Elan-Wyeth Alzheimer's Data Spook Bulls

The stocks sell off after the presentation of underwhelming and inconclusive study results of bapineuzumab. Let's break it down.
Publish date:

(Editor's note: Come see Adam Feuerstein at the Money Show in San Francisco. Adam will be speaking to attendees on Friday, Aug. 8, at 2:15 p.m. ("Biotech Investing for Individuals: How to Turn Geeky Science Into Fat Profits"); on a lunch panel on Saturday, Aug. 9, at 12:35 p.m. ("Tech and Biotech: Picks and Pans for 2008 and Beyond"); and on Sunday, Aug. 10, at 8 a.m. ("Biotech Investing for Individuals: How to Turn Geeky Science Into Fat Profits").

Updated from 8:35 a.m. EDT

CHICAGO -- The coronation of bapineuzumab as the future king of all Alzheimer's drugs has been put on hold.






, co-developers of bapineuzumab, entered Tuesday's presentation of phase II data at the International Conference on Alzheimer's Disease brimming with confidence.

After more than a month of speculation, researchers, the public and Wall Street were finally going to see the clinical data that would clearly establish bapineuzumab as the game-changing treatment for Alzheimer's disease. Skeptics, including short-sellers of Elan shares, were going to be vanquished once and for all.

By Tuesday night, however, the script had been rewritten. The

presentation of the bapineuzumab data

was weak, the results were

underwhelming and inconclusive

. Far more questions about the future of the drug were raised than were answered.

Investors reacted accordingly, selling Elan and Wyeth shares sharply in after-hours trading. In recent trading Wednesday, Elan was sinking $12.21, or 36.2%, to $21.54, and Wyeth was losing $5.05, or 11.2%, to $40.06.

As I did back in June when Elan and Wyeth

first announced top-line results

from the phase II bapineuzumab study, let's dissect the big issues affecting the drug Tuesday with a hypothetical Q&A.

Those are some strong words you use above to describe the bapineuzumab data presentation. Aren't you being a bit harsh?

Not really. Let's get real here. This phase II study failed its primary endpoint rather convincingly. Patients treated with bapineuzumab showed no improvement in either cognition or function compared to placebo.

So, what did Elan and Wyeth do with these negative results? They essentially threw them out and rewrote the rules of the study in order to find groups of patients who responded. First, the two companies changed the statistical analysis plan. Second, they broke down and analyzed the data for efficacy based on a post-hoc subgroup comparing patients who are carriers and non-carriers of the ApoE4 gene, a genetic variant that increases a person's risk for developing Alzheimer's.

Not content with one retrospective analysis, Elan and Wyeth went a step further. They also re-analyzed the results using a so-called "completers analysis" that only included patients who received the full complement of bapineuzumab injections.

Such extensive data-mining as this totally erodes any confidence that the "positive" findings being shown can be replicated in the ongoing phase III studies. There is way more risk in the bapineuzumab program now, which is why the stocks sold off so hard.

But Elan and Wyeth told everyone in June that they conducted a post-hoc analysis of the bapineuzumab data. At that time, most investors didn't seem to have any problems with this approach and the stocks went up. None of this should have surprised anyone Tuesday, so why the radically different reaction?

Because Tuesday afternoon, we finally got to see the actual data from the bapineuzumab study and it was wildly inconsistent.

Let me give you one example: There was absolutely no dose response with bapineuzumab in this trial. Typically, you like to see increasingly higher doses of a drug correspond with improved efficacy. Instead, what we saw from the bapineuzumab study just looked like so much randomness, which in clinical trials is definitely not a good thing.

The lowest doses of the drug worked better on some measures of cognition and function, while higher doses worked better on others. Sometimes, doses in the middle produced worse results than placebo.

It seems quite logical to characterize patients based on their ApoE4 status, and this is how Elan and Wyeth have set up the ongoing phase III studies. Didn't the bapineuzumab data Tuesday in the ApoE4 non-carriers look quite strong and convincing?

The ApoE4 non-carrier group of patients was Elan and Wyeth's best shot at making a convincing case for bapineuzumab, but once again, the details in the data Tuesday -- stuff not available in June -- tripped them up.

For starters, there was an imbalance in the baseline mental status of patients in the bapineuzumab and placebo arms that could have skewed the results in favor of the drug.

When you look at the actual performance curves of the study, bapineuzumab-treated patients reported a 5-point improvement over placebo on the ADAS-cog test, a measure of cognition and a co-primary endpoint of the study.

However, for nearly a year into the 18-month trial, the bapineuzumab and placebo patients were both losing cognition at the same rate. The benefit seen for bapineuzumab patients only came about because placebo patients suffered a steep loss of cognition at the very end of the study.

I was under the impression that a widening of the ADAS-cog curves between bapineuzumab and placebo over time is a good thing because it shows that the drug is slowing the progression of Alzheimer's.

True, but there is no good explanation for why bapineuzumab would take so long to be effective compared to placebo. The fact that the two ADAS-cog curves (bapineuzumab and placebo) were essentially identical for much of the first year of the study is a problem because most Elan "believers" thought the curves would separate early in bapineuzumab's favor and remain that way for the entire length of the study.

On Elan's conference call Tuesday night, there were questions raised about the performance of the placebo group in the ApoE4 non-carrier group. What's that all about?

The loss of cognition by placebo patients totaled 11 points on the ADAS-cog scale at 18 months, far worse than is typically seen in other Alzheimer's studies of this duration.

For instance, the placebo patients in the large phase III study of

Myriad Genetics'

(MYGN) - Get Report

Flurizan reported a 7-point loss of cognition on the ADAS-cog scale. These data, from a study much larger than Elan and Wyeth's phase II, were also presented Tuesday at the conference.

The concern here is that the improvement in cognitive function seen in the ApoE4 non-carriers is not a result of anything that bapineuzumab is doing, but is instead caused by poor-performing placebo patients from a small subset analysis.

You haven't touched yet on bapineuzumab's safety. Anything notable to report there?

Twelve patients in the study treated with bapineuzumab developed a potentially dangerous accumulation of fluid in the brain known as vasogenic edema. All of these cases resolved favorably however, and six of the patients were eventually able to continue treatment with bapineuzumab.

This makes the side effect profile of bapineuzumab appear relatively benign, and that was certainly the spin offered by Elan Tuesday.

There is another side to this safety story, however. Researchers were only able to detect the vasogenic edema by subjecting patients to numerous (and expensive) MRI scans. In the real world, Alzheimer's patients will never be monitored this closely, so what happens if vasogenic edema goes undiagnosed and untreated in bapineuzumab patients?

Other safety issues: Several instances of bapineuzumab patients experiencing "micro-bleeds" in their brains, plus higher rates of seizures and psychiatric events.

In June, you quoted an Elan short-seller who sounded bitter and a bit desperate. What's he saying now?

Needless to say, he's much more chipper. As a matter of fact, he emailed me Tuesday night. Here's what he wrote:

"Are you finding anyone who seriously thinks the bapineuzumab data show anything? I have never seen a better demonstration of randomness, even in the apoE4 non-carrier group, in my career. This is the most amazing intentional obfuscation of what is going on that I have ever seen. Truly astounding."

And what about the Elan bulls?

Again, I did hear from the same longtime Elan investor whom I quoted back in June. He writes:

"The data were quite interesting and encouraging. I cannot fathom the market reaction, but after-hours trading is not good at sorting through pretty complex stuff. Did the market think that Elan and Wyeth were going to announce an early filing and approval by the end of the year? "I think Elan and Wyeth will have an approved Alzheimer's disease drug within the next 30 months, and perhaps more than one. Bapineuzumab worked better than the multi-billion dollar approved Alzheimer's drugs in the non-carrier group at the dosages selected for further testing in phase III, and there is enough suggestion of efficacy in the carriers that the investment in two large phase III trials is warranted."

Sounds like a continued standoff.

Yup, but the playing field that was tilted in favor of the bulls appears now to be moving in the other direction.

As a very public bapineuzumab skeptic, you must be feeling pretty good today.

I feel a whole lot better and more confident than I felt on Monday, for sure.

Adam Feuerstein writes regularly for TheStreet.com. In keeping with TSC's editorial policy, he doesn't own or short individual stocks, although he owns stock in TheStreet.com. He also doesn't invest in hedge funds or other private investment partnerships. Feuerstein appreciates your feedback;

click here

to send him an email.