FDA’s new metabolic diseases committee gives hesitant nod to ultra-rare disease drug

Background: Niemann-Pick Type C (NPC)

• Niemann-Pick disease Type C (NPC) is an autosomal recessive lysosomal storage disorder that typically results in progressive neurologic symptoms and organ dysfunction. NPC is caused by variants in the genes encoding the NPC1 or NPC2 proteins, which are involved in cholesterol metabolism. Without functional versions of these proteins, lipids accumulate in tissues, according to the National Institute of Neurological Disorders and Stroke (NINDS).
• NPC is an ultra-rare disease, with prevalence estimates ranging from 0.35 to 2.2 per 100,000 births across countries. Although it is traditionally considered a childhood-onset disease, some patients display NPC later in life.
• Symptoms and outcomes vary considerably. NINDS explains that the associated neurological complications can include extensive brain damage that can cause movement difficulty and progressive loss of vision and hearing, among other symptoms. Generally, diagnosis earlier in life as well as having certain genotypes forecasts more severe disease.
• The heterogeneity of the disease, and its slow progression, have challenged therapeutic development. The FDA has participated in several discussions on how to overcome these obstacles. At a January 2022 workshop on NPC endpoint development, the FDA’s Director of the Center for Drug Evaluation and Research (CDER), PATRIZIA CAVAZZONI, explained that endpoints used in trials for ultra-rare conditions such as NPC may not reflect disease presentation or clinically meaningful improvements for all patients. In addition, “there are only a limited number of patients with NPC who are available to participate in clinical trials and this may restrict the study design,” Cavazzoni explained.
• The NIH-developed NPC Clinical Severity Scale (NPCCSS) used in most trials has some issues, according to participants in an August 2022 webinar that followed the January workshop. The scale measures clinician-reported outcomes in 17 domains, though the five most important domains are used in the abbreviated 5-domain NPC Clinical Severity Scale (5DNPCCSS). These domains are Ambulation, Swallow, Cognition, Speech, and Fine Motor Skills. At that webinar, JACKIE KARP, lead physician at the FDA’s Division of Rare Diseases and Medical Genetics (DRDMG), outlined validity concerns with the scale’s cognition domain for short trials because the “rate of decline does not align with trial durations.” [ Read full AgencyIQ analysis here.]
• As a result of these challenges, there are currently no approved therapies for NPC in the U.S. However, miglustat (brand name: Zavesca), which was approved by the EMA to treat NPC in 2002, has been reported to be used off-label in the U.S. to help preserve swallowing function in NPC.

Regulatory context: Arimoclomol aims to be the first NPC therapy to receive FDA’s nod

• Arimoclomol is an oral small molecule that is intended to amplify the body’s natural heat shock response (HSR); the drug crosses the blood-brain barrier. This is thought to prevent protein misfolding and facilitate lysosomal function, which in turn prevents cell death associated with NPC.
• In June 2021, Orphazyme received a Complete Response Letter (CRL) in response to its first New Drug Application (NDA) filing for arimoclomol. Side note: The product was acquired by KemPharm, Inc., in 2022, which changed its name to Zevra Therapeutics in 2023.
• That application relied on data from a single prospective randomized, double-blind, placebo-controlled, phase 2/3 study (NCT02612129). Patients were randomized 2:1 to receive arimoclomol or placebo as an add-on therapy to their current treatment. According to the results published in the Journal of Inherited Metabolic Disease, the 12-month study enrolled 50 patients ranging from age 2 to 18. “Given the ultra-rare condition, and expected rate of progression, sample size and trial duration were primarily informed by feasibility and not by a formal sample size calculation,” the paper states. Forty-two patients completed the study.
• The study reported modest but positive results in its primary endpoint, which was change in the 5DNPCCSS score from baseline to 12 months. The study found a treatment difference of −1.40 points (95% confidence interval [CI], −2.76 to −0.03; P = 0.046) favoring the arimoclomol group. In terms of safety, the researchers judged that, “The majority of TEAEs [treatment-emergent adverse events] were assessed by the investigators to be related to NPC rather than the investigational product.”
• According to a firm presentation, the CRL highlighted three information gaps: (1) evidence needed to support use of the NPCCSS as the primary instrument in measuring NPC disease progression, (2) additional analysis related to how missing data are handled for statistical analysis, and 3) additional support and data related to confirmatory evidence of efficacy. According to the presentation, “FDA did not request additional efficacy data in the CRL.”
• The presentation also summarized the discussion at a Type A End-of-Review Meeting that took place in October 2021. The firm writes that at the meeting, the agency agreed to three actions: (1) reanalysis of the 5DNPCCSS with the removal of the cognition domain, (2) rescoring and a reassessment of the swallowing domain including further validation, and (3) further discussions regarding NPCCSS.
• Zevra resubmitted the application in December 2023. While the application initially had a target PDUFA date of June 21, 2024, an updated Zevra press release explained that the subsequent review and associated information requests necessitated an extension to September 21, 2024. In the same press release, Zevra stated the agency’s intention to convene an advisory committee meeting regarding the application.

On August 2, FDA’s Genetic Metabolic Diseases Advisory Committee (GeMDAC) heard the case for arimoclomol’s second chance

• An upfront note: This was the first meeting of the newly minted GeMDAC. The committee was announced in December 2023, with a charge to provide “independent, knowledgeable advice and recommendations on technical, scientific and policy issues around medical products for genetic metabolic diseases.” [ Read AgencyIQ’s analysis of the announcement here.]
• The committee has slowly staffed up, with the roster indicating that six of its 10 spots have been filled. The August 2 meeting featured a significant number of temporary members. Temporary member ROBERT ALEXANDER, a psychiatry professor from the University of Arizona, served as the Acting Chairperson of GeMDAC. Alexander is no stranger to the agency’s advisory committee system, having previously served as the chair of the Peripheral and Central Nervous System Drugs Advisory Committee (PCNS) in an acting capacity.

The post hoc reanalysis

• As suggested by FDA, Zevra completed a post hoc reanalysis of the clinical trial results, with the primary endpoint changed from the 5DNPCCSS to a rescored 4-domain Niemann-Pick disease type C Clinical Severity Scale (R4DNPCCSS). This meant that the cognition domain was omitted from the score. NAOMI KNOBLE, Associate Director of FDA’s Division of Clinical Outcome Assessment, explained that published natural history study data showed that, “…it would take more than one year to observe changes in the NPCCSS Cognition domain score, making it unsuitable to show stabilization and/or changes as part of the primary efficacy endpoint. Ratings were also dependent on the patient’s environment, such as whether they were receiving educational services, which is not a clear indicator of cognitive functioning.” In addition, the agency raised concerns that the original swallow domain did not accurately reflect linear progression of the disease. For example, the original rubric did not differentiate between patients who require feeding tubes sometimes and all the time. The swallowing domain was therefore rescored based on consultation with experts.
• During Zevra’s presentation on the clinical background of NPC, Mayo Clinic child neurologist MARC PATTERSON further contextualized the endpoint. He pointed out that there are no established surrogate endpoints or measurable biomarkers for NPC. “It’s important to underscore that the NPC clinical severity scale is the only validated and appropriate tool to specifically measure Niemann-Pick disease type C severity and progression across multiple domains,” he said. He also provided perspective on its real-world use, saying, “We must bear in mind that we are dealing with neurologically impaired children across a broad range of ages. If you do not actually spend time sitting in an examination room with a child, particularly a neurologically impaired child, it may seem very reasonable to use a large number of scales to try to assess different aspects of the disease, but it’s important to recognize that is not a realistic goal for most children with NPC.”
• In addition, the reanalysis addressed missing data called out in the CRL. The initial analyses excluded data from one patient who died and other patients who took an “early escape” route due to rapid disease progression. “It is notable that the excluded data indicated disease worsening,” stated the agency. Various imputation methods were used to account for these data in the post hoc analysis.
• Taken together, these changes resulted in modest treatment differences favoring the arimoclomol arm, with effect sizes ranging from -1.5 to -1.2 points depending on the specific methods used. Importantly, FDA’s presentation states that “These analyses provide smaller treatment difference estimates and wider confidence intervals compared to those from the prespecified analysis.”
• Several committee members expressed reservations regarding post hoc analyses in general. KENNETH FISCHBECK, Distinguished Investigator Emeritus at NINDS commented, “My understanding from statisticians is that post hoc analysis [is] not statistically valid,” but also acknowledged that this was done at the request of the FDA. Zevra emphasized that the prespecified analysis reached statistical significance and favored arimoclomol, and that the new analysis did not change the direction of the results.

The confirmatory evidence

• Quick background: When seeking approval from the FDA, companies must demonstrate “substantial evidence,” as defined under the Federal Food, Drug, and Cosmetics (FD&C) Act, that their product is safe and effective. Traditionally, FDA has interpreted the need for “well-controlled investigations” to mean at least two clinical trials for applications for new drugs or supplemental indications. Over time (with the help of litigation and the passage of the 1997 FDA Modernization Act (FDAMA)), the definition has been updated so that one “adequate and well-controlled investigation” can be sufficient to demonstrate safety and effectiveness if it is supported by “confirmatory evidence.”
• FDA released draft guidance in September 2023 describing data sources that could serve as confirmatory evidence and the situations in which they would be appropriate. Importantly, the agency stated that the volume of confirmatory evidence needed is inversely proportional to the strength of the single adequate and well-controlled study. In addition, the single trial and confirmatory evidence are intended to be “considered together” to assess effectiveness. [ Read AgencyIQ’s complete analysis of the guidance here.].
• Over time, the FDA’s acceptance of single pivotal trials has become more common, especially given the increased submission of applications for products intended to treat, cure or prevent rare diseases or life-threatening conditions for which there is an unmet clinical need. In a recent analysis, AgencyIQ found that the majority of new molecular entities approved each year since 2020 have relied on a single pivotal trial.
• At the meeting, GeMDAC considered Zevra’s bolstered confirmatory evidence package, which consisted of results from the Open-Label Extension (OLE) phase of the original trial, plus additional nonclinical studies.
• Forty-one of the 50 total participants in the initial double-blind trial chose to continue to the OLE phase. Twenty-nine of those patients ultimately completed 48 months of treatment in the extension. Withdrawal reasons included caregiver preference, adverse events, death, and physician decision. The data alone “appears to show relative slowing/stability of disease progression with arimoclomol,” though a subset of patients declined rapidly on treatment. No single common characteristic was identified in the rapidly advancing subset, though factors such as early symptom onset and double-null mutations were explored as potential drivers.
• Given the lack of a control group in the OLE phase of the study, the data were compared to NIH natural history study data as an external control. Participants from the two groups were matched based on variables including baseline age, sex, and miglustat use. FDA reported that these comparisons “numerically favored arimoclomol” without reaching statistical significance.
• Multiple committee members, including Chair Alexander, were concerned that the OLE and natural history study represented an “apples to oranges” assessment given the updates to the endpoint made in Zevra’s study. Zevra clarified that it was able to convert the NIH data to the four-domain scale to enable comparison of scores. However, the rescoring of the swallowing domain could not be performed using the NIH data. As a result, Zevra confirmed that the same four-domain methodology was used on both groups.
• The CRL described the original nonclinical data as “weak and contradictory,” and Zevra conducted additional work in this area for the resubmission. The new nonclinical studies consisted of in vitro mechanistic data and in vivo studies in two mouse models.
• Arimoclomol is thought to work by amplifying the body’s natural HSR, and Zevra hypothesizes that this is done via upregulation of genes within the Coordinated Lysosomal Expression and Regulation (CLEAR) network. Specifically, arimoclomol is hypothesized to activate specific transcription factors that bind to CLEAR gene promoters during cellular stress response. The new cell culture studies in the resubmission demonstrated that arimoclomol treatment leads to small increases in nuclear localization of these transcription factors, as well as increased expression of CLEAR genes. However, these results were observed when the drug was given at 30 to 60 times the human clinical dose. Additional pharmacodynamic biomarker studies were limited due to their non-specificity to NPC and arimoclomol.
• The new animal studies involved mouse models of both gene interruption and point mutation, which are akin to infantile and delayed onset forms of NPC, respectively. At a high level, FDA described the results as “small, variable, and lacked a dose-relationship and/or failed to repeat when re-tested.” The studies assessed survival, motor function, and biochemical endpoints. Of note, the arimoclomol doses were administered through drinking water. FDA’s Acting Lead Pharmacologist SHAWNA WEIS explained that “…the applicant did not measure water consumption or [pharmacokinetics] to evaluate either the dose delivered or the exposures achieved, which make these data very difficult to interpret.” The GeMDAC members picked up on additional potential issues related to dosing arimoclomol in drinking water, as NPC itself is associated with swallowing difficulties that could further influence the dose.

The two discussion questions and the vote

Discussion Question: Discuss your assessment of the efficacy results of trial CT-ORZY-NPC-002 (NPC002). In your discussion, please comment on: (a) The 5-domain Niemann-Pick disease type C Clinical Severity Scale (5DNPCCSS) and the rescored 4-domain Niemann-Pick disease type C Clinical Severity Scale (R4DNPCCSS), and (b) Your assessment of whether the trial results demonstrate a treatment effect of arimoclomol on the treatment of Niemann-Pick disease type C (NPC).

• In general, the committee felt that the severity scales represented the best available option for an endpoint to assess efficacy. That said, committee members envisioned improvements to the scale in terms of increasing its utility, either by incorporating additional aspects of the full 17-domain scale or by increasing the granularity of the ratings within each domain. CHERYL COON, Vice President of the Clinical Outcome Assessment Program at the Critical Path Institute, summarized, “…We should not let perfect get in the way of good enough. And I think in this case, this is good enough. We can certainly get information out of this instrument as it’s administered.”
• The committee agreed that the trial results showed a small, incremental treatment effect, but members had varying opinions on the meaning and reliability of this effect. For some, the effect represented stabilization, which can be meaningful for progressive diseases. Consumer Representative SARAH CHAMBERLIN expanded on this, saying that “…that one-point difference to them is a bonus on top of what they’re really looking for—a treatment to keep their family members from getting worse than what they are.” On the other hand, concerns stemmed from the sample size of the study, which was necessarily small in the ultra-rare disease context. Notably, changes in a small number of patients contributed to large changes in the overall treatment effect.

Discussion Question: Discuss your assessment of other data (specifically the additional clinical and nonclinical data) with respect to support for the effectiveness of arimoclomol.

• “It’s all approximately the same,” said WENDY CHUNG, chief of pediatrics at Boston Children’s Hospital, who felt reassured by the similarity in the additional data regarding a small effect size and lack of new safety concerns. “For me, it was at least consistent internally,” she said. This sentiment was echoed by others. SUSAN ELLENBERG, professor emerita of biostatistics, medical ethics and health policy at the University of Pennsylvania, explained that while the new data did not help much, they also did not detract from the clinical trial results.
• Thomas Jefferson University Professor WALTER KRAFT pointed out that the lack of safety signals could have positive or negative ramifications. He spoke to the drug’s hypothesized mechanism of action, noting that a small molecule working through transcription factors would likely have off-target effects. “So, the fact that we don’t have side effects actually can be a little bit worrisome,” he said.
• Other committee members poked holes in the rigor and value of the additional clinical and nonclinical data. JEAN BAPTISTE LE PICHON from the University of Missouri-Kansas City pointed out that the NIH registry used as an external control included patients on a variety of treatment regimens, saying, “I don’t think it’s a comparison that should have happened.”
• Several members were frustrated by the lack of translatability of the mouse studies to humans. University of Michigan neuropathologist ANDREW LIEBERMAN was concerned by the need to use a much higher dose than that in humans to see effects in mice. Fischbeck felt that the data were limited because the treatment in mice was started prior to disease manifestation, which does not typically occur in humans. Chung found some utliity in these data, referencing family perspectives discussed during the public hearing regarding multiple siblings with NPC. “It was interesting through that lens to see the data on the mice in terms of what might be going on in the future,” she said.

VOTING QUESTION: Do the results of trial NPC-002 in concert with the other data (clinical and nonclinical in particular) support a conclusion that arimoclomol is effective in the treatment of patients with NPC? Provide a rationale for your vote. If you voted no, provide recommendations for additional data that may support a conclusion that arimoclomol is effective.

• The committee voted 11-5 in favor of arimoclomol’s effectiveness for NPC, with no abstentions. Prior to the vote, FDA’s Acting Director of the Division of Rare Diseases and Medical Genetics, CATHERINE PILGRIM-GRAYSON, clarified that the question is asking about arimoclomol “as it happened in the clinical trial,” meaning as an add-on treatment to the background or standard-of-care therapy
• The committee members who voted yes saw a small but beneficial treatment effect in the totality of the data. Le Pichon summarized, “It was not an enormous effect, but it was an effect that is clinically significant.” Duke Professor of Pediatrics PRIYA KISHNANI highlighted the importance of stabilization in the context of a progressive disease, and Ellenberg was particularly impressed by the improvement observed in a small group of patients, which she believes makes the treatment worthwhile to pursue.
• Other considerations contributing to “yes” votes included arimoclomol’s safety profile and the critical unmet need. Several members were assuaged by the safety profile, especially over the duration of the extension phase of the study. Retired Child Neurologist JOHN MINK explained the reasoning for his “very reluctant yes,” saying, “…the bulk of the data favored a slightly positive effect. I think the unmet need is very clear. I’m not sure that this meets that need, but again, on balance, I voted yes.”
• On the other hand, the “no” votes could not see past “problematic” aspects of the data. While most agreed that the clinical data suggested a modest effect, the confirmatory evidence was not compelling enough to support an overall conclusion. These members found noteworthy deficiencies in the data from the mouse models. Kraft assumed the regulator perspective, stating his view that the resubmission package “did not meet the evidentiary standards for approval.” He raised concerns about downstream effects, saying, “I do worry about approval of drugs for which there is not unequivocal evidence of efficacy that is not without harm, also in terms of diversion of resources and activity within the space.” Regarding next steps, the “no” votes pushed for more rigorous nonclinical studies.

Analysis

• FDA has more than six weeks to draw its final conclusion, with a Zevra press release indicating an action date of September 21, 2024, for the application. The agency is not required to align with the opinion of its advisors, and the mixed result of the vote makes FDA’s next steps less clear. FDA, the sponsor and the advisory committee agreed that NPC represents an area of high unmet need, but the first application’s CRL indicates that this alone is not enough to get arimoclomol across the finish line. While the additional confirmatory evidence was lackluster, it is notable that most advisors found the strength of the entire evidence base to be somewhat greater than the sum of its parts.
• If arimoclomol is ultimately approved, Zevra will receive a valuable voucher. The product has received Rare Pediatric Disease designation, meaning it is eligible for a priority review voucher (PRV) upon approval. PRVs, which can be redeemed for priority review of another product application, are a highly valuable commodity. Because the vouchers offer the potential for a company to bring their drug to market up to four months faster, they offer the potential to accelerate commercial sales, make use of more of a product’s useful patent life, and cement a lead for a product before a competing product can come to market. Individual vouchers regularly sell in excess of $100 million. [ See AgencyIQ’s Priority Review Voucher Tracker here.] The vouchers may rise further in value, as the program is set to expire at the end of September unless it is reauthorized by Congress.
• Back to the basic (science). A major takeaway from the meeting was that the agency and advisors struggled with issues stemming from an unclear mechanism of action and poorly designed nonclinical studies. While this reflects the brisk pace of development in the rare disease space, it may continue to throw a wrench in arimoclomol’s approvability. This sequence of events also underscores the importance of translational and regulatory science research to identify appropriate animal models and biomarkers.
• This meeting adds an interesting use case to FDA’s ongoing discussions related to the use of external controls. In late 2023, FDA finalized guidance offering extensive new recommendations on considerations for the development of drugs and biologics for rare diseases. That document expressed the agency’s willingness to accept external controls in cases where there are unmet medical needs, a well-documented and highly predictable disease course, and a large drug effect [ Read full AgencyIQ analysis here.]. The ultra-rare NPC certainly checks the unmet need box, though its course is notoriously heterogeneous. It remains to be seen how arimoclomol’s modest drug effect will factor into this calculus.

To contact the author of this item, please email Amanda Conti ( aconti@agencyiq.com).
To contact the editor of this item, please email Jason Wermers ( jwermers@agencyiq.com).

Key Documents and Dates

• August 2, 2024: Meeting of the Genetic Metabolic Diseases Advisory Committee (meeting materials and recording linked here)
• FDA Docket No.: FDA-2024-N-2930