Hold the Protease, Please
Ever since the 12th World AIDS Conference in Geneva, where researchers presented data from a study comparing the non-nucleoside reverse transcriptase inhibitor (NNRTI) Sustiva® (efavirenz) to the protease inhibitor (PI) Crixivan® (indinavir sulfate) -- both in combination with two nucleoside analogues -- there has been a great deal of speculation as to whether patients starting highly active antiretroviral therapy (HAART) for the first time necessarily need to take a protease inhibitor. True, PIs have been proven to offer a powerful antiviral punch and have for several years been the gold standard therapy. Yet their side effects -- which may include body composition changes, metabolic complications, and a slew of other problems -- have caused many physicians and patients to question their use for long periods of time.
A protease-sparing HAART combination is just that: a drug "cocktail" consisting of an NNRTI with two nucleoside analogues or, perhaps, a combination consisting of three nucleoside analogues. Now -- as reported at the Sixth Conference on Retroviruses and Opportunistic Infections in Chicago this past winter -- there are some data suggesting that such combinations may, in fact, be viable options.
Perhaps the most convincing data come from a DuPont study (with CRIA as a co-participant) involving the NNRTI Sustiva. Reported at the conference were data involving 450 patients who had been taking one of three combinations for approximately one year (48 weeks). The first combination consisted of Sustiva, AZT (Retrovir®) and 3TC (Epivir®); the second was made up of Sustiva and Crixivan, and the third consisted of Crixivan, AZT, and 3TC. Approximately 98% of the patients who were still taking Sustiva/AZT/3TC after 48 weeks had undetectable viral load (<400 copies/mL), compared to 86% of those who were still taking Crixivan/AZT/3TC. While this doesn't mean we can conclude that Sustiva/AZT/3TC is better than Crixivan/AZT/3TC, we can conclude that the two are equally effective.
Of particular interest are data from an "intent-to-treat" analysis. Such an analysis includes all patients enrolled in the study, even if they weren't taking their designated drugs at the time of the analysis, perhaps due to side effects or poor compliance. Unlike the "on-treatment" analysis reported above, an intent-to-treat analysis is much more accurate in terms of answering how the drug (or combination of drugs) will perform in the real world, as it reflects the total number of patients who started therapy with the intent of being successfully treated. Using this analysis, approximately 71% of patients who were originally randomized to receive Sustiva/AZT/3TC had undetectable viral loads after 48 weeks of therapy. Among those originally given Crixivan/AZT/3TC, however, only 48% had undetectable viral loads after 48 weeks. These data were statistically significant, which lends credence to the suggestion that Sustiva may, in fact, be more effective than Crixivan for patients who are starting HAART for the first time. There was, however, a higher than expected drop-out rate in the Crixivan/AZT/3TC arm of this study, and the fact that drop-outs count as treatment failures in the intent-to-treat analysis accounts for the poorer performance of this regimen compared to previous studies.
Another interesting analysis of this trial suggested that patients who had a viral load of more than 100,000 copies/mL upon entering the study did best if they received Sustiva/AZT/3TC. In terms of side effects, more patients taking Crixivan/AZT/3TC either reported side effects (particularly nausea and vomiting) or discontinued the drugs altogether than patients receiving Sustiva/AZT/3TC or Sustiva/Crixivan. However, more patients (approximately 55%) who received Sustiva/AZT/3TC reported central nervous system (CNS) problems including sleeping troubles, vivid dreams, and muddled thinking.
In a second report at the Chicago conference, preliminary results from the international ATLANTIC study comparing one PI-based combination with two PI-sparing combinations were presented. The specific regimen tested in this open-label randomized trial was the NNRTI Viramune® (nevirapine) in combination with d4T (Zerit®) and ddI (Videx®) (Group 1), Crixivan with d4T and ddI (Group 2), and a triple-nucleoside combination of 3TC, d4T, and ddI (Group 3). Data were presented with respect to the first 234 patients who completed six months (24 weeks) of therapy.
After 24 weeks of therapy, there appear to be some differences between the three groups. Yet, because the study is still in its early stage (the researchers hope to collect three years'-worth of follow-up data), it's still too early to tell if these differences are statistically significant. In the intent-to-treat analysis, approximately 78% of patients in Group 2 had undetectable viral loads (<500 copies/mL), compared with 69% in Group 1 and 71% in Group 3. As for the on-treatment analysis, 91% in Group 2 had undetectable viral loads, compared to 89% in Group 1 and 80% in Group 3.
Even though these results seem to suggest that a Viramune-containing regimen is equally as effective as one that contains Crixivan -- at least for 24 weeks -- it is important to remain cautious. For starters, patients appeared to enter this study with low viral loads. The average viral load before starting the treatment was approximately 15,000 copies/mL, which can often be pushed to undetectable levels for at least 24 weeks using two nucleoside analogues alone. Also, it will be important to see how well the combinations hold up for a prolonged period of time. It's important to keep in mind that not only do we want to keep viral load as low as possible, but also to keep it low for as long as possible.
Making the Switch
For patients who have already started a protease inhibitor-based combination and are enjoying good viral load and T-cell responses, but are experiencing lipodystrophy, a handful of studies suggest that switching the PI for an NNRTI may be a feasible option. One small study presented at the conference involved 21 HIV-infected patients, all of whom had received at least nine months of Crixivan, d4T, and 3TC, and had both undetectable viral loads and signs of lipodystrophy. The study either maintained them on this regimen or switched the Crixivan for Viramune. The study volunteers were monitored every four weeks using viral load tests, T-cell counts, and a slew of other tests to measure metabolic and body composition changes.
After 12 weeks, all patients -- which includes 10 patients in the Crixivan group and 11 patients in the Viramune group -- still have undetectable viral loads (<50 copies/mL). Among those who were switched to Viramune, cholesterol and triglyceride levels have significantly decreased. This is potentially an important outcome, as high cholesterol and triglyceride levels are associated with a heightened risk of experiencing heart disease or stroke over the long term. While most patients who were switched to Viramune also reported that their body composition has begun to normalize, none of the actual body composition tests have yet been able to detect a difference between the two treatment groups.
In another study, 12 patients who opted to drop their protease inhibitor for Sustiva did not appear to fare as well as those enrolled in the Viramune study. Although nobody who switched to Sustiva experienced an increase in their viral loads after dropping the protease inhibitor, most patients did see a rise in their cholesterol and triglyceride levels after 12 weeks. After 24 weeks of switching, however, the levels began to decrease once again.
These preliminary results -- at least those involving Viramune -- demonstrate a potential treatment alternative for some patients who are seeing good viral load response while on a PI-containing regimen but are also encountering signs and symptoms of lipodystrophy. Although the follow-up time is extremely short, the initial results in terms of declining triglyceride and cholesterol levels as well as improved patient-reported body image are very promising. Continued follow-up of patients in this trial will confirm whether switching from a protease inhibitor-containing regimen to an NNRTI-containing regimen will be a feasible management strategy for lipodystrophy associated with HAART.
Regimens that use NNRTIs as a foundation are not without their potential concerns. First, the chance of developing resistance to NNRTIs is much higher than with PIs. All three NNRTIs approved to date are associated with relatively high levels of resistance, even after just one mutation in the reverse transcriptase enzyme, as opposed to the PIs, which require multiple mutations in the protease enzyme to confer resistance. While combining NNRTIs with standard nucleoside analogues does appear to delay the onset of resistance, no one knows for how long. At the same time, combinations consisting of one NNRTI and two nucleoside analogues only prevent healthy cells from becoming infected, whereas a protease/nucleoside analogue combination also prevents cells that have already been infected from producing new virus. In turn, an NNRTI-based combination will not be active against long-lived, resting cells that can begin producing new virus if activated. Finally, as mentioned above, these studies have yet to produce long-term data.
PI- and NNRTI-Sparing Combinations
While the ATLANTIC study is in the midst of pitting its own all-nucleoside analogue combination against the PI- and NNRTI-based combinations, another study has demonstrated that a combination of ZiagenTM (abacavir), AZT, and 3TC is equally as effective as a combination of Crixivan, AZT, and 3TC, at least for 24 weeks. Unlike the ATLANTIC study, the abacavir study enrolled treatment-naive patients with high viral loads; the average viral load upon entering the study was higher than 100,000 copies/mL and the average T-cell count was 360. Using an intent-to-treat analysis, the researchers found that approximately 65% of patients in both groups had undetectable viral load (<400 copies/mL). As for the on-treatment analysis, approximately 85% in both groups had undetectable viral loads.
Should a combination consisting of three nucleoside analogues prove to be a truly effective option -- which will require follow-up data lasting a few years or more -- the number of treatment options available to patients would greatly expand. Given the lack of cross-resistance between many of the various nucleoside analogues, along with the fact that an all-nucleoside analogue combination will have no impact on the future sensitivity of HIV to both NNRTIs and PIs, future options are undoubtedly preserved. Yet, an all-nucleoside analogue combination is not without its problems, and suffers from many of the same potential drawbacks as NNRTI-based combinations.
This article was provided by AIDS Community Research Initiative of America. It is a part of the publication CRIA Update. Visit ACRIA's website to find out more about their activities, publications and services.