STEP's 12th World AIDS Conference Report

The 12th World AIDS Conference is the largest gathering of researchers, activists, drug company representatives, and doctors in the world. The last World AIDS Conference, held in Vancouver in 1996, gathered worldwide headlines with the arrival of protease inhibitors and the distant possibility of the eradication of AIDS. In comparison, the conference in Geneva seemed mild and sedate and seldom made the front page. But new and worthwhile information came out of the conference, which we will explore.

The conference was divided into four tracks: Basic Science; Clinical Science and Care; Epidemiology, Prevention, and Public Health; and Social and Behavioral Science. This review will focus primarily on the first two tracks. Rather than listing the most significant studies, we have listed the information as major themes, or pieces of the puzzle. We will focus heavily on the information that will have the greatest impact over the coming year on those living with HIV. This includes information on the efficacy, and possible long-term limitations, of combination therapy with protease inhibitors. We will also look at two valuable new tests -- phenotype and genotype resistance -- used to measure the sensitivity of HIV to various drugs. Additionally, we will discuss some of the implications that this new data has on the Public Health Services' guidelines about when to start anti-HIV therapy and what drugs to start with.

Long-term treatment does not lead to eradication

Three different research groups have now published and presented data showing that, even after several years on highly active anti-retroviral therapy (HAART), there remain in the lymph nodes a constant group of infected cells called latently infected CD4 cells. There are a relatively small number of these cells in the body, but their numbers remain the same even after several years of HAART.

Only one group, in New York, has presented data showing that these cells may decrease over a period of several years in people treated very early after HIV infection. Dr. David Ho postulated that either there is very slow decay, or that there is a significant decay of these cells, but that there is ongoing re-infection of new cells. He presented data that showed that even in people with levels of HIV replication below the levels of detection of very sensitive tests, newer and more sensitive research tests are showing there is still a small amount of HIV with ongoing replication.

This is not an irrelevant theoretical issue, because it has major implications on timing and intensity of therapy. If there is always going to be a pool of HIV-infected cells, which would cause a burst of HIV replication when therapy is stopped, then neither early nor long-term treatment has a chance of eradicating the virus completely. Thus, one should only need to start therapy before there is irreparable damage to the immune system. As discussed below, there was surprising data presented that demonstrates there is ongoing, and very significant, recovery of the immune system on HAART, even when treatment is started late, i.e. when CD4 counts are below 100 or 200. So the information presented at the Geneva Conference actually raised more questions about the timing of the initiation of therapy than it resolved.

HAART can produce long-term suppression of HIV replication (if viral replication is suppressed to very low levels)

Some data presented showed that people who were able to rapidly decrease viral load levels to below the level of detection (i.e., below 20 or 50 copies) were able to maintain that suppression for at least 2 to 3 years. The large majority (90 percent) of people in the Merck 035 protocol (which studied indinavir, AZT, and 3TC) who achieved a viral load of less than 500 by week 24, and who maintained that level for at least 12 weeks, were still below 500 viral load after 120 weeks on the trial.

HAART helps to rebuild the immune system

Several papers presented in Geneva showed that the immune system while on HAART can rebuild better than many researchers had expected. There are two general types of immunologic cells: naïve and memory cells. Naïve lymphocytes, or immune cells, are produced in the thymus, an organ in the chest that greatly decreases in size and activity with age. Naïve cells circulate in the lymph nodes and the blood, and when they encounter a foreign object or an invader such as a virus or fungus, they lead the body's attack on that invader. They reproduce and form large numbers of cells programmed to respond specifically to the invader, and thus become memory cells. So, in order to have a healthy immune system, you need adequate numbers of naïve cells, so you can protect yourself against new invaders.

However, with HIV infection, the body loses many of its naïve cells, and thus is poorly prepared to deal with new infections. Initially, after HAART, the immune cells that increase are predominantly the memory cells, the cells that are preprogrammed to respond to only one invader, and that are unable to respond to new types of invaders. In people who are on HAART, and who have very low viral load levels, there is a continual, gradual increase in naïve cells as well. When this occurs, it means that the immune system is healthier, and better able to respond to new infections.

The number of naïve cells is partially dependent on age, since after age 30 there is very little thymus gland remaining. This raises a serious question about the timing of anti-retroviral therapy. All the major advisory organizations recommend that HAART be started when CD4 cell counts fall below 500, based on the belief that there is irreversible immunologic damage if therapy is begun when the CD4 counts are much below 500. However, this recommendation needs to be re-evaluated in light of the immune reconstitution data presented at this meeting.

Alarming toxicities reported in many people receiving long-term HAART

The presentations at Geneva that received the greatest attention were those reporting on the long-term problems being increasingly reported by people taking protease inhibitor (PI) combinations. The studies of most concern were those reported by David Cooper and Andrew Carr, from Australia. David Cooper defined the signs of the syndrome that is now referred to as "lipodystrophy" associated with HAART. This syndrome is characterized by a shifting of fat from the face, arms, and legs to the belly, or abdominal area. This fat is deposited primarily within the abdominal cavity, rather than the abdominal wall. Fat is also increased in women's breasts. It may be deposited on the back of the neck, causing a dorsocervical fat pad, or a "buffalo hump."

Associated with this fat redistribution are elevations in blood fat levels, specifically triglycerides and cholesterol, and resistance to insulin, which may result in increased blood sugar, or diabetes. There have also been reports of heart attacks, or myocardial infarctions, in several relatively young men without other risk factors for heart disease. Also reported are problems with ingrown toenails, dry skin, and cracked lips, referred to as ectodermal dysplasia.

The Australian researchers reported that 74 of 116 people on PIs reported some signs of lipodystrophy, and 3 of 116 developed diabetes. Other researchers reported 7 of 29 people developed lipodystrophy on indinavir, and a French group reported that 8 of 32 people on indinavir developed lipodystrophy. A Swiss study reported that 18 percent of people had signs of lipodystrophy. Additionally, some women have had an increase of two bra sizes on HAART, and 21 of 116 women on HAART were concerned about changes in their body shape. Some women began to note changes after 4 to 6 weeks, but usually these changes take months to develop.

Other researchers have not reported as high a rate of lipodystrophy as Dr. Cooper's group, but in his study the changes were reported by patients themselves, and not their physicians. However, in Dr. Carr's study, which used body scans, 76 percent of people had signs of lipodystrophy, with a mean onset at 13 months of therapy. It should be noted that there is no standard definition for this syndrome and researchers are just beginning to look for these changes in a systematic way. Also, it is not clear if this is just a problem with PIs, or if some PIs are worse than others. In addition, the syndrome has been reported in some HIV-infected people not on any therapy, or on regimens which do not contain PIs. Overall, however, it is clear that this syndrome, in most people, is linked to PI use.

Triglyceride elevations as high as 200 to 300 percent above normal have been reported in as many as 65 percent of people on PIs. The greatest number of people with elevated triglyceride levels are on regimens that contain ritonavir. Dietary changes have not been very effective in lowering elevated triglycerides, and most drugs used to lower triglycerides cannot be used in conjunction with PIs because PIs change the way the liver metabolizes these drugs. One recommendation was clear: stopping smoking should help to lower the increased risk of heart disease caused by elevated triglycerides. Laboratory studies of insulin and glucose metabolism indicated that diabetes is a problem in only a few people, and these were people who already had a risk of diabetes, which may have been accelerated by the PI-containing therapy (see related story).

Protease-sparing effective alternatives

One of the more exciting presentations was a Phase II trial comparing a protease-sparing regimen, containing efavirenz, to a combination containing the protease inhibitor indinavir (IDV). Efavirenz (EFZ) was formerly known as DMP-266, and was marketed under the name Sustiva in the U.S. It is a member of the class of drugs known as non-nucleoside reverse transcriptase inhibitors (NNRTI), which currently includes the two FDA-approved drugs delavirdine (Rescriptor) and nevirapine (Viramune). EFZ is now available through expanded access programs. It is the most potent of the three NNRTIS, can be administered once a day, and appears to be very well tolerated.

The trial compared three treatment arms:

EFZ (600 mg once a day) plus IDV (1,000 mg every 8 hours) EFZ (600 mg once a day) plus AZT (300 mg twice a day) plus 3TC (150 mg twice a day) IDV (1,000 mg every 8 hours) plus AZT (300 mg twice a day) plus 3TC (150 mg twice a day)

A total of 450 participants were treated for at least 24 weeks. Entry criteria were CD4 greater than 50, and an HIV RNA count above 10,000. The best results were seen in the second group (EFZ plus AZT plus 3TC), where 94.5 percent had an HIV RNA result below 400 at week 24. This compares to 86 percent for the first group (EFZ plus IDV), and 64.9 percent for the third group (IDV plus AZT plus 3TC). When tested using the ultrasensitive viral load tests, 59 percent of the group treated with EFZ plus AZT plus 3TC registered a viral load below 50, compared to only 44 to 46 percent for the other two groups. In general, EFZ was well tolerated, with a minor rash being the most common symptom reported.

While only a preliminary report, it is the largest study reported to date comparing a regimen containing an NNRTI to a protease combination as initial therapy. Because there is cross-resistance among the four PIs currently available, there is a potential advantage to saving the PI combination for a second treatment combination. The long-term toxicities of lipodystrophy and elevated cholesterol and triglycerides have not been reported from NNRTI use, but it is too early to know whether these problems are PI-specific or not. Side effects reported from EFZ include difficulty with clear thought processes, disconcerting dreams, or depression for the first 2 weeks of treatment, but these problems usually resolve. There is also cross-resistance among the three NNRTIs, but EFZ is more potent than the two NNRTIs currently FDA-approved. DupontPharma has applied for FDA approval for EFZ and expects to receive approval later this year.

There are several large ongoing clinical trials comparing strategies for initiation of therapy. The trials compare three arms:

A PI alone An NNRTI alone A combination of a PI and an NNRTI

The AIDS Clinical Trials Group (ACTG) 364 protocol is studying the combination arm with nelfinavir and efavirenz in people who have been treated with only nucleoside analogs in the past. All participants also receive two reverse transcriptase inhibitors, such as d4T plus 3TC, or AZT plus 3TC. Preliminary data presented from this trial showed that the combination regimen and the NNRTI arm were better than the PI arm. However, it is too early to know if this difference will be maintained after the first 16 weeks. The ACTG has another trial, ACTG 388, which began in May 1998, comparing similar treatments in untreated people with HIV RNA counts greater than 80,000 or CD4 counts below 200.

Choosing the best therapy

Stefano Vella, of Italy, discussed current recommendations for when to start HIV therapy. He outlined the four major indications for therapy:

1. As a preventative after exposure (either after having sex or after occupational exposure)

2. To pregnant women to prevent transmission to the child

3. After primary (initial) infection

4. As therapy for people who have been infected for a longer time

Dr. Vella stated that with current therapy "HIV infection is potentially manageable on a decade time scale." Unfortunately, eradication is unlikely with currently available regimens. People must be prepared to make a long-term commitment, and decisions about when to start therapy should be individualized by level of risk, based on CD4 counts and viral load, because there is no new decisive information on when to start therapy. Other problems to be considered regarding when to begin therapy include the complexity of the dosing schedules, short- and long-term toxicity, cross-resistance, and drug-to-drug interactions. Rather than simply endorsing "Hit hard, hit early," Dr. Vella suggested that HIV infection should be approached like you would a shark. When dealing with a shark you must "Shoot early, but not too early, and remember that because you have one, maybe two shots, you must aim well."

Julio Montaner, of Vancouver, Canada, reviewed the currently available anti-retroviral combinations. He noted that it is difficult to compare regimens studied in different clinical trials because the trials use different inclusion and exclusion criteria for CD4, viral load, and prior anti-retroviral use. Because the entry CD4 counts, and viral load counts are all different, one cannot conclude that a combination that caused a fall in viral load to below 500 in 70 percent of participants is better than a combination that caused only a 50 percent suppression rate in another trial. Anther problem is that although some regimens may be better at reducing viral load to below 50, not all trials report data using the ultrasensitive viral load test that can detect levels this low. Also, viral load suppression does not account for all of the improvement seen in CD4 counts, because some people who have increasing viral loads while on HAART still experience significant increases in CD4 counts. So one must distinguish among virologic, immunlogic failure, and clinical failure.

Dr. Montaner strongly recommended that before therapy is initiated, the HIV-infected person discuss the specific goals of therapy with a health care provider who is very experienced in HIV. These goals may include how low a viral load is adequate and how to define treatment failure (for example, how high should the viral load go before a treatment is considered to have failed). The discussion should also look at the initial regimen, the next regimen after the first regimen fails, and how the requirements of these regimens fit into the person's lifestyle. Dr. Montaner stated that which regimen is chosen is less important than that the regimen chosen be one that can be adhered to for a long time. "There is too much at stake to leave it to those who are not properly qualified," he stated.

Easier-to-use anti-HIV regimens are on their way

Many combinations of the 13 approved drugs and the 4 drugs in expanded access were presented. Most of these trials are small, and the follow-up is short, but many of them are attempting to take advantage of the drug-to-drug interactions that may allow drugs that are currently prescribed three times a day, such as indinavir, nelfinavir, and fortovase, to be prescribed only twice a day. The following drugs all have favorable blood levels that may allow them to be administered only once a day: ddI, 3TC, nevirapine, efavirenz, and adefovir.

A very preliminary report compared the twice-a-day combination of efavirenz (300 mg every 12 hours) and indinavir (1,200 mg every 12 hours) to a three-times-a-day combination of efavirenz (600 mg in the evening) and indinavir (1,000 mg three times a day). At week 16, six of seven people in the twice-a-day regimen, and six of eight in the three-times-a-day regimen, had viral loads less than 400. This will be interesting to follow, as 71 people have been randomized on this trial so far. Another twice-a-day combination that appears to be effective in short-term trials is ritonavir (400 mg every 12 hours) plus nelfinavir (either 500 or 750 mg every 12 hours); however, 9 of 20 people experienced moderate to severe diarrhea in this trial.

Updates were presented of ongoing trials comparing indinavir and nelfinavir twice a day to three times a day. After 24 weeks, Nelfinavir at 1250 mg twice a day appears to be equally effective to 750 mg three times a day. The next 6 months should show if it is safe for all people on these drugs to change to twice-a-day dosing.

Wide-spread use of hydroxyurea in secondary treatment regimens (salvage therapy)

There were 12 reports at the conference regarding hydroxyurea (HU). Most reports found that HU could be safely added to ddI in salvage regimens, but one report found a significant increase in anemia and low white blood cell counts resulting from HU use in people with very low CD4. HU was reported to result in a greater decrease in viral load and many more clinicians are adding HU to second or third line regimens than at the Retrovirus Conference, only 5 months before this conference. HU does have a suppressive effect on the bone marrow, which can decrease CD4 cell counts, but this does not appear to have adverse clinical effects.

New anti-retroviral agents

There are four drugs currently in clinical trials that will probably be FDA-approved within the next 6 to 12 months. These are efavirenz, discussed above, adefovir, abacavir (formerly 1592), and amprenavir (formerly 141W94). Additionally there are three other drugs that are in clinical testing, but are not as close to FDA approval: ABT-378, taprinavir, and T-20 (fusin).

ABT-378 is a PI that is derived from ritonavir, and has the potential to be a second generation PI, or one which would be effective against HIV that is resistant to the currently available PIs. The drug is combined with a small dose of ritonavir, which greatly increases the availability, or blood levels, of ABT-378. A trial testing of this drug in people who are resistant to other PIs just began in May 1998. However, in a trial of people without prior PI use, Abbott Pharmaceuticals compared two doses of ABT-378 (400 mg and 200 mg twice a day), both containing 100 mg of ritonavir, plus d4T and 3TC. After 16 weeks, 26 of 27 people (94 percent) had viral loads less than 400 and at 20 weeks, 16 of 17 people had viral loads less than 400. To date, all participants were able to tolerate the drug and stay on the study. This is very preliminary data and the real test will be whether ABT-378 is truly a second generation PI, and able to suppress HIV replication in people resistant to the other PIs.

Adefovir, a nucleotide reverse transcriptase inhibitor (NRTI) appears to have a different resistance pattern than other NRTIs. However, it does not appear to be effective against HIV that is highly resistant to AZT. Recently it has been recognized that adefovir causes a type of damage to the kidney, known as proximal renal tubular dysfunction (PRTD), in many people on a dose of 120 mg for more than 6 months. The ACTG uses the 120 mg dose for 16 weeks, then decreases the dose to 60 mg, in an attempt to avoid this side effect. It is important that people on adefovir have their serum phosphate measured regularly to diagnose this problem before it causes serious kidney damage.

Amprenavir, like ABT-378, may be a PI that is effective against strains of HIV resistant to the current PIs. The combination of abacavir and amprenavir, which can be administered twice a day without any food restrictions, was reported to suppress viral load to below 500 in over 90 percent of anti-retroviral naïve people who had no prior therapy, at 20 weeks, and was well tolerated.

New tests identify drug resistance

Prior to the Geneva meeting, over 200 scientists and treatment advocates met in Italy to discuss HIV drug resistance and new treatment strategies. John Mellors, from Pittsburgh, presented a summary of the findings of this group in Geneva. There are two types of laboratory tests available to determine if HIV is resistant to a drug: genotype and phenotype testing. Genotype determines how the HIV virus looks genetically, then identifies the places in the virus where we know the mutations occur that cause the resistance that stops a drug's effectiveness. Phenotype testing actually measures the ability of the virus to infect cells and reproduce new viruses in the laboratory in the presence of known concentrations of drugs.

These two tests are complimentary, in that the genotype test will identify mutations that may precede actual resistance. (For some drugs, the virus needs to undergo mutation at several sites before it becomes resistant.) Phenotype testing will show directly whether or not the virus can grow in the presence of the drug. The major limitations of the use of these tests are that the virus must be isolated from a person's blood, which may be difficult if viral load levels are low, and these tests are very expensive, $300 for the genotype test and about $800 for the phenotype test. The genotype test can be done rapidly, and the phenotype test can now be done in about 2 weeks.

Neither test is FDA-approved, so insurance companies may or may not pay for them, on a case-by-case basis. Douglas Richman, from San Diego, in discussing the clinical use of drug resistance testing, stated that "prime time is coming quite soon" for drug resistance testing. However, he cautioned that you not get a laboratory test unless you know what you are going to do with the result.

Additionally, the resistance tests provide information on only the dominant clones, the groups of HIV viruses present in the blood at the time t hat the test was performed. There may be small numbers of HIV that are resistant to previously used drugs (referred to as "archived mutants") that will not be detected in these tests. Thus, it is still important to consider all prior treatment experience in designing a treatment plan, even when you have the genotype and phenotype data available.

The data from clinical studies presented Italy showed that these tests can be very helpful in predicting which drugs a person will not respond to when designing second or third line treatment for people who have failed one prior combination. However, some people fail to respond to a drug when these tests show they do not have resistance to that drug. There are many reasons for treatment failure, other than drug resistance, as discussed in the last two issues of the STEP Perspective. Nevertheless, while these tests are very expensive, they might in fact save money by reducing the use of drugs that the person has little or no chance of benefiting from.

Update on prevention of mother-child HIV transmission

Swiss researchers presented a study during the last morning of the conference that raised many eyebrows because it reported some major potential problems with the use of PIs in pregnancy. Swiss researchers reported that women who took PIs during their pregnancy had an increased rate of premature delivery (37 percent), an increased rate of bleeding into the brain of the newborn, and a rare blockage of the bile drainage tract from liver that required major surgery on the newborn child. Subsequently, the Pediatric AIDS Clinical Trials group in the United States has put a temporary hold (as of July 23, 1998) on their trials of PIs in pregnant women to further evaluate any possible risks from use of PIs during pregnancy.

Beware of how data is presented One of the most frustrating aspects of reporting on drug studies is that the groups of people in various trials are rarely comparable and data is often presented in very different ways. Trials always have specific entry values for minimal or maximal CD4 counts, viral load levels, and what type of prior therapy is allowed or not allowed. One trial might allow only people with a viral load less than 100,000, another might allow only viral loads greater than 5,000, and another any viral load above 500. Likewise, one trial may limit participation to CD4 counts above 500, another may allow only CD4 counts between 200 and 500, and another may allow any CD4 count. Additionally, one trial might exclude any prior anti-retroviral drug use, another might allow prior use of only AZT, and another might exclude only prior PI use. Thus, one trial studying AZT, 3TC, and protease inhibitor A cannot be directly compared to another study evaluating AZT, 3TC, and protease inhibitor B, or an NNRTI. These patient selection factors make it impossible to compare various trials, because the difference in results may have been caused by starting with healthier people, starting with less prior drug use, or other variables unrelated to the treatment the people actually received. Secondly, researchers and pharmaceutical companies often present their data in the best light. A poor study is less likely to get published in the medical literature and no company wants to have poor data circulating. There are various ways in which data can be presented. The best way to present data, statistically speaking, is to report the results for each treatment arm of the trial for all patients who began on that treatment arm, whether or not they completed the trial. (This is called "an intent to treat" analysis, as compared to an "on treatment" analysis). What good does it do to show data that reports that 90 percent of the people who completed treatment A had a good result, if only 50 percent of the people who began on treatment A were able to complete it? Another problem is how to handle missing data. For example, suppose some of the subjects in a trial were lost to follow-up or skipped appointments, and you are reporting on the percentage of people below 500 HIV RNA at week 52. How should you account for the missing subjects' data? You can either use the last known value for those people and assume they haven't changed (an optimistic assessment), or you can report them as a treatment failure (a more pessimistic assessment). the most conservative way to report data, meaning that the treatment is at least this good, is to report using an "intent to treat" analysis, with missing data handled as treatment failures. Conversely, the way to put the most positive spin on the data is to report using an "on treatment" analysis. These differences are not insignificant; a treatment regimen can result in a 70 to 80 percent suppression of viral load if reported by an "on treatment" analysis but a 50 to 60 percent suppression of viral load if reported as an "intent to treat" analysis. Additionally, one should be sure that the percentage of people with viral loads below 500, and below 20 or 50 using an ultrasensitive test, are reported. One regimen may result in more people being below 500, but fewer being below 50, and thus be less likely to result in long-term suppression of viral replication.

Back to STEP Perspective Fall 1998 contents page.