AstraZeneca's Efzimfotase Alfa Misses the Primary Endpoint in Adult HPP Trial, but Management Holds to $3 Billion to $5 Billion Peak Sales on Injection Burden and Tolerability
Meet the Management event at ASBMR 2026, Boston, October 12, 2026
AstraZeneca's Alexion unit presented Phase III HICKORY data for efzimfotase alfa, its second-generation enzyme replacement therapy for hypophosphatasia (HPP), and the headline is a miss. In 124 treatment-naive adolescents and adults, the drug did not separate statistically from placebo on the 6-minute walk test at week 25. Management used the call to argue that the miss reflects the trial's design and a handful of outlier patients rather than the biology. It also reaffirmed the $3 billion to $5 billion peak-year sales range first given in May 2024. The argument for that range now rests less on adult efficacy and more on a convenience and tolerability case against Strensiq, which management supports with a striking discontinuation figure.
The primary endpoint failed, and the adult-onset group is the problem
HICKORY randomized 84 patients to efzimfotase alfa and 40 to placebo in a 2:1 ratio. The treated arm walked numerically further at 25 weeks, but the difference was not statistically significant. Of the key secondary endpoints, lower extremity function, pain severity and FACIT-Fatigue all showed clinically meaningful improvement, but only fatigue reached statistical significance. Pain interference and treatment satisfaction also favored the drug with statistical significance.
The prespecified subgroup split is the core of the debate. In the 60 patients with pediatric-onset disease, the 6-minute walk test improvement was 35.6 meters, which is above the minimal clinically important difference. Sit-to-stand and pain severity were also nominally significant. In the 64 adult-onset patients, only fatigue was nominally significant, with a 7-point improvement that Gianluca Pirozzi, Alexion's head of development, regulatory and safety, called "highly clinically meaningful." The company is explicit that all of these subgroup p-values are nominal, because the primary endpoint failed.
Management offered several explanations for the adult-onset result. The placebo arm in that group improved by 28 meters, against 10 meters in pediatric-onset adults. Pirozzi attributed this to "a couple of patients" on placebo with improvements "in the range of 200" meters, which he described as "largely higher and uncommon in the HPP patient population." He added that the adult-onset group was older, with a mean age of 54, carried more comorbidities, and included patients over 64 using assistive devices. The placebo arm also started about 50 meters lower at baseline. Pirozzi acknowledged the limits of the 6-minute walk test, which is the only endpoint regulators have accepted for this setting, noting its variability and the potential for patients to learn the test. He conceded: "I cannot deny that the 6-minute walk test obviously did not reach significance." These explanations are plausible in a trial this small, but they are post-hoc, and they apply to an ultra-rare disease where two patients can move the result.
The supporting evidence is suggestive, not conclusive
Pirozzi outlined four lines of support for a treatment effect on walking distance. The first is a post-hoc responder analysis, in which 45% of treated patients reached the minimal clinically important difference against 30% on placebo, an odds ratio of 2.5 with a nominal p-value of 0.036. The second is the open-label extension, where the treated group's gain grew from about 35 meters at week 25 to 61.9 meters at week 49, and placebo patients who switched gained 43 meters. The third is the prespecified pediatric-onset subgroup, which maps to Strensiq's current label. The fourth is that every key secondary endpoint favored efzimfotase alfa.
Investors should weigh these accordingly. Open-label gains without a control arm are the weakest evidence in the package, and the responder analysis is post-hoc. The most credible case is the pediatric-onset subgroup, which was stratified at randomization, and the consistent direction of the secondary endpoints.
Rajan Sharma of Goldman Sachs raised a further concern: an earlier open-label Strensiq study in adults also showed no significant 6-minute walk benefit, which could point to a biological limit of enzyme replacement in this population. Pirozzi dismissed the earlier study as small, with 19 patients of whom 13 were adults, and said patients were underdosed. That answer addresses the design of the earlier study but not the underlying question, and it does not settle it.
Regulatory path: filed, under review, and no commitments
The company has filed in major geographies and declined to speculate on how regulators will read the data. In response to Michael Leuchten of Jefferies, Pirozzi confirmed that the FDA priority review was granted on the totality of the dataset, including both pediatric-onset and adult-onset patients, and covers patients over 2 years of age. That is a mild positive for the breadth of the filing, but it is not a signal on approvability, since priority review reflects the unmet need rather than the strength of the efficacy package.
The most likely downside is a label confined to pediatric-onset disease, which mirrors where Strensiq is approved. Asked whether an adult-onset rejection would prompt another trial, Marc Dunoyer, CEO of Alexion and AstraZeneca's chief strategy officer, would not commit. He said only that "in rare disease, it is very common to continue the development work or post approval," and that more work on the molecule should be expected. Dr. Raja Padidela, a pediatric endocrinologist at Royal Manchester Children's Hospital, said that clinicians in the U.K. do not prescribe expensive rare-disease drugs off-label, so a narrower label would translate directly into narrower use.
The commercial case now leans on Strensiq's discontinuation problem
The most useful new information for modeling came from Padidela, who said that about 40% of his patients stop Strensiq because of injection site reactions. In his words, "that is a huge number," and these are patients for whom the drug is clinically beneficial but the reactions become "so overwhelming" that they decide to stop. He said he would "definitely switch" patients to efzimfotase alfa given efficacy that appears similar and much better tolerability.
The Phase III data support the tolerability case. Patients had a median of 98.8% injection-site-reaction-free days, equivalent to 361 days a year. The annualized reaction rate was 5 times lower than with Strensiq. Of reactions that did occur, 80% resolved before the next injection, and 4 times fewer persisted beyond 28 days. Dosing is once every 2 weeks, roughly 12 times fewer injections per year than Strensiq's 3 to 6 weekly doses, with a smaller injection volume. In the pediatric Phase III studies, MULBERRY showed a median RGI-C score of +1.67 versus 0 on placebo (p=0.0003), and the CHESTNUT switch study showed comparable safety and maintained benefit.
Dunoyer said conversion from Strensiq should be "very rapid," and pointed to the 40% abandonment rate as the guide to speed. He also said that eliminating that abandonment would "have a very substantial impact on the average duration of treatment," and that investors could use the 40% figure to model longer duration of therapy. The company did not provide a conversion curve in response to Steve Scala of Cowen.
Peak sales range reaffirmed, with the label as the swing factor
Management put the addressable population at about 14,000 patients across the top 8 markets. Of these, 20% are pediatric, 60% are adults with pediatric-onset disease and 20% are adult-onset, and about 43% present with skeletal manifestations. Dunoyer confirmed the $3 billion to $5 billion peak-year range and said a more restrictive label would put sales "towards the $3 billion," while a broader label would bring them closer to $5 billion. He also framed the range as consistent with the previously stated goal of 4 to 6 times the 2022 Strensiq patient base. The drivers he cited were easier adoption, better retention, global expansion in markets where Strensiq is not available, and the possibility of adult-onset patients, though he said the sizing does not depend on the latter.
The range remains wide, and the low end of it implicitly assumes no adult-onset label. The reaffirmation does show that the company's commercial thesis does not depend on adult-onset efficacy. Dunoyer also noted that Strensiq has been growing well, helped by reimbursement and approval gains around the world, a playbook he intends to reuse for the new asset.
Pediatric bone data and open questions on variability
Christopher Uhde of SEB asked about the non-Gaussian distribution of RGI-C and Rickets Severity Score responses, which is based on only 29 pediatric patients. Pirozzi said every patient improved and that the spread reflects baseline severity, while Padidela noted that patients with more severe skeletal disease take longer to heal. Pirozzi said the treatment effect is comparable to Strensiq at a lower dose and frequency, though no head-to-head comparison exists, and he left open the possibility of testing additional dosing regimens later. Management declined to comment on the positioning of Ultomiris in IgA nephropathy, deferring to a later interaction.