Virologic and Regimen Termination Surrogate Endpoints in AIDS Clinical Trials [Gilbert PB, et al. JAMA 2001;285:777]: This is a "Special Communication" addressing the issue of HIV RNA levels as the appropriate surrogate endpoint for studies of antiretroviral agents. This has become the standard, although the authors suggest that the major issues are much more complex due to treatment-related toxicity, adherence problems, and drug resistance. They define two types of primary endpoints:

• Virologic endpoints: Frequency of virologic failure based on an arbitrary threshold such as 200 c/mL, lack of virologic response within 4 - 12 weeks, or decrease of some predetermined quantity such as 1 log, etc. They question the wisdom of the threshold of the 50 c/mL without further evidence of clinical relevance compared to 400 c/mL based on the observations of the Swiss HIV Cohort Study. [Lancet 1999;353:863]
  
• Regimen termination endpoint, which would evaluate outcome based on the time from randomization to the earliest "event" including virologic failure, permanent study treatment discontinuation, or an AIDS-defining diagnosis or death. Permanent study treatment discontinuation would be according to protocol-defined toxic effects.

The method to manage dropouts would be to censor the data and consider them successfully treated to the point of withdrawal or to consider dropouts to have reached an endpoint as failure. The authors conclude that the selection between these variables depends on the clinical objectives of the trial. The virologic endpoint is often preferred, but it may be appropriate to analyze both types of endpoints for study interpretation. Most important is the need for long-term clinical outcome studies to determine the utility of these various endpoints.
posted 3/1/2001

back to news table of contents