• Systematic Review of Nevirapine- Versus Efavirenz-Containing Three-Drug Regimens for Initial Treatment of HIV Infection (Poster 552)
  Authored by M.D. Neuwelt, O. Bacon, G.E. Kennedy, G.W. Rutherford

If making sense of many different clinical trials seems confusing, it is. The total body of evidence for the use of a drug is often made up of overlapping sets of data. Trials seem to give conflicting or inconclusive answers. The reasons for this are many and varied. They include:

• Design issues

  • Open label or blinded?

  • What type of comparisons were made?

  • Who were the patients?

• Analysis issues

  • What was the endpoint of the study?

  • How was it measured?

  • How many patients dropped out?

  • How were the results handled?

• Conduct issues

  • Where was the study done?

  • What was going on in the community?

  • What was the standard of care?

• Statistical issues, although boring, play a large role. Small studies may not show a difference simply because they are too small to find a difference if it exists.

One way that researchers can try to get a bigger picture of the evidence is through meta-analysis. This is a technique, or perhaps a dark art, of assessing trials based on quality and combining them for more robust statistical analysis. There are different types of meta-analysis and different statistical tools, but in the right hands, it can be a powerful tool.

A group from the University of California and the Cochrane Collaborative Review Group (a group of wizards of the dark art of meta-analysis) looked at the evidence comparing the efficacy of nevirapine (NVP, Viramune) with efavirenz (EFV, Sustiva) when combined with two nucleosides. The only large randomized trial, 2NN, found them to be no different, but there are other studies using efavirenz. Several retrospective studies, which although subject to many limitations, have suggested that efavirenz might have a more durable benefit.

By searching a variety of comprehensive databases of clinical studies (AIDSLINE, AIDSTRIALS, EMBASE, MEDLINE, the Cochrane Registry and the MetaRegister, not to mention the library at Hogwarts), the investigators found six comparative studies. Two were randomized trials and four were cohort studies. In the combined analysis, efavirenz and nevirapine were statistically equivalent at 24 and 48 weeks. Conclusions could not be drawn for later time points.

What does this add to the overall picture? Marketers of these two excellent drugs will struggle for a way to refine the spin, but the degree of fierce competition actually shows how useful both are. To patients and their doctors, the study sends a clear message. Both drugs are very effective and well tolerated. Within a reasonable degree of certainty, they are equally potent out to one year. Efavirenz has great data showing that this potency persists; similar studies for nevirapine have just not been done. Doctors and patients need to individualize the choice between these drugs based on the different toxicities, fertility, comfort with the data, convenience, patient preference, cost and regional availability.