The FDA's Antiviral Drug Products Advisory Committee met on November 6-8 in a marathon session to consider accelerated approval for Glaxo Wellcome's Epivir brand lamivudine (3TC) and Hoffmann-La Roche's Invirase brand saquinavir, as well as full approval for Bristol-Myers Squibb's Zerit brand stavudine (d4T). TAG was there, getting all the data and dish. Spencer Cox prepared this report.
Glaxo-Wellcome's Epivir brand lamivudine (3TC) is a nucleoside analogue, just like AZT, ddI, ddC and d4T. In combination with AZT, Epivir has been shown to produce increases in CD4+ T cells and reductions in viral load. The data come from four clinical studies, all of which showed that use of 3TC at a dose of 300 mg a day in combination with AZT produced CD4+ T cell increases of 30-70 cells/mm3 and reductions in HIVplasma RNA of about 1.75 logs. In these AZT naïve patients, the CD4+ response was maintained above baseline out to 52 weeks and the RNA reduction remained below baseline up to 24 weeks.
Toxicities in the controlled trials were those commonly seen before with the nucleoside analogues: headache, nausea, malaise, fatigue, diarrhea, neuropathy, anemia (low red blood cell counts), leukopenia (low white blood cell counts) and some nasal symptoms. Some pancreatitis was also seen in adult patients--and in over 15% of the pediatric patients receiving 3TC.
There was very little data on the 35,000 patients who have received 3TC through Glaxo Wellcome's industry-maligned expanded access program, due to the company's "passive" reporting system for adverse events. This was cause for considerable consternation
For the future, Glaxo Wellcome has planned a large, multinational study to examine the possible clinical benefits of antiretroviral combinations with 3TC. The study will involve 1,200-1,500 patients with CD4+ T cell counts 25-250 cells/mm3. Study participants will be randomized to receive either standard-of-care, standard-of-care + 3TC or standard-of-care + loviride, a TIBO-like non-nucleoside reverse transcriptase inhibitor. (If standard-of-care becomes AZT+ddI, will this study end up testing double combination reverse transcriptase inhibitor therapy to triple ?) The study is planned to be completed by the end of 1996. Finally, the company has planned a study in approximately 700 antiretroviral naïve pediatric patients comparing 3TC+ddI to 3TC+AZT+ddI. This study could yield important cross-resistance data on the combination of ddI and 3TC.
During the hearing, several aspects of Glaxo Wellcome's NDA troubled the advisory committee. To start, Glaxo Wellcome was applying for a pediatric indication based on new regulations that require only evidence of efficacy in adults, and then supplemental pharmacokinetic data to justify comparable dosing in children. The pediatric data, though, suggested that differing absorption rates in children often result in higher blood levels of drug in children than in adults, and a very high rate of pancreatitis (>15%). Much to our dismay, Dr. David Feigal, Director of FDA's Antiviral Drug Products Division, informed the committee that considering this clinical data on 3TC's effect in kids would violate the spirit of the regulations; the committee was to ignore the clinical data and focus only on the required pharmacokinetic data. (Babies don't vote and, after all, it is an election year.)
Also, there has been, for some time, data showing that the use of 3TC produces an RT mutation that also confers resistance in vitro to ddI and ddC (the 184 mutation). As the company was suggesting an indication for first-line therapy, this raised troubling questions about the wisdom of using 3TC before exhausting the demonstrated clinical benefit of ddI, AZT+ddI and/or AZT+ddC. In addition, granting a company an accelerated NDA for an indication where suitable therapy already exists required a decision by the committee that the treatment offered "meaningful improvement over existing therapy."
Given the recent findings of ACTG 175 and the Delta study, the committee was having some difficulty making this determination, until Dr. Feigal generously stepped in (yet again) to note that this clinical data (that is, Delta and 175) should be ignored, explaining that these studies had not yet been presented to the FDA. There were questions as to whether or not the planned pediatric study could, in fact, confirm the use of AZT+3TC in antiretroviral naïve adults. (This deal was apparently cut during a secret meeting between Glaxo Wellcome and the Advisory Committee weeks before the hearing.)
Finally, the committee voted to approve AZT+3TC for "use in HIV infection." TAG's Mark Harrington testified that the indication was "unwise, a violation of the accelerated approval regulations, a disservice to the public health, a disincentive to industry to conduct clinical validation studies and an invitation to Glaxo Wellcome to mount an unprecedented promotional campaign touting the AZT+3TC combination to a population broader than any accelerated approval past." TAG had supported an indication only for patients with advanced or progressive HIV disease.
On Tuesday, the FDA committee considered approval of an accelerated NDA for Invirase brand saquinavir, Hoffmann-La Roche's less-that-stellar first entry into the protease sweepstakes. The NDA requested approval of saquinavir for use in combination with any approved nucleoside analogues in adults with advanced HIV infection, and for use alone in adults with advanced HIV infection who are intolerant to or failing approved therapies.
The company presented data from uncontrolled phase I studies, as well as from two phase II studies: ACTG 229, a randomized trial comparing saquinavir+AZT vs. AZT+ddC vs. AZT+ddC+saquinavir; and interim data from NV14256, a study of 900 antiviral experienced patients comparing saquinavir vs. ddC vs. saquinavir+ddC. All told, saquinavir was pretty disappointing. As monotherapy, the surrogate marker responses were roughly comparable to AZT; in combination with AZT, the surrogate responses were not much better than what we've seen with AZT+ddC. Roche is currently conducting two studies which hope to provide clinical confirmation of saquinavir's usefulness: the first being simply clinical follow-up of the NV study; and the second, a study called SV14604, which will follow 3,200 antiviral naïve patients for 80 weeks, comparing AZT vs. AZT+ddC vs. AZT+saquinavir vs. AZT+ddC+saquinavir.
Some of saquinavir's disappointing surrogate marker responses (and also the relatively good toxicity profile) may be due to the drug's poor bio-availability. As currently formulated, only 4% of the drug makes it into the blood, which understandably prompted the committee to pose the question, "Why is Roche coming to us now with a drug that isn't even bio-available?" Furthermore, the committee was concerned that, while Roche was seeking an indication for saquinavir to be recommended for use with combinations of any of the approved nucleoside analogues, it had conducted drug interaction studies with only AZT and ddC. Finally, the conflicting claims from Merck and Roche regarding possible cross-resistance between their protease inhibitors have not yet been resolved. TAG supported accelerated approval of both combination and monotherapy indications for saquinavir, but the committee recommended approval only for use in combination. The FDA granted saquinavir licensure on December 5, and it was in pharmacies the following day--at an annual price tag of around $6,000.
As TAG predicted in May 1994, Bristol-Myers Squibb's 019 study, which was intended to confirm the clinical benefit of Zerit brand stavudine (d4T) compared to continued AZT in patients who had been extensively pre-treated with AZT, failed to find a statistically significant difference in time to clinical progression or death.
The FDA committee reviewed the BMS 019 study for consideration of full approval for d4T. The drug had been granted accelerated approval in May of 1994, over TAG's concerns that the planned follow-up study would not be capable of measuring the probable treatment effect. Only by using the study's initial "composite endpoints" (a combination of: CD4 T cell decline to < 50% of baseline, new or recurring opportunistic infection or death) or by post hoc stratification by AIDS diagnosis, could a d4T benefit be shown.
The committee explained that they could not recommend a broader indication for d4T based on this careful re-analysis of the BMS 019 study. NCI's Dr. Lawrence Freidman summed it all up the most succinctly: "I've learned over the years," he began, "that in order to get along with people I should never say 'I told you so.' But as I was not at the accelerated approval hearing for this drug, I can get away with it by saying 'he told you so,' and by 'he' I mean Mr. Gregg Gonsalves from the TAG organization." Ultimately the committee recommended full approval of the drug for use in patients who are "failing or intolerant to AZT," with no implication that the drug is better than AZT in these patients. Again, as TAGline goes to press, no FDA action has yet been taken.
In a separate and somewhat provocative presentation, data were shared comparing ddI+d4T to either drug alone, and the combination produced dramatic surrogate marker responses (up to a 2.5 log reduction in plasma HIV RNA and a mean CD4 count increase of 60-80 cells). Multiple studies are planned, by both BMS and the ACTG, comparing d4T to other agents alone and in combination.