ABSTRACT
The development of antiviral therapy is a recent phenomenon that has accelerated dramatically since the mid-1980s. The investigation of agents for treating viral disease began in the 1950s as an outgrowth of the search for antitumor compounds and the resultant interest in drugs that affected DNA synthesis in vivo. A number of
these drugs were shown to inhibit viral DNA synthesis in laboratory settings as well [1]. In the 1960s, Kaufman used topical idoxuridine to treat herpes keratitis [2] and Bauer employed thiosemicarbazone to prevent smallpox in patients exposed to this disease [3]. However, these successes were limited and it remained difficult to identify agents which could inhibit viral replication without affecting host cells and producing unacceptable toxicity in patients. There was a perception in the scientific community that it would be difficult to develop antiviral chemotherapy that was appropriately selective. Since viral replication occurs intracellularly and uses host cell mechanisms, the task of identifying effective and safe antiviral compounds was seen as nearly impossible. In the early 1980s, the successful development andmarketing of acyclovir [4] for treating a wide variety of herpes virus infections in a number of different patient populations changed the way antivirals were viewed and encouraged the subsequent development of numerous compounds. The AIDS epidemic has highlighted the continued need for antiviral chemotherapy and the last two decades have seen a plethora of new compounds developed for AIDS as well as other viral diseases. This chapter addresses the specific elements of the design and analysis of clinical
trials most relevant to the development of antiviral agents. While this discussion is relevant to all antiviral chemotherapy, it focuses on the diseases caused by the human immunodeficiency virus (HIV), the hepatitis B virus, and the hepatitis C virus. These diseases are the ones most studied and have the largest body of clinical and statistical research. Since they are manifested as chronic diseases, the assessment of their potential therapies will require the use of time-to-event endpoints. In contrast, treatment for other infectious diseases, such as skin infections in healthy individuals, is often assessed by a ‘‘cure=no cure’’ endpoint which requires statistical methodology appropriate to binary outcomes.
