Volume 7  Number 5  August 19, 2010
Second Opinions

Predicting Alzheimer's: Promises and Perils

Just because your doctor has a name for your condition doesn't mean he knows what it is.

Six Principles for Patients
Murphy's Law, Book 2

Everything is so dangerous that nothing is really very frightening.

Gertrude Stein

In medicine the term biomarker can describe any parameter or measurement, from blood pressure to proteins that correlate with disease presence, progress or susceptibility to treatment. Note the use of correlates. Correlation or association is not causation, a common misapprehension in the scientifically untrained. For example, body temperature is a well-known biomarker for fever. Fever is a "biomarker" for infection, inflammation, cancer, and a host of other conditions. Blood pressure is used to determine the risk of stroke, but is certainly not the ultimate cause of stroke, any more than high cholesterol, a risk factor for coronary disease is its cause. The P53 gene is a marker or predictor or a correlate of some cancers, but not their cause. (There are multiple correlations for just about any condition-in an endless regress, the bęte noir of all biologic research).

Sensationalizing an Old Biomarker

Originally reported in 2003 protein compounds called tau and beta-amyloid found in spinal fluid correlated with the presence of Alzheimers and other dementias. Clumps of these deformed proteins, commonly called plaques can now be detected by PET scanning. But the presence of these plaques has been the diagnostic standard of Alzheimer's pathology when Dr. Alois Alzheimer first identified them in a patient's brain in 1906. As misreported in one article "Researchers report that a spinal fluid test can be 100 percent accurate in identifying patients with significant memory loss who are on their way to developing Alzheimer's disease." The article goes on to say, "Although there has been increasing evidence of the value of this and other tests in finding signs of Alzheimer's, the study...shows how accurate they can be." (Actually, 90% of the Alzheimer's were positive and 39% of the control group.) In typical reportorial exaggeration, "The new result is one of a number of remarkable recent findings about Alzheimer's."

Since medical experts now agree that Alzheimer's starts a decade or more before people have symptoms, by which time it may be too late to save the brain, wouldn't it be a triumph if we could identify those patients at risk for developing the condition? As the news story unfolds, then we could "use those people as subjects in studies to see how long it takes for symptoms to occur and in studies of drugs that may slow or stop the disease." How much longer would the drug studies take, how would their success be measured other than possibly in brain pathology measured at post mortem?

A Second Take on the News

Because of the reports discussed above and the Archives of Neurology article, a panel of medical experts from the NIH National Institute on Aging and the Alzheimer's Association promptly proposed changes in the way doctors diagnose Alzeheimer's. to include the use of tests like PET brain scans to detect plaques, and analyses of spinal fluid tau and beta-amyloid. Serious misgivings were expressed in the Times the following day by Sanjay W. Pimplikar an associate professor in the department of neurosciences at the Cleveland Clinic.

Pimlikar points out that "roughly one-third of all elderly adults have such plaques in their brains yet function normally. And eleven clinical trials, recently made public by a group of drug companies, that were aimed at reducing these plaques in Alzheimer's patients all failed to show cognitive improvement, even when the brains were cleared of plaques. Thus, the presence of plaques cannot predict with any accuracy or specificity that an individual is going to acquire the disease." (My italics.) Moreover, the predictive potential of spinal fluid analysis of the levels of tau and amyloid-beta proteins "remains uncertain." This is obvious simply because it will take a decade or more and hundreds if not thousands of normal and suspect Alzheimer's patients-not to speak of tens of thousands of spinal taps to settle the issue by noting which patients finally develop the full-blown disease. It should be noted in this connection that questions will erupt over which patients with dementia in fact have –or had-Alzheimers, since other causes of dementia comprise up to 30% or more of all cases of the condition.

Moreover, spinal taps or lumbar punctures ("LP") as a screening test are not to be taken lightly, not only because of the expense and technical experience required, but because of pain, and potential misadventures, including but not limited to infection (meningitis.) Dr. Janis Petzel, a geriatric psychiatrist in Maine, observed how unfeasible this test is in nonacademic, rural or non-Western settings. Hopefully, a blood test could be developed.

As Pimlikar adds, there is also the psychological risk of false positives and misdiagnoses that greatly distress patients, at least until further tests can prove they do not have the disease. Which tests, indeed? There is no present reliable and specific test for Alzheimer's, except brain pathology (biopsy anyone?) unless we spend more millions or even billions, on the proposed research programs, wait for 10-30 years, and enough people are willing to wait in line and pay perhaps $1,500 for their lumbar punctures.

What the Future Holds

Would you like to know if you might get Alzheimers in 10 to 20 years, and if you start taking these new pills now you might forestall or prevent the disease? Certainly some people would elect to have the foreknowledge of a life ending in dementia and helplessness. My hunch, however, is that the number would be exceedingly small. A neurologist friend reminds us of the furor that initially erupted in the case of the 100% accurate gene-based prediction of Huntington's disease, a rare and ultimately fatal neurologic condition with dementia. This anxiety, widely covered in the media, ultimately subsided as offspring of Huntington parents with a 50% chance of developing the condition, chose privately whether to know or not to know their fate in advance. But is this really comparable to the uncertainty and fear enveloping the present arguments over predicting Alzheimer's via spinal tap? I have serious concerns about the wisdom, if not the feasibility and uncertainty of screening for a disease that may not appear for decades, although no one would question the need for research into the causes and possible prevention of dementias.

As the recent history of overhyped screening shows, using the PSA test as a screening test for prostate cancer has an alarming down side. A study last year found that in the two decades after the test was introduced, prostate cancer was detected in more than 1 million additional men, 95% or more of whom were likely overtreated. See the New England Journal articles reported.

The proposed research studies to be undertaken cooperatively at dozens of institutions, initially projected to cost hundreds of millions, leave no question that the drug companies would ultimately profit either way. Take Aricept with $3 billion in worldwide sales. Examine the documents supporting the FDA's approval of Aricept, and you will see "upon what a slim reed this drug's empire was built." In one study those taking the drug scored on average, three points better on a 70-item cognitive assessment scale. That's about a 4% difference.

Doctors and now economists and others like to quote the Latin "primum non nocere"-first, do no harm. How often do we pay lip service to that maxim while failing to heed its dire warning?


24 hours after finishing the above newsletter I read "Eli Lilly halted two late-stage clinical trials of an experimental Alzheimer's treatment on Tuesday (Aug.17, 2010), representing a setback to one leading theory on treating the degenerative disease and a new blow to Lilly's business prospects.

The company said patients who had taken the drug, intended to reduce plaque in the brain, actually showed worse cognitive functioning and less ability to perform daily living tasks than patients who had taken a placebo."

Stay tuned.

Martin F. Sturman, MD, FACP

Copyright 2010, Mathemedics, Inc.

EasyDiagnosis is an automated online service that analyses existing medical symptoms and predicts likely causes and conditions. Click here to find out more about this unique service and click here to try one of our modules for free.