How not to do “personalized medicine” to treat Alzheimer’s disease

With the aging of the population, one of the most feared potential manners by which more and more of us will leave this earth is through Alzheimer’s disease or other forms of dementia. And it is a scary thing, too. Having valued my intelligence all my life and in particular enjoying the intellectual stimulation that I derive from my job, not to mention from blogging and contemplating science outside my realm of expertise, like many people I fear Alzheimer’s disease at least as much as cancer or heart disease, possibly more. Imagining the slow decline in my faculties to the point where I can no longer take care of myself and no longer even recognize my loved ones terrifies me, as it does virtually anyone. Fortunately, we’re coming to understand more and more about neuroscience and how the brain works. So I at least have hope that, 20 or 30 years hence, when I’m in those dementia-prone years, there will be something medical science can do for me if I’m unfortunate enough to develop Alzheimers, in marked contrast to the situation now, in which we have nothing that can do much good for dementia.

Unfortunately, it is that very dearth of effective treatments for dementia that leads to quackery, and I’ve found a doozy of a pseudoscience-laden post at–where else?–that wretched hive of scum and quackery, The Huffington Post. It’s a post by someone whom we’ve met before, namely Dr. Mark Hyman, he of “Ultrawellness” and “functional medicine,” and it’s entitled 9 Steps to Reverse Dementia and Memory Loss As You Age. Not only doth the woo flow, but it does so with video:


Seldom have I found such a willful misapplication of the new science of genomics than here. Well, not quite. The way the Age of Autism misinterpreted the latest autism SNP study is up there, too, but that may have to wait to be a topic for another post. In the meantime, let’s see what our favorite woo-meister Dr. Hyman says:

There was a woman with mild cognitive impairment on the panel. Her condition is sort of like pre-Alzheimer’s disease. Everyone on the panel — including the Harvard neurologist — agreed that memory loss is NOT a normal part of aging. The sad part was that the panel didn’t have much to offer people in the way of prevention. Their only solution was just a very bad and pretty ineffective selection of drugs with lots of side effects.

It never ceases to amaze me how docs like Hyman can start out with a seemingly reasonable statement, namely that the drugs we have for the prevention and treatment of Alzheimer’s disease pretty much suck. They don’t work very well, and they’re chock full of side effects. Unfortunately, what separates science-based physicians from the woo are statements like the one Hyman makes next:

But there is another way to think about brain aging. The brain responds to all the same insults as the rest of the body — stress, poor diet, toxins, lack of exercise or sleep, nutritional deficiencies, and more. All we have to do is give the brain a tune-up and we can see miracles.

I have a rule of thumb. True, it’s just a rule of thumb, with no empiric evidence to back it up. I admit that. However, it does seem to hold true form an anecdotal perpective. That rule is: Whenever you see a health care professional promise “miracles,” run, don’t walk. Add to that references to “toxins” and “nutritional deficiencies,” and run harder and faster. Seriously, you’re about to be buried by woo or pseudoscience, and Hyman doesn’t disappoint:

Sometimes the practice of medicine lags behind the science, and sometimes the practice gets ahead of the science. Genetic testing puts us squarely in the middle of that dilemma. We are at a crossroads, where the old ideas we have about disease and diagnosis become less meaningful as we understand more and more about the importance of individual differences in determining illness. This a time when personalized medicine will replace medicine based on diagnosis and disease.

In fact, disease and diagnosis as we know it will soon be an obsolete concept, an artifact of medical history, like bloodletting or phrenology (the art of diagnosis based on the shape of your skull, popular in the 19th century). The reason is simply this: Naming a disease does nothing to help us identify and treat the underlying causes of the disease. We must address these causes if we have any hope of helping individuals heal.

Notice how Hyman paints current medicine as being obsolete, like bloodletting or phrenology. Truly, it’s hard to imagine more of a case of projection, given that his “functional medicine” is chock full of various forms of pseudoscience and woo. He also invokes the alt-med delusion that it, not “conventional” medicine, finds and adresses the “true causes” of disease. How I hate that lie! Does homeopathy address the “true cause” of disease? No. It bases its “provings” on symptoms. Does acupuncture address the “true cause” of disease? Of course not. It ascribes to all disease the cause of “blockages” in the flow of an imaginary “life force.” The list goes on. It’s really scientific medicine that seeks to find the cause of each disease and direct its therapies at that cause. Alt-med, on the other hand, often ascribes One True Cause to all disease.

He also misunderstands the nature of genetic diagnosis and genetic testing. I’ll tell you what I mean using the example of breast cancer. If there’s one thing we’ve learned analyzing the genetic makeup of breast cancer, it’s that breast cancer is several diseases, each with different, albeit overlapping, characteristics and behavior. Although we once divided (and still do divide) breat cancer into the broad categories of estrogen receptor-negative and estrogen receptor-positive cancers. Now, beginning with expression array profiling experiments published around the turn of the millennium, we now divide breast cancer into phenotypes known as luminal versus basal, all entirely based on gene expression profiles, not on the usual traditional characteristics that we used to use, in particular histology. These new molecular-based phenotypes have deepened our understanding of the disease known as breast cancer and allowed us to subdivide it into types based more on biology. True, we have a long way to go before this information is fully incorporated into how we diagnose and treat breast cancer, but you know what? Some of those subdivisions actually corresponded to subdivisions that we had already discovered on the basis of other characteristics.

Actually, what “personalized medicine” means to someone like Hyman is not a deeper understanding of disease that allows him to choose whatever woo he was going to choose anyway, in other words, as an excuse to make it up as he goes along. He then demonstrates this to be the case by trying to apply genetic information to a patient of his:

Even though no long-term studies have been done to look at treating dementia based on genes, there are so many scientific threads that weave together a picture of how and why our brains age and what genes are involved. This leads me back to George …

For this man, whose mind and life were evaporating, I looked deeply into his genes and the biochemistry his genes controlled and found places where we could improve things.

“I looked deeply into his genes and the biochemistry his genes controlled”? That line cracked me up, as if Hyman knew the first thing about “looking deeply” into anything, much less someone’s genes. In fact, the only thing Hyman’s good at “looking deeply” into is the woo. In fact, he “looks deeply” into the genes and biochemistry in the same way that Mark and David Geier or Andrew Wakefield “look into biochemistry.” In fact, what Hyman writes next would not be out of place on an “autism biomed” discussion board:

He had a gene called apo E4, which is a high-risk gene for Alzheimer’s disease(ii) and also made it hard for him to lower his cholesterol and detoxify mercury from his brain.(iii) He also had a version of a gene for detoxification of metals and other toxins (glutathione-S-transferase, or GST)(iv) that was very inefficient, making him accumulate more toxins over his lifetime. Having the combination of a problem with GST and apo E4 puts people at even more risk for dementia.(v),(vi) In another study, people with an absent GST gene were likely to have much higher levels of mercury.(vii)

George had another gene called MTHFR(viii) that made him require very high doses of folate to lower his blood levels of homocysteine, which is a substance very toxic to the brain. Lastly, he had a gene called CETP that caused his cholesterol to be high, which contributes to dementia. Combine this gene with the apo E4 gene and your risk of dementia goes way up.(ix)

We found that George had high levels of mercury(x) and helped him detoxify with foods such as kale, watercress, and cilantro, herbs such as milk thistle, nutrients such as selenium and zinc, and medications that helped him overcome his genetic difficulties by getting rid of toxins.

We lowered his cholesterol with diet and herbs. We lowered his homocysteine with high doses of folate and vitamins B6 and B12.

Yep, it’s all there, the obsession with mercury, glutathione, and “toxins.” In fact, if you changed the word “Alzheimer’s disease” to “autism” and left out the parts about decreasing cholesterol with diet and herbs (even I haven’t seen that on autism “biomed” boards), and Hyman’s post would be right at home on Age of Autism! I doubt any of the denizens would even notice the difference.

Of course, Dr. Hyman reports that his patient recovered spectacularly, providing the usual N=1 anecdote that is meaningless when you remember concepts such as regression to the mean and the placebo effect. he even has another anecdote about an elderly woman named Christine:

Her neurologist offered her words of comfort, but told her and her family there is no treatment truly effective to stop or reverse the progression of dementia. That’s when her daughter brought her to see me.

We discovered many subtle changes in her health that on their own wouldn’t explain dementia, but when added all together put a strain on her brain function. All we did was correct those problems — low thyroid function, mercury toxicity, inflammation, and deficiencies in vitamins B6 and D, folate, coenzyme Q10, and omega-3 fats — and improved her diet overall. I encouraged her to exercise, because exercise can help improve cognitive function and prevent dementia.

Six months later, she had the extensive memory tests repeated. Her psychologist was surprised to report that her scores got BETTER!

Note that he didn’t say how much better her report got. Was it dramatically better or just a little better? He also didn’t say if she kept getting better or if her condition stabilized. After all, dementia isn’t always a straight line decline. There are bumps along the way, times when it may slow down or even times of transient improvement. It’s quite possible that exercise and diet may have slowed Christine’s mental decline transiently. The question is: Is this just a fluke?

There’s another aspect of this article that’s interesting. One complaint I have about news articles is that they often don’t provide citations or even links to studies discussed. Dr. Hyman, to his credit, does. Unfortunately, a perusal of the reference failed to show them supporting the therapies that Hyman described. For instance, his fixation on apo E4 was overblown, as were his invocations of papers on glutathione. They’re articles that show that these may be biomarkers, but there’s one thing you have to understand about biomarkers. They only suggest the presence of disease or risk for disease. Depending on how tight the correlation is, they may or may not be reliable. More importantly, they may or may not be useful molecular targets for treatment of disease. Hyman assumes they are based on no strong evidence from interventional studies that they are. Like autism biomed “practitioners,” Hyman is extrapolating from correlational evidence that may or may not represent causation that is targetable with specific therapies.

Personal medicine, although it may well be oversold, does hold considerable promise for helping us tailor therapy to the specific characteristics of a patient’s disease, genes, and biochemistry. We’re nowhere near there yet; a lot of work remains to be done. Unfortunately, that doesn’t stop “practitioners” like Dr. Hyman from making wild leaps of speculation from correlational evidence and absent solid clinical trial evidence to guide their woo-filled therapies and then using anecdotes as evidence to declare their speculative “treatments” to be effective.

Clearly, Dr. Hyman has a different definition of “personalized medicine.”