Treatment of latent tuberculosis infection with isoniazid (INH) for nine months versus rifampin and pyrazinamide (RZ) for two months: A comparison of rate of liver injury in Baltimore City Chest Clinic during 1999--2001

For three decades, the antibiotic isoniazid (taken for at least six months) has been the recommended standard treatment for latent tuberculosis infection in the United States. In 1999, the combination of rifampin and pyrazinamide (taken together for two months) was introduced as an acceptable alternative treatment option for latent tuberculosis. However, in August of 2001, the CDC reported that rifampin and pyrazinamide were associated with 5 deaths from liver injury, and cautioned against the use of this regimen. Baltimore City Health Department had extensive use of rifampin and pyrazinamide for latent tuberculosis treatment from 1999 to 2001. The purpose of this study is to determine and compare the risk of liver injury among Baltimore City patients who were treated with the rifampin plus pyrazinamide regimen versus the standard isoniazid regimen.;The study reviewed medical records of patients treated from January 1999 to December 2001, in a public clinic in Baltimore City. Data for the study were extracted from the records of this public clinic. The number of patients found eligible for the study was 776. The number that received isonaizid was 444 or 37.8%. The number of patients that received rifampin and pyrazinamide was 322 or 35.7%. Logistic regression was used to estimate the risk of developing liver injury.;Three logistic regression models---one that estimated crude odd ratios, another that controlled for demographic variables, and another that controlled for both demographic and lifestyle variables---all showed statistically non-significant differences in the risk for liver injury between patients treated with isoniazid for 9 months and those treated with rifampin and pyrazinamide for two months. The risk was higher for the isoniazid group than for RZ, in all three models (OR: 3.7, 2.9, and 3.5 respectively).;However, our retrospective study has several important sources of potential bias, which limits its general izability, and which also limit our ability to make public health practice recommendations based on our study results. Further investigation of the risk factors for RZ-associated toxicity is needed. The apparent good tolerability of RZ in the population studied supports cautious use of RZ in carefully selected individuals.