Betting on alpha?


Betting on Alpha?


Is alpha protection something that really matters to clinicians? This Greek letter refers to a type 1 error in statistics (rejecting the null hypothesis when the null hypothesis is true). It should be sufficiently low (usually set at 0.05, or a 5% probability of falsely rejecting the null hypothesis) that you can believe the statistical result.

Now, readers of this column know well that I am not a fan of statistics, especially when they are required to explain a clinical study. So why am I talking about alpha protection in this issue of Retinal Physician? In the pivotal phase 3 clinical trials of ocriplasmin (Jetrea, Thrombogenics/Alcon/Novartis), there were only two alpha-protected end points: OCT-proven release of vitreomacular adhesion and total posterior vitreous detachment.

As you know, the study evaluated ocriplasmin against an intravitreal injection of a vehicle control. This is a key point because the FDA requires this control for any vitreolysis trial to compensate for mechanical effects of the intravitreal injection itself on the vitreous. In the MIVI-TRUST trials, both alpha-protected end points were met in both trials, and combined with safety, these end points form the basis for the FDA approval.

In my opinion, one of the most impressive findings from the ocriplasmin trials is the ~40% closure of full-thickness macular holes without surgery, gas, or facedown positioning, with a single ocriplasmin injection. This is especially true with smaller holes (≤250 μm), with which the rate was almost 60%.

I think this is a groundbreaking result, but you will never hear about it because although macular hole closure was a secondary outcome, it was not alpha-protected and therefore is not on the FDA-approved label and cannot be promoted.

There are two ways to look at this statement. The FDA is tasked with insuring the safety and efficacy of drugs for the US public, and because the outcome was not alpha-protected, we cannot be sure it is not due to chance. In other words, it may not work at all in macular hole patients. An additional alpha-protected trial would be required to determine whether ocriplasmin can successfully close macular holes and to expand the label to include this fact.

In the meantime, I see these data as additional information that I can use to decide how to manage my patients with macular holes. As with anything in medicine, it is important to understand the scientific basis of a treatment to best manage your own patients.