Sir

Your Editorials 'Broken promises' and 'Mismanaged measures' (Nature 452, 503 and 504; 2008) seem to be contradictory. In the former, you excoriate the research community for running human efficacy trials for AIDS vaccines, whereas in the latter you complain about the use of surrogate end points.

You were right in the second Editorial. Without human efficacy trials on AIDS vaccines, we have had to rely on surrogate — mainly preclinical — end points.

Until the Merck trial, we were searching in a wilderness of preclinical data. The Merck trial, although it unfortunately did not show efficacy, taught us more than all the surrogate experiments.

Without efficacy data in humans, particularly when we are dealing with a human viral disease that lacks a good animal model, we are always guessing. To have experimental correlates of human efficacy would be wonderful, but we are still a long way from that situation. Meanwhile, you should be supporting more human experimentation, not less.