A new antimicrobial agent is used to treat a febrile, infectious condition. Three groups of patients are treated with a range of dosages. In all groups, fever is significantly reduced compared with baselines. The reduction is greatest in persons receiving the highest dose, but so is the mortality.

An objective reader cannot help but raise two questions: 1) What is the rate of survival in the untreated controls? and 2) What is the effect of treatment with nonantimicrobial antipyretic agents such as cyclooxygenase inhibitors? In the recent article on treatment of M. avium complex infection in persons with the acquired immunodeficiency syndrome (AIDS) [1], these important controls were not done. In addition, the question of whether they should have been included was not even mentioned, either by the researchers themselves or the distinguished experts who were invited to write the accompanying editorial [2].

Granted, the issue is more complicated. The patients selected for study had M. avium complex bacteremia that was nearly eradicated by the therapy. After observing treatment of this condition first with antimycobacterial drugs and then with a five-drug regimen, I used neither because of their demonstrated inefficacy. I now treat patients with clarithromycin and ethambutol on the theory that patients are better off without bacteremia. But I am not the one doing the research. Studies that show improved quality of life in patients treated with clarithromycin are small-scale, preliminary, and, I believe, similarly uncontrolled; this point should have been addressed by the authors of the article and the authors of the editorial.

Chaisson and colleagues [1] conclude that "clarithromycin, 500 mg twice daily, was ... associated with better survival," that is, better than the survival for patients who received even higher doses, not those who were treated for symptoms. Because of the rapid emergence of antimicrobial resistance during clarithromycin therapy, the authors' implication is not that the entire subject be reconsidered but that two drugs be given; this is precisely what many practicing physicians are currently doing. Goldberger and Masur [2], however, indicate several problems that emerge during treatment with clarithromycin alone, and they state that it is unclear whether survival is better in persons treated with 500 mg twice daily or simply worse in those receiving higher dosages. However, in their plea for additional studies, they do not mention placebo or antipyretic controls; I hope that they intend to encourage the inclusion of these controls or at least explain their omission.