamide (2RZ) for the treatment of latent tuberculosis (TB) infection. This recommendation followed a prospective randomised trial in HIV infected persons showing that 2RZ had similar efficacy and safety to a 12 month regimen of daily isoniazid (12H). These two papers examine the incidence of 2RZ related hepatotoxicity.

The first is a retrospective cohort study from the Netherlands which compares liver function tests (LFTs) and hepatotoxicity in a 2RZ cohort matched with a cohort who received 6 months treatment of latent infection with isoniazid (6H) and a group of patients with active TB on treatment containing at least rifampin, isoniazid and pyrazinamide (2RHZ+). They found that 2RZ caused more severe hepatotoxicity than 6H (odds ratio (OR) 2.61, 95% confidence intervals (CI) 1.26 to 5.39, p = 0.012) or 2RHZ+ (OR 2.61, CI 1.21 to 5.59, p = 0.016), with an even greater OR in patients <25 years or those receiving >30 mg/kg pyrazinamide. The difference in hepatotoxicity between 2RZ and 2RHZ+ is peculiar and may reflect higher alcohol consumption in the 2RZ group. Another possibility is that isoniazid may confer some protection against pyrazinamide induced hepatotoxicity. This study is limited by incomplete LFT data but, nonetheless, indicates that 2RZ should not be recommended without more evidence.

The second paper is a detailed LFT analysis from the original study from this group (

) and compared the use of 12H and 2RZ in HIV infected individuals. They found that 2RZ caused a smaller increase in AST (OR 0.74, CI 0.54 to 1.08, p = 0.055) but a more frequent rise in bilirubin by >0.5 mg/dl (OR 1.73, CI 1.24 to 2.43, p = 0.002) than 12H. The authors conclude that 2RZ is as safe as 12H in HIV infected patients, although there has historically been more hepatotoxicity with 2RZ than 12H in HIV negative populations. Further studies are required to confirm and explain this observation.