Subscribe to eTOC

Medical Watchdog Groups Asks FDA to Withdraw Donepezil 23 Mg

DR. DAVID S. KNOPMAN: “The two primary end points they specified ahead of time were not met, but they glossed over that fact and emphasized that they achieved statistical significance on the Severe Impairment Battery.”

Shortly after patent protection on donepezil (Aricept) expired late in 2009, the manufacturer introduced a new version containing a higher dose, with advertisements claiming it was effective for treating mild, moderate, and severe Alzheimer disease (AD).

But several leading AD researchers have expressed serious doubts about the drug‘s effectiveness and safety at the higher 23 mg dose. And Public Citizen, a medical watchdog organization, has filed a petition with the FDA to request that the drug marketed as Aricept 23 be withdrawn immediately.

“Aricept went off patent last year, so in order to maintain a proprietary drug they came up with a dose formulation that you couldn‘t easily duplicate by taking two [standard dose] 10 mg pills,” said David S. Knopman, MD, professor of neurology at Mayo Clinic in Rochester, MN. “If the higher dose were more effective, I would have no objection, but the clinical trial conducted by Aricept‘s sponsors failed to meet basic standards for demonstration of efficacy. I would like to see the scientific leadership of the neurologic community go on record as objecting to this misleading interpretation of the clinical trial data.”

Donepezil, manufactured by the Japanese drug company Eisai in partnership with Pfizer, was recording sales of about $2 billion a year before the expiration of the patent opened the way for less expensive generic versions of the drug. Eisai predicted generic competition would cut sales by 60 percent within three years.


The FDA approved donepezil at 23 mg a year ago based on a head-to-head study — published last July in the journal Clinical Therapeutics — comparing the 10 mg dose to the new 23 mg formulation in more than 1,400 patients with moderate-to-severe AD.

The FDA often approves drugs that are similar to an existing formulation simply by requiring evidence of equivalence, but in an effort to demonstrate that the higher-dose drug would offer greater benefit, especially in later stages of AD, Eisai established two primary end points — statistically significant improvement on the Severe Impairment Battery (SIB) and on the Clinician‘s Interview-Based Impression of Change-Plus version (CIBIC-Plus).

“The two primary end points they specified ahead of time were not met,” said Dr. Knopman, “but they glossed over that fact and emphasized that they achieved statistical significance on the Severe Impairment Battery. They also claimed a post-hoc analysis showed that severely impaired patients showed a little bit of benefit on the global assessment. On that basis they claimed efficacy, and for some reason the FDA, which is usually so skeptical, gave them approval.” [For more on the study data, see “Donepezil 10 Mg Vs. Donepezil 23 Mg: Protocols and Results.”]


The study leader Martin R. Farlow, MD, professor and vice-chairman of research in the neurology department at Indiana University School of Medicine, defended the study. He said the findings showed significant benefit, and he attributed the controversy over approval to the fact that Eisai “set the bar high in terms of the design of the study. Rather than equivalence, they said there would be a significant benefit.”

Dr. Farlow said the study was based on the hypothesis that patients in the later stages of AD who are suffering from a larger deficit of acetylcholine would benefit from a higher dose of donepezil, which blocks cholinesterase, the enzyme that breaks down acetylcholine. “If you are able to block more cholinesterase, you can further compensate for the deficit in cholinergic functioning, providing greater benefits for the patient,” he told Neurology Today. “I would argue that the hypothesis has been proven. There clearly is a treatment signal, particularly in the US population, where significant treatment effects were seen for both the cognitive and global measures. The effect is modest, but this study would suggest that additional drug does provide additional benefit for some patients.”

Although adverse events were higher in the donepezil 23 mg group, that was expected since those patients had advanced to the higher dose, according to Dr. Farlow, whereas the donepezil 10 mg group was simply continuing on a medication dose they had been on for two years. “After the titration phase, adverse rates were the same for the 23 mg as for the 10 mg groups,” he said.


Two FDA reviewers recommended not approving the drug, however. Ranjit B. Mani, MD, FDA medical officer (neurology) observed in his review — online at — that results of the study “did not satisfy the pre-specified criteria for demonstrating the efficacy of the higher dose of donepezil.” Even though the primary efficacy analysis showed a small but statistically significant effect of donepezil on the SIB, the same was not true of the CIBIC-Plus. Therefore the study failed to provide “confirmation that the small effect on the SIB was clinically meaningful,” Dr. Mani said.

Statistical reviewer Tristan Massie, PhD, concurred, stating in his report — — that “the data from this trial does not seem to provide enough support for the efficacy of the 23 mg SR formulation.”

DR. MARTIN R. FARLOW: “There clearly is a treatment signal, particularly in the US population, where significant treatment effects were seen for both the cognitive and global measures. The effect is modest, but this study would suggest that additional drug does provide additional benefit for some patients.”

While recognizing the concerns of the two FDA reviewers, Russell Katz, MD, director of the FDA Division of Neurology Products, recommended approval of the higher dose. He wrote that because the 23 mg dose proved superior to the 10 mg dose on the SIB, “the 23 mg dose is very likely to also have an effect on overall functioning [CIBIC-Plus], despite this not having been demonstrated directly in this study.”

Public Citizen, in its petition to the FDA, found the reasons for approval provided by Dr. Katz “completely unacceptable,” alleging that Eisai sought approval for donepezil at 23 mg solely to keep one form of donepezil under patent protection.

DR. LON S. SCHNEIDER said other studies have suggested that the consequences of donepezil‘s side effects may be greater than recognized.

Adriane Fugh-Berman, MD, associate professor in the department of physiology and biophysics and the department of family medicine at Georgetown University Medical Center, supports that allegation. Dr. Fugh-Berman, founder of PharmedOut, which conducts research on the effects of pharmaceutical promotion on prescribers and consumers, called Eisai‘s effort to gain approval for Aricept 23 “a classic example of what the industry calls ‘evergreening‘ — extending the life on the drug by making minor changes in formulation, or by releasing a drug in an immediate release form, and then controlled release, and then extended release. Aricept is not very effective at any dose, and the higher dose increases uncomfortable side effects.”

And in a recent discussion of Aricept 23 in the “Journal Club” column of American Family Physician, Mark A. Graber, MD, clinical professor of emergency medicine and family medicine at the University of Iowa Carver College of Medicine in Iowa City, pointed out that of the four comparisons done between Aricept and Aricept 23, three were negative and one — the SIB —was positive by only 2 points on a 100-point scale. “So although this is statistically significant, it is clinically meaningless,” he wrote. “There is no discernible benefit for the patient or caregivers.”


The biggest concern expressed by opponents of approval involved the increase in side effects produced with a higher dose of donepezil. For example, 9.2 percent of the patients randomized to the 23 mg dose of donepezil experienced vomiting, according to the study, compared to 2.5 percent who continued to receive the 10 mg dose. In AD patients, vomiting can lead to pneumonia, massive gastrointestinal bleeding, esophageal rupture, or death, Dr. Mani stated in his review of the clinical data. So besides lacking any significant benefit, donepezil 23 “is also associated with a significant risk to patient safety,” he added.

Lon S. Schneider, MD, professor of psychiatry, neurology, and gerontology at the University of Southern California Keck School of Medicine, said other studies have suggested that the consequences of donepezil‘s side effects may be greater than recognized. For example, a 2009 Archives of Internal Medicine study — comparing 20,000 Canadian elderly demented patients who were taking donepezil with 61,500 demented adults who were not — reported that use of cholinesterase inhibitors “is associated with increased rates of syncope, bradycardia, pacemaker insertion, and hip fracture in older adults with dementia.” The study also pointed out that syncope can contribute to fall-related injuries such as hip fractures.

Even the study on which FDA approval was based found that those on the 23 mg dose of donepezil scored higher on all of the 18 side effects mentioned than those on the 10 mg dose, with particularly large disparities in nausea (11.8 vs. 3.4 percent), anorexia (5.3 vs. 1.7), and bradycardia/sinus bradycardia (2.8 vs. 0.8).

“Any efficacy and value you could find from the post-hoc analysis may not be worth the added risk,” Dr. Schneider said. “So what if you show a 2-point change on the Severe Impairment Battery? To go from 75 to 77 on a 100-point scale — is that meaningful? And especially when there are no effects on activities of daily living, clinicians‘ global impressions, and the MMSE [Mini-Mental Status Examination]? Fundamentally, 23 mg is equivalent to 10 mg, but with less tolerability and more adverse events. A few individual patients might benefit over six months, but not most. Many physicians prescribe 10 mg of Aricept as though it‘s safe as water, and if patients are dizzy, fall and have fractures, lose weight, or require a pacemaker, they have no idea that it can be from the drug.”

At press time, the FDA had not responded to the petition. In response to a query from Neurology Today, FDA spokesperson Sandy Walsh said: “We cannot comment on a pending petition. The FDA will review the petition and respond directly to the petitioner.”


  • The randomized, double-blind study was conducted at 219 sites in Asia, Europe, Australia, North America, South Africa, and South America from June 6, 2007 to March 27, 2009.
  • 1467 patients were randomly assigned to receive high-dose donepezil (23 mg once daily) or standard-dose donepezil (10 mg once daily) for 24 weeks.
  • The effectiveness analyses included 1371 patients (mean age, 73.8 years).
  • At study end (week 24), the SIB score was significantly greater with donepezil 23 mg/daily than with donepezil 10 mg/daily (difference, 2.2; p < 0.001). The between-treatment difference in CIBIC-Plus score was nonsignificant.
  • In post-hoc analysis, the SIB score and CIBIC-Plus treatment effect at end point were greater with donepezil 23 mg/d than 10 mg/daily in patients with more advanced AD compared with less impaired patients (SIB; [p < 0.001]; CIBIC-Plus, [p = 0.028]).
  • Thirteen deaths were reported during the study or within 30 days of study but all were considered unrelated to the study medication.


The Severe Impairment Battery is a multi-item instrument assessing elements of memory, language, orientation, attention, praxis, visuospatial ability, construction, and social interaction ranging from 0 (worst) to 100 (best).

Clinician‘s Interview Based Impression of Change rates the patient‘s condition, with input from the caregiver, from 1 (markedly improved) to 7 (markedly worse).


Farlow MR, Salloway S, Satlin A, et al. Effectiveness and tolerability of high-dose (23 mg/d) versus standard-dose (10 mg/d) donepezil in moderate to severe Alzheimer‘s disease: A 24-week, randomized, double-blind study. Clin Ther 2010;32(7):1234–1251.
Mani RB. FDA medical officer review for NDA application number 022568. July 23, 2010,, Accessed June 23, 2011.
    Massie TS, Jin K, Hung HM. FDA statistical review for NDA application number 022568. July 9, 2010., Accessed June 25, 2011.
      Fugh-Berman A. Prescribing evidence: the effectiveness and safety of new drugs. Amer Fam Physician 2010;82(1):53–57.
        Dachs R, Graber M A, Darby-Stewart A. Donepezil to manage Alzheimer disease: New vs. standard Dosing. Am Fam Physician 2011;83(6):742–744.
          Schneider LS. Treatment of Alzheimer‘s disease with cholinesterase inhibitors. Clin Geriatr Med 2001;17(2):337–358.
          Gill S, et al. Syncope and its consequences in patients with dementia receiving cholinesterase inhibitors. Arch Intern Med 2009;169:867–873.
          Birks J. Cholinesterase inhibitors for Alzheimer‘s disease. Cochrane Database of Systematic Reviews 2006;(1):CD005593.