AIDS Treatment News
Vancouver in Perspectiveby John S. James
The XI International Conference on AIDS took place in
Vancouver, British Columbia, officially July 7-12 (but almost
all of the working meetings occurred in only four days,
Monday July 8 through Thursday July 11). No one could have
missed the news reports of major treatment advances, the
messages of hope. It is true that great progress has occurred
within the last year; the Vancouver conference brought this
progress to wider attention, and also released additional
The media also reported some of the limitations and need for caution on the new treatment approaches. Much remains to be learned about how to use available treatments best. Patient compliance with physicians' instructions is both increasingly important and increasingly difficult. The high prices of the drugs and combinations used will prevent most people in the world from having any chance of access, and are likely to interfere with medically indicated treatment in all countries. The media largely failed to report that much (although not all) of the new data are from the patients who are easiest to treat -- persons with relatively early stage HIV disease and little or no prior treatment. (While the same strategy -- reducing the viral load to near zero -- also applies to those with advanced illness, on the whole those with more advanced HIV disease are more difficult to treat, because the virus is harder to control, the patients are more likely to have side effects from the drugs and combinations, and prior treatment may have produced viruses resistant to some of the available drugs. Better drugs and better information on how to use them are both required.)
In short, while there has been progress, much more is needed; and meanwhile, application of the existing knowledge to individual patients will often be difficult.
What Is New?
Several recent changes have contributed to the central good news coming from the Vancouver conference. (There is much additional news from the conference as well, most of it positive; we will look at some of those other developments later.)
Point #4 needs more explanation. For several years there has been laboratory evidence that HIV develops drug resistance much more slowly when multiple drug combinations are used. Biological theory supports this observation; resistance to multiple drugs usually requires multiple mutations in the same virus, and since mutations happen during replication, if replication is shut down, new mutation combinations do not have a chance to arise. Viral load measured in the blood is becoming accepted as a good indicator of the amount of new HIV being produced in the body. If viral load can be reduced hundreds of times or more by appropriate drug therapy, the mutations needed for multi-drug resistance are likely to take correspondingly longer to occur.
While the theory and laboratory data have long been available, confirmation from human trials has been delayed, for at least two major reasons. First, the antivirals in wide use until recently were not powerful enough by themselves to shut down viral replication sufficiently to allow this delay of resistance to be easily seen in clinical trials. Also, companies have been unenthusiastic about testing their drugs in combination with those of competitors; recently they have learned that they must do such trials, because probably none of the available individual drugs is powerful enough to have a permanent antiviral effect by itself.
What is new recently, and has come to widespread attention since the Vancouver conference, is human data from trials of antiviral combinations sufficiently powerful to keep HIV replication low enough to prevent rapid development of resistance:
(1) Drugs which cannot be used alone (because HIV develops high-level resistance quickly) can remain effective for a long time if used in the right combinations. (No one knows how long this will last, as the trials are still ongoing.) In one example, the recently-approved drug nevirapine plus AZT plus ddI provided viral suppression comparable to protease inhibitor combinations but without using any protease inhibitor, in patients with CD4 counts from 200 to 600 who had not been previously treated. But some patients decided to stop taking the ddI (the most difficult of the three drugs to use), and then their virus developed high-level resistance to the nevirapine. The take-home message is that viral resistance can be greatly delayed, but only if the drugs are used correctly. This might require suppressing viral replication to negligible levels.
(2) Even more importantly, in some of the new trials of powerful combination treatments, viral suppression seems to get better with time and then remain stable, as measured by the proportion of the study volunteers in whom virus is undetectable in the blood by any test (there is still virus in the lymph nodes, but it is being suppressed by the drugs and not apparently reproducing significantly). In previous clinical trials, there would be a peak antiviral effect, usually within a few weeks, but then HIV would return as it developed resistance to the drug or drugs used -- meaning that it was only a matter of time until the treatment stopped working, or had greatly reduced activity. Now it appears that antivirals may remain effective for a long time. No one knows how long, since few of the new trials have run for more than a year so far (and only a few patients are providing the longest data).
HIV-related symptoms have often improved or disappeared in these trials.
While most of this research has focused on patients who are easiest to treat (because companies want to make their products look good), an important Merck study was able to suppress HIV to undetectable levels in over 80% of the trial volunteers -- despite the fact that they started this trial with a median CD4 count of 142, viral load over 40,000, and prior AZT treatment of more than two years. (But none of the volunteers had prior experience with the other two drugs tested in this trial.) And Abbott's previous study of its protease inhibitor ritonavir (seeAIDS Treatment News #240, February 9, 1996) found a strong survival benefit from adding it to existing antiretroviral therapy, in persons with CD4 counts from zero to 100.
It is widely suspected that when changing therapies, adding a single drug to a failing combination is usually a mistake -- that at least two new drugs, without cross resistance to treatments previously used, should be started simultaneously.
The early trials of multi-drug combinations (usually including at least one protease inhibitor) have suggested that for many patients, probably most, it is now possible to suppress viral replication to undetectable levels. Assuming these people keep taking the drugs as directed, no one knows if their disease will ever progress; at this time there is no evidence suggesting that it will. Without treatment, almost all of these patients would eventually progress to AIDS. And if they become careless in taking the drug, resistance is likely to develop, making the treatment unable to control the virus.
A practical problem in widespread use of this strategy of essentially complete suppression of viral activity is that, for best results, it should start as early as possible -- like treatment for any infectious disease. This means that many patients will still be relatively healthy -- and likely to remain so for some time. Will they be willing to start a regimen of 10 to 15 or more expensive pills a day, taken on an exacting schedule, some requiring an empty stomach and some requiring meals -- and continue this regimen for years, perhaps for life? There are likely to be side effects of the medications, especially in the first few weeks. And missed doses, "drug holidays," or other failures to use the drugs as directed could result in the therapy doing more harm than good due to the development of resistant viruses -- which not only would harm the individual, but also could be transmitted to others, depriving them of the benefits of the new treatments. Sometimes it might be best to defer treatment until one is ready to make the commitment to use the drugs properly. These are some of the issues that physicians, patients, and others must work with, in applying the developing scientific knowledge in widespread medical practice.
New View of by John S. James
The previous article outlined the new view of antiretroviral
therapy which has emerged in the last several months, and
which has been responsible for much of the optimism emerging
from the XI International Conference on AIDS in Vancouver.
Here we examine some of the research leading to the new view
of antiretroviral treatments. For this purpose, we selected a
few of the "late breaker" published abstracts to review. (The
"late breakers," at this and some other conferences, are
reports of new research too recent to be submitted by the
regular deadline -- which is usually months in advance of the
meeting -- but were judged important enough to include anyway
in a special session.)
[We are preparing a full report on how to obtain abstracts and other documentation from the conference. One way to obtain selected abstracts is through the World Wide Web, http://sis.nlm.nih.gov/aidsabs.htm. If you have the printed Vancouver conference abstracts, note that the late breakers are not included in the regular two-volume set, but are available in a short PROGRAM SUPPLEMENT, which was quickly published on newsprint and distributed at the conference. If you have the conference abstracts on disk or CD-ROM, the late breakers are included.]
Note: (1) In this article we did not include studies combining two protease inhibitors, because these need more extensive coverage. (2) Most of the abstracts sited have multiple authors; we listed the presenting author here.
Bi-Directional Inhibition of HIV-1 Drug Resistance
Selection by Combination Therapy with Indinavir and Reverse
This report from Merck Research Laboratories looked at the emergence of drug-resistance mutations in two different clinical trials, both of which compared indinavir (Crixivan(R), the Merck protease inhibitor) alone, vs. nucleoside analog therapy alone, vs. the combination. In one case, the nucleoside analog therapy used was AZT; in the other trial, it was AZT+ddI.
In both trials, the use of the protease inhibitor greatly reduced the development of resistance mutations to the other drug or drugs. In the trial testing indinavir vs. AZT vs. the combination, only one of 22 patients on the combination developed AZT resistance mutations, vs. 11 of 17 patients who received the same dose of AZT alone. In the trial of indinavir vs. AZT+ddI vs. the three-drug combination, zero of 20 patients in the three-drug combination developed either AZT or ddI resistance mutations, vs. 10 of 16 of the patients who received only AZT+ddI. Both these results were highly statistically significant, showing that using indinavir in the treatment regimen greatly reduced the development of viral resistance to the other antiretrovirals.
Going the other way, the nucleoside analogs also reduced the development of resistance mutations to indinavir -- but here the effect was less. Using AZT with indinavir reduced the indinavir resistance mutations from 9 of 21 patients to 4 of 22 -- a trend, but not statistically significant. Using AZT+ddI with indinavir reduced the occurrence of indinavir resistance mutations from 13 of 24 patients to 2 of 20, which was highly statistically significant.
Comment: The two classes of drugs (protease inhibitors and nucleoside analogs) work at different stages of the viral life cycle; there is no reason to suspect that one drug could directly affect resistance mutations to the other. Instead, what seems to be happening is that adding a drug to the regimen results in a lower level of HIV replication overall, so there is less chance for mutations to develop. This effect was strongest with indinavir, because it is the most active antiretroviral used in these trials; it was weakest with AZT alone, which was the least active antiviral tested.
This result gives measurable data on why combination therapy is important -- and emphasizes the need to follow medical instructions carefully when using antiretrovirals, so that the drugs are in fact reaching the virus at effective concentrations in the combinations prescribed.
Potent and Sustained Antiretroviral Activity of Indinavir
(IDV), Zidovudine (ZVD) and Lamivudine (3TC)
This study is important because it shows a strong sustained effect (at 48 weeks so far) of one of the most popular new three-drug combination treatments.
The patients in this trial had CD4 counts of 50-400, viral load over 20,000 copies, and at least six months of prior AZT treatment, but no treatment with 3TC or any protease inhibitor. They were randomly assigned to either indinavir (Crixivan(R), the Merck protease inhibitor) alone, AZT+3TC alone, or the triple combination of all three drugs. Viral load was measured at 24, 32, and 44 weeks; the viral load test used had a lower limit of detection of 500 copies per ml. There was relatively little drop-out, with 90 of 97 patients continuing on the study after up to 52 weeks of followup.
With the triple combination, at week 24, 92% (22 of 24) patients had viral load below 500 copies (undetectable on the test used). At 32 weeks that was 83% (19 of 23 patients). At 48 weeks it was 83% (5 of 6 patients; the numbers are small because the trial is still continuing and most of the patients have not reached the 48 week point).
For those on indinavir alone, the percentages with undetectable virus are 38% at 24 weeks, 36% at 32 weeks, and 22% at 44 weeks. For those on AZT+3TC alone, with no protease inhibitor, the percentage with undetectable virus was zero at all time points.
A Merck press release dated July 11 gave slightly updated figures, but the overall picture did not change. The press release also gave median CD4 cell increases at 44 weeks: 218 for the triple combination (for 7 patients completing the 44 weeks so far), 158 for indinavir alone, and 14 for AZT plus 3TC (in these patients with much previous treatment with AZT).
Comment: In this and similar trials, the measurement of most interest is usually the proportion of patients whose viral load goes below the level of detection of the test (as close to zero as possible to measure in the trial). The log viral load drops were also reported, but these are likely to understate the effect of a drug which causes many patients' viral load to drop below the test limit. For example, in this case, since the lower limit of detection of the test was 500 copies, anyone with under 50,000 copies at baseline could not possibly record more than a two-log (100-fold) drop, since the test used could not record the resulting level.
This is one of several clinical trials now showing that certain protease inhibitor combinations can largely shut off viral replication for a long time at least in a large majority of patients. It is impressive that the volunteers entering this trial were relatively advanced (median CD4 count 142, median viral load 41,130 copies) and with considerable prior experience with AZT (median 31 months), and the combination still worked. In further research it will be important to check viral load with the newer "third generation" tests (not yet commercially available), which go down to about 25 copies; it is likely that if a combination is starting to fail, that may first be signaled by a small rise in viral load, which undetectable with the 400-copy cutoff of the currently approved viral load test, and which could serve as an early warning to change the drug regimen quickly to keep the virus under control. But it is not yet known whether or not the more sensitive test would make much practical difference.
Triple Therapy with AZT and 3TC in Combination with
Nelfinavir Mesylate in 12 Antiretroviral-Naive Subjects
Chronically Infected with HIV-1
This trial enrolled 12 patients to test triple combination therapy with the Agouron protease inhibitor nelfinavir. One withdrew due to apparent drug toxicity (grade 4 CPK elevation); all of the other 11 had viral load below the limit of detection (500 copies with the test used) at 12 weeks. These patients entered the trial with a median CD4 count of 253 and a median viral load of 4.91 log (about 81,000 copies). The median starting CD4 count was 253, and the median increase was 98 at 12 weeks.
Triple Therapy with AZT, 3TC, and Ritonavir in 12 Subjects
Newly Infected with HIV
This study recruited 12 patients with very early HIV infection; antiretroviral treatment was started an average of 65 days after the symptoms of primary infection began. (Primary infection is the initial stage of very rapid viral growth shortly after HIV infection occurs, when the body has not yet developed an immune response to the virus; often there are severe flu-like symptoms.) Baseline viral load varied greatly, but became undetectable after treatment in all but one of the volunteers (one was slightly over the cutoff). Virus could not be cultured from any patient.
Incidentally, 9 of the 12 believed they knew when they were infected; this varied from 6 to 20 days before the start of symptoms.
Comment: Many researchers believe that primary infection may be the best time to begin aggressive antiretroviral treatment. Unfortunately primary infection is difficult to diagnose; many patients have no symptoms at all, probably few suspect HIV, and most will see a general-practice physician if they see a doctor at all, and be sent home with a diagnosis of "viral syndrome." If it is confirmed that primary infection is the best time to begin therapy, locating these patients will be a challenge for the medical system.
Another late breaker reported a single case of apparently beneficial treatment of primary infection, with AZT plus 3TC plus saquinavir. (Rapid Viral Load Decrease in Primary Infection Associated with Aggressive Therapy. By C Workman and others, #LB.B.6021.)
How Long Should Treatment Be Given If We Had an
Antiretroviral Regimen that Completely Blocks HIV
This research, at Los Alamos National Laboratory and the Aaron Diamond AIDS Research Center, looks at the possibility that HIV might be eradicated just by completely suppressing its replication for long enough. (Whether this will really happen or not is unknown at this time.)
In this study, mathematical models were developed based on data from eight patients treated with the experimental Agouron protease inhibitor nelfinavir plus AZT plus 3TC. There is a two-phase decay of viral load when HIV replication is stopped by drugs. In the first phase, viral load drops about 100 fold in two weeks; this is due to the clearance of free virus and the loss of productively infected cells. The second, slower phase of viral load decay appears to be due either to longer-lived infected cells such as macrophages, or to the gradual activation of latently infected cells; the researchers do not know which. In either case, the mathematical model suggests that completely inhibitory treatment would need to be used for about 1.5 to 3 years before the second phase burns out.
The researchers note that there could also be a slower third phase, or a "sanctuary" site in the body, which would invalidate this projection.
Comment: If HIV can be eradicated from the body just by blocking replication completely for a long enough time, then it may be possible to cure many patients with drugs which are already available. But no one yet knows whether this theory will work in practice. No patient has yet had HIV replication completely blocked for 1.5 to 3 years, since the trials of combination treatments powerful enough to stop replication have not run that long. In the few who have had to go off treatment early, high levels of virus have returned.
Improved Survival and Decreased Progression of HIV in
Patients Treated with Saquinavir (Invirase(TM), SQV) Plus HIVID
In this large saquinavir study, named NV 14256, there were approximately equal numbers of patients assigned to the three treatment arms (314 for ddC, 318 for saquinavir, and 308 for the combination). But disease progression occurred in 85 on ddC, 77 on saquinavir, vs. 46 with the combination. There were 28 deaths among those assigned to ddC, 34 assigned to saquinavir, but only 9 with the combination. Both differences are statistically significant.
These patients entered the trial with CD4 count between 50 and 300, and at least 16 weeks of prior AZT therapy. Median viral load at entry was over 100,000 copies.
Comment: Most of the interest today is in at least three-drug combinations, but we included this analysis of a large two- drug trial (saquinavir, the Hoffmann-La Roche protease inhibitor, vs. ddC, vs. the two combined) because it shows that even a less desirable protease-inhibitor combination treatment can make a substantial difference in disease progression and death. The relatively poor result with saquinavir alone is probably because the current dose of that drug is widely recognized as being too low. A new formulation which should correct the dosage problem is now in clinical trials.
At this time, the newest studies of the protease-inhibitor combinations which are believed to be the most desirable are relatively small, and started fairly recently, so there are no results yet on reduction of progression to AIDS and death. It is widely suspected (but not yet known) that HIV disease will not progress at all if viral replication is stopped. And it is clear that much immune recovery can occur.
Expert Panel's New Recommendations by John S. James
Major new guidelines for antiviral treatment of HIV infection
were published July 10 in JAMA ("Antiretroviral Therapy for
HIV Infection in 1996: Recommendations of an International
Panel," JAMA, volume 276, number 2, pages 146-154); they were
announced July 6 in Vancouver at a press conference by the
American Medical Association, and presented in detail to
hundreds of physicians and other health professionals at a
satellite meeting of the International Conference on July 10.
The new treatment guidelines were prepared by a 13-member panel appointed by the International AIDS Society-U.S.A. All of the panelists (most of whom were on stage at the July 10 Vancouver presentation) are leading HIV/AIDS clinical researchers, and also are physicians who care for patients. While the new guidelines are much improved and far more aggressive than previous ones, they are already being criticized by some researchers as not aggressive enough, since they allow but do not necessarily require an attempt to completely suppress HIV replication (as indicated by viral load being reduced to undetectable levels). These guidelines were largely put together by January of this year, with some later revisions to reflect new data. They will be changed periodically as knowledge of HIV disease changes; however, the first revision seems unlikely before late 1996 or early 1997.
The guidelines focus on "when to start therapy, what to start with, when to change, and what to change to." They also look at primary infection (the 4 to 7 week period immediately following exposure, in which viral load goes extremely high and often causes flu-like symptoms), treatment to prevent infection in case of needlestick, and treatment to prevent mother-infant transmission of HIV.
The 9-page guidelines document cannot be meaningfully summarized; persons interested should see the original article (referenced above). But some of the key recommendations will give our readers a sense of the guiding philosophy.
The guidelines recommend therapy for all patients with symptomatic HIV disease (which "includes symptoms such as recurrent mucosal candidiasis, oral hairy leukoplakia, and chronic and unexplained fever, night sweats, and weight loss"). For asymptomatic patients, therapy is recommended for all with a CD4 count below 500. And for those who are asymptomatic with CD4 count greater than 500, therapy is recommended for those with a viral load greater than 30,000 to 50,000 copies or with rapidly declining CD4 cell counts, and "should be considered" for those with greater than 5,000 to 10,000 copies.
What to start with? The guidelines include extensive discussion, mostly based on published data from clinical trials. A table recommends a number of possible nucleoside analog combinations for initial use, with or without a protease inhibitor in addition. For an initial antiretroviral regimen, possibilities mentioned are AZT+ddI, AZT+ddC, AZT+3TC, or ddI alone (the latter may be less effective with more advanced disease). Also mentioned (although less data are available) are ddI+d4T, d4T+3TC, and d4T alone. Also discussed at Vancouver was the class of non-nucleoside reverse transcriptase inhibitors (of which only one, nevirapine, has now been approved; other possibilities are delavirdine and loviride). These did not get into the current recommendations, due to lack of information at the time these recommendations were put together.
The guidelines listed three reasons for changing therapy:
What to change to? The guidelines include a table suggesting options to change to in case of treatment failure after initial treatment with any of five different regimens now in common use -- or in case of drug intolerance to three of those regimens.
Throughout the Vancouver conference there was much concern about the difficulty of following the new antiretroviral regimens -- and the dangers of developing viral resistance if the medications are not used as directed. Although not stated in the new guidelines, there is a widespread belief that it may be better to wait until one is committed to using the treatments correctly, rather than beginning treatment but then pursuing it haphazardly.
Also, there is increasing discussion about the need for considering long-term plans when making antiretroviral treatment decisions -- including what treatment options remain available if the virus develops resistance to the first drugs chosen.
After Vancouver: by John S. James
Leading HIV physicians knew about the major results presented
in Vancouver well before the conference, and had been using
some of the new treatments. Already there are anecdotal
reports of fewer deaths in physicians' practices, compared to
a year ago or more. We and others involved in AIDS know many
people who would have been expected to have died long ago,
based on all experience from the previous decade of the
epidemic, who instead are alive and doing well.
How can this improved survival best be documented, to show that better treatments do make a difference, and to support advocacy to make them more widely available? Official death statistics usually take a long time to become complete, as records move slowly through the system. And the totals by city, state, and other regions -- as well as the large cohort studies -- reflect an aggregate of those who get excellent treatment, poor treatment, and no treatment; these are figures we certainly want to know, but because they mix different populations, they are likely to underestimate the effects of improved treatment.
Another approach could be a telephone survey of major HIV medical practices which have been using protease inhibitor combinations and other new treatments. Physicians might be asked how many deaths, opportunistic illnesses, or other events they had in the last year compared with previous years, as a proportion of their total patients. They would also be asked other questions which could indicate other changes in their practice which might explain differences in event rates (such as accepting healthier patients, or referring those with more advanced illness to other specialists). The physicians should be promised confidentiality, with only aggregate results being released, to avoid creating an incentive for those with good statistics to selectively participate in the study.
A small preliminary survey could be done quickly -- and there is an important political reason to do it now. During the last several months it has been difficult to get increased funding for ADAP (the AIDS Drug Assistance Program) and other Federal programs to pay for the new treatments -- even though the cost of treating AIDS is still a fraction of the cost of many other diseases, even when combination therapy and protease inhibitors are included -- and even though the new treatments are likely to reduce hospitalization, infusion, and other major costs. The best time to get the attention and commitment of politicians is before the November elections. Unfortunately, while a few organizations have been working around the clock on these funding issues, most of the AIDS community has not yet mobilized. Documentation of reduced deaths could help motivate community support, and justify funding to make the new treatments available to more people.
A recently published epidemiological study at Johns Hopkins University has documented striking survival improvements between the period of 1985-88, vs. 1989-93; it attributed the lower mortality to widespread use of antiretroviral treatment and opportunistic infection prophylaxis (C. Enger, N. Graham, Y. Peng and others, JAMA, May 1, 1996, pages 1329-1334). The new treatment advances might save even more lives than the previous ones, in the practices and localities where they are being used.
Other Vancouver News
This issue of AIDS Treatment News has focused on explaining the central message of optimism about antiviral treatments which has come from the Vancouver conference. There are many other developments to report as well. We plan to focus on this conference for the next several issues, including reports on: