Loss of Smell Linked to Brain Abnormality, Transition to AD

Susan Jeffrey

July 14, 2014

COPENHAGEN, Denmark — Two new studies link impaired olfactory function to elevated brain amyloid and greater neurodegeneration in cognitively normal adults, as well as progression to Alzheimer's disease (AD), among patients with mild cognitive impairment (MCI).

The findings suggest testing olfactory function may be a useful screening biomarker in the diagnostic work-up for AD in patients at risk for dementia, researchers say.

The new results, from the Harvard Aging Brain Study and the North Manhattan Study, were presented here at the Alzheimer's Association International Conference (AAIC) 2014.

Amyloid Deposition, Neurodegeneration

Previous research has shown that regions of the brain that process odor input, including the olfactory bulb and entorhinal cortex, are particularly susceptible to AD pathology, and are affected at early stages of the disease, lead author Matthew E. Growdon, MA, an MD/MPH candidate at Harvard Medical School and Harvard School of Public Health, Boston, Massachusetts, pointed out.

"Building on this, there has been interest in the field to develop clinical tests of odor identification," Growdon said. "Most notably, the one we're talking about today is the University of Pennsylvania Smell Identification Test, or UPSIT [Sensonics Inc], which is a very low-cost, easy-to-administer, 40-item scratch-and-sniff test that can be used easily in the clinic."

Matthew E. Growdon

In 1 study, Growdon and colleagues looked at whether there was an association between olfaction, memory performance, and biomarkers of neurodegeneration and amyloid deposition in clinically normal individuals participating in the Harvard Aging Brain Study. The sample includes 215 community-dwelling individuals aged 64 to 88 years, who undergo annual testing, including comprehensive neuropsychological testing, hippocampal volume and entorhinal cortex thickness — both structures important for memory and both measured using MRI — and amyloid burden using Pittsburgh Compound B positron emission tomography (PET).

He stressed that this analysis is entirely cross-sectional at this point, looking at associations at the beginning of the study. "We're hoping to reanalyze as we go forward longitudinally, but I'd like to emphasize we're just looking at a snapshot in time."

Among all the individuals in their study, smaller entorhinal cortex and hippocampal size were associated with worse odor identification using the UPSIT tool, and worse memory. "These were statistically significant and modest effects," Growdon noted.

When they looked at amyloid PET findings, they found that among participants with higher amounts of amyloid deposition in their brain, a smaller entorhinal cortex was similarly associated with a worse UPSIT score (P = .03 for interaction). "That suggests at least in this small dataset that there seems to be an interplay between the size of the subjects' brains and the amount of amyloid that they have," he noted.

In future, they hope to further study olfactory testing as a screening modality in 3 groups of these patients: those who did poorly just on the odor identification (lower 25th percentile for the UPSIT test, upper 75th percentile for the memory testing, 28% of the total patients); those who did poorly just on the memory testing (upper 75th percentile for the UPSIT test, lower 25th percentile for memory testing, 9% of the total patients) and those who did poorly on both tests (lower 25th percentile for UPSIT and memory testing, 16% of patients).

"This is a hypothesis at this point, but we would like to follow all 3 of those groups over the next 5 years at least, and see who might be at higher risk of developing cognitive symptoms," he said.

"I think what's interesting here in this cross-sectional study is finding significant associations between a relatively simple, low-cost tool like the University of Pennsylvania Smell Identification Test and these more established but costly and at times invasive biomarkers like the PET scan, and the size of the brain on MRI scan," Growdon said. "I think we can say there is potentially a role for something like the [UPSIT] in clinically normal, asymptomatic older individuals, and I envision it as something that could flag those who are at risk for developing Alzheimer's disease symptoms, or potentially as a gateway to more expensive or invasive procedures like a spinal tap or a PET scan, so it could be part of a whole panel of testing."

Still, he said, it's not yet "ready for prime time," and more longitudinal follow-up is required to confirm that poor olfactory performance does translate to higher AD risk. "So I find it very intriguing, but I'm also cautious at this point about its role as a screening tool."

Transition from MCI to AD

Dr. Devangere Devanand

In a separate study, Devangere Devanand, MBBS, MD, professor of psychiatry and neurology at Columbia University Medical Center, New York, New York, presented results using the UPSIT in a group of participants in the North Manhattan Study, an ongoing epidemiologic study of a multiethnic community sample.

"The olfactory bulb and tract are affected earliest in Alzheimer's disease; in brains going to pathology; in some cases it looks as if maybe even earlier than the entorhinal cortex and the hippocampus, which is why we looked at it," Dr. Devanand pointed out. He distinguished between odor sensitivity and detection, which is not affected in AD, and odor identification, which is the aspect affected by the disease, possibly because it requires an element of interpreting or "testing" the odor against a stored memory.

The UPSIT odor identification testing was added to the North Manhattan Study evaluation during 2004 to 2006; this report looks at 2 subsequent evaluations, at 2 years (2006 to 2008) and 4 years (2008 to 2010). The majority of participants were women (69%), the mean age was 80.5 years, there was a fairly even distribution of the 3 major ethnic groups in the United States: white, African American, and Hispanic.

The test was given by a research assistant in English or Spanish based on the participants' language preference. "We didn't see a big difference between English and Spanish speakers using the English test and the Spanish test, and there is other evidence from Italy and other countries that you can use it cross-culturally," he noted.

Of 757 participants followed in this analysis, 109 transitioned to dementia, and 101 of these transitioned to AD. They found that a lower score on the UPSIT test was associated with future transition to AD, Dr. Devanand said, and was "highly significant, even after adjusting for covariates," including demographic, cognitive, and functional measures (hazard ratio, 1.099 per point interval; 95% confidence interval, 1.067 - 1.131; P < .0001).

Receiver-operating characteristic curve analyses suggested that the UPSIT test, while less effective on its own to the Selective Reminding Test-total immediate recall (SRT-TR), was additive to results of the memory test, with the combination being more effective than the SRT-TR alone (P = .01).

Prediction of cognitive decline (defined as 1 standard deviation over 4 years or 0.5 standard deviation over 2 years in those with only 2 years of follow-up in composite measures of memory, language, or visuospatial ability) in people who were normal on cognitive testing was also seen with the UPSIT test in this cohort. At this time, however, the SRT-TR memory test and SRT delayed recall test were not predictive, Dr. Devanand said (hazard ratio, 1.067 per point interval; 95% confidence interval, 1.040 - 1.095; P < .0001).

The UPSIT then may be sensitive in this early clinical stage, "whereas for people who already clinically have mild cognitive impairment, memory [testing] is as good, or even maybe slightly better than the UPSIT," Dr. Devanand said. But in predicting cognitive decline in healthy individuals, the area under the curve was not as strong. "It's significant, but it's only a moderate effect, not as strong as predicting the transition from mild cognitive impairment to Alzheimer's disease," he said.

"We think of it, not as a gold standard test, but as something to add to other tests in making the prediction," he concluded. Other conditions that affect smell include smoking, allergies, Parkinson's disease, and schizophrenia, Dr. Devanand noted, "so if it's going to be used, one must recognize this. While most of these other conditions are relatively uncommon in the older population, that has to be taken into account if one is going to use it either as a screening test or for early diagnosis."

Some modifications to the test are needed, he added. "There are 1 or 2 odors in the test which don't work well and they need to be removed," he said.

Issues of Specificity

Asked for comment on these findings, David Knopman, MD, professor of neurology at the Mayo Clinic College of Medicine, consultant in neurology at the Mayo Clinic in Rochester, Minnesota, and a member of the Alzheimer's Association Medical and Scientific Advisory Council, who moderated a press conference on this topic here, pointed out that the link between declining olfaction and AD isn't new.

"It's been around for 20 years, and the problem with it is that it lacks specificity, as was indicated," Dr. Knopman told Medscape Medical News. "People lose smell for a variety of reasons that would render the test therefore useless, for example, if they'd had some kind of viral illness in their 40s or 50s."

The test performed about the same as the memory test, he added, but as the authors noted, "it does provide a little bit of additional substantiation that there might be brain disease under certain circumstances. I can't see it being used…as a free-standing screening test," he added. "It's one piece of the puzzle that might not make it into regular use because of the specificity problems."

Creighton Phelps, PhD, deputy director of the Division of Neuroscience at the National Institute on Aging, which supported both of these studies, also underlined the issue of specificity with this test.

"The reason I think it's so important is that the parts of the brain that are affected by Alzheimer's early on are related to olfactory function, so it would be an obvious place to go and look," he told Medscape Medical News. The fact that olfaction can be easily tested leads to the idea that declines in olfaction may predict decline in those areas of the brain that are affected by AD.

"So it's very logical and makes a lot of sense, but once again, we have to be careful because the brain is degenerating in other conditions too and it might affect some of these same areas," Dr. Phelps said. "But we need something to predict who is going to either progress or even detect somebody who hasn't shown any signs otherwise, so a noninvasive test is very important. The olfactory [connection] is a solid thing and it's been used so long now, we trust it. When you correlate with the changes in cognitive function," presented here, he said, "that's what's new."

At the press conference, 2 additional papers were highlighted looking at amyloid imaging in the retina and lens in the eyes of patients at risk for AD, underlining the ongoing search for noninvasive biomarkers that might flag those at future risk for inclusion in clinical trials, or someday to receive early interventions that may prevent progression to dementia.

"In the face of the growing world-wide Alzheimer's disease epidemic, there is a pressing need for simple, less invasive diagnostic tests that will identify the risk of Alzheimer's much earlier in the disease process," Heather Snyder, PhD, director of Medical and Scientific Operations for the Alzheimer's Association, said in a statement. "For now, these 4 studies reported at AAIC point to possible methods of early detection in a research setting to choose study populations for clinical trials of Alzheimer's treatments and preventions."

The Harvard Aging Brain Study is funded by the National Institute on Aging, the Massachusetts Alzheimer's Disease Research Center, and the Alzheimer's Association. The authors have disclosed no relevant financial relationships. The North Manhattan Study is supported by the National Institute on Aging. Dr. Devanand reports he is an advisor for Lundbeck and AbbVie, both unrelated to the data presented here.

Alzheimer's Association International Conference (AAIC) 2014. Abstracts IC-P- P4-057. Presented July 13, 2014.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: