Some people who scored badly on the test were later found to have early signs associated with Alzheimer's disease. Previous research has shown people's sense of smell gets worse as they get older. People with dementia seem to have an even worse sense of smell and ability to identify smells.
But simple odour identification tests do not account for variation in different people's sense of smell.
Researchers in the US tested 183 people to see if they could identify 10 common smells, including lemon, mint and strawberry.
They then did a second test to see whether people could identify 20 smells and remember the 10 they'd smelled in the first test.
The second test was better at identifying people with Alzheimer's as well as early symptoms of dementia.
It also picked out people who had no signs of Alzheimer's, but who carried gene variants connected with the disease.
We now need further research in more people to be sure that the findings are correct.
If you do lose your sense of smell (anosmia), you shouldn't panic – there could be a relatively trivial reason behind it, such as chronic sinusitis. But it is the sort of symptom you should get checked out by your GP.
Where did the story come from?The study was carried out by researchers from Massachusetts General Hospital, the University of North Carolina, Harvard School of Public Health, and Osmic Enterprises, all in the US.
It was funded by grants from the US National Institutes of Health, the Wilkens Foundation and Harvard Neurodiscovery Centre.
The study was published in the peer-reviewed journal, Annals of Neurology.
The Mail Online seems to have misunderstood some aspects of the study. It says the participants were "patients at the Massachusetts General Hospital" who "were deemed to have an increased risk" of Alzheimer's disease.
In fact, they were a mix of volunteers aged over 65 and living at home. Ten of them already had Alzheimer's disease, but most were healthy.
The point of the study was to see whether the test could pick out people at increased risk, not to test people already known to be at increased risk.
The Mail Online story also mistakenly said the test could pick out those with a build-up of amyloid protein in their brains, but this study found no link between amyloid protein and the test results.
What kind of research was this?This cross-sectional cohort study looked at how people performed on smell tests at one point in time.
Researchers wanted to see whether this was related to their mental health or other markers linked to Alzheimer's disease.
What did the research involve?Researchers recruited people taking part in a long-term study of ageing and dementia, and five people with dementia from a memory clinic.
They were given standard tests to identify dementia and early signs of dementia, known as mild cognitive impairment.
Some people also had brain scans and genetic testing for gene variants linked to dementia.
They took three tests to assess their sense of smell, memory for smells, and ability to discriminate between smells.
Researchers then looked at the results to see whether – taking account of possible confounding factors like age and education level or medical reasons for a poor ability to smell – the smell test results could predict people with dementia or at higher risk of dementia.
The three tests were:
- 10 common smells – people were asked if they recognised the smell and if they could identify it from a list of four names
- 20 common smells, including the 10 from the first test – people were asked if they'd smelled the odour in the first test and to identify it from a list of four names
- 12 smells – two smells were presented one after the other, and people were asked to say whether they were the same or different
Researchers used a battery of statistical tests to see which factors correlated with which. Their primary interest was whether the test results predicted people's diagnoses (normal, some concerns, mild cognitive impairment, or Alzheimer's disease).
They also wanted to see if the smell test results were linked to other early predictors of Alzheimer's disease, such as:
- degeneration of certain regions of the brain
- deposits of amyloid protein in the brain
- gene variants thought to be more common in people with Alzheimer's disease
What were the basic results?People who were cognitively normal or had only some concerns about their memory tended to do well on the POEM test.
And their results were significantly better than those of people with mild cognitive impairment or Alzheimer's disease.
When researchers looked at people who were cognitively normal but did worse than expected on the POEM test based on their results from the first (10 smells) test, they found these people were more likely to have:
- a gene variant associated with Alzheimer's disease
- thinner tissue in a part of the brain associated with memory (the entorhinal cortex)
- worse logical memory scores over time
We don't know if the people who did worse on smell tests went on to get Alzheimer's disease, as this was not part of the study.
A study with a longer-term follow-up period would be required to investigate this.
How did the researchers interpret the results?The researchers said the POEM test needed to be confirmed in longer studies and different groups of people.
However, they said if these results were confirmed, the POEM test "may identify a subset of clinically normal participants at greater risk for developing the progressive memory symptoms" of Alzheimer's disease.
They say this could identify suitable people to take part in research of treatments that might prevent the disease. They also suggest the tests could be used to screen for Alzheimer's disease risk in the general population.
ConclusionSense of smell varies greatly from one person to another, and tends to decline as we get older. Lots of people can lose their sense of smell – either temporarily or permanently – after illness or an accident.
Having a poor sense of smell does not mean you're going to get Alzheimer's disease, and that's not what this study found.
People who already had Alzheimer's disease, not surprisingly, did poorly at identifying smells.
But smell detection ability alone did not differentiate between healthy people, those with some memory concerns, and those with mild cognitive impairment.
Only the POEM test, which looked at people's ability to both identify and remember smells, could do that.
For people without dementia or mild cognitive impairment, those who did less well at remembering smells compared with their ability to identify them were more likely to have previously identified risk factors for Alzheimer's disease.
These include a genetic variant more common in those with Alzheimer's disease and physical evidence of some degree of tissue thinning.
But we don't know whether these people did go on to get dementia, as the study only looked at a snapshot in time, not at what happened to people over time. It's important to remember, too, that this was a relatively small study.
We need to have these POEM test results validated by larger studies that follow people over time before we can say whether it is a useful way of identifying older people likely to develop Alzheimer's disease before they have any symptoms of confusion or memory loss.
If you're concerned about symptoms that might be related to Alzheimer's disease or other forms of dementia, see your GP.
Read more about how dementia is diagnosed.