FDA Panel Endorses Insulin Degludec
A federal advisory panel has voted 8 to 4 in favor of licensing Novo Nordisk's novel basal insulin degludec for people with both type 1 and type 2 diabetes, with the stipulation that the company conduct a rigorous postmarketing study to evaluate cardiovascular (CV) safety.
A safety "signal" for a possible increase in major adverse coronary events (MACEs) with degludec arose in analyses of clinical trial data by both the company and the US Food and Drug Administration (FDA). The findings were presented yesterday to the FDA's Endocrinologic and Metabolic Drugs Advisory Committee.
All 12 committee members voted that the company should conduct a CV safety study, with the majority agreeing with the manufacturer that the study could be done on a postmarketing basis and the minority feeling that the issue should be resolved before FDA licensure. The FDA is not obligated to follow the recommendation of its advisory panels, although it usually does.
Novo Nordisk's insulin degludec (abbreviated IDeg, brand name Tresiba) is a long-acting basal insulin that forms soluble multihexamers on subcutaneous injection. It has a half-life of 25 hours, which is twice as long as currently available basal insulin products, with a 42-hour duration of effect.
Degludec was designed to meet the need for a flat, consistent basal insulin that provides steady coverage for more than 24 hours and from day to day for all patients, with the primary advantage being reduced hypoglycemia, particularly at night. In addition, because of its long half-life, the dosing regimen for degludec allows for flexibility if a patient misses or delays a dose, according to Novo Nordisk's Vice President for Regulatory Affairs Robert Clark.
The company is seeking approval for 3 different formulations: IDeg, a once-daily subcutaneous administration to cover basal needs either alone or in combination with mealtime bolus insulin or oral glucose-lowering medications; a premixed combination of 70% IDeg with 30% Novo's rapid-acting insulin analog aspart (IDegAsp, brand name Ryzodeg) for once- or twice-daily administration; and a U200 formulation of IDeg for use in patients with large insulin requirements, allowing for up to 160 units to be used in a single injection.
The original data set, included with the company's new drug application on January 31, 2011, contained a total of 3570 patients with type 1 and type 2 diabetes who were exposed to IDeg or IDegAsp and 1874 patients who were exposed to a comparator (insulin glargine in most of the studies). An updated data set, submitted at the FDA's request, included extension data from 6 studies and an additional trial, for a total 5153 patients exposed to IDeg/IDegAsp and 2563 patients exposed to comparators.
The safety database included a total 4939 patients with 6 months or more of exposure to either IDeg or IDegAsp and 1870 patients with 12 months of exposure or more.
Safety Signal
Much of the discussion surrounded an evident signal of an increased risk for MACEs. Injectable insulin products are exempted from the 2008 FDA guidance requiring that manufacturers of new diabetes therapies conduct trials specifically designed to assess CV risk. However, the FDA still asks companies to prospectively collect CV data in their phase 2/3 clinical trials.
The FDA and Novo Nordisk differed in their analyses of these data for IDeg/IDegAsp. According to Novo Nordisk's Corporate Vice President Anne Phillips, MD, the rates of prespecified MACEs from the original new drug application submission were similar between IDeg/IDegAsp and comparator, with a hazard ratio of 1.10.
However, with the addition of the extension trial data plus the new trial, and using a definition of MACE that included CV death, nonfatal stroke, and nonfatal myocardial infarction, but not unstable angina, the incidence rates for MACE at 30 days posttreatment discontinuation were 1.41 per 100 person-years for IDeg/IDegAsp vs 0.90 for the comparators, with a significant hazard ratio of 1.61.
Bo Li, PhD, from the FDA's Division of Biometrics, analyzed the MACE data both with and without unstable angina. She found that hazard ratios for an increased risk with IDeg/IDegAsp were 1.30 when unstable angina was included and 1.67 when it was not at 7 days after treatment discontinuation, both suggesting increased CV risk with IDeg/IDegAsp.
Panelists agreed that the small absolute number of MACEs (95 with IDeg/IDegAsp vs 37 for comparators, using the definition that included unstable angina) were too small to determine whether the signal was real.
Difficult Decision
In voting for or against approval for licensure of degludec before the company conducts a study specifically aimed at assessing CV risk, panel members weighed the benefits of the reduced hypoglycemia (for which Novo also presented data), ease of dosing, and ability to mix degludec with the short-acting insulin aspart against the possibility that the MACE signal was not just statistical noise.
Panelist William R. Hiatt, MD, professor of medicine in the Division of Cardiology at the University of Colorado School of Medicine in Denver, voted yes. "I struggled a bit and contemplated the implications of a preapproval safety study which would be for any sponsor, any indication, quite onerous.... The thing that swayed my vote to yes was the commitment that a properly powered and designed study will go forward and this question will be definitively answered." He added that interim analyses should be conducted and action taken if early signals suggest harm.
David W. Cooke, MD, associate professor of pediatrics at Johns Hopkins University School of Medicine in Baltimore, Maryland, also voted yes. "It was a bit of a difficult call, but my reason was that the pharmacokinetics of this insulin are an advance for our treatment of both type 1 and type 2 diabetes and would provide an additional option."
However, Brendan M. Everett, MD, associate physician at Brigham and Women's Hospital in Boston, Massachusetts, voted no. "I'm not an endocrinologist. I take care of patients with heart disease. The thing that I worry about most in my patients with diabetes is when and if they will develop heart disease.... For me, experience with prior antidiabetic medications with an adverse [CV] risk profile that get widely used put patients at increased risk for adverse outcome." Dr. Everett said.
"I don't know that what we've seen is real, but I have enough concerns about it that I would want to know the answer before a substantial number of patients with diabetes take this medication," he concluded.
Endocrinologist Robert J. Smith, MD, also voted no. "This was an extremely difficult decision for me.... I think this form of insulin is addressing a true need...I think it's likely to be useful and broadly used, but we do have alternatives now." He said although he believes degludec represents an "improvement" on current therapy, the improvement "is not of such magnitude" that it merits taking on the potential risk.
Committee chair Kenneth D. Burman, MD, voted yes. "Unfortunately, we're always in a precarious position, trying to the balance the efficacy and side effects of a new agent when inherently we have relatively short-term CV studies. The balance I have reached recommends postmarketing CV studies with definitive endpoints, such as lipids, [C-reactive protein], cardiac echo, and perhaps carotid dopplers."
Dr. Burman, chief of the Endocrine Section at Washington Hospital Center in Washington, DC, acknowledged that if the signal does turn out to be real, people would have been exposed to the risk. "On the other hand, it's usually impractical to require long-term definitive studies prior to approval. Therefore we have to render our best judgment; to me at present the benefits outweigh the possible adverse effects."
FDA advisory panels are vetted for financial conflicts of interest. One panel member not mentioned in this story was granted a waiver to participate.