Teplizumab got a reticent thumbs-up from FDA advisors for an indication in delaying progression of at-risk patients to clinical type 1 diabetes (T1D).
The agency’s Endocrinologic and Metabolic Drugs Advisory Committee on Thursday voted 10-7 that sufficient evidence of benefit outweighed the risks for this novel biologic.
However, a number on both sides indicated how narrowly they had come to making the opposite decision.
“I voted ‘yes’ but had my finger over the ‘no’ button for a long time,” said Jack Yanovski, MD, PhD, chief of the Section on Growth and Obesity at the National Institute of Child Health and Human Development in Bethesda, Maryland.
“This is a limited data set we’re asked to make a decision on,” he acknowledged, “but I came down on the side of ‘yes’ because this is the first ray of hope study that we’ve had in this disorder. There have been numerous attempts over the past 30 years to modify the course of type 1 diabetes and none of them have been successful.”
A First in T1D
If teplizumab ultimately gains approval, it would be a first for type 1 diabetes.
“Current therapies for type 1 diabetes focus on metabolic control in patients with clinical disease rather than the underlying autoimmune process,” said Eleanor Ramos, MD, chief medical officer of sponsor Provention Bio. “These are patients who are otherwise facing lifelong insulin dependence and devastating short- and long-term sequelae.”
Teplizumab is a humanized monoclonal antibody against CD3 given as a single 14-day course of infusions. Its mechanism is thought to be through expanding regulatory T cells and re-establishing immune tolerance.
The Evidence
The two pillars upon which its new drug application rests are the TrialNet 10 (TN-10) trial done in the proposed at-risk population and a set of other trials testing teplizumab in clinical T1D for safety data and to demonstrate impact on C-peptide as a surrogate marker of beta cell loss.
The TN-10 trial showed a relative 59% reduction in risk for developing T1D with teplizumab versus placebo, with a median time to diagnosis of 48.4 and 24.4 months, respectively. It included patients ages 8 to 45 years with preclinical (stage 2) T1D, as defined by two or more autoantibodies against beta cells but not yet the glycemic symptoms of the disease. Participants also had to have a relative with T1D.
The public comment period largely went to patients and clinicians who have T1D themselves, some whose children have it too. They testified to the intense burdens of care and how much difference a few extra years of childhood or young adulthood without this chronic disease would have meant to them and their families.
The Doubts
“I am swayed and influenced by the volunteers who shared their compelling stories,” said James de Lemos, MD, a cardiologist at UT Southwestern Medical Center in Dallas who voted against approval, “but it’s just not enough evidence for me of effectiveness. It’s not an orphan disease that can’t be studied. We’re being presented with a drug that looks like it may actually be a breakthrough drug… But I don’t think the evidence base is nearly sufficient to approve for widespread use based on what is essentially an NIH-sponsored phase II trial that didn’t follow the patients after they had the disease.”
TN-10 included only 44 patients exposed to teplizumab. While there were 750 patients exposed to the drug in other controlled clinical studies in the development program for T1D, those were a different population with clinically diagnosed disease.
On the other hand, Susan Ellenberg, PhD, a biostatistician at the University of Pennsylvania in Philadelphia, voted yes despite struggling with the same data. “I was pushed by the very encouraging results of what is admittedly a small study and something that I can’t really feel is completely definitive. But I would not like to deny people who are the kind of people we heard from today to weigh their own risks and benefits to try this.”
Despite the small number of patients, the trial was well designed with a strong primary endpoint that was hit by a large margin due to the proportionately large number of events and that also had internal consistency across sensitivity analyses conducted by the FDA, noted Marvin Konstam, MD, of Tufts Medical Center in Boston, who voted yes. “If my 15-month-old grandson were to be at high risk for type 1 diabetes, there’s absolutely no question in my mind I’d want him to have the drug.”
Michael Blaha, MD, MPH, a cardiologist at Johns Hopkins in Baltimore, also said he voted yes despite feeling “very conflicted.”
“My yes is very qualified,” he said. “The risk-benefit is very narrow here, and I would only approve this drug for the exact indication of the trial.”
He, like many others, called for a robust risk evaluation and mitigation strategy program, which the company representatives said they were planning with post-marketing surveillance.
The Risks
The TN-10 trial showed 16% of patients had a serious adverse event on teplizumab compared with 3% among placebo-treated patients. Most patients had transient lymphopenia and leukopenia, but 0.6% of patients had serious cytokine release syndrome and there was an excess in diabetic ketoacidosis (DKA, 2.3% vs 0.4% across the whole trial program).
The drug descended from the first monoclonal antibody approved by the FDA back in 1986, OKT3. That immunosuppressant drug caused malignancy and lymphoproliferative disorders.
The company argued that teplizumab was designed to avoid that risk as it’s not an immunosuppressant when used in a short course and that it does not engage effector cells or lead to T-cell depletion.
Still, Epstein-Barr virus reactivation was elevated with teplizumab, so there is a theoretical risk for later malignancy from lymphoproliferative disease, noted an FDA presenter. Also, there was one case of malignancy seen in the overall safety population with it.
It’s a safety issue that cannot reasonably be resolved in premarketing period because of long latency, noted FDA clinical reviewer Lauren Wood Heickman, MD.
Several advisory committee members suggested that a narrow indication population could allow for further randomized trials to permit replication of the results.
Erica Brittain, PhD, a statistician and deputy branch chief at NIAID, called her yes vote a “close call” given the safety signals.
“I would like to see more randomized studies, whether it’s a comparison of one dose versus two doses or perhaps versus placebo in the younger age group,” she said, noting that pushing T1D screening out into routine pediatric practice would certainly help recruitment.
In fact, having a drug approved in stage 2 T1D may actually push the field forward by giving families and physicians a reason for screening, noted the industry representative on the panel, Gary Meininger, MD, of Vertex Pharmaceuticals in Boston. Since some present clinically with DKA, he said, “part of the benefit might be an opportunity to prevent that.”
FDA is not required to follow its advisory committees’ recommendations, though it typically does. However, split votes are generally taken as no recommendation one way or the other.