A blood test to diagnose Alzheimer’s Disease

The progression of Alzheimer's Disease (AD) is usually gradual and unfolds over many years, beginning with a long pre-clinical phase where underlying brain changes, such as the buildup of amyloid plaques and tau tangles, occur without noticeable outward symptoms. As the disease advances into the mild or early stage, individuals may experience subtle but increasingly frequent memory lapses, trouble with planning or organizing, and perhaps disorientation in familiar places. These initial clinical symptoms gradually worsen, leading into the moderate stage, characterized by more pronounced memory loss affecting personal history, increased confusion, challenges with daily tasks like dressing or managing finances, personality and behavioral changes like agitation or suspicion, and difficulty recognizing familiar people. In the severe or late stage, cognitive function declines profoundly; individuals lose the ability to communicate coherently, require extensive or total assistance with personal care like eating and bathing, lose awareness of their surroundings and recent experiences, and undergo significant physical decline, including loss of mobility and difficulty swallowing. Ultimately, individuals in the final stages become bed-bound and highly susceptible to infections like pneumonia, which are often the direct cause of death as the body's systems fail due to the widespread neurological damage.

The timetable for Alzheimer's progression varies greatly, but typically begins with a pre-clinical phase lasting years or decades with no symptoms. Mild Cognitive Impairment (MCI) may follow, potentially lasting several years before progressing. Mild dementia, which comes after MCI, often lasts around 2-4 years (MCI does not significantly impact daily life whereas mild dementia does). The moderate stage is usually the longest, spanning 2-10 years. The severe stage typically lasts 1-3 years. While the average duration from dementia diagnosis to end-of-life is about 4-8 years, the total disease course can extend 20 years or more, heavily influenced by individual factors.

Diagnosing Alzheimer's Disease typically involves multiple approaches, as there isn't one single definitive test for a living person. Clinicians usually begin by taking a detailed medical history, including symptom onset and progression, often supplemented with information from family members. This is followed by neurological exams and cognitive and neuropsychological testing to evaluate memory, problem-solving, attention, language, and other cognitive functions. Standard laboratory tests and structural brain imaging, like MRI or CT scans, are essential to rule out other potential causes of cognitive impairment, such as strokes, tumors, vitamin deficiencies, or thyroid problems. More recently, biomarkers measured through cerebrospinal fluid (CSF) analysis (obtained via lumbar puncture) or specialized positron emission tomography (PET) scans, which can detect the signature amyloid plaques and tau tangles characteristic of Alzheimer's in the brain, are increasingly used to support the diagnosis. Historically, a definitive diagnosis could only be confirmed through post-mortem brain examination (e.g. amyloid plaques).

People are typically diagnosed with Alzheimer's Disease (AD) after the age of 65, with the risk increasing significantly with advancing age. This is often referred to as late-onset Alzheimer's, or just plain Alzheimer's Disease. While less common, early-onset Alzheimer's can be diagnosed in people younger than 65, sometimes even in their 40s or 50s. At whatever age the diagnosis, it most commonly occurs when the disease has progressed to the Mild Dementia stage. This is the point when cognitive decline (like memory loss, difficulty planning, language problems) becomes pronounced enough to consistently interfere with daily activities, work, or social life, prompting the individual or their family to seek medical evaluation. Studies indicate that around 40% to 60% of Alzheimer's cases go formally undiagnosed.

There are numerous benefits for earlier and more accurate diagnosis of Alzheimer's Disease. For one, the few existing treatments for AD (QH) seem to work best (i.e. slow progression) on patients at the earliest stages. For another, it has been shown that lifestyle changes can lower the risk and slow the progression of AD. However, the existing diagnostic tools are invasive (e.g. CSF biomarkers) and expensive (imaging). The holy grail would be an accurate blood-based biomarker, which indicates presence of the disease condition.

Dr. Eric Topol has written a detailed blog post on the potential of one such biomarker, p-tau217, which has been championed by Oskar Hansson and colleagues at the University of Lund in Sweden. They showed p-tau217 levels mirrored the progression AD and effectively discriminated AD from other neurodegenerative diseases such as Parkinson's Disease or vascular dementia.

p-tau217 measures tau protein that has been abnormally phosphorylated at the threonine 217 site. Tau protein is a microtubule-associated protein found primarily in neurons, crucial for maintaining the stability and structure of microtubules (MTs), which are essential for cell shape, transport, and neuronal communication. Phosphorylation of tau detaches the protein from MTs which can enable the dynamic restructuring of the cell cytoskeleton. The problem is that excessive phosphorylation (i.e. hyperphosphorylation) causes too much tau to detach detrimentally affecting MT function, and more importantly leading to formation of tau aggregates inside of the cell that give rise to neurofibrillary tangles (NFTs), a key feature of AD pathology.

The most widely recognized feature of AD pathology is APP plaques (QH). Amyloid Precursor Protein (APP) is cleaved into small peptides called amyloid beta (Aβ) by secretases. These peptides accumulate on the outside of neurons in the brain to form amyloid plaques (aggregate deposits) that can cause neuronal cell death as it accumulates. Both extracellular APP plaques and intracellular neurofibrillary tangles are toxic to cells.

What is the relationship between the two? According to the "Amyloid Hypothesis," the disease process typically begins with the aberrant cleavage of APP, leading to the accumulation and deposition of Aβ plaques in the brain, often decades before symptoms manifest, initially spreading broadly through the cortex. These Aβ plaques are believed to initiate a toxic cascade involving inflammation, oxidative stress, and disrupted neuronal signaling, which subsequently triggers the hyperphosphorylation of tau protein. This abnormal tau detaches from microtubules, aggregates inside neurons to form NFTs, initially appearing in memory-related regions like the entorhinal cortex and hippocampus before spreading. While Aβ deposition starts the process, it is the formation and spread of tau pathology (NFTs) that correlates more strongly with neuronal dysfunction, disruption of neuronal transport, synaptic loss, and the severity and progression of clinical cognitive decline. 

Given this pathophysiology, it is not surprising that amyloid plaque accumulation and the appearance of abnormal tau protein in the form of NFTs are important biomarkers for AD along with cognitive decline. The presence of amyloid plaques can be detected from a sample of cerebrolspinal fluid (CSF) or via PET scan. Aβ42 is the most aggregation-prone peptide fragment (42 amino acids long) produced by cleavage of APP. Because Aβ42 tends to aggregate and deposit into plaques in the brain, there is less of it in the CSF. Thus a biomarker for AD is low CSF concentrations of Aβ42, which is often normalized as the Aβ42/Aβ40 ratio because Aβ40 is not aggregation-prone like Aβ42 and hence is not decreased in the CSF by the onset of AD. By being in the denominator Aβ40 levels, can adjust for inter-individual variations in overall Aβ production. The CSF sample is typically obtained via a lumbar puncture (spinal tap) by a doctor who inserts a specialized needle into the subarachnoid space which is between the meningeal layers that surround and protect the spinal cord in the lower back. The lumbar puncture can be painful.

The second main method for assessing amyloid plaque formation is by Amyloid Positron Emission Tomography (PET) Imaging. Radioactive tracers are injected intravenously (no lumbar puncture); these tracers can cross the blood-brain barrier and bind specifically to fibrillar Aβ plaques in the brain. A PET scanner then detects the radiation emitted by the tracer, creating images that visualize the density and distribution of amyloid plaques.

Similarly, the presence and quantity of abnormal tau protein can be measured from the CSF. Elevated levels of total tau (t-tau) in the CSF reflect the amount of neuronal degeneration, as tau is released from damaged or dying neurons, but it is not specific to AD because other neurodegenerative diseases or acute neuronal injury (e.g. stroke) can also release tau. On the other hand, specific forms of hyperphosphorylated tau (e.g., p-tau181, p-tau231, and p-tau217) are more specific to the NFT pathology seen in AD, and reflect the ongoing process of tangle formation. In addition, just as with PET imaging of amyloid, one can inject radioactive tracers that specifically bind to aggregated tau (primarily NFTs) in the brain. PET scans then visualize this tau pathology.

Together the build up amyloid and p-tau deposits/tangles provide valuable information about AD status (stage) and prognosis. Amyloid is an earlier indicator while tau correlates more closely with the severity of cognitive impairment. The main diagnostic drawbacks are cost (PET imaging) and inconvenience (CSF spinal tap).

From this lengthy exposition, one can appreciate the benefits of a blood-based AD biomarker that is not overly expensive or invasive to measure. Amazingly, such a test may be within reach. As mentioned above, new assays can now detect specific p-tau species (i.e. p-tau217) in blood samples. Elevated levels of these plasma p-tau markers show a strong correlation with brain amyloid pathology and tau PET positivity, making them highly promising tools for inferring the presence of AD pathology. Plasma p-tau217, in particular, has shown high accuracy in discriminating AD from other neurodegenerative diseases.

A number of studies have demonstrated that the p-tau217 blood test is as accurate as CSF tau biomarker tests, and comparable or even superior to tau-PET scans . For example, p-tau217 levels change earlier than abnormalities appear on tau-PET scans, and can potentially be used for staging the disease instead of PET.

Even more importantly, p-tau217 appears to track amyloid plaque accumulation which as mentioned above occurs earlier than tau tangle formation. A Swedish study from the Hansson group studied 150 cognitively unimpaired participants and 100 patients with mild cognitive impairment (MCI). The p-tau217 blood biomarker was measured every two years for up to 6 years. CSF Aβ levels (to assess amyloid deposits) were measured at the start of the study, and from that subjects were categorized as Aβ+ or Aβ-. Over the course of the follow-up period, 104 individuals developed dementia; 48 of these cases were Alzheimer's disease, and 56 were other types. The remaining 145 individuals did not develop dementia.

Figure 1A (below) focuses on the cognitively unimpaired (CU, no dementia symptoms) group that were divided into Aβ+ or Aβ- based on the CSF measurement. Over the 6 year study period, presumably as amyloid deposits accumulated, blood p-tau217 increased dramatically for the Aβ+ subjects whereas there was little to no change in the Aβ- subjects. As early as year 2, the blood test could distinguish between the two groups. In addition, there was likely a correlation between rising p-tau217 levels and the disease phenotype; among the Aβ- group 0/88 converted to AD dementia compared to the Aβ+ group in which 8/62 developed AD dementia.

For those with mild cognitive impairment (MCI), the p-tau217 blood biomarker showed a similar trend. Among Aβ+ MCI patients, p-tau217 increased significantly (Figure 1B), whereas among Aβ- MCI patients there was once again little to no change. For MCI patients, compared to CU patients, there was a much higher probability of progressing to AD if you were Aβ+ (37/49); those who were Aβ- were much less likely to convert (3/50) thereby providing further evidence of amyloid deposits as a prognosticator of AD.

Given the close relationship between Aβ+ and AD conversion, it is not surprising that p-tau217 also correlated well with AD conversion (Fig 1C looks similar to Fig 1B). MCI patients who later converted to Alzheimer's Disease dementia exhibited rapidly increasing p-tau217, whereas it did not change much in patients with MCI who did not convert to AD. 

Thus in this study, plasma p-tau217 was highly effective in distinguishing between Aβ+ and Aβ− (Figs. 1A and 1B), and between AD conversion and AD non-conversion subjects (Fig 1C). Other studies have demonstrated the ability to distinguish between NFT+ and NFT− individuals. In other words, p-tau217 levels closely track the build-up of amyloid plaques and NFTs, reflecting the overall Alzheimer's Disease process, which is distinct from other neurodegenerative diseases.

Intriguingly, blood p-tau217 appears to be able to distinguish between those likely to be afflicted with Alzheimer's Disease (i.e. Aβ+) even (possibly) among clinically unimpaired (CU) individuals i.e. those not showing any dementia symptoms yet. If one examines the left side of Figure 1A, the blood test cannot quite distinguish between Aβ+ from Aβ- at the start of the study for CU subjects, but within the first year, p-tau217 increases enough in Aβ+ cohort to create significant separation between the two groups. The study does not mention when CU individuals begin showing signs of mild cognitive impairment during the 6 year study, but presumably within the first year many may remain unimpaired.

Because AD is a neurodegenerative disorder, the earlier one takes action the better. The p-tau217 biomarker provides a temporal window for flagging the risk of the disease possibly many years before symptoms. Interestingly, p-tau217 levels are dynamic, meaning they can change in response to interventions. Studies have demonstrated that treatments and lifestyle changes aimed at reducing amyloid burden can lower p-tau217. For example, exercise has been shown to lower p-tau217 levels. Because there is a close correlation between p-tau217 and AD cognitive decline (e.g. Fig. 1C), an important open question is whether decreasing p-tau217 can lead to slowing, halting, or even reversing AD progression. 

At the same time, it is important to acknowledge the limitations of p-tau217 whose diagnostic accuracy of AD is high (~80-90%), but not perfect. One reason why errors arise is the immense complexity of brain aging (e.g. mixed pathologies). Many individuals diagnosed with AD also have co-existing pathologies such as vascular dementia, Lewy bodies, or TDP-43 proteinopathy. Elevated p-tau217 indicates the presence of AD pathology, but does not rule out these other conditions. Thus, the p-tau217 biomarker should be interpreted within a clinical context, not as an infallible standalone test. A patient's dementia symptoms may be driven by a combination of factors.

In conclusion, a blood test for p-tau217 represents a major breakthrough, opening new possibilities for diagnosing, monitoring, predicting, and potentially delaying the progression of Alzheimer's Disease. This biomarker shows a strong correlation with established indicators of AD, such as amyloid plaques and tau tangles. Unlike fixed genetic risk factors like APOE4 or polygenic risk scores, repeated p-tau217 testing could provide insight into how an individual’s risk and disease progression evolve over time. Importantly, rising p-tau217 levels may signal the presence of AD years before noticeable symptoms, such as mild cognitive impairment, appear. This early detection window offers a valuable opportunity for preventive measures, including lifestyle changes or future treatments. Dr. Topol believes that such testing will be most useful for people with an elevated risk due to factors like genetics or family history, rather than as a routine screening tool for the general public, partly to avoid false positives. For now, he advises that testing should be voluntary, respecting each person’s choice about learning their risk status.

Figure 1. Blood plasma p-tau217 biomarker levels track progression in Alzheimer's patients in three groups of patients. The y-axis is blood plasma p-tau217 concentrations, and the x-axis is time in years. (A) Cognitively healthy (unimpaired, CU) individuals with brain amyloid (Aβ+, purple) versus those without (Aβ-, green). (B) Individuals with Mild Cognitive Impairment (MCI) with and without brain amyloid. (C) Individuals with MCI who later developed Alzheimer's dementia (purple) versus all other individuals with MCI (those who remained stable or developed other forms of dementia, green). Figure from Mattsson-Carlgren et al. Brain, 2020).