The Food and Drug Administration approved the first member of a new class of anti-HIV drugs yesterday, providing for the first time in seven years an entirely new mode of attacking the virus that causes AIDS.

The drug, enfuvirtide, belongs to a class of drugs called "fusion inhibitors," because they prevent the virus from fusing with immune system cells. It will be sold under the trade name Fuzeon, but is almost universally known as T-20, the name it carried in development.

While extremely expensive, in short supply and uncomfortable to take, the drug is expected to be embraced by tens of thousands of people harboring virus resistant to the nearly two dozen AIDS drugs now on the market.

Fusion inhibitors thus have been added to three other classes of anti-AIDS drugs. The last one, the protease inhibitors, which arrived in 1995, transformed patient care and led to a steep drop in AIDS mortality. While T-20 is not expected to have such dramatic effects, it may extend the lives of some late-stage AIDS patients by several years.

"I think it's a big deal. Anytime you have a new class of HIV drugs, it's a big deal," said Anthony S. Fauci, head of the National Institute of Allergy and Infectious Diseases. T-20 must be injected twice a day, like insulin, and will cost about $20,000 a year, more than double the price of any AIDS drug in use. It will be jointly marketed by Roche, the giant pharmaceutical house, and Trimeris, a small biotechnology company in North Carolina.

The companies will be able to produce enough of the drug this year to treat 15,000 patients. About 10,000 of them will be Americans, with the rest of production allocated to other countries. Production is expected to be sufficient to treat about 32,000 patients next year and 40,000 in 2005.

At least 20,000 Americans -- and possibly as many as 100,000 -- are candidates for T-20. The drug will be distributed from a central pharmacy, not yet named, on a first-come, first-served basis. Physicians prescribing it to more than one patient will be asked to rank them in order of need.

AIDS activists set up a mock graveyard at the gates of Roche's North American headquarters in Nutley, N.J., yesterday to protest the drug's price. Company officials say T-20 -- a 36 amino acid molecule synthesized through 100 separate steps -- is unusually expensive to make.

Current treatment of human immunodeficiency virus (HIV) infection consists of "combination therapy" -- three or more drugs from two or more classes. This helps prevent, or at least forestall, the emergence of drug-resistance traits in the virus, which mutates with notorious ease.

The three previous classes of drugs -- nucleoside reverse transcriptase inhibitors (NRTI); non-nucleoside reverse transcriptase inhibitors (NNRTI) and protease inhibitors -- all attack HIV after it enters cells of the immune system and begins to replicate. Fusion inhibitors block the first step in the complicated process of infection: the virus's attachment to the cell.

Some experts estimate that one-third of people who began AIDS treatment before 1995 carry virus resistant to at least one drug in all three drug classes. These patients are the ones most likely to benefit from T-20, which in most cases will be added to a combination of antivirals.

Studies have shown that T-20 suppresses the amount of virus in the bloodstream (the so-called "viral load") and boosts disease-fighting CD4 lymphocytes. Half of patients, however, develop resistance to it, too, within six months. Nevertheless, the pause in the disease's downhill course, however temporary, may offer real benefit.

"People don't just sort of fail like you flick off a switch. I would guess that it would be worth three to five years of survival, even if the benefit was only six months of response" as measured by viral load and CD4 count, said Michael S. Saag, a leading clinical AIDS researcher at the University of Alabama at Birmingham.

T-20's arrival may ultimately mark the start of a new wave of HIV treatments. Five other new classes of pharmaceuticals are under development. If one drug from each makes it to the market, doctors will have more choices in treating HIV infection than they have in treating all other viral infections, and most bacterial ones, as well.

This is a huge change from five years ago, when enthusiasm over the health-restoring effects of combination therapy was undercut by predictions that many people with HIV would soon develop drug resistance and run out of options again.

"We seem to be keeping up with the virus," John W. Mellors, a leading AIDS researcher at the University of Pittsburgh, said recently at the 10th Retrovirus Conference in Boston. "The pipeline of new drugs has an impressive number of new candidates in it. This is something we haven't seen in the last few years. I would say this is a bumper crop."

At the conference, about 3,500 AIDS researchers from around the world heard about research on compounds that block the cellular receptors CCR5, CXCR4 and CD4. HIV needs one of the first two -- and always the CD4 receptor -- to dock on immune-system cells, such as lymphocytes.

Two members of another new class of antiretroviral drugs, the integrase inhibitors, are being tested for safety and effectiveness in small numbers of patients.

Perhaps the most unusual new candidate described at the conference was a substance designated only as PA-457 by its makers at Panacos Pharmaceuticals in Gaithersburg. Derived from betulinic acid, a byproduct of the wood-pulping process, it inhibits the "budding" of newly assembled virus from the infected cell.

Unlike many of the compounds from new drug classes, this compound can be absorbed orally (at least by rats), which makes it immediately more appealing than an injected substance.