Results tagged “Alzheimer's disease” from Pathophilia

One part of a much-publicized 3-part Innogenetics assay appears to predict dementia in Parkinson disease. According to an early-release, prospective, cohort study (N = 45) from U Penn, low and declining levels of beta amyloid 1-42 in spinal fluid are associated with cognitive impairment in the disorder. There is no association, however, between dementia and CSF levels of total tau, phosphorylated tau, or any tau-to-beta-amyloid ratios. In fact, adding any tau measurement to the mix actually diminishes the predictive value of the beta-amyloid measure in PD patients.

During the 1-3-year PD study, baseline beta amyloid 1-42 levels of less than 192 pg/mL, which are considered diagnostic for Alzheimer disease, significantly increased the risk of dementia. On the basis of their data, the authors concluded that a PD patient with a low CSF beta-amyloid level would "progress from essentially normal cognition to a level consistent with [PD-associated dementia] within a 2-year period of follow-up." CSF beta-amyloid levels also correlated with apolipoprotein E ε4 carrier status among the small number of PD patients (15) who carried the AD risk factor.

Last week, the full 3-part Innogenetics assay (which measures CSF levels of beta amyloid 1-24, total tau, and phosphorylated tau) was promoted to predict AD in people with mild cognitive impairment (see Alzheimer Assay Deserves Qualified Embrace). In this study, there was a distinctive pattern change as MCI progressed to mild AD and then to advanced disease: CSF beta amyloid levels fell while tau levels rose. The former event appears to be true in PD patients who develop dementia. Lower levels of CSF beta amyloid reflect the underlying amyloid pathology of the dementing aspect of the disease and may be due to the sequestration of beta amyloid into developing amyloid plaques. At least that's the speculation.

The Penn authors concluded that the CSF level of beta amyloid 1-42 "may provide clinically useful prognostic information, particularly if combined with other risk factors for cognitive impairment in PD." Dementia may occur in the disorder in up to 80% of patients.

A Belgian company, Innogenetics was acquired by Solvay Pharmaceuticals in 2008 and is listed on Euronext Brussels (ticker, INNX)

Image of 1886 drawing of PD patient by neurologist Sir William Richard Gowers from Wikimedia Commons.

Eli Lilly announced today that it is pulling the plug on semagacestat, a gamma secretase inhibitor in phase 3 development for Alzheimer disease. Preliminary results from 2 large, ongoing, multinational phase 3 trials* showed that AD patients taking semagacestat actually performed worse on measures on cognition and activities of daily living than placebo-treated patients. Semagacestat treatment was also associated with a relatively increased risk of skin cancer (although the company's press release did not indicate what kind of skin cancer).

Lilly says it is still moving forward with its phase 3 clinical trials of another anti-beta-amyloid compound in AD: solanezumab, a monoclonal antibody. The drug works differently than semagacestat, an inhibitor of the enzyme (gamma secretase) that produces beta amyloid. The company also assures everyone that it will publish the results of its semagacestat trials.

By my estimation, Lilly is/was the frontrunner in clinical AD trials. Late-phase study of another anti-amyloid mAb, Pfizer/JNJ's bapineuzumab, has been complicated by underwhelming efficacy results and vasogenic brain edema.

* IDENTITY and IDENTITY-2.

Photograph of atrophied brain from person with AD: National Institute on Alcohol Abuse and Alcoholism.

A pattern of 3 markers in CSF—beta amyloid 1-42, phosphorylated tau, and total tau—appears to strongly predict Alzheimer disease in people with mild cognitive impairment (MCI), according to a newly published and highly publicized study from the US Alzheimer's Disease Neuroimaging Initiative. But the study findings, which can be found in the Archives of Neurology (subscription required), do not necessarily mean that the CSF assay should be incorporated immediately into clinical care—despite the fact that 1) the authors are lobbying hard for the test to be included in the diagnostic criteria for AD and 2) an accompany editorial urges clinical use of the CSF assay with the wince-inducing title, "Sharpen That Needle."

The "AD signature" of low beta amyloid 1-42 and elevated tau levels was found in 90% of AD patients (n = 102), 72% of MCI patients (n = 72), and 36% of cognitively normal subjects (n = 114). Consequently the sensitivity of the test is 90% (the rate of positivity in AD patients), and the specificity is 64% (the percentage of cognitively normal subjects who don't show the AD CSF pattern [ie, 100% - 36%]).

Among a subset of 68 people with autopsy-confirmed AD, 64 (94%) showed the AD CSF pattern, which correlates well with the 90% sensitivity measure in clinically diagnosed AD patients (via the MMSE). In another subset of 57 patients with MCI who developed AD during 5-year follow-up, all exhibited the AD CSF pattern for a sensitivity of 100%.

Rather than view a positive CSF assay in cognitively normal people as a failing of the test (ie, it has limited use in normal subjects), the authors conclude that these people are at risk of AD. This conclusion may well be true, but it seems that further study is warranted before such a bold statement can be made. The authors did find an "enrichment" of the apolipoprotein E ε4 allele, a recognized genetic risk factor for AD, in this population, but the finding was not uniform.

At least ABCNews, unlike many other news sources, found medical commentators who advise against the whole-hearted embrace of this assay. Cliff Saper, who is one of the most level-headed neurologists I've ever met, said there is no reason to perform the test until there is a successful treatment for AD. "This test shows up positive in presymptomatic individuals, and Alzheimer's disease is a common disorder," Saper was quoted. "The main value would be to detect [Alzheimer's disease] in atypical cases." For what it's worth, I agree.

The CSF assay currently appears to be most useful in the context of research—as an additional tool to reinforce the diagnosis of MCI or AD in clinical studies. Assay results may also inform the use of investigative therapies at different stages of cognitive decline. One of the most intriguing findings from the study is that there appears to be a distinctive pattern change as MCI progresses to mild AD and then to advanced disease: CSF beta amyloid levels fall while tau levels rise. Anti-beta-amyloid therapies like bapineuzumab, may provide relatively greater benefit in MCI or very early AD, while anti-tau therapies may have their place in more severe dementia. 

In the meantime, Innogenetics, which employs or employed several of the authors, provides the tripartite CSF assay (Inno-Bio AlzBio 3) for "research use only." Information on the cost of the assay, after several Google searches, remains elusive. 

CSF = cerebrospinal fluid.

Photograph of atrophied brain from person with AD: National Institute on Alcohol Abuse and Alcoholism.

If you're looking for a break from the FDA's much-covered review of Avandia,* try the refreshing In Vivo Blog. The latest from the armchair bloggers (talk about redundant phrasing, and heh, takes one to know one) is their follow-up of biomarker data from JNJ and Pfizer on bapineuzumab, the anti-amyloid mAb in phase 3 development for Alzheimer disease.

Presented at the Alzheimer's Association International Conference on Alzheimer's Disease 2010 in Honolulu (Altoona wasn't available?), the biomarker data don't concern a reduction of amyloid with bapineuzumab treatment. Instead the data (and they're phase 2 clinical data) show a drop in CSF levels of P-tau, at least when they're pooled.

The examination of P-tau, as a "downstream" biomarker, may be scientifically valid in studies of anti-amyloid therapies. P-tau is certainly an important marker of AD pathology generally, and investigators seem to be more enthusiastic about direct, anti-tau therapies than anti-amyloid therapies. But the tau angle is also commercially savvy. It's a way of bolstering interest in the continued investigation of the leading anti-amyloid drug, which has provided underwhelming clinical results to date.

* heartwire.org provides the most detailed description of yesterday's contentious agency meeting, IMO--just short of live blogging it.

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism.

CSF = cerebrospinal fluid; mAb = monoclonal antibody.

Yesterday's favorable news of Baxter's Gammagard (IV Ig) in Alzheimer disease (and the obligatory buzz) should be approached cautiously. Here's why.

First: The results are from a phase 2 study, not an all-important phase 3 study (two of which are typically required for FDA approval). As is common in a phase 2 study, different doses of the study drug were assessed.

Second: The study was small, enrolling only 21 subjects (14 with AD).

Third: The study was placebo controlled for a relatively limited time—6 months. Afterward all patients received IV Ig at various doses for 12 months.

Fourth: The difference in the ADCS-CGIC score change between treatment groups, while statistically significant, was not huge—1.36 points (-0.64 vs -2.0).

Fourth: The significant difference between the mean ADAS-cog scores at endpoint may be due to the substantial change in the mean score of control patients—15 points. It'd be nice to know when the bulk of this 15-point change in the control patients occurred: during placebo treatment or during IV Ig treatment?

Fifth: The results have not been peer reviewed.*

Perhaps the most intriguing result of Baxter's phase 2 AD study is the statistically significant associations among IV Ig treatment, brain atrophy on MR images, and clinical scores. Again, however, the study was small, and the outcome warrants confirmation in phase 3 investigations, which are ongoing or planned.

The rationale for administering IV Ig in AD is to provide nonspecific anti-amyloid antibodies. 

ADAS-cog = Alzheimer's Disease Assessment Scale, cognitive subscale; ADCS-CGIC = Alzheimer's Disease Cooperative Study, Clinical Global Impression of Change.

* While the Baxter press release indicates that these data were presented yesterday at the ongoing annual meeting of the American Academy of Neurology, I--for the life of me--cannot find the related abstract. The closest presentation from the same investigators is "Neuropsychological outcomes following 18-months of uninterrupted intravenous immunoglubulin (IVIg) treatment in patients with Alzheimer's disease (AD)." The data from this related trial are to be presented this evening.

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism.

Addendum: D'oh! Here's the AAN abstract (found in late-breaking science, which—Hmm—is to be presented as a poster this evening. It is Wednesday, isn't it?). Relkin N et al. Intravenous immunoglobulin treatment decreases rates of ventricular enlargement and cognitive decline in Alzheimer's disease. P03.294.

The 6-month data were presented at the 2008 annual meeting of the AAN. Among 23 completers, 19 had "fully evaluable data." The difference between the change in the ADCS-CGIC score of IV Ig-treated patients and that of placebo-treated patients was statistically significant (IV Ig, +0.25; placebo, -1.25; absolute difference, 1.50); however, the mean change in the ADAS-cog score was not (IV Ig, -0.38; placebo, +2.61). 

So most of the change in the ADAS-cog score of control patients at 18 months occurred while they were receiving IV Ig. It seems flukish that 6 months of upfront IV Ig therapy would provide that much neuroprotection in the treated group, whose ADAS-cog score changed by only 5 points at 18 months.

While reviewing trials of novel agents for Alzheimer disease at the NIH database, one thing is readily apparent: There are few candidates in the late-stage pipeline.

With 2 drugs in phase 3 development, Lilly is the clear frontrunner. Its sole competition appears to be a green tea extract, sunphenon EGCg, which is being assessed in 1 actively recruiting, phase 2/3 AD trial in Germany.*

Lilly's drugs of promise are semagacestat (LY450139), a gamma-secretase inhibitor, and solanezumab (LY2062430), a monoclonal antibody. For what it's worth (and that may not be much), both are intended to disrupt or break up the amyloid plaques of AD (like the woebegone bapineuzumab).

Two phase 3 company-sponsored trials of semagacestat are listed in the NIH database (here and here). The first trial, called IDENTITY, will be completed in June of next year. And two phase 3 trials of solanezumab, EXPEDITION and EXPEDITION 2, are recruiting patients with mild-to-moderate AD (here and here). These trials will not be completed until 2012.

A phase 2 safety trial of semgacestat in 51 patients with mild-to-moderate AD revealed possible "concerns" about skin and hair reactions. Although plasma levels of amyloid beta dropped significantly with semagacestat, there were no treatment-related changes in CSF levels and no between-group differences in cognitive measures; however, the trial was too short (14 weeks) to adequately assess any cognitive benefits. A more recently published study showed that single doses of semgacestat reduce amyloid beta production, but not clearance, in a dose-dependent fashion.

* The trial is expected to be completed in April of 2011.

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism.

The Alzheimer's Association has a new report on the large and growing effect of Alzheimer disease and other dementias on Americans. Here's the distillation:

• AD is the most common cause of dementia, affecting 5.3 million Americans (1.7% of the current US population). The vast majority (96%) of those affected are older than 64 years of age.
• Dementia affects women preferentially because they live longer, on average, than men. (Advancing age is the biggest risk factor for AD.)
• In 2006, AD was the 7th leading cause of death among all Americans. In those older than 64 years of age, AD was the 5th most likely cause of death.
• Next year, the first baby boomers will turn 65. In 30 years, all of the estimated 70 million baby boomers will be older than 65. The Association estimates that there will be 1 million new cases of AD by the year 2050.
• People with AD and other types of dementia need a lot of health care, including long-term and hospice care. In 2010, care costs for people with dementia will reach $172 billion. Medicare costs for people with dementia are nearly 3 times higher than those for the nondemented elderly. Medicaid costs are nearly 9 times higher. (Blogger's note: I suspect this is the result of Medicaid-funded nursing home care.)
• Older African Americans and Hispanics are about 2 and 1.5 times more likely, respectively, to have AD and other types of dementia than older whites.* Reasons for the higher risk of AD among these minorities include a greater prevalence of known risk factors for dementia, like hypertension, diabetes, and lower education and income levels. (Ethnically dependent genetic risks have not been identified.) Despite the higher risk of dementia among African Americans and Hispanics, the diagnosis is more likely to be delayed in affected minority patients. Delays in diagnosis mean missed opportunities for treatment, which is more likely to be effective in the early stages of AD.
• Last year, nearly 11 million family members or other uncompensated persons provided 12.5 billion hours of care for people with dementia. Unpaid care was assessed at $144 billion, which is more than the US government (ie, Medicare and Medicaid) spends on people with dementia. Costs of giving care are also measured in emotional and physical stress, declining caregiver health, and lost employment-related productivity. 

* This is the lead of today's press release from the Alzheimer's Association.

Photograph of atrophied brain from person with AD: National Institute on Alcohol Abuse and Alcoholism.

Someday I'm going to learn to capitalize on the overblown expectations and inevitable letdown of AD drug development.

Dimebon (latrepirdine), the nonselective antihistamine that Pfizer shelled out millions for, failed to improve cognition or global functioning in a placebo-controlled phase 3 study of patients with mild-moderate Alzheimer disease. The disappointing results were distributed today by way of press release from the drug's codevelopers, Medivation and Pfizer.

In a double-blind study of 598 individuals with AD in North America, Europe, or South America, Dimebon 20 mg tid was no better than placebo for improving or delaying declines in cognition (as measured with the ADAS-cog) or global functioning (as measured with the CIBIC-plus) at 6 months. Measures of secondary endpoints (eg, activities of daily living) were, likewise, similar between Dimebon- and placebo-treated patients.

As expected, the most frequently reported adverse events with the antihistamine were somnolence and dry mouth. A 5-mg treatment arm and another, large phase 3 safety study revealed consistent drug tolerability. For what it's worth, the phase 3 safety study showed that the drug can be taken with common AD medications, like cholinesterase inhibitors (eg, Aricept; Pfizer) or memantine (Namenda; Forest).

Results from the phase 3 efficacy study were to be used in conjunction with positive results from a phase 2 Russian study to support FDA approval. But, evidently, no more. 

In 2008, Pfizer agreed to pay Medivation $225 million upfront and another $500 million when the drug was approved. The agreement also conferred licensing rights to Pfizer for use of the drug in Huntington disease. Last month, results from a 90-day safety trial of Dimebon in HD showed that the drug may improve cognition.

The price of Medivation stock dropped nearly 30 points on release of the phase 3 trial results, from $40.12 to $12.88.

ADAS-cog = Alzheimer's Disease Assessment Scale-cognitive subscale; CIBIC-plus = Clinician's Interview-Based Impression of Change Plus Caregiver Input.

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism.

The defunct Russian antihistamine Dimebon may (may) improve cognition in Huntington disease, according to newly published results of a short-term, but well-controlled, trial in the Archives of Neurology. However, the efficacy outcome—which was just 1 of 3 efficacy measures made in the trial—was not the primary endpoint of this company-sponsored phase 2 study of the drug's tolerability.

Among 91 ambulatory enrollees with HD,* 90-day treatment with Dimebon, 20 mg tid, significantly improved the mean score on the 30-point MMSE by less than 1 point (0.97; P = .03, vs placebo). No significant differences between Dimebon and placebo treatment were noted, however, on the Unified Huntington's Disease Rating Scale (which assesses motor function, as well as cognition) or the 13-item ADAS-cog. Tolerability of Dimebon, the primary outcome measure, was comparable to that of placebo: 87% and 82% of patients, respectively, completed treatment. No significant treatment differences were noted with regard to the most commonly reported adverse events (eg, falls, headache, and dizziness).

The rationale for studying Dimebon in HD and other neurodegenerative disorders, like Alzheimer disease, is based on the drug's ability in the laboratory to stabilize mitochondrial membranes and possibly promote neuronal stability or even regeneration. According to the study authors, "Abnormal mitochondrial depolarization, which can lead to collapse of the mitochondrial membrane and ultimately neuronal apoptosis, has been implicated in the pathophysiologic progression of HD and other neurodegenerative diseases."

A placebo-controlled study of Dimebon in AD revealed significant improvement in cognition, behavior, and general function at 6 and 12 months. Results of this study evidently motivated Pfizer to give Medivation $725 million in 2008 for worldwide marketing rights to the drug. Corporate-funded phase 3 studies of Dimebon are currently recruiting patients with HD (here) or AD (for instance, here or here).

News of the phase 2 results was associated with a bump in Medivation's share price (but not Pfizer's).

N.B.--Dimebon is the trade name for dimebolin HCl or latrepirdine.

ADAS-cog = Alzheimer's Disease Assessment Scale, cognitive subscale; MMSE = Mini-Mental State Examination.

* Study-eligible patients were allowed to take antipsychotic agents (at stable dosages), but they could not take cholinesterase inhibitors (eg, Aricept) or N-methyl-D-aspartate antagonists (eg, Namenda).

Image of Dimebon structure from Wikipedia.

This is the final post in which randomized, clinical trials of the potentially active ingredients in the branded drink Souvenaid are reviewed. (For background on Souvenaid's development, go here, here, here, and here.) 

The remaining components to be examined in the nutritional supplement are phospholipids, choline, and uridine monophosphate (UMP).

In the case of UMP, a directed PubMed search reveals no randomized, controlled trial of the RNA nucleotide in Alzheimer disease. Therefore the rationale for assessing UMP in cognitive dysfunction appears to extend solely from animal studies (for instance, see a 2008 rat study by anchor author Wurtman, who is also coauthor of the recently published Souvenaid study).

The rationales for giving phospholipids or choline to AD patients are either 1) to bolster the integrity of neuronal membranes by providing the necessary building blocks, or b) in the case of choline, to supply a precursor for the neurotransmitter acetylcholine, which is deficient in AD. 

Barring the negative studies of lecithin,* a potential source of both phospholipids and choline, there are at least 9 randomized, controlled trials (published from 1986 to 2003) of these substances in dementia. Six studies, from the 80s and 90s, assessed the specific phospholipid phosphatidylserine; 3 studies assessed choline alfoscerate or citicoline. Most of these studies enrolled outpatients with mild-to-moderate AD. General study features are tabulated here:

Intervention	N	Duration	Outcome in a Nutshell
Phosphatidylserine
1986	42	6 wk	Trend toward improvement or significantly improved
1986	No abstract	—	—
1992	51	3 mo	Improvement observed in milder disease
1992	33	2 mo	Mixed improvement
1992**	40	6 mo	PET-correlated cognitive improvement
1994**	70	6 mo	Temporary improvement
Acetylcholine precursors
1993	126	?	Significant improvements with choline alfoscerate
1999	30	3 mo	Citicoline associated with improvements in APOE ε4 carriers; effect more pronounced in mild disease and associated with physiologic changes
2003	261	6 mo	Choline alfoscerate improved cognition or delayed cognitive decline

Cumulative data from the relatively small, controlled studies of phosphatidylserine in AD suggest that whatever cognitive improvements exist, they are short lived.

The most compelling support for the use of choline (ie, choline alfoscerate) comes from a sizeable, randomized, placebo-controlled, 6-month study performed at the Instituto Nacional de la Senectud in Mexico City. Cognitive assessments, performed with the ADAS-cog and other well-recognized tests, showed either a significant delay in cognitive decline or actual cognitive improvement. 

Nevertheless, these impressive results must be replicated. 

ADAS-cog = Alzheimer's Disease Assessment Scale, cognitive subscale; PET = positron emission tomography.

* A directed PubMed search reveals at least 5 studies of lecithin in AD, published from 1981 to 1987. Among the 3 small placebo-controlled studies in which an abstract is provided, none showed clinical improvement with supplementation (treatment duration, from 10 weeks to 6 months), despite a rise in plasma choline levels. These search results omit at least 8 trials of the defunct cholinesterase inhibitor tacrine, in which lecithin was provided as standard treatment.

** Treatment options were phosphatidylserine or pyritinol, a vitamin B6 analog.

Photograph of atrophied brain from person with AD: National Institute on Alcohol Abuse and Alcoholism.

Vitamin E and other antioxidants are included in the branded, nutritional drink Souvenaid, which is being developed as a treatment for Alzheimer disease. However, well-controlled clinical data to justify their use in AD or mild cognitive impairment (MCI) are scarce, and positive trial data require qualification.

The most prominent (and confounding) trial in which vitamin E was assessed in AD was the Alzheimer Disease Cooperative Study, published in 1997 in the NEJM. In this multicenter US study, 341 patients with moderately severe AD received randomized selegiline (10 mg/d), vitamin E (2000 IU/d), both, or placebo in a double-blind fashion.

In the unadjusted analysis, there were no significant differences among the 4 treatment arms with respect to the combined primary outcome of death, institutionalization, loss of the ability to perform basic ADLs, or severe dementia at 2 years. However, the trial results had to be statistically accommodated, because the mean baseline MMSE score of placebo-treated patients, despite randomization, was significantly higher than those of other enrollees. "[A]nd this variable," the authors predictably concluded, "was highly predictive of the primary outcome."

And so an adjusted analysis was performed, in which the investigators observed significant delays in the combined primary outcome with all 3 randomized treatments—selegiline, vitamin E, and both—when compared with placebo.

But these data are at odds with results from a 3-part Italian study of vitamin E and donepezil (Aricept; Eisai/Pfizer), published sequentially in 2001, 2002, and 2003. Among 40 patients with moderately severe AD who received randomized vitamin E (2000 IU/d), latencies of P300 recordings (a very rough, electrophysiologic indicator of cognition*), and cognitive scores declined significantly at 26 weeks (when compared with donepezil treatment [5-10 mg/d]).

These findings are supported by follow-up results, in which P300 latencies and cognitive test scores deteriorated with vitamin E supplementation at 3 months in patients with mild AD (n = 30). Cognition also significantly declined in subjects with moderate-to-severe AD (n = 30) who received vitamin E. Other electrophysiologic indicators of cognition worsened in a smaller and much briefer (ie, 30-day) study of vitamin E and donepezil in patients with "mild dementia."

To assess the potential benefit of vitamin E supplementation to prevent AD, investigators at the Mayo Clinic conducted a randomized, double-blind, placebo-controlled study of 769 patients with amnestic MCI. The results, published in a 2005 issue of the NEJM, showed no significant delay in the development of AD at 3 years with vitamin E (2000 IU/d) or donepezil (10 mg/d).** The annual rate of progression from MCI to AD was 16%, a rate that has been subsequently confirmed. 

The most recent trial assessing vitamin E in AD was published last year. In this small Spanish study, 57 patients with AD were assigned to placebo or vitamin E (800 IU/d) for 6 months. Among the 33 study completers, investigators correlated a controversial measure of oxidative stress (reduced glutathione) with maintained cognition in vitamin E-treated subjects. The authors concluded that "supplementation of AD patients with vitamin E cannot be recommended without determination of its antioxidant effect in each patient."

The American Psychiatric Association, as of 2007, no longer recommends vitamin E for the treatment of dementia because of limited efficacy data and potential safety concerns. The latter relate to reports of increased cardiovascular and cerebrovascular mortality. However, these risks may not be relevant in AD patients, according to a recent, longitudinal survival study.

So these clinical data for vitamin E in AD or MCI leave us firmly nowhere.

But there are other potential antioxidants in Souvenaid: namely, selenium and vitamin C. What are the clinical data to support their use in dementing disorders?

There is one semi-controlled Polish study (1999) in which selenium (100 µg every other day) was compared with placebo and some proline-rich substance isolated from sheep colostrum [yeesh] in patients with mild or moderate AD (N = 46). Psychiatrists blinded to treatment determined that cognition stabilized in 13 of 15 selenium-treated patients at 1 year, vs 8 of 16 placebo-treated patients. (The sheep colostrum product reportedly improved or stabilized cognition in all 15 treated patients.)

Controlled assessments of vitamin C, aka ascorbic acid, in AD are virtually absent, at least by my search. An Italian AD study, published in 1995, used ascorbic acid as a reference standard. In a later Czech study, 50 mg of ascorbic acid was given daily to all AD enrollees. 

* Increased latencies suggest delays in neuronal transmission.

** However, the rate of progression to AD at 12 months was reduced with donepezil treatment, and the AChEI prevented full-blown dementia at all time points in carriers of 1 or more APOE e4 alleles--a known genetic risk factor for late-onset AD. More recent, controlled studies of donepezil in patients with amnestic MCI suggest that the drug delays the progression to AD among depressed individuals but has only a modest effect on cognition in a more inclusive population with MCI.

AChEI = acetylcholinesterase inhibitor; ADLs = activities of daily living; MMSE = Mini-Mental State Examination; APOE = apolipoprotein E.

Ingredients in the proprietary nutritional drink Souvenaid, which its developers claim work synergistically to rejuvenate neuronal synapses in Alzheimer disease, include several B vitamins. (For necessary background on Souvenaid development and negative trials of omega-3 fatty acids [another ingredient in Souvenaid] in AD, go here and here.)

The rationale for using vitamins B6,* B9 (folic acid), and B12 specifically in AD is to reduce levels of homocysteine, an independent risk factor for AD. However, a directed PubMed search provides little evidence of the benefit of oral B-vitamin supplementation in dementing disorders. A total of 4 randomized, double-blind, placebo-controlled trials were discovered, in which 1 or more of these vitamins were assessed in patients with AD.

Published in 2007, a Thai study enrolled 89 AChEI-treated individuals with mild-to-moderate AD, all of whom had normal folic acid and vitamin B12 levels. The daily, oral administration of B12 500 mg and a multivitamin containing B6 (5 mg) and folic acid (1 mg) provided no statistically significant cognitive benefits at 26 weeks (measured by using the 11-item ADAS-cog); nor was the ability to perform activities of daily living (ADLs) affected. Negative clinical benefits were observed despite the reduction of homocysteine levels in treated AD patients.

A smaller UK pilot study, reported in 2008, revealed that folic-acid supplementation in 57 AChEI-treated AD outpatients improved the ability to perform instrumental ADLs and social behavior (a combined outcome) at 6 months; but daily folic acid did not significantly alter cognition (measured with the MMSE).

The same year, JAMA published results of an 18-month multicenter US study of 409 individuals with mild-to-moderate AD and normal folic-acid, B12, and homocysteine levels. Randomized, oral supplementation with B6 (25 mg/d), folic acid (5 mg/d), and B12 (1 mg/d) reduced homocysteine levels but had no effect on cognitive decline (measured with the ADAS-cog) in 340 "completers." Curiously depression was more likely to occur in supplemented patients.

Finally last year a multicenter Dutch study was printed, in which "vascular care" was assessed in 130 outpatients with AD and evidence of cerebrovascular disease on brain images. The combined, randomized treatment of aspirin, folic acid, and B6 provided no statistical benefits with respect to measures of cognition, dementia-related disabilities, or behavioral problems at 2 years. Rates of institutionalization were also comparable between patients who received vascular care and those who didn't. 

ADAS-cog = Alzheimer's Disease Assessment Scale, cognitive subscale; AChEI = acetylcholinesterase inhibitor (eg, Aricept [donepezil]); MMSE = Mini-Mental State Examination.

* It should be noted the B6 supplementation is associated with a dose-dependent peripheral neuropathy (for instance see Berger et al and Parry and Bredesen). The Food and Nutrition Board of the Institute of Medicine recommends that daily B6 intake should not exceed 100 mg.

Photograph: Atrophied brain of person with AD from National Institute on Alcohol Abuse and Alcoholism.

Developers of the proprietary drink, Souvenaid, claim its ingredients work synergistically to restore neuronal synapses in diseases like Alzheimer disease (for necessary background on Souvenaid clinical development, see yesterday's post). However, many of the individual components in the formula have been shown to be ineffective, or only marginally effective, in AD. Among these ingredients are omega-3 fatty acids—specifically docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA). One rationale for using DHA and EPA supplementation in AD is based on their potential to reduce disease-associated inflammation.

A directed PubMed search reveals exactly 1 randomized, double-blind, placebo-controlled study of DHA and EPA in AD. The 6-month study, with a 6-month open-label extension, was performed at the Karolinska Institute between 2001 and 2004 in 174 patients with mild-to-moderate disease who were receiving acetylcholinesterase-inhibitor therapy (eg, Aricept). Data were collected on cognitive decline, neuropsychiatric symptoms, weight gain, and plasma markers of inflammation and were parsed into 4 articles, published from 2006 to 2009.

Daily DHA (1.6 g) and EPA (0.6 g) supplementation did not affect cognitive decline (measured with the MMSE and ADAS-cog scales) at 6 months; however, in the subgroup with very mild AD (n = 32), supplementation was associated with a statistically significant reduction in the decline of the MMSE score. (This observation likely informed the enrollment of patients with very mild AD in the Souvenaid trial.)

Nor did DHA/EPA supplementation affect neuropsychiatric symptoms, the ability to perform activities of daily living (ADL), or caregiver burden at 6 months. In detailed subgroup analyses, however, supplementation reduced agitation in APOE ε4 carriers* (n = 125) and depression in APOE ε4 noncarriers (n = 49). Make of that what you can; I don't make much of it.

On the other hand, weight increased significantly with supplementation at 6 months and continued to increase for both treatment groups in the open-label extension phase. APOE ε4 noncarriers and high DHA plasma levels were independently associated with weight gain, reported the authors.

As expected, plasma levels of DHA and EPA increased with supplementation (n = 23). In addition, they were associated with reduced levels of some inflammatory cytokines (eg, interleukin-6).

* A genetic risk factor for late-onset AD.

ADAS-cog = Alzheimer's Disease Assessment Scale, cognitive subscale; APOE = apolipoprotein E; MMSE = Mini-Mental State Examination.

Image of cod liver oil capsules from Wikipedia.

Last week, various mainstream news sources reported the positive results of a randomized, double-blind, placebo-controlled, multicenter study of Souvenaid, a proprietary nutritional drink, in people with mild Alzheimer disease. The objective of the trial, printed in the latest issue of Alzheimer's and Dementia, was to restore neuronal synapses by providing "rate-limiting precursors for membrane phosphatide synthesis, such as...uridine, omega-3 polyunsaturated fatty acids, and choline," which Souvenaid contains. Citing rodent studies, the authors claimed, "These nutrients synergistically increase brain levels of phosphatide molecules that comprise the bulk of synaptic membranes, and brain levels of specific synaptic proteins, suggesting that they also increase synapse formation" [emphasis added].

Yes, it'd be nice to have a boosted smoothie that wards off cognitive decline, but...

To the credit of at least one news source, mainstream coverage of the Souvenaid study was tempered by reports of AD experts' "disdain" for the trial and its results.

Among the criticism:

• The relatively short length of the study, 12 weeks. The generally accepted minimum duration of an AD trial is 26 weeks; most extend to 1 year or longer.*
• The lack of benefit when measuring cognition with standard tests, like the ADAS-cog.
• The potential conflicts of interest among the authors, given their intimate connections to the company that makes Souvenaid. Notably coauthor Richard J. Wurtman, of MIT, owns patents on "uridine, omega-3 fatty acids, and choline to treat brain diseases by enhancing synapses," according to the study disclosures.

A review of the study article reveals additional problems.

• Because one unnamed study site failed to comply with ICH-GCP guidelines, data from 13 individuals (~6% of the study population) were excluded from the efficacy analysis.
• At baseline, approximately 40% of enrollees scored 0, the lowest score, on the delayed verbal-recall test (Wechsler Memory Scale-revised), one of the 2 primary outcome measures. This skew in the study population required the authors to use a nonparametric analysis (eg, use general categories of better, worse, and unchanged), instead of the planned parametric assessment (that is, use mean score changes in the study arms). Furthermore it's not clarified that the percentages of 0 scorers at baseline were similar in the 2 treatment groups. If they were not, then the potential for enrollees to decline on the basis of assigned treatment would certainly be affected.

Finally, although the authors claim synergistic effects of Souvenaid's active ingredients on synapse restoration, many of the individual ingredients in the formula have been shown to be ineffective (or only marginally effective) in AD. Watch for upcoming posts at the Pathophilia blog that review these AD studies.

* A 12-week blinded extension phase was conducted (n = 161). No benefit was seen with respect to the 2 primary outcome measures, but immediate verbal recall was significantly better in the group receiving Souvenaid.

ADAS-cog = Alzheimer's Disease Assessment Scale, cognitive subscale; GCP = good clinical practice; ICH = International Conference on Harmonisation; MMSE = Mini-Mental State Examination.

Photograph: Atrophied brain of person with AD from National Institute on Alcohol Abuse and Alcoholism.

Addendum: On the basis of the 12-week proof-of-concept study, the makers of Souvenaid are recruiting enrollees for a multicenter US study, called S-Connect. The trial, however, is not registered in the NIH database. The 40 US trial sites, 3 of which appear to be university affiliated, are listed here.

About 16% of people with mild cognitive impairment (MCI) progress to Alzheimer disease during the course of a year, according to new data from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Progression from MCI to AD, which correlated with baseline CSF levels of amyloid beta-42, occurred despite the prevalent use of available antidementia therapies (~50% of subjects).

ADNI investigators followed up more than 800 elderly subjects (mean age, ~75 years) with normal cognition, MCI, or AD. Baseline data were notable for remarkably high mean verbal IQ scores and education levels among all 3 groups and relatively high percentages of APOEε4 carriers among individuals with MCI or AD.

Baseline Feature	Normal   (n = 229)	MCI (n = 398)	AD (n = 192)	P Value
Mean education, y	16 ± 2.9	15.7 ± 3.0	14.7 ± 3.1	<.001*
Estimated mean premorbid verbal IQ	120	116	114	—
APOEε4 carriers, %	26.6	53.3	66.1	<.001*
Mean MMSE score	29.1 ± 1.0	27.0 ± 1.8	23.3 ± 2.1	<.001**
Mean ADAS-cog score	6.2 ± 2.9	11.5 ± 4.4	18.6 ± 6.3	<.001**
AChEI, %	0	43.7	84.9	—
Memantine, %	0	10.8	47.4	—
Combined therapy, %	0	8.8	40.6	—
Mean CSF Aβ-42 level, pg/mL	206 ± 5	164 ± 4	143 ± 4	<.001**

* Control vs MCI amd MCI vs AD.
** Control vs MCI; control vs AD; and MCI vs AD.

On the basis of a battery of cognitive and neuropsychological tests, the estimated rate of conversion from normal cognition to MCI was 1.4% during 12 months, and that from MCI to AD was 16.5%. Eight individuals with MCI reverted to normal cognition, and 2 subjects with AD reverted to MCI at 1 year. 

Among individuals with MCI, scores on a range of assessments (eg, MMSE, ADAS-cog, clock drawing) could be used to predict progression to AD. Lower CSF levels of amyloid beta-42, which did (and are known to) correlate with higher ADAS-cog scores, were also associated with cognitive decline in normal people and those with MCI. 

In July, the ADNI reported that structural abnormalities on MR images are more useful than known CSF biomarkers for discriminating among people with AD, amnestic MCI, or normal cognition, but that the use of MR imaging and CSF analysis is more diagnostically helpful than each measure individually.

AChEI = acetylcholinesterase inhbitor; ADAS-cog = Alzheimer's Disease Assessment Scale, Cognitive Subscale; APOE = apolipoprotein E; CSF = cerebrospinal fluid; MMSE = Mini-Mental State Examination; MR = magnetic resonance.

Photograph: Atrophied brain of person with AD from National Institute on Alcohol Abuse and Alcoholism.

No. 6: Big Trouble in the AD Pipeline

First suggested by last year's news, targeting amyloid buildup in the brains of people with Alzheimer disease doesn't seem to be the way to go.

In April, Elan and codeveloper Wyeth (now a part of Pfizer) announced that the highest dosage of bapineuzumab, an anti-amyloid mAb that Elan sold to JNJ in September,* would be discontinued in ongoing phase 3 trials. The reason: vasogenic brain edema in phase 2 assessment.

The peer-reviewed, phase 2 dose-ranging trial,* which was finally published in Neurology in September, revealed vasogenic edema in nearly 10% of bapineuzumab-treated enrollees with mild-to-moderate AD (n = 12/124). Vasogenic edema, possibly due to the effect of bapineuzumab on perivascular amyoid, was dose-dependent and more common in subjects who were ApoE4 carriers. Moreover, data supporting the efficacy of the drug were underwhelming, despite manipulation of the primary outcome and the use of a modified population for analysis. The authors concluded that the "trial provides insufficient evidence to support or refute a benefit" of bapineuzumab. Nevertheless, phase 3 study of the drug continues.

Then, 2 short weeks ago, Elan announced the deaths of 9 subjects in a company-sponsored phase 2 trial of another AD drug candidate, scyllo-inositol. The cause or causes of the deaths were not provided, but Elan (along with codeveloper Transition Therapeutics) discontinued the highest-dosage arms (1000 and 2000 mg bid) in development. All that's left now is the hope of blocking the accumulation of soluble amyloid plaques with a measly 250 mg of scyllo-inositol twice daily.

mAb = monoclonal antibody.

* JNJ bought Elan's AD immunotherapy program for $1.4 billion.

** The phase 2 results were originally made public in Elan's less-than-candid July 2008 press release, which the SEC is investigating.

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism. 

After 13 years as the number-one prescription treatment for Alzheimer disease, Aricept—a branded cholinesterase inhibitor that generated nearly $1.9 billion in US sales this year for its Japan-based developer Eisai and US comarketer Pfizer—is now (officially) available generically in the United States.

Yesterday the FDA announced the approval of a generic formulation of Aricept (or more correctly, donepezil) in orally disintegrating tablets. The generic manufacturer is the privately owned Mutual Pharmaceutical (URL Pharma), which will produce the drug in 5- and 10-mg strengths.

Other branded, FDA-approved treatments for AD include the NMDA-receptor antagonist memantine (Namenda; Forest) in tablets or solution and the cholinesterase inhibitors galantamine (Razadyne; Ortho-McNeil) in extended-release capsules, tablets, or solution and rivastigmine (Exelon Patch; Novartis). The patent for galantamine, originally marketed as Reminyl in the United States, expired in December 2008, and the drug is available generically in its various formulations.

According to the FDA's Orange Book, the following patent and exclusivity expirations apply to Namenda and Exelon*:

Drug	FDA Approval	Patent Expiration	Exclusivity Expiration
Namenda tablets	Oct 16, 2003	Apr 11, 2015	Oct 16, 2008
Namenda solution	Apr 18, 2005	Apr 11, 2015	Oct 16, 2008
Exelon patch	Jul 6, 2007	Aug 14, 2012	Jul 6, 2010

(Differences between a drug's patent and exclusivity are explained by the FDA here.)

On November 24, Eisai announced that the FDA accepted for review its NDA supporting the use of a 23-mg extended-release tablet version of Aricept in AD. The NDA is based on a head-to-head study (N > 1400) between the new formulation and the old Aricept 10-mg tablet. This study, apparently, has not been published yet in a peer-reviewed journal. In its 2009 report, Eisai also reveals that it is developing transdermal patch and jelly formulations of Aricept.

Very recent controlled studies of immediate-release Aricept in patients with amnestic mild cognitive impairment (aMCI) suggest that the drug delays the progression to AD among depressed individuals, but Aricept had only a modest effect on cognition in a more inclusive population with MCI.

NDA = new drug application; NMDA = N-methyl-D-aspartate.

* Rivastigmine is generically available in capsules; however, this formulation is no longer marketed by Novartis. It is also reported that Forest applied for a "patent term restoration" for Namenda, which would extend protection until September 2013.

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism. 

Nine deaths in an Elan-sponsored phase 2 trial of oral scyllo-inositol (aka ELND005 or AZD-103) in Alzheimer disease have led to the discontinuation of the highest-dose arms in clinical development. The cause (or causes) of the deaths was not provided in today's company press release or by news sources, like Reuters. The investigational drug is intended to prevent the buildup of amyloid in the brains of AD-affected patients.

According to the NIH registry, the trial (which is conducted in collaboration with Transition Therapeutics) is a randomized, double-blind, multicenter US and Canadian study that was comparing 3 doses of the experimental agent with placebo in more than 300 individuals with mild-moderate AD.* Elan states that the study completed enrollment in October; the study duration is 18 months (with results anticipated in May of next year, according to the NIH database). The discontinued dosages are 1000 and 2000 mg bid in both the blinded, phase 2 study and an open-label trial. All patients assigned to scyllo-inositol treatment will now receive the lowest dosage, 250 mg bid.

In August, Elan and Wyeth announced that the highest dosage of its anti-amyloid mAb, bapineuzumab, would be discontinued in 2 ongoing phase 3 studies of ApoE4 noncarriers (40%-70% of AD patients), because of the risk of vasogenic edema.

In September, JNJ closed its highly publicized buyout deal with Elan, in which JNJ acquires the rights to Elan's AD immunotherapy program (which Elan shared with Wyeth). The program includes the development of bapineuzumab but not scyllo-inositol.

A couple of previous Pathophilia posts about Elan's troubles:

• Despite Manipulation, Phase 2 Bapineuzumab Study Still Disappoints
• SEC Eyeballs Elan

mAb = monoclonal antibody.

* Approximately 353, per Elan.

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism. 

Authors of the peer-reviewed phase 2 study of bapineuzumab in patients with Alzheimer disease conclude that the "trial provides insufficient evidence to support or refute a benefit" of the anti-amyloid mAb. The results of the Elan/Wyeth-funded trial, which are now available in an online issue of Neurology, were originally made public in a less-than-candid June 2008 company press release and at the International Conference on Alzheimer's Disease 1 month later.

The major drawback of the multidose study now appears to be related to a change in its primary outcome. The trial was originally designed and powered to assess the safety of bapineuzumab in AD, but the protocol was amended on the basis of phase 1 data. The authors write, "It was recognized that the small cohorts in this study would provide sufficient power to detect only very large treatment differences; however, if successful, the urgency of delivering an effective treatment to patients with AD argued for making the efficacy the primary outcome."

In other words, the investigators changed the primary endpoint of the phase 2 study on the chance that differences in efficacy outcomes might be substantial. In addition, the authors report the results of a modified intent-to-treat (mITT) population (n = 227), which consisted of patients who received at least 1 dose of study drug and underwent 1 or more efficacy evaluations during the study.

Despite these modifications, there were no statistically significant primary-efficacy* differences between placebo- and bapineuzumab-treated patients at 18 months. However, "exploratory analyses" showed "potential treatment differences" in the completer population (ie, patients who completed bapineuzumab treatment and underwent week-78 assessment) and in the subpopulation that did not have an ApoE4 allele.

Reversible vasogenic edema was observed in nearly 10% of bapineuzumab-treated patients and was dose dependent. Vasogenic edema was also more likely in ApoE4 carriers. In August, Elan and Wyeth announced that the highest bapineuzumab dosage, 2.0 mg/kg, would be discontinued in 2 ongoing phase 3 studies of ApoE4 noncarriers (40%-70% of AD patients), because of the risk of vasogenic edema. (Protocols for phase 3 trials of the mAb in ApoE4 carriers remain unchanged; these trials are assessing only 1 bapineuzumab dosage, 0.5 mg/kg.)

During the 18-month phase 2 study, bapineuzumab was given IV every 13 weeks at 1 of 4 doses (0.15, 0.5, 1.0, or 2.0 mg/kg). The smallest dose was added to the study after investigators identified vasogenic edema on some brain MR images of patients receiving 0.5 mg/kg.

An aside: The disclosures of the authors, several of whom are industry employees, consume about 3 columns of small type.

N.B. Bapineuzumab now belongs to JNJ, instead of Elan. Wyeth now belongs to Pfizer.

ADAS-Cog/12 = 12-item Alzheimer's Disease Assessment Scale-Cognitive subscale; DAD = Disability Assessment for Dementia; mAb = monoclonal antibody.

* Measured by using changes in ADAS-Cog/12 and DAD scores. 

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism. 

While I was blogging about prions last week, Pfizer and Eisai semi-quietly avoided a mammoth conflict. The two companies announced a restructuring of their comarketing deal for Aricept (donepezil), the number-one prescription med for Alzheimer disease.

Back in May, the Japan-based Eisai informed Pfizer that it had a legal right to terminate their copromotion of the drug, in place since 1994, because of Pfizer's anticipated acquisition of Wyeth. But Pfizer understandably bristled at the threat, given that Aricept generated $482 million in sales for the US firm in 2008.

The new agreement, according to Eisai's press release, continues to allow the companies to comarket Aricept in the United States, Japan, and "key markets in Europe" until 2022. However, copromotion in Japan will cease by 2013, with rights there reverting back solely to Eisai (which discovered the drug).*

The new agreement also launches a long-term comarketing deal for Pfizer's blockbuster drug Lyrica (pregabalin), which is approved in the United States for the treatment of neuropathic pain, epilepsy, and the ever-controversial fibromyalgia. An NDA for the drug has reportedly been filed in Japan. According to the AP, Lyrica generated $2.6 billion in worldwide sales last year.

* Eisai says that the basic patent on Aricept expires in 2010.

When attempting to discriminate among Alzheimer disease (AD), mild cognitive impairment (aMCI),* and normal cognition, structural abnormalities on MR images are more useful than known CSF biomarkers; but the use of both together is more diagnostically helpful than each individually. This conclusion is based on data from the Alzheimer Disease Neuroimaging Initiative, now available in the latest issue of Neurology.

In a statistical analysis of data from 399 subjects, the degree and location of AD-related brain atrophy on 3D MR images (condensed into the so-called Structural Abnormality Index, or STAND, score) correlated with clinical measures of cognition (eg, the MMSE score). However, known CSF biomarkers for AD—total tau, amyloid beta(1-42), and phosphorylated tau(181P)—did not. Nevertheless, the diagnostic accuracy for AD was best with the combined use of the STAND score, CSF total tau, and CSF amyloid beta. (Measures of total tau are expected to directly reflect neurofibrillary tangle pathology in AD.) In a related study, the STAND score was also a better predictor of clinical cognitive decline—namely the progression of aMCI to AD—than CSF biomarkers.

The Alzheimer Disease Neuroimaging Initiative is a longitudinal, multisite, observational study of elderly individuals with normal cognition, aMCI, or AD at 56 US or Canadian institutes. Funding for the study originates from the National Institute of Aging, the National Institute of Bioimaging and Bioengineering, pharma, and "several foundations."

CSF = cerebrospinal fluid; MMSE = Mini-Mental State Examination; MR = magnetic resonance.

* Amnestic mild cognitive impairment.

Photograph: Atrophied brain of person with AD from National Institute on Alcohol Abuse and Alcoholism.

People with amnestic mild cognitive impairment* (MCI) develop Alzheimer disease (AD) at an annual rate of 10%-15%, and about half show AD-type brain pathology—neurofibrillary tangles and beta-amyloid plaques—at autopsy. However, clinical conversion is not a given. Consequently a number of investigators have attempted to identify biomarkers that predict the conversion of MCI to AD, such as hippocampal volume on MR images.

Now researchers show that MCI patients with a relatively increased brain load of beta-amyloid, measured by using positron emission tomography** (PET), are much more likely to develop AD than MCI patients with control levels of beta-amyloid. Results of the small UK and Finnish study (N = 31) are available in an advanced e-printing of Neurology.

During a 1-3-year follow-up, 14 of 17 (82%) MCI patients with increased beta-amyloid on PET images and 1 of 14 (7%) MCI patients with nonincreased beta-amyloid converted to AD. The speed of clinical conversion was positively associated with the beta-amyloid load and the APOE ε4 carrier status in MCI patients.

APOE = apolipoprotein E.

* Defined as subjective and objective memory impairment for age but largely preserved general cognition and normal activities of daily living (Petersen, 2004).

** By using the thioflavin-based radiotracer Pittsburgh compound B (PIB).

Photograph: Atrophied brain of person with AD from National Institute on Alcohol Abuse and Alcoholism.

A "highly accurate" index to predict the risk of dementia during late life has been developed by investigators in Northern California and Pittsburgh. The scale, which is available in the latest online issue of Neurology, is based on data from more than 3000 community-dwelling adults* (≥65 years of age) in the Cardiovascular Health Cognition Study and identifies 11 potential risk factors for late-life dementia. 

Points are allotted for the following risk factors, and the tallied score (range, 0-15) can be used to stratify dementia risk among subjects. The accuracy of the index (defined by the c statistic, which is similar to the correlation coefficient) was calculated at 0.81 (95% CI: 0.79, 0.83) and is close to that of a previously published dementia-risk index for middle-aged individuals.

Dementia Risk Factor	Points
Age 75-79 y	1
Age 80-100 y	2
Low Modified MMSE score	2
Low Digit Symbol Substitution Test** Score	2
Body mass index <18.5	2
≥1 APOE ε4 allele	1
MRI: white matter disease (grade ≥3)	1
MRI: enlarged ventricles (grade ≥4)	1
Internal carotid artery thickness ≥2.2 mm	1
History of coronary artery bypass	1
Time to put on and button shirt >45 s	1
Lack of alcohol consumption	1

Among subjects with low scores (0-3), about 4% developed dementia during the next 6 years. Moderate scores (4-7) were associated with a 6-year dementia risk of 23%, and 56% of subjects with high scores (≥8) developed dementia. Among those subjects who acquired dementia (n = 480 or 14%), 245 had Alzheimer disease, 62 had vascular dementia, 151 had mixed dementia, and 22 had other forms of dementia.

Because some of the identified dementia risk factors may be difficult to determine in routine clinical practice (eg, grading of MRI parameters, APOE ε4 status), the authors are performing additional analyses to determine if a simpler, but equally accurate, index can be generated.

In the meantime, drink up—in moderation, of course. Of course. 

* Mean age at baseline = 75 years; ~60% women, 15% African American.

** A measure of attention and mental-processing speed.

A major conflict looms between Eisai and Pfizer over the marketing of Aricept (ie, donepezil), the number-one prescription medication for Alzheimer disease. Last week, the Japanese-based company informed Pfizer that it had a legal right to terminate their comarketing agreement, after Pfizer's anticipated aquisition of Wyeth, reported the WSJ. Not surprisingly, Pfizer bristled at the warning, especially given that Aricept generated $482 million in sales for the US company last year.

The comarketing agreement, which has been in place since 1994, currently allows Pfizer to copromote Aricept (which was discovered and developed by Eisai) in the United States and several other countries, according to news reports. Pfizer also claims that is has an exclusive license to sell Aricept in certain other countries.

Pfizer's $68-billion acquisition of Wyeth is expected to be completed by the end of this year. Eisai's justification for getting out of its comarketing agreement may have something to do with Wyeth's development of a competitive product. Wyeth (with Elan) is currently investigating bapineuzumab, an anti-amyloid monoclonal antibody. However, results from phase 2 and 3 clinical trials in AD have been less than exhilarating—something that Pfizer probably didn't bet on.

Aricept, a cholinesterase inhibitor, was approved by the FDA for the treatment of AD in 1996. The basic patent on Aricept will expire in November of 2010.

At least 4 placebo-controlled trials of cholinesterase inhibitors* (ChEIs) have been conducted in patients with mild cognitive impairment (MCI), a likely precursor of Alzheimer's disease. And none has demonstrated a significant effect on primary outcome measures, either progression to AD or standardized, cognitive tests.

Now add another largely negative trial, the results of which are available in this week's issue of Neurology. In a multicenter, double-blind, randomized study of 778 patients with MCI, donepezil (at a maximum dosage of 10 mg/d) was no better than placebo for altering cognitive scores at 1 year. However, a small, but statistically significant difference was noted between the modified ADAS-Cog scores of donepezil- and placebo-treated patients. Other secondary measures also favored donepezil, but treated subjects were much more likely to discontinue the drug because of adverse events (18.4% vs 8.3% with placebo).

The authors speculate that standard measures of cognition, function, and global impairment may be insensitive to the more subtle, progressive changes observed in patients with MCI. In a related editorial, David Knopman of the Mayo Clinic suggests that biomarkers—like hippocampal atrophy, low CSF amyloid-beta/tau protein ratios, and APOE ε4 genotypes—may provide better endpoints in MCI trials. All have been associated with higher rates of progressive cognitive dysfunction in these patients.

N.B. The study was supported by Eisai and Pfizer. Two of the 8 authors are employees of Eisai; 2 are employees of Pfizer.

ADAS-Cog: Alzheimer's Disease Assessment Scale-cognitive subscale.

* Donepezil (Aricept; Eisai/Pfizer), rivastigmine (Exelon; Novartis), and galantamine (Razadyne; Ortho-McNeil).

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism.

In contradistinction to recent studies from Johns Hopkins and the Veterans' Affairs Administration, newly published data suggest that heavy, long-term use of NSAIDs actually increases the risk of Alzheimer's disease. Investigators at the University of Washington reported their data from a large observational study, Adult Changes in Thought (ACT), in the latest online issue of Neurology.

The study assessed the incidence of AD in more than 2700 elderly* dementia-free enrollees in a large healthcare delivery system, which provided access to pharmacy data on NSAID use from 1977 and later. Subjects were followed up every 2 years for up to 12 years (16,931 person-years total) to identify incident dementia, and specifically AD.

Those subjects defined as heavy NSAID users demonstrated an increased incidence of dementia and AD (adjusted HR = 1.66 [95% CI: 1.24, 2.24] and 1.57 [95% CI:, 1.10, 2,23], respectively), when compared with low-level NSAID users. The inclusion of data for self-reported NSAID exposure provided similar data. The authors also found no significant interactions with respect to age or APOε4 status. Likewise, incident dementia was increased when data were stratified by distant or recent NSAID use (adjusted HR = 1.51 [95% CI: 1.03, 2.20] and 1.85 [95% CI: 1.28, 2.66], respectively).**

In an accompanying editorial, Bennett and Whitmer (like the ACT study authors) argue that very wide age differences in the cohorts of accumulating studies may account for the conflicting results. The effect of heavy, long-term NSAID use on the risk of AD may depend on the timing of NSAID exposure during life—such that younger persons may experience an eventual protective effect with early and prolonged NSAID use, while older persons may experience an increased risk of AD with relatively later NSAID use.

AD = Alzheimer's disease; HR = hazard ratio; NSAID = nonsteroidal anti-inflammatory drug.

* 65 years of age or older.

** However, it should be noted that moderate or heavy NSAID users were more likely than light users to have comorbid medical conditions or engage in poor health-related behaviors.

The clinical development of bapineuzumab in Alzheimer's disease has hit another snag. Codevelopers Elan and Wyeth announced today that the highest dosage of the anti-beta amyloid mAb, 2.0 mg/kg, will be discontinued in 2 ongoing phase 3 trials (NCT00667810 and NCT00574132), owing to the associated risk of vasogenic brain edema detected on MR images. The decision was made in conjunction with a review of unblinded data from the phase 3 trials by an independent Safety Monitoring Committee.

The affected trials are assessing bapineuzumab in patients with mild-moderate AD who do not carry the risk factor of the apolipoprotein E4 allele (40%-70% of AD patients). Phase 2 data from last year indicated that the agent was no better than placebo in 240 AD subjects. However, a post-hoc analysis indicated that bapineuzumab provided "statistically significant and clinically meaningful benefits" in non-ApoE4 carriers. These data also indicated that ApoE4 carriers may be especially prone to drug-associated vasogenic edema.

Current phase 3 study arms assessing 0.5- or 1.0-mg/kg dosages of bapineuzumab in noncarrier patients will continue. Patients originally assigned to the highest dosage will be allowed to continue the trial with 1.0-mg/kg treatment. Protocols for phase 3 trials of the mAb in ApoE4 carriers (NCT00676143 and NCT00575055)—which are assessing only 1 bapineuzumab dosage*—remain unchanged.

mAb = monoclonal antibody.

* 0.5 mg/kg according to MedPage Today.

 ...before looking forward.

Pathophilia's Top 10 Medical Stories of 2008: A Recap

10. Gunvalson v. PTC Therapeutics

9. California v. Roozrokh and Cardiac-Death Organ Donation

8. Hand, Foot, and Mouth Disease in China

7. Continuing Backlash Against Pharma

6. Media Obsession With Delayed Results of ENHANCE Trial

5. Investigational Drugs for Alzheimer's Disease Disappoint

4. Milder Rotavirus Season Coincides With Vaccine Uptake

3. USAMRIID Scientist Identified as Sole Perpetrator of "Anthrax Letter Attacks"

2. Resurgence of Measles

1. Intentional Drug and Food Tampering in China

Other notable stories of 2008 that didn't make Pathophilia's totally arbitrary list:

• More cases of progressive multifocal leukoencephalopathy (PML) with Tysabri (natalizumab) use
• Pig-slaughter neuropathy
• US government compensates Poling family for vaccine-related autism
• Ted Kennedy diagnosed with glioblastoma multiforme

When considering drug development for Alzheimer's disease, 2008 was a colossal disappointment

.

In June, Elan and Wyeth pumped up their post-hoc data from a phase 2 trial of the anti-amyloid drug bapineuzumab in patients with mild-moderate AD, despite the fact that the mAb was no better than placebo for the primary endpoints of cognition and disability. Initial reaction on Wall Street was inexplicably optimistic; however, reality sunk in a month later when the very same trial results were presented at the International Conference on Alzheimer's Disease (ICAD) in Chicago. But Elan and Wyeth aren't giving up yet on the drug; the companies are recruiting patients for at least 4 phase 3 studies in AD,* according to the NIH database.

However, Elan did abandon work on its synthetic amyloid beta compound (AN1792), when a small, placebo-controlled UK study showed that immunization with the compound cleared plaques in the brains of AD patients but did not improve cognition or survival in the long term. A company-sponsored phase 2 study of the agent in AD patients at multiple US sites (NCT00021723) was terminated.

In June, Myriad Genetics announced that its experimental agent tarenflurbil (Flurizan) failed to affect cognition or activities of daily living in early AD in a large phase 3 trial. The company's CEO reported that further development of the compound, an NSAID enantiomer of flurbiprofen, would be discontinued.

In September, Targacept and AstraZeneca revealed that their novel molecule, AZD3480, did not significantly change cognition in a short-term, randomized, placebo-controlled, dose-finding study of patients with mild-moderate AD. The agent is a selective agonist of a neuronal nicotinic receptor subtype. According the Targacept web site, a decision regarding the future development of AZD3480 in AD is expected in the first half of 2009, pending completion of an adult ADHD study.

Perhaps going after tangle-related tau protein is more promising. Also at this summer's ICAD, results of a placebo-controlled phase 2 study revealed that a formulation of methylene blue (yes, the Wright's stain agent) significantly improved cognition in patients with moderate AD at 24 weeks. The developer, TauRx Therapeutics, plans a large phase 3 trial of the drug.

Given these mostly disappointing trial results, maybe Pfizer wasn't so crazy to give Medivation $725 million for the worldwide marketing rights to Dimebon, a nonselective histamine blocker in phase 3 development for AD. Pfizer is evidently betting on the chance that the 3-times-per-day Dimebon, a defunct antihistamine, will sail through clinical development, given recent results of a Russian, donepezil-controlled phase 2 study. According to Medivation, the Russian study can be used as 1 of 2 pivotal trials to support the FDA approval of Dimebon, as long as a "significant portion" of the current phase 3 trial occurs within the United States.

ADHD = attention deficit hyperactivity disorder; mAb = monoclonal antibody.

* Enrollees will be distinguished by their apolipoprotein E4 carrier status.

For more details on these AD trials, visit these Pathophilia posts of 2008:

• Elan, Wyeth Bury Missed Primary Endpoints in Phase 2 Trial of AD Drug
• More Missed Primary Endpoints in AD Trials
• Elan's AN1792: No Clinical Benefit With Clearance of Amyloid Plaques in AD
• Bapineuzumab and Share Prices of Elan/Wyeth: What a Difference Some Undefined Thing Makes
• Anti-Tau Agent Warrants Phase 3 Study in AD
• Pfizer Gambles on Three-Times-per-Day Antihistamine for Alzheimer's Disease
• Another AD Drug Disappoints (Although Placebo Looks Promising)

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism.

A novel, experimental molecule, AZD3480, did not significantly change cognition (per the ADAS-cog) at 12 weeks in a randomized, placebo-controlled, dose-finding study of 567 European or Canadian patients with mild-moderate Alzheimer's disease. Early results of the phase 2b study were revealed Monday by the molecule's codevelopers, Targacept and AstraZeneca.

In addition, the approved AD drug donepezil (Aricept; Eisai/Pfizer) did not significantly change the mean ADAS-cog score at 12 weeks. (The dosage of donepezil studied was not provided in the companies' press release.) Disappointing outcomes with AZD3480, a selective agonist of a neuronal nicotinic receptor subtype (α4β2), and donepezil, an AChE inhibitor, were due to unexpected cognitive improvement in those patients who received placebo, wrote the companies.

Patients who received 2 of 3 tested doses of AZD3480 did show modest cognitive improvement per secondary outcome measures, as did patients who received donepezil. AZD3480 appeared to be better tolerated than donepezil, with fewer episodes of diarrhea, nausea, or vomiting.

The companies indicate that analysis of the trial results is ongoing, and trial data will be presented in a scientific setting at some unspecified time and place. A decision regarding the continued development of the AD drug candidate is expected at the end of this calendar year. AZD3480 is also being evaluated for the treatment of cognitive dysfunction in schizophrenia (trial results are expected at the end of 2008) and adult attention deficit/hyperactivity disorder.

AChE = acetylcholinesterase; ADAS-cog = Alzheimer's Disease Assessment Scale-cognitive subscale.

Photograph: Atrophied brain from person with AD from National Institute on Alcohol Abuse and Alcoholism.

$725 million is a lot of money for an old Russian antihistamine.

But that's what Pfizer agreed to give Medivation last week for the worldwide marketing rights to Dimebon (dimebolin), a defunct, nonselective histamine blocker that is now in phase 3 clinical development for Alzheimer's disease: $225 million upfront and another $500 million when (or if) the drug is approved by the FDA.

The agreement—which splits development costs and profits on a 60-40 basis (Pfizer assumes the larger share)—also confers licensing rights to Pfizer for use of the drug in Huntingdon's Huntington's disease. (Dimebon is currently in phase 1/2 development for this condition.)

Pfizer is evidently betting on the chance that the 3-times-per-day Dimebon will sail through clinical development, given recent phase 2 results published in The Lancet. In a randomized, double-blind, placebo-controlled, multicenter study (N = 183) that was conducted entirely within Russia and funded by Medivation, patients with mild-moderate AD who received Dimebon* demonstrated significantly less cognitive decline at 6 months, as measured by the ADAS-cog (mean score difference, 4.0; P < .0001).

Specifically the mean ADAS-cog score of those who received Dimebon improved by approximately 2 points and the mean score of placebo-treated patients declined by more than 2 points. Secondary outcome measures (eg, MMSE) supported these findings. In a blinded, 26-week extension phase, the difference in the mean ADAS-cog score between patients who originally received Dimebon and those who originally received placebo increased further (6.9; P < .0001). The most common adverse effects observed with Dimebon treatment in this phase 2 study were dry mouth and depression.

By comparison, the mean ADAS-cog score difference between the once-daily blockbuster Aricept (donepezil; Eisai/Pfizer) and placebo was approximately 3 points at 30 weeks in patients with mild-moderate AD. The basic patent on Aricept, the most widely used AD drug, will expire in 2 years.

According to Medivation, Dimebon has been shown to inhibit brain-cell death in preclinical models of AD. A PubMed search reveals at least 4 rodent studies and 1 pilot study of the agent at the Moscow Center of Gerontology. The company further proposes that Dimebon confers benefits in AD by "improving mitochondria function."

Also according to Medivation, the Russian study can be used as 1 of 2 pivotal trials to support the FDA approval of Dimebon, as long as a "significant portion" of the current phase 3 trial occurs within the United States. A 6-month, placebo-controlled phase 3 trial of Dimebon (5 or 20 mg tid) in 525 patients with mild-moderate AD was initiated in the second quarter of this year; results are expected in 2010, reports The Street. Dimebon is also being assessed in combination with donepezil in a phase 1 study.

ADAS-cog: Alzheimer's Disease Assessment Scale-cognitive subscale; MMSE = Mini-Mental State Examination.

* 10 mg tid for 7 days, followed by 20 mg tid. 

Images of defunct Russian and defunct Russian antihistamine from Wikipedia.

While the effects of anti-amyloid agents in Alzheimer's disease (AD) have been less than overwhelming (see here and here, for example), a phase 2 study presented at the recent International Conference on Alzheimer's Disease suggests that going after tangle-related tau protein may be more fruitful.

A formulation of methylene blue—yeah, the Wright's stain agent—produced a significant 5.5-point difference (vs placebo) on the ADAS-cog at 24 weeks in patients with moderate AD. The proprietary drug, trade name "rember" (ugh), was assessed in an initial double-blind, randomized, placebo-controlled, dose-ranging study in mild-to-moderate AD (N = 321), followed by a blinded, extension phase. Pooled data after 19 months revealed a significant 6.8-point difference on the ADAS-cog in patients with either mild or moderate AD. Statistically significant cognitive changes at both time points were observed with the dosage of 60 mg tid, versus initial placebo treatment.

Rember was evidently well tolerated in the phase 2 trial, except for a high incidence of diarrhea, which may depend on the drug's intestinal absorption. Reports indicate that the efficacy and tolerability of a 100-mg dose may have been compromised because of formulation problems.

Rember is an investigational agent of TauRx Therapeutics and inhibits in-vitro tau aggregation, the foundation of AD's neurofibrillary tangles. The clinical study was headed and presented by Claude M. Wischik, MBChB, of the University of Aberdeen, who also happens to be founder and executive chairman of TauRx.

A large phase 3 trial of rember in AD is planned but awaits FDA approval.

Primary sources: Medpage Today; AP via Yahoo.

As if you need another reason to bang your head against a wall on Wall Street.

After phase 2 data of bapineuzumab in Alzheimier's disease (AD) were released on June 17, shares of Elan and Wyeth climbed 10.6% and 4.8%, respectively. However, after the same data were presented yesterday at the International Conference on Alzheimer's Disease (ICAD) in Chicago, the share prices dropped 19.6% and 11.2%, respectively. So what were the differences between the June press release and the meeting presentation? Not a heck of a lot.

 

Both presentations provided information about the missed primary endpoint in the trial, a change in cognition or disability at 18 months (although the primary outcome was buried in the second paragraph of the June press release). Both provided information about the post-hoc assessment of subgroup data—namely, that ApoE4 noncarriers performed statistically better on the ADAS-cog at 18 months (mean score difference: 5 points, bapinueuzumab vs placebo). And both reported a risk of vasogenic brain edema with the agent, especially in ApoE4 carriers.

In June, analysts talked up the bapineuzumab phase 2 study, despite the missed primary endpoint, and the companies' prospects. Yesterday, not so much. Nevertheless, Elan and Wyeth are apparently moving forward with phase 3 investigation of the amyloid buster in AD.

Women who live into their 90s are significantly more likely to be demented than similarly aged men, according to a report in this week's issue of Neurology. The finding supports data from previous studies, in which the prevalence of dementia among centenarians was higher among surviving women.

By using in-person examination data and telephone and informant questionnaires, US investigators assessed the all-cause dementia rate among participants in the so-called 90+ Study (N = 911). Among women, the overall dementia rate was 45% (95% CI: 41.5, 49.0), versus 28% (95% CI: 21.7, 34.2) among men. Moreover, the dementia rate in women doubled every 5 years after the age of 90 (Figure). Dementia in women, but not men, correlated inversely with education. From a glass-half-full perspective, dementia was not universal in women 100 years of age or older (61%).

In an accompanying editorial, David Hogan, MD, cites limitations of the study including differing methods used for the diagnosis of dementia; the characteristics of the study population (white, relatively well educated, upper middle class); insufficient pathologic data; a high lost-to-study rate (21%); and a low absolute number of dementia cases among men 95 years of age or older (n = 23). However, Hogan supports the overall conclusion and suggests that the lack of an age-associated increase in dementia among very old men may be due to their shortened survival with the condition.

Immunization with synthetic amyloid beta (AN1792; Elan) clears amyloid plaques in brain but is not associated with improved cognition or survival in the long term, according to a small UK study of patients with Alzheimer's disease (AD). Results of the placebo-controlled trial, funded by the Alzheimer's Research Trust, were published in this month's issue of The Lancet.

A total of 80 patients with probable mild-to-moderate AD received either randomly assigned AN1792 (50 or 225 ug) with adjuvant (n = 64) or adjuvant alone (n = 16).* Postmortem examination of 8 actively treated AD patients revealed a significantly lower Aβ load than that in nonimmunized controls.^† However, the investigators noted considerable variation of the Aβ load and the extent of plaque removal among immunized cases.

There was no relationship between the dose of AN1792 and Aβ load or postmortem plaque removal, but the mean antibody response during treatment showed a nonsignificant correlation with Aβ load. Seven of the 8 patients, including those who demonstrated "virtually complete plaque removal," exhibited severe end-stage dementia, and there were no survival or cognitive differences between the immunized and placebo-treated groups.

A phase 2a study of AN1792 was halted when 6% of patients developed meningoencephalitis. The study did not show relative cognitive improvement at 1 year with immunization, despite the presence of high levels of Aβ antibodies. In June, Elan and Wyeth announced favorable results with the anti-Aβ monoclonal antibody bapineuzumab in mild-moderate AD, despite the fact that the primary endpoint in the phase 2 trial was not met.

* In a protocol extension, patients received either AN1792 (n = 51) or adjuvant alone (n = 13) in a modified formulation to increase Aβ solubility.

† Because no patients in the placebo-treated group underwent postmortem examination, histologic findings in an age-matched, nonimmunized control group of patients with AD were used for comparison.

Image of wild-type amyloid precursor protein (left) and amyloid fibril (right) by David S. Goodsell from Wikimedia Commons.

There are no optimistic forward-looking statements here.

Today, Myriad Genetics announced that its experimental agent tarenflurbil (Flurizan) failed to affect cognition or activities of daily living in early Alzheimer's disease, the primary endpoints of a large (N = 1600), 18-month, phase 3 trial. Tarenflurbil is an NSAID enantiomer of flurbiprofen. The company's CEO reported that, consequently, the development of the compound would be discontinued.

This news follows on the heels of the missed primary endpoints in a phase 2 trial of bapineuzumab in mild-to-moderate Alzheimer's disease, which were nevertheless spun as favorably as possible by developers Elan and Wyeth.

For the first time, researchers showed that amyloid beta (Aβ) dimers from the brains of individuals with Alzheimer's disease induce several AD-like changes in normal rodents. The results of a series of related experiments were reported in the latest online edition of Nature Medicine.

AD-consistent pathophysiologic changes were observed in the normal rodent hippocampus after it was exposed to soluble (but not insoluble) Aβ. The investigators, from Boston and Ireland, discovered

1. the inhibition of long-term potentiation (LTP) of synaptic transmission (a cellular model for learning and memory) in a dose-dependent fashion;
2. enhanced long-term depression (LTD) (a marker of weakening synaptic transmission); and
3. reduced dendritic spine density (a marker of synaptic loss).*

The administration of the Aβ dimer also disrupted the learning of a standard avoidance task in rats. Antibodies to the N-terminus of Aβ prevented the dimer's effects on LTP and LTD (which requires metabotropic glutamate receptors); however, the effect of antibodies to other regions of Aβ was not as remarkable.

The fact that AD-like changes were not detected with insoluble Aβ or other oligomers of Aβ may explain the disconnect between relatively preserved cognitive function during life and a high burden of Aβ in some brains at autopsy—as Marcelle Morrison-Bogorad, director of the neuroscience division at the National Insititute on Aging (NIA), told the AP. The Nature Medicine study was funded, in part, by the NIA.

* The authors note that decreased synaptic density is the strongest neuropathologic correlate of dementia in AD.

Image of wild-type amyloid precursor protein (left) and amyloid fibril (right) by David S. Goodsell from Wikimedia Commons.

Elan and Wyeth announced "encouraging top-line results" from a randomized, dose-ranging phase 2 trial of bapineuzumab in 240 patients with mild-to-moderate Alzheimer's disease; although the monoclonal antibody was no better than placebo when assessing cognition or disability at 18 months, the primary endpoints. A press release from the companies contained the primary-endpoint data in the second paragraph.

Instead the companies chose to highlight results from the post-hoc analyses of subgroups, including patients who do not carry the high-risk apolipoprotein E4 allele (40%-70% of AD patients).* Data in this subgroup showed "statistically significant and clinically meaningful benefits" with bapineuzumab, per cognitive scales like the ADAS-cog. The treated subgroup also demonstrated relatively preserved brain volume on MRI, the press release reported.

Safety data in the phase 2 trial indicated that ApoE4 carriers may be especially prone to vasogenic edema with the agent. Consequently the dose of bapineuzumab, an anti-Aβ agent, is being modified for this subpopulation in phase 3 studies. According to ClinicalTrials.gov, two phase 3 studies of bapineuzumab in AD are currently recruiting subjects, and another two phase 3 studies are "active" but "not yet recruiting."

The Elan/Wyeth press release indicates that full details of the phase 2 study will be presented July 28 at the International Conference on Alzheimer's Disease in Chicago. Despite the mixed phase 2 results, the reaction of Wall Street was mostly optimistic, according to the WSJ Health Blog.

06/18/08 update: According to Forbes, analysts Leerink Swann and Davy Stockbrokers are talking up the study and the companies' prospects. For no particularly good reason, respective share prices of Elan and Wyeth shot up yesterday 10.6% and 4.8%.

Image: depiction of amyloid plaque from National Institute on Aging.

Aβ = amyloid beta.

*So, in this case, we may guess that approximately 12-20 patients per treatment group (4 for bapineuzumab; 4 for placebo) were noncarriers.

The aggregation of insoluble amyloid fibrils is an important pathologic hallmark of Alzheimer's disease (and a number of other conditions). The most fibrillogenic isoform of amyloid beta, Aβ42, is produced when the enzyme γ-secretase cleaves the transmembrane portion of amyloid precursor protein (APP). Therefore a potential strategy for the treatment of AD is the inhibition of γ-secretase with γ-secretase modulators (GSMs).

In this week's issue of Nature, Kukar and colleagues report the surprising finding that the target of their synthetic GSMs is the substrate APP and not the enzyme itself. Specifically the GSM binding site localizes to the 28-36 residue portion of amyloid beta, a region that is critical for amyloid aggregation. Therefore certain GSMs have the potential to suppress the production of Aβ42 and its aggregation. The authors propose that substrate-targeting GSMs for further investigation can be identified by their ability to bind Aβ and APP. One Aβ42-lowering GSM, tarenflurbil (Flurizan; Myriad), is in phase 3 clinical study in patients with mild AD.

The goal of suppressing Aβ42 formation, however, is at odds with recent population studies, which showed that the risk reduction of AD with NSAIDs is not dependent on the suppression of Aβ42. In an e-mail, Kukar acknowledged the disconnect but also stressed that epidemiologic studies cannot substitute for well-controlled clinical trials.

Image of wild-type APP (left) and amyloid fibril (right) by David S. Goodsell from Wikimedia Commons.

Last year, Johns Hopkins investigators found that use of NSAIDs reduced the risk of Alzheimer's disease by 37%. Closer examination of the data (from the Cardiovascular Health Cognition Study) revealed that the protective benefit of NSAIDs was confined to those individuals with an ApoE4 allele; however, the benefit was not attributable to the suppression of Aβ1-42 amyloid. 

Then earlier this month, a review of Veterans Affairs records showed that long-term use of NSAIDs reduced the odds of AD by 24%, and that a more substantial risk reduction was observed with ibuprofen and naproxen (odds reduction, 37%). However, like the Hopkins study, the Veterans records did not indicate a particular benefit with those NSAIDs that suppress Aβ1-42 amyloid.

Now comes a follow-up report from the same Hopkins investigators, who pooled data from 6 prospective studies* (N = 13,499). Again, the investigators found that the use of NSAIDs reduced the risk of AD by approximately one third. And also again, the risk reduction with NSAIDs that suppress Aβ1-42 amyloid was not higher than that with nonsuppressing NSAIDs. A substantial risk reduction with aspirin, but not acetaminophen, was also observed.

*Baltimore Longitudinal Study of Aging, Cache County Study, Canadian Study of Health and Aging, Cardiovascular Health Study, Framingham Heart Study, and Monongahela Valley Independent Elders Study.

Photo: iStockPhoto.

Data to date suggest that NSAIDs, including aspirin, do not affect the course of Alzheimer's disease in those with an established diagnosis (see, for instance, here and here). However, the potential of NSAIDs to prevent or delay the onset of dementing illnesses is another matter. Recently investigators at Johns Hopkins found that the use of NSAIDs, but not aspirin, lowered the risk of dementia overall—and in particular AD—among participants in the Cardiovascular Health Cognition Study.* Closer examination of the data revealed that the protective benefit of NSAIDs was confined to those individuals with an ApoE4 allele but did not appear to be attributable to the suppression of Aβ1-42 amyloid.

Now comes a review of Veterans Affairs records in the latest issue of Neurology, which supports the long-term use of NSAIDs to reduce the risk of AD. These data were particularly compelling for ibuprofen (or what I like to call God's gift to me). Also, like the Hopkins study, a distinction was not observed between NSAIDs that suppress the formation of Aβ1-42 amyloid (ie, ibuprofen, indomethacin, sulindac, and diclofenac) and those that don't (eg, naproxen, etodolac).

But most important, the benefit of NSAIDs appeared to be associated with the duration of their prescription (determined from VA pharmacy records). The authors found that the odds of AD decreased from 0.98 with ≤1 year of NSAID use to 0.76 with ≥5 years, in the nearly 250,000 individuals identified. A similar, but more pronounced, effect was observed with the use of arylpropionic acids, such as ibuprofen or naproxen (≤1-year OR = 1.00; ≥5-year OR = 0.63). However, no protective benefit was observed with the use of COX-2 inhibitors or nonacetylated NSAIDs (eg, salsalate).^†

The authors acknowledge a number of limitations in their review study, including potential errors in diagnostic coding, medical confounders (eg, education, tobacco use), and actual NSAID exposure. It's entirely possible that patients did not take their prescribed medications, or that others took undocumented over-the-counter NSAIDs. Also data were not available (understandably) for the ApoE4 allele status of the veterans.

Nevertheless, given these data, the authors conclude that a randomized, prospective trial of long-term NSAID therapy (and ideally ibuprofen) is warranted. The same can also be recommended to assess the risk reduction of Parkinson's disease with non-aspirin NSAIDs, according to a recent UCLA study.

*Curiously enough, the risk of vascular dementia was not affected by NSAID use.

†What's old is new again, or is this just the VA? I didn't know anybody prescribed Disalcid anymore.

Alzheimer's disease isn't particularly difficult to recognize clinically; although definitive diagnosis requires the pathologic examination of brain tissue—something not practical in life. Nevertheless, it's nice to have confirmation in the form of biomarkers, usually from the CSF, and particularly in cases of early disease. To this end, investigators from the University of Kentucky and the Oregon Health Sciences Center indicate that high levels of an aberrant protein complex of prostaglandin-d-synthase (PDS) and transthyretin (TTR) in CSF differentiate individuals with AD from normal control subjects. Their results were published online, April 30, in Neurology.

The conclusion was based on quantified levels of the CSF protein complex in patients who underwent autopsy and those in living individuals. The results are tabulated below:

Patient Type	Mean Age	Median Complexed TTR, ng/mL
Autopsy patients (ventricular CSF)
Normal control	83.3	<5
Diseased control	70.4	<5
Mild cognitive impairment	89.8	101.5*
Late-stage AD	81.6	74*
Living patients (archived lumbar CSF)
Normal control	71.0	<5
AD	68.4	26.5*

* P < .05.

In autopsied patients, a direct correlation was detected between the protein-complex level and measures of neurofibrillary pathology (Braak staging score) and neuritic plaque density (CERAD rating). There was, however, no significant correlation between the PDS/TTR level and the number of ApoE4 alleles.

Data from living subjects indicated that the level of CSF PDS/TTR was 100% sensitive and 93% specific for AD. Increasing PDS/TTR levels also correlated inversely with Mini-Mental State Examination scores.

It's important to note that the study's first 2 authors, Mark Lovell, PhD, and Bert Lynn, PhD, disclosed a financial interest in Scout Diagnostics, "a company that recently licensed intellectual property based on studies partially described in [the] report." More specifically, the company was founded by Lovell and Lynn in 2006, along with CEO John Beran, to develop and market the diagnostic test for AD. The authors conclude that larger numbers of patients are required to confirm the utility of the CSF level of PDS/TTR for the diagnosis of AD in life.