Chapter 4

Embodied Risk Made Visible

The relation between what we see and what we know is never settled.

—John Berger, Ways of Seeing1

At the 2010 Honolulu meeting of ICAD (International Conference on Alzheimer’s Disease), a news briefing was held on July 13 in which information that had previously been embargoed from release was first made public. The briefing stated that by 2009 a broad consensus had become apparent among experts working on dementia that the criteria for an AD diagnosis needed to be revised on the basis of scientific advances made in the field over the past quarter century. With this in mind, the National Institute on Aging (NIA) and the Alzheimer’s Association set up advisory meetings composed of individuals from both academia and industry, and it was decided that three working groups should be formed. The Honolulu briefing reported on the progress these groups had made over the preceding year.

The tasks that the groups had been assigned were to examine and make recommendations on three specific conditions: “preclinical Alzheimer disease” (to be detected through biomarker testing); “mild cognitive impairment (MCI) due to Alzheimer disease,” and “Alzheimer disease.” Two of the groups were composed of 11 members and the third had 12. It was striking that virtually every involved expert was associated with an American university and or drug company. Of the two exceptions, one was the neurologist Bruno Dubois of the Hôpital de la Salpêtriére in Paris, the driving force behind the European working group convened in 2005 to redefine what would count as Alzheimer disease. Despite the dominant American presence in these newer groups, it was explicitly asserted that a balance of expertise existed in and among the three groups, and that they had international representation. Following the Honolulu meeting, the recommendations for changes to be made to the AD diagnosis by these groups were posted on the Alzheimer’s Association website and other sites, allowing for a period of public commentary. The plan was that the comments were to be assessed and incorporated, if appropriate, into the revised documents before publication in the journal Alzheimer’s & Dementia in the spring of 2011.

Both critical and supportive commentary, scientific and otherwise, appeared online. One oft-repeated statement, to the effect that the groups were clearly not international in representation, resulted in a small modification to their final constitution. Immediately before the final round of discussion and revisions took place in preparation for drafting the prepublication documents, two experts from the United Kingdom and one each from Canada, Japan, Australia, and the Netherlands were distributed among the three groups. These researchers, together with Bruno Dubois, 7 in all, formed the international component among the more than 30 Americans. It is significant that particular emphasis was given toward the end of the prepublication discussion in the fall of 2010 to the necessity of striving for a “harmonization” of biomarkers.

Public Critique of the New Proposals

An op-ed appeared in the New York Times on July 19, 2010, written by Sanjay Pimplikar, professor in the department of neurosciences at a Cleveland research institute. Pimplikar had been present when the news briefing took place in Honolulu, and he had also read Gina Kolata’s article of July 14 in the New York Times in which she reported on the release, citing researchers who claimed that these new guidelines would mean earlier diagnoses and that they represented a “major advance” in the Alzheimer field. Pimplikar’s op-ed, titled “Alzheimer’s Isn’t Up to the Tests,” takes strong issue with the notion that the new proposals represent a clear advance in the AD field; above all he expresses concern about the move to systematic use of biomarker detection in order to assist with early diagnoses. He writes, “Although these recommendations are well intentioned, evidence suggests that it would be a mistake to adopt them at this time.”2 Pimplikar points out with respect to the proposed use of PET scans to detect amyloid plaque in vivo, that patients would be systematically exposed to radiation. Furthermore, he reminds readers, plaque has repeatedly been found to be present in the brains of about one-third of normal elderly individuals. He adds for good measure that 11 drug trials had failed to show cognitive improvement in individuals after plaque had been removed from their brains.

Pimplikar goes on to consider routine use of spinal fluid analyses, a second biomarker recommended in the new guidelines. He notes that spinal taps apparently show promise, but the findings nevertheless remain uncertain and the cost is considerable. Moreover, the procedure can be painful, is liable to have side effects, and is not easy for practitioners to perform unless they are very skilled. One of the greatest difficulties in connection with biomarker detection emphasized by Pimplikar is the possibility for misdiagnosis and false positives. He makes an analogy to the routine use of the PSA test used to detect prostate cancer, in which it is now widely agreed that overdiagnosis and overtreatment have been common.3 In conclusion, the op-ed reminds readers that no cure exists for Alzheimer’s, with the result that early detection tests may well do more harm than good to public well-being.

Prior to and immediately following the appearance of this op-ed and associated articles, the following letters appeared in the Times:

To the Editor:

Re “Drug Trials Test Bold Plan to Slow Alzheimer’s”

As many researchers have pointed out in the last 15 years, it is unlikely that amyloid is the causative agent of neurodegeneration, the real cause of dementia. Therefore, it is unlikely that removal of amyloid would stop the progression or reverse the disorder.

Unfortunately the amyloid hypothesis has acted too long as a red herring, robbing the Alzheimer disease field of precious financial resources needed to develop new avenues of investigation.

The persistent focus on amyloid may be best explained by both the lack of alternative scientific theories and the reluctance of the pharmaceutical industry to abandon years of sunk costs in amyloid-based research and development.

Nikolaos K. Robakis

New York, July 17, 2010

The writer is a professor of neuroscience and Alzheimer disease research at Mount Sinai School of Medicine.

To the Editor:

Sanjay W. Pimplikar is absolutely correct. Even assuming that the new diagnostic test for Alzheimer’s is 100 percent accurate, what good does that knowledge do? There is no drug that cures the disease, only ones that mitigate the symptoms at an early stage.

My wife died four years ago from Alzheimer’s at age 69. She and her family suffered with the disease for seven years after the initial diagnosis. Had we known earlier, everyone would have suffered even longer.

While early detection may have some merit, the real breakthrough will come when we know and understand the cause of this disease.

William Eisen

Philadelphia, July 20, 2010

To the Editor:

Over the last two months, your outstanding series of articles and an Op-Ed essay have emphasized the mysterious nature of Alzheimer disease. A particularly lively argument focuses on the role of amyloid, sticky clumps of material that build up between brain cells.

There are amyloid believers (like Dennis J. Selkoe and me, both of us quoted in your July 17 front-page article “Drug Trials Test Bold Plan to Slow Alzheimer’s”) and amyloid agnostics or naysayers (like Sanjay W. Pimplikar, whose Op-Ed article, “Alzheimer’s Isn’t Up to the Tests,” appeared on July 20).

… The amyloid hypothesis may ultimately fail (although I suspect it is right), but it is backed by solid science and worthy of proper evaluation. We must not abandon the quest halfway, without performing prevention trials and getting a definitive answer to the amyloid question.

Sam Gandy

New York, July 21, 2010

The writer is a professor of neurology and psychiatry and an associate director at the Mount Sinai Alzheimer disease Research Center and the James J. Peters V.A. Medical Center.

To the Editor:

Scientists’ understanding of Alzheimer disease may not be clear enough to develop tools for diagnosis and treatment. But Sanjay W. Pimplikar falls into the trap into which many of us physicians find ourselves: thinking that without a medical treatment, “many individuals would simply prefer to be spared the emotional trauma of a diagnosis.”

This runs contrary to the spoken wishes of many people with memory loss. They are grateful to hear that their disorienting and frightening experiences have a name. And if the Alzheimer disease is diagnosed early … they can actively plan for their future.

Susan Czapiewski

St. Louis Park, Minn., July 21, 2010

The writer is a geriatric psychiatrist who treats Alzheimer’s patients.

These opinions, strikingly at odds with one another, express common tensions at work in the AD world today that have been evident in earlier chapters. When I interviewed Sanjay Pimplikar in Cleveland two months after the publication of the op-ed, he stated ruefully that he had been wondering whether writing it had been “a smart thing to do.” He had received negative feedback from several of his colleagues that disturbed him:

[I]t really feels like being on a rubber dinghy out on the ocean in a storm—one moment you are high up, and at the next moment you are down there. I get so many emails and phone calls from people from all over. It’s probably a ratio of 50:1—the 50 being people I don’t know. But most of the messages make me feel good. There is a reward, but I haven’t been able to get on with my science—that’s a big price.

Pimplikar describes himself as someone who works on “membrane trafficking,” and believes that he is in many ways an outsider to the AD field because initially he was interested in the APP (amyloid precursor protein) as a marker in a general way, but not in connection with AD per se. In his early work he studied a small cytoplasmic fragment that results from enzyme cleavage of APP that had not at that time been researched by anyone. His team found that this fragment has an effect on the expression of several genes, and the group eventually produced a transgenic mouse that expresses the cytoplasmic fragment alone; such animals show “many of the characteristics of AD—tau pathology, neurodegeneration and inflammation—but no plaques,” says Pimplikar: “You have to ask a philosophical question; if they don’t have plaques, do they have Alzheimer’s?” The neurologist Peter Whitehouse was present at the interview, and he interrupted at this juncture to say, “I’m not sure that people have ever seen tangles in mice.” Pimplikar responded, “The question is how do you define a tangle. By twelve to fourteen months, as the mice start to get old—they live up to two years [he inserted for my benefit], we start to see tangles with Bielschowsky’s stain, but if you want them confirmed using electron microscopy—we haven’t done that yet.” This settled the matter for Whitehouse, exhibiting neither plaques nor having confirmation of tangles, he said emphatically, “no mouse gets Alzheimer disease.”

When discussion turned away from mice to humans, it was clear that Pimplikar and Whitehouse have a fundamental difference of opinion that extends beyond the utility of animal models. Taking friendly issue with Whitehouse and the title of his recent book, The Myth of Alzheimer’s, Pimplikar insisted, “In practical, real terms, Alzheimer’s is a disease and by no means a myth … once we understand it, it will be preventable. Maybe not curable once it becomes fully a disease, but preventable … if I thought it was not preventable then I would probably be studying something else. … For me, Alzheimer disease is like a person who’s not physically fit who finds him or herself on the expert ski slope. They go down, because they have no choice, but in a very bad way; out of control, and there’s nothing you can do to stop it. … We all age, but some of us age in a graceful manner whereas others simply lose it.” Pimplikar is in agreement with the position taken by his neuropathologist colleague, the late Mark Smith, that inflammation may well be a crucial component of AD, and this is why he believes the condition may be preventable.

While talking with me, Pimplikar added that two principal concerns had made him write the op-ed: first, that the Times publications were, in his opinion, making the AD problem much too “black and white,” whereas in real-life research “there is never a ‘yes’–‘no’ answer—it’s always, always a shade of gray.” He added that Kolata’s article “seemed to be trying to promise people something that is really not true.” The second and yet more troubling concern for Pimplikar was the “enormous financial burden that would be placed on society” by moving to biomarker detection when the “outcome would be really miniscule.”

Sanjay Pimplikar’s op-ed had pleased Peter Whitehouse and brought about a restoration of their good relationship that, although it always remained congenial, had suffered somewhat following the publication of Whitehouse’s book. In The Myth of Alzheimer’s one of Whitehouse’s major claims is that “so-called Alzheimer disease cannot be differentiated from normal aging … there is no one biological profile of Alzheimer’s that is consistent from person to person, and all the biological hallmarks of AD are also hallmarks of the aging brain.”4 His argument is also one about inappropriate taxonomic classification in which he insists that numerous conditions, and not only several kinds of dementia, are routinely inappropriately sorted into the Alzheimer’s category. His purpose in writing The Myth is to expound more successful ways of aging, based largely on lifestyle changes, family support, and community interactions, including cross-generational activities involving the young and the elderly.

As a basic scientist who works with mouse models, Sanjay Pimplikar’s discussion of AD tends to be constrained by what can be produced and reproduced consistently in the laboratory. His approach and vision are very different from those of Whitehouse. He stated again and again during the interview that he is not a clinician and his task is not to tend to troubled patients; his work is to get the science right. But he added that he dislikes drug companies and, because they need payback, “we are stuck with amyloid because such a lot of money has been put into it … it takes a long time for science to be self-correcting. … And for those people who believe in the amyloid hypothesis, no amount of negative evidence is going to convince them to change their minds. It just doesn’t … the human mind just doesn’t work that way … it’s going to be like the Copenhagen debate between Heisenberg and Niels Bohr. … Heisenberg, smart as he was, never got the probability part of quantum theory … but after a while, everyone moved on [it was actually fission theory that was at issue].” In contrast to Whitehouse, Pimplikar believes that we must struggle until we better understand the involved molecular biology; only then will we be able to tackle AD prevention. On the other hand, Whitehouse argues that because of the inherent, seemingly irresolvable entanglement of aging and dementia, we should focus on improving the quality of life of people as they age, a strategy that may well reduce the incidence of dementia. An incisive normal/pathological distinction cannot be made as far as he is concerned, and putting the bulk of our resources into drug development is not appropriate.

Many clinicians and researchers, regardless of their competing positions about AD ontology, have been quick to produce cautionary responses about the highly publicized move toward systematization of a presymptomatic diagnosis of AD. Even so, the AD juggernaut is proceeding inexorably toward identifying and attempting to standardize embodied biomarkers and associated estimates of AD risk. In order to carry out this feat, a new definition of Alzheimer disease is indispensible. The earlier formulated European publications discussed in the previous chapter were not sufficient; the NIA and the Alzheimer’s Association in the United States had also to make their position very clear, with the result that their revised three-part report was published in April 2011, and a fourth part, on the neuropathologic assessment of AD, was published in 2012.5

Among the numerous authors of the companion articles that elaborate on the new recommendations, by far the majority had to report “conflict of interest disclosures,” because they receive money from, consult for, and/or have stocks in drug companies. The exceptions are those individuals employed by the Alzheimer’s Association or the government of the United States, and two university-based researchers.

Unhooking the Clinical-Pathological Entity of an Outdated AD

In previous chapters it has been pointed out repeatedly that over the years, and with increasing frequency in the past decade, research has shown that no tight correlation exists between clinical symptoms associated with AD and the neuropathology believed to underlie Alzheimer disease. The introduction in the journal Alzheimer’s & Dementia that precedes the three-part companion articles not only is explicit about this lack of fit, but also makes it clear that this incongruence has been the wedge that has opened up a new approach to AD. With the purpose of moving AD research forward, the recommendations set out revised NIA–Alzheimer’s Association criteria in which a “semantic and conceptual distinction is made between AD pathophysiological processes (AD-P) and the various clinically observable syndromes that result (AD-C).”6

The authors of the essay emphasize that research over the past quarter century has shown repeatedly that distinctive molecular changes in the form of biomarkers consistently occur in individuals prior to the onset of dementia due to AD. They add that these changes are often accompanied by mild clinical symptoms, changes that do not warrant a diagnosis of AD, but are characteristic of “an intermediate stage between normality and dementia.”7 The new recommendations state that from now on these intermediate changes—biomarkers and mild clinical symptoms—will be systematically incorporated into the diagnostic procedure.

The authors affirm that evidence of neuritic plaques containing beta-amyloid present in some parts of the neocortex, together with regional distribution of neurofibrillary tangles, will continue to be the defining features of AD-P.8 An additional comment in the introduction about genetic findings makes the claim that “the available genetic risk data overwhelmingly point to the β-amyloid pathway as the initiating, or at least a very early pathophysiological event in the disease cascade.”9 Emphasis is also given to a time dimension associated with the development of AD pathology—individuals do not suddenly become demented, nor do they exhibit all the molecular changes at the same time. Notably, amyloid pathology and neurodegenerative pathology in the form of tangle formation and neuronal/synapse loss usually take place on different time scales.10 β-amyloid pathology develops first, during a long presymptomatic phase of AD now well documented by neuroimaging. In contrast, neurofibrillary pathology and synapse loss begin to accelerate shortly before the appearance of the symptomatic phase of AD, during the “intermediate stage” between normality and dementia (emerging research suggests that this timeline is by no means as clear-cut as is claimed in these revised guidelines).11

img

Figure 4.1.

The continuum of Alzheimer disease (AD). Model of the clinical trajectory of AD. The stage of preclinical AD precedes mild cognitive impairment. Note that this diagram represents a hypothetical model for the pathological-clinical continuum of AD but does not imply that all individuals with biomarker evidence of AD-pathophysiological process will progress to the clinical phases of the illness. Reprinted from Alzheimer’s & Dementia: 7, no. 3, Sperling et al., “Toward defining the preclinical stages of Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease”, 280-292, © 2011, The Alzheimer’s Association, with permission from Elsevier.

If the diagnostic gold standard of plaques and tangles is to remain in place as the embodied representation of AD, has a significant change in AD conceptualization taken place or not? Given that a long presymptomatic time dimension of up to two decades or more is now acknowledged, during which supposedly significant molecular changes inherent to AD formation take place, AD is now undeniably understood as a process—a continuum—one that results in localized pathology. The current move to prevention, involving a temporal backtracking in order to bring about a nondemented future, is described in the medical literature as a “paradigm shift,” and the emphasis on a continuum of long duration not fully appreciated until now by involved experts is indeed new. Biomarker technology has made this insight possible. However, the dominant model of AD causation survives unchanged, and amyloid deposition continues to be the prime event of interest—there is no new theory about AD causation—suggesting that Thomas Kuhn would almost certainly disagree with the neurologists now making a claim for a paradigm change.12 Rather, the new approach to Alzheimer disease is perhaps an example of a shake-up in “normal science” in which the dominant model remains, although subject to considerable criticism and modifications.

However, given that it is freely acknowledged that persistent anomalies in research findings have been “the wedge” that has opened up a new approach to AD, it is possible that the field is entering a preparadigmatic stage prior to a major shift in orientation. Kuhn argued that “a proliferation of competing articulations, the willingness to try anything, the expression of explicit discontent, the recourse to philosophy and to debate over fundamentals” constitutes a crisis in a scientific field, and it is quite conceivable that out of such ferment new ideas may well arise, leading to new methods, and finally to a new theory—this would then constitute a paradigm change.13 As Ian Hacking notes, for Thomas Kuhn, novelty is the hallmark of science, “without revolution, science would degenerate.”14 Kuhn was, of course, initially a physicist, and he had physics in mind when he wrote The Structure of Scientific Revolutions. We will return to this discussion in the concluding chapter with reference to the complexities of the biology of Alzheimer disease, in which the law-like characteristics of physics are nowhere apparent.

Striving for Radical Change

Following the introduction, the first of the three companion pieces in the report titled “The Diagnosis of Dementia Due to Alzheimer Disease” outlines diagnostic criteria that are flexible enough to be used by both general health care practitioners and specialized investigators. It is argued that several features of the original NINCDS-ADRDA criteria that have been in extensive use over the past 27 years should be modified. In contrast to the European recommendations noted in chapter 3, the categories of “possible” and “probable” AD are retained in order to distinguish such patients from individuals with other kinds of dementia. The article concludes that a diagnosis of AD dementia should be made by means of clinical assessment; in everyday clinical practice, such a diagnosis is indispensible, and biomarker support, if used at all, should be secondary—this emphasis on clinical assessment differs from the European position.

The second of the companion articles deals with the diagnosis of MCI. This article stresses that it is particularly challenging for clinicians to identify transition points when patients progress from a presymptomatic to a symptomatic predementia phase because this progression is one of a slow steady change with no sudden events to define it. The core clinical criteria for determining MCI outlined in this article are expressly designed to be of use in every kind of clinical setting (again a position that differs from the European recommendations). Like the first of these companion pieces, it is emphasized that biomarker testing should be limited to research settings, but even so they can provide confirmation of a clinical diagnosis. The authors use the term “mild cognitive impairment (MCI) due to AD” to distinguish this state from cognitive impairment due to other forms of dementia. They stress that the degree of cognitive impairment diagnosed in this category is “not normal for age” (see Appendix 1 for a summary of clinical and cognitive evaluation for MCI due to AD).

The final companion article sets out by warning readers once again of the negative economic consequences looming on the horizon if nothing is done to lower the global incidence of dementia. Citing data from the Alzheimer’s Association, it is noted that more than 13.5 million individuals in the United States will manifest AD by the year 2050 if the disease incidence continues at the present rate, with projected Medicare costs of $627 billion. A reduction in onset of the disease by five years would halve this bill.

This article makes an argument for moving to a preventive mode because of the dismal results to date in attempting to deal with AD, notably a lack of effective therapeutic interventions. Once again it is stressed that these recommendations are being made for use in research settings alone: “The extent to which biomarkers of AD-P predict a cognitively normal individual’s subsequent clinical course [to AD-C] remains to be clarified, and we acknowledge that some of these individuals will never manifest clinical symptoms in their lifetime.”15 Subsequently it is claimed, “Recent advances in ante-mortem biomarkers now allow us to test the hypothesis that many individuals with laboratory evidence of AD-P are indeed in the preclinical stages of AD, and determine which biomarker and cognitive profiles are most predictive of subsequent clinical decline and emergence of AD-C.”16 It is suggested that once tested, individuals who are “cognitively normal” but biomarker positive be classified as “asymptomatic at risk for AD.” Furthermore, it is pointed out,

The difficulty in the field of AD is that we have not yet established a firm link between the appearance of any specific biomarker in asymptomatic individuals and the subsequent emergence of clinical symptomatology. If we can, however, definitively determine the risk of developing AD dementia and the temporal course of clinical progression associated with AD-P in individuals without dementia or MCI, we will open a crucial window of opportunity to intervene with disease-modifying therapy. Although we hypothesize that the current earliest detectable pathological change will be in the form of Aβ accumulation, it is possible that Aβ accumulation is necessary but not sufficient to produce the clinical manifestations of AD.17

Following this comment, the authors write, “[W]e acknowledge that the etiology of AD remains uncertain, and some investigators have proposed that synaptic, mitochondrial, metabolic, inflammatory, neuronal, cytoskeletal, and other age-related alterations may play an even earlier, or more central, role than Aβ peptides in the pathogenesis of AD.”18 They also note that arguments persist as to whether the problem of Aβ lies with its “abnormal processing” or, alternatively, with its ineffective “clearance,” and, furthermore, that it is now known that different forms of Aβ exist, some of which can have a protective mechanism on healthy brain functioning, although these forms may become faulty under circumstances as yet unknown. In conclusion, the authors remind their readers that genetics is implicated in AD etiology and, moreover, epidemiological data suggest that “there are significant modulating factors that may alter the pace of the clinic expression of AD-P.”19

A list of caveats appears toward the end of this third companion article: First, many of the studies on which findings about biomarkers are based are likely to have “cohort biases” and are probably not representative of the general population, particularly when the “oldest-old” are implicated. Second, caution is noted because the biomarkers of interest are merely “proxies” for the disease, and may not fully reflect the processes taking place in the living brain. Last, but not least, it is reiterated that “the role of Aβ as the etiologic agent in … AD remains to be proven.”20

The excerpt cited above together with the additional comments make it glaringly clear how many unknowns are involved in this so-called paradigm shift, and it takes little imagination to appreciate the many years of research that lie ahead to establish even a modicum of certainty as to what is involved in AD etiology and, above all, who among us is “at risk.”

Even if biomarker evaluations are limited to research settings, individuals who become research subjects will be left in limbo, a heightened state of anxiety that can only intensify once they have tested positive for one or more biomarkers, and then must undergo repeat tests at regular intervals. Given that none of the biomarkers in use are “static” (in researcher language) and rates of change in each biomarker are not consistent over time and follow a nonlinear trajectory, it is unlikely that it will ever be possible to give individual research subjects, and eventually patients, information about their risk status or prognosis on which they might act. Whatever individuals are told could well be thoroughly misleading and, if imparted with appropriate caution, should be filled with equivocations. Even so, when I attended the 2011 meeting of ICAD in Paris four months after the new recommendations had been published, it came as no surprise to hear from one California neurologist as he delivered his research paper that his patients were asking for biomarker tests for AD and that, in his opinion, they have a right to pay for such tests and be given their results, if they so wish.

In February 2012 John Morris, long unwilling to accept the concept of MCI, published an article that clearly signals the extent of the difficulties, clinical, ethical, and social, that will arise as these new guidelines begin to be put into practice. Morris examined the records of over 17,000 individuals who had been evaluated at Alzheimer disease centers in the United States and whose files were submitted to the National Alzheimer’s Coordinating Center. He concludes, under the revised guidelines for MCI (see above), “Almost all (98.8%) individuals currently diagnosed with very mild AD dementia and the large majority (92.7%) of those diagnosed with mild AD dementia could be reclassified as having MCI using the revised criteria. This would be based on their level of impairment in the Clinical Dementia Rating domains for performance of instrumental activities of daily living in the community and at home.”21 Morris concludes that the categorical distinction between MCI and milder stages of AD dementia that have been in place until recently are now compromised, particularly by a new expansive use of the idea of “independence in functional activities.”22 He is particularly concerned because the distinction will now depend on the individual judgment of clinicians that will “result in nonstandard and ultimately arbitrary diagnostic approaches to MCI” that will “confound clinical trials … complicate diagnostic decisions … and research comparisons.”23 Morris highlights the ambiguity already present in the MCI concept, and the repeated efforts over the years to revise and standardize the diagnosis. He insists that with the revised criteria in use

the distinction of MCI from dementia now simply is a matter of an individual clinician’s threshold for what represents one condition vs. the other. … Already many individuals with MCI are treated with pharmacological agents approved for symptomatic AD, indicating that clinicians often do not distinguish the two conditions when faced with issues of medical management. It is now time to advance AD patient care and research by accepting that “MCI due to AD” is more appropriately recognized as the earliest symptomatic stage of AD.24

Morris notes that in 2010, 1,220 publications appeared in which MCI was listed in the title, keywords, or abstract. It seems likely that publications will continue to appear at a great rate on this subject as experts work to sort out the ambiguities raised by the new criteria, and it is quite possible that diagnostic creep will take place in research settings due to the broadening of MCI diagnostic criteria. The question remains as to whether increased diagnoses of MCI will also take place in most clinical settings—a possibility that calls for systematic inquiry.

Future Prognostications

More than 15 biomarkers are already used in connection with AD research, subdivided into three primary divisions according to their molecular activity: (a) biomarkers of Aβ deposition, (b) biomarkers of neuronal injury (including changes in the protein tau associated with tangle formation, hippocampal volume, and evidence of brain atrophy, and (c) biomarkers of biochemical changes associated with AD (including inflammatory biomarkers, signs of oxidative stress, and/or cell death) (see Appendix 2). In addition many other posited biomarkers are being researched. It is assumed that if a plasma biomarker can be found in the near future, ensuring that a simple blood test is all that will be required to assess who is at risk for AD, it will then be possible to implement a public health strategy in which the prevention of AD takes center stage—but so far this is not in the cards. The two biomarkers that have been primarily examined to date are markers in cerebrospinal fluid (CSF), and the use of PET (positron emission tomography) neuroimaging to demonstrate deposits of plaque in vivo. In addition genotyping for the APOE gene is a third well-researched biomarker (see chapter 5).

A consensus report of the Working Group called together to consider “Molecular and Biochemical Markers of Alzheimer Disease” was published as early as 1998, sponsored by the Ronald and Nancy Reagan Research Institute, an affiliate of the Alzheimer’s Association and an NIA working group. This report set out criteria for defining, developing, and assessing biomarkers for AD and stated that their detection should be, above all else, simple to perform, inexpensive, noninvasive, and precise, with high levels of sensitivity and specificity.25 Genetic testing was included, but the group devoted most of its attention to CSF. In the late 1990s it was not yet possible to detect amyloid in the brain using imaging methods.

It is perhaps not surprising that attention was given early on to the possibility of CSF as a biomarker for AD; CSF has been described on occasion as a “biochemical window” into the brain. It is a clear liquid that surrounds the structures of the central nervous system (the brain and spinal cord), has more physical contact with the brain than any other type of fluid, and is not separated from the brain by the tightly regulated blood/brain barrier. Normal CSF contains Aβ in its usual soluble form, but it has been repeatedly shown that levels of this soluble peptide drop in the CSF when conversion into its insoluble form starts to take place in the brain, resulting in deposits of the plaque characteristic of AD and certain other conditions. It is assumed that a drop in soluble Aβ in the CSF is the first detectable biomarker associated with the onset of the AD continuum, and that it can be consistently detected prior to evidence of any clinical symptoms.

A second biomarker found in CSF associated with the onset of AD is an increase in the total amount of tau protein, and also of phosphorylated tau linked to tangle production. These changes become apparent later than the reduction in soluble beta-amyloid levels, and are associated with a stage in the continuum when MCI or early signs of AD become clinically apparent. A recent editorial in the Archives of Neurology titled “Sharpen That Needle” argues that what is measured in these CSF changes “reflects the biochemical composition of the anatomic lesions” of AD.26 But the editors are quick to point out that other diseases bring about similar changes. It is suggested that these biomarker determinations should be used only as a “diagnostic aid.” They add that the Reagan requirement for noninvasiveness is in the eye of the beholder, and that CSF taps are no more invasive than, for example, endoscopies. CSF tests have already been made use of in research settings for a couple of decades.

The writers of this editorial argue for patient education about the worth of spinal taps, and note that until relatively recently granting permission for autopsies was not regarded with favor, although this has changed in recent years—hence public attitudes about procurement of CSF can also be changed. Readers learn that physician training is needed, and it is suggested that specialist clinics might be set up in neurology departments because primary care physicians are neither “skilled nor experienced” in such procedures. These editorialists, “gazing into the future,” as they put it, envision a time when CSF analysis will be implemented as a routine screening program to identify clinically healthy individuals at risk for MCI and AD, although the NIA–Alzheimer’s Association guidelines specify that, for now, use of these biomarkers should continue to be limited to research settings.

In the same issue of Archives of Neurology an article appeared reporting results of a longitudinal study designed to identify biomarker patterns typical for AD without knowledge of information on the clinical diagnosis, using samples of “cognitively normal persons, patients with AD, and individuals with mild cognitive impairment.” An AD “signature” was found in 90%, 72%, and 36% of individuals in the AD, MCI, and cognitively normal groups, respectively. The cognitively normal group proved to have a high proportion of people estimated to be at increased risk for AD due to their genetic makeup (see chapter 5). In a related data set of individuals diagnosed with MCI who exhibited these biomarker changes, 100% progressed to AD during the course of the following five years. The explicit assumption on the part of the researchers is that those normal controls who exhibited biomarker changes are without doubt on their way to MCI and then to AD. The authors argue, “AD pathology is active and detectable earlier than has heretofore been envisioned.”27 Emphasis is given to the value of being able to assess more than one marker at the same time, and it is concluded that “the AD signature appears to be naturally present in the data and is expressed as a homogenous group, consistent with a single pathological process underlying AD.”28 Their findings corroborated earlier findings that the first changes of the AD footprint appear to be a drop of soluble β-amyloid in the CSF, followed thereafter by changes in tau. The authors of this article argue that it is not surprising that their cognitively normal research subjects have both amyloid-containing plaques and tau-containing neurofibrillary tangles in their brains, and they cite two other earlier studies that have shown these findings.29

This particular article has been cited repeatedly during ongoing discussions in recent months about a preventive approach to AD. No caveats or cautionary notes appear in the article and no mention is made of recent findings indicating that β-amyloid is more complex than has usually been acknowledged to date, or that the changes in both β-amyloid and tau, assumed to be the AD signature, are regularly detected in patients with other neurological conditions.

In contrast, other researchers have voiced criticisms about the claims being made about the value of tracking CSF biomarkers, the most common being that there is, as yet, a lack of standardization, and no adjustment is made for the age of patients, even though the significance of β-amyloid levels in CSF may be quite different in 80-year-olds than in 50-year-olds. Those researchers who support the idea that there are many roads to dementia and do not recognize AD as a single disease entity are very uncomfortable with the assumption that CSF biomarkers are the undoubted predictors of future AD. One participant at the 2011 Paris ICAD meeting, David Brooks, a neuroimager based at Imperial College, London, argued that we should think of amyloid as equivalent to high blood pressure; not as a sign of a specific disease but, rather, as an indicator of risk for many possible conditions. He argued that the revised criteria should not become a new religion because the presence of β-amyloid deposition in the brain is not yet standardized among individuals or populations, nor is it known what its presence signifies.

Amyloid In Vivo

The second biomarker to receive a great deal of attention in the AD world, and frequently cited as exceptionally promising, is a recently developed application of functional neuroimaging with positron emission tomography (PET). PET is a technique where radioactive ligands (molecules) are injected intravenously into subjects, and then the distribution of the ligands in the brain is imaged using a highly sophisticated three-dimensional sensing machine. Among many other uses, PET imaging can be used to demonstrate amyloid plaque in the living brain. This is a technique that rules out AD if no plaques are demonstrated, but, if plaques are found, interpreting their meaning is challenging, as we will see below.

In 2002, University of Pittsburgh scientist Chet Mathis and geriatric psychiatry colleague William Klunk reported the first results obtained by using a radioactive PET tracer molecule known as PIB (Pittsburgh compound B) on humans. This report, demonstrating that PIB can bind to amyloid in the brain and then be imaged using PET scanning, was based on a single case from Uppsala University, Sweden. Researchers from Pittsburgh partnered with colleagues in Uppsala for this initial study. A second report by the team using the PIB tracer molecule was published in 2004. These findings showed that when injected into 16 patients diagnosed with “mild” AD, PIB was retained in areas of the cortex known from autopsies to be associated with the accumulation of amyloid plaques, and also in the hippocampus (part of the limbic system associated with memory and emotions) where amyloid also accumulates. In 9 healthy controls little uptake of PIB was demonstrated.

The findings, described as “proof of concept,” encouraged the authors to claim that further research was justified in connection with the relationship between β-amyloidosis and AD, and the following conclusions were drawn:

As the technology of amyloid imaging moves forward, it will be important to avoid the circular reasoning inherent in the association of amyloid deposition with both the diagnosis and the cause of AD. Therefore, at the outset, it may be best not to equate amyloid deposition to clinical diagnosis. Rather than as a method of diagnosis, it might be best to first think of PIB retention more fundamentally as a method to detect and quantify brain β-amyloidosis.30

The authors note that the term β-amyloidosis was first used in 1983,31 and they argue that plaque buildup can usefully be described this way. They point out that several basic, unbiased questions can be asked about amyloidosis: First, how does it correlate with a clinical diagnosis? Second, what is the natural history of β-amyloidosis and its onset relative to clinical symptoms of dementia? And third, can β-amyloidosis serve as a surrogate marker of efficacy for anti-amyloid therapeutics? In the future, the authors conclude, individuals diagnosed with MCI should be included in research protocols, as well as subjects at very high risk for AD due to familial AD genetic mutations. In addition, larger samples of individuals diagnosed with AD should be enrolled as research subjects as well as “normal” subjects for purposes of comparison.

In a 2007 article Klunk and colleagues, making use of a mixed sample of AD and MCI patients and healthy controls—nearly 100 subjects in all—using PIB scans, demonstrated that amyloid buildup did not change significantly over two years. The researchers argue that in nondemented individuals diagnosed with MCI, PIB binding and impaired episodic memory loss are clearly associated, but with progression to AD, the amyloid accumulation appears to reach a “plateau.” They conclude that their findings support the proposal that Aβ imaging can detect the preclinical phase of AD, but that further longitudinal study is required.32

An article published in 2008 by Klunk’s group, in which what was characterized as a community-based sample was made use of, reported results of PIB scans on 43 “nonimpaired elderly” volunteers aged 65 to 88 years. Once again amyloid is detected, but in healthy elderly subjects. The small sample size is acknowledged, but even so the researchers conclude that the findings show that elderly people with significant amyloid deposition can remain cognitively normal. It is noted that longitudinal follow-up of these subjects is needed, and that findings should be replicated using larger cohorts, with the objective of determining if amyloid deposition alone is insufficient to cause AD within some specified period of time.33 Caveats such as this have not stopped the media from commenting on this research positively: “PET scans with PIB clearly distinguish people with Alzheimer’s from healthy people. They may also help identify people with the progressive form of MCI.”34 Mathis and Klunk were honored in 2008 with the Potamkin Prize—the highest award for Alzheimer disease research.

A burgeoning number of research projects have drawn similar conclusions to the findings of Klunk and colleagues, and it is now commonly stated that between 20% and 40% of healthy individuals, also expressed as “up to a third of cognitively normal elderly subjects,” carry plaque in their brains,35 an anomaly that demands explanation. It is evident that these findings are dependent upon age and genotype (there is a greater likelihood of having amyloid plaques in the brain as an individual ages, or if that individual carries the allele of the APOE gene known to increase risk for AD) (see chapter 5), but this does not explain away the finding. It is also recognized that a nonconcordance exists between amyloid deposition, which takes place at a slow, constant rate but does not usually result in detectable cognitive decline, and neurodegeneration, which is commonly associated with the appearance of clinical symptoms and usually accelerates quite rapidly once it commences.

A 2011 summary article concludes that amyloid deposition can take place up to a decade or more before other changes take place. Furthermore, interindividual differences exist in the time lag before neurodegeneration starts, if indeed it commences before death. Such variation is “likely caused by differences in brain reserve, cognitive reserve, and coexisting pathologies.”36 The article points out that neurodegeneration usually both “precedes, somewhat, and parallels cognitive decline,” and can be detected in the form of brain atrophy, notably atrophy of the hippocampus, using MRI. Furthermore, the “neurodegenerative element is the direct substrate of cognitive impairment and the rate of cognitive decline is driven by the rate of neurodegeneration.”37

Luc Buee, from the University of Lille, in France, presenting research at the Paris ICAD conference, noted that tau, the protein associated with neurodegeneration, comes in six isoforms. He added that there is “good” tau and “bad” tau, the second kind being implicated in 20 neurological disorders. Buee used the yin/yang icon to depict how tau is in constant flux, and he explained that it has many more functions than had formerly been recognized, including bringing about changes in the cell nucleus—something that had been ruled out until recently. In short, there is a great deal more to be learned not only about amyloid but also about tau and, above all else, about their roles in neurodegeneration.

Neurodegeneration cannot be detected through PIB PET imaging, and researchers must couple detailed structural analysis of MRI images with PIB PET imaging (figure 4.2) to ascertain a more complete picture of what is taking place over time in the brains of individuals and/or in small population samples (by definition limited in size due to cost). Clifford Jack and colleagues at the Mayo Clinic argue that for the purposes of research trials both types of imaging are indispensible. PIB detection of amyloid is needed if removal of amyloid is the target of the trials. MRI detection of neurodegeneration is key if one is trying to track the relationship of neurodegenerative pathology to clinical manifestations of the disease.38 One other application of PIB imaging, perhaps the most widely used at present, is to refine the AD diagnosis and rule out false positives—many researchers have told me that to date this is the only valid use for this technology in the clinic.

img

Figure 4.2.

A neuroimage composed of multiple reconstructed images of one patient’s MRI (magnetic resonance imaging) scan together with a PIB PET (positron emission tomography with Pittsburgh B compound) scan. The MRI images have been reconstructed with multiple surface renderings and appear as though gray. The PIB PET results are overlaid on the surface, with amyloid-rich areas appearing as whitish coloration. The PIB is distributed in regions of the brain typically known as the “association cortex,” and primary sensory and motor areas are not affected. Reproduced with the permission of Howard Chertkow MD.

The New Phrenology

It was quickly recognized by involved researchers that the number and quality of plaques manifested in a living brain by means of PET imaging needed independent verification. A consortium of researchers decided that the best method to do this would be by comparing imaging findings with those obtained at postmortem—the AD gold standard. An ingenious, rather disturbing method of carrying out this comparison involved the voluntary recruitment of patients in hospice, long-term, and community health care facilities who were likely to die within six months. These individuals were invited to participate in a study in which they agreed to undergo PET imaging and then to donate their brains for autopsy once deceased. An article published in JAMA in 2011 was based on the first 35 individuals, average age of 79 years, to come to autopsy, all of whom had undergone a PET scan on average three months before their death.39 Only 17 of these subjects had been diagnosed with AD. A control sample of 74 individuals aged 18 to 50 years of age (so-called young individuals) whose brains, it was assumed, would be free of amyloid also underwent scans.

This multicenter study, 23 sites in all, enrolled people from several parts of the United States and made use of a new PET imaging ligand. The compound, known as florbetapir, has a similar chemical mode of action to PIB in binding to amyloid plaque in the brain, but it has a distinct advantage over PIB because it has a longer half-life and can, therefore, be used in imaging centers that do not possess a cyclotron. The conclusions drawn from this study were that this first prospective study of florbetapir-PET imaging correlated well with the presence and density of β-amyloid at autopsy and therefore provided evidence of the effectiveness of a molecular imaging procedure that can identify β-amyloid in the brains of living individuals. It is noted in a final comment that individuals may have differing abilities to “tolerate aggregated amyloid” in their brains and that this depends upon “genetic risk factors, lifestyle choices, environmental factors, and neuropathological co-morbidities.” The researchers acknowledge the small sample size of this study.40

This particular research project has received extensive criticism on several grounds, in addition to that of sample size, although none of the objections are based on ethical issues as far as I can ascertain. Writing an editorial in JAMA, the epidemiologist Monique Breteler points out that the participants in this study

were not a random sample of individuals with cognitive problems who might be tested for presence and extent of amyloid pathology, nor were they a random selection of asymptomatic persons who might be evaluated to determine whether some of them already had amyloid pathology developing. Rather, they were a convenience sample of terminally ill persons who, for the most part, had advanced dementia and non-dementia disorders, mostly cancer. The reported agreement of 97% between florbetapir-PET and the postmortem pathologic diagnosis is likely inflated because of the selection of participants in the study.41

Breteler, in striking contrast to the 20 authors of the original article, who have more than a column worth of conflict of interest disclosures, has none. She goes on to note that the absence of any scans rated amyloid-positive in the young control group was assumed to reflect a very high degree of specificity of the imaging. But she adds, “[J]ust as the observation of a flock of white swans does little to support the notion that no black swans exist, so the observation of amyloid-negative scans in young persons does little to establish the specificity of this imaging mode in persons aged in their 70s and 80s, who may or may not have varying amounts of coexisting brain atrophy and other pathological changes.”42 Breteler adds for good measure, “β-amyloid accumulation in the brain is likely not the culprit but rather an epiphenomenon of developing AD.”43

A second major critique is to do with interrater reliability. In a letter to the editor published by JAMA,44 two physicians, Carome and Wolfe, both involved with citizen awareness groups, pointed out that there was substantial variability in rating among extensively trained readers of the PET scans in this study and that the assessments of the three readers had been averaged out in the final publication. It was noted that should use of PET neuroimaging become more widespread, it would be very unlikely that readers would have had the extensive training given to these three individuals, with the result that “the variability and unreliability is likely to get worse,” suggesting that false positives and false negatives would be the outcome. The authors of the study attempted to rebut the criticism of their publication and to justify at length the reasons for averaging out the reader findings in a letter published together with that of Carome and Wolfe in a later issue of JAMA.

When writing about the use of PET scanning in schizophrenia research, Joseph Dumit remarks that publications about illustrative cases show a clear differentiation between people labeled as schizophrenic and those without “even though there are many people diagnosed with schizophrenia whose brains look like those of people without, and people without schizophrenia whose brains look like those of people with it.” The scans of volunteers are labeled as “normal controls” and the scans of those diagnosed with schizophrenia prior to scanning are labeled as schizophrenia. Dumit concludes, “[T]he image is thus labeled as showing the ‘disease’ itself, rather than a correlate symptom of someone found to have schizophrenia. Hence, the symptom has been collapsed into the referent.”45 Currently, considerable efforts are being invested in the AD world to replicate and standardize findings resulting from PET scanning of in vivo amyloid deposits in research settings. And other research is attempting to show the relationship between lowered β-amyloid in CSF fluid and an increase in plaques detected by means of this type of neuroimaging.46 The findings associated with schizophrenia are clearly problematic and much more elusive than amyloid. The “realness” of amyloid is not in dispute, particularly because it can be shown at autopsy, but what it signifies in vivo, especially for the future health of individuals, is debatable. A bias may well creep into the interpretation of images of amyloid in the living brain if it is categorically assumed that amyloid predicts an AD future, and such an assumption will in turn influence interpretations of the significance of biomarker findings that in effect come to assume the trappings of the disease itself.

Summarizing thus far, “slippage” between apparent neuropathology and cognitive well-being is overwhelmingly evident as a result of neuroimaging, but the interpretation of these findings continues to be disputed. Although the position taken in the new diagnostic guidelines calls for caution and argues explicitly that this anomaly cannot yet be explained, a good number of the individual authors who contributed to these guidelines, notably John Morris and colleagues, persist with their belief that every individual who exhibits preclinical biomarkers is on the way to Alzheimer’s. The assumption is that only death can intervene in this unavoidable process. In contrast, other researchers focusing on elderly individuals who show few if any signs of cognitive impairment, even when neuropathology is evident in their imaged brains, assume that although death is clearly inevitable, AD is not; nor can it ever be shown to be so. Many among this latter group of researchers believe that greater efforts should be made to clearly establish, to the extent that it is possible, what it is that protects people from becoming demented well into old age. Yet other researchers equivocate (see below).

Rethinking Amyloid

One thing seems to be clear from the findings of epidemiologic community studies as well as PIB scanning thus far: those of us old enough to have trouble retrieving a word or two now and then can safely assume this is not a definitive sign of approaching Alzheimer’s, although it is quite possible that β-amyloid may be beginning to clutter our brains and/or the effects of one or two small strokes may contribute to the situation. Plaques appear rather unsightly when neuroimaged, but these garbage bags (as one geneticist described them to me) may well simply accumulate unobtrusively little by little, and have never been proven to be definitive signs of a precarious future, even though they are almost universally present in the brains of patients diagnosed with AD. Furthermore, it is increasingly clear that, up to a point, the proteins that produce plaques are protective of brain function.

When I interviewed Mark Mintun, a neuroimaging specialist who was part of the well-known group researching Alzheimer’s at Washington University in St. Louis (he has subsequently become an employee at the company that makes florbetapir), he made it clear that probably something “goes wrong in amyloid, or at least related to amyloid” at the beginning of the long decline into Alzheimer disease. But, he added,

[S]omebody pointed out that the association between amyloid and AD may just be circular logic, although I don’t think that’s right. It seems now that first you get the plaques but they’re not AD, so there has to be a second step. It seems that in some people, especially those who are homozygous for the APOEε4 gene [see chapter 5] the plaques buildup quickly and become more extensive and will just always push the next step to happen. But for a lot of people, I think that they can either develop some sort of equilibrium in the sense that plaques don’t actually trigger the next step, or perhaps they don’t ever get a high enough numbers of plaques. Plaques are there, but there are not enough of them. Or they’re not of the right type. One of the things we are learning is that there are plaques that PIB binds to and there are plaques that it doesn’t bind to. And you can actually have quite a lot of diffuse plaques and not any PIB binding. The neuritic plaques, dense extracellular plaques, have long been associated with AD, and have been the diagnostic gold standard since Alzheimer’s day.

Mintun continued,

[I]t’s true you almost never have AD if you don’t have neuritic plaques. But if you lived in a moderately toxic environment this might affect diffuse plaques, at this stage no one can say that those diffuse plaques are safe; it’s just that they aren’t, at the point that we look at them, leading clearly to AD. But they could transition quickly under certain circumstances. So we may be missing a lot of the iceberg that’s really close to us.

For all we know if everyone lived to 118 we’d all get AD. But some people get there much earlier, and your genetic load and the environment can accelerate it. Diabetes and heart disease and not doing crossword puzzles can accelerate things. But the person who exercises and is mentally active may live with amyloid in his brain for 20 years because he has sort of, you know, he’s growing more neurons and synapses as fast as the amyloid is killing them, or something like that. And his immune system is properly balanced, and he doesn’t have a run-away inflammatory process. But what we have now is a gap between the two anchors involved with AD. We now know much more about the first anchor—the neuritic plaques, but we don’t know what sets off the second anchor. Right now I’m looking at the relationship between plaque and free radicals. (May 2009)

Discussion then turned back to the PIB findings, and Mintun reminded me that PIB is negative in 60% to 70% of the people who become experimental controls in their research. But he added that this does not necessarily mean that they do not have any plaques at all: “[T]hey may have some that I simply cannot see. Or they may just have diffuse plaques.” When conversation turned to the 30% of “normal non-demented controls” who test PIB-positive, Mintun made it very clear that he believes they are indeed “drifting” toward AD, as opposed to individuals who exhibit no plaques. He argued that cognitive testing carried out regularly over 10 or more years begins to show a drift downward in the plaque positive group, whereas those without plaques retain their cognitive status. Mintun is quick to add that it is not possible to see this with only a 5-year follow-up; a longer time frame is essential, and the sample followed must be large enough—several hundred individuals—such as they are currently studying at Washington University. In an email exchange several months later, Mintun made the following comment: “But, this is not a slam-dunk. There is very much the possibility that people who are in the normal range for cognition but have amyloid in their brains may not all end up with Alzheimer disease symptoms. [An analogy is that not all people with atherosclerosis and coronary artery disease end up with heart attacks]. However, it is reasonable to hypothesize that the fraction that go on to be impaired may be very high.” Mintun supported this final comment on the basis of research in progress, but noted the shortcomings of the studies to which he refers.47

Peter Nestor, who does neuroimaging at Cambridge University, also believes that those people who are healthy but positive for PIB scans are on their way to becoming demented, but have not, as yet, reached a “threshold” for symptoms: “[I]t’s an insidious process, and you can see from biomarkers that it’s a pathology years before that threshold is reached.” He adds,

[O]f course, a positive PIB scan doesn’t tell you when its going to start, and so if you’re completely healthy and you’re 75 years old and you get a positive scan, how useful is it to know that? I find that I can often cheer people up when they worry if they are going to get Alzheimer’s because their mother and grandmother had it. I remind them that their relatives were both in their 90s when they died, but I also point out that no one in my own family has been diagnosed with AD, and that’s because no one in my family has ever survived beyond 80—my family all die relatively early of something else! For some reason this seems to cheer some people up—it makes them think that maybe they are lucky in a way! I think it’s really helpful to remind people gently that they have to die of something and, yes, their chances of getting AD may be a bit higher than someone else because of their family history, but it’s not necessarily destiny. We still have so much more research to do before we can make big claims about AD. (September 2010)

Howard Chertkow had the following to say when asked what he thinks is the significance of findings indicating that approximately 30% of healthy elderly people have tested PIB-positive:

I think it can mean one of several things: that some of these people are indeed in a presymptomatic stage of dementia; alternatively, it may be that some, or many, have sufficient cognitive reserve for one reason or another that they will eventually die of something else without crossing the threshold that precipitates dementia. And, most important, we simply don’t know enough as yet. It’s only six or seven years since we have been doing PIB scanning. Insisting that everyone with amyloid in their brains is going on to get AD is an act of faith—we have no proof of this. You would have to follow a cohort to age 100 and if any of them were not demented then you’d have to follow them to age 120 to be sure! You see the problem! So, it’s a bit of a religious assertion to insist that “every older person with plaques will eventually go on to get AD.” You can never prove this kind of statement. Marsel Mesulam [a Chicago-based neurologist] and others say the opposite: “the brains of old people have a lot of abnormalities in them and since most of these individuals never get demented, it is obvious that the mere presence of some plaques or tangles, or even infarcts, is not enough to guarantee eventual dementia.” This is actually a more scientific viewpoint. (October 2010)

Chertkow went on to remind me that one has to ask, “What is normal?” (He might have said, “Normal for what? Normal for whom?” as did Redlich 30 years ago.)48 And he made it clear that this question cannot be considered in abstraction from aging: “All 20-year-olds are PIB-negative, and 50% or more of people aged 90 and over are PIB-positive. If one chooses only 75-year-olds as normals for research, and their reaction times to cognitive testing are average, then about 95% of this population will, by definition, be excluded, because they have slowed down a bit and no longer count as ‘normal.’ And should people with diabetes and obesity be excluded? These conditions are now well recognized as risk factors for AD.” Chertkow pointed out that researchers such as himself who make a rigorous selection of “normals” for PET scanning by drawing only on volunteers from the population at large who have no subjective memory complaints and who cope well with cognitive testing find that only 10% are PIB-positive. This is in striking contrast to research where normals are chosen from among people who have voluntarily presented themselves to a memory clinic or similar facility for testing, in which case the number of PIB-positives can be as high as 45% or 50%. This is the situation with the ADNI (Alzheimer’s Disease Neuroimaging Initiative) study, to which we now turn.

Toward Neuroimaging for All

The idea of a national initiative to prevent Alzheimer disease has been on the table since 1987, but despite ongoing effects to reduce its incidence and standardize diagnoses, relatively little progress was made, as we have seen. By 2009 a renewed effort, focused on early detection and prevention, was set in motion. The promotional banner this time around is the “Campaign to Prevent Alzheimer Disease by 2020.” This latest national strategic goal, argued Zaven Khachaturian, the editor-in-chief of Alzheimer’s & Dementia, is “mandated by the looming financial catastrophe facing the U.S. national health care system.”49 In an editorial the previous year, Khachaturian, whose mother died of probable Alzheimer disease, wrote the following passage:

The array of challenges for the mission to prevent AD within a decade are no less daunting than those faced by similar national endeavors such as the Apollo space program, the Manhattan Project, or the Human Genome Project. Ultimately, the execution of this national enterprise will require decisive actions by both public and private entities, as well as bold public policies that foster radical changes in: 1) the governance and organization of research, 2) mechanisms or programs of research funding, 3) deployment of resources and infrastructure, and 4) paradigms for developing interventions/treatments. … One of the most critical challenges for the national mission to prevent AD is the need for centralized control and coordination of all AD-related activities.50

Khachaturian went on to list the programs and consortia that have been put into place over the past three decades—11 of them in all, many managed by different governmental agencies. And yet other assemblages exist that are globally distributed.

ADNI, a public–private partnership, was created in 2004 and is central to the “goal of the proposed National Strategic Plan … to create a new paradigm for planning and supporting the organization of worldwide cooperative research networks to develop new technologies for the early detection and treatment of various forms of memory impairments.”51 The objective of the ADNI partnership is to research changes in cognition, brain structure and function, and biomarkers in elderly individuals selected as research controls, subjects with MCI, and subjects with Alzheimer disease. The technologies made use of by ADNI include MRI and PET scanning, as well as the systematic examination of CSF and blood biomarkers. The data generated by this research are available to researchers worldwide without embargo and may be neither patented nor owned. The governance and administration of ADNI are exceptionally complex, and will not be elaborated on here.52 William Thies of the Alzheimer’s Association made the following observation about the onus that will fall onto ADNI research subjects:

I think in the Alzheimer world we are blessed with having people with the disease who are remarkably willing to give back in the hopes of helping the next generation. You know, they’re not only going to get lumbar punctures but some of them will get three kinds of scans every six months … that’s not trivial. You sit in this machine and it’s claustrophobic and it’s banging away and noisy. With the MRI machines you can hear everything bumping and crashing and sometimes people spend a total of almost 60 minutes in the three scanners and they will be exposed to radiation, of course … this is really an amazing gift to give back to society. (April 2010)

John Morris made it clear just how difficult it is to recruit volunteers from among “normal” people who will become controls for AD research:

We were inundated with phone calls when we put out the call for research subjects, but these came from offspring of individuals who had Alzheimer’s. We had a much harder time with people who don’t have Alzheimer’s in their family. Why would they come in? So we tried to get people from Alzheimer families to bring their spouse in as a control, or a friend. We used the families to help recruit the controls. Now we have the largest group anywhere of cognitively normal people who have undergone PIB studies. (May 2009)

A 2010 article in the journal Alzheimer’s & Dementia explicitly states that the ADNI project is not built around, nor does it depend upon, the amyloid hypothesis, despite evidence “in favor” of this hypothesis. It is also made clear that to date there is “limited information concerning the pathophysiological sequence of events of AD in human beings from autopsy studies and studies measuring only cognition.”53 The overarching goals of ADNI are to “determine the relationships among the clinical, cognitive, imaging, genetic, and biochemical biomarker characteristics of the entire spectrum of AD as the pathology evolves from normal aging through very mild symptoms, to MCI, to dementia, and to establish standardized methods for imaging/biomarker collection and analysis for ultimate use in clinical trials.”54 Of the 58 involved sites, 53 are in the United States and 5 are in Canada. Globally, sites in Europe, Australia, and Japan are affiliated with the project, but have their own neuroimaging initiatives, and Chinese and Korean sites are about to be added. The number of subjects involved, over 1,000, is large for neuroimaging studies.

The limitations of ADNI have been clearly set out. First and foremost, the population studied is not an epidemiological sample, but represents what is termed “a clinical trial population” in which individuals who have had a cortical stroke, cancer, heart failure, substance abuse, and so on are excluded. Second, the age range of subjects is from 55 to 90 years, and hence, given what is now known about amyloid deposition in the brain, the lower age limit should, ideally, be younger by a decade. A third limitation is that no lifestyle information is being collected from subjects, and a fourth limitation is that not all of the measurements taken thus far have been made on all subjects, limiting the possibilities for comparison. The principal objective is that of standardization to provide “an infrastructure for world-wide clinical trials by the pharmaceutical industry.”55

To date, amyloid imaging showed 89% PIB positivity in patients diagnosed with AD in the ADNI study. Other PIB studies have shown that 10% to 20% of AD patients are PIB-negative—a finding that is not yet explained, and may or may not be due to a doubtful clinical diagnosis. Among the “normals,” a very large number—47%—tested PIB-positive. This number is considerably higher than in other studies to date, but is not regarded as an anomaly because it correlates well with the findings of CSF β-amyloid measures in the same subjects. The plan is to follow these individuals over the coming years to determine whether they begin to evince changes in cognitive function.

Gina Kolata reported on ADNI in an August 2010 article in the New York Times. When asked by Kolata to comment on the ADNI approach to sharing of data, John Trojanowski, a specialist in the pathobiology of neurodegenerative disorders with a prodigious publication list, replied, “It’s not science the way most of us have practiced it in our careers. But we all realized that we would never get biomarkers unless all of us parked our egos and intellectual property noses outside the door and agreed that all of our data would be public immediately.”56 Kolata reported that companies as well as academic researchers are using the data, and that there have been more than 3,200 downloads of the entire massive data set and almost a million downloads of the data sets showing brain scans. This consortium, between January 2010 and September 2011, published 135 articles.

ADNI is a work in progress—a prescient example of the biosciences of the future organized to defeat AD—a complex amoeba-like public–private consortium designed principally to bring about success with drug development for this baffling condition. Readers of the 2010 New York Times article by Kolata learn that in the interests of cost containment the entanglement of aging, lifestyles, sociopolitical conditions, and dementia have been set to one side. It was also pointed out that biomarker measures “have not yet been demonstrated to have high value as either predictors or outcomes,” and, furthermore, the “subject burden” in the form of repeated tests is high, and additional tests will impair enrollment and increased dropout rates.57

Brain Identity

The sociologist Nikolas Rose has hypothesized that the very idea of what constitutes “self” is being transformed in light of developments in molecular neurobiology. He has documented evidence for what he describes as “neurochemical selves” in which “[m]ind is simply what the brain does. And mental pathology is simply the behavioral consequence of an identifiable, and potentially correctable, error or anomaly in some of those elements now identified as aspects of that organic brain.”58 With respect to Alzheimer nosology, mind effectively disappeared from view from the latter part of the 19th century, displaced by a powerful discourse about localized neuropathology, although in certain eras efforts were made over the years to challenge or dislodge this dominant discourse. A focus on neuropathology permitted both the medicalization and gradual destigmatization of dementia and, at the same time, bolstered the common belief that once individuals become demented they are no longer their former “selves” and that a radical, irreversible rupture has taken place. Initially, early findings about the genetics of AD reinforced this situation, as we will see in the next chapter, with the result that a fundamental materiality has pervaded Alzheimer discourse over the past two decades. This position is consolidated further because cognitive functioning is today comprehended as the direct product of neurological activity, although a good number of experts believe that this situation will hold only until such time as the neural networks and pathways that activate consciousness and individual awareness are spelled out.

The findings of Simon Cohn, a medical anthropologist who interviewed psychiatric patients who had undergone brain scans and were given copies of the scans to take home, alert us to what may be in store with increasing routinization of neuroimaging. The patients in this study told Cohn in no uncertain terms that a scan is, for example, a “picture of who you really are. On the inside. I tell people it is my self-portrait.” Further investigation by Cohn made it clear that the brain scan in effect verifies the diagnosis for patients, thus making them into living cases of the label they have been given.59 It is not yet known if a similar transformation is likely to take place when people undergo repeated PIB scanning, individuals who, on the basis of their selection for trials, will already be confronting the possibility of dementia in their lives. It appears from the literature and the interviews I have conducted that perhaps the majority of neurologists believe that evidence of amyloid in the brains of healthy individuals is a definitive sign of incipient AD. They take this position despite repeated findings that very large numbers of old people with amyloid deposits are cognitively intact when they die, but simply assume that patients would have died of AD if they had not previously been picked off by some other condition. It is quite possible that neurologists and healthy research subjects together may well interpret findings of a positive PIB scan as a definitive sign of approaching AD with, no doubt, significant effects on trial subjects and their families.

A neurologist who comes from a family where AD is present told me that if he found amyloid in his own brain he would be very worried indeed (although he has no intention of checking), although finding out about his genotype would not concern him much, whatever the result. The question arises as to whether or not people who are informed about amyloid deposits in their brains will have the fortitude to situate the risk estimates handed out to them in the broader social context of their life experiences. Or will they, in effect, be transformed into Alzheimer patients years before symptoms are detectable, if indeed symptoms ever do manifest themselves—living out a slow, unrelenting death that in reality may not be taking place at all? Alternatively, will some of these research subjects, whatever their own test results, be able to sustain a conviction that the trial may eventually be successful, in which case they will have contributed to research that has potential benefit to millions, should an effective drug be developed? Such individuals will be stoic corporeal citizens—those who choose to make their bodies available for the betterment of society at large. Given the enormity of what is involved, it will be important to watch carefully for dropout rates from drug trials that carry out repeated biomarker testing. And presumably some subjects monitored over the months and years in connection with incipient Alzheimer’s will be vulnerable to serious depression.

Pasting Over Incommensurabilities

It is just possible that AD research is at the beginning of the long road to a paradigm shift, because acknowledgment of stubborn anomalies is often a first step to a major change of orientation. The work of Julie Schneider and Carol Brayne, together with their respective colleagues, among other recent research findings, highlights incommensurabilities that can no longer be brushed to one side.60 And as early as 1997 Zaven Khachaturian noted,

The life of a neuron hangs on many interdependent systems, particularly the systems for neural intercommunication, metabolism and repair. Those three systems ordinarily work in sync, like highly trained acrobats. But internal factors such as changes in a person’s nutrition, immune response or neuroendocrine status, and external factors such as toxins, trauma or infection can shake one of the systems and upset the entire delicate balance.61

Despite commentary such as this for more than a decade, a dramatic transformation in the order of a conceptual revolution is not in sight. On the contrary, the primary hope associated with the application of technologies that allow assessment of biomarkers is, it seems, that the findings will confirm and consolidate the causal model centered on the amyloid cascade hypothesis. This model is the one that continues to attract drug companies as they plan the direction of future research. And, as William Thies notes,

People have the idea that the biomarker puzzle and the therapeutic puzzle are sort of independent. But they’re not. They go hand in hand. They’ll be solved at about the same time because one really needs the other to really come to closure … ultimately therapies will be the key tools because they’ll give us the ability to then confirm our biomarker or imaging markers of particular subsets. At the moment you can hypothesize all you want about a certain biomarker indicating a certain kind of change but until you can change that biomarker [by “treating” it] and see the disease change you really don’t have the ability to confirm what biomarkers are doing. (April 2010)

Strategic use of the concept “paradigm shift” is very frequent in the AD world today. This seductive hype, above all else, is a means of raising money and of sustaining hope among the public for a cure for AD and, ultimately, for its prevention. Formerly researchers often made claims that AD would be defeated before too long. One researcher broadcast in 2001 that “within 10 years drugs will be developed to prevent Alzheimer’s.”62 In recent years, although such hype has continued to appear in the media at times, it is striking among researchers that this kind of talk has become relatively unusual, and caution and even skepticism have been quite often evident as people confront the slew of unknowns that riddle the field.

The next chapter introduces the complexities associated with the genetics of Alzheimer disease and discusses genotyping as yet another type of biomarker. In contrast to the biomarkers discussed above, in a small number of families worldwide, genetic knowledge can predict AD occurrence with a high degree of certainty, but, for by far the majority of aging individuals, genotyping for AD provides no definitive answers.

Appendix 1

Summary of clinical and cognitive evaluation for MCI due to AD


Establish clinical and cognitive criteria

Cognitive concern reflecting a change in cognition reported by patient or informant or clinician (i.e., historical or observed evidence of decline over time)

Objective evidence of impairment in one or more cognitive domains, typically including memory (i.e., formal or bedside testing to establish level of cognitive function in multiple domains)

Preservation of independence in functional abilities

Not demented

Examine etiology of MCI consistent with AD pathophysiological process

Rule out vascular, traumatic, medical causes of cognitive decline, where possible

Provide evidence of longitudinal decline in cognition, when feasible

Report history consistent with AD genetic factors, where relevant


Abbreviations: AD, Alzheimer’s disease; MCI, mild cognitive impairment.

Reprinted from Alzheimer’s & Dementia: 7, no. 3, Albert M.S. et al., “The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging and Alzheimer’s Association workgroup”, 270-279, © 2011, The Alzheimer’s Association, with permission from Elsevier.

Appendix 2

Biomarkers under examination for AD


Biomarkers of Aβ dispositionCSFAβ42

PET amyloid imaging

Biomarkers of neuronal injury

CSF tau/phosphorylated-tau

Hippocampal volume or medial temporal atrophy by volumetric measures or visual rating

Rate of brain atrophy

FDG-PET imaging

SPECT perfusion imaging

Less well-validated biomarkers: fMRI activation studies, resting BOLD functional connectivity, MRI perfusion, MR spectroscopy, diffusion tensor imaging, voxel-based and multivariate measures

Associated biochemical change

Inflammatory biomarkers (cytokines)

Oxidative stress (isoprostanes)

Other markers of synaptic damage and neurodegeneration such as cell death


Abbreviations: Aβ, beta-amyloid protein; AD, Alzheimer’s disease; BOLD, blood oxygen level-dependent; CSF, cerebrospinal fluid; FDG, fluorodeoxyglucose; fMRI, functional magnetic resonance imaging; MR, magnetic resonance; MRI, magnetic resonance imaging; PET, positron emission tomography; SPECT, single photon emission tomography.

Reprinted from Alzheimer’s & Dementia: 7, no. 3, Albert M.S. et al., “The diagnosis of mild cognitive impairment due to Alzheimer’s disease: Recommendations from the National Institute on Aging and Alzheimer’s Association workgroup”, 270-279, © 2011, The Alzheimer’s Association, with permission from Elsevier.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
34.229.173.107