There have been some scary headlines in the last week regarding MannKind's (MNKD) Afrezza - let's address the arguments underlying the headlines.
The latest negative piece is courtesy of the anonymous Equity Options Guru:
"I don't expect these trials to be successful. Back in 2008, the trials didn't show great improvement vs. the placebo and I don't expect that they will this time."
That 2008 study is also used to justify Brian Wilson's stance in an article published a few days earlier:
This trial is similar to the original Afrezza trial in T1d (Study 009), although it might have an even more difficult time due to the fact that enrollment criteria for HbA1c levels has been upped to a minimum of 7.5% versus the 7.0% in Study 009. There are other factors that hurt Affinity 1 (we will focus on later), although MannKind investors can analyze vital information of this trial for themselves.
Study 009 suggested that technosphere insulin (Afrezza) has inferior efficacy versus placebo. It is very likely that data from Affinity 1 will have the same end result, which would prevent Afrezza from receiving FDA approval. Since Afrezza seems to be the primary constituent of MannKind's valuation, the stock should see a drastic revaluation to the downside once this is realized.
So both the Equity Options Guru and Mr. Wilson judge that Afrezza will not gain approval based on 2008's study 009.
Let's address Mr. Wilson's first quoted sentence, which is doubly misleading.
First, the blood sugar. Mr. Wilson is suggesting that enrolling patients with higher blood sugar than previous studies could somehow make Afrezza less likely to be approved - the very opposite is true. Patients with higher blood sugar are more likely to show improvement than patients with lower blood sugar. The American Diabetes Association:
However, older drugs have typically been tested in clinical trial participants with higher baseline HbA1c, which is itself associated with greater treatment emergent glycemic reductions, irrespective of therapy type.
You could argue that both Afrezza and the subcutaneous insulin would both benefit equally - but the patients in the Affinty 1 trial are already on subcutaneous insulin, and it is not working. Enrolling patients with higher blood sugar makes Afrezza more likely to be approved, not less.
Perhaps Mr. Wilson was concerned that MannKind could not find enough patients with such high blood sugar to complete the trials? If so, it is important to note that both Affinity studies are nearly complete and tracking well, CEO Alfred Mann from MannKind's latest conference call in February:
One of our concerns had been potential patient dropout because of the substantial demands on the patients in those trials. We therefore overenrolled to better ensure that we have adequate statistical power. We are pleased to report that both studies are tracking well. At this point, we are quite confident that the trial will be completed as scheduled with more than enough patients. As of last week, 297 patients already completed the treatment phase for Affinity 1, that's about almost 75% of the total, and 167 for Affinity 2 or approximately 2/3 -- a little over 2/3 of the target.
Next, the assertion that study 009 is similar to the Affinity 1 study ignores two very important differences between the studies - aggressive treat-to-target titrations, and titration monitoring teams. Let's first look at the treat-to-target titrations.
Afrezza has completed 59 clinical trials, but it has never done a study that is treat-to-target. In our previous article on MannKind we noted that in past trials, Afrezza has not always shown how effective it can be. We noted that every type 1 diabetes trial combined Afrezza with a basal insulin - all type 1 diabetics need basal insulin. The problem is that at too high a dose, basal insulins can cause hypoglycemia, which is dangerous in the near-term. For this reason, clinicians administering trials can tend to use too little basal insulin to optimize blood sugar at fasting times, especially when a meal-time insulin will be introduced later. This results in a raised fasting glucose level.
This skews Afrezza's mealtime data as measured by HbA1c - HbA1c is a measure of average blood sugar over the past several months. CEO Alfred Mann on the third quarter conference call:
Our problem in early trials is that our fear for hypos, it has been difficult to get fasting glucose levels much reduced. The most significant cause of hypos with insulin therapies today is the late postprandial hyperinsulinemia seen with current prandial products. That excessive insulin causes a postprandial plunge of glucose below baseline that offsets the higher prandial rise, producing a minimal average change from baseline. A1C, which effectively reflects the average glucose level over 2 to 3 months is largely determined for this deficient insulin by that high fasting level.
And Dr. Mann on the latest conference call:
"Since our earlier trials, basal insulins were actually not titrated, the fasting glucose for the AFREZZA patients were thus excessive. What is different in this trial is the protocol very clearly defines the proper titration of the basal insulin. As a result, the AFREZZA patients end up with much lower fasting levels"
These very clearly defined titration protocols are treat-to-target - this means that proper titration of the basal (non-mealtime) insulin is not reached until blood sugar is under control.
MannKind had a similar problem with the Afrezza titrations. Clinicians who were not familiar with Afrezza were not dosing patients high enough to properly control blood sugar - the mealtime insulins currently on the market carry a very high risk of hypoglycemia at too high a dose, whereas Afrezza's risk of hypoglycemia is much lower because it leaves the blood much faster, like the insulin delivered from a healthy pancreas.
So again, the titrations for Afrezza are spelled out explicitly, and again, are treat-to-target for the first time.
The aggressive and explicit treat-to-target titrations are an important difference between Affinity 1 and study 009.
The other major difference is that, remarkably, the FDA permitted MannKind to hire an independent monitoring team to make sure those aggressive titrations are followed by clinicians. Dr Mann on the third quarter conference call:
Yet even so because of the fear of hypos is so great -- because it's so great, the basal insulins have not been increased in our earlier studies even for AFREZZA patients. For example, in our earlier MKC-117 trial, fasting glucose was supposed to be adjusted to under 120 milligrams per deciliter. However, even though required in the protocol, average basal insulins in that study were not increased. And as a result, the fasting glucose remained too high, masking the advantages of AFREZZA. The average A1C reduction in the 117 trial for AFREZZA cohort was actually slightly better than for the insulin aspart that is NovoLog, but not enough to show superiority.
What is truly different in 171 is the FDA understands this and has authorized us to retain an independent monitor who sees the e-Diary data, and who is charged to contact the clinician with any noncompliance. With this tool, there's a much greater likelihood that fasting glucose for those on AFREZZA will be lowered below the target of 120 milligrams per deciliter. The only non-inferiority in this trial is required for FDA approvals, with such compliance, AFREZZA patients should perform exceptionally well in this study, reaching significantly lower A1Cs. Such improvement would be unlikely for patients on rapid-acting analogs.
Let me provide an example to illustrate the significance of what this compliance and the 171 study should make possible. With basal insulin titrated for an AFREZZA patient, to yield an average fasting level of, say, 110 milligrams per deciliter and with prandial glucose likely rising an average of, say, about 40 milligrams per deciliter over 2 hours for each meal, the result in A1C at 3 months would be about 5.8%, essentially normal for a nondiabetic. That would really be outstanding. For those using current prandial insulins lowering fasting glucose significantly would not be safe, and the A1Cs would, thus, not be substantially reduced.
The aggressive titrations and titration monitoring teams are important differences between 2008's 009 study and the current Affinity 1 study.
Along with patients with higher blood sugar, Afrezza should show clear superiority in Affinity 1 - but it does not even have to. In order to gain approval Afrezza just has to show non-inferiority, which as Dr. Mann noted, it already has done in trial 117.
These new measures, along with clinical evidence of efficacy in past studies like 117, make Afrezza very likely for approval.