As Amgen's Prolia Goes, So Goes The Market: Reaction To AC Recs

FDA and the Reproductive Health Drugs Advisory Committee's focus on small imbalances in Amgen's trial data for Prolia (denosumab) sparked disparate reactions on the Street, with some analysts trimming their estimates and others holding tight.

The committee focused much of its attention on the risk of malignancies and the potential for tumor progression in patients with prior history of malignancy or in patients with existing tumors. While the concern was sparked by three deaths in a dose-finding trial due to neoplasms - all of which were new malignancies - treatment arms throughout Amgen's four pivotal trials showed comparable safety profiles with only slight imbalances.

The panel voted unanimously to recommend approval of the RANK ligand inhibitor for the treatment of osteoporosis in postmenopausal women - a critical win for Amgen - and also for treatment of bone loss due to hormone ablation therapy (HALT) in men with prostate cancer, but Amgen lost out on two additional indications it was after (¹ (Also see "Amgen's Mixed Advisory Committee Votes On Prolia Still Offer A Good Start" - Pink Sheet, 13 Aug, 2009.)).

As a result of Amgen's failure to gain the panel's endorsement for the prevention of osteoporosis in postmenopausal women, as well as treatment or prevention of bone loss due to HALT in women with breast cancer, analysts such as UBS, Lazard, and Cowen trimmed their sales forecasts for the drug.

But Deutsche Bank's Mark Schoenebaum stressed "Amgen got what they really needed at the panel: a vote for approval of d-mab in PMO treatment." Schoenebaum believes the negative votes will have only minimal impact on peak Prolia sales estimates and thinks there is significant confusion around the market sizes for the different indications.

The advisory committee's mixed reaction to Prolia also means Amgen must be on its game at the upcoming European Society for Medical Oncology conference in September, where it will present data on Prolia's effect in late-stage cancer patients with existing bone metastases to treat skeletal-related events (SRE), since this is the market that will be of utmost importance to Amgen's future growth.

On July 27, Amgen announced Prolia will be co-marketed by GlaxoSmithKline in Europe, Australia, New Zealand and Mexico. Under terms of the deal, GSK has rights to sell Prolia for PMO indications and in exchange, it will make an initial payment, including near-term milestones, totaling $120 million, and pay royalties down the road.

Amgen keeps all rights to the drug in the U.S. and Canada and will sell it in the oncology market in Europe and other markets. GSK will have the rights to register and commercialize Prolia for all indications in countries where Amgen does not currently have a commercial presence - including some fast-growing emerging markets such as China, Brazil, India and South Korea (² (Also see "GSK Denosumab Deal Gives Amgen A Leg Up In Primary Care" - Pink Sheet, 3 Aug, 2009.)).

The drug's application is pending regulatory review in the European Union and several other countries (³ (Also see "Why Doesn't Pharma Get Smaller?" - In Vivo, 1 Jun, 2009.)).

Market Reactions Mixed

Despite securing an approval recommendation for PMO, Amgen's Prolia won mixed reviews from the Street because of the exceeding attention paid to the small imbalances in the drug's safety profile. Cowen, for instance, trimmed its 2015 projection below consensus estimates in the bone loss setting from $1 billion to $865 million even as analyst Eric Schmidt professed confidence that Amgen shares would outperform the market by 15 to 20 percent in the next 12 months.

Lazard, meanwhile, maintained a hold rating for Amgen, and reduced its original 2011 estimate of $1.33 billion for Prolia. UBS' 2013 Prolia sales estimates likewise moved south from $2.7 billion to $2.5 billion ($1.88 billion in PMO, $12.0 million in HALT, and $637 million in SRE). Baird also noted that the harping of the panel on safety issues may bring the 2010 consensus down to $375 million from $410 million.

"While questions surrounding the malignancy issue were largely expected, we were somewhat surprised at the extent of discussion given the statistical insignificance of the safety data," noted UBS analyst Mageh Shenouda in a report.

In a Deutsche Bank analyst call-in, Schoenebaum said the advisory committee meeting may have eliminated about $250 million in peak Prolia sales, which is 7 percent to 9 percent of the peak potential, but predicted worst case that Prolia would still be a billion dollar-earner for the big biotech.

All Eyes On Amgen For SRE Indication

Clearly some analysts are concerned that safety issues raised at the Aug. 13 advisory committee might spell trouble for Amgen as it attempts to gain approval for Prolia in the SRE indication, a market that could be worth as much as $2.4 billion. Schoenebaum noted in his call to investors that data slated to be presented at the ESMO meeting will provide a critical read on the drug's chances of success. Amgen is conducting three major SRE trials, each with more than 1,500 patients, using Novartis' Zometa as a comparator.

During Amgen's second quarter call, company reps said the study in women with breast cancer showed superiority to Zometa for both delaying time to the first on-study SRE and delaying time to the first-and-subsequent SRE. Both results were statistically significant (⁴ (Also see "Amgen, GSK To Team Up On Denosumab Commercialization in Europe" - Pink Sheet, 27 Jul, 2009.)).

They also said Prolia had similar time to first SRE compared to Zometa, which was statistically significant for non-inferiority in a solid tumor study. However, while the delay in time to first SRE and time to first-and-subsequent SRE were numerically greater for patients taking Prolia, they were not overall statistically significant.

In the studies, infectious adverse events, overall survival and time to cancer progression, however, were all balanced between the two treatment arms - although it's hard to say what that exactly means since Amgen claimed the same thing about the four pivotal trials.

"Although yesterday's panel made it clear that the FDA will hold drugs for supportive care conditions in oncology to high standards, our confidence in denosumab's approvability and commercial potential for the treatment of SREs remains high," Cowen predicted, adding that it believes Amgen's studies were rigorously conducted and establish no worsening of underlying disease in patients treated with denosumab vs. Zometa.

It noted that consultants view Prolia as modestly superior to Zometa in terms of efficacy, safety, convenience, and profit potential, and expect the drug to capture and grow beyond Zometa's $1.4 billion world-wide market.

The REMS Effect

Some analysts also cringed at the prospect of a Risk Evaluation and Mitigation Strategy given yesterday's vote 12 to 1 in favor of a REMS. It's almost a certainty that that added oversight will be tied to Prolia's approval.

Thanks to its experience with NPlate , Amgen is no stranger to REMS, and has proposed extensive pharmacovigilence and risk plans for Prolia, including an observational study in approximately 380,000 women with postmenopausal osteoporosis for five years.

The REMS will include a Medication Guide and a health care provider communication plan focused on hypocalcemia, skin infections and osteonecrosis of the jaw (an infection common to other osteoporosis medications).

Furthermore, the sponsor is still culling safety data from five studies in postmenopausal women, three studies in HALT and five additional registrational studies in advanced cancer.

The Cowen note said that while it is not concerned by the negative panel vote on PMO prevention or HALT-induced bone loss, the 12-1 vote in favor of establishing a REMS program could curtail commercial uptake.

Shoenebaum, however, noted that it would be hard to predict the impact of any REMS until the actual monitoring details are revealed. He added that he would be "shocked" if the REMS was as extensive or limited as Tysabri's TOUCH program, and that it is unlikely to involve restrictive limitations.

- Lauren Smith ( ⁵ [email protected] )