ARTICLE IN BRIEF
Researchers described findings from a new algorithm used to study clinical differences among subtypes of hippocampal-sparing Alzheimer's disease patients.
Atypical Alzheimer's disease (AD) has long baffled patients and neurologists, often masquerading as other conditions, including frontotemporal dementia, corticobasal syndrome, or primary progressive aphasia. In an analysis presented at the AAN Annual Meeting in Philadelphia, Melissa Murray, PhD, an assistant professor of neuroscience at the Mayo Clinic in Jacksonville, FL, and colleagues, set out to examine the clinical relevance of hippocampal sparing (HpSp) in Alzheimer's disease patients compared with the more common form of the disease.
From the brain bank at the Mayo Clinic in Jacksonville, FL, Dr. Murray and colleagues identified more than 1800 neuropathologically confirmed cases of AD. They then identified the HpSp AD cases using an algorithm, which assessed the relative distribution of hippocampal and neocortical neurofibrillary tangles. [Read about the group's earlier work here: http://1.usa.gov/1qA952h.]
In the current analysis, they designed a new algorithm that further stratified the HpSp patterns according to the region of the brain most affected, based on the relative proportion of neurofibrillary tangles.
“I have a huge interest in atypical Alzheimer's disease especially because the hippocampal sparing forms are so often misdiagnosed. It's probably one of the questions I get the most from neurologists — how can we better identify these patients? We found that all of these patients have both tangles and amyloid plaques, which are necessary to neuropathologically diagnose as Alzheimer's. Interestingly, they don't have different amyloid patterns from typical AD patients; it's just the tau that varies,” Dr. Murray told Neurology Today.
Patients with HpSp AD represent about 11 percent of all AD patients, and around 50 percent or more can be misdiagnosed, she said. “We really tried to look at different presentations — similar to stroke, where depending on where the infarct occurs, you'll have a specific set of clinical presentations.”
Their findings, she explained, emphasize the much earlier age of disease onset in these atypical patients. It suggests, for example, that “if you have somebody coming into your clinic, and they're having one of these focal-cortical syndromes,” Alzheimer's disease should not be ruled out as a diagnosis. Dr. Murray hopes that use of the algorithm will help improve the understanding of atypical AD patients and eventually increase diagnostic accuracy.
STUDY METHODOLOGY, FINDINGS
Using thioflavin-S microscopy, Dr. Murray and colleagues identified 1821 neuropathologically confirmed cases of AD through their brain bank. They then applied an algorithm based on hippocampal and cortical neurofibrillary tangle pathology to classify HpSp AD, typical AD, and limbic-predominant AD. Within the final cohort of 187 HpSp AD cases, a new algorithm was created based on the patterns of focal cortical neurofibrillary tangle pathology to further divide the individuals into three groups according to the cortical region of the brain most affected — frontal, parietal, or temporal. The researchers then observed differences in these groups.
Unlike typical AD, which is believed to be more common in women, HpSp AD is significantly more common in men. According to this analysis, the gender distribution was relatively similar among the three HpSp groups. A frequent misdiagnosis for those patients with frontal HpSp AD was frontotemporal dementia (47 percent); for parietal HpSp AD, corticobasal syndrome (60 percent); and for temporal HpSp AD patients, primary progressive aphasia was diagnosed in 40 percent (p=.023).
The investigators did note striking differences in the age of onset for cognitive dysfunction among the groups: frontal and parietal presented at an average age of 59 and 61, respectively, compared with the temporal group, which presented at an average age of 69 (p<0.001).
The duration of the disease did not differ significantly among the three groups, although it was significantly shorter than typical AD (p=.009). There were, however, some differences in the average age at death: 70 and 69 for the frontal and parietal groups, but 78 in the temporal group (p<0.001).
“These atypical hippocampal sparing patients tend to be younger and have a much faster decline — losing up to 5 points per year on the MMSE [Mini-Mental State Examination]. So, for example, if you have a patient presenting with behavioral changes in their late 50s or early 60s, your first inclination may be to think frontotemporal dementia. But if you have other information such as amyloid imaging, it may help with the differential diagnosis, given these patients all have amyloid pathology.”
Our next step, said Dr. Murray, “is to understand what the early symptoms are that might actually help a neurologist identify these patients” in the clinic. The other important piece, she added, is the genetics. “Our algorithm operationally classifies these patients, highlighting the neuropathologic and clinical heterogeneity of Alzheimer's disease, and may for the first time point us in the right direction of why there is selective vulnerability — is it genetics that's pointing to it? Is there a difference in their risk factors?”
David A. Wolk, MD, an assistant professor in the department of neurology at the University of Pennsylvania and assistant director of the Penn Memory Center, called the analysis “very nice work” with a “remarkable cohort” in terms of the number of autopsy cases included.
What was striking, he added, but consistent with previous literature, “was that 10 or 11 percent of the population had these atypical versions of Alzheimer's disease, meaning this atypical presentation is much more common than people realize.”
Further, the patients included in this analysis only represent the extremes, he said. “My suspicion is that even within the 89 percent that didn't fall into these hippocampal sparing groups, there probably are varying levels of involvement in cortical regions relative to hippocampal involvement.”
Another important point from this abstract, said Dr. Wolk, was the high level of misdiagnosis based on the symptoms and the cognitive phenotype in this group. He noted that this is “not surprising given that these patients wouldn't present with the typical memory problems that we would expect for Alzheimer's disease.”
Dr. Wolk believes that these findings may allow researchers to unlock genetic and other underlying factors involved in these presentations and the variability of the disease. “I think defining these cases and trying to learn more about how they differ in regard to demographic information, genetic details, disease course — all of these may actually help us understand the heterogeneity of Alzheimer's disease, unlock keys to disease pathophysiology, and ultimately help in tailoring more personalized therapeutics.”
Additionally, this analysis may be helpful in selecting “more pure populations in clinical trials …and the study of these patients may be enhanced by use of in vivo molecularly based markers of AD to see how the course transpires over time.”
Perhaps, Dr. Wolk continued, it may also shed light on how AD pathology moves within the brain. “There have been arguments about the notion that these misfolded proteins are spread in kind of a synapse-to-synapse or cell-to-cell manner. So understanding why in some cases you would have limbic-predominant or medial-temporal-predominant cases that don't spread to cortical areas, or why sometimes the pathology is contained in the cortex and doesn't move into the medial temporal lobes, may help us gain insight into what drives this process. This may then ultimately point to how to slow down or stop that kind of transmission from occurring in the brain, if that's really the mechanism.”
One limitation of the study is that an autopsy dataset will not allow for contemporaneous phenotypic information, “so trying to understand what these patients were like at earlier stages of the disease, before they expired, and where the pathology started and what the symptoms were initially present is always hard to do retrospectively.”
Dr. Wolk said the analysis raises these among other questions: “How often do patients who start out with more cortically-based diseases turn into more typical AD, or even vice-versa? Are there patients who start off with more hippocampally-based disease, but over time, the degeneration accelerates cortically relative to the medial temporal lobes? That's something you can never completely get from a cross-sectional analysis of autopsy data.”
Ultimately, this research serves as an important reminder that “there may be different mechanisms at play that drive these differences in disease presentation, so it's really useful to not to lump all of AD together in one group and to start to try tease apart these differences,” Dr. Wolk said.