The Conference on Retroviruses and Opportunistic Infections is the most comprehensive yearly scientific meeting covering research on HIV and the treatment of HIV disease. Yet the thing that most impressed me at this year's meeting, in San Francisco, was not a data presentation or posters on an exciting new class of drugs or any other scientific breakthrough. It was the sight of my friend Luis, coming toward me in the conference hotel with black eyes and a bruised and bandaged face. I thought he had been mugged or savagely beaten. No, he cheerfully reassured me, he had just had cosmetic surgery in which fat was liposuctioned from his waist and injected into his cheeks. He paid to have it done because he had begun to notice his face grow a little gaunt -- a change that has been observed in many patients on antiretroviral therapy. Later that week, while eating at restaurants along the Castro, I watched people streaming up and down the street. The body composition changes associated with anti-HIV therapy, collectively referred to as lipodystrophy were everywhere, often among the most robust gym bodies. Mind you, just five years ago, this neighborhood was strikingly different. The very ill were a common sight, in wheelchairs, with catheters, often blind or wasting. Little of that is in evidence today. Thanks to antiretroviral therapy, the Castro's health in 2000 appeared dramatically better than the Castro in 1995. But even so, it was not quite right.
But what is more worrisome is that this is just the tip of the iceberg. There are potentially far more serious consequences of long-term antiretroviral therapy lurking unseen beneath the surface: metabolic changes leading to compromised organ systems, and now (fresh news at the conference) osteoporosis. Most grave are the increasing reports of heart disease, which has been observed in people with HIV before. While there is no definitive proof of an increase in incidence on HAART and there are always other risk factors, the cluster of metabolic changes associated with anti-HIV therapy is consistently associated with an increased risk of heart disease. One poster funded by Agouron noted that there was no increase in heart disease in patients taking different protease inhibitors over the course of 48 weeks, but the brevity of the study makes a mockery of long-term follow-up.
Meanwhile, I keep noticing heart attacks in the HIV community's obituaries. At the conference, I discovered that Steve Whitson, editor of the treatment news magazine Positively Aware, and a beautiful soul, had died of a heart attack just a few days earlier. While it is not clear whether antiretroviral-related toxicity contributed to his death, as with many such cases, one can never be completely sure.
No one can question that antiretroviral therapy has dramatically improved both the longevity and the quality of life of people with AIDS. A number of studies have shown that the benefits of antiretroviral therapy far outweigh its risks in people with advanced disease. However, in our initial euphoria over the great advance that HAART represented for people with AIDS, antiretroviral therapy has been routinely recommended and given to patients with HIV who were several years away from even being in danger of opportunistic infections
Now the tide seems to be turning against this practice. Many experts are questioning whether patients are being treated far too early and raising the alarm that antiretroviral therapy may be harming these patients more than helping them. At the Retrovirus Conference, Washington University's Dr. William Powderly surprised many in the audience by saying, "we now are looking at treating patients for a long period of time with drugs that have at least the potential for serious morbidity and possible mortality. Therefore, we should be asking the question of whether somebody who has no symptoms and is in no immediate risk of clinical progression should be exposed to these drugs at all or at least until they need them." A couple of weeks later, in an article in the Annals of Internal Medicine (February 15, pages 306-11), Dr. Keith Henry of the University of Minnesota presented a strong case against early treatment, concluding that "it is time for clinicians to rethink their approach to the treatment of HIV infection." Some speak even more plainly. In March, one very controversial article in Gear magazine by AIDS denialist journalist Celia Farber quoted UCLA's Dr. Steven Miles as saying, "large numbers of people are being inappropriately treated with drugs they don't need. And their lives are probably being shortened, yes" (see "Celia Farber and Denial in the Streets of San Francisco"). Still, when later interviewed by Treatment Issues, Dr. Miles worried that already the pendulum has swung "too far in the other direction" and that people who probably should be on treatment have quit or are afraid to start (see "Has the Pendulum Swung Too Far in the Opposite Direction?").
Hitting Hard or Striking Out against the Virus Early?
It's a central tenet of medicine to treat infectious disease as soon as possible. However, HIV is a very slow disease that takes a very long time to progress to life-threatening illness in most patients. It would be a mistake to treat it like a bacterial infection.
There were a number of reasons why early treatment of HIV disease initially seemed like, and may still be, a good idea. However, on close inspection most of these reasons were hypothetical and are yet to be clinically validated.
The major hypothesis now seems more than a little outdated.
"Perhaps more than any other assumption," writes Dr. Henry, "the hope that potent antiretroviral therapy could cure HIV infection within a reasonable time frame powered enthusiasm for immediately beginning treatment." This was the hope of the Vancouver AIDS Conference in 1996 -- that viral replication could be completely stopped by HAART -- leading in time to eradication as posited by Dr. David Ho. However, low levels of viral replication have since been found to persist in most patients on HAART, replenishing a small pool of infected cells. With the addition of immunotherapeutic agents, therapeutic vaccines, and other experimental approaches, eradication may yet occur, but again, these are experiments. "It is safe to conclude that a cure is extremely unlikely with the current approach to treatment," Dr. Henry wrote. Dr. Powderly concurred: "it is not for the foreseeable future an achievable goal."
However, access to viral load testing at roughly the same time as Vancouver played an important role in clinicians' preference for early treatment. Large cohort studies such as the Multicenter AIDS Cohort Study (MACS) demonstrated that viral load levels (or set points) were predictive of the time a patient had until progression to illness (see table below). Higher levels of virus predicted more rapid progression rates, and some clinicians used these data to argue for reserving treatment until absolutely necessary (just before the evolution of clinical events). But another important take-home message was that all but those with the lowest viral loads eventually progressed, although it could take more than a decade.
At the same time, these tests could be used to see the effect of treatment on the virus and when that treatment failed. At the very least, having an undetectable viral load gave the appearance that the disease was under control. But the difficulties in taking the medications (adherence and toxicity) and the rates of failure and development of resistance were much worse than initially hoped. Says Dr. Miles, "The problem is that the community did not weigh the pros and cons of treatment. We underestimated the risk of failure and toxicity and overestimated the benefit. At the same time, the commercial message from the pharmaceutical industry was that if you get the viral load down to zero, these drugs would last forever."
But they don't. And he continued, "zero was not zero; it was merely undetectable."
The truth is, treatment of patients whose immune systems are not seriously deteriorated and who have several years to clinical progression often leads to a year or so of undetectable virus and then the development of resistance. This runs counter to another one of the postulated reasons for beginning early -- that treating earlier gives the virus less chance to diversify in the body and develop mutant strains resistant to drug therapy. However, numerous studies suggest that, while mutant strains occur with each replication cycle, the mutations so slow down viral replication that mutant strains rarely get a strong foothold in the body in the absence of treatment. Only under the selective pressure of therapy does much resistance become established. In fact, this may even happen in patients responding well to treatment. One piece of bad news at the Retrovirus Conference, presented by Dr. Richard D'Aquila, was the detection of resistance in latently infected cells, in patients whose virus was normally suppressed to below 50 copies/ml on therapy, but who had experienced temporary 'blips' of viral activity (abstract 238). Such 'blips' are common, and now the data suggest that they may eventually lead to failure, even in those patients who appear to be having some of the best responses to therapy. Dr. Henry writes, "in my clinic some patients who were never at imminent risk for AIDS-related clinical events harbor virus strains that are resistant to all available drugs."
Resistance is made all the more likely because adherence to treatment is difficult, and the consequence of anything less than perfect adherence can be significant. In one study last year, patients on protease inhibitor-based regimens who described themselves as being at least 95% adherent had a 81% chance of having an undetectable viral load at week twelve, while those who were just a little less adherent (80 to 90%) had only a 50% chance of being undetectable.1 Naturally, the chance of a patient not being fully adherent only increases over time. Furthermore, people with HIV often have complicated lives which makes the likelihood of poor adherence greater, with problems such as poverty and disenfranchisement, low self-esteem, depression, alcoholism and illicit drug use. While these are extremely difficult issues to deal with, it may make more sense to treat the "context" of HIV infection before treating the virus with drugs. "Working with the patient for a long time about adherence and waiting for less complicated therapies may be a better long-term strategy," says Dr. Henry. Since HIV is often slow to cause illness, many patients could have the time to try to tackle these issues. The consequences of not doing so and being poorly adherent are too great.
One outcome could be exhausting the activity of treatment before patients really need it. Dr. Miles thinks we've been here before. For example, although AZT monotherapy had shown clinical benefit in people with AIDS, the Concorde study showed that the early use of AZT, a drug with limited effect and substantial toxicity, in patients with asymptomatic disease was no better, and in fact, was even slightly worse than delayed treatment. "But that study's message was lost on the medical community. That message was that if your drug's benefit is of a finite time, and you are not going to get sick during that time, then it is very hard to show clinical benefit," said Dr. Miles.
That benefit is even more difficult to demonstrate when it is being undermined by toxicity, which in many patients on HAART sets in long before danger of progression. Since HAART has only been in use for four years, the adverse events reported so far might be just a foretaste of what is yet to come. "The patients that I see today -- there's no way they are going to stay on these drugs for ten years," says Dr. Miles.
Dr. Powderly thinks that we have confused the short-term goals of antiretroviral therapy with the long-term goal of treatment: "The goal of antiretroviral therapy is to maximally suppress the virus with the best tolerated regimen, but that is not the goal of treatment. The goal of treatment is that we prolong meaningful survival for our patients." This would be easier to do if we had better drugs. "If you had perfect drugs that lasted forever, we would say treat everybody," says Dr. Miles. "But if our drugs aren't perfect and don't last forever, then we should only treat those at immediate risk of progression. Our dilemma is that we are somewhere in the middle."
Still, it would be a mistake to think that treatment at some point before progression offers no benefit. Therapy can delay AIDS if employed at the right moment. Studies have shown that HAART delays immunologic deterioration in patients and that should delay subsequent clinical progression. Even a dual nucleoside analog regimen delayed progression in ACTG 175 when used by patients with up to 350 CD4 cells at baseline. But studies have also shown that HAART can produce a miraculous immunologic recovery in more advanced patients. Ironically, such a positive response in people with AIDS provides theoretical support for delaying therapy until just before developing AIDS -- the more effective therapy is at reconstituting and sustaining health, the more likely it is that a patient will develop serious long-term complications of therapy before experiencing clinical progression. If this is the case, it might make sense to postpone treatment until the last possible moment.
However, other studies indicate that more sustained responses to therapy are seen in patients with higher CD4 cell counts and lower viral loads when treatment begins. These include two of the papers cited in Dr. Henry's article as evidence that the drugs don't work as well in the real world as in drug company sponsored studies. Henry writes, "In clinical practice, the rates of virologic success for potent regimens are lower than those seen in published studies; after one year, viral suppression often remains below standard detection limits in 37% to 60% of patients." However, the low rates of success in the studies cited by Henry were partly because therapy was delayed in some patients. In one of the studies, following patients in Southern Alberta, the study specifically reported that predictors of failure were having a higher viral load and lower CD4 cell count at the initiation of therapy.2 The same was true in the second study, conducted in an inner-city setting, which also found that factors such as missed clinic visits (which could reflect poor adherence) were predictive of drug failure.3 Both studies suggest that waiting to begin therapy increases the risk of early failure.
So when is the better time to start? How far from clinical progression, at what viral load, at what CD4 cell count does the clinical payoff from starting therapy justify the risks of toxicity or becoming resistant to the drugs? Unfortunately, we just don't know.
"Studies are needed to show how to use antiretroviral therapy to the best long-term advantage for the patient," said Dr. Powderly. Activists have been saying this for some time; in fact, one day just before the Retrovirus Conference the Treatment Action Group sponsored a symposium on "Long-Term Effectiveness Clinical Trials in HIV/AIDS" to address this issue (see "Long-Term Effectiveness Clinical Trials in HIV/AIDS").
Unfortunately, the question of when to start treatment cannot be adequately addressed without first characterizing the long-term side effects of HAART regimens. The exact contribution of each drug and each class of drug to the emerging toxicities needs to be mapped out. We need studies that look at all different types of regimens, including protease inhibitor and nucleoside analog-sparing regimens. For example, if nucleoside analogs are found to contribute more to long-term toxicity than protease inhibitors, then a dual PI regimen with, say, one of the more benign nucleosides such as 3TC may wind up being the best first-line approach to treatment. Or a different strategy could be to use more toxic but more potent regimens for first-line treatment, but only for a limited time. "We might increasingly see doctors choosing less toxic regimens or using the more toxic regimens for a while and then switching," says Dr. Miles.
We also need to recognize that the regimen that is most effective at 24 or 48 weeks, the current standard for regulatory approval, may not be the best regimen for a patient in the long run. That same regimen may cause greater toxicity or even a greater fatality rate after, say, four years.
"Let's say you have a study between two drug regimens," said Dr. Miles. "Ninety percent of patients are undetectable on one regimen at week 48 but only 50% are at year two because of drop-out from long-term side effects. On the other regimen only 70% are undetectable at week 48, but they stay undetectable until year two because there are no drop-outs because of long-term toxicity." It is common in drug studies today for the first regimen to be declared the victor at week 48, and patients rolled over onto that "more effective" regimen. The possible difference at two years goes unevaluated and is rarely even suspected.
What About Me? Choosing the Frying Pan or the Fire
Calling for more clinical trials is fine, but what if you are a person with HIV now considering or on early antiretroviral therapy? Dr. Miles in particular had much to share on this topic: "Now there is a clear recognition that these drugs cannot be taken for the duration. That is a given. But given this toxicity, where is it appropriate to intervene?"
If you have just recently seroconverted, the rationale for beginning treatment is slightly different than for those who have been infected longer. At this stage of infection, early treatment might be capable of preserving a virtually undamaged immune system, complete with an anti-HIV immune response that seems impossible to recover when treatment is postponed. Having an immune system that is primed to fight the virus may lead to more sustained treatment responses. Also, this immune response may be part of the reason why some people are "long-term nonprogressors" -- their immune system may better contain the virus.
But it is not at all clear yet that early treatment will turn a recent seroconverter into a long-term nonprogressor, or that these immune responses are clinically relevant. In the vast majority of patients who opt to discontinue therapy despite being undetectable for a year or two, the virus quickly rebounds (see "Structured Treatment Interruptions and HIV Immune Response").
"Post-seroconversion treatment is a wholly unsettled question because you are committing them to lifelong therapy without any evidence of benefit," said Dr. Miles. "Despite the preservation of HIV-specific immune responses, in the majority of patients who discontinue treatment, viral loads rebound, so it is not clear what those immune responses are doing. They should not be treated outside of an experimental protocol such as those at Aaron Diamond."
Most people do not find out that they are HIV-infected quite so early. And most have years to contemplate when to start treatment.
This decision should, of course, be made with the advice of your doctor but clearly treatment should begin before you are in danger of getting a life-threatening opportunistic infection such as PCP. This would mean starting therapy while you still have more than 200 CD4 cells. In the US, however, most doctors prefer not letting CD4 cell counts fall so low that there might be a danger of dropping to below 200 between clinic visits. Recently, the International AIDS Society modified its treatment guidelines, recommending starting treatment when the CD4 cell count falls below 350. However, some patients' CD4 cells counts could take years to fall from 350 to 200 cells. Again, data from the MACS suggests that both CD4 cell count and viral load should be consulted (see table below). A very high viral load (e.g. over 100,000/ml), may indicate which patients are in greater danger of a sudden drop in CD4 cell counts and at subsequent risk of clinical progression. Most patients with lower viral loads take years to progress. Frequently monitoring CD4 cell counts and viral load may be a better course of action than going on treatment.
"One of my patients with a low viral load is progressing but very slowly," continued Dr. Miles. "He will probably not progress to symptoms until he is 85. Should he go on treatment now? No. And if you were my patient, and I showed you the tables that came from the MACS. If that table says you have a four percent chance of getting sick over the next nine years -- would you go on treatment? Probably not."
Some people who have early signs of disease such as shingles or other conditions which can occur at higher CD4 cells may wish to begin treatment sooner. Dr. Miles cited one example. "I have an older patient around 57 with a pretty low viral load, but his CD4 cell count had been slowly dropping. Recently, it dropped to about 350 CD4 cells, and I suggested to the patient that maybe this was the time to do something. The patient wasn't sure that he wanted to stay on therapy, but decided to give it a six-month trial. I put him on a once-a-day regimen with nevirapine, ddI, and 3TC. Over the next couple of months, his viral load went undetectable, and his CD4 cell count began to creep back up to 500. At one of his visits, the patient said, 'You know, I actually feel a lot better on the medicine.' Little things that he never knew were being caused by HIV had gone away, and he had a lot more energy. I asked him whether he thought he would still go off therapy at six months. Now he said that he thought he might want to stay on the drugs."
As Dr. Miles' example illustrates, it is also possible to start early treatment on a trial basis and see how it goes. A decision to begin early therapy is not necessarily a commitment to stay on it indefinitely. Many patients do not seem to be suffering serious side effects. You might be one of them, and your response to therapy may be better than if you wait until your viral load is higher. You can always quit if you don't tolerate therapy well, or if your lifestyle can not presently accomodate the drug dosing schedules. But if you cannot adhere to therapy right now, it would be be better to stop the drugs entirely, as long as you are not in danger of clinical progression.
Many patients who have been experiencing serious side effects from therapy have already quit. But this is not a decision to be made lightly either. Chances are that your viral load and CD4 cell count will return to the point where you initiated therapy. Some patients should not even consider it. "For example," said Dr. Miles, "take two patients. One has a CD4 cell count of 43, and a viral load over 400,000. You put him on triple drug therapy and his viral load goes undetectable, and his CD4 cell count goes up to 750. Another patient has a viral load of 50,000 and a CD4 cell count of 600. You put him on therapy and he also goes undetectable, and his CD4 cell count goes to 750. Which one should not go off therapy?"
The patient who started out with advanced disease could be at great risk of risk of progression if he quits treatment. Unfortunately, many are doing this anyway (see "Has the Pendulum Swung Too Far in the Opposite Direction?"). But there are other options to consider first, such as a different treatment regimen or side effect management, or participating in a toxicity management study. The latter could have the dual benefit of possibly treating your condition, and providing solutions for which many other people with HIV are now waiting.
This may be the best reason to wait before beginning treatment. Tomorrow's antiretrovirals should be better. At the very least, there should be more answers about how to manage the toxicity.
1. Paterson D et al. 6th Conference on Retroviruses & Opportunistic Infections, Chicago, 1999 Abstract 92.
2. Mocroft A et al. AIDS. 1998;12:2161-7.
3. Lucas GM et. Ann Intern Med. 1999;131:81-7.
|Table: The Relative Risk of Progression to AIDS Within Three Years in the Multicenter AIDS Cohort Study
Viral load levels relate to results using the Roche Amplicor RT-PCR test. **Indicates a lack of data
Back to the GMHC Treatment Issues Spring 2000 contents page.