What Causes Alzheimer’s?: An Exchange

To the Editors:

I have long been an admirer of the contributions Jerome Groopman makes to The New York Review, but when commenting on my book The Alzheimer Conundrum: Entanglements of Dementia and Aging [NYR, May 22] he came up short. Nothing in my book indicates that there is an “easy way” to prevent Alzheimer’s disease (AD), as Groopman suggests I am arguing. He recommends that we learn from the recent “stunning success” obtained in combating with drugs the viruses known respectively to cause HIV/AIDS and liver failure resulting from infection with hepatitis C.

But Groopman’s argument is flawed. The causal agents of HIV/AIDS and hepatitis C are known. The prime candidates for proximate AD causation are proteins—amyloid precursor protein and tau—inherent to normal body functioning. Despite years of research it is still not known under what circumstances these proteins become altered in humans and hence toxic or malfunctioning, although aging is undeniably implicated. It is plausible that their toxicity is the result of other as yet poorly understood bodily perturbations. Furthermore, there are a growing number of AD researchers who doubt that these particular proteins are the major initiating cause of the disease at all.

In addition, many researchers today argue that “pure” AD is a researcher-produced artifact because the numerous potential research subjects with comorbidities are deemed ineligible to participate in trials. Groopman assumes that recent findings about a protein termed REST that appears to be absent in individuals with AD may well be the key (easy way?) to determine who among us will retain neural integrity and who will degenerate with age. But among the critics of this study published in Nature, one points out that we do not know if this protein is a key driver factor or just one of many factors that contribute to the degenerative process.

My purpose in writing The Alzheimer Conundrum was to make clear that the Alzheimer field today is replete with discussions about uncertainties, unknowns, failed trials, and the shortcomings of animal models. Further basic science research is clearly needed, probably driven best by new hypotheses—I did not dispute this. I also cited epidemiological research that showed that the brains of individuals over seventy-five are remarkably similar, whether or not these people were demented in life. In addition I posited the need for a largely neglected public health approach to AD prevention, one that would entail a reduction in inequalities and toxic environments, better social support, improved education, and so on, all associated with increased risk for AD. Such an approach is the only realistic way in which AD prevention can be implemented globally. Jerome Groopman has given readers a greatly impoverished impression of what my book is about.

Margaret Lock
Marjorie Bronfman Professor Emeritus in Social Studies in Medicine
McGill University
Montreal, Canada

Jerome Groopman replies:

Professor Lock is a partisan in the debate over the causes and potential remedies of Alzheimer’s disease. She contends that researchers are myopic when they focus on misshapen proteins as the primary cause of the dementia. I noted in my review that such scientists carry monikers: those who favor the beta-amyloid protein as a primary cause are termed “baptists,” while those who believe it is the tau protein are called “tauists.” Partisans who share Lock’s view might be termed “environmentalists,” certain that ambient toxins and other factors outside the brain are key in the development of the disorder.

I am not a partisan. Professor Lock must know this, because she states that she is a regular reader of my work; furthermore, footnote reference 6 in my review is a five-thousand-word essay (“Before Night Falls: Alzheimer’s Researchers Seek a New Approach”) I wrote on the spectrum of views, including those of her camp. I concluded that it is premature to be a partisan in the face of so much conflicting evidence.

Readers may recall that in my review of her book, I first highlighted its merits. The strongest challenge to the partisans of misshapen proteins is the finding of widespread deposition of beta-amyloid and tau in the brain of some elderly individuals whose cognition is intact. This point was the lead in my assessment. (Lock reiterates it in her letter.) I also found her pages on epigenetics, how genes may be modified by environmental and other factors, a valuable contribution. But my appreciation of these aspects of her work appears not to be satisfactory, because I neither rejected Prusiner and others who pursue misshapen proteins nor endorsed her proposed interventions to prevent Alzheimer’s disease.

Lock’s letter reveals the pitfalls in thinking that diminish the value of her book. Specifically, she mistakes correlation for causation, and advocates paternalistic interventions in the absence of compelling data. Lock lists poverty, educational deficiencies, other inequalities, and fast food as factors that correlate with an elevated risk for dementia. None of these is shown to be causative. Junk food is not linked to Alzheimer’s disease as cigarettes are to lung cancer. Nonetheless, she asserts with great certainty that the “only realistic way” to combat Alzheimer’s disease is by alleviating these social, political, and cultural ills.

While all of us can agree that such deficiencies affect our world, there is no scientific evidence that alleviating them will limit the burgeoning global number of cases of dementia. Public policy should not be grounded on the cardinal error of mistaking correlation for causation. Lock also is annoyed about my use of the term “easy.” Solutions to problems like poverty and poor education are complex. The one measure easy to implement through legislation is limiting access to fast foods.

Now, on to Lock’s point about the REST protein. Contrary to her statement, I did not “assume” its ultimate importance in the genesis of Alzheimer’s disease. Rather, I purposefully called it a “correlate” and noted that the researchers “proposed” a role for it in protecting neural integrity. It is a hypothesis. Lock rewrites my prose, changing what I call a correlation into an assumed cause, a hypothesis into a conclusion. Her reframing of my words is indicative of a desire to prove me a partisan. Caveats that the finding was a correlation were included in Nature and in the lay reporting of the discovery.

Lock further takes me to task when I draw lessons from the AIDS epidemic. Here, she missed my point because of hindsight bias. As a researcher and clinician who saw the first patients with that disease in California in the early 1980s, I lived the history of identifying its cause and developing its treatments. Before HIV was discovered, numerous epidemiological studies highlighted environmental toxins, particularly party drugs like amyl nitrate (“poppers”) and exposure to foreign antigens present in semen. All of these correlated with risk for developing AIDS among gay men. Intravenous drug users were similarly believed to suffer an immune collapse from injected toxins, based on epidemiological data. None of these correlations proved causative or pointed in the direction of a new virus.

Rather, what proved quite informative in unraveling the etiology of AIDS was its uniquely high incidence among those with hemophilia. Hemophiliacs were relatively homogeneous as a group, receiving transfusions of the blood product factor VIII concentrate, and typically not exposed to illicit drugs, semen, and other putative environmental toxins. A transmissible agent persisted in the factor VIII preparations, and that strongly suggested a microbe like a virus.

So why is the AIDS analogy so apt? In the absence of a proven etiology, it showcases the temptation to transmute correlations into scenarios of causation. Further, as AIDS researchers were importantly informed through their study of hemophiliacs, researchers in Alzheimer’s disease may be most informed by studies of populations where social and cultural factors are more homogeneous. Currently, the most substantial support for the misshapen protein hypothesis comes from its occurrence among young men and women in Colombia who carry a genetic mutation related to beta-amyloid. My analogy also extends to the social, economic, and cultural settings in which AIDS is amplified, specifically poverty, educational gaps, and other “inequalities,” all of which Lock enumerates in Alzheimer’s disease. But the most dramatic change in the trajectory of the AIDS epidemic has come from targeted intervention against its root cause, HIV.

I’m glad that Professor Lock has written to summarize her thinking. Readers can now judge who falls short in their objectivity about the current state of understanding of Alzheimer’s disease, who transmutes correlation into causation, and whether she has determined the “only realistic way” to address this devastating disorder.