Understanding Alzheimer’s

Although Alzheimer’s disease was first named in the early 1900s, the phenomenon of “senility” has been remarked on since ancient times. Plato believed that because advancing age seemingly “gives rise to all manners of forgetfulness as well as stupidity,” older men were unsuited for leadership positions requiring acumen or judgment. William Shakespeare gave us an unforgettable portrayal of an old man struggling with his failing mind in King Lear.

The notion that this might be a disease was first suggested by Dr. Alois Alzheimer, a psychiatrist who worked as medical director at the state asylum in Frankfurt, Germany. In 1906, while performing an autopsy on a patient named Auguste Deter, a woman in her midfifties who had suffered from memory loss, hallucinations, aggression, and confusion in her final years, he noticed that something was clearly wrong with her brain. Her neurons were entangled and spiderweb-like, coated with a strange white dental substance. He was so struck by their odd appearance that he made drawings of them.

Another colleague later dubbed this condition “Alzheimer’s disease,” but after Alzheimer himself died in 1915 (of complications from a cold, at fifty-one), the disease he had identified was more or less forgotten for fifty years, relegated to obscurity along with other less common neurological conditions such as Huntington’s and Parkinson’s diseases as well as Lewy body dementia. Patients with the kinds of symptoms that we now associate with these conditions, including mood changes, depression, memory loss, irritability, and irrationality, were routinely institutionalized, as Auguste Deter had been. Plain old “senility,” meanwhile, was considered to be an inevitable part of aging, as it had been since Plato’s day.

It wasn’t until the late 1960s that scientists began to accept that “senile dementia” was a disease state and not just a normal consequence of aging. Three British psychiatrists, Garry Blessed, Bernard Tomlinson, and Martin Roth, examined the brains of seventy patients who had died with dementia and found that many of them exhibited the same kinds of plaques and tangles that Alois Alzheimer had observed. Further studies revealed that a patient’s degree of cognitive impairment seemed to correlate with the extent of plaques found in his or her brain. These patients, they concluded, also had Alzheimer’s disease. A bit more than a decade later, in the early 1980s, other researchers identified the substance in the plaques as a peptide called amyloid-beta. Because it is often found at the scene of the crime, amyloid-beta was immediately suspected to be a primary cause of Alzheimer’s disease.

Amyloid-beta is a by-product that is created when a normally occurring substance called amyloid precursor protein, or APP, a membrane protein that is found in neuronal synapses, is cleaved into three pieces. Normally, APP is split into two pieces, and everything is fine. But when APP is cut in thirds, one of the resulting fragments then becomes “misfolded,” meaning it loses its normal structure (and thus its function) and becomes chemically stickier, prone to aggregating in clumps. This is amyloid-beta, and it is clearly bad stuff. Laboratory mice that have been genetically engineered to accumulate amyloid-beta (they don’t naturally) have difficulty performing cognitive tasks that are normally easy, such as finding food in a simple maze. At the same time, amyloid also triggers the aggregation of another protein called tau, which in turn leads to neuronal inflammation and, ultimately, brain shrinkage. Tau was likely responsible for the neuronal “tangles” that Alois Alzheimer observed in Auguste Deter.

Scientists have identified a handful of genetic mutations that promote very rapid amyloid-beta accumulation, all but ensuring that someone will develop the disease, often at a fairly young age. These mutations, the most common of which are called APP, PSEN1, and PSEN2, typically affect the APP cleavage. In families carrying these genes, very-early-onset Alzheimer’s disease is rampant, with family members often developing symptoms in their thirties and forties. Luckily, these mutations are very rare, but they occur in 10 percent of early-onset Alzheimer’s cases (or about 1 percent of total cases). People with Down syndrome also tend to accumulate large amounts of amyloid plaques over time, because of genes related to APP cleavage that reside on chromosome 21.

It was not a huge leap to conclude, on the basis of available evidence, that Alzheimer’s disease is caused directly by this accumulation of amyloid-beta in the brain. The “amyloid hypothesis,” as it’s called, has been the dominant theory of Alzheimer’s disease since the 1980s, and it has driven the research priorities of the National Institutes of Health and the pharmaceutical industry alike. If you could eliminate the amyloid, the thinking has been, then you could halt or even reverse the progression of the disease. But it hasn’t worked out that way. Several dozen drugs have been developed that target amyloid-beta in one way or another. But even when they succeed in clearing amyloid or slowing its production, these drugs have yet to show benefit in improving patients’ cognitive function or slowing the progression of the disease. Every single one of them failed.

One hypothesis that emerged, as these drugs were failing one after another, was that patients were being given the drugs too late, when the disease had already taken hold. It is well known that Alzheimer’s develops slowly, over decades. What if we gave the drugs earlier? This promising hypothesis has been tested in large and well-publicized clinical trials involving people with an inherited mutation that basically predestines them to early-onset Alzheimer’s, but those, too, failed in the end. A broader clinical trial was launched in 2022 by Roche and Genentech, testing early administration of an anti-amyloid compound in genetically normal people with verified amyloid accumulation in their brains but no clear symptoms of dementia; results are expected in 2026. Some researchers think that the disease process might be reversible at the point where amyloid is present but not tau, which appears later. That theory is being tested in yet another ongoing study.

Meanwhile, in June of 2021, the FDA gave approval to an amyloid-targeting drug called aducanumab (Aduhelm). The drug’s maker, Biogen, had submitted data for approval twice before and had been turned down. This was its third try. The agency’s expert advisory panel recommended against approving the drug this time as well, saying the evidence of a benefit was weak or conflicted, but the agency went ahead and approved it anyway. It has met a tepid reception in the marketplace, with Medicare and some insurers refusing to pay its $28,000 annual cost unless it is being used in a clinical trial at a university.

This cascade of drug failures has caused frustration and confusion in the Alzheimer’s field, because amyloid has long been considered a signature of the disease. As Dr. Ronald Petersen, director of the Mayo Clinic Alzheimer’s Disease Research Center, put it to The New York Times in 2020: “Amyloid and tau define the disease….To not attack amyloid doesn’t make sense.”

But some scientists have begun to openly question the notion that amyloid causes all cases of Alzheimer’s disease, citing these drug failures above all. Their doubts seemed to be validated in July of 2022, when Science published an article calling into question a widely cited 2006 study that had given new impetus to the amyloid theory, at a time when it had already seemed to be weakening. The 2006 study had pinpointed a particular subtype of amyloid that it claimed directly caused neurodegeneration. That in turn inspired numerous investigations into that subtype. But according to the Science article, key images in that study had been falsified.

There was already plenty of other evidence calling into question the causal relationship that has long been assumed between amyloid and neurodegeneration. Autopsy studies have found that more than 25 percent of cognitively normal people nevertheless had large deposits of amyloid in their brains when they died—some of them with the same degree of plaque buildup as patients who died with severe dementia. But for some reason, these people displayed no cognitive symptoms. This was not actually a new observation: Blessed, Tomlinson, and Roth noted back in 1968 that other researchers had observed “plaque formation and other changes [that] were sometimes just as intense in normal subjects as in cases of senile dementia.”

Some experts maintain that these patients actually did have the disease but that its symptoms had been slower to emerge, or that they had somehow masked or compensated for the damage to their brains. But more recent studies have found that the reverse can also be true: some patients with all the symptoms of Alzheimer’s disease, including significant cognitive decline, have little to no amyloid in their brains, according to amyloid PET scans and/or cerebrospinal fluid (CSF) biomarker testing, two common diagnostic techniques. Researchers from the Memory and Aging Center at University of California San Francisco found via PET scans that close to one in three patients with mild to moderate dementia had no evidence of amyloid in their brains. Still other studies have found only a weak correlation between the degree of amyloid burden and the severity of disease. It appears, then, that the presence of amyloid-beta plaques may be neither necessary for the development of Alzheimer’s disease nor sufficient to cause it.

This raises another possibility: that the condition that Alois Alzheimer observed in 1906 was not the same condition as the Alzheimer’s disease that afflicts millions of people around the world. One major clue has to do with the age of onset. Typically, what we call Alzheimer’s disease (or late-onset Alzheimer’s disease) does not present in significant numbers until age sixty-five. But Dr. Alzheimer’s own Patient Zero, Auguste Deter, showed severe symptoms by the time she was fifty, a trajectory more in line with early-onset Alzheimer’s disease than with the dementia that slowly begins to afflict people in their late sixties, seventies, and eighties. A 2013 analysis of preserved tissue from Auguste Deter’s brain found that she did in fact carry the PSEN1 mutation, one of the early-onset dementia genes. (It affects the cleavage of the amyloid precursor protein, producing loads of amyloid.) She did have Alzheimer’s disease, but a form of it that you get only because you have one of these highly deterministic genes. Our mistake might have been to assume that the other 99 percent of Alzheimer’s disease cases progress the way hers did.

This is not all that uncommon in medicine, where the index case for a particular disease turns out to be the exception rather than the rule; extrapolating from this one case can lead to problems and misunderstanding down the road. At the same time, if Auguste Deter’s illness had appeared when she was seventy-five instead of fifty, then perhaps it might not have seemed remarkable at all.

Just as Alzheimer’s disease is defined (rightly or wrongly) by accumulations of amyloid and tau, Lewy body dementia and Parkinson’s disease are associated with the accumulation of a neurotoxic protein called alpha-synuclein, which builds up in aggregates known as Lewy bodies (first observed by a colleague of Alois Alzheimer’s named Friedrich Lewy). The APOE e4 variant not only increases someone’s risk for Alzheimer’s but also significantly raises their risk of Lewy body dementia as well as Parkinson’s disease with dementia, further supporting the notion that these conditions are related on some level.

All this places high-risk patients like Stephanie in a terrible predicament: they are at increased risk of developing a disorder or disorders whose causes we still do not fully understand, and for which we lack effective treatments. That means we need to focus on what until fairly recently was considered a taboo topic in neurodegenerative disease: prevention.

Can Neurodegenerative Disease Be Prevented?

Stephanie was terrified. I had treated other patients who had two copies of the e4 gene, but none of them had responded with this much fear and anxiety. It took us four long discussions over a period of two months just to get her past the initial shock of the news. Then it was time to have a talk about what to do.

She had no obvious signs of cognitive impairment or memory loss. Yet. Just as a side note, some of my patients freak out because they lose their car keys or their phones from time to time. As I keep reminding them, that does not mean they have Alzheimer’s disease. It generally only means that they are busy and distracted (ironically, often by the very same cell phones that they keep misplacing). Stephanie was different. Her risk was real, and now she knew it.

With my highest-risk patients, like Stephanie, I would at that time typically collaborate with Dr. Richard Isaacson, who opened the first Alzheimer’s disease prevention clinic in the United States in 2013. Richard remembers that when he first interviewed with the dean of Weill Cornell Medical College and described his proposed venture, she seemed taken aback by his then-radical idea; the disease was not considered to be preventable. Also, at barely thirty, he did not look the part. “She was expecting Oliver Sacks,” he told me. “Instead, she was sitting across the table from Doogie Howser.”

Isaacson had gone off to college at age seventeen, to a joint BA-MD program at the University of Missouri in Kansas City and had his medical degree in hand by the time he was twenty-three. He was motivated to learn as much as possible about Alzheimer’s disease by the fear that it ran in his family.

When he was growing up in Commack, Long Island, Richard had watched his favorite relative, his great-uncle Bob, succumb to Alzheimer’s disease. When Richard was three, Uncle Bob had saved him from drowning in a swimming pool at a family party. But about a decade later, his beloved uncle had begun to change. He repeated his stories. His sense of humor faded. He developed a vacant stare and seemed to take longer to process what he was hearing around him. He was eventually diagnosed with Alzheimer’s disease—but he was already gone. It was as if he had vanished.

Richard couldn’t stand to watch his favorite uncle be reduced to nothing. He was also worried because he knew that Alzheimer’s disease risk has a strong genetic component. Was he also at risk? Were his parents? Could the disease somehow be prevented?

As he went into medical practice at the University of Miami, he began collecting every single study and recommendation he could find on possible ways to reduce risk of Alzheimer’s disease, gathering them into a photocopied sheaf that he would hand out to patients. He published the sheaf in 2010 as a book titled Alzheimer’s Treatment, Alzheimer’s Prevention: A Patient and Family Guide.

He soon found out that in Alzheimer’s disease circles, the very word prevention was somewhat off limits. “The Alzheimer’s Association said, you can’t really say this,” he remembers. That same year, a panel of experts assembled by the National Institutes of Health declared, “Currently, firm conclusions cannot be drawn about the association of any modifiable risk factor with cognitive decline or Alzheimer’s disease.”

That led Isaacson to realize that the only way to show that a preventive approach could work, in a way that would be accepted by the broader medical community, was in a large academic setting. Cornell was willing to take a chance, and his clinic opened in 2013. It was the first of its kind in the country, but now there are half a dozen similar centers, including one in Puerto Rico. (Isaacson now works for a private company in New York; his colleague, Dr. Kellyann Niotis, has joined my practice, focusing on patients at risk for neurodegenerative diseases.)

Meanwhile, the idea of preventing Alzheimer’s disease began to gain scientific support. A two-year randomized controlled trial in Finland, published in 2015, found that interventions around nutrition, physical activity, and cognitive training helped maintain cognitive function and prevent cognitive decline among a group of more than 1,200 at-risk older adults. Two other large European trials have found that multidomain lifestyle-based interventions have improved cognitive performance among at-risk adults. So there were signs of hope.

---

To a typical doctor, Stephanie’s case would have seemed pointless: She had no symptoms and she was still relatively young, in her early forties, a good two decades before any clinical dementia was likely to develop. Medicine 2.0 would say there was nothing for us to treat yet. In Medicine 3.0, this makes her an ideal patient, and her case an urgent one. If there were ever a disease that called for a Medicine 3.0 approach—where prevention is not only important but our only option—Alzheimer’s disease and related neurodegenerative diseases are it.

Frankly, it might seem odd for a practice as small as ours to employ a full-time preventive neurologist such as Kellyann Niotis. Why do we do this? Because we think we can really move the needle by starting early and being very rigorous in how we quantify and then try to address each patient’s risk. Some patients, like Stephanie, are at obvious higher risk; but in a broader sense, all of us are at some risk of Alzheimer’s disease and other neurodegenerative disease.

Viewed through the lens of prevention, the fact that Stephanie had the APOE e4/e4 genotype was actually good news, in a way. Yes, she was at far higher risk than someone with e3/e3, but at least we knew the genes we were up against and the likely trajectory of the disease, if she developed it. It is more worrying when a patient presents with an overwhelming family history of dementia, or the early signs of cognitive decline, but does not carry any of the known Alzheimer’s risk genes, such as APOE e4 and a few others. That means that there could be some other risk genes in play, and we don’t know what they might be. Stephanie was at least facing a known risk. This was a start.

She had two additional risk factors that were out of her control: being Caucasian and being female. While those of African descent are at an overall increased risk of developing Alzheimer’s disease, for unclear reasons, APOE e4 seems to present less risk to them than to people of Caucasian, Asian, and Hispanic descent. Regardless of APOE genotype, however, Alzheimer’s disease is almost twice as common in women than in men. It is tempting to attribute this to the fact that more women live to age eighty-five and above, where incidence of the disease pushes 40 percent. But that alone does not explain the differential. Some scientists believe there may be something about menopause, and the abrupt decline in hormonal signaling, that sharply increases the risk of neurodegeneration in older women. In particular, it appears that a rapid drop in estradiol in women with an e4 allele is a driver of risk; that, in turn, suggests a possible role for perimenopausal hormone replacement therapy in these women.

Menopause is not the only issue here. Other reproductive history factors, such as the number of children the woman has had, age of first menstruation, and exposure to oral contraceptives, may also have a significant impact on Alzheimer’s risk and later life cognition. And new research suggests that women are more prone to accumulate tau, the neurotoxic protein we mentioned earlier. The end result is that women have a greater age-adjusted risk of Alzheimer’s, as well as faster rates of disease progression overall, regardless of age and educational level.

While female Alzheimer’s patients outnumber men by two to one, the reverse holds true for Lewy body dementia and Parkinson’s, both of which are twice as prevalent in men. Yet Parkinson’s also appears to progress more rapidly in women than in men, for reasons that are not clear.

Parkinson’s is also tricky genetically: While we have identified several gene variants that increase risk for PD, such as LRRK2 and SNCA, about 15 percent of patients diagnosed have some family history of the disease—and are therefore presumed to have a genetic component—but do not have any known risk genes or SNPs.

---

Like the other Horsemen, dementia has an extremely long prologue. Its beginnings are so subtle that very often the disease isn’t recognized until someone is well into its early stages. This is when their symptoms go beyond occasional lapses and forgetfulness to noticeable memory problems such as forgetting common words and frequently losing important objects (forgetting passwords becomes a problem too). Friends and loved ones notice changes, and performance on cognitive tests begins to slip.

Medicine 2.0 has begun to recognize this early clinical stage of Alzheimer’s, which is known as mild cognitive impairment (MCI). But MCI is not the first stage on the long road to dementia: a large 2011 analysis of data from the UK’s Whitehall II cohort study found that subtler signs of cognitive changes often become apparent well before patients meet the criteria for MCI. This is called stage I preclinical Alzheimer’s disease, and in the United States alone over forty-six million people are estimated to be in this stage, where the disease is slowly laying the pathological scaffolding in and around neurons but major symptoms are still largely absent. While it’s not clear how many of these patients will go on to develop Alzheimer’s, what is clear is that just as most of an iceberg lies beneath the surface of the ocean, dementia can progress unnoticed for years before any symptoms appear.

The same is true of other neurodegenerative diseases, although they each have different early warning signs. Parkinson’s may show up as subtle changes in movement patterns, a frozen facial expression, stooped posture or shuffling gait, a mild tremor, or even changes in a person’s handwriting (which may become small and cramped). Someone in the early stages of Lewy body dementia may exhibit similar physical symptoms, but with slight cognitive changes as well; both may exhibit alterations in mood, such as depression or anxiety. Something seems “off,” but it’s hard for a layperson to pinpoint.

This is why an important first step with any patient who may have cognitive issues is to subject them to a grueling battery of tests. One reason I like to have a preventive neurologist on staff is that these tests are so complicated and difficult to administer that I feel they are best left to specialists. They are also critically important to a correct diagnosis—assessing whether the patient is already on the road to Alzheimer’s disease or to another form of neurodegenerative dementia, and how far along they might be. These are clinically validated, highly complex tests that cover every domain of cognition and memory, including executive function, attention, processing speed, verbal fluency and memory (recalling a list of words), logical memory (recalling a phrase in the middle of a paragraph), associative memory (linking a name to a face), spatial memory (location of items in a room), and semantic memory (how many animals you can name in a minute, for example). My patients almost always come back complaining about the difficulty of the tests. I just smile and nod.

The intricacies and nuances of the tests give us important clues about what might be happening inside the brains of patients who are still very early in the process of cognitive change that goes along with age. Most importantly, they enable us to distinguish between normal brain aging and changes that may lead to dementia. One important section of the cognitive testing evaluates the patient’s sense of smell. Can they correctly identify scents such as coffee, for example? Olfactory neurons are among the first to be affected by Alzheimer’s disease.

Specialists such as Richard and Kellyann also become attuned to other, less quantifiable changes in people on the road to Alzheimer’s disease, including changes in gait, facial expressions during conversations, even visual tracking. These changes could be subtle and not recognizable to the average person, but someone more skilled can spot them.

The trickiest part of the testing is interpreting the results to distinguish among different types of neurodegenerative disease and dementia. Kellyann dissects the test results to try and trace the likely location of the pathology in the brain, and the specific neurotransmitters that are involved; these determine the pathological features of the disease. Frontal and vascular dementias primarily affect the frontal lobe, a region of the brain responsible for executive functioning such as attention, organization, processing speed, and problem solving. So these forms of dementia rob an individual of such higher-order cognitive features. Alzheimer’s disease, on the other hand, predominantly affects the temporal lobes, so the most distinct symptoms relate to memory, language, and auditory processing (forming and comprehending speech)—although researchers are beginning to identify different possible subtypes of Alzheimer’s disease, based on which brain regions are most affected. Parkinson’s is a bit different in that it manifests primarily as a movement disorder, resulting from (in part) a deficiency in producing dopamine, a key neurotransmitter. While Alzheimer’s can be confirmed by testing for amyloid in the cerebrospinal fluid, these other forms of neurodegeneration are largely clinical diagnoses, based on testing and interpretation. Thus, they can be more subjective, but with all these conditions it is critical to identify them as soon as possible, to allow more time for preventive strategies to work.

One reason why Alzheimer’s and related dementias can be so tricky to diagnose is that our highly complex brains are adept at compensating for damage, in a way that conceals these early stages of neurodegeneration. When we have a thought or a perception, it’s not just one neural network that is responsible for that insight, or that decision, but many individual networks working simultaneously on the same problem, according to Francisco Gonzalez-Lima, a behavioral neuroscientist at the University of Texas in Austin. These parallel networks can reach different conclusions, so when we use the expression “I am of two minds about something,” that is not scientifically inaccurate. The brain then picks the most common response. There is redundancy built into the system.

The more of these networks and subnetworks that we have built up over our lifetime, via education or experience, or by developing complex skills such as speaking a foreign language or playing a musical instrument, the more resistant to cognitive decline we will tend to be. The brain can continue functioning more or less normally, even as some of these networks begin to fail. This is called “cognitive reserve,” and it has been shown to help some patients to resist the symptoms of Alzheimer’s disease. It seems to take a longer time for the disease to affect their ability to function. “People that have Alzheimer’s disease and are very cognitively engaged, and have a good backup pathway, they’re not going to decline as quickly,” Richard says.

There is a parallel concept known as “movement reserve” that becomes relevant with Parkinson’s disease. People with better movement patterns, and a longer history of moving their bodies, such as trained or frequent athletes, tend to resist or slow the progression of the disease as compared to sedentary people. This is also why movement and exercise, not merely aerobic exercise but also more complex activities like boxing workouts, are a primary treatment/prevention strategy for Parkinson’s. Exercise is the only intervention shown to delay the progression of Parkinson’s.

But it’s difficult to disentangle cognitive reserve from other factors, such as socioeconomic status and education, which are in turn linked to better metabolic health and other factors (also known as “healthy user bias”). Thus, the evidence on whether cognitive reserve can be “trained” or used as a preventive strategy, such as by learning to play a musical instrument or other forms of “brain training,” is highly conflicted and not conclusive—although neither of these can hurt, so why not?

The evidence suggests that tasks or activities that present more varied challenges, requiring more nimble thinking and processing, are more productive at building and maintaining cognitive reserve. Simply doing a crossword puzzle every day, on the other hand, seems only to make people better at doing crossword puzzles. The same goes for movement reserve: dancing appears to be more effective than walking at delaying symptoms of Parkinson’s disease, possibly because it involves more complex movement.

This was one thing that Stephanie, a high-performing, well-educated professional, had in her favor. Her cognitive reserve was very robust, and her baseline scores were strong. This meant that we likely had plenty of time to devise a prevention strategy for her, perhaps decades—but given her increased genetic risk, we could not afford to delay. We needed to come up with a plan. What would that plan look like? How can this seemingly unstoppable disease be prevented?

We’ll start by taking a closer look at changes that might be happening inside the brain of someone on the road to Alzheimer’s. How are these changes contributing to the progression of the disease, and can we do anything to stop them or limit the damage?

Once we begin looking at Alzheimer’s disease outside the prism of the amyloid theory, we start to see certain other defining characteristics of dementia that might offer opportunities for prevention—weaknesses in our opponent’s armor.