Read this topic

Basic Neurosciences · Paper A

Memantine in Alzheimer's disease

The idea

Memantine is not another cholinesterase inhibitor — it works by calming excess glutamate, not by boosting acetylcholine. Cholinesterase inhibitors (donepezil, galantamine, rivastigmine) are the guideline's first drugs, for mild-to-moderate Alzheimer's disease. Memantine has its own separate slot: moderate disease when a cholinesterase inhibitor cannot be used, and severe disease, where it is often added to what came before rather than replacing it. The exam question is rarely 'is this an Alzheimer's drug' — it is which specific trigger, severity, tolerance or progression, points to memantine rather than a cholinesterase inhibitor.

How it works

Memantine is an NMDA (N-methyl-D-aspartate) receptor antagonist: it blocks the receptor in a voltage-dependent, use-dependent way, preferentially shutting down the low-level, chronic glutamate signalling that becomes pathologically excessive in Alzheimer's disease, while still letting the brief bursts of physiological glutamate signalling needed for learning and memory pass through.
Cholinesterase inhibitors (donepezil, galantamine, rivastigmine) work on a completely different system: they block the breakdown of acetylcholine, raising levels of the neurotransmitter lost as cholinergic neurons die off in Alzheimer's disease.
Guideline place in therapya cholinesterase inhibitor is the option for mild-to-moderate Alzheimer's disease; memantine monotherapy is for moderate disease with cholinesterase-inhibitor intolerance or contraindication, and for severe disease.
When disease progresses to moderate-to-severe on an already-tolerated cholinesterase inhibitor, guidance favours ADDING memantine to it, not switching within the cholinesterase-inhibitor class or replacing it with memantine alone.
Rivastigmine carries a specific extra licence — dementia associated with Parkinson's disease — that donepezil and galantamine do not, and is also available as a transdermal patch.
For behavioural and psychological symptoms of dementia (agitation, aggression, distress), non-pharmacological management and addressing triggers comes first; an antipsychotic is reserved for severe distress or real risk of harm, at the lowest dose for the shortest time.

How to spot it

A vignette describing newly diagnosed mild-to-moderate Alzheimer's disease, independent for daily activities, with no drug started yet, is pointing to a cholinesterase inhibitor as the first treatment.
A vignette describing severe Alzheimer's disease, or moderate disease with a genuine contraindication or intolerance to a cholinesterase inhibitor (e.g. new heart block, intolerable nausea), is pointing to memantine.
A vignette describing a patient who has tolerated a cholinesterase inhibitor well for months or years but has now declined to a more severe stage is pointing to adding memantine to the existing drug, not switching or stopping it.
A vignette naming Parkinson's disease alongside a new dementia diagnosis is pointing to rivastigmine specifically, not donepezil or galantamine.
A vignette describing mild agitation with an identifiable trigger and no risk of harm is pointing to non-pharmacological management, not a drug.
A vignette describing vascular dementia with no coexisting Alzheimer's, Parkinson's or Lewy body features, asking about drug treatment, is testing whether you know there isn't one — the answer is vascular risk-factor control.

Tell it apart

Mechanismmemantine blocks excess NMDA-receptor glutamate signalling; cholinesterase inhibitors raise acetylcholine by blocking its breakdown — two unrelated systems, not competing versions of the same drug.
Severity triggera cholinesterase inhibitor is for mild-to-moderate disease; memantine monotherapy is for moderate disease with cholinesterase-inhibitor intolerance/contraindication, or for severe disease.
Progression on treatmentworsening despite a well-tolerated cholinesterase inhibitor means ADD memantine, not switch to a different cholinesterase inhibitor and not switch to memantine alone.
Poor tolerance of the drug itself, not disease progression, is what justifies switching within the cholinesterase-inhibitor class instead.
Coexisting Parkinson's disease dementia specifically favours rivastigmine over donepezil or galantamine, on licensing, not on any difference in cholinergic mechanism.
Vascular dementia with no coexisting Alzheimer's, Parkinson's or Lewy body features has no specific drug treatment — management is vascular risk-factor control.
Agitation or aggression in dementia is managed non-pharmacologically first; an antipsychotic is reserved for severe distress or real risk of harm, not the default first response.

Watch for

⚠️ Assuming memantine is simply 'a stronger version' of a cholinesterase inhibitor, rather than a drug acting on a completely different receptor system.
Reaching for memantine the moment a stem mentions Alzheimer's disease, without checking whether the disease is actually mild-to-moderate (where a cholinesterase inhibitor is the guideline answer) rather than moderate-with-intolerance or severe.
Switching to a different cholinesterase inhibitor, or to memantine alone, when a patient's disease has simply progressed on a WELL-TOLERATED cholinesterase inhibitor — guidance says add memantine to it instead.
Treating donepezil, galantamine and rivastigmine as fully interchangeable and missing that rivastigmine alone carries the licence for dementia associated with Parkinson's disease.
Starting an antipsychotic as the automatic response to any agitation or aggression in dementia, rather than checking for an identifiable trigger and trying non-pharmacological measures first.
Assuming a drug must be indicated for every dementia diagnosis — vascular dementia with no coexisting Alzheimer's, Parkinson's or Lewy body features has no specific pharmacological treatment at all.

Use it again

First askis this actually Alzheimer's disease, and if so, what severity — mild-to-moderate points to a cholinesterase inhibitor, moderate-with-intolerance or severe points to memantine.
If the stem describes a cholinesterase inhibitor already in use and being tolerated, check whether the change described is disease progression (add memantine) or poor tolerance of the drug itself (switch within the cholinesterase-inhibitor class) — these two triggers demand opposite actions.
If a coexisting neurological diagnosis is named (Parkinson's disease, Lewy body dementia), check whether it points to a specific licensed drug (rivastigmine) rather than the general Alzheimer's pathway.
If the stem centres on behaviour (agitation, aggression, distress) rather than cognition, resist jumping to a drug at all — check for an identifiable trigger and non-pharmacological options first, and reserve an antipsychotic for severe distress or real risk of harm.
If the dementia described is vascular with no coexisting features, do not force a dementia-drug answer — the guideline answer is often 'no specific pharmacological treatment', with vascular risk-factor management instead.
Never let 'this patient has dementia' alone decide the answer — the decisive detail is always severity, tolerance, progression, coexisting diagnosis, or the presence of a genuine risk.

In detail

Hook — the pill that isn't "more of the same"

A junior doctor on a memory clinic list assumes every dementia drug does the same job, just with different names. Then a consultant asks her to explain why a patient who has taken donepezil for two years, and tolerated it perfectly well, is now being started on a second tablet rather than switched to a stronger one. She has no answer, because she has been treating "Alzheimer's drug" as one category instead of two completely different jobs. This page builds the second job — memantine's — from zero, and shows exactly where the two jobs meet.

Plain definition

Memantine is a drug for Alzheimer's disease that works on a completely different system from the more familiar cholinesterase inhibitors (drugs such as donepezil, galantamine and rivastigmine that raise levels of acetylcholine, the chemical messenger nerve cells use to communicate). Memantine instead blocks the (N-methyl-D-aspartate receptor, a docking site on nerve cells that responds to glutamate, the brain's main excitatory chemical messenger). In Alzheimer's disease, damaged nerve cells leak glutamate constantly at a low level, and that constant leak slowly damages the cells further — a process called excitotoxicity (cell damage from too much excitatory stimulation). Memantine dampens that constant leak.

The critical caveat — it blocks the noise, not the signal

Here is the trap: assume memantine blocks glutamate signalling altogether, and you would expect it to wipe out memory rather than protect it, since glutamate is also how the brain forms new memories moment to moment. That is not what happens, and the reason is how memantine binds. It sits in the receptor's channel only when the channel has been open too long — exactly what happens during the pathological, constant leak of excitotoxicity — and lets go quickly when a brief, normal burst of glutamate arrives for real learning. So the constant background noise gets blocked while the real signal still gets through. This is why memantine can be added to daily life without blunting the very memory function the disease is destroying.

First principles — why the guideline splits into two drug classes

Start from what is actually failing in Alzheimer's disease: the acetylcholine-producing nerve cells die off early, and separately, ongoing glutamate-driven excitotoxicity keeps damaging surviving cells throughout the illness. A cholinesterase inhibitor answers the first problem: by blocking the enzyme that breaks acetylcholine down, it makes the acetylcholine that remains last longer, propping up communication between the cells that are left. Memantine answers the second, unrelated problem: it does not restore any chemical, it slows the ongoing damage.

Because the two drugs solve two different problems, NICE guidance (NG97) does not treat them as steps on the same ladder, where you simply move to memantine when a cholinesterase inhibitor stops working well enough. Instead, each is matched to a specific situation: a cholinesterase inhibitor for mild-to-moderate disease, and memantine specifically for moderate disease where a cholinesterase inhibitor cannot be used, or for severe disease.

Structured walk-through — placing one patient on the pathway

Take, for example, a 78-year-old woman diagnosed with mild Alzheimer's disease three years ago. At diagnosis, independent and mildly forgetful, she was started on donepezil — the guideline's mild-to-moderate answer. She tolerated it well, and her disease slowly progressed anyway, as Alzheimer's disease does, to moderate-to-severe over the following two years.

Now ask what should change. If she were newly intolerant of donepezil — new bradycardia, unmanageable nausea — the answer would be memantine as a like-for-like replacement, matching the "cannot tolerate a cholinesterase inhibitor" trigger. But she is not intolerant; she has simply declined further on a drug that still suits her. That is the trigger for adding memantine to the donepezil she already takes, not for stopping it. The distinction — intolerance versus progression — is the entire exam question dressed up as a clinical vignette.

Look-alikes & differentials

Looks like a memantine case	How it actually differs
Mild-to-moderate Alzheimer's disease, newly diagnosed	Needs a cholinesterase inhibitor first — memantine is not the starting drug here
Dementia associated with Parkinson's disease	Rivastigmine specifically holds this licence, not memantine and not the other cholinesterase inhibitors
Vascular dementia with no coexisting Alzheimer's, Parkinson's or Lewy body features	No specific drug treatment at all — the answer is vascular risk-factor control
Agitation or aggression in dementia	Managed non-pharmacologically first; an antipsychotic, not memantine, is the drug reserved for severe distress or real risk of harm

Diagnosis

There is no blood test or scan for "which drug does this patient need" — that decision rests on a careful clinical assessment that has already established the Alzheimer's disease diagnosis (cognitive testing, a history of gradual decline, and imaging or other investigations used to exclude reversible causes). From that assessment you read off three things: how severe the disease is right now, whether a cholinesterase inhibitor has been tried and how it was tolerated, and whether any coexisting condition, Parkinson's disease for instance, points to a specific licensed drug.

Management

NICE NG97 (dementia: assessment, management and support) recommends a cholinesterase inhibitor — donepezil, galantamine or rivastigmine — as the option for managing mild to moderate Alzheimer's disease. Memantine monotherapy is recommended for moderate disease in people who cannot tolerate, or have a contraindication to, a cholinesterase inhibitor, and for people with severe disease. For moderate-to-severe disease progressing on an already-tolerated cholinesterase inhibitor, memantine can be added to it rather than switching drugs.

Suppose the same patient later develops evening agitation: guidance still asks you to look for a trigger and try non-pharmacological measures first. Any antipsychotic is kept for severe distress or genuine risk of harm, at the lowest dose for the shortest time, because of the raised risk of stroke and mortality antipsychotics carry in this population.

Exam pearl

Whenever a vignette hands you a dementia patient and a drug question, resist naming "an Alzheimer's drug" and instead name the trigger: severity at diagnosis, of what has already been tried, whether the change described is progression or intolerance, and whether a coexisting diagnosis points to a specific licence. Before you move on, test yourself: if a patient has tolerated donepezil well for two years and simply declined further, can you explain why the answer is adding memantine rather than switching to it?

In one line

Memantine calms excess glutamate signalling rather than boosting acetylcholine, and guidance reserves it for moderate disease with cholinesterase-inhibitor intolerance, for severe disease, or as an add-on when a tolerated cholinesterase inhibitor stops being enough on its own.