On Friday September 18, the FDA granted accelerated approval to DuPont Pharmaceuticals' efavirenz, also known as Sustiva, a potent non-nucleoside reverse transcriptase inhibitor which was all the rage at this year's two big international AIDS conferences. FDA approval for Glaxo Wellcome's abacavir (Ziagen) is also on the horizon. While there seems little question that both drugs are deserving of the federal imprimatur, the difference between them is stark: efavirenz appears to represent a substantial therapeutic advance, while abacavir is the quintessential "me-too" tag-along. Spencer Cox prepared this analysis on his way to ICAAC.
Efavirenz is the third in a class of anti-HIV drugs known as non-nucleoside reverse transcriptase inhibitors (NNRTIs), which include nevirapine (Viramune) and delavirdine (Rescriptor). Because efavirenz seems to be more potent than its NNRTI siblings -- and can be administered once a day with no eating restrictions -- it might offer patients a substantial advantage over existing products.
Activity and Efficacy
In a number of studies conducted in several distinct patient groups, efavirenz has been shown to provide potent, sustained suppression of HIV replication.
In study 006, which enrolled 450 antiretroviral-naive patients, the combination of efavirenz (EFV) with AZT and 3TC was superior to a standard regimen of indinavir with AZT and 3TC in reducing participants' plasma HIV RNA to less than <400 RNA copies. (74.7% for AZT+3TC+EFV vs. 56.2% for AZT+3TC+IDV). In the same study, a combination of efavirenz with indinavir was comparable, although not significantly superior, to the standard regimen in achieving undetectable viral load after 24 weeks. Due to some limitations in the data set, it is probably not safe to assume that AZT+3TC+EFV or IDV+EFV are, in fact, better than AZT+3TC+indinavir, but these combinations are clearly comparable in their antiviral potency over the twenty-four week time period studied.
In study 004, 137 antiretroviral-naive patients were treated with AZT+3TC plus one of three doses of efavirenz, or with AZT+3TC and an efavirenz placebo. After 36 weeks of therapy, 88-96% of patients treated with efavirenz-containing regimens had <400 HIV RNA copies as compared to about 50% of patients taking only AZT+3TC.
In study 005, 63 patients, both naive (n=30) and experienced (n=33) to therapy were treated in an open-label fashion with a combination of efavirenz and nelfinavir. Only prior nucleoside experience was allowed; all patients were NNRTI and protease inhibitor-naive. After 16 weeks of treatment, study participants experienced a mean1.5log drop in plasma HIV RNA levels, and the majority of patients had plasma HIV RNA levels that were below the limit of detection.
SafetySerious adverse events in patients treated with efavirenz are relatively rare, and are unlikely to substantially diminish the clinical benefit predicted by the drug's activity against HIV. The expanded access database provides the largest safety data set available. When the expanded access safety database was queried on June 1, 1998, there were 1,193 patients for whom the first month of case report forms had been submitted to the database. A new query will be submitted prior to approval with more information, presumably including data on a larger percentage of the more than 11,000 patients worldwide who have received efavirenz through expanded access programs.
Serious adverse events occurred in approximately eight percent of participants. Only rash and diarrhea occurred in more than 0.5% of the study population. Due to the lack of an appropriate control arm, it is impossible to determine the causal relationship between these events and the use of efavirenz. 103 (8.6%) of patients discontinued treatment due to adverse events, whether or not those events were related to the use of efavirenz.
Central Nervous System Symptoms
A syndrome of CNS symptoms seems to occur with some frequency in patients treated with efavirenz, although rarely are those symptoms classified as "severe." According to the manufacturer, these CNS symptoms are most commonly described as "mild dizziness and disorientation," and have been reported after doses of 200, 400 and 600 mg. Episodes reoccur upon daily dosing. Intensity, though, is said to decrease over time, and generally seems to pass after about 2-3 weeks. Intensity of these symptoms is dose dependent, and may be minimized with dosing in the evening just before sleep. Only 1.2% of 1,193 patients in the expanded access program treated with more than one month of efavirenz reported serious CNS symptoms, while approximately 2.7% of patients in ongoing clinical trials classified the syndrome as "severe."
RashA mild rash has also been associated with use of efavirenz, however investigators note that both the frequency and severity of the rash is less than those seen with use of other NNRTIs. Most patients are successfully "treated through" the rash, with anti-inflammatories used to treat symptoms if necessary.
TeratogenicityEfavirenz may also be teratogenic: of thirteen pregnant monkeys treated with efavirenz in doses comparable to those used in humans, three had progeny with serious birth defects, including a cleft palate, small eyes. One was born without a brain and missing one eye. No pregnancies were seen in the expanded access program, which required use of barrier contraception. Although this raises troubling concerns about the teratogenicity of efavirenz, it should be noted that a single monkey study does not in and of itself prove that efavirenz is unusually teratogenic. However, the label for efavirenz should carry a black-box warning that studies of the drug in non-human primate models suggest that use of efavirenz during pregnancy carries a high index of suspicion with respect to teratogenicity.
In vitro studies suggest that HIV requires multiple mutations in the reverse transcriptase in order to develop resistance to efavirenz, and the emergence of highly resistant virus only develops after multiple passages in tissue culture. Normal dosing should produce concentrations sufficient to suppress replication of virus with a single mutation; however, the presence of two or more resistance mutations will probably be sufficient to overcome the antiviral potency of efavirenz at normal doses. Virus that is resistant to efavirenz is likely to be cross-resistant to the other currently marketed non-nucleoside reverse transcriptase inhibitors, including delavirdine and nevirapine.
Although efavirenz is metabolized through the hepatic cytochrome p450 pathway, and therefore interacts with a number of other drugs used in the treatment of HIV infection and its sequelae, a number of these interactions have been well-studied, and are, in general, within the range of a number of other marketed therapies. Studies are still needed evaluating interactions between efavirenz and methadone, ritonavir, midazolam (Versed), lorazepam (Ativan), and paroxetine (Paxil). FDA should require the timely completion of these studies as a condition of approval. Studies of efavirenz in combination with abacavir and amprenavir should be strongly recommended.
Although data on the use of efavirenz in pediatric populations were not made available, we anticipate that such data will be available to the FDA, and are likely to meet the accepted standards for approval of pediatric products. DuPont Pharmaceuticals is to be congratulated on its commitment to rapidly evaluate the use of its product in this special population.
Efavirenz penetrates readily into the cerebrospinal fluid, achieving CSF concentrations that are well above the drug's minimum inhibitory concentration. In a study of eight patients treated with efavirenz in combination with indinavir in patients with a mean CSF viral load of 52,963 copies/mL, all patients achieved undetectable CSF viral loads following 17-35 weeks of therapy.
According to representatives of DuPont Pharmaceuticals, the company will file for "full approval" based on 48 week follow-up data from DMP-006.
Other Procedural Issues
FDA's decision to approve efavirenz without an open public presentation of the available evidence submitted to the agency in support of the drug's labeling means that advocates for people with HIV, physicians and patients have limited access to important information about the product that has not come from the sponsor's public relations department. Rapid publication of the agency's interpretation of the data set is vital, but especially so when there is to be no public presentation of the data set. FDA should post the Executive Summaries of its analyses of data sets supporting applications for approval of new drugs on its World Wide Web home page within a week after the notification of a product approval is mailed to the product's sponsor.
Studies of abacavir suggest no distinguishing benefit for abacavir compared to other currently available marketed therapies. The drug is clearly active against HIV in a variety of different patient groups; however, no evidence suggests clinically greater utility than existing nucleoside analogue combinations -- especially in nucleoside pre-treated patients. Approval of abacavir will offer patients another "choice" in antiretroviral therapy, but data so far suggest that the choice offered is not a functionally important one; rather, it is more like the choice between a blue pill and a red one. Furthermore, use of abacavir has been associated with a potentially serious hypersensitivity response which seems to resolve rapidly with permanent treatment discontinuation.
Although Glaxo has conducted a number of studies of abacavir, two key groups of studies stand out: studies that compare the combination of AZT+3TC to the combination of AZT+3TC+abacavir (Guess what? Three Glaxo drugs are better than two!) and studies that evaluate the virologic effects of abacavir in combination with amprenavir, Glaxo's protease-in-development. Additional on-going boutique-style studies are evaluating the quadruple combination of AZT+3TC+abacavir+amprenavir. If Glaxo's goal is to establish simultaneous administration of four Glaxo products as the standard-of-care for HIV infection, then approval of abacavir at this point could well be regarded by the company as the next step toward that aim. But until the quadruple-combination is thoroughly evaluated, patients have little reason to use abacavir following the drug's approval.
All this raises questions about on-going regulatory policies regarding approval of new drugs. While Glaxo has technically met the accelerated approval requirement that its drug demonstrate "preliminary evidence of efficacy," the regulations also require that drugs receiving accelerated approval demonstrate properties that existing therapies do not have. Abacavir has not met this standard. Consequently, advocates for people with HIV should think long and hard about the wisdom of bringing "me too" drugs to the market under reduced standards of safety and efficacy. It is worth asking whether the benefit of having an unremarkable drug like abacavir on the market outweighs the increased risks to patients of marketing a drug whose safety and efficacy have not been well-characterized. Ultimately, this may be a case in which it is best to let the market sort out available therapies. But in doing so, physicians and patients should be aware that existing double-nucleoside combinations may be just as useful as abacavir, and at substantially lower prices.
Late every August, the once-and-future greats (and some not-so greats) of AIDS research pay homage to HIV co-discoverer Robert C. Gallo at the annual laboratory meeting once, in palmier days, hosted in Bethesda by the NCI Laboratory of Tumor Cell Biology, now hosted by the Institute for Human Virology (IHV) and held at the tourist-friendly Hyatt Regency Harborside in glamorous, sweltering Baltimore, Maryland. Mark Harrington attended the first four days of the week-long research gab-fest. Some highlights:
Zacarias stated that AIDS cases may be declining in Brazil. The government there is spending $700,000 to provide antiretroviral treatment, including protease inhibitors, to 50-60,000 people with HIV there. He discussed a gradient of HIV clinical management, dependent on a country's given health care infrastructure, the quality of usage, etc. It would begin with HIV counseling and testing, ancillary care (diarrhea, pain management, aspirin, etc.), scale up to opportunistic infection treatment (especially for TB), sexually transmitted disease treatment, and prevention of vertical transmission, and finally culminate in comprehensive antiretroviral therapy. Obviously different countries have different resources.
Is Subtype HIV-1C More Virulent? Max Essex of Harvard presented an unconvincing paper attempting to demonstrate that HIV-1 subtype C, which is increasingly prevalent in parts of Africa, is more virulent than other subtypes and more easily transmitted heterosexually. He claimed that some C subtypes have additional long terminal repeat (LTR) and NF-kappa-B viral transcription promotor regions than other strains. This would, he claimed, enable subtype C to grow faster than other strains. However, "there are different Cs; all Cs are not alike."
Laurie Garrett, behaving just like television's relentless Murphy Brown, approached the microphone and decimated Essex's paper with a historical/epidemiological explanation for the greater preponderance of subtype C in the new South African pandemic. Patterns of guerrilla travel, social disruption and migrant labor led to HIV transmission along supply routes, leading to a dramatic increase in HIV in southern Africa. The fact that it was subtype C was just a matter of chance.
Accelerating Decay of the Third Compartment. Several teams are wasting no time exploring immuno-stimulants which may accelerate decay of the third compartment (the pool of 100,000 to 1 million resting, latently infected CD4+ T cells which represent the current principal stumbling block to total eradication of HIV in the body). At NIH, Cliff Lane and Tony Fauci took 13 patients on HAART who had maintained a viral load below the limit of quantification and administered relatively high-dose interleukin-2 (IL-2). IL-2 is postulated by some to activate resting T cells, although some (e.g., Siliciano) maintain that the resting cell population does not express IL-2 receptors. In any case, Fauci reported that provirally-infected cells could no longer be isolated from 3 of the 13 IL-2+HAART recipients, even when as many as 300 million cells were extracted. Obviously the "limit of detection" for eradication experiments will be compromised by the impossibility of sampling such large numbers of cells.
Joep Lange of the University of Amsterdam administered an even more daunting regimen of five antiretrovirals plus IL-2 and the mouse monoclonal antibody OKT3, which targets and activates all the body's T lymphocytes through the CD3 receptor. The result is a clinical condition resembling toxic shock, as all the cells release pro-inflammatory cytokines and many die of apoptosis. The hope was that the coadministered HAART would prevent the viruses produced by waking latently infected cells from infecting another generation of cells. While the patients became very sick (requiring admission to the intensive care unit), infectious virus could no longer be cultured from two of the three. One declared himself cured and disappeared from follow-up. However, in both Lange's and Lane's experiments no one has yet to go off HAART. Lane will start enrolling a study at NIH this fall in which anyone who has maintained a CD4 count below 5 copies/mL for over one year will have their antiretroviral therapy terminated. Researchers will then follow the brave volunteers for evidence of viral re-emergence (this approach does not seem to make sense unless the patients have received immunostimulatory therapy). Researchers were reluctant to use the "C-word" in their presentations, but the halls were abuzz with the slightly surreal gossip about very toxic, potentially fatal, possibly curative regimens.
Frequent HIV Reinfection Demonstrated in Chimpanzees. Addressing an issue vexing HIV treatment and prevention, Pat Fultz of the University of Alabama at Birmingham (UAB) presented compelling evidence that infection with a primary HIV strain does not prevent subsequent reinfection by another strain in the chimpanzee model. In some animals, both strains were readily detectable at later time points, while in some cases the earlier strain (or strains, in one animal reinfected twice) predominated. The ability to detect superinfection depends on using strain-specific PCR primers (which is not done with standard commercially-available viral load assays used in humans). This work has chilling implications for the dissemination of drug-resistant HIV strains among indivudals already infected. Superinfection is not necessarily detectable by increasing antibody titers, suggesting that there is not always a "booster effect." Fultz achieved 100% (9/9) superinfections within viral clades, and 9/13 between clades.