Repatha Was Almost Derailed By Formulation Bridging Concerns

FDA's worries about sufficiency of long-term safety data were exacerbated by 'weak' bridging data between two formulations; agency decided against refuse-to-file action but directed Amgen to immediately begin new PK/PD study.

The one-month gap between FDA approval of Sanofi and Regeneron Pharmaceuticals Inc.'s Praluent and Amgen Inc.'s Repatha could have been far longer had the agency not shown considerable flexibility in reviewing Amgen's new clinical pharmacology data during the PCSK9 inhibitor's review.

FDA's pre-submission concerns about the sufficiency of the safety database and duration of exposure to support licensing of Repatha (evolocumab, also known as AMG 145) were exacerbated by "weak" data in the BLA bridging between two formulations of the LDL-cholesterol-lowering drug, agency review documents show.

FDA's decision to file Repatha's BLA initially and its speedy review of new clinical pharmacology data may reflect an appreciation for the competitive implications in the PCSK9s' race to market.

So dismayed was FDA that it came close to refusing to file Amgen's BLA. The agency required Amgen to conduct a new bridging pharmacokinetics/pharmacodynamics comparability study during the review and cautioned that submission of the study results potentially could delay FDA's action date for the drug.

Praluent (alirocumab) beat Repatha to the market by one month in the US solely because Sanofi and Regeneron bought and redeemed a priority review voucher for their PCSK9 inhibitor. Had Amgen been on the receiving end of either a refuse-to-file letter or a three-month user fee date extension for a major amendment, it would have put Repatha's launch much further behind that of Praluent and likely caused considerable angst for Amgen.

FDA's decision to file the evolocumab BLA and its speedy review of data from the newly conducted clinical pharmacology study may reflect an appreciation for the competitive implications in the PCSK9 inhibitors' race to market.

Heavy Reliance On Phase II Safety Data

It is not uncommon for biologic product sponsors to use a formulation made by one process for some clinical studies, and a different formulation for later trials or commercial marketing. In the case of Amgen, however, its failure to bridge between the formulations – to FDA's satisfaction – came close to undermining the application.

Amgen used an earlier formulation (referred to in review documents as Process 1) in Phase I and II studies, and the to-be-marketed formulation (Process 2) in most of the Phase III studies. However, the BLA's long-term safety database relied heavily on the Phase II studies. In particular, the DESCARTES trial – the only placebo-controlled, long-term study in the development program – used the Process 1 formulation.

DESCARTES began as a Phase II trial. During an end-of-Phase II meeting on July 20, 2012, Amgen informed FDA that it intended to reclassify DESCARTES as a Phase III study. "Amgen confirmed that the Phase III trials will be performed with the formulation of AMG 145 intended for the market," FDA's meeting minutes state.

FDA's view on the sufficiency of the safety database to support approval changed as it gained a better understanding of the nature of the exposure data.

In written materials prepared for that same meeting, Amgen included exposure estimates at the time of BLA submission based upon the number of patients treated for at least three, nine, 12, 18 and 24 months.

These estimates are redacted from FDA's publicly available meeting minutes, but the agency's preliminary written response states, "The anticipated safety database should be sufficient to support approval."

FDA Wanted More Long-Term Data

The agency's view changed, however, as it gained a better understanding of the nature of the exposure data.

In written questions and responses ahead of an April 10, 2014, pre-BLA meeting, FDA raised concerns about the size of Amgen's long-term safety database.

"As we have previously discussed, we are most concerned about the potential for long-term adverse effects of either PCSK9 inhibition or very low LDL-C," FDA said. "We note that your current description of your long-term safety database is substantially different than the safety database you described at the EOPII meeting."

Amgen's pre-BLA meeting package estimated that 1,045 patients would be exposed for at least 12 months, 630 for at least 18 months and "only" 160 for at least two years, FDA said, requesting that Amgen "please provide us with an estimate of when you will have accrued the long-term safety experience that you described" at the EOPII meeting.

The pre-BLA meeting minutes reflect discussion between Amgen and FDA about "the discrepancy from the exposure projected" during the July 2012 meeting, details of which are redacted. However, FDA remained dissatisfied with Amgen's explanation.

"The agency stated that we have serious concerns whether this safety database, [REDACTED], would constitute a complete application for a first-in-class product intended for chronic administration with a potential for use in a large patient population, if marketed," the minutes state.

In addition, the proposed data cut-off for two long-term, open-label studies "seemed unnecessarily distant from the projected date of submission," FDA said.

"The agency voiced concern that approximately half of the safety data would be planned for submission in the 120-day safety update. The spirit of PDUFA V legislation is that a complete application will be submitted, including all information required for a regulatory decision in the initial submission," the minutes state.

FDA said its concern "was not with the cut-off date per se but with the amount of safety data that will be submitted four months into the review, as it would be very challenging to integrate all of this volume of new information into the safety review with the PDUFA timeline."

Several days after the pre-BLA meeting, Amgen sent FDA a revised proposal for its safety database using an April 1, 2014 data cut-off. The agency said it did not anticipate this proposed cut-off date would preclude a BLA filing, although the sufficiency of the safety database to support the proposed indications would be a review issue.

A 'Very, Very Weak' Bridge

The agency did raise concerns, however, about Amgen's reliance on safety data from studies using the Process 1 formulation and its ability to bridge between formulations.

Observing that a proportion of subjects in the safety database with at least 361 days of exposure would come from the Phase II and open-label extension studies, FDA said: "We note that you administered evolocumab differently in Phase II (total volume per administration drawn from six sterile vials) with a formulation (70 mg/ml) that you do not intend to market and that you did not use in Phase III. Please explain how you plan to bridge your Phase II and Phase III programs for the evaluation of clinical safety."

In response, Amgen said it planned to include in its BLA a comprehensive analytic comparability assessment; PK/PD dataset and analyses from 23 clinical trials; and individual and integrated clinical efficacy and safety data from 23 trials.

"These data will demonstrate similarities in drug substance, a consistent PK/PD profile, and consistency across all key safety and efficacy parameters across the studies,” Amgen said, according to review documents.

FDA told Amgen it was "extremely close to refusing to file" the application.

Yet, FDA considered the bridging data inadequate.

FDA convened a teleconference with Amgen to discuss the bridging issue on Oct. 24, one day before the BLA's filing date.

"The relevant disciplines have reviewed what you have submitted with regard to a bridge between Process 1 and Process 2. In short, our assessment during the filing review is that your bridge is very, very weak," FDA said, according to the partially redacted minutes from the teleconference.

"In addition, as we noted around the time of the pre-BLA meeting, you have very little long-term data for your to-be-marketed product; it appears, for example, that only 16 patients treated with evolocumab have completed year 1 – the controlled, albeit open-label, period – of your Phase III extension study."

FDA said this would be a "substantial review issue" and "underscores the reliance our review will have on a drug substance process that will not be marketed, which is a concern of its own."

"We strongly suggest that you begin designing a clinical PK/PD study that could be used to bridge Process 1 and Process 2 if needed," FDA told Amgen.

"This would need to performed in an appropriate population, with an appropriate duration, and using a dose expected to be on the steep portion of the dose-response curve with respect to a relevant pharmacodynamic biomarker or biomarkers so that you would have assay sensitivity for formulation-related differences in PK/PD," FDA said.

"If, in fact, these data are needed and you are able to submit the data during this review cycle, it is certainly possible that this would be considered a major amendment to the application," the agency continued.

Although not expressly stated in the minutes, a major amendment often will lead to a three-month extension in a product's user fee date.

FDA officials also made clear that unless the bridging concerns were clearly resolved, Amgen could forget about receiving a first-cycle advisory committee review.

"If we determine that you have not adequately bridged your two formulations with data that have been submitted to the BLA, leaving us uncertain to what extent your Phase II efficacy and safety data reflects your to-be-marketed product, we would not anticipate taking this application to an advisory committee for discussion," FDA said.

"If the majority of your long-term data derive from a drug formulation that will not be marketed, we must be certain that these data are relevant to the marketed product."

Oversight Board Weighs In

The agency said it was "extremely close to refusing to file this application, and these issues have been discussed with senior management across relevant offices reviewing your application."

"Ultimately, we have decided to file the application given that you have at least provided some data to review, which you believe should bridge your Phase II and Phase III programs," the agency said. "Our filing of the application, however, should not be reflected as our acceptance that these data are adequate; it simply indicates that we are willing to subject the application to further review."

During that same teleconference, FDA informed Amgen that evolocumab would receive a standard review, rather than the priority review requested (Also see "PCSK9 Sponsors Looked For Regulatory Advantages In Race To Market" - Pink Sheet, 21 Dec, 2015.).

Clinical pharmacology review documents note that FDA's concerns about the weak bridging data and the inadequacy of Amgen's bridging strategy were presented to the Office of Clinical Pharmacology's Biologics Oversight Board (BOB) on Oct. 30, 2014 (see box).

At A Glance: The Biologics Oversight Board

• Created in January 2011
• Comprises representatives from the Office of Clinical Pharmacology and other entities within CDER
• Meets once a week
• Develops policies and standards for biocomparability and biosimilars that provide consistency for the clinical pharmacology review of proposed biosimilar products and biocomparability related to manufacturing changes of innovator products and assure high quality scientific recommendations
• Provides scientific and regulatory input on specific proposed biosimilar product reviews at all stages of development and resolves differences in the interpretation of scientific and policy related issues within OCP and between OCP and other FDA units
• Maintains a growing knowledge base and repository on biocomparability and biosimilarity within OCP

"While there was discussion within BOB whether a PK study would at all be needed, or, if population PK assessment would suffice, in the end BOB did not object to the requirement for a bridging PK study," clinical pharmacology reviewer Suryanarayana Sista said in a Jan. 15 memo.

Amgen submitted a protocol synopsis for the PK/PD bridging study on Nov. 5, and the agency provided feedback on the study design in its Nov. 7 filing communication (Also see "Praluent Vs. Repatha: Interactive Timeline Of Regulatory Milestones" - Pink Sheet, 21 Dec, 2015.).

The 350-subject, parallel design study evaluated PK equivalence, safety, tolerability, immunogenicity and changes in LDL-C and PCSK9 following a single dose of evolocumab manufactured by Process 1 and Process 2.

FDA's clinical pharmacology review of the bridging study found that the mean serum unbound evolocumab concentration-time profiles following subcutaneous administration from either syringe (Process 1 material) or auto-injector/pen (Process 2 material) were similar and that statistical evaluation of the PK data indicated that the two formulations were comparable. In addition, the mean reduction in LDL-C between the two groups was nearly identical.

In response to questions from "The Pink Sheet" about the formulation bridging issues and the challenges of conducting a new PK/PD study during the review, Amgen said that switching from clinical to commercial manufacturing of drug substance is standard during drug development.

"Studies using clinical and/or commercial drug substances were in the evolocumab/Repatha BLA," Amgen said. "During the filing review, FDA recommended we perform a head-to-head study. In response, Amgen provided PK, PD, and safety data to support the bridge from the existing data and conducted a head-to-head study comparing the bioequivalence of both drug substances."