The Eleventh International Conference on AIDS, held last month in Vancouver, British Columbia, summarized the progress made during a watershed year for HIV treatment. As few as twelve months ago, many people with HIV were offered AZT monotherapy as initial therapy and switched to other nucleoside analogs only if they experienced a large drop in CD4 cell counts or the development of clinical symptoms. This sequential administration only produced a slight delay in the relentless progression to HIV infection. But some researchers forcefully contended in Vancouver that combining several of the available antiviral treatments could completely stop disease progression in its tracks and perhaps even clear the virus from the bodies of some patients.
Such optimism was the product of a number of developments that have occurred over the last year. First, studies such as ACTG 175 and Delta established that combination therapy works and can even prolong life when given to patients relatively early in the course of disease. Five new antiretrovirals have been approved since last fall by the Food and Drug Administration. Three of these are protease inhibitors which, according to Scott Hammer, M.D., of Boston's Deaconess Hospital, "are at the heart of the current optimism in this field."
During the same period, scientists have found that viral load is a very accurate predictor of one's risk for progression to AIDS and have tracked in greater detail how lowering HIV levels preserves health. "Viral load is important as a direct measure of how many cells are being infected and killed by HIV each day,"said John Coffin, M.D., of Tufts University,"CD4 count can be imagined as a measure of the current distance from progression to AIDS and viral load as a measure of how fast the patient is getting there."
Shutting Down the Virus
Indeed, treatment-associated "reductions in viral load correlate with protection from AIDS and AIDS-related death," said David Katzenstein, M.D., of Stanford University, based upon an analysis that he presented of the virologic data from ACTG 175 (abstract Mo.B.293). Supportive data also were presented at the conference from the Delta studies (which compared AZT monotherapy with AZT plus ddI or ddC) (abstract Mo.B.292). Both studies found that a one log reduction in viral load from baseline after eight weeks of treatment was associated with a long-term decrease in disease progression.
The antiviral effect of the nucleoside analog combinations in those studies was slight, though, compared to what was found in other trials reported at the conference. In two small studies conducted by Martin Markowitz, M.D., of New York's Aaron Diamond Research Center, viral load in treatment-naive patients taking more advanced combinations such as AZT/3TC with a protease inhibitor (ritonavir or nelfinavir) becomes undetectable even with the new ultrasensitive viral load assays that can detect as few as 25 copies of HIV RNA in a milliliter of blood plasma. If there is any ongoing HIV production, it is minuscule compared to what was seen before treatment. Such findings seduced even formerly skeptical clinicians and activists into contemplating the possibility of eradicating HIV from infected individuals -- if the virus cannot reproduce itself, then it is only a matter of time until cells infected with HIV die out.
There are some suggestions that extinction of the virus could be achieved in a few people with primary HIV infection or within a few months of seroconversion, even when treatment does not include a protease inhibitor (see article, Nipping HIV in the Bud; August 1996). But, HIV "eradication"may be an unfeasible goal for those in whom the virus has been replicating for years and spread to every possible nook and cranny in the body. In one presentation, David Ho, M.D., of the Aaron Diamond Research Center, calculated that it would take a year and a half to three years of therapy to eliminate all the cells in the body harboring HIV (abstract Th.B.930; see also Treatment Issues, article Throw Down Your Crutches and Walk; June/July 1996, ). This calculation, based on a mathematical model of the decay in viral load after the start of therapy with protease inhibitors, may not take into account small undetected reservoirs of virus that drugs are not penetrating sufficiently. Residual amounts of long-lived chronically infected cells might survive in such hard-to-reach tissues as those in the brain or testes. Additionally, the HIV trapped on the surface of follicular dendritic cells in the lymph nodes might reignite widespread HIV infection after treatment is stopped.
Avoiding Drug Resistance
Eradication is a worthy goal, but most clinicians and people with HIV would be glad if antiviral therapy could at least extend life by decades, not just the current couple of years. Long-term benefit of antiretrovirals can only be sustained if the emergence of resistant HIV does not cause drug failure. If viral replication is essentially halted, though, the development of resistance may not be inevitable. Participants in one study who had profound drops in viral load on the combination of nevirapine/AZT/ddI remained susceptible to nevirapine, a drug that loses its effectiveness through just one resistance-conferring mutation in the reverse transcriptase enzyme (abstract Mo.B.294)
Dr. Coffin has previously stated that resistance to three-drug combinations was likely (see Treatment Issues, article Throw Down Your Crutches and Walk; June/July, 1996). But even he conceded that from what he has seen of the data, the combinations appear to "provide long term suppression of virus replication without permitting the appearance of resistance." If he is correct, one of the main concerns about early treatment would be obviated -- physicians and patients would need not fear wasting their therapeutic arsenal before it is really needed in symptomatic disease. Nevertheless, the nature of the putative HIV reservoirs is poorly understood. Only time will tell whether viral hideouts eventually become the source of resurgent HIV.
Those failing on current therapies due to resistant HIV face a great dilemma. As Margaret Fischl, M.D., of the University of Miami noted, "merely adding one new agent to a failing regimen is doomed to failure."One way to overcome HIV's evolutionary propensities is to increase the number of novel agents that could be introduced to the therapeutic mix. At least nine new drugs are entering clinical trials, potential complements to the nine now for sale in the U.S. Among these are new protease inhibitors and exceptionally potent reverse transcriptase inhibitors requiring unique, frequently multiple, mutations for full-scale HIV resistance. Compounds exploiting other antiviral targets such as HIV's zinc fingers or its integrase enzyme also are approaching full-scale human testing.
Another possible option could be the new genetic assays that help determine whether a person's virus has become resistant to a drug. Finding which drugs are failing in a combination regimen would allow a person to drop just those agents and not the entire regimen. On the other hand, such tests cannot show whether a drug fails for reasons other than resistance, such as a change in the way the drug is absorbed or processed in the body.
A Note of Caution
The optimism of the conference was rather intoxicating. It is clear that a decade of AIDS research is beginning to come to some fruition, but the fruit has yet to fully ripen. "The degree of meaningful clinical benefit of the newer regimens needs to be defined," Dr. Hammer, M.D. said in the conference's opening plenary. "The current optimism is warranted but needs to be tempered with caution and, most importantly, complacency that we have 'arrived' needs to be avoided."
There is a danger that in the excitement over those who have had an unprecedented response to antiviral therapy, we will overlook those who do less well. Already it is apparent that not everyone who takes the combination regimens enjoys the same dramatic reductions in viral load or indefinite stabilization. People with advanced disease commonly harbor HIV resistant to many of the nine approved drugs, due to prior treatment with sequential monotherapies or weak combinations consisting of two nucleoside analogs. There are growing ranks of those who have become resistant to protease inhibitors after adding them to regimens of nucleoside analogs. Cross-resistance and side effects may render useless anti-HIV agents patients have not yet tried.
Continual compliance with three different dosing schedules may turn out to be next to impossible over the long run, and exposure to suboptimal drug levels provides HIV with a prime opportunity to improve its resistance capabilities and rebound. Furthermore, the limits of immune system recovery are becoming evident. Development of immune reconstitution therapy has been even more laggard and haphazard than the search for antiviral agents, many of which have languished for years. Both fields need to be developed in tandem.
Finally, in Dr. Hammer's words, "access and cost are major obstacles to underserved populations throughout the world," including many here in the U.S. In her presentation (abstract Mo.B.532), Dr. Fischl estimated that the cost of three-drug therapy (including a protease inhibitor) plus CD4 and viral load monitoring would cost her hospital in Miami and its 2,200 patients with HIV a total of $21,850,000 per year, an amount she termed, "an insurmountable figure" Meanwhile, a physician from Thailand, Chaiyos Kunanusont, recounted how a portion of the infected population in Thailand could afford old-guard nucleoside analog therapy "but they would have to eat less" (abstract Mo.B.533).
For people with HIV, but not in imminent danger of progression (i.e., those with low viral loads), one option is to wait for another six to twelve months while the medical and economic issues sort themselves out. Beginning a treatment regimen that is too difficult or costly to sustain or that does not essentially eliminate HIV replication may be more dangerous than putting off a decision. But the delay cannot be indefinite, as early, effective treatment will preserve a more functional immune system.
People with low CD4 counts or higher viral loads may not have the option of waiting nor should they need to. Whatever the caveats, the Vancouver conference left participants with the feeling that a major step forward has been taken. Ninety-nine percent and more of the HIV can be eliminated in most people. The question remaining is how many steps more need be taken to reach 100 percent elimination in everybody. Until that goal is approached, long-term management of HIV will continue to be a desperate juggling act in which the particular drug combinations prescribed to each individual are frequently adjusted to keep pace with viral evolution.
Antiviral treatment "could ultimately lead to the clearance of HIV" in people recently infected with the virus, according to Bradford Saget, M.D., a general practitioner from San Francisco. He and other researchers at the International Conference presented dramatic results on the treatment of those recently infected with HIV. "These are the easiest patients in whom to test the hypothesis [that HIV can be eradicated]," Martin Markowitz, M.D., of New York's Aaron Diamond Research Center told the Conference, "because early in infection, we're dealing with homogenous virus populations that we hope are susceptible to all the drugs we're using."
Dr. Markowitz has not yet had a chance to verify this hypothesis. Although all the eight recently infected individuals who could comply with his hard-to-tolerate AZT/3TC/ritonavir study regimen have undetectable viral loads (under 25 copies per ml), proof will have to wait until some of them choose to stop treatment. The earliest Dr. Markowitz's study permits this is when, after nine months of undetectable viral load, patients elect to undergo optional lymph node biopsies to ensure that infectious HIV is not harbored there. Based on what they find in the lymph nodes, the patient will make the decision whether to stop therapy. This will be several months away, at best.
Other researchers presented more mature data on less potent combinations that nevertheless achieved very impressive results in the same population. According to Luc Perrin, M.D., of Geneva, Switzerland, "treatment with AZT/ddI reduces [viral load] to undetectable levels in the majority of patients [with primary infection]." This has been sustained for twelve months in nine out of twelve people in Dr. Perrin's study. Many of these patients are gradually losing antibodies to HIV. Four subsequently became negative in a test for proviral DNA, indicating the absence of latently infected cells containing nonreplicating HIV hidden inside. Nor could the virus be detected in the lymph nodes of these patients.
Similar outcomes have not been reported in untreated patients, and these are clearly superior responses than AZT/ddI achieves later in the course of HIV disease. It still is not clear whether the virus has been eliminated in any of Dr. Perrin's patients, though, as he has not yet risked taking them off therapy.
Other clinicians who took slightly different approaches found similarly impressive results with AZT/3TC and AZT/ddI/3TC treatment (abstracts We.B.531 and Mo.B.1331, Mo.B.1332, Mo.B.1337). A team from Australia decided to use everything but the kitchen sink. They started one patient with AZT/3TC, added high dose saquinavir (7,200 mg per day) after 26 days, and then threw in d4T 53 days later. Viral load dropped by more than six logs (one million-fold) to below detectable limits by week ten, and the patient's CD4 cells increased by more than 200 cells.
In a similar vein, but without a protease inhibitor, Dr. Saget excited conference attendees with case studies in six patients treated within six months of seroconversion using the improbable combination of AZT/ddI/ddC/interferon-a plus, sometimes, 3TC. (The doses of all of the drugs were rather low, and individual drugs were discontinued if the patient could not tolerate them.) As a result of this treatment, all of the study participants had undetectable viral loads (below ten copies per ml), down from a baseline average of 62,000 (range 11,000-227,000). CD4 cells increased, and other measurements of the immune functions normalized.
One of Saget's patients went off treatment after 87 weeks. Unfortunately, there was an immediate rebound in viral load, which subsequently went below detection when treatment was reinitiated with AZT/3TC. This patient previously had an indeterminate result on a proviral DNA test. Dr. Saget concluded from this experience that "if we are going to remove someone from therapy, their proviral DNA loads must be negative."
Other researchers differ on this point, however. Dr. David Ho of the Aaron Diamond Research Institute notes that proviral DNA levels decreased in Dr. Markowitz's ritonavir/AZT/3TC study but only very slowly. He believes that most of the residual infected cells detected by the proviral DNA test are infected with defective forms of the virus that cannot replicate. Such cells will not be destroyed prematurely because the virus within them will remain inactivated. The cells could have a normal lifespan, and the defective HIV DNA will continue to be measurable as long as the cells survive.
The actual test is still whether the patient can be taken off therapy without a rebound in viral load. Aside from Dr. Saget's failure, there was only one group with a report about terminating therapy. One French team using AZT/ddI/3TC (abstract Mo.B.1332) stopped a patient's therapy after 52 weeks. This patient's viral load remained undetectable for several months, then blipped up, only to become undetectable again without reinitiating treatment.
What this anomalous result means is anyone's guess. But whether this patient is "cured" or not, much of the data presented made many clinicians consider whether they should be actively screening patients for primary HIV infection, and immediately offering treatment. Dr. Perrin clearly thinks so: "As a private physician, shall you look at every patient with a flu syndrome [to see] if he has a primary HIV infection? I think you should, and you should treat all these patients."
Such advice is based on a presumption that early potent chemotherapy will eventually lead to eradication of the virus in these patients, and that treatment can then be discontinued. Otherwise, the long-term toxicity of the drugs and the potential risk of drug-resistant HIV may far outweigh any benefit offered by early treatment. It is also worth noting that these studies so far cover only a few dozen persons, a weak basis for establishing medical care for thousands. But those currently in the throes of primary infection and accepting treatment can always quit later if these advance studies do not produce the hoped for results. At the conclusion of one Conference session, Treatment Issues' own Gabriel Torres, M.D., noted from the floor that, "despite the caveats and the caution about the fact that these are short-term experiments, those of us who take care of people who are recently infected are going to take back this information and apply it."
Study Regimens (N†): nelfinavir/AZT/ 3TC (N=11)
Baseline Viral Load: mean: 156,000, range: 14,000- 648,000
Baseline CD4: 245, range 26-504
Viral Load Response: reduced to below 25 RNA copies/ml* in all participants (wk 16, N=11)
CD4 Count Response: Up more than 100 cells/ml (wk 16, N=11)
Notes (conference abstract): Participants were all treatment-naive (Th.B.930)
Study Regimens (N†): ritonavir/AZT/ 3TC (N=12)
Baseline Viral Load: median: 10,423, mean: 91,389
Baseline CD4: median: 564, mean: 633
Viral Load Response: reduced to below 25 RNA copies/ml* in all participants (wk 16, N=9)
CD4 Count Response: Up more than 200 cells/ml (wk 16, N=9)
Notes (conference abstract): People with early infection. Three persons lost due to non-compliance and intolerance to ritonavir. (Th.B.933)
Study Regimens (N†): Simultaneous ritonavir/AZT/ 3TC vs. adding AZT/3TC (N=33)
Baseline Viral Load: 251,189
Baseline CD4: Over 500
Viral Load Response: ~ 3 log (1,000-fold) reduction
CD4 Count Response: +170 cells vs. +128 cells (immed. vs. delayed AZT/3TC)
Notes (conference abstract): AZT/3TC added after 22 days in delayed group. (Source: Sven Danner, M.D., Ph.D, during satellite symposium.)
Study Regimens (N†): indinavir/AZT/ 3TC vs. indinavir alone vs. AZT/3TC (N=97)
Baseline Viral Load: median: 41,130
Baseline CD4: 144
Viral Load Response: 2 log (100x) drop vs. 1 log (10x) drop vs. 0.4 log (2.5x) drop (wk 36, N=77)
CD4 Count Response: +120 cells vs. +100 cells vs. +40 cells (wk 36, N=77)
Notes (conference abstract): 80% AZT resistant at baseline. Between 80-90% of patients on three drug combination have viral loads below 500*. Majority of patients have completed 36 weeks, small number at 48 week s. (Th.B.931)
Study Regimens (N†): saquinavir/AZT/3TC N=33
Baseline Viral Load: mean 95,499
Baseline CD4: mean: 326
Viral Load Response: 1.96 log drop at week 4 (N=30)
CD4 Count Response: +107.5 cells
Notes (conference abstract): (Mo.B.171)
Study Regimens (N†): ritonavir/AZT/ ddC (N=32)
Baseline Viral Load: 50,100
Baseline CD4: mean: 170, median 152
Viral Load Response: 2 log drop, 50% under 200* (both week 24, N=21, and week 60, N= 17)
CD4 Count Response: +141 cells week 24, +182 cells week 60
Notes (conference abstract): Eleven patients quit during the first six months, most due to ritonavir toxicity. At 24 weeks, four more people dropped out of the study. (Mo.B.175)
Study Regimens (N†): nelfinavir/d4T (3 doses) vs. d4T (N=36)
Baseline Viral Load: 56,000- 81,000
Baseline CD4: 310-390
Viral Load Response: day 60 1.7-2.4 log drop for nelf. arms (75% under 500*), 0.9 log drop for d4T, 5 months: 1.3-1.8 log below baseline
CD4 Count Response: day 60: +105 to 130 cells 5 months: +87 to 111 cells
Notes (conference abstract): d4T naive. After two months d4T monotherapy arm added nelfinavir. (Mo.B.413)
Study Regimens (N†): saquinavir/d4T (N=14)
Baseline Viral Load: median: 158,500
Baseline CD4: median: 33
Viral Load Response: 1.6 log drop week 4, 0.9 log drop week 8
CD4 Count Response: +57 cells week 8
Notes (conference abstract): Saquinavir added to d4T after median of 4 months of d4T pretreatment. (Th.B.945)
Study Regimens (N†): ritonavir 400 mg bid/saquinavir 400 mg bid, or ritonavir 600 bid/saquinavir 400 mg bid
Baseline Viral Load: means: 31,600, 47,900
Baseline CD4: means: 274, 279
Viral Load Response: 2.14 log decrease at week 4 (N=60) and 2.42 log decrease at week 6 (N=43)
CD4 Count Response: +76 cells at week 4; +98 cells at week 6
Notes (conference abstract): Batched data from both doses. 86 % of participants have had at least 2 log reduction or undetectable viral loads at week 6. Toxicities include diarrhea, nausea, facial tingling, fatigue -- only two people have dropped out. (LB.B.6041)
†N= number of individuals studied.
*Limit of detection for viral load assay used in this trial
The impressive data from the International Conference on protease inhibitor combination therapies are summarized in this data. As is evident from the table, the results are still scanty. They represent only a few people, generally on therapy for relatively brief periods. Especially remarkable are the first two studies, conducted by Martin Markowitz, M.D., at New York's Aaron Diamond Center. Dr. Markowitz's studies, the initial ones to try triple combination therapy in persons without any previous drug exposure, indicate that it is possible to essentially shut down HIV replication in such people. Note, too, the striking preliminary results from the ritonavir/saquinavir combination study.
The table omits most of the toxicities reported for these therapies. Side effects particularly appeared to plague ritonavir, according to the Conference reports. Concluding his presentation on the sustained clinical efficacy observed after a year, Abbott Laboratories' watershed study ritonavir added to current therapy in people with advanced disease, William Cameron, M.D., of the University of Ottawa tellingly noted that, "ritonavir is safe and well-tolerated in some. " [emphasis added]
At the Fifth International Workshop on HIV Drug Resistance held prior to the Conference, Amy Patick, Ph.D., of Agouron Pharmaceuticals reported on the development of resistance to the company's protease inhibitor nelfinavir in patients on monotherapy. Early data suggest that HIV that became resistant to nelfinavir in these study participants may remain susceptible to subsequent therapy with the other protease inhibitors, including ritonavir and indinavir.
Meanwhile, Glaxo Wellcome continues to assert that many people who fail on ritonavir or indinavir will remain sensitive to treatment with its protease inhibitor 141W94. If the company finally demonstrates that its drug is worth taking and if Agouron's data is confirmed, there is a possibility that, if used in the proper sequence, protease inhibitors could be included as part of an initial antiviral regimen early in disease with little risk of developing HIV cross-resistant to all other available protease inhibitors. One may be able to take at least three protease inhibitors sequentially without losing the demonstrated benefit this class of drugs has in advanced HIV infection.
Protease inhibitors may have grabbed the spotlight in Vancouver, but there also was good news from those old standbys, the nucleoside analogs. The presentation concerning the new nonnucleoside reverse transcriptase inhibitor delavirdine was less impressive, but the data on nevirapine (see page 8) indicates that it may not have been tested under optimum conditions -- combined with two other drugs in treatment naive patients with comparatively early HIV infection.
1592U89 Succinate's Striking Results
Further data on the unexpectedly strong antiretroviral effects of 1592U89 succinate, a guanosine nucleoside analog being developed by Glaxo Wellcome, were presented at the Vancouver conference (abstract Th.B.294). "1592" is synergistic in the test tube with other nucleoside analogs and protease inhibitors and is active against AZT-resistant HIV. This new agent is absorbed well when taken orally, and it also has significant (18 percent) central nervous system penetration. Since it is not metabolized by the liver, 1592 should not have serious drug-drug interactions with the protease inhibitors.
The phase 1-2 study presented at the International Conference comprised three groups of 20 persons each with CD4 counts between 200 and 500 and less than twelve weeks experience on AZT. Participants were sequentially enrolled into three 1592U89 monotherapy cohorts (200 mg thrice daily, 300 mg twice daily or 400 mg thrice daily). After four weeks, AZT (600 mg/day) or placebo was added for an additional eight weeks of follow-up. At baseline those in the 200 mg thrice daily cohort had a higher median viral load (125,000, or 5.1 logs) than those in the other two cohorts (about 32,000, or 4.5 logs). All groups had similar average entry CD4 counts (355 to 381).
After four weeks of follow-up, the three groups had similar reductions in viral load (97 to 99.4 percent, or 1.5 to 2.2 log). This increase was stable through week twelve: An analysis of those receiving 1592U89 monotherapy during the entire study period compared to those who also received AZT during weeks four through twelve found that there was no added viral load benefit from the second drug. Median increases in CD4 counts were also similar among the three groups -- an average of 100 cell from either 1592U89 alone or in combination with AZT.
Four patients discontinued the study, two because of fever and rash and two others because of fatigue and nausea. A fourth cohort using a higher dose (600 mg thrice daily) is fully accrued and being followed.
No data are yet available on resistance, although one trial participant did develop a mutation on the 184th codon of the reverse transcriptase gene. This is the mutation that confers resistance to 3TC, but in the test tube, at least two additional mutations (at codons 65, 74, or 115) are required to reduce the sensitivity of HIV to 1592. Glaxo Wellcome plans to develop 1592 during 1996 and 1997 by studying it in combination with its other nucleoside analogs, AZT and 3TC, in three studies (one pediatric) that look primarily at changes in CD4 count. Disease progression (clinical endpoint) studies are also in the planning stage. The company anticipates obtaining accelerated FDA approval for this surprising drug by late next year.
Lobucavir Finally Gets a Little Attention
Lobucavir, formerly known as cyclobut G, is a guanosine nucleoside analog, like 1592U89. First developed in the late eighties, the compound has languished at Bristol-Myers Squibb while sister compounds such as d4T were rushed to market. Now it appears that lobucavir has activity against most viruses in the herpes family and hepatitis B and is a potent inhibitor of HIV as well.
In an understated and little noticed presentation at the International Conference on AIDS (abstract Th.B.943), Lisa Dunkle, M.D., of Bristol-Myers, reported on a four-week study in which lobucavir eliminated CMV shedding in the urine of people with AIDS. Six individuals receiving lobucavir at the highest dose, 400 mg four times daily, saw their plasma HIV levels fall by 98.7 percent after four weeks. No drug-related toxicities were observed. Dr. Dunkle also cited observations indicating that lobucavir can cause oral hairy leukoplakia and drug-resistant herpes simplex lesions to resolve. In animals, lobucavir reduced hepatitis B levels by 99.99 percent.
Despite the recently evident HIV activity, Bristol seems to be still focusing on its use against other viruses. In an interview, Dr. Dunkle commented, "There's only a very tiny amount of HIV data, which need confirmation. The HIV activity may be an important component for treating opportunistic infections in AIDS patients." Although Bristol now claims to be committed to rapidly developing lobucavir, no specific anti-HIV trials, not even a short viral load pilot study, are in the works. Bristol is planning a large CMV retinitis trial to commence late this year. That study will enroll 100 to 200 participants and will take a year and a half to complete. Also under discussion is a year-long opportunistic infection prophylaxis study in 750 to 1,000 people with AIDS that could start early in 1997.
ddI/d4T Exhibits Safe, Potent Effect
ddI/d4T combination is a popular regimen that many have been using since a preliminary report presented by Richard Pollard, M.D., from the University of Texas at the Third Conference on Retroviruses and Opportunistic Infections last January described an enhanced antiretroviral effect from the two drugs together without the feared additive neurotoxicity. At the Vancouver meeting, Dr. Pollard updated the results of the study (abstract Th.B.293) by presenting the one year follow-up of 79 patients who were randomized to receive one of five different dosage regimens -- ddI at 200 or 400 mg/day combined with d4T at 20, 40, or 80 mg/day.
Trial participants had entry CD4 counts between 200 and 500 (average of 343) and a mean viral load of 16,000 (4.2 logs). After one year of follow-up, the five groups as a whole had an average viral load reduction of 96 percent (1.4 logs) and CD4 increases of 80 to 100. Only one patient has developed a grade II peripheral neuropathy. There were no significant differences among the five groups in terms of antiretroviral response, although it seemed that those receiving the higher doses of both drugs had a trend towards an enhanced antiretroviral response (31 to 33 percent achieved viral load reductions of more than 99 percent (2 logs) as compared to 18 to 20 percent of those on the lower doses).
Peter Carberry, M.D., of Pharmacia & Upjohn, presented data on the effect of adding delavirdine to AZT or ddI (abstract Mo.B.295). Sixty percent of the 800 patients in the dose-ranging AZT/delavirdine study had never taken AZT. In the study, the combination at higher doses of delavirdine (300 or 400 mg three times daily) reduced viral load significantly more than the AZT and low dose AZT/delavirdine combination. The high dose delavirdine arm sustained a 70 percent reduction in viral load, out to week sixty, which was accompanied by a durable 20 to 30 cell rise over baseline CD4 cell counts (which averaged about 330).
In the ddI/delavirdine study, all of the 870 participants were AZT-experienced, with a quarter having had up to four months of ddI. Their baseline CD4 count averaged only 135. Those on combination therapy had CD4 count rises of only five to twenty more than those on ddI monotherapy and a "two-fold" greater decrease in viral load. Neither difference is statistically significant.
Pharmacia & Upjohn has filed a new drug application with the FDA for delavirdine. A meeting of the antiviral advisory committee has been scheduled in October to review the application.
Results from several ongoing European studies of hydroxyurea presented at Vancouver confirm this drug's capacity to induce sustained viral inhibition when used in combination with the nucleoside analog ddI. The most notable of these studies, however, did not find any significant increase of CD4 cells associated with these substantial viral load drops. Although some alopecia (hair loss) occurred (mostly in patients with very low CD4 counts), no major adverse effects were reported.
Franco Lori, M.D., Ph.D., director of the Research Institute for Genetic and Human Therapy (RIGHT) in Pavia, Italy, presented the data of the first randomized, controlled clinical trial of ddI versus ddI and hydroxyurea. In this study, 60 HIV-positive patients with CD4 counts greater than 250 were randomized to either ddI alone (twenty patients), or ddI plus hydroxyurea (40 patients). The doses used were 200 mg of ddI twice a day and 500 mg of hydroxyurea, also twice a day. The objectives of the study were to assess drug toxicity, as well as its effect on plasma viral load and CD4 count (abstract Th.B.942).
According to Dr. Lori, both treatment groups registered a similar initial drop in the plasma viral load -- approximately one log (90 percent) during the first two weeks of treatment. Starting at week four, the ddI monotherapy group experienced a rebound in viral load, whereas in the combination group, viral load continued to decrease at a very slow rate (1.13 log and falling). Of the eighteen patients in the combination arm who had reached week 24 at the time of the report, only one showed a rebound in virus levels. On the other hand, the average viral load at week 24 in the ddI monotherapy arm was 52 percent below baseline, and 25 percent of the patients had returned to baseline values.
There was no significant CD4 cell increase to accompany the reduction in viral load observed in the study. Dr. Lori contended that this is probably due to the "cytostatic" nature of hydroxyurea. The compound inhibits cellular DNA synthesis and suppresses the growth and multiplication of cells. Two of the other four studies presented at Vancouver, Dr. Jorge Vilas' study in Lyon, France (abstract Th.B.291), and Dr. Bonaventura Clotet's study in Barcelona, Spain (abstract Mo.B.1143), did report significant increases in CD4 counts. In Dr. Clotet's study of seventeen patients, viral load was reduced by more than 93 percent and CD4 counts rose by 60 cells (from a baseline of 145 cells) at week 24. In Dr. Vila's study of 25 patients, CD4 counts jumped from a baseline mean of 525 to 610 cells (day 180) and were sustained at 601 cells after one year of treatment. Viral load was undetectable in the half of the twenty patients who have completed one year of therapy. Among a subset of six, five had no infectious virus detectable in their lymph nodes.
Study researchers suggested three possible reasons for the sustained viral inhibition. The cytostatic properties of hydroxyurea may be responsible for some of the virus suppression. Another possibility is that since hydroxyurea inhibits a cellular enzyme essential for the creation of the nucleotide units, it creates a shortage of the building blocks needed to form DNA. The defective nucleoside analogs like ddI are then incorporated more readily into the DNA chain that HIV is building as it infects new cells. This could result in a reversal of HIV resistance to ddI and other nucleoside analogs, and extend the duration of those drugs' antiviral effect. Hydroxyurea's sustained effect may also be due to the fact that cellular proteins, as opposed to viral proteins, are less prone to change that overcomes a drug's inhibitory activity.
In an interview with Dr. Lori after his presentation, he suggested that the lack of CD4 increases associated with ddI plus hydroxyurea may actually be good for patients with HIV, since CD4 proliferation results in more viral replication. When questioned about individuals who have lost some of the CD4 repertoire that protected them against specific microbial antigens, he argued that there was no evidence yet that a severely damaged immune system could be restored and that the objective of treatment in these cases presumably is to prevent further immune deterioration. Still, there are strong indications coming out of the protease inhibitors studies that some level of immune reconstitution has been achieved, and at least one study of IL-2 (abstract We.B.290) presented at the Conference showed it is possible to dramatically raise CD4 counts without an associated large increase in viral burden, albeit in patients in early HIV infection.
ACTG 307, a trial of ddI plus hydroxyurea versus ddI alone, is currently recruiting across the U.S. The American Foundation for AIDS Research also is conducting a study of hydroxyurea/ddI. For more information on either trial, readers may call 800/TRIALS-A.
Protease inhibitors in combination with other antiretrovirals offer new possibilities for treating HIV. The unprecedented effects seen in recent clinical trials underscore the promise of modern therapeutics. Little is known, however, about the long term benefits of protease inhibitors or how to optimally use them given complicating issues like toxicity and resistance.
Coincidentally, these issues became apparent as NIH's Community Programs for Clinical Research on AIDS (CPCRA) network was planning its next antiretroviral trials. At the urging of community activists, a CPCRA collaborative seminar took place on June 8-11 to explore new treatment options and revisions in HIV health care standards. Attendees were a unique mix of clinicians, activists, community members, statisticians, epidemiologists and representatives of pharmaceutical companies with the aim of designing scientifically valid research that also would be acceptable to the community.
Eighty participants met in Denver, Colorado for an intensely focused working meeting. Participants stayed in the same hotels and shared meals so as to encourage maximal collaboration. Eight focus groups formed to design trials addressing the four questions described below. Each topic was assigned to two groups. The groups came together at the end of the seminar to discuss how their proposals could be integrated into the CPCRA scientific agenda.
The two teams addressing this issue agreed that a treatment strategy trial was an appropriate design and recommended that a " screening" study precede a larger clinical endpoint study that would compare rates of disease progression and death.
Group One proposed a study in which antiretroviral naive participants with CD4 cell counts greater than 500 and viral burdens greater than 10,000 were randomized to six or twelve weeks of triple therapy. Treatment would then stop if viral burden was undetectable. If viral burden became detectable again, the same course of treatment would be readministered. Based on those results, a clinical endpoint study would be designed comparing intermittent triple therapy to deferred therapy. The intermittent arm could become continuous if intermittent treatment was not effective in controlling viral burden.
Group two proposed two studies, each with two parts. Antiretroviral naive participants with CD4 cell counts greater than 500 and " low" viral burden would be randomized to no treatment, best antiretroviral background treatment (BAT), or BAT plus protease inhibitor. " High" viral burden participants would receive BAT or BAT plus protease inhibitor. Each stratum would be separately analyzed. The primary endpoint for part one of each study would be the time till a change in therapy based on laboratory markers. Part two would compare strategies for survival.
Both teams working on this question recognized that the study would have to involve a sequence of treatments, because over a long-term study, many participants would need or want to change treatments. Both teams also allowed the decision of what constitutes " best" antiretroviral background therapy (BAT) to lie with individual physicians and patients.
Group Three proposed a three-arm clinical endpoint trial randomizing protease inhibitor naive participants with CD4 cell counts less than 300 to BAT plus ritonavir or indinavir, BAT plus saquinavir or nelfinavir (the specific protease inhibitor in each strategy being at patient/clinician discretion), or BAT plus ritonavir plus saquinavir.
Group Four proposed a clinical endpoint trial in which protease inhibitor naive participants with CD4 cell counts less than 100 were randomized to three groups that added ritonavir, indinavir or saquinavir to BAT. In the event of toxicities or treatment failure, participants would be re-randomized among the other drugs.
Both groups felt that it would be preferable to combine the fewest possible drugs to achieve an antiviral effect. Group Five proposed a screening study that measured drug toxicity and viral load changes. Group Six proposed a large clinical endpoint study.
According to Group Five's proposal, participants with CD4 cell counts less than 100 would all receive nelfinavir. They would then be randomized to add either one or two nucleoside analogs, with or without an NNRTI (non-nucleoside reverse transcriptase inhibitor). Nelfinavir was specified because it was assumed resistance to nelfinavir might be lower than to the protease inhibitors already on the market and because there was less risk of drug-drug interactions. In the large strategy trial to follow, the concurrent drugs would be specified, too.
Group Six's study compared one or two background treatments. Participants ready to start protease inhibitor treatment would be randomized to one or two reverse transcriptase inhibitors, the choice of which would be at patient/clinician discretion.
In the closing discussion, it was noted that the designs proposed by these groups might logically be linked to those proposed by groups three and four. Why not randomize participants to a protease inhibitor and background treatment simultaneously?
Groups Seven and Eight proposed treatment strategy trials that followed clinical endpoints. Group Seven elaborated a design that compared making decisions based on viral load changes versus the current practice. A second part of the study would compare three possible responses when changing that combination therapy -- change the protease inhibitor, change the non-protease inhibitor component or change the entire regimen.
Group Eight proposed a study where participants already on protease inhibitors would be randomized to usual care with viral burden monitoring or to plasma genotypic antiretroviral resistance (GART) testing whenever a treatment change was considered.
Making Best Use of Limited Resources
Group Eight's GART study, the Phase 2 study proposed by Group One and Group Five's studies of protease inhibitors, nucleosides, and NNRTIs were considered high priority in the closing discussion. The HIV Therapies Committee of the CPCRA was asked to review these protocols immediately. Important general points came out which should help the Committee establish research priorities for its limited resources:
The report begins: " We must learn how to optimally use available treatment regimens, including when it is most appropriate to start and change treatment; how aggressively to treat; when if ever, treatment can be stopped; and if optimal treatment can reduce the transmission of HIV and the occurrence of opportunistic conditions. There is a particular need for studies that assess important clinical outcomes such as survival, quality of life, and the long-term consequences of treatment (e.g., toxicity to drugs). We must now evaluate various treatment strategies using different combinations and sequences of potent drugs from several classes over longer periods of time"
The group envisioned strategy trials enrolling several thousand people and lasting several years. Since the group could identify no mechanism capable of conducting such trials, its report proposes as a first step the formation of a public/private " Forum for Collaborative HIV Research." The Forum would first survey the existing trials and identify gaps in research and elaborate the necessary protocols for new trials. It would then proceed to pull together the resources to conduct these trials. It might possibly, for example, find pharmaceutical companies willing to aid the CPCRA mount that strategy trials it is proposing.
After the International Conference was over, new data were released from a 1,892-person disease progression study of 3TC. This clinical endpoint trial was terminated prematurely, at the end of July. Known as the CAESAR trial because it took place in Canada, Australia, Europe and South Africa, the trial compared remaining on previous therapy (AZT, AZT/ddI or AZT/ddC) to previous therapy plus 3TC or previous therapy plus both 3TC and the experimental NNRTI loviride. Eighty percent of trial participants had prior exposure to AZT -- for an average of two years -- while 20 percent had no prior treatment. (Those who were treatment-naive were first started on AZT or combination therapy for several weeks and then assigned to one of the trial arms.)
Adding 3TC was found to have reduced the incidence of new opportunistic infections and death by 54 percent after an average of nine months follow-up. No supplemental benefit from loviride was observed. Surrogate marker results paralleled the clinical endpoint data: Baseline values for CD4 count averaged about 130, and viral load was at about 100,000 copies/ml (5 logs). By week two, both 3TC-containing arms had registered CD4 count rises of about 30 and viral load drops of 90 percent (one log). These figures bounced back somewhat, but at week 28, the CD4 counts were still 20 points above baseline and viral load remained 70 percent (0.5 log) below the initial value. For those remaining on their previous therapy, there was essentially no change in CD4 count or viral load during the study.
A number of further analyses of this study are planned, including one that breaks out the results according to whether individuals had received prior treatment or not. Another important analysis is to relate the magnitude of individuals' viral load reduction to the extent to which their disease progression was delayed. It is unfortunate that viral load data was collected only for about 300 participants since an exact delineation of the relationship between viral load response and clinical outcome would allow for smaller and quicker trials than CAESAR. As it is, it will be difficult to enroll and conduct a trial of this nature in the future. In the words of John Bartlett, M.D., who was on CAESAR's Data and Safety Monitoring Board, " You couldn't do this trial again. An AZT monotherapy arm is no longer acceptable, and viral load has become the standard of care. People will want to know their viral loads and possibly drop out of the study based on the results."
The results of the CAESAR study were corroborated by two presentations at the International Conference. Schlomo Staszewski, M.D., from the Goethe University in Frankfurt, Germany, presented clinical efficacy data from a meta-analysis of four independent trials of AZT/3TC combination involving 972 individuals, both treatment-naive and -experienced (abstract Th.B.948). The meta-analysis found a 49 percent reduction in progression of HIV disease (including outbreaks of minor opportunistic infections like thrush or oral hairy leukoplakia) and a 65 percent reduction in progression to a first new AIDS-defining event. In another study (abstract Mo.B.290), reported by Andrew Phillips from the Royal Free Hospital School of Medicine in London, 620 patients who were enrolled in NUCA 3001 (therapy-naive patients randomized to either AZT/3TC, AZT or 3TC) and NUCB 3001 (therapy-experienced patients randomized to either AZT/3TC or AZT/ddC) were analyzed to determine if their viral load effects correlated with clinical disease progression. In both trials, there were significant effects on both viral load and CD4 counts in those receiving AZT/3TC as compared to those who received the control regimens.
As Treatment Issues goes to press, the recently approved nonnucleoside reverse transcriptase inhibitor nevirapine (brand name Viramune) was expected to reach drug store shelves by the middle of August. Its wholesale price will amount to $2,511 per year.
Nevirapine must be used in combination with other potent agents because HIV easily mutates to resist the nevirapine's effect when nevirapine is administered alone. Unfortunately, there still is no data on the effect of nevirapine on indinavir or ritonavir levels in the blood. (Nevirapine induces liver metabolism of some drugs -- early data suggests it can cause a slight reduction in levels of saquinavir, but Boehringer Ingelheim believe interaction with the other protease inhibitors is less likely.)
Even in combination with AZT/ddI, though, nevirapine can have a sustained potent effect on viral load. Follow-up virologic data was presented at Vancouver, on the use of AZT/ddI/nevirapine in 151 treatment-naive patients (Mo.B.294). The three drug combination produced a sustained reduction in viral load, averaging over 99 percent (2 log) at one year. By comparison AZT/ddI effected a sustained 90 percent (one log) reduction. Viral load was undetectable (below 200 copies per ml) in more than 50 percent of those on the three-drug combination, compared to just 20 percent on AZT/ddI. Two-thirds of those undetectable on the three-drug combination were also undetectable as measured by Roche's ultrasensitive assay (which goes down to 20 copies per ml).
On June 27, 1996, the Food and Drug Administration granted Gilead Science's cidofovir (brand name: Vistide) approval for the treatment of CMV retinitis. Gilead quickly began shipping drug at a price of $589 per vial, which is enough for a two-week course of maintenance therapy. The price per vial is high, but Gilead's marketing department calculates that the total for yearly therapy is $16,000. This is less than for competing intravenous CMV treatments, when one includes the expense of drugs used to manage these drugs' toxicity. The cost of administration also is much less for cidofovir than for the older IV drugs because there is no daily administration or any need for a permanent in-dwelling catheter. The cost of cidofovir turns out to be very similar to the current price of oral ganciclovir.
Still, the cost of all of any of these CMV treatments is unbearably high for people with AIDS. Gilead does have a financial assistance program to help patients pay for cidofovir. Call 800/Gilead 5 (800/445-3235) for more information.
At Treatment Issues' press time, FDA approval was imminent for Serono Laboratories' version of human growth hormone (brand name Serostim) as a therapy for AIDS-related wasting. FDA officials slammed the company's application last March at an FDA advisory committee hearing (see Treatment Issues article; Growth Hormone Cut Down at the Pass, March 1996, page 6), but discussion between the agency and Serono have continued ever since. These discussions culminated in a meeting on July 25 at which the FDA essentially agreed to grant "conditional"approval to Serono's product provided proper confirmatory trials are conducted.
Hammering out the details of such trial or trials is now the main stumbling block to issuing an official approval letter. All along, the FDA has criticized the data Serono has provided. The company contends that human growth hormone marks a major breakthrough for people with AIDS wasting by increasing their proportion of lean tissue rather than just increasing their weight by adding fat. FDA officials have argued about the benefits that added lean tissue provide and even whether that addition really occurs. They also say that Serono has failed to establish an optimal dose and length of administration for human growth hormone.
To settle all the doubts, Serono is proposing a year-long trial involving about 750 persons with HIV who have lost at least ten percent of their normal weight and do not have active opportunistic infections. All participants would receive nutritional and fitness counseling and support, with one group getting growth hormone at a dose of 0.1 mg/kg of body weight every day while another group would get this dose every other day. A third group would receive placebo for the first twelve weeks and then be reassigned to the other two groups. Data would be collected for weight, lean body mass, physical function and quality of life.
Community support for growth hormone has wavered in the past because of the agent's enormous cost -- over $1,000 per week. Enthusiasm for Serostim improved when Serono promised to cap the annual price any individual has to pay at $36,000 and establish a financial assistance program for those unable to pay.
Three community representatives were present at the July 22 meeting. In the words of one of them, Jeff Getty, from ACT UP/Golden Gate, " It is amazing that the FDA listened to people with the condition in deciding the way to go rather than just examining the empirical data. Approval never would have happened without the activists."
Potentially serious bleeding under the skin, in the muscles and joints, has been observed in people with hemophilia taking protease inhibitors, according to a Food and Drug Administration alert. So far, the fifteen reported cases have all occurred in Europe.
Eleven cases were noticed by a national drug monitoring system in France. The bleeding usually starts within the first month, and some patients had to be given increased doses of Factor VIII. No cases have been fatal. The FDA recommends that patients with hemophilia should not discontinue treatment or be hesitant to initiate treatment with protease inhibitors but that they should be more closely monitored.