While drug companies have been working hard to come up with new options for the treatment of dementia, there have been no new FDA approvals since memantine (Namenda) hit the scene in 2003. The currently available meds include four cholinesterase inhibitors (donepezil, galantamine, rivastigmine and tacrine) and one NMDA receptor antagonist (memantine). Frankly, they have only modest efficacy, but these are the only pharmacological tools we have—so let’s review their clinical utility.
Before starting, however, there is an important point to be made about the nature of the clinical trial evidence. Although these medications can sometimes temporarily improve the symptoms of dementia, they usually work by modestly slowing the decline of cognition. It is important to explain this point to family members, who might otherwise expect to see actual improvement in cognition, which is unusual.
Remembering how cholinesterase inhibitors (CIs) work is helpful in remembering their side effects. Acetylcholine, one of the neurotransmitters involved with memory, is inactivated by the enzyme acetylcholinesterase. The four CIs work by blocking the acetylcholinesterase and increasing acetylcholine in the synaptic cleft—leading to the nasty side effects of nausea, vomiting and dizziness that we find with these medications. In addition, tacrine (Cognex) can cause hepatotoxicity, and is so rarely used that we won’t discuss it in detail.
All three of the commonly used CIs perform equally well, according to a meta-analysis of 13 randomized double blind placebo-controlled trials (Birks J., Cholinesterase inhibitors for Alzheimer’s Disease. Cochrane Database of Systematic Reviews 2006).
Donepezil (Aricept) is generally well tolerated, easy to dose and has an added bonus in that it is the only CI that is FDA approved to treat severe dementia as well as mild to moderate dementia. Start at 5 mg every morning and go up to 10 mg after four weeks. Dosing it in the morning may prevent the side effect of insomnia and vivid dreams some patients report on the medication. According to the Cochrane Review, donepezil causes fewer side effects than rivastigmine.
Galantamine (Razadyne) comes in both immediate-release tablets (with initial dosing starting at 4 mg twice a day), or an extended-release tablet (an initial dose of 8 mg once a day). The dose can be increased every four weeks to a maximum of 24 mg per day. Galantamine’s original boast was that it has a dual mechanism of action, working on nicotinic receptors as well. However, accumulating evidence seems to show that this does not make any difference in efficacy (Loy C et al., Galantamine in Alzheimer’s Disease. Cochrane Database of Systematic Reviews 2002. Updated 2006). One piece of bad news is that galantamine may cause a slightly higher mortality rate than placebo when given off-label to patients with mild cognitive impairment (ibid).
Both donepezil and galantamine are substrates of CYP2D6 and CYP3A4, so medications that inhibit these enzymes (such as some SSRIs and duloxetine) can increase levels of both CIs, potentially worsening side effects.
Rivastigmine (Exelon), indicated for mild to moderate AD, is the only CI FDA-approved to treat dementia associated with Parkinson’s disease. Unfortunately, the oral form causes particularly high rates of nausea, vomiting, weight loss and syncope, and the titration to higher doses is difficult for many patients to tolerate (Birks J et al., Rivastigmine for Alzheimer’s Disease. Cochrane Database of Systematic Reviews 2000. Updated 2009). The good news is that rivastigmine is now offered in a transdermal patch which is better tolerated. The target dose of the patch is 9.5 mg per 24 hours, which is similar to oral dose of 6 mg twice a day. Start with a 4.6 mg per 24 hours patch for four weeks and then move up to the 9.5 mg per 24 hours patch. Unlike donepezil and galantamine, rivastigmine is not metabolized by the liver and therefore presents little risk of drug interactions.
Since all of the CIs are equally effective, we recommend starting with donepezil, given its once a day dosing and good tolerability profile. Donepezil will soon be available in a cheaper generic form (probably by the end of 2010).
The perpetually vexing issue in dementia treatment is how long to continue CIs once you’ve started. Most consensus guidelines recommend trying another CI if there is no clear slowing of cognitive decline in the first six to nine months, although the research on this is sparse (Tariot P et al., J Clin Psychiatry 2006;67(suppl 3):2002–2013). Most experts believe that you can switch from one CI to another without a washout period, but there is one case report of a fatal adverse event during a transition from donepezil to rivastigmine. (For a review of the literature on this topic, see Burns A et al., J Psychopharmaco 2006; 20(6):732–755).
Memantine (Namenda), the fifth medication used to treat AD, is an Nmethyl-D-aspartate (NMDA) receptor antagonist, and may work by slowing the activation of NMDA receptors caused by excessive glutamate release.
Memantine is FDA-approved for the treatment of moderate to severe AD only, and a 2006 Cochrane Review showed that it has a small beneficial effect on both cognition and behavior after six months in such patients (McShane R et al., Memantine for Dementia. Cochrane Database of Systematic Reviews 2005. Updated 2006). Memantine may also be effective as an augmentation agent added to donepezil in patients with moderate to severe AD. One large randomized, placebo-controlled study of 404 patients with moderate to severe AD showed significant but modest improvements in cognition, activities of daily living, and behavior in those receiving donepezil and memantine as compared to those receiving donepezil and placebo. Confusion was the most common adverse event attributable to memantine (7.9% on memantine vs 2.0% on placebo; P = .01) (Tariot et al., JAMA 2004;(3)291: 317–324).
We recommend adding memantine when dementia has progressed to the moderate or severe level. Usual dosing starts at 5 mg per day for the first week, 5 mg twice a day for the second week, 10 mg in the morning and 5 mg at bedtime for the third week, and a final increase to 10 mg twice a day in the fourth week. A recent study showed that switching from donepezil to memantine (for reasons of tolerability or functional decline) can be done safely either abruptly or through cross titration (Waldemar et al., Int Jrn of Geriatric Psychiatry 2008;23(9): 979-981).
Another pressing question is whether these medications work for agitation or behavioral issues in dementia patients. While atypical antipsychotics are often used for agitation, they are associated with a small excess risk of mortality and are best used sparingly. Recently, some retrospective pooled analyses have shown that memantine can be helpful for agitation and aggression in AD. But these are not clinical trials, and are therefore less compelling. (For recent research, see Wilcock GK et al., J Clin Psychiatry 2008;69(3):341–348). The only clinical trial of a CI for agitation was negative: a 2007 study of 272 patients with AD randomly assigned to either donepezil or placebo showed no significant differences between the two for behavioral symptoms (Howard RJ et al., N Eng J Med 2007; 357(14) 1382–1392).
TCPR VERDICT: Start with Aricept, add Namenda if needed.