The relationship between the use of antiretroviral drugs and the development of hepatic abnormalities has been documented in both randomized controlled trials (RCTs) and observational database studies. Both types of study design are known to have limitations when addressing this issue. Whereas RCTs may enroll a population that is at lower risk for the development of hepatotoxicity, thus underestimating the possible effect of antiretroviral therapy on hepatic abnormalities, observational databases may encompass information from a more heterogeneous group of patients, allowing the drugs to be assessed in a more realistic situation. However, a number of possible biases associated with the use of observational data may limit the conclusions that can be drawn from such studies. I describe some of the benefits and limitations of RCTs and observational data sets when drawing conclusions about the relationship between antiretroviral therapy and the development of hepatic abnormalities.