The Virus at the End of the World

By Laurie Garrett
Esquire 1 March 1999

We thought the worst epidemic in history had been subdued. But this is not the same AIDS. This is a monster that is mutating faster than we can keep up. It's alive. It's healthy. And the worst is yet to come.

For the past several years, Spencer Cox, who is thirty years old, has lived behind a dam built to contain the catastrophe that is the most threatening virus of our time. It is a dam made of drugs designed to keep him from dying of AIDS. Diagnosed at the age of nineteen, Cox, a southerner who lives in Manhattan, became involved with the Treatment Action Group, which put him in daily touch with the cutting edge of HIV science. As each new drug was discovered, Cox was among those at the front of the line to take it.

And he was among the first to unwittingly cultivate viruses that were resistant, successively, to each of the new drugs he took. So for the last two years, as most HIV patients have been gleefully grabbing the latest combination therapies, Cox has already been experiencing problems with them, as well as with just about every other anti-HIV drug.

By late 1997, Cox's HIV-positive friends and fellow activists were, for the most part, feeling the best they had in years. Some even looked upon the gleaming dam and happily declared the end of AIDS. But Cox had to concede that he had run out of options. Nothing worked against the mutant microbes in his blood.

"None of us would say that having access to the best available scientific information every step of the way has helped us particularly," he says. "Professional judgment has consistently been way out in front of evidence. We're going to have to rethink everybody."

When Cox says everybody, he means the two hundred thousand Americans and at least an equal number of Western Europeans, Australians, and Canadians who are living behind the dam with him. And the rest of the 33.4 million in the world who are desperate to get their hands on these drugs. Treatments consist of more than 250 combinations of three or more anti-HIV drugs that were first introduced in January 1996 and popularized in the summer of that year. For the past two years, this dam has blocked the mighty HIV river. But there are now grave problems.

"The worst-case scenario is that I'm back in 1987," Cox says, referring to a time when there were no effective treatments for HIV. He insists that he's always taken his medicines religiously, carefully, by the book. And yet mutant viruses never fail to emerge eventually. As this past Christmas approached, Cox anxiously studied his worsening test results, praying that some new experimental treatment would be available before HIV overwhelmed him completely.

Spencer Cox represents the future of this virus.

At the University of Alabama in Birmingham, Dr. Michael Saag supervises research and the care of more than one thousand AIDS patients. The treatment is positively state-of-the-art, often better. Drug companies eagerly offer their newest innovations to Saag for clinical trials.

As physicians venture into ever wilder frontiers of HIV treatment, the grand experiment with combination therapies, called Highly Active Anti-Retroviral Therapy, or HAART, is rushing forward without any data. No one is keeping track. Indeed, as a result of the "AIDS is over" rhetoric, many AIDS service organizations have seen donations drop. So they've cut back on their policy and research staffs. And all over America, acute-AIDS-care facilities have shut down, breaking up teams of scientists, physicians, and nurses that used to monitor patient outcomes on a scale that offered vital information.

Saag's is one of the few remaining facilities, and his latest results are not good. Talk of a cure is over. "Perhaps the biggest difference between the cure paradigm and whatever paradigm we're in now is, we now should expect failure with whatever [HAART cocktail] we first use. We should plan on it. We should prepare for it. Clinicians should expect failure."

Far from imagining patients feeling triumphant while their viruses disappear, Saag sees a world in which "patients are only on regimens for three to four months before they switch," either because resistance emerges or side effects become unbearable. In one year, 157 of Saag's patients collectively took 189 different drug formulas, with only three patients taking the same mix of HAART drugs.

And despite such rigorous, individualized medical attention, Saag says, the HAART dam "is already leaking, and there's high danger of it collapsing altogether. There's no question that we're in a honeymoon right now. Failures are occurring right and left."

The very first person treated with HAART was Saag's patient. In August 1993, a young Alabamian started taking a drug called indinavir in combination with two older anti-HIV drugs. The patient experienced a miraculous recovery. By every measure, Saag says, he appeared to have been "cured." But in early 1994, his viral load started to crawl back up; the HIV became resistant to indinavir. For nearly four years, with careful switching of his HAART treatment, the man survived.

"He died last April," Saag says with a heavy voice. "It tells me that when people begin to fail virologically, they will eventually do so clinically." By the end of 1998, Saag's massive data pool was yielding alarming numbers. He could see that May 1997 had been the nadir for AIDS deaths in his population. Since that time, death rates have clearly been on the rise. "They aren't dying of a traditionally defined AIDS illness," Saag says. "I don't know what they're dying of, but they are dying. They're just wasting and dying.

"It is sobering," he continues. "While we are making good guesses, they are just guesses. We don't know what we are doing. Hubris! Hubris! We need to be scientists every step of the way and do our darndest to seek what reality is."

Reality is that by their very nature, viruses, and particularly HIV, have the better of human beings. Though this century has witnessed the development of successful vaccines against such viruses as smallpox and polio, and a handful of antiviral drugs have been invented, the course of viral infection is usually determined by one thing: the human immune system. And as the body musters its forces, viruses continually and rapidly mutate. The greater the pace of this viral transformation, the more difficult the immune system's work. It would take a revolution in scientific thought to come up with a drug that would shift the odds in favor of humans against the deadly elite of the viruses.

Every revolution has its Paul Revere, who shouts the news that rallies despairing masses. For HIV, it was Emilio Emini, of the Merck Research Laboratories in West Point, Pennsylvania. On January 29, 1996, Emini stepped to the podium in an enormous hall in Washington, D.C., a hall overflowing with thousands of scientists and physicians attending the annual Conference on Retroviruses and Opportunistic Infections. He summarized several small clinical studies conducted on fifteen patients here, four there, another dozen over there. He was describing the effects of then-unlicensed protease-inhibitor cocktails on people with HIV.

As Emini's slides, depicting plummeting HIV levels in the blood of patients taking the novel cocktails, flashed before the gathering, the excitement was unmistakable. And the audience went electric when Emini described one of Michael Saag's patients who, after taking the Merck protease inhibitor indinavir in combination with the older drugs AZT and ddI, had had no detectable HIV in his blood for nearly two years.

"If this can be achieved with one patient, there is no reason to believe it can't be achieved for many more patients," Emini told the crowd. He was struggling to stifle a grin. "And what we're facing here is a paradigm shift in the goals of HIV therapy. The goal now should be to lower the viral load as much as possible, for as long as possible, in as many patients as possible."

The usually staid scientists in attendance were beside themselves with enthusiasm. So was Wall Street, which within an hour had pushed Merck stock to a fifty-two-week high, gaining more than a dollar a share.

It's a home run, declared the ecstatic scientists.

But the greatest elation was felt among HIV patients and their physicians. In San Francisco and New York City, where large numbers of doctors handle HIV patients exclusively, the medical community eagerly prescribed the new cocktails. By the summer of 1996, word of miraculous deathbed recoveries had spread worldwide, and at the International Conference on AIDS, convened in Vancouver that August, mountains of evidence supported Emini's battle cry.

Word from Vancouver was simple and upbeat: New tests allowed physicians to measure, with exquisite specificity, how many HIVs were in a droplet of an individual's blood. The number of viruses, or the viral load, correlated well with the patient's prognosis; high viral loads--more than ten thousand HIVs per milliliter of blood--spelled very bad news. But numerous studies showed that the new HAART cocktails enabled the body to slaughter HIVs by the billions, bringing viral loads down to below the limits of technological detection, which at the time was about five hundred HIVs per milliliter.

Using even more-sophisticated technology, clinical researchers such as New York's Dr. David Ho, Alabama's Dr. Saag, and Emini showed that "zero detectable" HIVs in blood actually equaled no viruses in the bloodstream. And on the eve of the Vancouver meeting, a small group of top HIV researchers from all over the world gathered to cautiously discuss one word: eradication. They dared not say "cure"--that word was far too emotionally charged. But the group thought that powerful HAART cocktails might actually eradicate all vestiges of HIV from an individual's body.

And that--whatever one called it--would free the individual to go off the drugs and live a full life.

By the time thousands of scientists and clinicians gathered in Vancouver, Ho, Tufts University's John Coffin, and the University of Alabama's Dr. George Shaw had calculated that every day, an untreated individual's body made about ten billion viruses and a hundred million new immune-system CD4 lymphocytes to battle them. Over the course of the disease, this daily standoff shifted in HIV's favor, CD4 levels fell to zero, and viral loads soared to upwards of half a million HIVs per droplet of blood.

But in the presence of HAART, new HIVs were produced much more slowly, proponents said, and the immune system had a chance to recover. Eventually, the only remaining HIVs would be sequestered in hiding places throughout the body, and after three to five years, they predicted, those lethal reservoirs would die out.

Thus, they said, it might be possible to put thousands of people on HAART for just a few years, watch HIV disappear, and then withdraw the drugs. Voila. Eradication. An epidemic that by the close of last year would claim 13.9 million lives--and imperil more than 30 million others--gone.

Eradication. For the fifteen years of the epidemic before the emergence of these drugs, hope was unknown for someone diagnosed with AIDS. So a little excitement was understandable. And suddenly, the science of AIDS itself seemed to mutate--from a science of inevitable loss into something very different. It was a science of hope.

Propelled by the jubilant Vancouver news, tens of thousands of Europeans and North Americans started taking HAART. And when the international AIDS community reconvened last year in Geneva, results, overall, still looked great. For example, in 1996, the number of people in New York City who died of AIDS fell to five thousand, down from a 1994 total of nearly seven thousand. And rough data for 1998 shows about thirty-five hundred deaths recorded. Similar trends have been seen in most major U.S. cities. And nationally, AIDS death rates between 1996 and 1997 fell a breathtaking 47 percent.

Owing to the fact that there are forty-three million uninsured Americans whereas Western Europe has near-universal health care, Europeans have seen an even more dramatic trend. Between 1994 and 1998, death rates for HIV patients fell an astonishing 80 percent. The successes in Europe may also be a result of greater initial conservatism there in the use of experimental anti-HIV drugs during the 1980-95 period, which could explain the lower rates of multidrug resistance seen in European HIV patients.

But something was happening. HIV was not being eradicated. It would hide. It would mutate. It would, it seemed, wait for the toxicity of the new drugs to make its host sick in some other way. Wait for the sick person to quit the drugs or reduce them or take them in some slightly different way. And then it would come back.

Now, while drug companies tout genuine improvements in HAART--new drug combinations that are easier to take and perhaps somewhat more potent--they concede that earlier optimism about knocking out unique HIV actions, such as integration of the virus's genes into human DNA, is completely unfounded.

"I think we're probably as far away from cures as we are from vaccines," says Peter Young, vice-president of HIV therapeutic developments for the Glaxo Wellcome pharmaceutical corporation. In this epidemic, we have historically witnessed peaks and troughs of research optimism. Young acknowledges that the peaks felt last year are now disappearing. "The image that comes to mind is a lot of people filling up sandbags."

The declining efficacy of HAART drugs is, Emini now says, "a stochastic event. It's a matter of chance" that cannot currently be predicted in any individual patient. "There are patients who, even though they have suppressed the virus for years, will fail," Emini says with deep regret in his voice. "So what does that tell you? It tells you it's not an easy consideration to 'cure' a patient even after a long time of suppression. The virus will cycle up."

HAART didn't lose its luster suddenly. Rather, it's been fading slowly--so slowly, in fact, that it's easy for the denial-inclined to ignore the new reality.

Last year, Dr. Neal Nathanson became head of the National Institutes of Health's Office of AIDS Research, overseeing an annual scientific budget of $ 1.7 billion. "My sense is that it won't be possible to keep someone on HAART for a lifetime," Nathanson says. "Every death that didn't occur last year is not a cure; it's just a postponed death. I don't hear much optimism. I'm afraid that the death rate may start to climb back up. The decline in mortality could be used to argue that we should cut back in our research. And that would be disastrous."

Disastrous, Nathanson says, because there will soon be a need for fundamentally new treatment strategies for AIDS. Most of the drugs now in development at the approximately twenty-five companies targeting the $ 5 billion U. S. HIV market, though, are simply variations on the HAART theme. No viable drug that combats HIV in an entirely new way, and no vaccine, is likely to find its way to the marketplace before 2005 at the earliest. Even the rosiest forecast fails to envision a genuine HIV vaccine in widespread use before 2015.

Meanwhile, for more than 95 percent of the 33.4 million people living today with the disease, the $ 10,000 to $ 60,000 annual HAART price tag is utterly out of reach. Per-capita health-care spending averages ten dollars a year in each of Africa's hardest-hit countries--nations in which 20 to 26 per cent of all men and women aged fifteen to forty-nine years are HIV-positive.

Many researchers--including those originators of the eradication hypothesis of 1996--now say that the reservoir of hidden HIVs in apparently successful HAART patients is large and long-living. David Ho, who was a pioneer of the combination therapies, now thinks patients would have to take the difficult drugs for twenty-five to thirty years to eliminate those hidden viruses. Some other scientists put the figure even further out, at forty to fifty years.

It doesn't really matter, because either number is impossible. The HAART drugs are difficult to take and increasingly are seen to cause a range of nasty side effects. Chief among them is a mysterious change in lipid metabolism that results in redistribution of body fat and elevated cholesterol counts. The extent of the lipid problem in HAART users, and the reasons for it, remain matters of vigorous debate. But there can be no doubt of the disorder's severity when thirty-year-old men develop cholesterol counts above 400, soaring triglyceride levels, skinny limbs, and protruding paunches. The effect on women's bodies is even more profound, as they develop large, waistless trunks and sticklike limbs. So dramatic is the physical change that some people elect to discontinue HAART and face death rather than spend longer lives in unrecognizable bodies. After a year on HAART, some patients have their parents' physiques. Or their grandparents'.

Many respond by dieting, taking hormones, working out at the gym. But this is not the sort of problem that readily responds to such effort. And no one in the field thinks cholesterol counts of 300 to 900 can be tolerated for long, even by a strong, youthful heart.

Anybody who takes ten to twenty even mildly toxic drugs a day is giving his liver, kidneys, intestinal tract, and bowels a real beating. Over time, these vital organs become less able to absorb and process the medicines, so the antiviral agents never get to their targets. Some HAART drugs cause damage to the nervous system. There are reports of drug-induced hearing loss. And while such acute side effects appear in less than 10 percent of all patients who stay on HAART for one to two years, it is widely assumed that toxicity will increase over time, simply because analogous approaches to treatment, such as those used to fight cancer or drug-resistant bacteria, become increasingly toxic the longer they are employed.

And the drugs are very difficult both to prescribe and to take. Some can't be taken together; some exacerbate the effects of others; some enhance others; and certain cocktails simply don't work well.

For patients, HAART can become a full-time lob. Some drugs must be taken four times a day, some once, some twice. Some must be taken on a full stomach, others before eating. And all well-managed HIV patients also take a host of prophylactic drugs that prevent common opportunistic infections such as parasitic pneumonia, a blindness-inducing virus, and a bacteria similar to tuberculosis that's normally found in birds. It is not uncommon for an HIV patient to be taking some type of pill every three hours.

Keeping track of all those pills, injections, salves, and the long list of herbs popular in the HIV community is a daunting task. And failure either to prescribe correctly or to follow doctor's orders precisely promotes the emergence of drug-resistant viruses. In many cases, the ever-mutating HIV is capable of making changes that allow it to resist not just one drug but an entire class of them.

HIV develops resistance to antiviral drugs roughly the same way bacteria become resistant to antibiotics, but HIV does it with a frequency that is orders of magnitude faster than that seen with bacteria. Any use followed by interruption and then reuse of the same drugs gives HIV the opportunity to mutate and clone an enormous colony of resistant viruses. And in the case of HAART, very brief interruptions--on the order of days--are enough to shift the advantage to the deadly viruses.

Companies have responded by developing quick resistance tests that physicians can perform routinely on virus samples from their patients. If a patient is found, for example, to have HIVs that have mutated to be resistant to indinavir, the physician might then switch the client to a cocktail that has a different protease inhibitor.

Unless, or until, the virus is resistant to all protease inhibitors.

No other disease, either infectious or chronic, is treated for years on end with so complex a drug regimen. And the first drugs you take are probably the most important, because everything you do after that is going to be less effective.

There are no HAART cookbooks, and experts differ on every fundamental aspect of standards of care: When should patients begin HAART? With which combinations should they start? When the first round fails, what second-line cocktails should be used? Every leading HIV organization and national medical-research institute in the United States and Europe has tried to answer these questions. They have all reached the same conclusion: Answers must be found by each physician in consultation with each patient.

In other words, because each individual case is an experiment, the physician and patient must approach treatment choices in collaboration, acting as a unit that embodies the traditional experimental research team of guinea pig, scientist, doctor, and ethics review panel.

It's tailor-made medicine, surfacing at a time when managed-care companies think that any general physician can handle HIV treatment. A national HIV-treatment survey conducted last year by the University of California San Francisco found that one out of four HIV patients was receiving incorrect, out-of-date drug treatments for his infection. The survey, which polled 476 American physicians, found that doctors who had in their careers handled fewer than 100 HIV patients were the most likely to prescribe the wrong combinations and to delay treatment until their patients were too sick.

One way to get around this--decreasing both toxicity and complexity--would be to simplify the regimens of HAART. Much hope was placed on the possibility that a full HAART cocktail would be needed only for a few months: Once the patient's viral load dropped to below detectable levels, the patient could switch to an easier, two-drug therapy. Large studies employing this strategy were conducted in Europe and the United States. The grim results were published last October in The New England Journal of Medicine. Patients who went on the simple therapy were six times more likely to suffer dangerous viral rebounds after six months than patients who remained on the more complex standard treatment. No one can safely diminish the drug cocktail.

In fact, some physicians are moving radically in the other direction, giving patients extraordinarily complex regimens. Manhattan HIV specialist Dr. Howard Grossman determined that several of the drugs used in HAART were prompting cross-resistance against other drugs, and he was beginning to see an alarming number of failures among his patients.

"So I talked about it, thinking, Should we add in other combinations and recycle back in some old [anti-HIV] nucleosides?" Grossman says. The result: Twice a day, Grossman's patients now take two protease inhibitors, three nucleoside-class anti-HIV drugs, including a new anti-HIV drug called abacavir, an old drug used for cancer care called hydroxyurea, a nonnucleoside HIV blocker, and another new anti-HIV drug called efavirenz. That's eight anti-HIV drugs, in addition to a plethora of medicines aimed at preventing opportunistic infections, plus growth hormone and testosterone to combat HIV-induced wasting. "And it's really well tolerated," Grossman says. "It's amazing."

This "mega-HAART," as Grossman calls it, is now being taken by at least forty patients, all of whom were failing standard treatment due to the emergence of drug-resistant HIVs. Half have kept their viral loads down to below the limits of detection, Grossman says.

Of course, every single one of them is suffering lipid-metabolism side effects, and Grossman says he is "worried about the cardiovascular stuff. Are we going to have ten thousand people who are having myocardial infarctions before age forty?"

The data has caught up with "cure" and "eradication": Both concepts are, for now, dead. The new vogue is "remission," a term taken from another currently dismal field of medicine, cancer care. HAART, it is argued, pushes patients into a remission in which the virus is forced into hiding. The strategy then is to use HAART simply to bring viral loads way down, then hit patients with something that packs a wallop to their immune systems--often an agent that the body mistakes for an infection--forcing cells that are harboring hidden HIVs to wake up, release the viruses into the bloodstream, and thereby expose the microbes to HAART drugs. There's no easy way to do this. Provoking the immune system with such force makes people feel, at best, as if they have the flu. With stronger hits, patients must be hospitalized in an intensive-care unit during treatment.

Radical as it may sound, the approach has been tried in several places, yielding mixed results to date. Dr. Joep Lange at the University of Amsterdam put three patients through ICU care after placing them on a five-drug HAART regimen. One of the patients left the hospital free, according to tests, of any HIV. He pronounced himself cured and disappeared, lost since to Lange's team. Other scientific groups have had similar experiences, finding that a single patient here or there seems to experience a remission that is roughly equivalent to going cancer-free after chemotherapy. Of course, no responsible oncologist tells such a patient that he or she has been cured of cancer.

The most positive remission results so far have been obtained at the National Institute of Allergy and Infectious Diseases in Bethesda, Maryland. There, director Dr. Anthony Fauci and his colleague Dr. H. Clifford Lane have used a powerful human-immune-system stimulant called interleukin-2, or IL-2, on fourteen HAART patients. The patients got infusions of the chemical for five days at a time every eight weeks for at least a year.

Most of the patients didn't get much more out of the IL-2 experiment than several rounds of feeling as if they had lousy cases of the flu. But six patients have had bona fide remissions. Fauci's group has found no HIV in their blood--in more than 330 million cells analyzed per person or in biopsied lymph nodes.

But what does this mean? Putting aside the astronomical costs and practical difficulties of conducting the same treatment on a million HIV patients in North America and Europe, does it work? Fauci is cautious. He avoids using many words--not only cure and eradication, but also remission.

Dr. Robert Siliciano at the Johns Hopkins School of Medicine in Baltimore uses a new technology for searching for evidence of HIV in seemingly "cured" patients, and his findings consistently point to continued presence of HIV hidden deep in reservoirs in the body. Siliciano searches for the presence of HIV genetic material integrated into the chromosomes of human cells. These seemingly innocuous chunks of DNA can, when stimulated, commandeer the genetic machinery and make millions of copies of themselves that are released into the bloodstream the moment HAART drug levels, for any reason, fall.

In Chicago, Dr. Steven Wolinsky of Northwestern University employs similar techniques to search for HIV genes in seemingly effectively treated patients. His new findings also suggest that viruses in these patients are actively reproducing and, worse yet, mutating. Wolinsky conducts painstaking DNA and RNA analysis of HIV samples drawn from several HAART patients at various points over the years they are on the therapy. He uses an ultrasensitive assay first to look for whole viruses in the patients' blood. Then he looks for pieces of viral genes.

He consistently finds that certain HIV genes are being produced even when a patient's viral load is below zero for more than a year. And one key viral gene, called nef, never goes away, even when all else disappears. In every case, nef levels begin to increase in the patients about six months to a year before the rest of the viral markers rise and well before a clinician would recognize that treatment was failing.

"After eighteen months on therapy, the patient's back to where he started. And that's the reality," Wolinsky says. "The virus is not gone--it's still there years out. So the question is, is this an evolutionary question? Is there ongoing replication?

"Why do we always see nef RNA? The virus is telling me something, but I'm not smart enough to see it," Wolinsky says with a shrug. "Is the sky falling? I wish I knew."

HIV is made up of many genes; nef is perhaps the most tantalizing but least well understood. A few years ago, Ronald Desrosiers of Harvard's New England Primate Center caused tremendous excitement when he demonstrated that monkeys vaccinated with a live, genetically altered AIDS virus that had no nef were immunized against all subsequent exposures to the virus. It looked as though Desrosiers had found the Holy Grail: an AIDS vaccine. But later it was shown that baby monkeys and chimps and mice did get AIDS when vaccinated with nef-deleted viruses--it just took longer. How and why have not been discerned. Nef-deleted viruses appear to stay in latency. And nef-active viruses, Wolinsky believes, eventually cycle up in HAART patients.

Mario Stevenson of the University of Massachusetts Medical Center in Worcester is one of the world's experts on nef. He says the mysterious gene "is mind-boggling to scientists," who simply cannot agree about its role in the course of disease events.

"It's like mud just now," Stevenson says. "Many of the things we know about the virus just don't make sense."

Stevenson sees nef and other HIV genes in successfully treated HAART patients, too. And he thinks that HIV is always reproducing and changing, despite the seemingly effective therapy. Merck's Emilio Emini says that there is "no doubt about it at all," that HIV replicates and mutates in seemingly effectively treated patients. But he thinks more research will be needed to see whether Wolinsky's nef hunch is true.

But that, in the end, isn't nearly as important clinically as the overall consensus now reached among scientists: HIV will reproduce and mutate. Period. Unless fundamental breakthroughs are made in anti-HIV science, the miracle of 1996 is finished.

Denial and fear of facing failure can be seen among scientists, says Steven Wolinsky. "I don't understand where all this optimism comes from. The problem is hope-driven science. The problem is reading your data through a prism of hope, rather than seeing it for what it is."

"We've said from the beginning, this is a nasty little virus," says Emini. "My fundamental hope is that in the end we'll be able to make a sincere shot at a vaccine here."

A shifting emphasis to a vaccine, of course, means giving up on the 33.4 million infected. Burnout and death have decimated the ranks of traditional AIDS activists. Those who remain recognize that HAART isn't the answer, and they rage over the medical community's tendency to blame the patients for somehow creating their poor fates by failing to adhere to their drug schedules. They claim that the biomedical community feels no urgency in the matter and isn't searching for radically new drugs. Since 1994, Dave Gilden has edited Treatment Issues, published by Gay Men's Health Crisis in New York. He's been organizing new protests aimed at the drug companies. "Look, the drugs are getting more expensive, half the people in the U.S. aren't getting any, all those people in Africa are getting written off, and there's no attempt to find simpler, more effective treatments," Gilden says.

And most problematic, activists feel that their own HIV communities cast vitriol against them for speaking about the truly dire state of AIDS medicine.

Well-known New York activist Michelangelo Signorile says that a sort of mass denial has set in. "People were furious that I would be saying that AIDS did not go away. I've been accused of causing panic, being hysterical. People are embarrassed to talk about the fact that the drugs aren't working for them, and even to say that their lover recently died of AIDS--because of that sense of failure." So strong has been the attack on Signorile that he is giving up and moving to New Zealand.

Another loud voice, Gabriel Rotello, author of Sexual Ecology, has opted for the "calmer" world of network television production.

There's another small problem. New HIV infections have continued to mount, but often in different social groups than before. While new infection rates appear to have declined among gay white men in North America, for example, they have risen steadily elsewhere. A striking illustration of the change can be seen in the Bay Area, where until the early 1990s, by far the majority of all new HIV infections occurred among gay white men living in four key neighborhoods of San Francisco. By last year, the new-infection rate in that group, and in the city overall, had fallen.

But across the bay in Alameda County, officials declared a crisis in late 1998, when number crunching revealed that 41 percent of all newly diagnosed HIV cases were African-Americans, although blacks make up just 18 percent of the county's population.

Further surveys published since November of last year reveal that far more than half of all people infected with HIV in America do not receive HAART drugs, primarily because they are outside the circles of sophisticated HIV care. In fact, the Rand group has concluded--after surveying thousands of health--care providers and HIV patients--that fewer than a third of those infected have the sort of private insurance that would allow access to specialized HIV treatment; about one in every five of them has no health insurance at all, and the remainder are on Medicare or Medicaid. A substantial majority of uninsured and Medicaid patients are African-American or Hispanic.

Estimates are that approximately 200,000 Americans are being treated with HAART drugs today, although somewhere between 650,000 and 900,000 Americans have been infected with the virus, according to the U.S. Centers for Disease Control and Prevention. Health analyst Derek Link of GMHC estimates that there are 50,000 people in New York State alone who should be in treatment but are not, either because they lack access to appropriate health care or--as is especially common in the communities now being hardest hit--because they are unaware of their HIV status. Dr. Perry Smith of the New York State Department of Health thinks that's a reasonable guess, adding that his office "gets about 51,000 undiagnosed as an upper limit" on the number of New Yorkers who are unaware that they are HIV-positive. Dr. Helene Gayle of the CDC says, "Our estimates are that a third of the people who are HIV-positive don't know it."

Given that most experts on combination therapies advocate use of the drugs soon after infection, this seems to indicate an enormous gap between the number of Americans who ought to be on HAART versus the roughly one third of those infected who actually are, and who already shoulder what experts estimate was in 1998 a $ 5 billion national bill for anti-HIV medicines.

All over America, state and federal health officials are planning big campaigns for 1999 aimed at encouraging people to get tested for HIV and, if positive, to get on HAART.

Which could be a mighty hard sell, given the mounting evidence that these therapies ultimately will not work.

What exactly lies in the months ahead? Overall, the experts say, it appears that in the absence of an unexpected breakthrough, most, if not all, HAART patients will eventually run out of treatment options. They will then be in Spencer Cox's position--in limbo, waiting and hoping for a new drug.

Given that most of the two hundred thousand or more Americans currently on HAART started taking the drugs during the last quarter of 1996, the bulk of that population will likely reach the limits of their therapy at about the same time. Already, more than a third of all patients who started the drugs during the exciting days of 1996 have failed on the therapy. And for those still sitting comfortably behind the dam, the reservoir is increasingly turbulent, filled with bodies that are rebelling against drug toxicity or being overwhelmed by viruses that are evolving to outwit the HAART cocktails.

If Dr. Saag in Birmingham is correct and the first cracks have already appeared in the HAART dam, 1999 could be the Year of Great Disappointment as clinic after clinic across America finds itself trying to cope once again with acute AIDS cases. It will pose remarkable challenges, as few cities still have centers of excellence for AIDS care. Further, the demographics of the HIV population have shifted radically since the American epidemic's mortality peaks in 1994. But it will be worse this time: More and more, HIV hits the country's most alienated sectors. These are people who lack the resources and skills to demand proper treatment. The fastest-growing group is African-American women, most of whom acquire the virus through heterosexual contact. Many are impoverished mothers. The new HIV population doesn't have the same organization and political clout that successfully confronted three presidents, Capitol Hill, the world's largest pharmaceutical conglomerates, and the scientific research establishment during the period between 1984 and 1994.

They are less like Howard Grossman's gay white clientele. And more like Dr. Wafaa El-Sadr's Harlem Hospital patients.

"Fifty-three percent of my patients fail therapy by twenty-four weeks," El-Sadr told the twelve thousand scientists at last summer's international AIDS conference in Geneva. Lester, a reformed IV-drug user, is typical, she said. After complaining about how much he needs to read in order to decide what to do about his HIV, Lester simply asked, "Should I take these drugs?"

"We don't know," El-Sadr said. "Will these drug regimens work? How can I answer?"

All anyone can do is look at the numbers, chart the graphs, map the projections: The trajectory doesn't look good. With time notched out in weeks, extending from July 1996 through the end of 2000, and graphed against numbers of HAART patients whose treatments are failing, it takes a heap of denial to reach anything but sobering, even grim, conclusions.

And time, once again, is the enemy, for it favors the virus. In the time it has taken you to read this article, millions of new HIVs were made in the body of a typical unmedicated patient, thousands in a HAART patient. Some of the viruses will have developed capacities that they didn't have when you began reading.

Minute by minute, HIV gains advantage. Our best defense against it is slowly giving way. And it will take a lot more hard work, money, and creative human thinking to come up with a way to shore up the dam--or build a better one.

After doing graduate work in immunology at UC Berkeley, Laurie Garrett decided to become a journalist. Today, not only is she the medical and science writer for New York Newsday, but she is the only writer ever to win the top three prizes in her field: the Pulitzer, the Polk, and the Peabody. Garrett has been covering AIDS since 1981, in the days before the scourge had a name, so this month's cover story, "The Virus at the End of the World" (page 102), was a long time coming. "We're currently in an acute optimism phase in which the public, most HIV patients, government agencies, and the majority of HIV caregivers feel that genuine advances have been made that have all but eliminated AIDS," says Garrett, author of The Coming Plague (available in paperback from Penguin) and The Death of Health (to be published this year by Hyperion). "But this optimism doesn't jibe with what I feel the scientific data supports. And unless some serious scientific breakthroughs are made swiftly, we'll all be hurting terribly in the months ahead."