by Gustavo Alva, MD, and Jeffrey L. Cummings, MD

Dr. Alva is from ATP Clinical Research, Costa Mesa, California, and Dr. Cummings is from the University of California, Los Angeles, Alzheimer’s Disease Research Center, Los Angeles, California

Psychiatry (Edgemont) 2008;5(11):27–36

Financial Disclosures: Dr. Alva has received financial support from Forest Laboratories, Inc., Janssen, Novartis Pharmaceuticals, Astra-Zeneca, Corcept Therapeutics, Myriad Genetics, Inc., Bristol-Myers Squibb/Otsuka Pharmacuticals, Takeda Pharmaceuticals, Pfizer, and Organon; Dr. Cummings has received financial support from Forest Laboratories, Inc., Janssen, H. Lundbeck A/S, Merz Pharma, Myriad Genetics, Inc., Novartis Pharmaceuticals, Pfizer, Avanir Pharmaceuticals, Eisai, Neurochem Inc., Ono Pharmaceutical Co., Sanofi-Aventis, Sepracor Inc., Takeda Pharmaceutical, and EnVivo Pharmaceuticals, Inc.

Key Words: Alzheimer’s disease, cholinesterase inhibitor, memantine, tolerability


Objective: In the US, approved therapies for mild to moderate Alzheimer’s disease (AD) currently comprise three cholinesterase inhibitors (ChEIs: donepezil, galantamine, and rivastigmine), while the N-methyl-D-aspartate (NMDA) receptor antagonist memantine and the ChEI donepezil are approved for moderate to severe AD. The purpose of this study is to review the safety and tolerability of the ChEIs and memantine, based upon manufacturers’ data found in prescribing information (PI) documents.

Design: Current PI documents for donepezil, galantamine, rivastigmine, and memantine were obtained from American manufacturers’ websites, accessed in September, 2007. Adverse events (AEs) data for each drug versus placebo were compiled and analyzed using odds ratios.

Results: A review of PI data indicated that all three ChEIs are associated with cholinomimetic effects. Nausea (2–8%) and vomiting (1–5%) were reported across all ChEI trials as the most common reasons for trial discontinuation. Dizziness, anorexia, and diarrhea were also commonly experienced; however, a recent study suggests improved tolerability with transdermal administration of rivastigmine. The most frequently reported AEs in memantine trials were dizziness, headache, and confusion. There were no AEs that lead to trial discontinuation in more than one percent of memantine-treated patients and at a frequency greater than that observed in placebo-treated patients.

Conclusions: Data from this review suggest that gastrointestinal side effects are typical of ChEIs. Problems with rivastigmine tolerability may be reduced by transdermal administration. Memantine provides a distinctive tolerability profile. It is important to note that this study sought to overcome the lack of direct-comparison trials by analyzing the data presented by each company in its own PI material; however, caution should be exercised when comparing values obtained from different trials or trial groups.


In the United States, the cholinesterase inhibitors (ChEIs) donepezil,[1,2] galantamine,[3,4] and rivastigmine[5–7] are approved for the treatment of mild to moderate Alzheimer’s disease (AD), while donepezil[8–11] and memantine,[12,13] an uncompetitive antagonist of N-methyl-D-aspartate receptors, are indicated for patients in moderate to severe stages of the disease. All four drugs are administered orally; in addition, rivastigmine has also earned a Food and Drug Administration (FDA) approval for transdermal application.[6]

The ChEIs differ slightly from each other in their pharmacological profiles, which can conceivably translate into differences in drug tolerability.[14] The mechanism of action of memantine is unrelated to that of the ChEIs,[15] which provides for a distinctive safety and tolerability profile. Considering the variety of therapeutic options, and the fact that a typical patient with AD is also likely to receive concomitant medications for both related and unrelated illnesses,[16,17] we believed it is useful to provide physicians with a compilation of safety and tolerability data for each anti-AD drug, based on the prescribing information documents provided by manufacturers.


Current US prescribing information (PI) documents for donepezil,[1] galantamine,[3] rivastigmine,[5,6] and memantine[12] were obtained from American manufacturers’ websites in September 2007.[18–22] The information contained in these documents was accepted as presented, and the data were not reconciled with individually published trials or meta-analyses. Quantitative data about adverse events (AEs) from each PI document were reviewed. In addition, the odds of each AE occurring in the active versus placebo group for each drug were compared by means of odds ratios (ORs),23 which are commonly used to assess the odds of experiencing a particular AE.[2,4,13,24] Since the numbers of patients who experienced a particular AE were not given exactly, but provided in form of percentages rounded to the closest integer, we calculated the ORs using m-values, derived through multiplying available percentages by the total number of patients (e.g., the m-value of an AE that has been observed in 3% of the total of 1000 patients would be taken to be 30 [0.03 x 1000], although, due to the rounding, the actual number could have been as low as 25 and as high as 34). Statistical analyses were based on a simple linear logistic regression model, with treatment as the only explanatory variable. Descriptive safety data are presented without statistical analysis.


All of the information presented in this manuscript was taken directly from the PI documents for each drug, with the exception of the odds ratios and P values in Tables 2, 4, 6, and 7, which were calculated specifically for this manuscript.

Cholinesterase inhibitors. Donepezil. Donepezil is a reversible antagonist of acetylcholinesterase (AChE), with a plasma half-life of about 70 hours and a time to peak plasma concentration of 3 to 4 hours. Doses of 5 and 10mg, administered once daily, have been shown to be effective in patients with mild to moderate AD; for patients with severe AD, the recommended daily dose—also administered QD or once daily—is 10mg.[1]

The most frequent AEs leading to discontinuation in donepezil trials, occurring in at least two percent of patients receiving up to 10mg/day of the drug and at twice the incidence seen in placebo patients, are shown in Table 1.

Individual AEs experienced during clinical trials, reported in at least two percent of patients receiving up to 10mg/day donepezil and at a higher frequency than in placebo-treated patients are shown in Table 2 Table 2, Continued. There is evidence of a lower incidence of the most frequent gastrointestinal symptoms in patients who were titrated to the target dose of 10mg/day over six weeks, compared to the group titrated to the same dose within one week.[1] In Table 2 Table 2, Continued, the data from studies in mild-to-moderate AD are presented separately from studies in severe AD, following the approach used in the PI document. The statistically significant odds ratios of drug versus placebo in each set of data are highlighted gray.

In most cases, gastrointestinal effects have been mild and transient, usually lasting less than three weeks. Diarrhea, nausea, and vomiting were shown to occur more frequently with the 10-mg daily dose than with the 5-mg dose.

Galantamine. Galantamine is a ChEI with the plasma half-life of approximately seven hours and a time to peak plasma concentration of about one hour. Recommended doses of 16 and 24mg are administered twice daily as an immediate-release tablet or oral solution, or once daily as an extended-release formulation.[3]

Discontinuations due to AEs recorded in a single, three-arm, five-month trial of galantamine, titrated in 8-mg increments every four weeks, are shown in Table 3. In two additional six-month trials of galantamine titrated more rapidly (8–32mg/day in 8-mg increments/week), the rate of discontinuation due to AEs among the patients treated with galantamine was about three times greater than the rate observed in the placebo group. The PI document for galantamine does not compile the information for AEs leading to discontinuation across multiple trials.

Individual AEs reported in at least two percent of patients receiving 16 or 24mg/day galantamine and at a higher frequency than placebo-treated patients in four placebo-controlled trials are shown in Table 4. These data are based on studies of the immediate-release tablet formulation. In clinical trials with the extended-release capsules, once-daily treatment was well-tolerated and AEs were similar to those seen with the immediate-release tablets (this information is not reported separately in the galantamine PI document). The statistically significant odds ratios of drug versus placebo are highlighted gray.

The median duration of nausea in patients who experienced this AE was 5 to 7 days. The galantamine PI document notes that the majority of AEs occurred during the dose escalation period, and that administration of galantamine with food, the use of anti-emetic medication, and ensuring adequate fluid intake may reduce the impact of side effects. Dose reduction is recommended in patients with moderately impaired hepatic or renal function, and the use of galantamine is not recommended in patients with severe hepatic or renal impairment. There were no important differences in AE rates related to dose or sex, and there were too few non-Caucasian patients to assess tolerability by race.

Rivastigmine. Rivastigmine inhibits both AChE and butyrylcholinesterase in the central nervous system. It has a very short plasma half-life of 1 to 2 hours, thus requiring twice-daily oral dosing. Recommended oral doses are 6, 9, and 12mg/day. Due to significant gastrointestinal side effects, patients should always be started at 1.5mg BID (twice daily), and titrated to the tolerable maintenance dose.[5] A recently approved transdermal patch requires one application every 24 hours (starting dose: 4.6mg/day; recommended dose: 9.5mg/day), which allows for a slower development of the peak plasma concentration (8–16 hours vs. approximately 1 hour) and an increased half-life of 3.4 hours.[5,6] The two modes of delivery provide different safety and tolerability profiles, as demonstrated in Tables 5 and 6.[5,6]

The most frequent AEs leading to discontinuation for each method of administration, occurring in at least two percent of patients and at twice the incidence seen in placebo patients, are shown in Table 5.
Individual AEs reported in at least two percent of patients receiving 6 to 12mg/day rivastigmine as a capsule or 9.5mg/day in form of a patch, and at a higher frequency than placebo-treated patients, are shown in Table 6. Data from studies using a supratherapeutic transdermal dose of 17.4mg/day are also included in the rivastigmine PI document, but are not included in this report. All statistically significant odds ratios of drug versus placebo are highlighted gray.

For the rivastigmine capsule, weight loss of at leas seven percent of the baseline value was experienced by 26 percent women and 18 percent men on high doses of rivastigmine (>9mg/day), compared to six percent women and four percent men taking placebo. It is not clear how much weight loss was associated with other gastrointestinal problems, such as anorexia, diarrhea, nausea, and vomiting. A case of severe vomiting, including esophageal rupture, was reported following improper reinitiation of treatment (4.5mg/day) after interruption of eight weeks. In clinical trials, nausea, vomiting, and weight loss were more frequent in women than in men.

In patients with hepatic impairment and moderate renal impairment, clearance of orally administered rivastigmine is decreased by up to 65 percent. For unexplained reasons, patients with severe renal impairment have a 43-percent higher clearance of the drug than healthy subjects. Dose adjustments are reportedly not necessary in patients with renal or hepatic impairment, since the drug dose should be individually titrated to tolerability. No specific pharmacokinetic studies were conducted to investigate the effect of gender and race on the disposition of rivastigmine; however, clearance is not affected by race or sex.

The majority of participants in the rivastigmine patch trials had either no skin irritation or it was of a mild to moderate degree (it should be noted that the same site for patch application should not be used within 14 days). The incidence of severe skin irritation was very low and unrelated to the dose. No studies involving subjects with renal or hepatic impairment were conducted for the rivastigmine patch, but the PI document maintains the note that dosage adjustment is not necessary since the dose of the drug is individually titrated to tolerability. Caution should be exercised in patients with low body weight (<50kg) as these patients may experience more AEs and may be more likely to discontinue due to AEs; monitoring of patient weight is also recommended during rivastigmine patch therapy. For all methods of rivastigmine application, the drug should be titrated as prescribed, and reinitiated at the lowest dose if treatment is interrupted for more than a few days. NMDA receptor antagonists. Memantine. Memantine is an uncompetitive inhibitor of NMDA receptors, with a half-life of 60 to 80 hours. It is administered orally twice a day, with the recommended daily dose of 20mg.[12]

In clinical trials, there were no AEs that were associated with the discontinuation of treatment in one percent or more of memantine-treated patients and at a rate greater than placebo, and no individual AE occurred at a frequency of five percent or more and twice the placebo rate.

AEs reported in at least two percent of patients receiving up to 20mg/day memantine and at a higher frequency than placebo-treated patients are shown in Table 7. The statistically significant odds ratios of drug versus placebo are highlighted in gray.

The memantine PI document notes that memantine dosage should be reduced in patients with severe renal impairment. No dose adjustment is recommended for patients with mild to moderate hepatic impairment, but the drug should be administered with caution to patients with severe hepatic impairment. In addition, substances that alkalinize the urine (e.g., carbonic anhydrase inhibitors, sodium bicarbonate) could lead to a possible increase in AEs in patients taking memantine. The PI document for memantine does not indicate whether there are differential effects based on patients’ gender, age, or race.


ChEIs. Based on regulatory information from the package inserts of currently available AD therapies, several issues must be considered when optimizing treatment approaches. In general, cholinesterase inhibitors are associated with gastrointestinal side effects, particularly nausea, vomiting, and diarrhea. Across pivotal studies with donepezil, rivastigmine, and galantamine, fixed dose and forced titration have been shown to result in a higher frequency of gastrointestinal side effects, compared to studies employing a flexible-dose, flexible-titration design.[25] Therefore, a flexible, tolerability-driven approach should be taken by physicians when increasing the dose of these medications. In clinical practice, at least one month of titration is often necessary before reaching the recommended, clinically beneficial dose, which is in accordance with the guidelines presented in the PI documents. In the case of galantamine and the rivastigmine oral formulation, taking the medication with a meal may decrease the likelihood of nausea or emesis. In addition, dosage adjustments of both donepezil and galantamine should be kept in mind when treating patients with hepatic impairment or those taking several medications at a time. Data available for the transdermal rivastigmine patch suggest that its tolerability and caregiver preference are superior to those of the orally administered drug,[26–28] although it should also be noted that this formulation is the most recently introduced and the least studied of the anti-AD treatments.

In addition, it should be mentioned that, due to their mechanism of action, all three ChEIs should be monitored for an exaggerated, succinylcholine-type muscle relaxation during anesthesia, vagotonic effects on sinoatrial and atrioventricular nodes (bradycardia or heart block), an increased gastric acid secretion, gastrointestinal bleeding, urinary obstruction, and seizures. ChEIs also have the potential to interfere with the activity of anticholinergic medications, may act synergistically with other cholinomimetic drugs, and should be prescribed with care in patients with a history of asthma or obstructive pulmonary disease.[19–22] Similar to other cholinomimetics, ChEIs may induce or exacerbate extrapyramidal symptoms.[20,21]

Finally, although all ChEIs have the same molecular target, chemically they are sufficiently different to possess different pharmacokinetic and pharmacodynamic properties. Thus, major enzymes involved in metabolic processing of donepezil and galantamine are the CYP 450 isoenzymes 2D6 and 3A4. No clinical trials have investigated the effects of donepezil on the clearance of drugs metabolized by those two enzymes, but a study conducted in healthy volunteers showed that ketoconazole, an inhibitor of CYP 450 3A4, reduced donepezil clearance.[19] Similarly, drugs that are potent inhibitors of the CYP 450 isoenzymes 2D6 and 3A4 (ketoconazole, paroxetine, amitriptyline, fluoxetine, fluvoxamine, and quinidine) have been shown to reduce clearance of galantamine.[22] On the other hand, rivastigmine is primarily metabolized by esterases, undergoes only minimal modification by CYP 450 isoenzymes, and has no known metabolic drug-drug interactions.[20,21]

Clearance of all three ChEIs is decreased in patients with moderate to severe hepatic impairment; moderate to severe renal impairment is associated with a decreased clearance of galantamine and rivastigmine, but not of donepezil.[19–22] These subtle differences in metabolism and clearance should be kept in mind when making prescribing decisions and when monitoring patients for adverse reactions.
Memantine. Since memantine does not inhibit AChE, its use is not associated with cholinomimetic tolerability issues, and gastrointestinal AEs are rare. The three most common AEs associated with memantine use are confusion, dizziness, and headache, with only headache occurring at a frequency twice the one observed among the placebo-treated patients. In a 24-week pivotal trial of memantine involving patients with moderate to severe AD patients who were already taking stable doses of donepezil,[29] the rates of AEs in patients treated with both drugs were mostly similar to those observed in patients treated with donepezil alone. In addition, in the group treated with memantine and donepezil compared to the placebo-donepezil group, fewer patients experienced certain gastrointestinal AEs (diarrhea, fecal incontinence); however, statistical analyses of tolerability were not performed in this study and further studies would need to be performed to determine whether adding memantine actually improves tolerability to ChEIs. Similar to cholinesterase inhibitors, memantine requires a gradual titration (achieved in 5-mg weekly increments) before reaching the recommended dose of 20mg/day.

Memantine undergoes minimal hepatic metabolism, which makes interactions with the inhibitors of CYP 450 enzymes unlikely.[18] The drug is primarily excreted in urine, largely as the unmetabolized parent compound. The clearance of memantine is diminished in individuals with moderate to severe renal impairment, and dosage adjustment is recommended in patients with severe renal impairment.[18] Pharmacokinetic studies demonstrate that dosage adjustment is not necessary in patients with mild to moderate hepatic impairment, but caution should be used when prescribing memantine to patients with severe hepatic impairment, as metabolism in this patient population has not been studied.[18]

The principal limitation of our report is the fact that there were no head-to-head comparisons of the four drugs within a single controlled trial. We attempted to make our approach more objective by (1) limiting data sources to the AE data from PI documents only, thus using only the material that was selected by the manufacturers or FDA, and (2) performing an odds ratio analysis for each AE in order to indicate which ones were significantly more likely to occur during active treatment. The analysis was not weighted (i.e., it did not take into account differences in the number of patients reported for each drug), a consideration that should be kept in mind when comparing the statistical significance of odds ratios calculated from data sets with different numbers of patients (lower numbers of patients are associated with larger confidence intervals). Thus, the odds ratios calculated for this study, while useful for making general comparisons within each study, should be used with caution when comparing trials. It also should be noted that each PI document reports data in a slightly different format. Finally, the PI documents typically do not reference or provide details about the trials that are compiled in the data tables. The authors took the information contained therein as being representative of the most accurate data for each drug.


Since patients with AD are typically frail, have several comorbid conditions, and take multiple medications, physicians should be aware of the tolerability and safety profiles of the currently prescribed antidementia drugs. Direct, head-to-head clinical trials would be an ideal method for assessing differences in tolerability and efficacy between AD medications; unfortunately, such trials have not been performed and are not likely to be sponsored by the drug manufacturers. This study sought to overcome the lack of direct-comparison trials by analyzing the data presented by each company in its own PI material; however, caution should be exercised when comparing values obtained from different trials or trial groups.


The authors are grateful to Hai-An Hsu, PhD, of Forest Research Institute, who performed the statistical calculations and to Michael L. Miller, PhD, and Vojislav Pejovi´c, PhD, of Prescott Medical Communications Group, who provided data verification and editorial assistance.


1. Aricept®. US Prescribing Information. New York, NY: Pfizer, 2006:November.
2. Birks J, Harvey RJ. Donepezil for dementia due to Alzheimer’s disease. Cochrane Database Syst Rev. 2006(1):CD001190.
3. Razadyne ER. US Prescribing Information. Titusville, NJ: Ortho-McNeil Neurologics, 2007:April.
4. Loy C, Schneider L. Galantamine for Alzheimer’s disease and mild cognitive impairment. Cochrane Database Syst Rev. 2006(1):CD001747.
5. Exelon®. US Prescribing Information. East Hanover, NJ: Novartis, 2006.
6. Exelon Patch®. US Prescribing Information. East Hanover, NJ: Novartis, 2007:July.
7. Birks J, Grimley Evans J, Iakovidou V, et al. Rivastigmine for Alzheimer’s disease. Cochrane Database Syst Rev. 2000(4):CD001191.
8. Black SE, Doody R, Li H, McRae T, et al. Donepezil preserves cognition and global function in patients with severe alzheimer disease. Neurology. 2007;31;69(5):459–469.
9. Feldman H, Gauthier S, Hecker J, et al. A 24-week, randomized, double-blind study of donepezil in moderate to severe Alzheimer’s disease. Neurology. 2001;57(4):613–620.
10. Winblad B, Kilander L, Eriksson S, et al. Donepezil in patients with severe Alzheimer’s disease: double-blind, parallel-group, placebo-controlled study. Lancet. 2006;367(9516):1057–1065.
11. Seltzer B. Donepezil: an update. Expert Opin Pharmacother. 2007;8(7):1011–1023.
12. Namenda®. US Prescribing Information. St. Louis, MO: Forest Pharmaceuticals, Inc., 2007:April.
13. McShane R, Areosa Sastre A, Minakaran N. Memantine for dementia. Cochrane Database Syst Rev. 2006(2):CD003154.
14. Cummings JL. Alzheimer’s disease. N Engl J Med. 2004;351(1):56–67.
15. Parsons CG. Memantine: a NMDA receptor antagonist that improves memory by restoration of homeostasis in the glutamatergic system—too little activation is bad, too much is even worse. Neuropharmacology. 2007;53(6):699–723. Epub 2007 Aug 10.
16. Pitkala KH, Laurila JV, Strandberg TE, et al. Behavioral symptoms and the administration of psychotropic drugs to aged patients with dementia in nursing homes and in acute geriatric wards. Int Psychogeriatr. 2004;16(1):61–74.
17. Lacro JP, Jeste DV. Physical comorbidity and polypharmacy in older psychiatric patients. Biol Psychiatry. 1994;36(3):146–152.
18. Namenda®. US Prescribing Information. Accessed September, 2007.
19. Aricept®. US Prescribing Information. Accessed September, 2007.
20. Exelon®. US Prescribing Information. Accessed September, 2007.
21. Exelon Patch®. US Prescribing Information. Accessed September, 2007.
22. Razadyne ER. US Prescribing Information. Accessed September, 2007.
23. Bland JM, Altman DG. Statistics notes. The odds ratio. BMJ. 2000;320(7247):1468.
24. Birks J. Cholinesterase inhibitors for Alzheimer’s disease. Cochrane Database Syst Rev. 2006(1):CD005593.
25. Anand R, Hartman R, Sohn H, et al. Impact of study design and patient population on outcomes from cholinesterase inhibitor trials. Am J Geriatr Psychiatry. Mar-Apr 2003;11(2):160–168.
26. Winblad B, Grossberg G, Frolich L, et al. IDEAL: a 6-month, double-blind, placebo-controlled study of the first skin patch for Alzheimer disease. Neurology. 2007;69(4 Suppl 1):S14–22.
27. Winblad B, Kawata AK, Beusterien KM, et al. Caregiver preference for rivastigmine patch relative to capsules for treatment of probable Alzheimer’s disease. Int J Geriatr Psychiatry. 2007;22(5):485–491.
28. Winblad B, Cummings J, Andreasen N, et al. A six-month double-blind, randomized, placebo-controlled study of a transdermal patch in Alzheimer’s disease—rivastigmine patch versus capsule. Int J Geriatr Psychiatry. 2007;22(5):456–467.
29. Tariot PN, Farlow MR, Grossberg GT, et al. Memantine treatment in patients with moderate to severe Alzheimer disease already receiving donepezil: a randomized controlled trial. JAMA. 2004;291(3):317–324.