This site is intended for health professionals only

At the heart of general practice since 1960

pul jul aug2020 cover 80x101px
Read the latest issue online

Independents' Day

Memory problems

When is a patient's memory loss due to Alzheimer's and

when does it have another cause?

Dr Melanie Wynne-Jones and Dr Paul Bannister discuss the problems

What's in this article

 · Who can benefit from referral to old age psychiatry service

 · Distinguishing multi-infarct dementia from Alzheimer's disease

 · Counselling patients and the worried well

What's in this article

Who can benefit from referral to old age psychiatry service

Distinguishing multi-infarct dementia from Alzheimer's disease

Counselling patients and the worried well

Referral criteria for memory problems

How do you approach memory problems?

Age-related memory decline is common, although you see some phenomenal 90- year-olds. If someone develops memory impairment as they get older you need to ask two questions. First, is there something else physical going making their memory dysfunctional (sepsis or other causes of brain failure)? This poses less of a problem for GPs than hospitals because you usually see people over time and know what their premorbid condition was like.

Other processes mimic memory loss, but as part of a more widespread dysfunction such as depression, schizophrenia or Parkinson's disease. These have specific diagnostic criteria and need to be excluded.

If you're happy nothing else is going on, and that you're dealing with some sort of age-related memory decline, you can move to the second question: can you apply a specific psychiatric diagnosis, for example multi-infarct dementia, Alzheimer's disease or mixed picture ­ and if you can, do they meet the therapeutic criteria to benefit from the old-age psychiatry service?

This may include drugs1 but even people too severely demented for anticholinesterases can still get help and support from the clinic service.

Given many people complain about memory as they get older, what referral criteria should we use?

The 30-point Mini Mental State Examination is sufficient for screening. If they score 28-30 they certainly won't meet entry criteria for medication. We are most enthusiastic about treating people who score between 20 and 26, but below this they may still benefit. It's worth referring anyone with a score of 27 or below, as the MMSE is a 10-minute test that simply flags up patients who warrant more detailed memory testing that may take several hours.

What is the typical rate of memory decline?

Both multi-infarct dementia and Alzheimer's disease are hugely variable. Some have quite aggressive disease, whereas in others it's fairly benign. You can often get a fairly historical feel if you talk to patients and their families; you can extrapolate back to when they were functioning well and gauge how dysfunctional they have become, before even doing full formal testing.

It's usually a downward progression

from losing their keys to stopping looking after themselves. Wandering is always a bad sign, of course, and so is poor personal hygiene.

Multi-infarct dementia or Alzheimer's disease?

How do you tell multi-infarct dementia from Alzheimer's disease?

Alzheimer's is ultimately a pathological diagnosis but the Hachinski ischaemic score2 (see box) can help distinguish it from vascular dementia in people with previous stroke (a Hachinski score of five or more would make you less suspicious that the patient has Alzheimer's).

You should take presence of CHD or peripheral vascular disease into account and a CT scan may demonstrate multi-infarct dementia. An MR scan showing shrinkage of temporal lobe volume is more suggestive of Alzheimer's, although this investigation is not performed routinely on the NHS.

It's often a mixed picture in individual patients and they can still be treated if they meet the criteria for true Alzheimer's. Anticholinesterase drugs are not licensed for multi-infarct dementia at present but I suspect they will be in the future and at that stage it will probably come down to an assessment of the end-user ­ in other words, whether they make the brain work a bit better, regardless of what has damaged it.

If you suspect multi-infarct dementia you may need to give them the stroke work-up, including an ECHO and carotid dopplers, to ensure you're not missing treatable causes for embolic disease (and we still see people with AF who would benefit but have not been anticoagulated).

You also need to check other risk factors ­ hypertension is by far the biggest ­ and consider using aspirin and/or a statin.

How do you run the memory clinic?

I run it jointly with one of the psychiatrists so we both flag up people we have screened in our other clinics. We also accept referrals.

I investigate the physical aspects and the psychiatrist looks at the mental health side, and then we decide whether the patient warrants more detailed assessment by the clinical psychologist.

She uses a number of tests such as the CANTAB3 and is responsible for monitoring if drugs are prescribed.

Multi-infarct dementia accounts for 40 per cent of the patients we see ­ of those with Alzheimer's, between a half and a third get treated.

How is treatment managed?

If a patient meets the suitability criteria for for treatment with anticholinesterases they get a home visit from the dementia specialist nurse who checks their health, arranges blood tests, a chest X-ray, CT scan and an ECG (to check for prolonged Q-T interval).

If their results are suitable they can have treatment, but even this is not a simple decision because they have to be able to take multiple daily doses reliably and be willing to have their LFTs monitored.

They also have to agree to be assessed over the next six months and the drug will be stopped if it doesn't seem to be working.

What are the benefits and risks of treatment?

About a third of suitable patients do moderately well on anticholinesterases, a third are perhaps marginally improved, and a third are not helped at all.

If their score stays the same after three months they are allowed to remain on the drug, because you would expect there to have been some decline otherwise, if the original diagnosis was robust.

Those are about the numbers you would expect from the trials. It's important to ensure a patient meets the trial criteria of mild to moderate disease and most people are very good about not letting the entry criteria creep.

You can give drugs a bad name if you use them for patients they were not designed for, and most places take the National Institute for Clinical Excellence guidelines very seriously, only prescribing at an MMSE score of 20 to 25.

Counselling patients and dealing with the worried well

How do you counsel patients?

The words Alzheimer's disease can strike fear into patients' hearts and many with AD actually become depressed as well, so it needs to be handled sensitively. I tend to say we all get memory impairment as we get older, but that we're very keen to find out which people with moderate memory impairment may have a treatable cause or explanation for it. That's relatively positive, as is saying the drugs have a one-in-three chance of improving your memory, or tests show you do not actually need any drugs.

Advice about driving causes more problems than anything, as stopping driving can be literally disabling. You have to get a feel for how people handle new information, and how quickly they can respond to it. If you have any concerns you must notify the DVLA which can arrange formal assessment, but patients don't like that.

If you suspect memory loss might be due to depression, would a trial of antidepressants be reasonable, and would any delay in diagnosing Alzheimer's be significant?

I would expect some improvement within six to eight weeks of starting an SSRI and in terms of the dementia timescale that's not going to be critical ­ the process of assessment is fairly lengthy anyway.

What about stopping treatment?

This can be a problem and carers are often more reluctant than the patient, even when assessment has not demonstrated any benefit. But we don't automatically withdraw the rest of service when we withdraw the drugs. We do what we always used to do before cholinesterase inhibitors came along, and for the majority who do not actually turn out to be suitable for therapy ­ we provide an extensive hospital and community service. This is based on societal, environmental and behavioural treatments; patients who are discharged retain quick access through CPNs who provide very high levels of skilled support.

What about the worried well?

They present an increasing challenge. True familial Alzheimer's is uncommon, although people carrying the apolipoprotein allele ApoE4 are more at risk4. There are no proven preventive measures for Alzheimer's although NSAIDs and iron chelation therapy have been suggested.

A balanced healthy diet and avoiding hypertension are important for

maintaining cognitive function. Continued mental exercise is also believed to be protective so we should advise people to stay active and sociable.


1 Alzheimer's disease ­ donepezil, rivastigmine and galantamine (NICE appraisal no. 19,

January 2001)


2 Molsa PK et al. Validity of clinical diagnosis in dementia: a prospective clinicopathological study.

J Neurol Neurosurg Psychiatry. 1985 Nov;48:1085-90

3 Canterbury Neuropsychological Test Automated Battery


4 Alzheimer's Society


Melanie Wynne-Jones is a GP in Marple, Cheshire

Paul Bannister is consultant in care of the elderly, Manchester Royal Infirmary, Central Manchester Hospitals Trust

Hachinski Ischaemia Score

Stroke Internet Centre

A Hachinski score of 5 or more would make you less suspicious that the patient has Alzheimer's

Feature Score

Abrupt onset 2

Stepwise deterioration 1

Fluctuating course 2

Nocturnal confusion 1

Relative preservation of personality 1

Depression 1

Somatic complaints 1

Emotional incontinence 1

History of hypertension 1

History of strokes 2

Evidence of associated atherosclerosis 1

Focal neurological symptoms 2

Focal neurological signs 2

Total score

Rate this article 

Click to rate

  • 1 star out of 5
  • 2 stars out of 5
  • 3 stars out of 5
  • 4 stars out of 5
  • 5 stars out of 5

0 out of 5 stars

Have your say