Ten Years of HAART

The year 1996 has brought a sea change in AIDS research and treatment. Three major factors have contributed to this sea change: a new understanding of viral pathogenesis, new and powerful antiretroviral treatment regimens, and new, more powerful tools for managing HIV levels in the blood and elsewhere in the body. Indeed, it was due to the sensitivity of these new viral load assays that researchers were able to determine the kinetics of HIV replication and immune system clearance within the infected human host, and devise new therapeutic approaches to reduce viral replication. The impact of viral load assays on HIV pathogenesis and treatment research can be compared to the impact of the Hubble Space Telescope on cosmology: both allowed researchers to see their subject with unprecedented resolution.

-- Mark Harrington, Viral Load in Vancouver, 1996

Next year will mark a decade since the introduction of highly active antiretroviral therapy (HAART), which ushered in one of the most startling transformations in the history of medicine.

The Advent of HAART

By late 1995, the AIDS epidemic had been going on for 15 years. Despite years of activism and research, just 4 drugs -- all of them nucleoside analogs (AZT, ddI, ddC, and d4T) had been approved to treat HIV. While AZT and its chemical cousins could delay HIV progression, none of them alone or in 2-drug combinations could durably control HIV, let alone reverse its associated immune suppression and threat of opportunistic diseases. The death toll from AIDS in the United States was approaching 50,000 per year and seemed to be rising.

HIV disease management appeared to be advancing at a snail's pace. Most people living with HIV would progress to AIDS and die, their progression only temporarily halted by 1- or 2-drug nucleoside analog therapy. At best, a cocktail of opportunistic infection (OI) prophylaxis drugs such as Bactrim, fluconazole, and azithromycin might be hoped to forestall the development of Pneumocystis carinii pneumonia (PCP), toxoplasmosis, cryptococcal meningitis, and Mycobacterium avium intracellulare (MAC), while a number of other OIs remained horribly undertreated or untreatable.

The first reports suggesting that protease inhibitors (PIs) might be different were thought to be drug-company hype. TAG's coverage of the 1995 Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) conference described a small Abbott study:

Participants were given AZT/ddC/ritonavir. Their CD4s went up by 100 and their plasma RNA went down by 2.5 logs at 20 weeks. Over the subsequent weeks, he claimed, an increasing proportion of participants became viral culture negative -- which is to say, they could not culture infected cells from the blood. "Some became PCR and culture negative, which suggests that the viral reservoir was empty."

At the time, we described this as an "irresponsible, unsubstantiated claim" (TAG Does ICAAC, 1995).

The 8 months after November 1995, however, witnessed approval by the US Food & Drug Administration (FDA) of 5 new anti-HIV drugs -- the nucleoside analog 3TC, 3 new PIs -- saquinavir (Invirase), ritonavir (Norvir), and indinavir (Crixivan) -- and the non-nucleoside reverse transcriptase inhibitor (NNRTI) nevirapine (Viramune). Moreover, that year saw the introduction of new, quantitative viral load assays -- such as Roche's RNA PCR and Chiron's branched-chain DNA (bDNA) test -- which could reliably measure the amount of HIV in the bloodstream.

Many forces contributed to the HAART revolution. AIDS activism certainly played a part in multiple ways -- by demanding faster AIDS research, mobilizing national awareness, and pressuring Congress to increase funds for clinical trials. Basic science on the biology of HIV infection clarified the key role of HIV in depleting CD4 cells. Studies funded by the National Institutes of Health (NIH) demonstrated both the uses and the limitations of available nucleoside analog reverse transcriptase inhibitors (NRTIs). Drug companies invested in a variety of new treatments for both HIV and its associated infections and cancers. Among the most crucial factors was the role of the FDA in permitting expanded access to experimental AIDS therapies beginning with AZT in 1986, which perhaps most critically provided a framework for accelerated approval of AIDS drugs, beginning with ddI in 1991 and codified in federal regulations in 1992. Accelerated approval allowed drugs to be approved based on favorable changes in surrogate markers such as measurements of CD4 cells and later, viral load. Indeed, ddI was approved in 1991 based on a very small rise in CD4 cells (11 cells over baseline in the ddI group versus a continued decline in those on AZT) in an AIDS Clinical Trials Group (ACTG) study, 116B/117, which compared ddI to AZT among AZT-experienced individuals. The clinical benefits of ddI were confirmed in the same study by early 1992. Soon after, ddC was approved in mid-1992 based on similar changes, which did not however confer clinical benefit.

The problem with CD4 cell changes as a surrogate marker was that CD4 cell count was a direct marker of a person's immune status, but only an indirect marker of anti-HIV drug activity (it rose when HIV levels fell). The CD4 cell changes associated with NRTIs were modest and often transient. And the available blood tests for HIV levels in the early 1990s were primitive.

By 1995, however, pivotal papers by groups from the University of Alabama and the Aaron Diamond AIDS Research Center (ADARC) showed that ritonavir -- a member of a new class of drugs, the PIs, which interfered with a different stage of HIV's life cycle than the NRTIs -- could reduce HIV levels in the blood by 2 logs (99%) -- as measured by new quantitative HIV RNA tests such as the polymerase chain reaction (PCR).

While the longer-term benefit of triple-combination antiretroviral therapy (ART) was not yet clear, in the short-term, falling HIV RNA was associated with rising CD4 cells and reductions in incidence of AIDS and death. In Abbott's pivotal ritonavir study, which randomized 1,090 HIV-infected persons with CD4 counts below 100 cells/mm³ to ritonavir or placebo over a background of NRTIs, 119 ritonavir recipients developed AIDS or died versus 205 (37.5%) placebo recipients during a median follow-up of just 29 weeks (Cameron 1998). In March 1996, ritonavir became the first AIDS drug since AZT in 1987 to win full approval based on a reduction in mortality. The problem with the approach used in the Abbott study became clear within a few months, however -- simply adding a PI into a background of what later would be considered a failing NRTI background. Despite the potency of the PI, the emergence of resistance to ritonavir became widespread. Even worse was the fact that HIV resistant to ritonavir was often resistant to indinavir -- Merck's equally potent PI -- as well.

Nonetheless, the short-term clinical benefit and longer-term promise of the PIs appeared clear. The FDA approved Roche's saquinavir in December 1995. It was weak and had low bioavailability (4%), yet when combined with two NRTIs (AZT and ddC, again) was clearly superior to AZT+ddC on their own. Perhaps more important for AIDS treatment in the long run was FDA's approval of lamivudine (Epivir, 3TC) in November 1995. Because of its low toxicity, good pharmacokinetics (eventually being approved for once-daily use) and the fact that 3TC-resistant HIV was less fit and mutated more slowly than wild-type HIV, 3TC eventually became the most commonly used drug in HIV combination therapy.

When FDA Commissioner David Kessler -- an appointee of the first President Bush who stayed in office until the mid-Clinton years -- heard about the favorable results of the Abbott study (due to Abbott's FDA filing for full approval), he asked Merck to submit its PI, indinavir, to the FDA for accelerated approval at the same time. Both drugs were approved in March 1996. Each was a very potent drug with significant drawbacks. Ritonavir, taken at the full dose of 600 mg twice daily, had high gastrointestinal (GI) toxicity, while indinavir had to be taken thrice daily on an empty stomach, requiring people to undergo 3-hour, food-free windows each day, while downing liters of water to prevent the development of kidney stones (nephrolithiasis).

By the time of the 11th International Conference on AIDS, held July 8-11, 1996 in Vancouver, British Columbia, researchers from a variety of groups -- the ADARC, Boehringer Ingelheim, Merck, and Abbott -- showed that various 3-drug antiretroviral combinations -- AZT and 3TC plus ritonavir or nelfinavir (Viracept), AZT and ddI plus nevirapine, or even ritonavir plus saquinavir without any reverse transcriptase inhibitors -- could reduce HIV levels in the bloodstream by over 99% to below limits of detection (ranging from less than 500 to less than 25 HIV copies per milliliter of blood) -- in a substantial majority (70% to 100%) of those treated. The studies, while small and short-term, demonstrated that dramatic reductions in viral load could be seen among individuals with either acute or chronic HIV infection, among antiretroviral-naive people, and even among those who had been exposed to prior therapy. In Vancouver, John Mellors from the Multicenter AIDS Cohort Study (MACS) presented the famous study demonstrating that viral load in chronic infection predicts rate of progression to AIDS (Mellors 1996), and David Ho and Marty Markowitz presented their famous eradication hypothesis, suggesting that if HIV levels in the peripheral blood could be suppressed by combination antiretroviral therapy (ART) for several years, residual amounts of integrated HIV provirus in human cells might eventually be eliminated from the body (Ho 1996, 1998).

At the time, TAG pointed out the unanswered questions:

The duration of long-term viral suppression remains to be determined. ... The most potent, tolerable regimen(s) for long-term suppression need to be determined, and rational sequences ... need to be defined. Studies in patients with higher CD4 levels and relatively low viral load need to compare partial versus apparently complete plasma RNA suppression as an initial strategy, or immediate versus deferred suppression. Simpler, more convenient regimens need to be developed to enhance compliance, eg with the use of time-release drugs which can be taken once daily or less often. Studies to optimize long-term compliance with complicated, inconvenient regimens need to be conducted, covering a range of strategies in a diverse set of populations. Interventions which may contribute to replenishment of holes in the immunologic repertoire ... need to be designed and implemented. When and whether people experiencing a significant CD4+ T cell rise after combination therapy can be removed from opportunistic infection (OI) prophylaxis and maintenance needs to be addressed. Finally, research needs to be conducted to determine the feasibility of treating or eradicating HIV disease among infected persons in developing nations (TAG, Viral Load in Vancouver, 1996).

What is remarkable is that even in the first months of 1996, before anyone really knew how to best use PIs in combination therapy, AIDS deaths began to drop dramatically in the US, as shown in these data from the HIV Outpatient Study:

Common AIDS-related opportunistic infections (OIs) dropped dramatically as well:

As the following data show, the decline in AIDS deaths in the US was dramatic, falling almost 3 fold between 1995 and 1998, and stabilizing thereafter.

Spurred in part by HAART and perhaps even more by the introduction of AZT to prevent mother-to-child transmission (MTCT), perinatal HIV infection rates declined by over 90% and pediatric AIDS diagnoses went from 952 in 1992 to just 59 in 2003.

Consolidation of the HAART Approach

Following the euphoria of Vancouver, with a regrettable but predictable media over-emphasis on the possibility of eradication of HIV with HAART, came the much harder work of developing and disseminating solid, evidence-based guidance for physicians, care providers, people living with HIV, and insurers. This work began in late 1996 when the Department of Health and Human Services (DHHS) set up not one but 2 panels to develop "Principles of HIV Therapy" and "Guidelines for Antiretroviral Therapy." The first panel was set up by NIH's Office of AIDS Research, the second by the Office of HIV/AIDS Programs (OHAP) in the Department. Their work was published in the Morbidity and Mortality Weekly Report (MMWR) in 1998. The DHHS Guidelines Panel became an ongoing body, meeting by conference call monthly and in person annually at the Retrovirus Conference. The Guidelines were updated frequently and sometimes significantly.

At the time, no one knew the duration of ART-induced viral suppression, how complete immune reconstitution might be, or whether there would be long-term side effects, and if so, what they would be.

By the late 1990s, however, reports began emerging, first in a trickle, then in a flood, of new and apparently horrible side effects associated with antiretroviral therapy. These ranged from so-called "Crix-belly" to "buffalo hump" to facial lipoatrophy/visceral fat accumulation/peripheral fat wasting ("lipodystrophy"), to a complex of possibly mitochondrial-DNA-dysfunction-associated effects such as lactic acidosis and hepatosplenomegaly to -- also seen before HAART -- peripheral neuropathy and myopathy. It was initially unclear and, in some cases still, is less than completely understood to what extent these side effects were related to individual drugs versus drug classes. Certainly, however, in combination with the increasing evidence that HAART could sometimes reverse (and not just delay) the progression of HIV-related immune suppression, the pendulum began to swing from "hit early, hit hard" to "hit HIV-1 hard, but only when necessary" (Harrington & Carpenter 2000).

Many clinicians were still taken with the "hit early" approach, but in the absence of controlled clinical trials defining the optimal point at which to start therapy, and the emergence of data from several large cohort studies including the British Columbia cohort, the EuroSIDA cohort, and the CASCADE collaboration, it appeared that the benefits of HAART were clearest when started before CD4 counts dropped below 200 cells/mm³. However, the benefits of beginning at various CD4 strata above that level were less clear. Greater knowledge about toxicity and the importance of life-long adherence also promoted a change in the starting strategy. The DHHS Guidelines Panel changed its recommendations on "When to start" in 2002, and by 2005 they were rather more conservative:

The debate about whether to continue recommending use of 2 NRTIs without a potent PI backbone was more or less ended when the results of ACTG 320 were announced in early 1997. The study showed that among 1,156 HIV-infected persons with fewer than 200 CD4 cells, the combination of AZT with 3TC and indinavir reduced the incidence of AIDS and death by 50% compared with those taking AZT with 3TC alone. The study provided clear evidence that the HAART approach prolonged health and life in this population, and that its benefits were associated with sustained viral suppression and increases in CD4 counts, often to above 200.

Later evidence came from multiple studies showing that among persons whose CD4 counts rose above 200 cells/mm³, it was safe to stop taking prophylaxis and even maintenance therapy for Pneumocystis carinii (now Pneumocystis jiroveci) pneumonia (PCP), toxoplasmosis, cytomegalovirus (CMV) retinitis, fungal infections, and MAC. Thus, there was clinical evidence that HAART could in fact restore holes in the immune repertoire, even if not always completely.

Major additions to the drug armamentarium came in 1998 with efavirenz (Sustiva), in 2000 with lopinavir/ritonavir (Kaletra), in 2001 with tenofovir DF (Viread), and -- arguably -- in 2003 with enfuvirtide (Fuzeon, T-20), the first fusion inhibitor, and with atazanavir (Reyataz). Efavirenz catapulted the NNRTI class to first-line therapy along with the PIs; its sponsor, DuPont, boldly compared efavirenz to the gold-standard PI, indinavir, in its pivotal study DuPont 006, which demonstrated that efavirenz was as potent as indinavir while being less toxic and, at once rather than 3 times daily, easier to take as well. Lopinavir/ritonavir also emerged as a new first-line drug after proving to have greater potency, better pharmacokinetics, and a higher barrier to resistance than other PIs. Tenofovir had the benefits of a long half-life and apparently reduced toxicity compared with other NRTIs. T-20 was a breakthrough drug from a scientific perspective, but its clinical use was limited by being a twice-daily injectable, with painful injection site reactions. And although BMS's atazanavir reduced cholesterol levels compared with other PIs, and could be taken once daily, its potency was inferior unless combined with ritonavir, which mitigated its cholesterol benefits for some, while increasing its potency overall.

Regimen simplification came with the introduction of fixed-dose combinations including Combivir (AZT+3TC, 1997), Trizivir (AZT+3TC+ABC, 2000), Epzicom (ABC+3TC, 2003), and Truvada (FTC+TDF, 2003); however, these advances were blunted to some extent by their being -- in Joep Lange's words -- "incestuous combinations" developed because they had the same manufacturer, rather than because they made intrinsic therapeutic sense.

By 2005, results from several long-term studies indicated clear favorites among the NNRTI and PI classes for initial therapy, along with better NRTI backbone regimens. With this amount of choice in first-line regimens, it is perhaps not a surprise that there is less clarity with respect to second-line and salvage therapy regimens. Most treatment switching currently occurs because of side effects and involves changing one drug within a regimen or simplifying regimens. For instance, changing to a once-daily regimen to simplify the regimen, rather than for virologic failure or the emergence of drug resistance.

The use of resistance testing is another novelty which interceded over the past 5 years; however its benefits in clinical practice as compared with the intelligent use of treatment history guided by expert opinion have yet to be clearly demonstrated in clinical trials.

Feasibility of HIV Eradication

Starting in 1997, a number of research teams from Baltimore, San Diego, and elsewhere demonstrated that despite HAART's potent effects in limiting HIV replication, slowing the emergence of drug resistance and supporting immune reconstitution, the impact on integrated proviral DNA that existed in more than 100,000 resting but infected CD4 cells was negligible, and would not lead to viral eradication within a normal human lifespan. Drastic therapeutic techniques such as whole body T cell ablation "therapy" followed by immune cell transplants (carried out in an understandably small number of patients) failed to yield anything to justify their toxicity. The possibility of HIV eradication awaits a breakthrough in our understanding of HIV pathogenesis and immune therapy.

Future Therapy for HIV

The best new drugs to emerge since 1996 are all members of the first 3 established drug classes, the NRTIs, NNRTIs, and PIs. However the best new drugs to emerge in the coming decade are likely to be members of new drug classes such as the several entry inhibitors (EIs) -- CCR5 (R5), CXCR4 (X4), gp120, and fusion inhibitors -- and possibly even newer approaches such as integrase, budding, and maturation inhibitors. The coreceptor blockers -- R5 and X4 antagonists -- must surmount fears of their effects on HIV tropism and demonstrate lack of significant toxicity. In addition, the introduction of either an R5 or an X4 antagonist might require the addition of an expensive new diagnostic test -- the viral tropism assay that currently costs over $1,000 -- to the standard of care. As the experience of the past decade shows, adding a new test -- such as viral load, resistance, or even therapeutic drug monitoring (TDM) -- is expensive, time-consuming, and beset by difficulty, particularly if the additional clinical benefit attributable to the test is difficult to demonstrate as in the case of resistance testing or controversial as with TDM, widely used in Europe but not in the United States.

Ultimately, lifelong combination chemotherapy for HIV is far from ideal. As yet unproved are approaches to strengthen the immune system by using cytokines such as interleukin-2 (IL-2) or interleukin-7, therapeutic vaccines (which are beset by the same problems afflicting preventive HIV vaccine research), or other therapeutic strategies such as intermittent therapy (which might reduce total drug exposure while preserving drug benefit).

HIV research is well funded and the new antiretroviral pipeline is fairly robust -- TAG's antiretroviral drug pipeline chart, available online here, shows up to 20 drugs currently in clinical trials that may make it to FDA review within the next few years, with many more in pre-clinical stages of development. However, as the long saga of AIDS research to this date indicates, the most promising approaches will come from new insights derived from basic science. Unfortunately, the next few years do not look as promising as the last decade for AIDS research. Funding at the NIH, the engine of global funding for HIV research, will be climbing by just 2% next year, compared with the more than 100% increases registered since 1992. More AIDS funds will go to vaccine research, which is an urgent priority, but less likely to result in immediate breakthroughs. Industry investment, while still healthy, depends on a robust basic science base whose future is no longer guaranteed. Finally, the US and global healthcare systems are in a state of flux, far from ideal, and deeply unjust for many here and most internationally. And the task of providing treatment to those who need it worldwide, and keeping them on therapy for decades to come, has only just begun.

Mark Harrington is a founder and Executive Director of Treatment Action Group (TAG) in New York (aidsinfonyc.org/tag). He was awarded a MacArthur Fellowship in 1997.

References

Cameron DW, Heath-Chiozzi M, Danner S, et al. Randomised placebo-controlled trial of ritonavir in advanced HIV-1 disease. The Advanced HIV Disease Ritonavir Group. Lancet. 1998;351(9102):536-7.

Centers for Disease Control and Prevention (CDC). HIV/AIDS Surveillance Report, Vol. 15, cdc.gov/hiv/stats/2003SurveillanceReport.htm.

Harrington M, Marco M, Cox S, Horn T. TAG Does ICAAC: AIDS research highlights from the 35th Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), San Francisco, California, September 17-20, 1995. Treatment Action Group, New York, 1995.

Harrington M. Viral Load in Vancouver: A Report from the 11th International Conference on AIDS, Vancouver, British Columbia, July 8-11, 1996. Treatment Action Group, New York, 1996.

Harrington M, Carpenter CCJ. Hit HIV-1 hard, but only when necessary. Lancet. 2000;355:2147-52.

Ho DD. Viral counts count in HIV infection. Science. 1996;272:1124-5.

Ho DD. Toward HIV eradication or remission: the tasks ahead. Science. 1998;280(5371):1866-7.

Mellors JW, Rinaldo CR, Gupta P, White RM, Todd JA, Kingsley LA. Prognosis in HIV-1 infection predicted by the quantity of virus in plasma. Science. 1996;272:1167-70.

Palella FJ, Delaney KM, Moorman AC, et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. N Engl J Med. 1998;338(13):853-860.