The argument that people infected with the AIDS virus can safely stop taking their medicines for extended periods of time -- and may benefit from the breaks -- has been alternately championed and rejected over the past five years. A group of studies presented here today makes clear that the question is not settled and is actually more muddled than ever.

The new research suggests that longer and more open-ended interruptions of treatment for some patients may be more useful than presumably "safer," shorter breaks. But whether there is a net benefit to taking the breaks is uncertain.

"It's a hodgepodge. It's much more complicated than we think," said John W. Mellors, an AIDS physician at the University of Pittsburgh and one of the organizers of the 10th Retrovirus Conference taking place here this week. "The message is, people shouldn't be willy-nilly interrupting treatment."

There are many appeals to the idea of periodically stopping the three- or four-drug combinations of antiretroviral medicines that are the standard treatment of human immunodeficiency virus (HIV) infection. Patients can forget about taking pills for a while, and they avoid the drugs' immediate side effects and possibly the long-term ones, too. So-called "structured treatment interruptions" are also viewed by some as a way to make AIDS medicine go farther in poor countries, where the burden of the disease is greatest.

A study from Thailand compared three strategies: continuous treatment with three drugs; week-on/week-off interruptions, and an interruption that was allowed to proceed as long as the patient's CD4-cell count (a key measure of immune-system robustness) remained above a specified threshold.

There were no differences in mortality, complications or quality-of-life measures among the three groups at the end of a year.

However, the people on the CD4-guided strategy took drugs for only about one-third of the year, and they actually did better in controlling "viral load" in the bloodstream than the people in the week-on/week-off group.

A study from Spain compared the experience of patients randomly assigned to either continue antiretroviral treatment or to stop it and restart only when either CD4-cell count or viral load cutoffs were reached.

Over a year, about 60 percent of the treatment interruption group had to start taking medicines again (after an eight-week break, on average), while about 40 percent stayed off the drugs with no obvious or immediate ill effects.

A U.S. study compared seven rounds of eight weeks on/four weeks off treatment with continuous antiretroviral therapy. One purpose of this experiment was to test a theory that the immune system can be boosted by periodic exposure to swarms of HIV (as occurs when treatment is stopped), after which the body may be able to suppress the virus more forcefully or without the help of medicine.

The researchers found no evidence that this was happening. They also found no evidence that people regularly interrupting treatment had avoided the increases in bloodstream cholesterol and triglycerides that often occur from the medicines. However, the patients interrupting therapy were more likely to develop drug-resistant virus -- a problem worrisome enough to lead the researchers to stop enrolling patients in the trial.

Overall, the results of the various studies suggest that frequent interruptions may promote the emergence of drug-resistant virus. This may be especially true when a person is taking certain drugs -- such as nevirapine -- that persist in the bloodstream for weeks after the last pill is taken. In that circumstance, interrupting treatment in effect puts someone on single-drug therapy for a significant period of time -- a situation long known to encourage the emergence of drug-resistant virus.

On the other hand, there seem to be people whose immune systems can suppress virus adequately (if rarely completely) for long periods of time. Treatment interruptions guided by CD4 or viral load thresholds may be able to identify such people.

A large study done through the government-sponsored network of HIV clinics in the United States is now testing this idea.