Thirteen antiretroviral agents are now approved for use in the U.S. Another, the protease inhibitor amprenavir, will probably receive approval in April, and as many as a dozen other agents will likely come before the FDA in the first half of the next decade.
With only half of HIV-positive individuals in the U.S. now receiving treatment, choosing the best initial therapy will continue to be a major issue for the foreseeable future. Although there is a dearth of solid clinical trial data on secondary or salvage regimens, there is now a body of knowledge useful for deciding on the proper first-line regimen.
Protease Inhibitor-Containing Regimens
The morbidity and mortality associated with HIV disease were dramatically decreased as protease inhibitors came on the scene. Although most clinicians still consider protease inhibitors optimal for first-line combination therapies, there are legitimate reasons to question whether a regimen that includes a protease inhibitor is in fact the best choice.
Arguments that strongly favor starting with a protease inhibitor in an initial regimen include:
- The follow-up results from the Merck 035 study include the best durability data to date. In this study, 30 patients started receiving AZT/3TC/indinavir three years ago. Twenty out of the 30 still have a viral load below 50 copies/ml, the limit of quantification on ultrasensitive viral load assays.1
- The vast majority of the currently completed studies showing combination therapy's capability to restore lost immune function and/or prevent loss of immune function, have included a protease inhibitor.2,3,4,5 An article worth noting that recently appeared in the journal AIDS compared suppression of persistent HIV replication in lymphoid tissue between protease inhibitor-containing regimens and nucleoside analog-only regimens. The authors' conclusions were that protease inhibitor-containing regimens "had a greater effect on the lymph nodes than those containing nucleoside analogues alone despite comparable levels of suppression of plasma HIV RNA. Furthermore, protease inhibitor-containing regimens appeared to more effectively reverse HIV-induced immunopathological changes in lymphoid tissue."6
- Preliminary data suggest a lack of viral fitness in protease inhibitor-resistant HIV. In other words, if you include a protease inhibitor in your initial regimen and HIV levels rebound after achieving an undetectable viral load, the virus that now predominates may be less capable of causing immune destruction. This would help explain why individuals who have seen viral rebound while on protease inhibitor therapy have not always seen an accompanying decline in CD4 cells.7
- A series of recent studies suggest that protease inhibitor resistance may not be an issue for individuals who experience viral rebound while on protease inhibitor-containing regimens. The predominant virus that rebounds is often resistant to the other drugs in the combination but not the protease inhibitor.8,9 This raises the possibility that protease inhibitors can be rescued in the case of treatment failure by switching the other drugs in the combination or by adding an extra agent as an attempt at therapy "intensification."
- Yet another report presented at the 6th Conference on Retroviruses showed the ability of a protease inhibitor-containing regimen to change viral phenotype in at least 70% of children infected with syncytium-inducing variants to non-syncytium-inducing variants.10
Arguments against including a protease inhibitor in first-line antiviral combinations include:
- One might wish to save the potent protease inhibitor class of drugs for future antiviral attack when other drugs have failed and the immune system is severely compromised.
- High pill burden, frequency of dosing and dietary restrictions make protease inhibitor-containing regimens challenging even for the most diligent. A recent study found that there was a highly significant association between virologic suppression and adherence three months after the start of therapy.11 Eighty-one percent of those with greater than 95% adherence to dosing schedules had complete viral suppression (defined as viral loads below 400 copies/ml). This success rate rapidly trailed off with decreased adherence: 64% of the subjects with 90% to 95% adherence had complete viral suppression; 50% of the subjects with 80% to 95% adherence had complete viral suppression; and 25% of the subjects who had 70% to 80% adherence successfully suppressed HIV. Only 6% of the subjects with less than 70% adherence were able to achieve complete viral suppression. While the number of pills or the hours between dosages are not the only reasons individuals are nonadherent, they arguably contribute to lack of adherence.
- At present, body shape and metabolic changes are most commonly reported by individuals taking protease inhibitors. Such changes include facial fat wasting, loss of fat in the arms and legs, prominent veins, increased abdominal girth, upper dorsal fat pad ("buffalo hump"), increased breast size in women, lipomas (fat nodules), glucose intolerance and high blood lipids. Although the reason for these changes are not completely understood, a number of studies have recently shown a strong association between taking protease inhibitors and the development of these abnormalities.12,13,14,15,16,17
Sparing Protease Inhibitors with NNRTIs
As with the protease inhibitors, there are pros and cons to including nonnucleoside reverse transcriptase inhibitors (NNRTIs) in initial regimens.
Arguments for including an NNRTI in initial regimens:
- Easier dosing. Two drugs in this class, nevirapine and efavirenz, offer low pill burden and once- or twice-daily dosing. The entire class, which also includes delavirdine, has no restrictions as to timing of meals in relation to taking the drugs.
- NNRTIs originally were thought by most to have less antiviral potency and durability when compared with the protease inhibitors. Ongoing trials are calling into question those assumptions. The DuPont 006 study now has 48-week data showing that patients who start with high viral loads (over 100,000 copies/ml) can have sustained viral suppression on a protease-sparing regimen of AZT/3TC/efavirenz.18 Using an intent-to-treat analysis that counts trial dropouts as treatment failures, 66% of those with baseline viral loads (44 persons) over 100,000 copies/ml and treated with AZT/3TC/efavirenz had viral loads under 50 copies/ml. The figures for the other two treatment arms, efavirenz/indinavir (46 persons) and AZT/3TC/indinavir (44 persons) were, respectively, 35% and 34%. (The analogous figures for the overall trial, 65%, 48% and 43%, were nearly the same.)
Arguments against including an NNRTI in an initial regimen:
- Side effects. Fewer people have been taking this class of drugs and for less time than protease inhibitors. Less is known about their longer-term side effects. Skin rashes can occur with these drugs, especially at the beginning of therapy. With nevirapine, these rashes can develop into life-threatening Stevens Johnson syndrome, although attempts are being made to prevent nevirapine-related skin reactions through the use of concurrent corticosteroids. Efavirenz has effects on the brain: dreams, headache and lack of concentration. For some, these side effects can be overcome, but for others, the NNRTI must be discontinued. And although lipodystrophy and metabolic changes have been mostly associated with protease inhibitors to date, there have been some reports on the similar effects with other antivirals, including NNRTIs.19 The DuPont 006 trial, too, noted significant blood cholesterol elevations in people taking efavirenz. Total cholesterol increased by 25% in the AZT/3TC/efavirenz arm during the first 24 weeks, while the heart-protective HDL (high-density lipoprotein) cholesterol fraction went up 15%. HDL increases at all have not been noted with protease inhibitors, so the added risk of heart attacks and strokes created by the efavirenz-associated may be somewhat less than with the protease inhibitors.
- Resistance: Worries about use of NNRTI center on the ease with which resistance to this class of drugs can occur. A single mutation, at position 103 on the reverse transcriptase enzyme, can nullify the whole class. Other single mutations effect individual NNRTIs. Reports that these mutations impair viral activities are lacking. Protease inhibitors, in contrast, usually require multiple mutations for high-level resistance, including accessory mutations to compensate for loss of viral fitness. The problem is that reverse transcriptase is a much larger and more flexible enzyme than protease. It is easier to adapt to attack by the NNRTIs, which, because of their mechanism of action, all have to bind to the same spot on the enzyme.
Nucleoside Analog-only Regimens
The concept of using an all nucleoside-analog regimen has been boosted by the recent approval of abacavir (Ziagen). Two studies reported at the 6th Retrovirus Conference, combining abacavir with AZT/3TC found that this combination may be a reasonable alternative in first line treatment.
In the first study, CNA3003, 56% of volunteers on abacavir/AZT/3TC for 48 weeks had viral loads below 50 copies/ml in an intent-to-treat analysis that counted therapy switches as treatment failure.20 The second study, CNA3005, found an equivalence between abacavir/AZT/3TC and indinavir/AZT/3TC. At 16 weeks, 45% of those starting on abacavir/AZT/3TC and 42% of those on indinavir/AZT/3TC had viral loads below 400 copies/ml, according to an intent-to-treat analysis covering 498 trial participants.21 Blood cholesterol levels had risen by 15% in the indinavir while no increase had occurred in the abacavir arm.
The abacavir/AZT/3TC combination is attractive because it consists of just two pills twice each day. An all-nucleoside analog therapy is also attractive because of the therapy options it allows in case of viral rebound. If one were to fail on this combination, he or she would still have two fresh classes of drugs with which to construct a rescue combination. One previous caution about abacavir-based nucleoside analog-only regimens was that their potency seemed limited in patients who start therapy with viral loads over 100,000 copies/ml. This caveat arose because of data from the CNA3003 study, which enrolled only 18 persons with baseline viral loads over 100,000. In the larger CNA3005 trial, the proportion of people with viral loads under 400 was virtually the same at 24 weeks -- about 60% -- regardless of initial viral load or treatment assignment.
The DuPont 006, Glaxo CNA3005 and Atlantic (see box) trials are a good start, but more head-to-head trials in treatment-naïve patients are needed to firmly establish protease inhibitor-sparing as legitimate for first-line therapy. With 14 drugs on board, it will be hard to determine whether all combinations containing the same drug classes give the similar results. Certainly, not all drugs in a given class are equivalent in effectiveness, and such differences will alter the conclusions reached in different trials.
For example, DuPont's 006 trial may well have reached a different conclusion had the exceptionally potent new protease inhibitor ABT-378 -- a one-pill, twice-a-day drug -- replaced indinavir, with its sometimes marginal blood levels and its strict dosing schedule (in this trial, five capsules every eight hours). Conversely, the potential for hydroxyurea to increase the effectiveness of nucleoside analogs may tip the balance toward protease-sparing regimens.
Another enduring question is long-term safety. The causes and ultimate implications of elevated blood lipids and abnormal fat deposits have yet to be sorted out for protease inhibitors. Research into whether these same phenomena occur in protease-sparing regimens, or whether other as-yet unrecognized toxicities occur, has hardly begun.
Until all this information is available, doctors and people with HIV will be relying on spotty data, personal impressions and companies' marketing efforts to make vital decisions on managing the virus. Such decisions will not only influence their short-term success at suppressing HIV but may also close off various options for backup therapy should their virus get out of control.
1. Gulick R et al. Treatment with IDV ZDV and 3TC: Three year follow-up. 6th Conference on Retroviruses and Opportunistic Infections (CROI). Chicago IL. Jan. 31-Feb. 4 1999: abstract 388.
2. Hardy G et al. Kinetics of lymphocyte proliferation IL-2 and IL-4 production in HIV-1 infected individuals initiating HAART. 6th CROI: abstract 322.
3. Plana M et al. Immunologic reconstitution after 12 months of antiretroviral therapy in very early stages of chronic HIV-1 infection. 6th CROI: abstract 323.
4. Schrier R et al. Immune reconstitution of advanced HIV patients (following HAART). 6th CROI: abstract 325.
5. Talal A et al. Immune response to antiretroviral therapy in gut associated lymphoid tissue (GALT). 6th CROI: abstract 328
6. Ruiz L et al. Protease inhibitor-containing regimens compared with nucleoside analogues alone in the suppression of persistent HIV-1 replication in lymphoid tissue. AIDS 1999 13(1):F1-8.
7. Ledergerber B et al. Clinical progression and virological failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study. The Lancet. Mar. 13 1999; 353(9156):863-8.
8. Havlir D et al. Viral rebound in the presence of indinavir without protease inhibitor resistance. 6th CROI: abstract LB12.
9. Descamp D et al. Virologic failure resistance and plasma drug measurements in induction maintenance therapy trial (ANRS 072 Trilège). 6th CROI: abstract 493.
10. Essajee S et al. The effect of HAART on syncytium-inducing variants of HIV. 6th CROI: abstract 396.
11. Paterson D et al. How much adherence is enough? A prospective study of adherence to protease inhibitor therapy using MEMSCaps. 6th CROI: abstract 92.
12. Gharakhanian S et al. Frequency of lipodystrophy and factors associated with glucose/lipid abnormalities in a cohort of 650 patients treated with protease inhibitors. 6th CROI abstract 642.
13. Tsiodras S et al. Effects of protease inhibitor use on hyperglycemia and hyperlipidemia: a five year analysis. 6th CROI: abstract 643.
14. Lee D and Matthews WC. Prevalence and risk factors for hyperglycemia dyslipidemia and coronary disease among HIV infected patients on initial protease inhibitor therapy. 6th CROI: abstract 644.
15. Behrens G et al. Lipid evaluation and glucose metabolism in HIV-1 positive patients treated with protease inhibitors. 6th CROI: abstract 647.
16. Thompson M et al. The prevalence of hypertriglyceridemia in HIV infected individuals who have been exposed to various antiretroviral drug classes since January 1996. 6th CROI: abstract 649.
17. Justman J et al. Association of diabetes and protease inhibitor use in a large natural history cohort of HIV-positive women. 6th CROI: abstract 661.
18. Manion D et al. Efficacy of efavirenz-containing regimen in patients with baseline plasma HIV-1 RNA viral loads exceeding 100,000 copies/ml. 6th CROI: abstract 383.
19. Madge S et al. Lipodystrophy syndrome in patients on reverse transcriptase inhibitors. 6th CROI: abstract 654.
20. Fischl M. et al. Ziagen combined with 3TC & ZDV is highly effective and durable through 48 weeks in HIV-1 infected antiretroviral-therapy-naïve subjects. 6th CROI: abstract 19
21. Staszewski S et al. Ziagen/Combivir is equivalent to indinavir/Combivir in antiretroviral therapy (ART) naive adults at 24 weeks (CNA3005). 6th CROI: abstract 20.
The Atlantic Study
The Atlantic Study is a major ongoing multicenter European and North American trial directly comparing protease-containing and protease-sparing regimens. Preliminary 24-week data were presented at the 6th Conference on Retroviruses and Opportunistic Infections. Study participants receive d4T/ddI/3TC, d4T/ddI/nevirapine or d4T/ddI/indinavir in open-label fashion.
|Atlantic Study: Intent-to-Treat Analysis at 24 weeks|
|Number of currently evaluable participants
|Baseline median CD4 count
|Baseline median viral load
|Percent with viral load below 50 copies/ ml
|Median increase in CD4 count
At 24 weeks the nevirapine and indinavir arm look similar in their ability to raise CD4 cell counts and reduce viral load while the 3TC-containing arm appears inferior (C. Katlama et al, slide presentation 18). This difference does not reach statistical significance, however. In a related study of 12 individuals enrolled in Atlantic, researchers looked at whether initial HIV clearance rates differed by on regimen. The conclusion was that it did not (R. van Heeswijk et al., poster 634).
The criticism most often leveled at this study is that baseline viral loads were only about 15,000 in all three arms. This low level makes it difficult to compare the different treatments' potency.
Atlantic has a targeted enrollment of 300, of which 287 had entered the study by last fall. The ddI is taken once daily between meals while the d4T is taken in its usual twice-a-day schedule, as is the 3TC. The indinavir is administered three times a day, and the nevirapine is given at a 400 mg once daily dose.
So far, nevirapine-induced rashes resulted in nine switches to other arms of the study, two individuals reported severe peripheral neuropathy (probably from the d4T and ddI) and 12 individuals in the nevirapine arm reported grade 3 or 4 liver function abnormalities.
New Nonnucleoside Reverse Transcriptase Inhibitors
Five NNRTIs still under development were presented at the 6th Conference on Retroviruses and Opportunistic Infections. Currently the NNRTIs have cross-resistance to each other. A single mutation (K103N -- a lysine to asparagine at amino acid 103 in HIV's reverse transcriptase enzyme) can occur rapidly and negates the activity of all currently approved NNRTIs. Drug companies are now attempting to develop NNRTIs that either will not trigger this mutation or, even better, will still be potent against HIV with K103N. Such activity would allow the new agent to replace one of the present NNRTIs should treatment failure occur.
DuPont, maker of the NNRTI Sustiva (efavirenz), is hoping to bring a "second generation" NNRTI to market within a few years. The company has two compounds in development, DPC 961 and DPC 963 (S. Erickson-Viitanen et al., slide presentation 13). Both show strong activity in the test tube against wild type HIV. Their activity against the K103N mutant virus is reduced but is greater than that of the established NNRTIs. Initial pharmacokinetic data show a long half-life for both drugs in the body. This will most likely allow once-a-day dosing. DuPont is planning to select the better of these drugs for fast-track development. It hopes to start testing the chosen agent in humans by the end of the year.
Glaxo Wellcome presented GW420867X, one of a family of NNRTIs purchased from Hoescht-Bayer that includes the previously reported HBY097. The GW compound exhibits better antiviral activity than HBY097 and is less affected by binding to blood proteins (JP Kleim et al., poster 599). Test-tube data on the resistance profile of GW420867X found results comparable to efavirenz (JP Kleim et al., poster 600). Based on its pharmacokinetic profile, GW420867X, like the DuPont compounds, should permit once-a-day dosing. Glaxo has done some pharmacokinetic (drug level) studies in healthy HIV-negative adults, and clinical studies in HIV-infected patients are to commence soon.
Agouron, which produces the protease inhibitor Viracept (nelfinavir), also has a new NNRTI, AG1549 (K Potts et al., slide presentation 12). AG1549 remained active in cell cultures against laboratory-generated strains of virus resistant to present NNRTIs. Significant reductions in virus susceptibility to AG1549 did occur when two or more reverse transcriptase mutations arose. Phase I/II clinical trials are underway to determine optimal dosing.
Sarawak Medichem pharmaceuticals has a singular set of related NNRTIs, calanolide A and B (respectively, J Ruckle et al., poster 606 and T Jenta et al., poster 602), derived from the latex of a tree indigenous to Sarawak, Malaysia. In tests that were once again highly preliminary, the calanolides exhibited an attractive resistance mutation profile. Their activity in the test tube against nevirapine-resistant HIV with the Y181C, normally a resistance-conferring mutation, was actually better than against wild type HIV, although the K103N mutation was detrimental. In early dosing trials of calanolide A, the most common reported adverse events were dizziness, headache and an oily taste. Since the calanolides readily combine with fats (are lipophilic), they may penetrate tissue membranes more readily and reach areas difficult to get drugs into, such as the brain.
A word of caution concerning all of these compounds: The data from cell culture experiments look promising, but test-tube data on resistance has often differed in the past from what was observed once an agent is tested in the human body.
Back to the GMHC Treatment Issues February 1999 contents page.