Prostate Cancer Prevention Endpoint May Have Been Key Problem As ODAC Votes Down BPH sNDAs
Executive Summary
The problems that GlaxoSmithKline and Merck had in trying to convince FDA's Oncologic Drugs Advisory Committee to add labeling on prostate cancer risk reduction to Avodart (dutasteride) and Proscar (finasteride) may have been avoided had the firms selected a stronger mortality endpoint for their trials.
|
The problems that GlaxoSmithKline and Merck had in trying to convince FDA's Oncologic Drugs Advisory Committee to add labeling on prostate cancer risk reduction to Avodart (dutasteride) and Proscar (finasteride) may have been avoided had the firms selected a stronger mortality endpoint for their trials. The pivotal studies for both drugs took biopsies of patients over a number of years, and, although they both showed significant reductions in cancer rates overall, they suffered a pair of lopsided "no" votes at the Dec. 1 meeting, based primarily on worries that they increased high-grade prostate cancers.
ODAC Votes Against Avodart, Proscar
That may be the end of the story for efforts to broaden the labels for Avodart (which might go generic as soon as 2013) and Proscar (which went generic in 2006). But it is worth considering how a different study design might have produced different results, especially since there are several other benign prostatic hyperplasia products still on patent, although there's no indication that any are being studied for cancer risk reduction. How best to develop prevention claims for drugs will be especially important for industry as health care reform is implemented because the legislation mandates zero-copay coverage of qualifying preventive therapies (Also see "Adult Vaccines Stand To Gain From Preventive Care Rule, Outreach Is Next" - Pink Sheet, 19 Jul, 2010.). Avodart and Proscar are 5-alpha reductase inhibitors currently approved to treat BPH. Only GSK explicitly asks in its sNDA for an indication of reduction in prostate cancer risk (which would be a first in the U.S.), but Merck's request to include the results of the Prostate Cancer Prevention Trial of Proscar in the drug label is tantamount to the same thing, FDA noted in background materials prepared before the ODAC meeting (Also see "Prostate Cancer Risk Reduction sNDAs Go Before ODAC, Pitting Statistical Versus Clinical Significance" - Pink Sheet, 29 Nov, 2010.). Had the companies opted for a mortality endpoint instead of biopsies, the trials would have been much longer and more expensive, but would likely have avoided the debate that the sponsors were forced to have about the high-grade cancer signal. GSK and Merck argued that the detection of more high-grade cancers in the treatment arm was a function of better diagnoses because the products reduced prostate volume, not because the products themselves caused the cancers. The committee and FDA didn't buy it, leading to some surprisingly loud exchanges between the presenters and the panel (see (Also see "Tempers Flare At ODAC Meeting, And Pazdur Must Call "Order In The Court"" - Pink Sheet, 6 Dec, 2010.)). If the sponsors had outcomes data and their theory was correct, they might have been able to avoid raising their voices and instead could have pointed to slides showing that patients lived longer with treatment, since presumably those hidden high-grades would eventually fell patients in the control group. GSK declined to say whether it is considering outcomes trials with a mortality endpoint for Avodart for prevention of prostate cancer, although none are under way at present. "We feel Avodart will continue to play an important role in the treatment of BPH, which is a condition that affects more than 50% of men over age 50," the company said. Prevention Needs Bigger Numbers Even if the sponsors had such outcomes data, they would have likely had to clear another big hurdle: a "heightened" level of certainty required for products that reduce risk (see (Also see ""Risk Reduction" Indications Face A Higher Bar For Approval, May Require Specialized Prescribers" - Pink Sheet, 13 Dec, 2010.)). The proposed chemoprevention indication for Avodart (and the implied one for Proscar) called up thoughts of the oncology drug tamoxifen, which clinical studies have shown to have a preventive effect against breast cancer, in the mind of the committee's nonvoting industry representative, Gregory Curt, U.S. medical science lead for emerging products at AstraZeneca Oncology ("STAR Trial Results Support Chemoprevention in Primary Care," Health News Daily, June 6, 2006). Clinical data showed that when using tamoxifen, "you had to treat 6,000 patients to prevent 60 cancers. So that's about 100 patients treated for one that's prevented. And I would submit that tamoxifen is a drug that has a worse safety liability than [Avodart and Proscar]," Curt said. AstraZeneca originally marketed tamoxifen, which is now available as a generic. Historical data are useful, responded the National Cancer Institute's Wyndham Wilson, the ODAC chair, but added, "I think what is of primary concern is whether or not [Avodart] may actually be leading to a worse grade, and given everything we know about grade, a worse outcome. So I think that's a fundamental difference between the tamoxifen study and this one; I don't think there was any evidence you were possibly making things worse [with tamoxifen]." Had the panel and reviewers been more impressed with the drugs, it's possible to imagine the advisory committee would have focused on risk management options to make sure they were used wisely rather than approvability in general. But that high level of certainty eluded the two drugs, which showed clear efficacy against low-grade prostate cancer – tumors with Gleason scores of 6 or less, which may not have any clinical relevance for men who harbor them – while seeming to increase the risk of high-grade tumors, those with Gleason scores of 7 or 8 to 10 that may lead to metastasis and death (Also see "Nothing Benign About It: BPH Drugs' Bids For Prostate Cancer Risk Reduction Voted Down By Advisory Committee" - Pink Sheet, 1 Dec, 2010.). The apparent increase in high-grade cancers did not instantly doom the applications; the advisory committee members discussed how the risk trade-offs might play out if the new labels were approved. "In the end, if we accept the facts, the numbers the FDA has given us, for every 200 patients you treat [with Avodart], you'll get one additional case of high-grade disease. We have to balance that against, I would say in the worst case, you have to treat 60 patients to get one decrease in what they call low-risk cancers," David Penson, Vanderbilt University Medical Center, said. "In my office, when I'm following patients and monitoring their PSA [prostate-specific antigen], I'm frankly comfortable with that risk profile." But Penson shared the concern of Richard Pazdur, director of FDA's Office of Oncology Drug Products, who questioned how effectively the products would be used by a broader spectrum of doctors. "We're talking about literally hundreds of thousands of patients being exposed to this drug, with many different practitioners having the right to prescribe this drug," Pazdur stressed. That was enough to persuade most of the waverers, and Avodart's sNDA went down in defeat 14-2, with two abstentions. The changed label indication Merck had requested was defeated 17-0, with one abstention. Building A Case On Biopsies In the aftermath of the sound rejection GlaxoSmithKline received for a prostate cancer risk reduction indication for its BPH drug Avodart (dutatsteride), the design of the pivotal REDUCE trial seems questionable. The first issue FDA highlighted in its background document about the drug "is the questionable relevance of the results of this trial to clinical practice, where only 'for-cause biopsies' are performed. The high prevalence of latent prostate cancer at autopsy (30-70%) in men 50-80 years of age makes time-mandated, repeated biopsies without a clinical indication of questionable clinical relevance." In an e-mail sent after the Oncologic Drugs Advisory Committee vote, GSK said, "The REDUCE trial was purposely designed to minimize for-cause biopsies in order to ensure that subjects in both treatment arms had an equal opportunity to be diagnosed with prostate cancer so that an accurate treatment effect could be ascertained." "In other words, REDUCE was designed to investigate a scientific question using rigorous standards in a way that would give health care providers clinically applicable insights," GSK continued. "Men who have one negative biopsy often will be closely monitored and re-biopsied in following years, so it’s very plausible that they will be re-biopsied at some point in two to four years following the first negative biopsy. In this respect, REDUCE replicated existing clinical practice." If that's what doctors are doing, FDA doesn't think too much of their practice either. "It is apparent that if men are biopsied frequently without cause, a high incidence of latent low-risk prostate cancers will be detected, most of which pose little threat to men during their lifetime," the agency's briefing documents states. "A decrease in incidence of these latent cancers is of limited value to patients." Another aspect of the trial design that proved quite problematic – GSK's decision to break out the data for the high-score Gleason tumors, which turned out to increase for subjects taking Avodart – "was always part of the REDUCE study protocol," the company said. Adjucating the Adverse Data The extent to which Avodart was associated with an increase in high-grade prostate cancers provoked intense debate at the advisory committee, with Yang-Min Ning, a medical officer in FDA's Division of Drug Oncology Products, repeatedly questioning Christopher Logothetis of the University of Texas MD Anderson Cancer Center, who delivered the presentation on the incidence of high-grade prostate cancers in the Avodart studies for GSK. Ning disagreed with the sponsor's interpretation of the numbers in the pivotal REDUCE study and the supporting data from the COMBAT and REDEEM studies, and asserted that the sponsor had artificially depressed the number of high-grade tumors in the REDUCE study in particular by removing patients with low-grade tumors from the study during the first and second years who otherwise would have progressed on to develop high-grade tumors. "We did not remove any patients," Logothetis responded. "All we did was to say the 558 patients on placebo and the 417 patients on dutasteride [who were diagnosed with Gleason score 5, 6 or 7 tumors in the first or second year] discontinued the study at two years because they had the primary event of the study. If 5% of them … had continued the study and been biopsied at four years, we would have expected 28 Gleason 8-10s in the 558 placebos and 21 in the 417 dutasteride." Neither Ning nor Brent Logan, associate professor of biostatistics at the Medical College of Wisconsin, was happy with that explanation. "There's a lot of assumptions in those imputed additional cases that may have come out of those five-to-sevens that were removed from the study in years one and two," Logan said. "The assumption is that the rates of conversion from five-to-seven to eight-to-10 in the following years are the same between those two groups, and that may not be the case because we have differential upfront diagnosis. And then we're assuming, of course, that it's 5%. That's from an external study; we don't know if it's generalizable to this particular data set." Addressing Paolo Paoletti, senior vice president for oncology R&D at GSK, Wilson put the question in stark terms. "I understand that you agree what the data shows, but you're trying to rationalize, justify, statistically analyze it away. And we are faced with patients that have no medical problems. And we are being asked to take this data, see it clearly and then try to accept a bunch of post-hoc analyses to try to rationalize and justify why the data isn't showing what it's showing, and that's got me very concerned." "What matters, as best we can tell, is trying to reduce high-grade tumors, because those [patients], all of the clinical data shows, are most likely to die from prostate cancer," Wilson said. "We're trying to prevent prostate cancer deaths. We're not trying to prevent a histological diagnosis." Moreover, he said, the trials contained no evidence that Avodart kept tumors from progressing from lower to higher Gleason scores. |