Chapter 2

Striving to Standardize Alzheimer Disease

Depending on where you set your sights, Alzheimer disease is a scientific puzzle, a medical whodunit, a psychosocial tragedy, a financial disaster or an ethical, legal and political dilemma.

—Zaven Khachaturian, “Plundered Memories”¹

The historian of medicine Charles Rosenberg writes, “[T]he modern history of diagnosis is inextricably related to disease specificity, to the notion that diseases can and should be thought of as entities existing outside the unique manifestations of illness in particular men and women.”² Rosenberg adds, “Diagnosis labels, defines, and predicts and, in doing so, helps constitute and legitimate the reality that it discerns.”³ In elaborating on this “specificity revolution,” Rosenberg highlights, as have many other researchers writing about changes since the late 19th century, the way in which the medical world has with ever-increasing acceleration developed an array of technologies and “objective” tools designed to reliably replicate and standardize diagnostic procedures.⁴ These days it is common practice to have a clinical diagnosis confirmed by laboratory examinations of specimens procured from patients. If the laboratory findings do not support the clinical assessment, then, usually, the clinician’s diagnosis is trumped, and clinician and patient must adjust their understanding of what has gone amiss. But entities understood as irrefutable signs of “disease” also come to light for the first time with laboratory testing, usually because they cause no symptoms, and may never do so. Hepatitis C is one obvious example and HIV another. In both these instances the virus can be measurably present in the body, but the individual remains perfectly healthy. Multiple sclerosis is a further example, one that signals diagnostic problems similar to those documented in connection with the dementias. Neuroimaging has shown that the characteristic lesions associated with this condition—multiple sclerosis plaques—may be present in individuals who are entirely unaffected by the disease.⁵ Conditions such as these make the task of delineating disease specificity extraordinarily difficult,⁶ and it is abundantly evident that simply demonstrating the presence of so-called pathological entities or features in a body is not sufficient, and may actually be very misleading, as the previous chapter made clear. Questions of ontology and epistemology come to the fore rather dramatically in these instances: When is a “disease” not a disease? When is “pathology” normal? Under what conditions does a “natural” entity become “pathological”?

The first section of this chapter briefly reviews the literature that points out inconsistencies in both the clinical and neuropathological diagnosis of AD despite repeated efforts over the years to standardize them. A diagnosis of AD involves the demonstration at autopsy of neuritic amyloid plaques, neurofibrillary tangles, and also cell loss or shrinkage of brain tissue. For more than half a century, notably once the electron microscope was made available, efforts have been made to understand why and how plaques and tangles form in the brain, at times to excess, and what exactly they consist of. For two decades the preeminent model to account for plaque buildup has been the “amyloid cascade hypothesis” that, it is argued, initiates the eventual formation of tangles and other neuropathological changes. In the second part of this chapter I discuss this paradigm, on the foundation of which the majority of efforts to develop drugs to combat AD have been based. The model is currently being questioned by an increasing number of key researchers, but has by no means been overturned, and continues to be a driving force, even as the entire Alzheimer enterprise moves to include prevention as a major goal. The third section of this chapter returns to the question of dual diagnostics—symptomatic and neuropathology—and to current efforts to restandardize the criteria yet one more time in both Europe and North America. In the United States a new clinical diagnosis, made public in the spring of 2011 by the National Institute on Aging (NIA) and the Alzheimer’s Association of the United States, is distributed along a time dimension that extends backward through the life span of individuals and has three phases: a presymptomatic phase, an early symptomatic, predementia phase, and a full-blown dementia phase. This most recent round of standardization is explicitly designed to move the management of Alzheimer’s to a point in life well before the assumed amyloid cascade would have started to gather momentum.

Standardizing a Diagnosis of Alzheimer Disease

The dementias present a special challenge because no blood, urine, or other laboratory test of human secretions or fluids can, as yet, affirm or refute a clinical diagnosis in the living patient. It has long been agreed that confirmation of an AD diagnosis can be obtained only at autopsy. Hence, verification of accuracy is necessarily after the fact of death. Confirmed autopsy results are of use in creating medical records and for epidemiological, public health, and basic science research, but such neuropathologically confirmed diagnoses have little or no direct relevance for the clinic. However, families often wish to be informed about autopsy reports. Confounding matters, a “shortage” of neuropathologists exists in the United States, Canada, and the United Kingdom, and no doubt elsewhere; relatively few cases of dementia ultimately go to autopsy, and families may have to wait up to three years to obtain verification of the diagnosis of their deceased relative.⁷ Furthermore, as one American neuropathologist noted, autopsy costs exceed financial returns to the hospital. Establishing AD specificity in individual cases is, therefore, exceptionally challenging: because autopsies are not routinely carried out and, when they are conducted, the findings do not always provide evidence that is congruent with the initial clinical diagnosis. As we saw in the previous chapter, what exactly constitutes an “AD brain” remains in question, especially because research makes it increasingly evident the extent to which normal and pathological aging are entangled.

Despite these difficulties, since the 1970s, AD has steadily become an uncontestable reality, in effect, a social actor in its own right⁸—an entity in association with which a specific language, images, estimates of incidence and prevalence, and concerns about a looming epidemic are mobilized. The concept of Alzheimer’s is an artifact of value, one which has taken on a life of its own among medical practitioners, the public, and advocacy groups alike. And yet, every medical specialist with whom I have talked readily agrees that accurately estimating the number of Alzheimer cases leaves a great deal to be desired. Some believe that a dramatic underdiagnosis is probably the case, particularly when one keeps in mind the global situation,⁹ while others, focusing on findings about the heterogeneity associated with AD, believe that what passes for “pure AD” may well be overdiagnosed, and others note that the presence of rarer forms of dementia, such as frontotemporal dementia, can rather easily be overlooked.

Given that no decisively effective medication exists for AD, there is little need for primary care physicians to be overly concerned at present about the accuracy of their diagnoses. Moreover, certain clinicians have told me that it is well known that some doctors in Canada may diagnose a patient as having AD based on neuropsychological testing scores in which the figures are “fudged” to make them fall under the cutoff point below which individuals are eligible for cheaper medication as part of the socialized health care system. On occasion families may actually push doctors to confirm an AD diagnosis even when the physician continues to have doubts. Likely Canada is not alone in this practice. Evidently, as a result of these confounding factors, national and global estimates of AD cases cannot possibly provide an accurate estimate of the number of individuals at increased risk because statistically reliable population databases do not exist. Those few communities that have been consistently researched over the years using well-designed epidemiological methods are an exception. In these instances the resultant data are much more reliable,¹⁰ but strictly speaking such findings apply only to the studied communities, making extrapolations from them about the nationwide number of elderly people at risk for AD inappropriate, unless it is assumed that age and sex are the only significant variables implicated in AD incidence.

A brief purview of the massive literature on AD diagnoses over the past two and a half decades highlights two things. First, there have been repeated efforts over the years to refine the standardization of the diagnosis both for use in the clinic and for epidemiological purposes. And, second, in practice, the criteria for making a diagnosis, whether clinical or based on neuropathology, are not entirely uniform, despite the effort put into standardization. This situation makes very clear just how challenging is the pinning down of the AD phenomenon, with enormous consequences for estimations of AD prevalence and incidence.¹¹

At a research workshop sponsored in part by the NIA in 1984, the organizer, neurologist Dr. Zaven Khachaturian, made the following comment:

We are still uncertain whether AD is a specific, discrete, qualitative disorder such as an infectious process, endogenous or exogenous toxic disorder, or biochemical deficiency, or whether it is a quantitative disorder, in which an exaggeration and acceleration of the normal aging processes occur and dementia appears when neural reserves are exhausted and compensatory mechanisms fail.¹²

Khachaturian noted that incorrect diagnosis of AD is “quite common” and may be as high as 30% in the general population. The goal of workshop participants, then, was to better define the criteria for AD, which would, it was assumed, improve diagnostic validity. Perhaps not surprisingly, there was no discussion about the ontological status of AD—the material “facticity” of AD was assumed to be unassailable, if only the diagnosis could be improved.

Later that same year, what has since become the most commonly used criteria for the clinical diagnosis of Alzheimer’s were put out jointly by the National Institute of Neurological and Communicative Disorders and Stroke (NINCDS) and the Alzheimer’s Disease and Related Disorders Association (ADRDA; abbreviated shortly thereafter simply to the Alzheimer’s Association), both in the United States. These criteria, known as NINCDS-ADRDA criteria, were extensively revised between 2007 and 2011. But there are several other widely used criteria that have been available for many years, among the best known in the English-speaking world being the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders (DSM), which has also undergone repeated revisions; the World Health Organization’s International Classification of Diseases (ICD), which too has been revised; and the Cambridge Examination for Mental Disorders of the Elderly (CAMDEX). These criteria all require that evidence of cognitive impairment indicating dementia be established by neuropsychological testing. The most commonly used screening test for cognitive impairment is the 10-minute Mini-Mental State Examination (MMSE),¹³ supplemented by further more comprehensive tests of cognitive function requiring up to several hours to administer, when a clinician seeks to confirm the diagnosis. Repeat testing over the course of months or even years then follows, during which time possible comorbidities are ruled out and a steady, irreversible progression of cognitive decline is carefully documented, following which a clinical diagnosis of probable Alzheimer’s is made.

The usual assumption has been that these various diagnostic classifications are essentially interchangeable. But one project carried out in 1997 involved an examination of 1879 men and women aged 65 years or older who were originally among more than 10,000 individuals enrolled in the Canadian Study of Health and Aging, a project designed to ascertain the prevalence of all types of dementia across Canada in both community and institutional settings.¹⁴ The results showed that the criteria used for diagnosis differed by a factor of 10 in the number of subjects classified as having dementia; this depended on which among six of the standardized diagnostic criteria were used.

Additional problems have been brought to light as a result of systematic research; prominent among these are the very numerous cases of so-called mixed dementias, atypical dementias, and “cognitive impairment no dementia,” all of which conditions have been highlighted as confounding factors in diagnostic standardization.¹⁵ Current neuropathological findings indicate strongly that 30% or more of dementias may be a mixture of vascular dementia and Alzheimer disease, but in the clinic, up to a decade ago, it was the case that mixed dementias were not usually recognized. Interrater reliability has also been shown to be a problem—neuropsychologists and neurologists do not always agree on diagnoses.¹⁶

A clinical diagnosis of AD is frequently described as one of exclusion, that is, in clinical practice AD has been and often continues to be the diagnosis that remains after all other known forms of dementia and many other conditions, including hypothyroidism, stroke, vitamin B12 deficiency, folic acid deficiency, head trauma, Parkinson’s disease, Huntington disease, and several further conditions, are first ruled out.¹⁷ A diagnosis of AD has often been described to me as a “wastebasket” category by general practitioners. When I met with David Bennett, a Chicago-based neurologist, and inquired what he thought about this idiom, he replied,

What I would say is that historically, Alzheimer disease was truly a wastebasket. So if you go back a few decades, Alzheimer disease in the textbooks was a diagnosis of exclusion. You would take somebody with dementia, and you would say, they don’t have this, they don’t have that, they don’t have the other thing, and therefore they must have Alzheimer disease. But, significantly, the wastebasket didn’t have all the Alzheimer disease in it because—well, strokes don’t prevent you from getting Alzheimer disease. So if you throw out your strokes, okay, very many of those people have Alzheimer disease, but are missed. And if B12 deficiency and hyperthyroidism are thrown out, you often throw out AD as well.

So basically these researchers were left with their dementia cases after they’d thrown out all the comorbidities. And of course they didn’t throw out these conditions in their controls who weren’t demented. And then they compared these cases and controls, violating every epidemiologic principle. So this wastebasket was a mess. And so what’s happened over the years with the recognition of mixed dementias is the change from a diagnosis of exclusion to a diagnosis of inclusion.

When I asked exactly what he meant by this, Bennett explained,

Well, today when we train people, they learn that Alzheimer disease coexists with other common diseases. And these diseases probably actually contribute to dementia and make it more likely. We train people to ask themselves, does the person have the characteristic temporal and spatial pattern of a cognitive deficit? And so if you have somebody who moves from a progressive amnesia [memory loss], to profound amnesia, followed by problems of orientation and so on, I mean that’s a pattern that’s just so unmistakable, that 90 plus percent—there are some things that can mimic it, but there’s a pattern that 90 percent plus most of the time—these people have Alzheimer disease. We see this pattern just from cognitive testing. In our research here we keep ourselves blinded at each year of follow-up and, even so, 90 plus percent of them prove to have the AD pathology when they come to autopsy. So this approach to diagnosis is not a wastebasket.

But a problem remains, and the reason why I agree with you, is that along with the Alzheimer’s pathology comes all kinds of other things. So, for example, infarcts [local tissue death from stroke], Lewy bodies [which characterize a rare form of dementia known as Lewy body dementia], and other things, are contributing to whether someone tips over the edge. Now we’ve moved beyond the wastebasket and we’ve got an Alzheimer’s bucket and in that Alzheimer’s bucket is mostly Alzheimer’s—and we’re pretty good at that—but in that bucket are all kinds of other crap that also tags along. And the difficulty is that, even if you can separate these things out, the value of doing this from a research point of view is dubious at best, and probably introduces bias. So from a research point of view, you don’t want to be studying “pure” Alzheimer disease because you’re going to find some really weird things. But people still do this.

Some of these people are friends of mine and they take their brain tissue, for example, and they say, I’m going to throw out everyone who had infarcts and Lewy bodies and blah blah blah, and I’m just going to study my pure Alzheimer disease. And it’s kind of like, it’s not the way it works in the world. So they find weird things. They find sometimes the opposite of what we find when we keep everyone in, and it’s probably because of the bias that they’ve introduced by throwing stuff out. So you can find diabetes cases who have infarcts, but if you throw out all your infarcts and you just try to understand the relationship of diabetes to AD pathology, it’s not clear what you’re studying anymore. (April 2010)

Studies have shown that in Alzheimer disease research centers and in memory clinics—frequently, but not exclusively attached to teaching hospitals—AD diagnoses are confirmed at autopsy in over 90% of the cases, as David Bennett points out is the situation for his unit.¹⁸ But, clearly, for research purposes, attempting to isolate “pure” AD introduces difficulties because a “normal” aging brain is full of “crap” (a word used surprisingly often by AD scientists). But, so-called brain “crap” or “junk” is a reality of aging, particularly in old age, and should be dealt with systematically in research protocols rather than ignored. Cases of “pure” AD neuropathology are, in effect, researcher-created artifacts that can introduce serious difficulties into studies designed to investigate risk for Alzheimer’s, as Bennett’s comments make clear.

A further major difficulty is that by far the majority of patients with dementia are never seen at memory clinics or similar specialist research centers. One Montréal family practitioner commented,

Do I send people to memory clinics? No, I don’t initiate referrals to memory clinics. I have only done it on one or two occasions where it was an atypical presentation that I wasn’t clear about. I don’t initiate, but I do respond to families asking about it. And when families ask, what I reply is that there’s an eight month to a year waiting time; that the clinics are—for the most part—funded and designed to gather research data; that the treatment modalities that the patient will be offered would likely be no different from what I would do here, and that the choice of medication might be a little bit better informed at a memory clinic but that it probably wouldn’t make any difference. And then I say, “you make your choice.” I give them all that I think they need to know, and in my experience, the vast majority doesn’t go to the memory clinic. And my patients are mostly solidly middle class.

The ones who I’ve seen go and who hang on do it because they believe in research. They do it because it gives them something to do that makes them feel a bit better. They tend to have slightly higher IQs and therefore intellectually, they’re interested in what goes on in the memory clinic, and it gives them a sense of control that they may feel would be lacking if they were only followed by a family doctor. And often it’s the spouse who encourages this rather than the patient himself. I take a lot of time listening to the family as whole; talking about daily life, following up on them carefully over the years, and it seems most patients really appreciate this, even if it’s clear you can’t cure the problem. (May 2011)

This practitioner and other primary care practitioners whom I interviewed (12 in all) are concerned that once patients go to specialty clinics they essentially become research subjects, and that their quality of life and that of their families may not be attended to adequately. A good number of these practitioners have little faith in the medications currently on the market and administer them sparingly. Should a highly effective drug become available, this situation would no doubt change—under these circumstances these clinicians would ensure that a finely tuned differential diagnosis be carried out.

Another Montréal-based family practitioner, Robert Diez d’Aux, who recently returned to the city after working for nearly 25 years in a small town in Tennessee, said,

So far all we’ve really got to offer are platitudes, more or less. You know: get lots of exercise, good nutrition, do crosswords (which I do myself!), and watch your weight. My position is, when the family comes to see me, if they don’t identify a problem and they are managing well at home, then I don’t want to upset the apple cart. So, I certainly don’t rush into a diagnosis of AD, because, even if the patient doesn’t do too well on the animal fluency test, for example [naming as many four-legged animals as one can in one minute], or on the Mini-Mental, I don’t really know if they have some dementia until I’ve followed them for months. I very, very rarely sent families to the Memory Clinic when I was in Tennessee, because they would have had to drive for two and a half hours to Vanderbilt in Nashville. I thought it was better to support the family rather than trigger a consultation cascade. I can live with uncertainty, especially if the family isn’t pushing for a diagnosis of the “A” word. It makes little difference to patient well-being. (May 2011)

As part of his practice, Tom Dening, formerly a Cambridge University–based old-age psychiatrist, routinely makes house calls in the villages and communities that surround Cambridge. He agreed to take me along with him on two half days when he was visiting patients and their families in their homes. Dening states that the majority of people referred to his clinic for memory impairment are older than 80 years of age. “Some of them are quite a bit more than that,” he adds, “it just so happens that I’m going to see someone who is 95 this morning.” Dening is emphatic that when an individual is very old it is impossible to be precise when making a diagnosis because, inevitably, one is confronted with a mixture of clinical features that may or may not contribute to dementia: “So many factors are in play that my most common diagnosis for people age 80 and over is one of mixed dementia.” Moreover, he adds,

The dementia concept is quite rocky with old people because there’s a psycho-social component that must be taken into consideration—one has to ask, is this person “normal” for their age, or not? Even normal Olympic athlete types who are 95 years old are not going to score 30 on an MMSE. It’s important to recognize that there may be very, very few 95-year-olds who are not MCI [mild cognitive impairment]. What is a normal baseline at this age? We really have little idea, and the situation is that the condition is so heterogeneous it’s impossible to have a gold standard for [diagnostic] sensitivity and specificity—at least among the numerous older patients one sees these days. (September 2010)

A very busy general practitioner working in a relatively economically deprived part of Montréal emphasized that in his opinion among his patients aged 75 and older he is usually dealing with individuals who have what he describes as a “syndrome.” Even with a socialized medical system in which no fees for service are charged, David Dunn finds that his patients come to see him “very late in the game,” whatever their problem (unless it is an emergency), at which time it is virtually impossible to help them change lifelong habits that bring about poor health. Such patients may well have dementia, he stated, but in addition they have diabetes, high blood pressure, and other problems. Among this mixed patient population, including recent immigrants and some refugees from Eastern Europe, Russia, China, Senegal, the DRC, and many other locations, few of them, says Dunn, have good social supports, and a few are in hiding from authorities or members of their own “ethnic group.” When they come to see a doctor their purpose is very often to obtain medication and little else.

An alert practitioner might well suspect early signs of dementia among some of these older patients quite frequently, as does Dunn, but when he attempts to administer a modified version of the MMSE he finds that he is confronted with language barriers that even his trilingualism cannot surmount and, in addition, often suspicion and resentment on the part of some of his patients at being made to do a “mental test.” Dunn notes that a good number of his patients do not return for follow-up visits, making it impossible to diagnose anything other than “possible dementia”—a diagnosis that may well not be correct, especially if the patient is depressed. Under the circumstances, Dunn is most likely to write a diagnosis of diabetes or another common problem that he can verify on the clinical chart, and leave it at that.

Trained in epidemiology in addition to general medicine, David Dunn agreed with me without hesitation that official statistics on the incidence and prevalence of AD are without doubt highly inaccurate. But when he makes a differential diagnosis in his practice, his intention is not to discard everything he finds until he exposes “pure AD.” In primary care, there can be no wastebasket for leftover “real” cases, Dunn argues, because comprehensive treatment of patients and families is the primary objective, and clinicians note all the significant observations they have made in their examination, otherwise they would be failing in their care of patients. Dunn sees some positive signs of change with the formation of group practices made up physicians, nurses, and other health care practitioners, all of whom are specialists in geriatrics. Primary care practitioners can refer their patients to these practices specifically for assessment of dementia and other common disorders among the elderly. But, inevitably, not all patients are willing to go.

Neuropathology as the Gold Standard

As we have seen, clinical diagnoses of AD can be confirmed only at autopsy—neuropathological findings have been the gold standard for over a century. But here again, problems associated with accuracy and replication abound. Not only are autopsies infrequently carried out, inconsistencies have been shown in the findings, notably in mild to moderate cases, and especially for older subjects. One study in the early 1990s showed that quantitative estimates of the number of plaques and tangles in the same fixed slices of cortical tissue sent to 12 different neuropathological laboratories varied by a factor of 10.¹⁹ Inaccuracies have also been demonstrated where autopsies show that mixed dementias are involved.²⁰

The British neuropsychologist C. J. Gilleard summed up the matter in 2000 as follows:

The clinical diagnosis of Alzheimer’s assumes a distinct neuropathology to “confirm the diagnosis.” But it is apparent that the rather stylized “types of dementia” beloved by textbook writers are expressed more clearly in text than in tissue.²¹

Gilleard charges, “[T]here is confusion over what constitutes the common disease process termed Alzheimer’s,” and he goes on to argue that a degree of ambiguity and arbitrariness has been tolerated in the field.²² Specialists working in memory clinics are liable to insist that in recent years this kind of uncertainty has been eliminated, but epidemiologists in particular challenge this position, claiming that the complexity of the situation continues to be underestimated.

John Morris, one of the doyens of Alzheimer researchers in the United States, in a carefully crafted review article published in 1995, cautioned,

The fundamental relationship … of plaques and tangles to AD phenotype is only incompletely understood. The presence of these lesions in the neocortex remains the cardinal feature for both the qualitative and quantitative assessment of AD, but neither lesion is pathgnomonic [that is, a conclusive diagnosis cannot be made from such findings].²³

And John Breitner, an epidemiologist and psychiatrist, has pointed out how the DSM (both DSM-III and DSM-IV) and the ICD use criteria for establishing dementia “cases” that conflate clinical signs with what are assumed to be signs of irreversible brain pathology. In other words, in Breitner’s estimation, criteria are used that are “at once syndromic and neuropathological.” He continues, “It is an important goal in dementia research, as in all of medicine, that we characterize the underlying pathological or etiologic entity responsible for the clinical picture.”²⁴ Before this can be done, he argues, a reliable, replicable understanding of what exactly constitutes the dementia syndrome is needed. A 2010 draft of DSM-5 proposed that the term “dementia” of the Alzheimer’s type should be dropped entirely, in large part due to its stigmatizing qualities, and be replaced by the term “major neurocognitive disorder”—now formally incorporated into DSM-5, that does little to clarify matters.²⁵

The research cited in the previous chapter by Carol Brayne and colleagues suggests that in the case of AD-type dementia we are not dealing with one syndrome, but with at least two and possibly several types, each of which is closely associated with aging, an unavoidable process that Brayne insists is “the elephant in the room.” It is patently clear that aging completely dwarfs all other variables associated with risk for AD and dementia (aside from rare, early-onset AD).²⁶ Furthermore, as more and more research findings emerge, including a good number that demonstrate the remarkable plasticity of the human brain, the difficulty of characterizing even a “diffuse clinical syndrome” on the basis of a high degree of concordance between standardized clinical and neuropathological diagnoses may not, ultimately, be possible. Current research findings repeatedly destabilize the situation and challenge dogged assumptions anchored in a lingering belief about localization theory that goes unquestioned by many researchers. In recent years, for example, an abnormal form of a protein known as TDP43 associated with ALS (Lou Gehrig’s disease) and also with the relatively rare form of dementia known as frontotemporal dementia, from which patients with Lou Gehrig’s disease may also suffer, has been found in up to 20% of AD patients.²⁷ It has been suggested in this case that AD diagnoses may be “masking” frontotemporal dementia.

In summary, in specialty clinics, an assumption is often made that it is possible to diagnose “pure” AD in the clinic with a high degree of specificity and sensitivity. This diagnosis is based on observed behavioral changes associated with a decline of cognitive capacities in individuals who are monitored over months and years, and who continue to decline steadily until a rather dramatic downturn frequently takes place in the final months before death. As Zaven Khachaturian puts it,

The disease quietly loots the brain, nerve cell by nerve cell, a burglar returning to the same house each night. Typically the first symptom to appear is forgetfulness. Then comes more severe memory loss, followed by confusion, garbled speech and movements, hallucinations, personality changes and moods that can swing from anger to anxiety to depression. As the brain loses mass the rest of the body gradually shuts down.²⁸

No one who has cared for a relative afflicted by AD could take issue with this description, even if some might be uncomfortable with the use of dramatic metaphors in scientific papers. Khachaturian admits that the speed of this deterioration can differ greatly among patients, and that death can take place between 2 and 20 years after the first symptoms appear, suggesting that molecular complexity is at work that varies greatly from one patient to another. And Khachaturian also acknowledges that mixed dementia is very prevalent. Epidemiological and basic science findings both add support to an argument that a diagnosis of “pure” AD is an artifact of attempts at standardization. Recently Carol Brayne has reiterated the position she has taken for many years, namely that the dominant assumption about neuropathology as the cause of AD had resulted in “a reification of particular diagnostic entities without the necessary empirical evidence from relevant populations.”²⁹ Should such robust evidence ever be amassed, even then, it may ultimately prove to be impossible to consistently link clinical diagnoses of AD with standardized postmortem neuropathological findings due to the very complexity, plasticity, and adaptability evident in aging brains. Such an argument does not, of course, diminish the reality of dementia.

A well-known Cleveland neurologist, Peter Whitehouse, with 30 years of specialized clinical practice and neurological research under his belt, feeling increasingly ill at ease with the situation in the AD world, published a book in 2008 titled The Myth of Alzheimer’s: What You Aren’t Being Told about Today’s Most Dreaded Diagnosis.³⁰ When I talked with him, Whitehouse made it clear that he continues to diagnose his patients using what he describes as “an American Academy of Neurology standard approach.” He quickly notes that he does not use tests such as spinal taps (as is the case for many neurologists (see chapter 4), but he does use CT scans, and occasionally an MRI (magnetic resonance image). He adds,

It sounds a little strange, but most people come to me—about 70% I would say—having seen somebody else and certainly having read something about memory problems, so I use the first part of the examination to kind of get a sense of what they think about the issues. I usually spend the time with the patient and caregiver together, offering them an opportunity for separate time if they wish it. I spend a lot of time talking about language. I tell the patient and family that we still don’t understand very much about dementia. I believe that the most accurate statement you can make to an older person coming to a clinic with dementia is that they have a “mixed dementia.” I also use the word “aging” a lot, and “brain aging” and talk about multiple factors that affect brain aging. It’s trivial just to say that AD is heterogeneous—no disease is ever exactly the same in two bodies; what you really have to ask is what is the relationship of dementia to aging in the particular person in front of you.

When people ask if they might have Alzheimer’s, I reply that we don’t really know how Alzheimer’s can be separated out from aging. So I try to undermine the concept—to get them away from thinking that they are beginning the “funeral that never ends,” or that they’ve got a progressively degenerative disease. There are things one can do, as I write in my book, to make aging much more “successful” even when someone in the family has dementia. I think we need to focus much more on brain fitness.

Of course, the national Alzheimer’s Association criticized my book, particularly because of its title. In the public forum they try to scare people to raise money, and my book is a threat. I think this is particularly so in America, and the links with drug companies push things in this direction. The amount of money involved is a powerful incentive to keep advocating for an Alzheimer cure and not to look around for new ways of approaching this massive problem. (June 2010)

I have asked several of Peter Whitehouse’s colleagues what they think of his book. Predictably the answers have been varied, ranging from “Peter has lost his marbles” to “I haven’t time to look at that kind of thing.” But the most common answer has been that a new vision is needed in the dementia world, and that a good shake-up by Peter Whitehouse is not at all a bad thing—“[H]e’s pushing the boundaries, challenging people and making them think,” replied one eminent neurologist.

When I asked William Thies, until recently the medical and scientific officer of the Alzheimer’s Association, what he thought of The Myth of Alzheimer’s, he responded,

Let me say at once that I know what Peter has written and I disagree with it. It’s perfectly clear to me that there are people who have Alzheimer disease and there are people who don’t. And there’s a pathological process involved. Now, is it a single monolithic disease? We’d be ignoring some of the most famous lessons of history if we thought it was such, and I think there’s interesting work being done now that suggests that many of the people that show dementia actually have mixed dementia and what you probably have is a continuum. Just to put two points on the continuum (there will be more) but just to put two, it would be amyloid pathology and vascular disease. And you are going to have some people who are at one end of the continuum and others at the other end, but most will be in the middle. But we don’t really know what the distribution curve looks like yet. So I don’t have any problem with the idea that if you look across the whole population of people who are currently labeled as Alzheimer disease, that there will be some sub groups in there and, if I had to make a guess, I’d say that therapy for AD will end up looking like therapy for hypertension … it will be tailored to the individual.

In response to my question about AD being a wastebasket category, Thies said,

It’s no more a diagnosis of exclusion than many other transient diagnoses are. The first day one comes up with a diagnosis for anything it may prove to be wrong, but I think we’re getting better and better at identifying AD and the pathological studies really do bear that out, as long as good clinicians are involved, … further down the road I wouldn’t be at all surprised if we found it best to create subcategories. Ultimately I think therapies will be the best tools for allowing us to make distinctions into subsets of Alzheimer disease. In the end, one only gets confirmation from therapeutic outcomes. (April 2010)

It is understandable that William Thies, as scientific officer of the Alzheimer’s Association, whose working life was closely associated with raising money to defeat Alzheimer’s disease, would be very cautious about entertaining the idea that AD might not be “real.” But his willingness to accept the likelihood of emerging subcategories on the basis of drug development is not at odds with his position in the Association. Even if the Alzheimer category fragments, the overarching concept will continue to be used to promote fund-raising, political activities, and policy making.

To my inquiry into what kind of research the Alzheimer’s Association generally funds, Thies replied, “a fair amount of amyloid science—well, that’s kind of the state of the science, you know,” but he stressed that efforts are made to fund across the spectrum of research applications, and to give young researchers with innovative ideas a chance to get established. Thies emphasized that rigorous Alzheimer research has a very short history of only 30 or 35 years to date, and he insists that it remains “an open question” as to whether “intervention into amyloid” is likely to be effective. He then added a further proviso: “and another thing that complicates matters is that our cognitive tools are so dull that we can’t really tell how cognitively intact people are,” adding greatly to diagnostic problems.

Summarizing, despite inherent problems with standardization of both clinical and neuropathological AD diagnoses, and explicit doubts expressed by certain researchers about the very concept of AD, the basic science paradigm that has held sway for over 20 years in the Alzheimer’s research world continues to be dominant—it is a model in which amyloid deposition in the form of plaques is for all intents and purposes regarded as the master key.

The Amyloid Mafia and the Prevailing Paradigm of Alzheimer Disease

Arguably the most influential article associated with Alzheimer research appeared in 1992 in Science, one author of which was the well-known British neurogeneticist John Hardy. Hardy is today the most cited Alzheimer researcher in the United Kingdom and has on occasion appeared in a Hawaiian shirt and other unconventional dress to deliver plenary addresses to audiences sometimes of well over 1,000 people. The article, coauthored with Gerald Higgins, a neurobiologist formerly associated with the NIA, is subtitled “The Amyloid Cascade Hypothesis,”³¹ and the theory it postulated quickly became the dominant paradigm for understanding Alzheimer disease causation at the molecular level. The authors emphasized at the outset, “Alzheimer’s is characterized by various pathological markers in the brain—large numbers of plaques surrounded by neurons containing neurofibrillary tangles, vascular damage from extensive plaque deposition, and neuronal cell loss.”³² The second paragraph then set out the cascade hypothesis: “deposition of amyloid β protein (AβP), the main component of the plaques, is the causative agent of Alzheimer’s pathology and … the neurofibrillary tangles, cell loss, vascular damage and dementia follow as a direct result of this deposition.”³³

The cascade hypothesis states that what is known as the amyloid precursor protein (AβPP), when “cleaved” by a specific enzyme, y-secretase, itself composed of four proteins, can result in an excess amount of toxic Aβ protein, resulting in a cascade of negative effects in the brain. The authors generated this hypothesis on the basis of what were at the time newly discovered findings in connection with the genetics of early-onset Alzheimer disease in which mutations of the APP gene are implicated (see chapter 5). They also noted that people with Down syndrome, once adults, are highly susceptible to Alzheimer disease because the region of chromosome 21 that contains the APP gene is involved in that condition. Hardy and Higgins acknowledged that several upstream pathways no doubt exist that can set off processes contributing to AD, but their position when the hypothesis was first formulated was that the final common AD pathway is precipitated by the toxic effects of a buildup of Aβ protein, and it is this buildup that causes the disease.

Since the publication of this article, a vast amount of research has been directed toward solving the question of what exactly leads to excess precipitation of the Aβ protein, and by far the majority of drug trials in connection with Alzheimer’s, costing billions of dollars, have been designed to eliminate excess plaques in the brains of experimental mice and humans—with no real success to date.³⁴

In 2000, Mark Smith, a Cleveland-based neuropathologist, and three of his colleagues published a letter in the prestigious medical journal The Lancet, titled “Amyloid-β Junkies.” The letter started out, “Evidence indicates that, in the very near future, by either snorting or injecting antibodies to amyloid-β, it will be possible to remove amyloid-β-laden senile plaques from the brains of individuals with Alzheimer disease.” These letter writers were referring to a 1999 article claiming that by means of amyloid-beta immunization Alzheimer-disease-like pathology had been attenuated in mice. Smith and colleagues ran through all the reasons why they were certain that such an immunization might do more harm than good in humans, including the fact that animals with massive doses of amyloid-β more extensive than that found in human AD do not replicate the full range of AD pathology, nor the dementia associated with Alzheimer disease. They concluded that attempting to remove amyloid-β is playing a dangerous game; in their opinion, amyloid may well play a role in “defense of the aged brain.”³⁵

In contrast to Smith’s position, in 2002, John Hardy together with Dennis Selkoe, a Harvard-based neuronal cell biologist, robustly defended the amyloid cascade hypothesis in an article cited to date well over 5,000 times. They argued that basic science findings in the 1980s that had provided a genetic framework that enabled the postulation of the hypothesis “marked the beginning of the modern era of research” in connection with AD.³⁶ However, it was also noted in this article that research had shown that the number of amyloid plaques in the brain does not correlate well with “the degree of cognitive impairment.” It was also pointed out that cell culture systems on which most of the research has been carried out may well not adequately reflect “the complexity of Aβ economy in the human brain.” Two summary points were made by these authors: first, although gaps in knowledge are evident, none of the currently perceived weaknesses of the amyloid hypothesis provide good enough reasons to abandon it; and, second, drug development based on the cascade hypothesis to date has not slowed cognitive decline in patients. Even so, it was argued, given that no other alternative hypothesis had been proposed, the amyloid cascade hypothesis should not be abandoned. Numerous research projects are designed to unravel some of the persistent problems associated with amyloid analyses, notably by further molecularizing Aβ into specific subtypes and researching their independent effects on cell death in the brain.³⁷

Scientists whose focus of investigation is tangle formation have long felt themselves overshadowed by the so-called amyloid mafia, but on the whole such in-house debates have remained reasonably friendly. However, as Mudher and Lovestone, neuroscientists at the Institute of Psychiatry in London, have pointed out, these debates “assumed religious overtones” when those whose investigations concentrate on amyloid activity were dubbed “baptists” (β-amyloid protein) and those whose efforts are devoted to an explication of the tangles became known as “tauists,” because excess production of a hyperphosphorylated form of the protein tau is associated with tangle buildup.³⁸ Certain tauists have argued that researchers devoted to the amyloid cascade hypothesis have been inclined to think of neurofibrillary tangles as an “epiphenomenon,” and take issue with this position.³⁹ They point out that it is clear that tau-related pathology alone can cause neurodegeneration in the brain; that plaque formation by itself does not induce tau pathology; and that in transgenic mouse models the cascade hypothesis does not apply because plaques alone are produced in mice, and the brains of these animals exhibit no neuronal loss, no tau pathology, and no tangles.⁴⁰ It is also of note that it is human APP that is inserted into the mice, and that mouse APP is not toxic to mice.

An article in Science titled “Tauists and Baptists United—Well Almost!” reminds readers that β-amyloid plaques have never been proven to cause neurodegeneration—rather, an association has simply been established.⁴¹ However, research findings may have “cooled down” the argument because two research groups working with new models of transgenic mice have shown what had not previously be demonstrated—that plaque and tangle pathways are themselves “entangled.” This suggests that too much emphasis may indeed have been given to amyloid alone.⁴²

Criticisms about the amyloid mafia have been in print since the early part of this century, as The Lancet letter makes clear, but have gone largely unheeded until very recently. A set of brief articles published in 2006 in Nature Medicine are the product of a call by the journal editors to 34 leading scientists to set out what they believe are the most important research findings in AD research since 2003. The editors summarize, “Notably, of the papers our experts chose as the most important, 95% are related to the processing or pathogenesis of amyloid-β.” They continue, “Does this mean that amyloid-β is accepted by all as the primary causative factor in Alzheimer disease pathogenesis …? Perhaps, but there are some indications from experts that the time may be ripe for the influx of new ideas into Alzheimer disease research.”⁴³

The lead article in this collection, written by the science writer Apoorva Mandavilli, is titled, “The Amyloid Code.” This article initially airs the idea of an amyloid conspiracy, but quickly drops this suggestion to embrace a slightly more jocular tone by embarking on a discussion about the “Church of the Holy Amyloid,” an expression coined by Mark Smith. Mandavilli points out that for well over a decade every aspect of Alzheimer research has been dominated by the amyloid hypothesis, ranging from the search for a cure to animal modeling, most funded grants, and the most-cited articles.⁴⁴

Mandavilli notes that many of the scientists to whom she talked who are associated with the nature/genetics collection are less passionate in their criticisms of amyloid than is Smith, but even so they repeatedly made comments about the dangers of “putting all of your eggs in one basket.” The interviewed scientists warn that with such a complex disease “it is foolhardy, perhaps even dangerous, to focus exclusively on one theory.” One eminent researcher concludes, somewhat paradoxically, that evidence for the amyloid hypothesis is overwhelming but, even so, “[t]he whole field is governed by an old boy’s network that is not very positive. It needs new blood, new movement, new ideas.”⁴⁵ The work of Mark Smith and his colleagues highlights new directions that, in fact, are not so new,⁴⁶ but are currently being given more coverage than previously. Smith and colleagues insist that investigators should move upstream and consider more generalized bodily changes such as oxidative stress, inflammation, mitochondrial changes that are part of aging, and cell cycle dysfunction—all factors that may set off a chain of events leading eventually to amyloid and tau deposition.

Indirectly in support of Mark Smith’s position, a Finnish researcher, C.P.J. Maury, published an article in 2009 in the Journal of Internal Medicine in which he argued that although amyloid has usually been thought of as a “pathological structure,” growing evidence indicates that amyloid may well contribute to “normal physiology.” Maury points out that β-amyloid is “highly conserved,” meaning that it evolved long ago, and is present in numerous organisms ranging from bacteria to mammals and spans all aspects of cellular life. For proteins such as β-amyloid to be biologically active they must be appropriately folded into a three-dimensional structure, and in vivo are subject to strict “quality control” for this purpose. Maury goes on to describe in detail the diverse and essential positive functions of amyloid deposition in organisms of all kinds, including the preservation of homeostasis in mammalian cells. While acknowledging that amyloid may well be a sign of disease, his argument, like that of Smith and others, is that without doubt it also has indispensible positive functions.⁴⁷

A 2010 article by Smith and coauthors, published in the Journal of Alzheimer’s Disease and cited as the best article of that year, comments, “[T]he continued targeting of end stage lesions in the face of repeated failure, or worse, is a losing proposition.”⁴⁸ When I talked with Mark Smith shortly before his death at age 45 as the result of a shocking hit-and-run car accident, he suggested that amyloid should be understood as “scar tissue” and that it could be best thought of positively—as a sign of healing. Plaques are not the “cause” of AD, Smith insisted; rather, they are a sign of the body fending off troubles that commence further upstream.

Smith understood AD as being on a continuum with normal aging, and under certain circumstances plaques and tangles become irrefutable evidence that a threshold has been crossed into pathology. However, he insisted, given that plaques and tangles are not limited to AD alone, they simply cannot be regarded as the defining features of AD. Smith was in no doubt that the neuropathology he observed routinely at autopsies is evidence of age-related disease, but, he emphasized, as do other neuroscientists, that the number of neurons present and signs of cell shrinkage (figure 2.1) are of more significance than are the mere presence of plaques and tangles. His estimation as an experienced neuropathologist was that “educated people” can lose up to 50% of their brain tissue and continue to function reasonably adequately. Almost as an afterthought, Smith added, “[I]n older people a ‘normal’ brain looks very like an AD brain at autopsy,” and in common with perhaps the majority of researchers today, Smith noted that “mixed dementias” are most frequently found at autopsies of older individuals.

Figure 2.1.

Cell loss in a brain diagnosed with Alzheimer’s disease. Reproduced with the permission of the Alzheimer’s Association, from Inside the Brain: An Interactive Tour, Alzheimer’s Association, http://www.alz.org/alzheimers_disease_4719.asp. This tour is available in 14 different languages.

Rudolph Tanzi, a neurogeneticist, holds an endowed chair at Harvard and is director of the Genetics and Aging Unit associated with the Massachusetts General Hospital. Tanzi is billed on the Internet as a “rock star of science,” and is one of the best known and most outspoken Alzheimer specialists of the day. His position on the role of amyloid is not unlike that of Mark Smith, and when I met him in early 2010 in his research unit he had this to say:

[T]o this day, plaques don’t confirm dementia. What is replicated over and over again is that if you grind up part of the brain at autopsy and measure the amyloid-β load—that always correlates with dementia. But plaques that a neuropathologist can sit there and count using a microscope and extrapolate from—that doesn’t correlate. Why? Because the plaque’s not the problem!

I think of plaques as the end-game. I call them brain pearls because, just like an oyster, when you put a piece of sand in, it makes a pearl to entrap the sand to stop the pathogens. An oyster makes a pearl not because the sand is an irritant. The sand is covered with microbes and pathogens and the oyster protects its gut from them by producing a pearl to entrap that piece of sand and then you get a vicious cycle and it keeps going. You should leave plaques alone because they are sequestering the amyloid-β away into a neat little ball that cannot get into the synapse. If Aβ oligomers get in the synapse the synapse dies, the dendrites start to prune back and the nerve cell body dies. Plaque stops that happening as best it can.

It doesn’t bother me when somebody shows me a case with a lot of plaques. I am interested though, genetically, in understanding how someone with a lot of plaques comes through cognitively intact. What protects them?

Tanzi’s thoughtful, occasionally provocative comments on the current situation in connection with AD genetics will be discussed further in later chapters.

Anti-amyloid Therapeutics

The historian Jesse Ballenger has recounted the way in which efforts to develop antidementia drugs have played a crucial role over the years in the transformation of old age. His has researched the American situation, but his comments apply equally to the majority of European countries, Australia, New Zealand, Japan, several countries in Central and South America, and increasingly elsewhere. The first drugs designed to treat dementia began to appear in the mid-1950s in the United States and the United Kingdom, in an era when psychiatrists and social gerontologists were the specialists primarily concerned with dementia. These drugs were designed with the hope that patients would be able to continue an active social life and avoid further mental deterioration. As Ballenger notes, “By the early 1980s, the social and cultural transformation of aging had worked seismic changes in the landscape of dementia.”⁴⁹ The very term “senility” was under attack, and Robert Butler coined the phrase “ageism” in making a powerful argument that broad social, political, and medical changes were needed to deal with the aging society. It was at this time that AD was rediscovered as a discrete entity and real investment began to be made in finding a cure for this condition that was increasingly termed in the professional literature as SDAT (senile dementia of the Alzheimer type), making it clearly discontinuous with normal aging.

Ballenger emphasizes that the labeling of AD as a discrete entity was politically effective and allowed the medical world and advocacy groups to stress that a hunt for a cure should be launched forthwith. Even if a cure was not found, a delay in onset of the disease by five years or so would result in enormous economic savings (and incidentally reduce considerable suffering).⁵⁰ The sociologist Tiago Moreira argues that efforts to standardize an AD diagnosis were intimately related to drug development, and that drug development in turn, from the 1980s on, has been managed within a regulatory framework in which free enterprise, scientific innovation, and the interests of insurance companies have all played a part.⁵¹

By the mid-1970s it had been shown by several research groups that the brains of patients diagnosed with Alzheimer’s exhibited a pronounced deficit in a neurotransmitter known as acetylcholine, a deficit associated with memory loss.⁵² In the early 1980s Peter Whitehouse and colleagues at Case Western Reserve University discovered a “cholinergic center” in the cerebral cortex in which a marked decrease in acetylcholine was associated with cell loss and was presumed responsible for much of the symptomatology of AD including amyloid plaques. This finding became formulated as the “cholinergic hypothesis” of AD symptomatology—a model for the proximate cause of AD.⁵³ Since that time, the majority of drugs that have shown sufficient efficacy to be marketed to combat AD have been known as cholinesterase inhibitors. These drugs, although they improve some clinical symptoms of AD, have mixed results at best, are effective for only a short time in some patients, do not halt the progression of the condition, and are often accompanied by unwanted side reactions.

The story of Alzheimer drug development is complex (for a comprehensive discussion, see works by Ballenger, Moriera, and Whitehouse, among others),⁵⁴ but one thing is strikingly evident: despite widespread recognition for many years by a good number of eminent researchers that AD is heterogeneous, the amyloid cascade hypothesis has dominated the search for a drug to “cure” this condition. This search has been spurred on by a remarkable sustenance of hope with repeated predictions that Alzheimer’s would be defeated within five years. Only in the past several years has this language been toned down as numerous drug trials costing “untold millions” have consistently resulted in failure.⁵⁵

In the summer of 2009 Forbes reporter Robert Langreth published an interview with Mark Smith, in which he labeled him as a “renegade” because Smith had argued that anti-amyloid therapies could actually harm patients by precipitating them into a downward spiral. But, by 2010, shortly before Mark’s death, the tone of reporting had changed. Numerous drug companies, including Pfizer, Elan, Bristol-Myers Squibb, and Eli Lilly, had all been working on drugs designed to counter the buildup of β-amyloid plaque. In 2010 two clinical trials were stopped in the same week. One carried out by Eli Lilly and Co. was hastily withdrawn when it was shown that patients taking the drug were having more cognitive problems than were members of the control group taking a placebo. The drug also appeared to increase the risk of skin cancer. John Hardy was asked to comment online about the Eli Lilly failure. He stated that there are two key questions that have to be asked: First, did the drug hit the target but still not work? If this was so, then this suggests that the entire class of drugs is simply not appropriate. Second, Hardy questioned whether the scientists had neuroimaging data to show that the amyloid had actually been cleared in the trial subjects. If the amyloid was cleared and the patients became worse, this would be a very bad sign and clearly would put the amyloid cascade hypothesis into question.⁵⁶ Chapter 5 includes a discussion of several trials about to get under way that are in part designed to answer the latter of Hardy’s queries.

The second failed trial was designed to create a vaccine that would teach the immune system to recognize amyloid as though it is a foreign body. The researchers saw no measurable clinical gains in the treated group; the immunized patients did not live longer, nor did they take longer to reach a stage of severe dementia than did members of the control group. Two patients, even though their amyloid plaques were virtually cleared from their brains, died of profound end-stage dementia. These findings are based on a small number of patients, but nevertheless the researchers concluded that “the presence of plaques does not seem to be a prerequisite for progressive cognitive impairment in AD.”⁵⁷ Other vaccine trials caused encephalitis, resulting in death in a few subjects. Despite these findings further such trials are in progress.

The FDA has recently set up a Coalition Against Major Diseases in the United States. The chief executive of this endeavor stated with respect to Alzheimer’s and Parkinson’s diseases, “We really believe drugs are failing because we honestly don’t understand the disease.” The hope is to remedy this situation by pooling findings obtained from drug companies and academic institutions, to form a massive computerized database.

The cholinergic hypothesis had a second, quite different impact on the approach to AD, in addition to emphasis on amyloid removal from the brain. Due to the commercial availability over the past fifteen years of medications available for symptomatic therapy, a more “precise” clinical diagnosis is necessary, and physicians have been taught to base the diagnosis largely on the presence and extent of measurable memory loss that is far more easy to document than are changes in personality, social behavior, and emotional tone. Hence, demonstrated memory loss has in effect become a synecdoche for AD itself, thus reducing the concept to one associated largely with memory function alone.⁵⁸ Furthermore, clinical testing for AD has become heavily dependent on just one aspect of memory—that of recall—rather than giving equal attention to the other well-recognized forms of memory function.⁵⁹

The Move to Prevention: Finding a New Lexicon

In 2005, 15 international dementia experts were brought together to attend a workshop in Florence designed to discuss the need for revised diagnostic criteria for AD. The resultant 19-author article appeared in 2007 in The Lancet Neurology. This article was the first definitive sign of a sea change in connection with Alzheimer diagnostics. The article, the senior author of which is Bruno Dubois of the Hôpital de la Salpêtrière in Paris, sets out by stressing that in the clinic no definitive “diagnostic biomarker” exists for AD, and readers are reminded that a reliable diagnosis can only be made by means of histopathological confirmation. The article claims that in recent years the clinical diagnosis for AD has improved, is “no longer described in exclusionary terms,” and can now be taken further and “characterized more definitively on a phenotypic basis.”⁶⁰ Dubois and colleagues argue that potential patients can be assessed for “distinct markers,” including structural brain changes assumed to be definitive signs of prodromal, presymptomatic AD. They insist that prodromal AD “must be distinguished within the broad and heterogenous state of cognitive functioning that falls outside normal aging.”⁶¹

The proposal in this article is to retain core diagnostic criteria for probable AD as they are currently set out, in which patients or family report a gradual and progressive change in memory function for the worse for at least six months. This complaint must be confirmed by psychological testing. In addition, at least one “support feature” is called for, including positive results from one of the following: an MRI scan, cerebrospinal fluid analysis, positron emission tomography (PET) neuroimaging (see chapter 4), and/or demonstration of an autosomal dominant gene mutation for AD. These four “support features” are described as “biological footprints of the disease.” It is emphasized that because the “clinical phenotype of AD is better known than its biological phenotype,” for the time being at least, the clinical diagnosis should take priority. Validation studies of the biomarkers will be needed and can be applied retrospectively to large cohorts of diagnosed cases that are already part of ongoing research. Validation should also be carried out prospectively on people with and without dementia, who will then be followed to postmortem. The scope of this proposed research enterprise takes away one’s breath, and the need for extensive financial resources and technical skills is acknowledged.⁶² The authors argue that these criteria “represent a cultural shift requiring a more biologically focused work up,” one that is justified because AD diagnosed in the clinic has already progressed so far along an irreversible pathway, no pharmacological intervention will ever be effective.

Many adherents of the amyloid cascade hypothesis think of the move to an earlier diagnosis as positive. An article in Nature in 2008 insists, “Researchers have very strong reasons for believing that plaques and tangles are the cause of Alzheimer’s symptoms, rather than just the markers of it,” and the author cites well-known research findings to back up this assertion.⁶³ Simon Lovestone, even though a committed tauist, is quoted as saying, “It could well be that there has been too much focus on early events of the amyloid cascade. … But let’s not throw out the baby with the bathwater—there is just too much convincing preclinical research that supports the amyloid-cascade hypothesis.”⁶⁴ More than 30 trials were under way in 2008 for treatments based on the amyloid cascade hypothesis, and the assumption is that if “surrogate” biomarkers (“biological footprints”) are tested for, rather than waiting until the disease is diagnosable in the clinic, clinical trials will be of shorter duration and cheaper.⁶⁵ The conclusion of the Nature article is that the biggest tests of the amyloid hypothesis are yet to come.

Perhaps this is indeed the case if the 2011 remarks John Hardy made at a symposium sponsored by the pharmaceutical industry are anything to go by. The symposium was titled “The Amyloid Cascade Hypothesis Has Misled the Pharmaceutical Industry,” a title deliberately selected by Hardy to generate debate among several AD experts. As part of the introductory talks at this gathering, Hardy said, “I am convinced that, for early onset [AD] of the familial type [referring to the rare single gene form of AD—see chapter 5], the [amyloid cascade] hypothesis is proven. The situation remains less clear for … late-onset disease [the “late-onset” condition commonly diagnosed as AD].” During the debate Simon Lovestone asked John Hardy, “Why don’t mice get AD?” And Hardy replied, “I don’t know. The system that worries me most with respect to … AD is the vascular system. … The vascular system of a mouse [is] very different to that of humans and therefore this may have a strong effect on why mice don’t get AD.”⁶⁶

In the following two chapters I discuss at greater length the move to transform the clinical diagnosis of Alzheimer’s into one to be made at the “prodromal” stage, with the ultimate objective of directing a substantial amount research toward the goal of prevention. A new round of standardization involving the condition known as MCI (mild cognitive impairment) and the so-called biological footprints of prodromal AD, including those found in cerebrospinal fluid and by means of PET scanning, is under way. For the foreseeable future these investigations will be limited to research settings alone. This new approach is attractive to many because it enables detection of amyloid changes in cerebrospinal fluid, and the assumption is that if amyloid changes are spotted very early, then its deposition may well be reversible, or at least its accumulation may be slowed down. In other words, the amyloid cascade hypothesis continues to be the center of attention. But, as the article in Nature suggests, and as a large number of the researchers whom I interviewed indicated, this approach as far as they are concerned may well be mired in outmoded thinking.

The move to identification of biomarkers is likely to compound the difficulty of disassociating normal and pathological aging. As we will see in the next chapter, biomarkers do not reliably identify AD, but merely indicate probabilities for becoming demented. If one assumes that aging and AD are on a continuum and acknowledges that AD cannot be characterized by one or more features unique to it, then the paradox of entanglement cannot be resolved by means of biomarker testing. Given this situation, we may well finally be forced to confront the question repeatedly set to one side throughout the history of Alzheimer disease: Why are certain individuals “protected” from the ravages that amyloid deposition and/or other intimately related molecular changes apparently precipitate in many others?