CRANBURY, NJ ( TheStreet) -- Amicus Therapeutics ( FOLD) presented an alternative theory Friday to explain the failure of its Fabry disease drug Amigal in a late-stage clinical trial. Not surprisingly, investors reacted with skepticism and Amicus shares sank 25% to $2.89. It's easy to dismiss Amicus' analysis for what it looks like on the surface: Just another example of a drug company resorting to retrospective data mining in order to stave off the painful reality of a negative clinical trial. But this time, what Amicus is saying about Amigal might be right: The drug works against Fabry disease. The challenge -- more difficult than the company anticipated -- is designing a clinical trial to prove it. Here's what happened Friday: Amicus dissected the Amigal phase III study and found slightly more than half of the Fabry patients had relatively mild disease at the start of the trial, way more than the company intended to enroll. Mild Fabry disease is harder to measure so in these patients, the study was unable to detect a significant difference in kidney response between Amigal and placebo. However, in Fabry patients with severe disease at baseline and therefore more measurable, the difference in kidney response detected between Amigal and placebo was greater. Had the study enrolled more patients with severe disease, as originally planned, the difference in response between Amigal and placebo might have been statistically significant, the company said. Now, to be clear, Amicus conducted this six-month analysis on a post-hoc basis, meaning separating patients by disease severity was not part of the original statistical analysis plan. However, Amicus is speaking to FDA about amending the study design before the next scheduled look at the Amigal data at 12 months. Amicus expects 12-month data to be ready late in the second quarter. Investors have largely written off Amigal today -- and it's hard to blame them -- but skepticism about the drug could turn into optimism if the 12-month data are strong. If that happens, Amicus shares could rebound and more. To understand why Amigal chances shouldn't be discarded quite yet, it's important to go back over some background on Fabry disease and review the design and initial results from the phase III study.
Fabry is an inherited disease in which a genetic mutation stops an enzyme from breaking down a fatty substance known as globotriaosylceramide, or GL3. The buildup of GL3 in cells throughout body causes severe damage to kidneys, heart, brain, and other organ systems. By definition, Fabry patients lack sufficient enzyme to break down GL3, but their inability to do so is highly variable. Amicus, like other companies conducting Fabry disease clinical trials, tried to enroll patients with severe disease burden, defined as higher levels of GL3 in kidneys at baseline. The reason is fairly simple: Patients with more measurable disease at the start of the study tend to have a greater magnitude of response to treatment. The primary endpoint of Amicus' study compared response to Amigal versus placebo at six months, with response defined as a 50% reduction in GL3 levels measured via a kidney biopsy. Enrolling patients with more GL3 in their kidneys gave Amigal a better chance of demonstrating a superior response rate over placebo. As reported last December, the Amigal study failed. Forty-one percent of Amigal patients were counted as responders (meaning they had a 50% decrease in GL3 in their kidneys. By comparison, 28% of placebo patients were responded. The response trend clearly favored Amigal but the difference was not statistically significant. Amicus also measured the absolute percent change in kidney GL3 from baseline to six months. The result: Amigal 41% reduction compared to placebo 6% reduction. Again, the difference favored Amigal but was not statistically significant. Which brings us back to the new data presented Friday at the Lysosomal Disease Network World Symposium. To get a better understanding of why Amigal missed the primary and secondary endpoints of the study, researchers used the results of the kidney biopsies taking after six months to divide patients into groups by the level of GL3 in their kidneys at the the start of the trial. The results of this analysis are illustrated in the chart below:
Thirty-five of 60 (58%) Fabry patients in Amicus' study had baseline GL-3 levels between 0 and 0.3, representing low disease burden that is difficult to measure. In these patients, 32% treated with Amigal were classified as responders compared to 44% of patients treated with a placebo. This is where the Amigal study failed, Amicus believes.
The results for the 25 Fabry patients with baseline GL-3 levels of 0.3 or higher -- more severe, and measurable disease -- tell a different story. Here, 64% of patients (7/11) treated with Amigal were classified as a responder compared to 14% of patients (2/14) treated with a placebo. The separation in response rates between Amigal and placebo continues for GL-3 cut points above 0.3 as well. Amicus designed the phase III study to meet its primary endpoint based on the assumption of a 60-70% response rate in the Amigal arm compared to a 20% or less response rate in the placebo arm. In other words, the study would have succeeded had the overall patient population resembled the 25 patients with baseline GL-3 levels of 0.3 or higher, Amicus believes. At this point, you should be wondering skeptically if Amicus just picked this GL-3 level cut-off level of 0.3 because it makes the Amigal data look the best. The company says it did not, adding that FDA and independent pathologists have all agreed previously that a GL-3 level of 0.3 represents a starting point for measurable, clinically meaningful Fabry disease. The only reason Amicus didn't exclude patients with GL-3 levels below 0.3 was that kidney biopsies are not routinely performed to screen patients for clinical trials. Amicus tried to use urine levels of GL-3 as a surrogate marker for higher disease burden but it proved to be too permissive. Looking ahead, Amicus hopes to bolster its argument supporting Amigal's efficacy with an analysis of the trial after 12 months of treatment. The best case for the company would be if patients treated with a full 12 months of Amigal maintain the the response rate observed at six months, or improve further. According to the study design, placebo patients crossed over to receive Amigal at six months. Amicus hopes to show that the response rates for the placebo-delayed treatment patients at 12 months look similar to the six-month results recorded by the original Amigal-treated patients. If that happens, Amicus and its partner GlaxoSmithKline ( GSK) expect to seek U.S. Approval for Amigal despite the study failing to hit its primary endpoint. Amicus believes FDA will look at the totality of the data, including the benefit seen in patients with more measurable disease. -- Reported by Adam Feuerstein in Boston. Follow @AdamFeuerstein