FDA announces public hearing to discuss interim CVOT data disclosure – July 16, 2014

Executive Highlights

  • The FDA will hold a public hearing on the disclosure of interim results from ongoing cardiovascular outcomes trials (CVOTs) for type 2 diabetes drugs on August 11 at its White Oak campus in Silver Spring, MD.
  • A public docket open to receive comments can be found here and registration is open here until July 28 (on a first-come first-served basis).
  • The ripple effects of the FDA’s 2008 CV Guidance are being felt increasingly throughout the field, and we expect this meeting to prompt a broader discussion on the current regulatory hurdles in type 2 diabetes drug development.

On Tuesday, the FDA announced that it will hold an August 11th public hearing on the confidentiality (or lack thereof) of interim results from ongoing cardiovascular outcomes trials (CVOTs) for type 2 diabetes therapies. Registration for the meeting, which will be held at the agency’s White Oak campus, is open until July 28. We were very pleased to learn of this hearing – of course, interim data disclosure is not the only problem with the CVOT paradigm, but it is a big piece of the puzzle that has led to the postponement of drug approvals already and runs counter to accepted practices and protocols for clinical research. We are not optimistic that a straightforward solution can be found.

This public hearing comes about six years after the FDA passed its now-infamous 2008 CV Guidance (based on an infamous Advisory Committee meeting in July 2008). The guidance stipulates that a drug sponsor must rule out a certain level of cardiovascular safety risk (defined as a hazard ratio for CV events with an upper 95% confidence interval less than 1.8) before approval, either through pooled phase 2/3 clinical data or from interim analysis of an ongoing CVOT. In most cases, even if approval is granted on the basis of phase 2/3 data, a dedicated outcomes trial will be required post-approval. As a result, many pharmaceutical companies have started CVOTs pre-approval and have relied on (or plan to rely on) interim data to support approval. We think the stipulations around the requirements are a bit random and weren’t particularly well thought through; in any case, the requirements have contributed to incredible uncertainty and unpredictability in the field. While the upside might be seeing cardioprotection stem from one of the drugs or classes, we think there is a very good chance that the trials are all too short to show this and that the additional expense for making them longer is too much for most companies to pay given that they had not forecast expense of the multi-thousand person long-term trials in the first place.

In terms of the focus of the day, the disclosure of interim data from a CVOT is potentially problematic in a number of ways. Disclosure of interim data can alter the behavior of trial participants, investigators, and sponsors in a way that jeopardizes or entirely compromises the integrity of the full-length trial. As a result, internationally agreed upon standards on clinical trials recommend strongly against interim data disclosure in most cases. Two recent case studies demonstrate how problematic interim disclosure can be. In the case of J&J’s Invokana (canagliflozin), which was approved in early 2013 using interim data from the CANVAS CVOT, unblinding the trial essentially precluded J&J from testing for CV superiority (read our report on the drug’s FDA Advisory Committee meeting). Later that year, Sanofi withdrew its New Drug Application for its GLP-1 agonist Lyxumia (lixisenatide) to avoid interim data disclosure from its ELIXA CVOT, which pushed the product’s timeline back by around three years.

The goal of the FDA’s public hearing appears to be to initiate constructive discussion among regulators, researchers, health care providers, pharmaceutical industry representatives, health care organizations, the general public, and other stakeholders about the appropriate handling of interim analysis results of these ongoing CVOTs. The FDA hearing notification provides comprehensive background on the topic and lists five key questions (see below). We are looking forward to a constructive discussion with plenty of representation from a range of stakeholders. We also hope the FDA’s decision to hold this meeting is indicative of a willingness on the agency’s part to more carefully consider whether every element of the 2008 CV Guidance is really in patients’ best interests. We highly encourage anyone who can to attend to be present at the hearing, and to submit comments on this issue, as we see this as a most valuable opportunity to work with the FDA to make the regulatory side of diabetes drug development a smoother and more effective process.

  • The FDA announcement presents five questions intended to guide the conversation during the meeting, which are listed below. We believe the answer to the first question is a fairly clear yes, but that the latter four should stimulate richer discussion.
    • 1. When a trial to evaluate CV safety of a new treatment is ongoing at the time a drug is approved, do stakeholders agree that disclosure of detailed analyses (such as point estimates of hazard ratios and the associated confidence intervals) could undermine the integrity of an ongoing trial and jeopardize its continuation, potentially eliminating or substantially delaying the Agency's ability to obtain needed long-term safety information?
    • 2. What interim findings, if disclosed, would represent the greatest risk to trial integrity or jeopardize trial continuation?
    • 3. Can partial disclosure of interim findings at the time of approval, essentially disclosing only that the standard for approval has been met, offer protection of trial integrity and also provide health care practitioners with the essential scientific information needed to inform their use of the drug?
    • 4. If the detailed interim results were disclosed at the time of approval, and the ongoing study was discontinued at that time, would it be feasible to conduct a new large trial as a postmarketing requirement that would fulfill the original study objective?
    • 5. Are there other, alternative trial designs that would allow for disclosure of interim results on safety risks at the time of product approval while also allowing for further information to be obtained postmarket?
  • The FDA has opened a public docket for comments on the issue of interim CVOT data disclosure. Comments should be submitted by July 28 to be ensured consideration, but comments will be accepted after the meeting through October 10. For those who will be unable to attend the meeting in person, this docket represents an excellent way to voice opinions on the issue. The docket page also lists the number of comments received – we hope to see that number climb substantially during the next two weeks.
  • The issue of interim data disclosure from CVOTs has come up at scientific meetings, though not as frequently as we might have liked. In a stellar talk at EASD, cardiologist Dr. Sanjay Kaul (Cedars-Sinai Medical Center, Los Angeles, CA) discussed the problems involved with interim CVOT data disclosure in the context of J&J’s CANVAS trial. He noted that disclosure of interim analyses slows enrollment; reduces adherence in the experimental arm (for HR >1.2); encourages cross-ins in the control arm (HR <0.8); and minimizes retention. To use interim analysis results and maintain trial integrity, Dr. Kaul suggested that data access be restricted to an unblinded “firewalled group” formed with input from the regulator, DMC, and steering committee; discussions should occur in a “closed” session to preserve data integrity; and regulators should only disclose publicly if a hazard ratio of 1.8 has been ruled out (implying a point estimate of <1.26 without loss of equipoise).
  • For a time, it appeared that the FDA was unmoved by the growing number of negative ripple effects of its 2008 CV Guidance. While architects of the guidance and some others (including Dr. Kaul) have suggested that the guidance has not stifled innovation, pointing to the seemingly rich landscape of diabetes drugs in development, the long duration of the drug development process means that the ripple effects would not be felt until years after the implementation of the new policy. Indeed, in the past year alone, we have seen a number of examples of how the CV Guidance has posed major problems for diabetes drug development and approval.
    • Novo Nordisk’s ultra-long-acting basal insulin Tresiba (insulin degludec) received a Complete Response Letter (CRL) from the FDA due to CV safety concerns. Our disappointment with the FDA was heightened by the fact that the drug was approved without any major restrictions in Europe.
    • BMS exited diabetes in late 2013, ostensibly (in part) due to the challenges involved in diabetes drug development. Roche and Genentech have also (at least for now) departed the diabetes arena.
    • As mentioned above, interim data disclosure (or the risk of it) precluded J&J from testing for superiority in CANVAS, and caused Sanofi to withdraw its NDA for Lyxumia.
    • The delays with Tresiba and Lyxumia have not only delayed those two candidates, but have also pushed back an entire class of GLP-1 agonist/basal insulin combination therapies. Regulatory approval of Novo Nordisk’s IDegLira (insulin degludec/liraglutide) depends on Tresiba’s approval, while approval of Sanofi’s LixiLan depends on Lyxumia’s approval.
    • During its F4Q13 update, Takeda management shared that the company will likely not continue development of the once-weekly DPP-4 inhibitor trelagliptin in the US due to the high cost of phase 3 trials and outcomes studies.
    • GSK’s Avandia (rosiglitazone) was exonerated by the FDA following a CV safety scare and FDA-mandated REMS program that effectively wiped the drug out. Earlier this year, the FDA approved a label change that eliminated many elements of the REMS program, but the drug is used by so few patients in the US that the move has little impact.


-- by Melissa An, Manu Venkat, and Kelly Close