How do you solve a problem like Afrezza: An FDA dilemma
This article was originally published in Scrip
The FDA is facing a dilemma.
For the third time, Mannkind is seeking approval of its experimental inhaled insulin drug-device product Afrezza (insulin human [rDNA origin]) after previously being rejected twice.
Of concern is the fact that the only inhaled insulin to ever make it onto the US market – Pfizer's Exubera – was found to have an imbalance in lung cancers.
Pfizer later withdrew Exubra, but it blamed the action on poor sales, not the cancer concern.
But as far as Mannkind’s product goes, perhaps the FDA's Dr Jean-Marc Guettier, acting director of the agency's Division of Metabolism and Endocrinology Products, said it best: "Afrezza has had a complex regulatory history."
So the FDA wants advice from its Endocrinologic and Metabolic Drugs Advisory Committee (EMDAC), which is convening for an 1 April meeting, where Afrezza will finally get its day in the public spotlight.
But ahead of the meeting, the FDA drug reviewers raised some safety issues in briefing documents posted on 28 March, specifically bronchospasm, pulmonary function decline and lung malignancies, which gave investors cause for concern, triggering an 8.6% selloff of Mannkind's stock.
New data also showed that while Afrezza is effective in types 1 and 2 diabetes, the FDA questioned the robustness, particularly in the type 1 disease.
The Valencia, California-based biotech's shares closed at $4.83, a loss of 37 cents, or 7.12%.
But JP Morgan analyst Cory Kasimov said he didn't see any obvious "smoking guns" or a "knockout punch" in the briefing documents.
Sagient Research's BioMedTracker, an affiliate of Scrip, however, lowered its likelihood of Afrezza winning the FDA's nod in the in type 1 diabetes indication by 1%, but raised the drug’s chances from 80% to 83% in the type 2 indication, which is 2% above the average probability of US approval for the same indication based on the historical performance of medicines in the same development phase.
The Exubera effect
The FDA said pulmonary safety issues are of particular concern with Afrezza because of the experience with Exubera.
At the time of the new drug application was filed for Exubera in 2006, there was a known imbalance in lung cancers in study participants exposed to the drug, the FDA acknowledged.
Pfizer conducted a follow-up study in patients exposed to Exubera and comparison medications in pre-approval clinical trials and to standard of care after trial completion.
In July 2012, FDA received the final study report for the FUSE trial, which showed there were significant imbalances in lung cancer mortality.
"Because Afrezza also directly administers insulin into the lung, this safety concern with Exubera may be relevant to Afrezza and other inhaled insulin products," regulators said.
BioMedTracker analysts noted that 11 of the 15 cases of lung cancers reported in the Exubera study were former smokers, though they noted details on those patients’ histories were not provided, so it is unknown how much they smoked.
The analysts also pointed out there was only a small imbalance in the Afrezza program, with two versus zero patients when looking at all trials over 14 days, though there were an additional two cases in non-smokers spontaneously reported after they had completed participation in the studies.
In addition, the BioMedTracker analysts also noted that while the FDA's reviewers said those data were of concern, they said the interpretation must be taken with caution, due to the possibility of reporting or detection bias.
But in light of the Exubera experience, the FDA said the consideration of the adequacy of Afrezza's pulmonary safety data by the EMDAC and what further data may be required – pre- or post-approval – "will be an important topic of discussion."
The FDA said its staff concluded Mannkind's registry proposal had insufficient detail.
"Given that pulmonary malignancy risk is heavily confounded by smoking and that Afrezza users may have more thorough or frequent pulmonary assessments than the proposed comparator, the proposed postmarketing approach would be inadequate to evaluate risk of pulmonary malignancies with Afrezza use," the agency’s reviewers said.
They said a registry with methodology to reduce detection bias and a large randomized controlled study might better assess pulmonary malignancy risk with Afrezza.
Successful, but not robust
Mannkind has, nonetheless, remained undaunted by New York-based Pfizer's or the California biotech’s own experience, and has pressed on in seeking the FDA's approval to market Afrezza as an adjunct to diet and exercise to improve glycemic control in adults with type 1 or type 2 diabetes, despite two previous rejections – coming back to the agency with new data in hand from two requested Phase III studies and a newly designed inhaler device, known as Gen2 (scripintelligence.com, 20 January 2011, 9 August 2011).
To gain US approval for type 1 and type 2 diabetes, products must typically demonstrate they offer superior versus placebo or non-inferior versus active comparator HbA1c reduction at about six months in adequate and well-controlled trials.
In Mannkind's Phase III 171 study, Afrezza was tested in type 1 diabetes against insulin aspart, which is sold by Novo Nordisk as NovoLog.
The FDA said the primary analysis met the criterion that Afrezza delivered with the Gen2 inhaler was non-inferior to insulin aspart in lowering HbA1c after 24 weeks of treatment in patients whose disease were suboptimally controlled with their current basal insulin regimens – insulin glargine, insulin detemir or NPH insulin.
But the reviewers said the comparative efficacy shown was not compelling since the upper bound of 0.37% of the 95% confidence interval of the treatment difference in change from baseline in HbA1c at week 24 was almost right at the boundary of the pre-specified margin of 0.4%, and the mean reduction in the Afrezza patients was actually statistically significantly worse by an estimate of 0.22% versus the insulin aspart-treated patients.
There were 25% and 11% dropouts in the Afrezza and insulin aspart treatment arms, which the FDA's reviewers said could have potentially impacted the primary non-inferiority analysis.
Among the sensitivity analyses Mannkind conducted, all showed similar findings to the primary analysis, except for the multiple imputation under the noninferiority null method, where 0.4% was added to every discontinued patient in the Afrezza group, failing to satisfy the noninferiority criterion, regulators said.
The 95% confidence intervals for the primary and sensitivity analyses were all above zero, demonstrating that Afrezza was inferior to insulin aspart in the HbA1c change from baseline to week 24.
About 55% of Afrezza patients had improved HbA1c levels at 24 weeks, versus 73% in the insulin aspart group.
The firm's Phase III 175 study examined Afrezza compared to placebo in patients with type 2 diabetes.
Data from study demonstrated Afrezza was statistically superior to placebo in lowering HbA1c after 24 weeks in patients whose disease were suboptimally controlled on optimal/maximally tolerated doses of metformin only or two or more oral antidiabetic drugs.
But the FDA said the treatment difference in change from baseline in HbA1c at week 24 was modest.
In addition, regulators said there were 21% and 30% dropouts in the Afrezza and placebo arms, respectively, which they said could have potentially impacted the primary superiority analysis.
The FDA said 86% of the Afrezza patients experienced an improved HbA1c level after 24 weeks of treatment, versus 72% placebo group.
Type 1 approval unlikely
While treatment with Afrezza was shown to be effective in lowering HbA1c versus placebo in type 2 diabetes, the FDA reviewers said the robustness of the analysis is an issue because of missing data.
"Since there was only one confirmatory study submitted for the indication of type 1 diabetes mellitus, this makes drawing a solid conclusion regarding efficacy for this type of diabetes mellitus problematic," the agency's reviewers said.
Among the topics the FDA wants its EMDAC to discuss is the impact the missing data may have on the efficacy determination for Afrezza.
The agency also plans to ask the committee to vote separately on whether the agency should approve Afrezza in type 1 or type 2 diabetes.
Cowen analyst Dr Simos Simeonidis predicted the FDA would refuse to grant the type 1 diabetes approval, although he said the "door is possibly open" for the type 2 indication for Afrezza.
"But a positive vote and approval would be a surprise," Dr Simeonidis said.