Dimebon Disappoints: Is There Hope for Novel Alzheimer's Agent?

Susan Jeffrey

March 24, 2010

March 12, 2010 ( Updated March 25, 2010 ) — Top-line results of 2 phase 3 trials of latrepirdine (Dimebon, Medivation/Pfizer) show that the drug failed to meet either primary or secondary endpoints vs placebo for the treatment of Alzheimer's disease (AD).

The results have caused not only disappointment but some skepticism now about the original phase 2 results with latrepirdine, published in the Lancet in 2008, which had shown extremely positive response to the drug.

The CONNECTION trial, a randomized, phase 3, multicenter trial of almost 600 patients with AD, showed no difference between those receiving 20 mg 3 times daily of latrepirdine vs placebo on the coprimary endpoints of change from baseline on either cognition, measured with the Alzheimer's Disease Assessment Scale–cognitive subscale (ADAS-cog), or global function, measured with the Clinician's Interview–Based Impression of Change Plus Caregiver Input (CIBIC+), after 6 months of treatment.

Results of a separate phase 3 trial with the aim of assessing the safety and tolerability of latrepirdine in combination with other approved AD medications, including cholinesterase inhibitors, memantine or both, were released at the same time. The results confirmed the safety of combining latrepirdine with other medications, although now to an uncertain end.

"The results from the CONNECTION study are unexpected and obviously a major disappointment for all of us, especially for Alzheimer's disease patients and their caregivers," David Hung, MD, president and chief executive office of Medivation, which is developing the drug in collaboration with Pfizer, said in a telephone conference call.

The top-line results were released March 3, and full data are expected to be presented at an upcoming medical meeting, Dr. Hung noted, although he did not specify which.


CONNECTION was a phase 3, multinational, double-blind safety and efficacy trial conducted at 63 sites in North America, Europe, and South America. A total of 598 patients with mild to moderate AD were randomized to 1 of 2 doses of latrepirdine (20 mg or 5 mg 3 times daily) or placebo. The 5-mg 3 times daily dose had been included to define the effective dose range, the company noted.

Patients included had a mean age of 74 years and a mean score of 17.7 on the Mini-Mental State Examination (MMSE) at baseline.

After 6 months of treatment, no difference was seen between the group receiving latrepirdine, 20 mg 3 times daily, and placebo. On the ADAS-cog, treated patients achieved a nonsignificant 0.1-point difference from those taking placebo (P = .86), and neither group was significantly changed from baseline. No difference either was seen in independently rated global function on the CIBIC+; 64.9% of patients taking latrepirdine, 20 mg, showed improvement or no change at week 26 vs 65.4% among those taking placebo (P = .81).

Similarly, there were no differences on secondary efficacy endpoints. Patients taking latrepirdine showed a nonsignificant difference of 0.4 point from those taking placebo on the Alzheimer's Disease Cooperative Study Activities of Daily Living Scale (P = .61), although, again, neither group changed significantly from baseline.

The Neuropsychiatric Inventory (NPI) showed a 1.6-point improvement in behavior for treated patients over placebo (P = .17); both groups improved over baseline on this measure, but the change was only significant for the latrepirdine group.

However, on the MMSE, both groups improved significantly over baseline, and the improvement was actually greater, although not significantly, with placebo (latrepirdine, 0.7; placebo, 1.2; P = .10).

The 20-mg dose was well tolerated, with similar numbers of patients having at least 1 adverse event (72.0% with latrepirdine vs 74.2% with placebo). The most frequently reported adverse events with latrepirdine occurring more commonly than in placebo were somnolence, dry mouth, headache, dizziness, constipation, cough, and depression.

In the separate phase 3 safety and tolerability study, 742 patients with mild-to-moderate AD from the United States and Canada were randomized to receive either 20 mg 3 times daily of latrepirdine or placebo, with treatment for either 3 or 6 months. Approximately 85% of patients were already taking an approved AD medication.

The drug was well tolerated, the companies reported. The most frequently reported adverse events in the latrepirdine group occurring more commonly than with placebo were somnolence and fatigue.

Lancet Phase 2 Study?

The new results are at odds with those previously reported in a phase 2 trial published in the Lancet in 2008 by lead investigator Rachelle Doody, MD, PhD, professor of neurology at Baylor College of Medicine in Houston, Texas, and colleagues (Doody R, et al. Lancet. 2008;372:207-215).

The drug had originally been developed as an antihistamine in Russia, and so the phase 2 trial was performed at 11 centers there, although under the direction of US clinical trialists.

In that trial, patients with mild to moderate AD who were taking the drug showed impressive and significant improvements across all domains measured; cognitive function, memory, ability to perform tasks of daily living, global function, and behavior, which increased and then were sustained over time.

"At this point we are keenly focused on 2 key questions," Dr. Hung said on the conference call. "First, why are the results from CONNECTION so different from those seen in the Lancet trial, and second, what are the implications of these for the ongoing development of Dimebon for Alzheimer's and Huntington's diseases?"

To the first question, he noted that "the 2 biggest differences we saw between the 2 trials were, first, response to Dimebon treatment was materially weaker in CONNECTION than in the Lancet, and second, we saw a placebo response in CONNECTION that we did not see" in that previous paper.

For example, in the 20-mg 3 times daily dose studied in both trials, latrepirdine patients were significantly improved over baseline scores on all 5 efficacy endpoints after 6 months of treatment. In CONNECTION, treated patients were significantly improved over baseline in only 2 of the 5 endpoints, the NPI and the MMSE.

On the ADAS-cog, a coprimary endpoint, latrepirdine-treated patients improved by 1.9 points in the phase 2 trial vs a decrease of 0.2 point in CONNECTION, he noted. On the CIBIC+, the other coprimary endpoint, 81% of latrepirdine-treated patients were improved or unchanged in the phase 2 trial, whereas the corresponding number in CONNECTION was 65%.

In the phase 2 study, placebo-treated patients decline significantly from baseline scores on all 5 efficacy endpoints by 6 months, but in the phase 3 study, there was no decline in the placebo patients on any of these endpoints and, in fact, some improvement in the MMSE. Decline in the placebo patients was consistent between the studies for mild AD patients, but those with moderate disease had decreased by 3.9 ADAS-cog points in the phase 2 study vs only 1.1 points in CONNECTION.

"A placebo response in the normal range would not have changed the outcome of the CONNECTION study," Dr. Hung said. "However, we believe it is worthwhile to analyze the reasons behind this response as part of our process for determining appropriate next steps for the development of Dimebon."

Trials Ongoing, for Now

Four phase 3 trials are still ongoing and enrolling at this point in AD and Huntington's disease, Dr. Hung said. These include the following:

  • The CONCERT trial, a 12-month study testing latrepirdine in patients with mild-to-moderate AD who are taking donepezil;

  • The CONTACT and CONSTELLATION trials, 6-month trials of latrepirdine in patients with moderate-to-severe AD also taking donepezil and memantine, respectively.

  • The HORIZON trial, a 6-month study of latrepirdine in patients with Huntington's disease.

In February, results of another trial using latrepirdine in Huntington's disease, the DIMOND trial, showed some positive effect on cognition in this population (Kieburtz K, et al. Arch Neurol. 2010;67:154-160).

Dr. Hung concluded by pointing out that Medivation and Pfizer have already undertaken more comprehensive analysis of the CONNECTION data, with an eye to determine whether they will have any impact on the ongoing development program for the drug.

"This will be a complex analysis that will encompass scientific, clinical, and business issues," he said. "We recognize this as an urgent task. We will proceed accordingly and update you when decisions are reached."

"Absolutely Zero Effect"

Medscape Neurology polled some AD experts on their reaction to the new results.

Dr. Samuel E. Gandy

Samuel E. Gandy, MD, PhD, from Mount Sinai School of Medicine in New York City, who has also done some investigation with this drug, told Medscape Neurology that he was always skeptical of the meteoric rise of this agent.

"This was a drug with no plausible mechanism that emerged from an incomprehensible series of screens that then had a 'better than anything ever' effect in a Russian trial and then gave absolutely zero effect in a US replicate trial," Dr. Gandy said.

Dr. Gandy pointed to another situation now unfolding, relating to results reported last month with a compound called NOV-002 (Novelos) for lung cancer. Used in combination with chemotherapy, results of a phase 2 study conducted in Russia showed treatment with NOV-002 increased 1-year survival from 17% to 63%, representing an 80% improvement over the US standard of care of 35%. However, in top-line results reported February 24 from a pivotal phase 3 trial, which had been conducted under a Special Protocol Assessment and Fast Track designation by the Food and Drug Administration, the drug failed to improve overall survival in patients with advanced non–small cell lung cancer.

The company is now facing a class action suit claiming that Novelos violated section 10(B) of the Securities Exchange Act in connection with "alleged disclosures" related to the phase 3 trial of NOV-002.

"While formally one should consider differences in drug, in subjects, and in trial design and conduct, it is worth noting that this is the third or so trial within the last few years to be spectacular in Russia and totally ineffective in the US," Dr. Gandy added.

"I would not dismiss the possibility of execution irregularities in the Russian study, and forensic review of that study would go a long way toward reassuring already skeptical experts like myself."

At last summer's International Conference on Alzheimer's Disease, Dr. Gandy presented animal data showing that, despite the promising clinical results, latrepirdine actually increased levels of β-amyloid in mouse brain tissue up to 2-fold, a confusing finding that fueled discussion at the time about what is really known about the role of β-amyloid in the AD process.

Results Overhyped?

Lon S. Schneider, MD, professor of psychiatry, neurology, and gerontology at the Keck School of Medicine, University of Southern California in Los Angeles, pointed out that he has been on record for some time suggesting that the phase 2 results were "overhyped." "Its results have been exaggerated, and we were not doing ourselves or our patients any good by having spoken about this drug as if it's really great," he told Medscape Neurology.  

In this case, a number of differences are clear between the phase 2 and phase 3 trials beyond just sample size, he noted, among them the formulation of the drug used. The phase 2 trial used tablets, and it's known that the medication has a bitter taste and numbing effect on the tongue, he said. The new trial used film-coated tablets rendering the medication tasteless and that might have better maintained blinding, although it's also possible that the substance itself, its formulation and pharmacokinetics, may have been materially different.  

"The generality is that small studies are not terribly reliable," Dr. Schneider pointed out. "They're subject to play-of-chance results and other influences that aren't wholly predictable, so I think the focus should be on the present study, the CONNECTION study, not showing any effect whatsoever."

It could be vetted, but one could ask, vetted towards what end?

Although results of the phase 2 study raised lots of questions from the time it was first presented in 2007, it is not clear that it will now be necessary to answer them. "It could be vetted, but one could ask, vetted towards what end?" he said. "Because you've got this much larger subsequent study, a well-designed trial, that appears to be well conducted and really shows no signal at all, it essentially negates the previous study."  

However, there are trials ongoing in AD with latrepirdine, and if the results of any of those are positive, the question of the conduct and results of the phase 2 study could be raised again. In the setting of a new drug application, the divergent results of these 2 studies would have to be explained. So it might be worthwhile to examine this more closely now.  

Furthermore, it is just not clear how the drug works, "if it works at all," Dr. Schneider pointed out. There are little pharmacokinetic data available, for example, no published information on how well the drug is absorbed or what the relevant concentrations are, and the concentrations are important because preclinical work shows the drug affects myriad neuroreceptors, including serotonergic, α-adrenergic, and dopaminergic subtypes in addition to histamine receptors.  

"So this creates a high degree of uncertainty," he said. "You don't know about the drug, but you know it didn't work in a large mild-to-moderate AD trial. It's possible it could still have some effect in the ongoing trials where it’s added on to donepezil or memantine background therapy, but now, what would the effect be, a drug that has efficacy on top of donepezil but not alone, and how would it be interpreted?"

He does not give any weight, though, to the lack of decline in the placebo group as an explanation for the lack of drug effect seen in CONNECTION. "The authors of the phase 2 study emphasized that the effect of Dimebon was improvement over baseline in all 5 outcomes and was not due to the placebo decline. The placebo group in the CONNECTION study actually worked the way monotherapy placebo groups work in clinical trials," he noted.

"Counterintuitively," he said, AD trial patients who are not taking donepezil or other cholinesterase inhibitors tend to deteriorate only slightly if at all over 6 months, "despite conventional wisdom, while those AD patients who enter clinical trials already taking these medications generally worsen on average by about 1.5 points on the ADAS-cog over 6 months."

Too Good to Be True Usually Is...

Dr. Steven DeKosky

Steven DeKosky, MD, vice president and dean of the University of Virginia School of Medicine in Charlottesville, pointed out that although many phase 2 findings do not translate into good outcomes in phase 3, this situation was a bit different.

"The fact was that the data were really good across every domain in the Russian study. The idea that a much bigger, well-regulated study here would not show anything is casting doubt back on those phase 2 data," Dr. DeKosky said. "It's an easy cliché, but when data are too good to be true, they usually are, and this case, they were," he said. The next question naturally is — why weren't they true?

Even the Food and Drug Administration had agreed to consider the phase 2 trial as 1 of 2 pivotal trials toward approval of latrepirdine for AD if the phase 3 trial had been positive, he noted. He speculated that the agency was willing to accept a study done in another country where such trials were not commonly done toward approval of a drug here "perhaps related to the fact that the investigators were trained by an American trialist who is very good," he pointed out.

This is a big loss for us.

Although there had been some skepticism about this drug, there was also great hope. "There's huge pressure to produce a medication that works in this disease, and there's huge money available to someone who comes up with a successful treatment," he said.

The negative outcome here means some in the business and clinical communities appear to be disappointed, even angry, and want to revisit the phase 2 results, he said. "I'm just upset," Dr. DeKosky said. "This is a big loss for us" in the AD community.

"I wince every time one of these medications doesn't work because not only has it gotten the hopes of all of our patients and families up, but pharma could look at this again and say 'this disease may be too tough for us, we're going to get out of the field,'" he told Medscape Neurology.

No Lessons

Robert Green, MD, from Boston University School of Medicine in Massachusetts, was also disappointed with the outcome of this promising lead. No stranger to disappointing trial outcomes, Dr. Green presented negative phase 3 results with another promising AD agent, tarenflurbil, in 2008.

"So I'm sympathetic, but I'm not sure what anybody can say," he said. "I'm disappointed, as we all are, that what appeared to be a promising new line of treatment may not in fact be efficacious. I commend the company and the investigators for following up on the promising early data and subjecting it to a rigorous and exacting trial, which appears, with a larger number of people in more experienced sites, to be unable to demonstrate efficacy."

The urge to revisit the phase 2 data though seems to him "not entirely fair, given that we've now had several phase 2 trials done in western centers that have appeared to demonstrate efficacy and then were demonstrated not to have efficacy in phase 3."

"It's easy to point fingers, and I was one of those who had concerns that the quality of a trial done in less experienced hands might not be as rigorous," Dr. Green acknowledged. "But this is the way science works. You get exciting leads, and a lot of them don’t pan out."

This is the way science works. You get exciting leads, and a lot of them don’t pan out.

"I don't know that there's any generic lesson to be drawn from this except that findings need to be replicated, especially phase 2 findings," he said.

He is less concerned about the lack of definitive information about the drug's mechanism. It's not unusual in drug development, he says, to "stumble onto some kind of efficacy and figure out why later."

Dr. DeKosky agreed on this point, saying he doesn't "give much traction" to the lack of solid information on mechanism. "My own view of it is, if a drug works, you may have to chase for a while to find the mechanism, but that doesn't mean you shouldn't move it forward and find out if it will be functionally good even if you're not sure exactly why it works."

Investigators Respond

In response to some of this discussion of the phase 2 vs the phase 3 results, Dr. Doody told Medscape Neurology that she was involved in the phase 3 and the phase 2 studies and has been privy to subsequent review of the new data.

"I can only talk about what's currently in the public domain, but of course we're all really sad for the patients," Dr. Doody said. "That's really the biggest feeling anybody had about it. But on top of that, we've got a study that's so completely different in outcomes, it bears a lot of investigation, and how much can you do in a couple of days before you're required to get a press release out? The companies did their absolute best and I hope will continue to do so, but there are a lot of differences between the 2 studies and they will all have to be looked at."

There may in the end be a dominant theory, she added, but it will only be a theory. "I don't know of any failed study where you say, 'this is clearly the one and only reason'."

She finds it unreasonable, though, to think the only reason is geography, that the phase 2 study was conducted in Russia. "There are so many differences between the 2 studies, including geography, but since we don't have any a priori reason for picking one of the reasons as being more important, we have to consider them all."

Other differences include trial size; the variability introduced by a more uniform population in the phase 2 trial vs a multinational, multiethnic cohort in phase 3; different points in time; different healthcare systems; and different backgrounds for the investigators. "There are so many differences between the 2 trials that any of them, or a combination of them, could turn out to be very important."

And in the end, the very positive results in phase 2 were just that, phase 2 results. "Everything in phase 2 pointed to efficacy, but that's all it pointed to," Dr. Doody pointed out. "As I said many times when I was asked (about the earlier results), what you look for in phase 2 is a signal of efficacy. You can't say that much about the signal if you see one — you either agree there is a signal or there isn't a signal, and if there is, we go to phase 3.

"There's a reason why regulatory authorities ask for more than 1 study; I agree with them," she added, "but there's also a reason that companies do more than 2."

She disagrees with Dr. Scheider, though, about the importance of the placebo effect, which she called "quite notable" in the phase 3 data. "I think in this study we were expecting both improvement over baseline on drug and decline in the placebo. You had to plan for placebo decline in your study design, and it didn't happen."

Still, further examination of the data may or may not yield any explanation for this, Dr. Doody added.

"Some people think that Alzheimer's is behaving a little differently in the modern world, and if that's true, that's great, because patients may have more time, more time for us to find new treatments before they decline."

Chronic Underinvestment

The Alzheimer's Association expressed regret at the findings but were quick to move on and took the opportunity to point to the wider issue of general underfunding in AD research.

"The Alzheimer's Association is disappointed to learn of the negative results from the phase 3 clinical trial of latrepirdine," William Thies, PhD, the association's chief medical and scientific officer, said in a statement. Nonetheless, the statement adds, the association "remains optimistic about the future prospects for better Alzheimer's treatments and prevention strategies," pointing out that several dozen other compounds are still in the pipeline that address the disease from a variety of angles.

To get better diagnosis, treatments, and prevention of AD, they write, there is an urgent need to address the "chronic underinvestment" in Alzheimer's research and that people participate in AD trials. At the upcoming International Conference of Alzheimer's Disease in Hawaii in July this year, the Alzheimer's Association plans to launch a tool meant to match people with AD with trials for which they may be candidates.

The association was also meeting this week in Washington, DC, for the Alzheimer's Action Summit to encourage legislators to increase research funding.

Dr. DeKosky disclosed that he has been a consultant on drug development with Esai, Psychogenics, Merck, Wyeth, and Novartis and an advisory board member for Pfizer (Academic Health Center-Pharma Relationships). Dr. Schneider disclosed that he has consulted for Medivation and Myriad and received research support from Myriad. Dr. Gandy disclosed that he has served as an adviser or consultant to Amicus, Epix, Diagenic, Wyeth Pharmaceuticals Inc, and Elan Pharmaceuticals Inc. He has received grants for educational activities from Forest Laboratories Inc. Dr. Green disclosed that he has received grants for educational activities from Myriad Pharmaceuticals.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: