BOSTON (AP) — A new study finds that temporarily stopping AIDS treatment is ineffective and even dangerous as a strategy for controlling drug-resistant HIV.

Many people being treated for AIDS carry mutant versions of the virus that are resistant to drugs in the three major classes of medicines used to subdue HIV. While many often stay outwardly healthy despite this, they are at increased risk of AIDS-related complications, and doctors would like to eliminate all traces of the virus from their blood.

One experimental approach, called structured treatment interruption, is intended to make the virus easier to kill. Doctors withdraw all medicine for a few months, then resume it. In theory, the drug-resistant viruses die off and are replaced by ones that are susceptible to the medicines.

At least three studies have tested the idea. The result of the largest, directed by Dr. Jody Lawrence of the University of California, San Francisco, were published in the Aug. 21 issue of the New England Journal of Medicine.

The doctors studied 270 volunteers, all of whom had significant levels of resistant HIV. Half were immediately switched to a different combination of the standard medicines. The rest took no drugs at all for four months before making the switch.

Just as doctors hoped, nonresistant virus emerged in two-thirds of the patients taking no drugs. But that offered no benefit in the end.

After a year of follow-up, 22 of the patients who took the four-month break had died or gone on to developing typical AIDS symptoms, compared with 12 who made the immediate switch to a different drug regimen.

“We’ve shown that the treatment interruption strategy was not of benefit and actually was detrimental,” said Lawrence.

One of the two other smaller studies concluded treatment interruption may work, while the other found the opposite.

Dr. Bernard Hirschel of Cantonal University Hospital in Geneva wrote in a journal commentary that treatment interruptions may eventually prove to have some role, but when study results conflict, doctors “must go with the results of the large study.”