Even skeptics can find something to enjoy in the surprising results of AIDS Clinical Trials Group (ACTG) protocol #175. Last month's TAGline ("Combo Jumbo") predicted that the study's principal investigators would "be more or less forced
to recommend combination therapy--regardless of the actual outcome of the trial." Imagine our surprise, then, when the ACTG 175 principal investigators, Scott Hammer and David Katzenstein, presented a fair and balanced picture of the study's results at the 35th Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) in San Francisco on 18 September 1995. And in a strange reversal of fortune, it was the British principal investigators of the Delta study, released on September 26 in Copenhagen, who over-hyped their study (meanwhile trashing ACTG 175), while the Americans displayed a newfound sense of modesty and balance.
ACTG 175 randomized 2,467 HIV-infected participants (with starting CD4+ T cell counts between 200 and 500) to receive either AZT alone, ddI alone, combination AZT+ddI or combination AZT+ddC. They were then followed for three years to watch for the development of 1) a 50% decline in CD4 T-cell count; 2) AIDS; or 3) death. Those whose CD4 cell counts dropped by more than 50% were re-randomized to combination therapy. Participants were stratified by prior AZT experience: 1,067 (40%) were AZT-naïve; 1,400 (60%) were AZT-experienced.
The overall results of ACTG 175 showed that any
of the alternative regimens was superior to AZT monotherapy in preventing a decline in CD4 to below 50% of baseline. However, when only clinical endpoints (AIDS or death) were measured, only the ddI-containing regimens were superior to AZT alone. Being switched to ddI monotherapy or to AZT+ddI combination therapy reduced the risk of death from 9% (AZT alone) to 5%--a 50% reduction in relative risk. This was a dramatic benefit which had not been anticipated.
The surrogate marker results were not as clear-cut. CD4 counts were above baseline at 60 weeks in all arms (rising the most in the AZT+ddC combination arm) except in the AZT monotherapy arm. Viral load did not appear to drop in the AZT arm either, but dropped significantly in the combination arms--and less significantly in the ddI monotherapy arm. So there is (once again) a dissociation between changes in CD4 counts and clinical outcomes (otherwise, people in the AZT+ddC arm--which showed the greatest CD4 cell increases--would have done better clinically) and a dissociation between magnitude of viral load reductions (which was greater on combination therapy) and clinical outcomes: ddI monotherapy didn't reduce viral load as much as combination therapy, but clinical outcome was the same.
The study investigators presented additional analyses based on the two (AZT experienced, AZT naïve) strata. In general, ddI alone and the combinations appeared better than AZT alone in both strata. AZT+ddC appeared to be the best in first-line, AZT-naïve patients, but the number of clinical events was so small that this strata lacked the statistical power to be convincing. The results of ACTG 175 were mainly driven by the larger number of participants--and endpoints--in the AZT experienced stratum.
ACTG 175 also incorporated a second comparison, between "immediate" and "delayed" combination therapy, looking to see which strategy was more effective at delaying clinical progression. Although the study's designers hoped that immediate combination therapy would prove superior to delayed combination therapy, ACTG 175 showed no difference
in clinical outcomes between those who started combination therapy immediately and those who switched to combo only after a 50% CD4 cell decline. Despite pressure to announce the superiority of combination, Drs. Hammer and Katzenstein, supported by statistician Michael Hughes, firmly noted that combination was not proved, per se
, superior to monotherapy in ACTG 175.
The Australian-European Delta study, by contrast, enrolled 3,214 mainly AZT-naïve HIV-infected persons with CD4 counts from 50-350. They were randomized to receive AZT alone or combination therapy with either AZT+ddI or AZT+ddC and followed for two years. In the Delta results, both combinations were found superior to AZT monotherapy, and these combination regimens reduced the relative risk of AIDS or death by about 25%. In the AZT-naïve group (60% of the total), 10% died on AZT+ddI and 12% died on AZT+ddC, while 17% died on AZT monotherapy. Overall, 18% of AZT+ddI takers progressed or died, 23% of AZT+ddC takers progressed or died, and 28% of AZT monotherapy takers progressed or died. The Delta investigators hyped their results as "the most important clinical trial results ever" and indulged in a needless round of ACTG 175 bashing.
The similarities between the two studies are deceptive, and the take-home message is not as simple as "combination is better than monotherapy." ACTG 175 enrolled a much healthier population, with higher CD4 counts (median= 350) yet more AZT experience, reflecting American practice. The Delta investigators enrolled a more symptomatic population (mean CD4= 200) with less extensive AZT experience, reflecting European practice. Follow-up was longer in ACTG 175, but drop-out rates and loss-to-follow-up were both higher. In addition, most ACTG 175 participants reached a CD4 endpoint
rather than a clinical one, whereas the Delta study monitored only clinical endpoints. Most importantly, the Delta study lacked the ddI monotherapy arm, whose results provided the most surprising part of ACTG 175's results. So the Delta's investigators touted the news that "combo is better" while the ACTG 175 team maintained that therapy with ddI alone, AZT+ddI and
possibly AZT+ddC are all better than long-term AZT monotherapy.
Neither study, it is important to observe, answers the question of when
is the best time to start antiretroviral therapy. Both studies suggest, however, that AZT monotherapy should not necessarily be the first-line therapy of choice, as it is now. Thus these studies have the potential to change the standard of care.
As noted, ACTG 175 participants started with around 350 CD4 cells and did just as well starting on ddI as on combination therapy. Delta participants started with the more traditionally European 200 CD4 cell level and did better on combination therapy, but ddI monotherapy was not an option in the Delta study. Future trials will have to untangle the still elusive "benefit" of combination therapy--which is more expensive, inconvenient and toxic than monotherapy, but which offers a greater magnitude of viral load reduction, though this greater magnitude has yet to translate convincingly into a greater clinical benefit.
ACTG 175 and Delta are a strange kind of vindication for those of use who, for years, have been arguing that large-scale, post-marketing clinical endpoint studies are both necessary and feasible. Nonetheless, the AIDS Clinical Trials Group (ACTG) was sorely taxed by 175, and future such studies will have to be done across network lines, using the CPCRA and perhaps even individual physicians' offices to follow patients and gather data.
The ACTG is no longer capable of supporting such large trials by itself, and its method of conducting research is too expensive for such phase IV studies. In the future, the ACTG should enroll a subset of participants for intensive laboratory monitoring while community-based research units and clinics follow patients for clinical outcomes. Thus it is a pity that the National Institute of Allergy and Infectious Disease (NIAID) is rumored to be considering pulling the plug on the Community Program for Clinical Research on AIDS (CPCRA) just when the need for large-scale, low-tech community-based trials is becoming ever clearer.