|
|
|
|
||
Baby Hold on to Me; Whatever Will be Will be but Makena Revenue Still Likely Ours to SeeBaby Hold on to Me; Whatever Will be Will be but Makena Revenue Still Likely Ours to See Raymond James ($3.37) March 10, 2019 We reiterate our Strong Buy rating on Antares and would utilize Friday's weakness on news flow related to AMAG's confirmatory trial for use of Makena in prevention of pre-term birth to accumulate the name. While confirmatory trial failure introduces a possibility the product could eventually be withdrawn, a risk long relegated to the back pages of AMAG risk disclosures given Makena's 2011 approval, we believe the risk, though now highlighted and reintroduced into discussion, is minimal. Friday's 12% decline effectively eliminated half of Makena's contribution to our sum-of-the-parts valuation framework and given continued momentum in other key, larger dollar contribution value drivers such as Xyosted as well as partnered programs such as generic EpiPen, we think the share price swoon should be used opportunistically ahead of an anticipated top line growth spurt in 2019-2020. Makena Controversy Reemerges Once Again: ● Friday AMAG announced the the 1,700 patient PROLONG trial, a large confirmatory trial intended to support the 2011 accelerated approval of Makena for use in preventing preterm birth prior to 37 weeks gestation in women with a prior spontaneous birth failed to reach either co-primary endpoint; incidence of pre-term delivery prior to 35 weeks or reduction in neonatal morbidity/mortality index. ● Conversion to full approval via the PROLONG study was required as part of the original 2011 accelerated approval and introduces the possibility that Makena could eventually be withdrawn from the market. ● Historically FDA has been patient if not flexible in enabling sponsors to eventually convert accelerated to full approvals. Only 16 of 194 total accelerated approvals granted since 1992 have actually been withdrawn with 110 applications converting to full approval after an average of four years. There were 92 accelerated approvals granted prior to Makena's 2011 FDA Ok; 71 have been converted and perhaps more importantly at least eight remain outstanding. ● Lack of separation on PTB rate and relatively low incidence is surprising in light of substantially higher incidence rates and placebo separation in original 463-patient study. Confirmation of infant morbidity/mortality benefit was a big unknown in PROLONG. AMAG has suggested subgroup analysis as a next step. A benefit in either PTB rates or mortality/morbidity in either U.S. patients or high-risk women would give us high confidence in Makena's eventual full approval. ● A competitor also dog-piled on the weakness by highlighting a risk disclosure from AMAG's recently filed 10K revealing reports of product issues with ATRS supplied auto-injectors. We think this was overblown as FDA adverse event data suggest that less than 3% of Makena AE's in 2018 are related to device issues, mainly pain, and none actually call out failure. Valuation: Makena represents 16% of our SOTP fair value estimate for ATRS shares of $5.65. Friday's move more than adequately discounts slight risk Makena revenue stream could eventually disappear in our view. See pages 4-5 for details. ATRS shares swooned 12% in sympathy with partner AMAG post release of that company’s phase III data from the PROLONG study, a confirmatory trial designed to convert Makena’s (hydroxyprogesterone caproate) accelerated approval status into a full approval per conditions outlined in FDA’s 2006 approval of the drug. Makena is approved for reducing the risk of preterm birth in women with a singleton pregnancy who have a history of singleton spontaneous preterm birth. Risk had seemingly not existed prior to Friday, buried deep in risk section of financial filings and long forgotten since 2011 approval. By failing to provide confirmatory data to support the accelerated approval, in theory at least Makena could be at risk of being withdrawn assuming that AMAG cannot eventually convince the FDA some form of sub-group analysis focusing on differences between U.S. and ex-U.S. patients or in patients at high risk of pre-term delivery. However, withdrawal of drugs approved under accelerated pathway is extremely low (16%) and the time to eventual conversion can extend for years suggesting the agency is patient if not flexible. Makena failed to hit either of the co-primary endpoints; incidence of preterm delivery at 35 weeks (11% vs. 11.5% placebo rate) or a neonatal morbidity and mortality composite (5.4% vs. 5.2% placebo rate) in the phase III PROLONG (Progestin’s Role in Optimizing Neonatal Gestation) study. PROLONG is by far the largest trial ever conducted evaluating the role of progesterone in the prevention of pre-term delivery in woman with a history of spontaneous preterm delivery. The trial began enrolling in 2009, prior to Makena’s February 2011 FDA approval, and included a total of approximately 1,700 woman, the vast majority of which were located outside the U.S. While the disappointing topline results of AMAG’s PROLONG study – as part of an approval commitment with the FDA’s subpart H accelerated approval process – is raising concerns over Antares-partnered Makena sales, we believe the selloff is overdone as there are a few factors that we need to consider before jumping into any premature conclusions. These factors include the likelihood of FDA actually pulling the drug out of the market and whether the data is strong enough to support the withdrawal. Key will be subgroup analysis of high risk and U.S. patients. If there is a morbidity/mortality benefit in either subgroup, there is no risk of withdrawal in our view. Even assuming subgroup data analysis is similar to overall PROLONG results, there is strong supportive analysis for both positive benefits in PTB and infant mortality/ morbidity outcomes that support FDA not requesting Makena withdrawal. Idea that FDA will allow cGMP produced HPC (Makena) to remain on market is due to risks around compounded formulations which is not the safety net many assume in our view. AMAG’s Sub-Q Makena utilizes Antares’ QuickShot auto-injector. Under a license, supply and distribution agreement entered in 2014 with Lumara Health (subsequently acquired by AMAG), Antares has granted AMAG an exclusive worldwide, royalty-bearing license to develop a variant of the QuickShot auto-injector for use in a subcutaneous formulation of Makena. Per the agreement, Antares is eligible to receive payment from AMAG for each device and tiered high single to low double digit royalties and sales milestones related to sales of subcutaneous Makena. We believe it’s also important to point out that AMAG had previously indicated that Sub-Q Makena sales of $40M to be a sustainable quarterly sales figure and as a good proxy for long-term revenue stream expectation of annualized rate between $160M-$200M. Sub-Q Makena generated sales of $35M in 4Q18. We estimate that Sub-Q Makena has peak sales potential of roughly $173M driving stand-alone EPS of $0.10-0.12 in 2022. On NPV basis, we ascribe value of roughly $142.0M to the asset, which equates to a per share value of $0.90. Another concern that was raised alongside the development was possibility of malfunctioning of Antares’ auto-injector device that are being used in Sub-Q Makena. However, after carefully examining FDA’s Adverse Event Reporting System, of roughly 250 Makena related AEs reported after March 2018 less than 3% were device-related incidences. Backgrounder The original formulation of hydroxyprogesterone was approved in 1956 under the trade name Delalutin for use in pregnant woman for uses including threatened miscarriage. Though evidence emerged of a benefit in PTB, use of progestogen therapy for the indication was largely abandoned until a meta-analysis published in 1990 showed a benefit in decreasing pre-term labor, PTB rates as well as reducing low birth rates. Makena was approved in February 2011 to reduce the risk of preterm delivery before 37 weeks of pregnancy in pregnant woman with a history of at least one spontaneous preterm birth (PTB) in a landmark trial was conducted by the National Institute of Child Health and Human Development Maternal-Fetal Medicine Units (NICHD-MFMU) Network. Approval was granted on the basis of randomized double-blind trial in 463 women ages 16 to 43 who had a history of PTB in which Makena (HPC) treated patients experienced pre-term delivery at a rate of 36% vs. 55% in the placebo group. The trial was originally expected to enroll a total of 500 patients though was stopped early by its DSMB after positive interim analysis. As a condition of the accelerated approval, the FDA required a confirmatory trial with 5% of subjects enrolled prior to FDA approval. In addition to a reduction in PTB, the confirmatory PROLONG trial was powered to show both a direct clinical benefit, i.e. a reduction in prespecified neonatal morbidity and mortality index. Given the requirement to show a neonatal benefit, the trial enrolled some 1,700 patients to provide 90% power to detect a 35% reduction in the rate of neonatal morbidity/mortality. It was always contemplated that given the expected standard of care in the U.S., the majority of patients would be from outside the U.S. and FDA only required that 10% of patients in the trial actually be from U.S. trial sites. FDA Accelerated Approvals Makena was approved under a 505(b)(2) NDA and was granted Accelerated Approval under the provisions of 21 CFR 314.510, Subpart H. Since enactment of the accelerated approval pathway back in the early 1990’s, the FDA has approved a total of 194 drugs/indications using surrogate endpoints as the basis. While there have been several recent high profile withdrawals of drugs approved under the accelerated approval pathway, generally FDA appears to have been extraordinarily patient and flexible in terms of allowing sponsors to fulfill their commitments to generate confirmatory data. Only 16 drugs/indications out of a total have actually been withdrawn and there are at least eight products with outstanding commitments longer than those of Makena. Of the 194 total approvals, only 16 have actually been withdrawn though not always for reasons of failing to deliver the required confirmatory data. The majority of these accelerated approvals, 110 in total, have been converted to full approvals based on confirmatory trial data. The average length of time to convert from accelerated approval to full approval is 1,461 days or four years. Prior to Makena’s 2011 approval, the FDA had granted 92 accelerated approvals of which 71 have been converted and eight remain approved but not yet converted on the basis of confirmatory data. Thirteen applications for either drugs or indications were granted accelerated approval prior to Makena and then subsequently withdrawn though it took an average of nearly nine years before the withdrawal decision was made. Subsequent to Makena’s 2011 approval, the FDA has granted 88 accelerated approvals of which only 28 have yet to converted. The average conversion times of these drugs has been 2.2 years. PROLONG Implications While the PROLONG study is the largest PTB study every conducted, the original Makena approval study itself contained nearly 500 patients and should have served as a reasonably good proxy for at least endpoints based on the incidence of PTB at less than 35 weeks. The composite mortality index is a completely different story as this evidence was never explored in earlier pivotal trials. At first glance, the most surprising aspect of the of the PROLONG study results is the markedly different Makena and placebo response rates between the original pivotal trial and PRLONG. Though PROLONG enrolled nearly four-fold the number of patients, the placebo PTB incidence rate of 55% in the pivotal vs. 11.5% in this trial seems to suggest other factors influenced the outcome. The difference between PTB rates on Makena active, 36% in the pivotal vs. 11% in PROLONG similarly suggests other factors likely influenced the outcomes. Difficult to discern exactly why the results of PROLONG were so different than prior studies. As suggested elsewhere, skeptics of HPC therapy have long considered the 55% placebo event rate in the original trial supporting approval as artificially high. It was well known at the outset of the trial that the majority of patients would in fact be recruited from outside the U.S. so the high prevalence of non U.S. patients itself was not a surprise however the high concentration of non U.S. patients would likely yield vastly different baseline characteristics in terms of facial makeup and baseline PTB rates. Additionally, potential changes in neonatal care over the long trial period including supplemental oxygen management, cooling for encephalopathy could have influenced the outcome measures. Also worth noting that PROLONG, unlike the NICHD-MFMU trial, allowed subjects to receive progesterones other than HPC with a four-week washout period before receiving Makena; a measure that could also have led to vastly different results. The Nitty Gritty Use of progesterone therapy to prevent spontaneous pre-term delivery has always been controversial. Despite the positive outcome in the original NICHD-MFMU trial, skeptics have long pointed to what they have viewed as an artificially high PTB rate (55%) occurring in the placebo arm as a basis for doubting the efficacy of the HPC as a preventative measure for PTB. It is important to note however that larger non-pivotal meta-analysis including a July 2013 Cochrane review do confirm a multitude of positive outcome measures using progesterone therapy in woman with a past history of spontaneous pre-term birth. The use of progesterone in this patient population from eleven randomized trials including 1,900 woman and infants confirm statistically significant reductions in pre-term birth as well as in perinatal mortality. There are also confirmed statistically significant benefits in terms of infant outcomes such as birthweight, use of assisted ventilation, necrotizing enterocolitis, intensive care admission, intensive care admission and death. Initial reaction to the PROLONG study results seemed to fall back on the idea that if Makena were removed from the market, use of HPC would continue thus forcing physicians to revert back to compounded product which satisfied usage requirements until Makena’s 2011 approval. The idea here is that given the problematic history of compounding operations, particularly involving sterile injectables, that FDA would somehow provide a safety net around Makena because the agency would rather have physicians using HPC manufactured under cGMP conditions than by compounders especially given the nature of the patient population. Though we would agree that this could be an eventual consideration, we think relying on this argument alone will lead to a false sense of security around Makena’s future. If the agency doesn’t believe the clinical evidence warrants use of the product in the approved indication in the specified population, the risk of withdrawal is real irrespective of prior usage of largely unregulated compounded formulations. Lilly’s Lartruvo, in combination with doxorubicin, was the first drug approved as a first line treatment in patients with advanced soft tissue sarcoma in decades and yet was withdrawn following a failed confirmatory study; it can happen. That said, while the PROLONG study did not show the expected benefits, it is our view that significant evidence exists, both in the prevention of PTB as well as infant outcome measures to warrant continued usage of the drug as well as support continued marketing approval for Makena. AMAG intends to conduct sub-group analysis in both high risk-patients as well as U.S. patients in order to determine in the lack of consistency with prior trials is in fact due to treatment and protocol differences as well as evolving standards of care in ex-U.S. markets. We believe that positive or statistically significant benefits in either or both of these sub-groups would likely be sufficient to support ongoing marketing approval. Valuation In terms of our valuation approach, we have deviated from our usual “mature” company comparable valuation framework, which utilizes peer group multiples on out-year financials at the appropriate stage of growth when a company’s financial metrics most closely resemble those of peers, and instead opted for a sum-of-the-parts (SOTP) approach. Given the complexity of Antares’ partnership and revenue structures for individual products, we have elected to value Antares using this approach. Importantly, we would note that outside of XYOSTED (prior to FDA approval), our SOTP analysis does not probability-adjust each of the individual assets for the likelihood of approval. We believe that the ultimate approval likelihood for each individual product is certainly higher than typical phase III assets, and if pressed for a number we would suggest 85-90%. Thus, probability adjustments will have little impact on the SOTP framework. In light of the FDA approval, we have increased the probability of approval for XYOSTED to 100%, from a previous conservative 50% rate in our valuation framework, which was derived from last year’s setback in the XYOSTED regulatory process and the possibility of lengthy delays or additional data generation required to obtain approval. Our SOTP framework is derived based on detailed cash flow assumptions for each of Antares’ key individual assets, including Otrexup and XYOSTED, on the proprietary branded side and epinephrine, teriparatide, exenatide, and the recently added sub-q Makena on the partnered generic front. Our estimates for the individual product revenue streams, associated EBITDA and NOPAT levels, and ultimate net present value (NPV) with respect to each individual pipeline asset are detailed on page 4 of this report, where the first Exhibit summarizes our SOTP framework based upon the individual product NPVs. Each of these assets will follow very different individual pathways. Essentially, the path of each asset is completely independent in terms of the likelihood of approval, commercialization dynamics, timing, and duration. We thus believe the individual asset valuation or embedded option approach best captures the uniqueness of each pipeline or commercial product opportunity. In terms of terminal period multiples, we believe we have been relatively conservative in our assumptions on each of Antares’ key partnered programs, given the long horizon of our forecast period and the substantial uncertainties surrounding out-year market dynamics, both in terms of the potential for additional generic entrants as well as decreases in utilization of the existing molecules owing to therapeutic advances or next-gen branded drugs. Our NPV calculations, which utilize a WACC of 12.5%, incorporate terminal multiple assumptions of 6.0-8.0x NOPAT vs. the 9.0-11.0x adjusted EPS multiples of profitable pharma majors and emerging growth BioPharma companies. Given the possibility, although low likelihood, of generic competition on either Otrexup or XYOSTED by the end of the forecast period, or 2028, we believe we are being even more conservative in our terminal value assumption by utilizing end-period NOPAT multiples of 3.0-7.0x, compared again with the current peer group range of 9.0-11.0x. |
return to message board, top of board |