It was a rare moment in Cathy Olufs' life -- she felt good. Diagnosed with HIV in 1995, Olufs had made peace with her illness two years later and was living a "normal," symptomless, 600-CD4-cell life. Olufs had landed a new job at an AIDS outreach center in Los Angeles, her best gig ever. "I was completely fine," says Olufs, now 38. "Everything was going along great." There were even days when she forgot she had a disease.

Then she went to the doctor.

In 1997, most AIDS docs were pushing their patients to start triple-combination therapy, and Olufs' was no exception. "The buzz was 'Everyone needs to be on three drugs' and 'Get your viral load to undetectable,'" Olufs says. She went on a standard regimen: two nukes (ddI and d4T) and nelfinavir (Viracept), one of the new class of protease inhibitors.

Right away, she felt lousy. She had less and less energy. Her once-shapely legs became ridden with large, ugly veins. And then came the coup de grâce: daily diarrhea, usually without any warning. "I hate to be too explicit, but it would just explode out of my butt," Olufs says. "Quite often I would have accidents." Olufs took to carrying around an air freshener in her purse.

More insidious, her feelings about having HIV changed dramatically. On the one hand, she felt grateful for the meds -- without them, her doctor reminded her, she might get sick. But her future was growing darker by the pill. "I was thinking 'Oh my God, am I going to have to continue this forever?'" she says. "I was really despairing." After a year of this, Olufs read a lot more about the drugs; learning of a radical new experiment called structured treatment interruptions (STIs), she was this close to trying one. A conversation with a friend clinched it. When she told him her CD4 count was around 700 when she started the meds, he said, "What the hell are you doing on drugs? You're healthy enough that you don't have to be taking them." Olufs immediately made an appointment with her doctor and told him (she didn't ask) that she was quitting meds. Four years later, Olufs is still drug-free and feels fine.

Still, Olufs can't help wondering why she was put on drugs in the first place. She's not angry, she says, but she recognizes that she didn't exactly get the best medical advice. "Everybody was going on antiretrovirals, so I got them too. They didn't treat me as an individual," she says. "I wound up just having this horrible year that I never want to have again in my entire life."

This is the official story: Protease-powered combos worked a miracle. Since their widespread use in 1996, the number of Americans dying each year from HIV has dropped from 50,000 to around 16,000. AIDS, the worst plague in history, was transformed in the West into a chronic, manageable disease. "The success of HIV therapies is one of the great accomplishments of modern medicine," says Martin Markowitz, MD, who, along with colleague David Ho at New York City's Aaron Diamond AIDS Research Center, did landmark research on protease inhibitors.

This is the revisionist version: Protease inhibitors (PIs) were hypermarketed and overprescribed to tens of thousands of healthy patients who, like Cathy Olufs, went on to develop vexing side effects and viral resistance. These patients exhausted future drug options long before they truly needed them. Now many are looking down the barrel of salvage therapy. According to Mark Harrington of Treatment Action Group (TAG), AIDS doctors made this "anti-intellectual scientific extrapolation": If one group of HIVers -- the deathbed Lazaruses and their slightly less sick sisters and brothers -- benefited greatly from triple-combo therapy, then all HIVers would.

According to an unpublished study by Julio Montaner, MD, an HIV researcher at the University of British Columbia, 50 percent of U.S. and Canadian HIVers put on HAART since 1996 didn't need these meds: Their clinical health -- how they actually felt -- would have remained the same even without attaining "undetectable" viral loads or raising CD4 counts that were already above 200. "All these people could have waited," Montaner says. "They could have deferred the cost of the drugs and the side effects." If, as Harrington estimates, as many as 300,000 HIVers went on antiretrovirals, that's 150,000 wasted regimens.

Now, no one is advocating a class-action suit. But there is a growing belief among AIDS docs that a certain group of patients -- those with CD4s above 500 who had never been on meds -- were done significant harm. One after another, these HIVers were rushed onto the protease-combo conveyer belt when they would have been much better off, well, off. If nothing else, these treatment newbies could have waited a few years for kinder, gentler options. "We gave antiretrovirals to people who weren't going to get sick anytime soon," says Keith Henry, MD, director of HIV clinical research at Minnesota's Hennepin Medical Center. "We were naïvely and foolishly prescribing these drugs."

To look back in anger may seem a bit uncharitable. After a decade of watching their patients die, despite heroic efforts, docs can be forgiven if they became fanatical about protease combos. They desperately wanted to heal their patients -- and perhaps themselves, as well. "We were coming out of a decade of death," says Roger Pomeranz, MD, an HIV researcher at Philadelphia's Thomas Jefferson University. "It made sense to say, 'Let's throw every drug we've got at the virus.'"

Many researchers say that the new understanding of when and how to start HAART came out of a natural, if bumpy, learning process. "The retrospectoscope is a powerful tool used best by critics," Markowitz says. "Had women with breast cancer had [drugs like] PIs available and lived to tell the story, no one would question their value, even if better drugs came along."

But the "trial and error" argument only goes so far. As early as the mid-'90s docs did have more tolerable drug alternatives available. It's just that studies showing that these regimens were effective were widely ignored, as were the small group of researchers arguing for greater restraint in mass HAART prescription. "I kick myself now because I fell for the party line and put patients on treatment when they didn't need it," Keith Henry says. "What we were doing was just not good medicine."

Julio Montaner was at a conference in 1995 when a panicked colleague pulled him over in the hallway. "We've got a major problem," Montaner was told. "Our blood samples are all messed up." For two years, Montaner had been working on a clinical trial of nevirapine (Viramune), the first in the class of non-nucleoside reverse transcriptase inhibitors (NNRTIs). Nevirapine had been a bust in several previous studies, and Montaner's persistence had earned him a reputation as, he says, a "non-nuke fanatic" -- even in his own lab. So when one of his team picked up, for the first time, a significant decrease in the viral loads of patients on an AZT/ddI/nevirapine combo, the scientist assumed there had been a major screw-up. "We must have mixed up the blood samples," he told Montaner. "We're going to have do this thing all over again."

Montaner, a stubborn Argentinean, shot back with a sarcastic "Could it be that the nevirapine is working?" The experiment would continue as scheduled.

A few months later, some 15,000 scientists, docs, media and advocates descended on Montaner's hometown of Vancouver for the 1996 International Conference on AIDS. It was -- and remains -- the most exciting and optimistic AIDS confab ever.

The buzz, of course, was all about protease inhibitors. America's top AIDS researchers presented a slew of papers showing that PIs, when combined with two nukes, could bring HIVers back from the brink of death. The "high-tech" PIs -- at that point, Abbott's ritonavir (Norvir), Merck's indinavir (Crixivan) and Roche's saquinavir (then marketed as Invirase) -- were actually developed from a failed kidney drug that was found to jam HIV's protease protein. It was dusted off, painstakingly tweaked and then pushed through FDA approval in record time. News of such fabulous new drugs had been unthinkable even three years earlier at the Berlin conference, where final data on AZT, the first HIV med, showed that the drug was a bust. "We have traveled from the despair of Berlin to the hope of Vancouver," then-Health and Human Services secretary Donna Shalala said in her opening remarks at Vancouver. "Not false hope, but hope grounded in science."

Montaner was also among the presenters at Vancouver. In a small auditorium, in the shadow of David Ho, whose tantalizing talk of HIV eradication made him the man of the hour, Montaner presented his nevirapine data, showing that an AZT/ddI/nevirapine combo lowered viral loads to undetectable. His study focused on a fairly narrow group of patients -- treatment naïve with CD4s between 200 and 600 -- but it showed that for them a nevirapine regimen worked fine. The drug had the added bonus of being cheaper than PIs by half and easier to take (two pills a day compared to as many as three every eight hours). Finally, nevirapine's side effects -- fevers, nausea, headaches and, above all, rash -- were well established, and they generally cleared up in a few weeks. PIs, on the other hand, had been rushed to market so fast that little was known about potential side effects.

Montaner's research had some significant limitations. His subjects were healthier than many in the PI trials, leaving it unclear whether a non-nuke would benefit patients with full-blown AIDS. His study was also the first to show an NNRTI's potency, whereas PIs had been proven effective in numerous trials. But none of these caveats justified the cool reception to Montaner's work, according to Robert Murphy, MD, a Northwestern University researcher. "Julio's study just got overlooked," he says.

Montaner blames the "overwhelming power of the PI lobby" for nevirapine's neglect. These lobbyists -- mainly protease-makers Merck, Abbott and Roche and the docs who had made their names on protease research -- were deeply invested in seeing PIs succeed. They pulled out all the stops to insure maximum media interest in what was soon dubbed the Protease Revolution.

In contrast, Montaner's collaborators were from Canada, Europe and Australia -- outsiders to the American medical establishment. And nevirapine's German maker, Boehringer Ingelheim, was only mildly promoting a product that had repeatedly produced underwhelming data. BI was also having a hard time overcoming doctors' concerns about nevirapine's tendency to cause a rash, known on rare occasions to send patients into shock.

Ironically, the most damning evidence marshaled against nevirapine grew out of a "mistake" Harvard University virologist Martin Hirsch says that he and several colleagues made while performing the same experiment on nevirapine as Montaner. While Montaner's participants were all treatment naïve, Hirsch's had already been on nukes. Hirsch's group showed little benefit from the non-nuke regimen -- a devastating blow. In Vancouver, Montaner's critics cited the Hirsch study as proof that Montaner's findings were a fluke.

But as Hirsch sees now, his clinical trial -- ACTG 241 -- was doomed from the start because most of his pre-treated HIVers had already developed resistance to the nukes in their experimental combos. Even though they were downing combos such as nevirapine, AZT and ddI, they were, in effect, on nevirapine monotherapy. "It wasn't going to work," Hirsch says.

Hirsch's study is one of the many great "What if" moments in HIV treatment: What if he'd thought to give nevirapine to treatment-naïve patients? Nevirapine would likely have succeeded, and docs would have had an effective triple-combination therapy a year or two before the PIs hit the market. TAG's Harrington wrote in last March's TAGline about how "tempting, but ultimately futile" it is to "wonder how many lives could have been saved" if Hirsch had caught the flaw in his approach from the start. Montaner, not surprisingly, believes that Vancouver would have turned out quite differently: Scientists would have been less intoxicated by the protease class and far more reasonable about considering alternatives. "We could have given the PIs a real run for their money," he says.

Keith Henry got home from Vancouver "so excited," he says, "I couldn't see straight." He had just witnessed the Protease Revolution. A convert, he began prescribing PI combos to patients at his Minnesota AIDS clinic. He knows he saved lives. But he also noticed that the drugs were poorly tolerated -- that is, in the short term they made his patients feel sicker. Many of his patients suffered an ever-growing list of side effects -- stomach aches, fatigue, nausea, diarrhea, vomiting. The meds also seemed to cause serious spikes in cholesterol and triglycerides. "I sent blood samples out to be tested -- they started messing up the machines because the levels were so high," Henry says. "Common sense was telling me that PIs weren't going to live up to all the hype."

Many clinicians were beginning to experience a hangover from the Vancouver euphoria. Faith in David Ho's much-hyped, mathematical model that protease combos could "cure" AIDS by stopping replication entirely, allowing your immune system to reboot, began to dim. But the mantra of "Hit early, hit hard," regardless of your health, had already taken hold. That there was almost no scientific evidence to favor this aggressive approach in healthy people mattered little. Steven Miles, MD, the head of UCLA's AIDS clinic, says he gave more than 50 lectures in 1997 and 1998 urging doctors to consider holding off on treatment or at least using a more tolerable, protease-sparing regimen. "No one," he says, "would listen."

Across the Atlantic, the European medical community was taking a sharply different approach. Treatment typically wasn't initiated until patients' CD4s fell below 350 -- whereas the U.S. guideline, set in 1997, put the threshold at 500. When European HIVers did finally begin treatment, they were often given a non-nuke regimen. "We were pretty keen on the NNRTIs over here from the get-go," says Brian Gazzard, MD, a leading professor of HIV medicine at Chelsea Westminster Hospital in London. Some 80 percent of his patients have stayed on non-nuke regimens for at least two or three years.

According to Gazzard, it wasn't that Europeans knew better -- they just knew what they didn't know. The assumption in the U.S. was that HIV was like any other infectious disease: You threw as much medicine at it as you could, and eventually it would be subdued -- or eradicated. "We realized that there was no evidence that this was true for HIV," says Gazzard. So the Brits doled out drugs more sparingly. Treatment strategies were designed from the first with an eye on the long term.

Little of this Old World wisdom managed to cross the pond. Here, not only magazines (such as POZ) but bus shelters and subway cars were papered with protease ads. Physicians too were heavily marketed. Antiretrovirals became a $1 billion-a-year business. AIDS doc Keith Henry recalls being told repeatedly at pharma-sponsored seminars that the goal of treatment was undetectable viral load. "They were pressuring us to give out these drugs," he says. "They made it seem like it was almost malpractice not to put people on PIs if they had a detectable viral load."

A spokesperson for Merck, maker of indinavir, told POZ, "We had a breakthrough treatment for AIDS and we did everything we could to get it to the people who needed it."

The real problem, though, was that no studies were being done to pit the PIs against the two other HAART classes -- due in great part to drug company resistance to supporting head-to-head comparisons of their product with a competitor's. As a result, there was almost no data to test the assumption that PIs were truly more effective than non-nukes. "So much of our approach," says Emory University AIDS researcher Mark Feinber, MD, "was driven by marketing hype, not data."

In 1998, Julio Montaner was finally vindicated. An international team of researchers reported that DuPont's new non-nuke, efavirenz (Sustiva), was as effective as the PI indinavir in treatment newbies. While DuPont argued that its product was more potent than nevirapine, many scientists suspected that the drugs were equally powerful. The efavirenz study, in effect, confirmed the non-nuke results Montaner had reported with nevirapine three years before. Only this time around, doctors were willing to listen. "Once the efavirenz study came along it was like 'Gee, Julio, you were right,'" says Miles.

Soon after the efavirenz study was published, the tide began turning against PIs. It became standard practice to start treatment-naïve HIVers on a non-nuke combo, rather than a PI. The new openness was made possible by a sea change in HIV treatment. The dream of HIV eradication was finally undone when researchers discovered those "latent reservoirs" of virus, out of reach of the drugs. Looking forward to a lifetime on HAART and a growing list of side effects ranging from deformed faces to clogged arteries to tingling toes, HIVers began to revolt. If they didn't exactly throw their pills away en masse, they did compel the medical community to take seriously the risky, radical concept of drug holidays, or STIs.

Last year, the U.S. government lowered the "start HAART" time from 500 to 350 CD4s -- finally bringing us in line with the cautious Brits. "Who was right in retrospect? The British," says Miles. "Were we too proud to recognize this in 1997? Probably. Should patients take what we say with a grain of salt now? Yes."

Next May, the new PI, atazanavir (Zrivada), is due to hit the market. Relatively easy to take -- just two pills, once a day -- its main attraction is that it has so far shown none of the classic protease problems -- rising cholesterol and triglycerides and lipodystrophy. Some scientists predict that atazanavir will return PIs to the top of the heap. "If you could get rid of the toxicity problems," says Johns Hopkins researcher Joel Gallant, MD, "then you could restore PIs to the first line of defense."

But there are signs that the new skepticism will remain. It's by no means certain that atazanavir will remain free from the serious side effects of its predecessors. Many researchers now advocate making better use of drugs currently on the market, rather than putting a lot of blind faith in new ones that lack long-term data. And at press time, Robert Gallo, MD, stunned an overflow audience at the International AIDS Conference in Barcelona by opening his lecture on the future of HIV meds with this statement: "They are toxic and we are seeing resistance, and my guess is in two or three years, protease inhibitors will go away."

Britain's Brian Gazzard starts out almost every treatment-naïve patient on a non-nuke regimen. "If their CD4 count is 3 or 300, I put them on an NNRTI," he says. He argues that it's best for HIVers to ease into combo therapy with the easiest-to-tolerate regimen possible. More and more U.S. doctors are now following his lead. They are far more willing to start out relatively healthy, treatment-naïve HIVers on a non-nuke combo, reserving PIs -- still widely viewed as the most potent class -- as a first-line regimen for sicker patients.

The new watchword is humility. As Cornell University's Roy Gulick, MD, points out, "If you pose the question, 'Which are more potent, PI or NNRTI regimens?' we still don't know." According to Gulick, who did pioneering work on PIs in the mid-'90s, "We're voting more often against PIs right now. We're all now looking for less toxic, simpler regimens to start patients on."

There's even emerging evidence that an all-nuke regimen may work just as well as a PI-based cocktail. Last year, at the International AIDS Society conference in Buenos Aires, an international team of researchers reported that a combo of AZT, 3TC and a third nuke, abacavir, was as effective as one containing AZT, 3TC and the PI indinavir. While some remain skeptical, this study has sparked speculation about new treatment-sequencing options: Start with a nuke/non-nuke combo, switch to an all-nuke, and save protease drugs for last.

This study is so significant because it further clarifies that the true "miracle" of the so-called Protease Revolution was less the PIs themselves than the concept of using three or more drugs in one regimen. And this discovery was less rocket science than common sense. So what took so long? "Our simplistic little minds," says Northwestern University's Robert Murphy. "There's so much we just didn't know."

Cathy Olufs can sympathize with such uncertainty. "I've looked back many times and wondered if I made the right decision, going off meds," she says. "Who knows? I feel extremely lucky to have had the chance to make an informed decision that has worked out for so long. But I know I can't ride it out forever. I'll have to go back on meds pretty soon." The veins in her legs didn't disappear, she adds with a smile, but more important, they didn't get any worse. In a world where the mistakes outnumber the miracles, sometimes small favors are enough.