The US Food & Drug Administration (FDA), which will soon be under the control of Acting Commissioner Dr. Michael Friedman when Commissioner Dr. David Kessler leaves to run the Medical School at Yale University, is facing a controversy over approval of a promising new anti-HIV drug. On Friday, March 14th the US Food and Drug Administration approved Agouron Pharmaceuticals' Viracept brand nelfinavir, a new HIV protease inhibitor, without a public hearing of its Antiviral Drugs Advisory Committee.
AIDS activists are concerned that, without a hearing, there will be no publicly available objective review of Agouron's data set. In the past, such review has turned up discrepancies between data that were presented to FDA, and data that had been presented in earlier community meetings. For instance, in presentations to AIDS activists, Serano Laboratories which manufactures recombinant Human Growth Hormone for treatment of AIDS-related wasting, claimed a survival advantage. However, as FDA pointed out, no such advantage could be shown based on the company's data. Activists emphasize that the concern is a procedural one, and is not related to the particular drug. Early studies have suggested that nelfinavir is a relatively safe and potent new anti-HIV therapy.
In three early studies, more than 700 patients were treated with nelfinavir alone or in combination with either d4T or AZT+3TC. Subjects included both both antiretroviral-naïve and antiretroviral-experienced patients. Average CD4+ cell counts at baseline ranged from 200 to 300 cells/mm3, and average HIV RNA levels were just below 100,000 copies/mL. Patients taking nelfinavir in combination with AZT+3TC had an average reduction in HIV RNA levels of about 2.5 logs (using an assay with a sensitivity of 100 copies/ml). Sixty-five to 81% of these patients had reductions in viral load to < 500 copies/ml when the Chiron second generation bDNA assay was used. By comparison, patients treated with AZT+3TC experienced an average viral load drop of about 1.5 logs. In a study of patients taking nelfinavir in combination with d4T, the average viral load drop was also about 2.5 logs when the more sensitive assay was used and about 1.5 logs when a cut-off of 1,200 copies was used. 76% of patients in the d4T+NFV group had plasma viral load reductions to levels < 500 copies/ml. By contrast, patients receiving only d4T monotherapy experienced about a .5 log reduction in viral load on both assays.
In general, nelfinavir has an acceptable safety profile: the main side effect is diarrhea, which the company describes as loose stools, "like what you'd get after eating a Mexican meal." Overall, about eleven percent of study participants discontinued treatment due to side effects. More common, it seems, per reports from several nelfinavir pioneers is a nasty case of ozone-withering flatulence, which may abate after the first couple of weeks but often returns -- uncontrollably so -- with little notice.
Nelfinavir, like other marketed protease inhibitors, is processed through a pathway in the liver known as CYP3A. Many drugs are processed in this way, and these drugs may interact with nelfinavir. In particular, rifabutin levels in the blood are increased two-fold by nelfinavir, and therefore rifabutin doses should be reduced by one half when used in combination with nelfinavir. Several other commonly prescribed drugs, terfenadine (Seldane), astemizole (Hismanal), cisapride (Propulsid), triazolam (Halcion) and midazolam (Versed), should not be used when taking nelfinavir because competition for CYPA3 by nelfinavir could result in "serious or life-threatening cardiac arrhythmias or prolonged sedation." Rifampin decreases nelfinavir levels by 82%, so the drugs should probably not be combined. Nelfinavir reduces levels of ethinyl estradiol, a common birth control pill, by almost one-half.
On the flip side, nelfinavir interacts with other protease inhibitors in ways that are potentially useful: in combination with indinavir, nelfinavir levels rise by 83% and indinavir levels rise by 51%. In combination with ritonavir, nelfinavir levels rise by 152% and ritonavir levels are unchanged. Some dose adjustments may be called for when combining these drugs. There are no data on the interaction between nelfinavir and Invirase (the currently available formulation of saquinavir), but in a small study of nelfinavir in combination with the new soft-gel formulation of saquinavir, levels of saquinavir were increased by nearly 4-fold with little effect on nelfinavir levels. Unfortunately, the soft-gel capsule formulation of saquinavir is still experimental and unavailable.
There are at present no data on interactions between nelfinavir and approved non-nucleoside reverse transcriptase inhibitors (NNRTIs), such as nevirapine and delavirdine. But in a combination study with DMP-266, an experimental NNRTI, researchers are raising the dose of indinavir from 800mg to 1,000mg q8h -- because of DMP-266's ability to speed up the rate at which indinavir is processed and cleared by the liver. In general, because the NNRTIs are also processed by the CYP3A pathway, interactions with the nelfinavir are expected.
As for resistance (and cross-resistance), Agouron claims that use of nelfinavir will not produce cross-resistance to the other protease inhibitors, but that use of the other protease inhibitors will produce cross-resistance to nelfinavir. In other words, "Use ours first." This convenient assertion is based on the fact that the major amino acid substitution conferring reduced sensitivity to nelfinavir is a D30N mutation. But D30N is only the first mutation to develop in the presence of nelfinavir, and is quickly followed by amino acid substitutions at the more familiar codons: 84, 46, 36 and 71 -- all of which would be expected to reduce sensitivity to the other protease inhibitors. Agouron's self-serving admonitions could still prove warranted, but without any clinical data to go on it is premature at best to make such claims publicly. Since day one, Roche has heavily promoted a similar scenario with saquinavir based on less-than-convincing evidence. Predictably, competing manufacturers of other anti-protease drugs (Merck in particular) have vociferously disputed this. Recent clinical findings (see "Sweating the Switch," in this issue) suggest that, in this particular case, Merck might just be right.
Results of AIDS Clinical Trials Group (ACTG) 320, one of the most important studies of triple-drug antiviral therapy in people with AIDS, were recently released by the National Institute of Allergy & Infectious Diseases (NIAID). The principal investigator was Scott Hammer, M.D. of Harvard Medical School. ACTG 320 is the first study to prove that the protease inhibitor indinavir (Crixivan), when added together with a second new drug (in this case 3TC), to an underlying AZT or d4T treatment regimen, prolongs AIDS-free time and extends survival in people with fewer than 200 CD4 T cells and over three months of AZT experience -- compared to simply adding one new nucleoside analogue (in this case, 3TC).
The study is also the first to prove clinically that combination anti-HIV treatment strategies -- adding at least two new drugs when making a decision to switch therapy -- are more effective than adding single new drugs to antiretroviral treatment regimens. The study was also innovative and flexible in allowing participants to change underlying regimens based on toxicity -- those experiencing AZT toxicity were allowed to use d4T instead. Finally, the study may be the last of its kind, demonstrating as it does that AZT+3TC alone, which only last year proved its superiority to AZT monotherapy, is inadequate as a treatment for advanced HIV disease. Double nucleoside combinations are no longer appropriate control arms in people with AIDS or fewer than 200 CD4 T cells.
ACTG 320, combined with data from the Abbott advanced study M94-247 and Roche saquinavir+ ddC study NV14256 (which demonstrated a survival benefit to the protease inhibitors ritonavir and saquinavir, respectively) when added to nucleoside analogues, further strengthens support for changing the standard of care for people with AIDS and under 200 CD4 T cells to triple-drug therapy including a potent protease inhibitor plus two nucleoside analogues, at least one of which is initiated when the protease inhibitor is begun.
Starting in January 1996, ACTG 320 enrolled 1,156 HIV-infected persons who had fewer than 200 CD4 T cells, more than three months of AZT experience, less than seven days of 3TC experience and no protease inhibitor experience. They were randomly assigned to receive AZT+3TC or AZT+3TC+ indinavir and followed for a mean of 38 weeks (maximum one year). Persons intolerant to AZT at baseline were allowed to take d4T, and persons experiencing moderate, non-AIDS defining disease progression were allowed to switch underlying nucleoside analogue therapy to any combination of the nucleosides AZT, d4T, ddI, or 3TC. They were followed for an AIDS-defining illness or death.
Participants were stratified at baseline by CD4 above or below 50 (38% below, 62% above). 83% of participants were male and 17% female; 27% were black, 19% Hispanic, 3% had hemophilia and 16% had a history of injecting drug use. The mean CD4 count at baseline was 86, and mean age was 39. Virologic analyses are not yet available. The dataset was closed on January 27, 1997. An independent Data & Safety Monitoring Board (DSMB) analyzed the data on February 18, 1997 and recommended that ACTG 320 be stopped early. Study participants were unblinded (told which arm they were on) and offered a number of follow-up treatment options.
ACTG 320 is the third, and perhaps the most well-designed, study to demonstrate a survival benefit by adding a protease inhibitor to nucleoside analogues. Previously, in January 1996, Abbott showed that the protease inhibitor ritonavir, when added to underlying nucleoside analogues (AZT, ddI, ddC, or d4T), prolonged survival during six months of study in people with under 100 CD4 T cells. Later in 1996, Roche showed that its protease inhibitor, saquinavir, when added to the nucleoside analogue ddC, prolonged survival compared with saquinavir or ddC alone.
Like the Roche study and unlike the Abbott study, all ACTG 320 participants received at least one new drug, and those on triple therapy received two new drugs. Unlike the Roche study, however, whose control arms were ddC or saquinavir monotherapy, all ACTG 320 participants were on two nucleosides or two nucleosides plus indinavir. Therefore, ACTG 320 more clearly defines a new standard of care than either of the two previous studies, but all three demonstrate that protease inhibitor regimens confer a survival benefit (a 50% reduction in death) in advanced, AZT-experienced people followed over six to twelve months.
ACTG 320 also proves the benefit of changing more than one antiretroviral therapy at a time. ACTG 320 was the first controlled study to prove survival benefit not just for the protease inhibitors as a new class of drugs, but to prove the survival benefit of a new and revolutionary anti-HIV treatment strategy. In ACTG 320, all participants had received over three months of AZT, and all participants were given at least one new drug -- 3TC -- and half of the participants were given two new drugs -- 3TC and indinavir. This study proves what, after Vancouver, we only guessed and hoped: that, when people with HIV make a decision to switch therapy, the best way is to add at least two new drugs, in order to reduce the chance that the virus can develop resistance to any one new drug. This means that the era of sequential monotherapy, as practiced since 1987, is over, and that the new treatment strategy, at least for people with AIDS, should attempt to administer potent antiviral combination therapy designed to reduce viral levels beneath the limit of detection.
Particular praise goes to the study investigators who are planning a state-of-the-art salvage trial for ACTG 320 participants now on AZT+3TC, rather than simply offering them open-label indinavir, as was done in the past. ACTG 320, like its predecessor protease studies, offers new hope for antiretroviral-experienced people with HIV.
ACTG 320 also suggests that people taking therapy likely to cause maximal viral suppression have superior clinical outcomes. The results of ACTG 320 confirm earlier suggestions from surrogate marker trials (trials measuring viral load and CD4, but not clinical endpoints) Merck 035 and Merck 039, which showed that AZT-experienced persons adding 3TC and indinavir experience a high likelihood (65-85%) of having their plasma HIV levels go below the limit of detection (400 copies of HIV per milliliter of plasma). ACTG 320 suggests that this benefit -- of having virus go "undetectable" -- is associated with significant clinical benefit -- a 50% reduction in AIDS-defining events and death. However, virological analyses of ACTG 320 are incomplete.
AZT+3TC+indinavir is just one of many combination therapy regimens which can drive viral load below detectable limits, and thus slow down progression to AIDS or death. There is nothing unique about the particular combination of AZT+3TC+indinavir. What is new is the clinical confirmation of the ability of maximally-suppressive antiviral combinations to extend health and prolong life. What remains to be determined is whether there is some sort of qualitative difference between going "undetectable" with a potent protease inhibitor combination and going "undetectable" with an all RTI (either nucleoside or non-nucleoside) combination. People who are already taking AZT+3TC should be encouraged NOT to rush out and merely add indinavir to their regimen. This is not what ACTG 320 tested: all 320 study participants were essentially 3TC and protease inhibitor naïve. That is, they had never taken them. When protease therapy is begun, at least one underlying drug-- and possibly both -- should be switched at the same time.
ACTG 320 may (and probably should) be the last trial of its kind in people with AIDS. Although AZT+3TC only last year proved clinical benefit in PWAs (in the CAESAR study), and AZT+ddI only did so in 1995 (with ACTG 175), double nucleoside combinations with approved agents should no longer be the standard of care for PWAs. Triple therapy with two nucleosides [at least one of them new] and a potent protease inhibitor should be recommended as the standard-of-care for people with AIDS or fewer than 200 CD4 T cells. The Abbott and Roche studies plus ACTG 320 demonstrate significant benefit for the class of drugs plus nucleosides versus nucleosides alone.
ACTG 320, like the Abbott and Roche studies, does not prove clinical benefit for people with over 200 CD4 cells. We still don't know when is the best time to start antiretroviral therapy. We don't know enough about resistance and cross-resistance among the protease inhibitors; in particular, whether choosing indinavir as "first-line" protease therapy will make it less likely to later benefit from second or third choice protease inhibitors. Healthy asymptomatics not currently on antiretroviral therapy who wait to begin therapy may have additional treatment options in six months or a year from now. In any case, people initiating triple-drug regimens need to strongly consider how their first regimen will affect future treatment options.
When it comes to corporate positioning for protease inhibitor market share, every company has a yarn to spin. Merck loyalists insist that theirs be used first-line because it's so powerful and, "after all, really requires multiple mutations in order to significantly alter viral sensitivity." Roche (and later, Agouron, in lock step) claim that their protease is the only one to deserve a first-line indication because the mutations elicited with saquinavir and nelfinavir are unique and not nearly as predisposing to cross-resistance as are, say, the indinavir mutations. Since scientists at all the protease outfits seem capable of pulling what ever color rabbit out of their hats deemed most conducive to a successful marketing campaign (and since all cross-resistance analyses to date have been conducted in test tube experiments), trying to sort through the morass of claims and counter-claims has been at times Herculean; at others, Sisyphean. With the stunning and unexpectedly early results of ACTG 333 -- the first study to actually look at the effects of switching from one protease inhibitor to another in patients -- we may finally be getting an idea of how all this mumbo jumbo translates to clinical practice. We may finally be approaching objective truth.
ACTG 333 was a study designed to see how patients with long-term use (median 112 weeks) of saquinavir fared when they switched either to indinavir or to the new formulation of saquinavir, keeping their underlying nucleosides stable. While one could question the ethics of switching only one of a patient's drugs in a "failing" combination, ACTG 333 was terminated early when an interim analysis of the 72 patients (median baseline viral load= 21,000 copies/ml) showed only a negligible virologic benefit to the new protease inhibitor. The mean change in plasma HIV RNA for those switched to the saquinavir soft gel or indinavir was -0.23 log and -0.58 log, respectively. Since this fell short of the -0.7 log minimum acceptable viral load drop stipulated in the protocol design, participants will now be given the option of real-time viral RNA and CD4 measurements and switching to either open-label indinavir or saquinavir soft gel. A letter to patients also recommends that they consider changing their non-protease drugs as well when making the protease switch. (Monday morning QB!) Virology samples were collected on all study participants and the results of genotypic and phenotypic studies are eagerly awaited. Will it turn out that a mutation at 90 or 71 or 84 was responsible for the switching failure? Or will something even more nefarious (i.e., "cross-class" mechanisms) turn out to be the culprit, such as improved activity of the protease enzyme or reduced reliance on protease activity for replication? Remember ACTG 116B/117 and BMS 019? It may be time to beef-up our codon driven conception of antiviral drug resistance...