Hacker News new | comments | ask | show | jobs | submit login
PET scans show many Alzheimer’s patients may not actually have the disease (washingtonpost.com)
158 points by DrScump on Aug 2, 2017 | hide | past | web | favorite | 65 comments



Pet scan: $3000 - $4000 Peanut butter: $3 - $10

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3823377/pdf/nih...


A study from 2014 could not replicate this:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4167392/pdf/nih...


> Methods: Participants with probable AD (N=18), mild cognitive impairment (MCI, N=24), other causes of dementia (OD, N=26) and matched controls (OC, N=26) were tested, with closed eyes, for their ability to detect an odor, one nostril at a time. A container of 14g of peanut butter was opened, held medially at the bottom of a 30 cm ruler, and moved up 1cm at a time during the participants’ exhale. Upon odor detection, the distance between the subject’s nostril and container was measured.

> Results: The mean odor detection distance of AD patients’ left nostril (5.1 cm), and not their right (17.4 cm), was significantly less (F(3,90) = 22.28, p < 0.0001) than the other groups. The mean, standard error, and 95% Confidence Interval of the L R nostril odor detection difference (cm) for AD was −12.4 ±0.5, (−15.0, −9.8); for MCI was −1.9 ±1.2, (−4.2,0.4); for OD was 4.8 ±1.0, (2.6,6.9); and for OC was 0.0 ±1.4 (−2.2,2.1).


Tiny pilot study. Where are odds ratios? Where are the multiple measurements per patient?


Great.

But we'll also have machines that measure out the dosages. And possibly a few controls. Plus we'll have to adjust for folks that have an exceptional sense of smell. And come to think about it, we'll need to make sure we have baseline sensitivity levels at different ages to base the test result on. We'll have to account for folks that have sensitive sense of smells, which could be a large problem for controls. Not only that, but we'll need to check for signs of a stroke and the many other things that can cause one to lose one's sense of smell.

On second thought, I'm much more likely to support the more expensive test and hope they invent a simple blood test as it seems to narrow things down further.


We would also have to take into account simple cases of the flu or a cold.

Yet, as long as the errors are overwhelming at the false positive side, it can be used to reduce the number of the more expensive test.


Peanut butter is even cheaper than that -- 10 people can smell the same peanut butter but they can't fit in the same Pet scanner.


What..I can't... I am speechless.


How is this test methodology affected by phenomena like: https://en.wikipedia.org/wiki/Nasal_cycle


That's just the initial bill. I bet insurance gets this down to no more than a few hundred bucks, same as with MRI.


Historically, a bunch of symptom-defined diseases (e.g. hepatitis, diabetes, etc) have turned out to actually be a set of different diseases with different causes, different treatments but similar symptoms - and understanding that split has been a big milestone in treating them. This now might be the case for Alzheimer's as well.


prob. AD as defined at the moment is a set of "criteria" based on a number of things like time scales, and also some specific cognitive measures. I've always found things to be a little vague and there's probably a whole spectrum of different biological entities that can do these sort of things to you.

You can throw PETs, CSF tests etc at it but I never really have a high level of confidence calling it.

Basically people lose memory and certain cognitive functions...but anything could really do it.

At the moment the hodgepodge way I fudge it is: if it is a clearly progressive syndrome year over year, with an amnestic memory syndrome (vs. poor attentional state/registration), with language/naming involvement, and evidence of visuospatial neglect/apraxias, I tell them it may "likely" be AD. Otherwise I can't say if it is AD vs. vascular dementia.

There is a paper looking at accuracy of how we call things vs. post-mortem, and it is about 50/50 eyeroll. Sad but this is the limitation of science vs. mother nature at this state in time.


It's not much better than searching for the cure for all illness.

Drink enough, sleep enough, eat well, do some sports, meditate. If that doesn't help, good luck with going to the doctors.


> meditate

Any clinically proven benefit of meditation though? Don't throw it together with sleep and exercising...


A ton of benefits. It's a very good habit to pick up. Not a wonder drug though.

>In 2013, researchers at Johns Hopkins identified 47 studies that qualify as well-designed and therefore reliable. Based on these studies, they concluded that there is moderate evidence that meditation reduces anxiety, depression, and pain, but there is no evidence that meditation is more effective than active treatment. [1]

1. https://en.wikipedia.org/wiki/Research_on_meditation


It's been thoroughly proven to be beneficial, in many ways. https://nccih.nih.gov/health/meditation/overview.htm


Sorry, what? From the site: "Some research suggests that practicing meditation may reduce blood pressure, symptoms of irritable bowel syndrome, anxiety and depression, and insomnia. Evidence about its effectiveness for pain and as a smoking-cessation treatment is uncertain."

I see "may" for some things, and "uncertain" for others. Nothing on that page says that it's been "thoroughly proven to be beneficial".


How about this: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4203918/

(mindfulness being a form of meditation)


No, I wouldn't consider that at all definitive as a study, and the authors specifically note the following:

"This study is limited by the lack of a control group or active comparison clinical intervention that would provide a basis for making a stronger inference about how MBSR might modify the behavioral and neural bases of different types of emotion regulation."

So while MBSR may help some people, this tested one specific method and did not include a control group or test against any other treatment methods.

Now, I'm not suggesting that meditation - MBSR or other - can't be helpful for some people (possibly even the majority of people under certain conditions), but it certainly has not been studied enough to show clear benefits compared to other treatment methods.


> Don't throw it together with sleep and exercising...

Actually, if you read recent studies (go through Pubmed), you'll find that it should in fact be lumped with sleep and exercise.


Does that invalidate any studies on treatments for Alzheimer? If it turns out many people who show symptoms don't actually have it, it's no wonder that a treatment is so hard to find.


There seem to be a lot of loose ends in this article: how many of these patients will go on to develop Alzheimer's? On what basis are the current treatments for Alzheimer's considered to be effective, if we did not previously have a test to determine who has developed it? On what basis would the physicians whose patients were enrolled in this study change their care plans, and in what way?

At one point, the article says "A positive test for amyloid does not mean someone has Alzheimer’s... But a negative test definitively means a person does not have it". To me, there seems to be a hint of "these patients' condition does not fit the dictionary definition of Alzheimer's, so clearly it would be wrong to treat them as if they had it", yet the article seems to leave open the door to the possibility that it is the progression of Alzheimer's that is not fully understood. I hope an expert can chime in and set me straight.


Not an expert, but I read Derek Lowe's blog[1] and he tends to report on pretty much all the drug trials in the Alzheimer's space.

The amyloid hypothesis (that the cause of Alzheimers is the buildup of beta amyloid plaques) seems pretty weak at this point. The drugs that remove or reduce amyloid buildup don't seem to stop the progression of the disease. The current best guess seems to be that the amyloid plaques are a symptom of whatever the underlying cause is. And the key takeaway is that we have NO IDEA what the actual cause is. We've got a bunch of guesses. One of them might even be right. Or possibly multiple, there are probably several diseases that cause the same symptoms.

So if we can detect the non-presence of amyloid plaques but detect symptoms that would be good as a way to distinguish between the different diseases. Of course there might be more than two variants...

[1] http://blogs.sciencemag.org/pipeline/archives/category/alzhe...


I'll admit I wasn't aware in the 1980s when the hypothesis took hold, but it never seemed like a good one to me just because amyloids are so thermodynamically favored. There must be constant and diverse measures functioning correctly to avoid their formation.


Huh?


nonbel 469 days ago [-]

>"Of all the myriad way a protien can fold, it happens to find one that induces the same malformation when it interacts with another protein."

It doesn't really "just happen", amyloids consist of peptides folded into beta-sheets and aggregates of these seem to be the most thermodynamically stable structures it is possible for polypeptide chains (regardless of sequence) to form:

"From a wide range of in vitro experiments on peptides and proteins we now know that the formation of amyloid structures is not a rare phenomenon associated with a small number of diseases but rather that it reflects a well-defined structural form of the protein that is an alternative to the native state — a form that may in principle be adopted by many, if not all, polypeptide sequences

[...]

These observations, therefore, have led to the remarkable conclusion that, at the concentrations present in living systems, the native states may not always represent the absolute free energy minima of the corresponding polypeptide chains — the native form of a protein could in some cases simply be a metastable monomeric (or functionally oligomeric) state that is separated from its polymeric amyloid form by high kinetic barriers" http://www.ncbi.nlm.nih.gov/pubmed/24854788

https://news.ycombinator.com/item?id=11535494


It reduces these studies' power to detect results. If you thought you had 100 cases and 100 controls, it turns out perhaps you really had 30 cases and 100 controls, and 70 controls misclassified as cases. Power is going to be substantially hurt.


I could see it having more implications than just measurement error reducing power. It's entirely plausible that a drug which has an effect on Alzheimers would come with side-effects that would damage those who just have senile dementia; imagine a drug which dissolves amyloid plaques while doing some moderate damage, in a trial with a mix of dementia and Alzheimers, the Alzheimers getting better is masked by the dementia ones getting worse, for a net average zero compared to the placebo group.


However, it may significantly increase the size of measured effect if you can go back and re-classify the patients - in your example, it may turn an earlier study where only 30 of 100 cases showed improvement to a study where 30 of 30 cases responded to treatment well.


Those 70 are not controls but each may have a different disease.


Not surprising to me.

A family member just had a stroke... the entire system is designed around maximizing billing. If someone is in a longer term situation and a Alzheimer's diagnosis makes a buck, you'll get that diagnosis.

Never mind the mental condition is probably a result of the institutional environment and drugs.


> the entire system is designed around maximizing billing

Our entire world is built that way, it isn't unique to medicine.


Except for businesses that compete on price.


And every business does everything it can to not have to do that, it works when there's choice and price transparency, markets fail without the right ingredients and healthcare is certainly one of the markets the free market has failed.


This is amazing news. I love the advances we're making as a society.

In that $3-5k cost of running a PET scan, I wonder how much of that is paying professional operators and physicians to interpret the results?

Could ML be applied to this problem to reduce the cost of a PET scan?


The cost isn't just in dollars though. In a PET scan they inject a radioactive dye into your bloodstream. It "exposes you to around the same amount of radiation that you would receive from the general environment over about three years." [1]

The high dollar cost may also be related to newer, more sensitive machines that don't require as much radiation exposure to the patient.

[1] https://www.betterhealth.vic.gov.au/health/conditionsandtrea...


And it's not just any old radioactive source, it's a positron emission!

I believe the way it works is that the injected source emits positrons as it decays (having swiped them from the vacuum), shortly thereafter, these annihilate with the electrons of your body, and the scanner looks for the signature pair of gamma rays thus created, using the relative time of arrival to determine their common origin in 3-space. That reduces the problem of scanning to arranging for the decaying source to accumulate in the place we wish to scan which apparently can be done.

That we scan ourselves by annihilating antimatter with our bodies in targeted ways just seems so awesome to me.


OK you've sold me. Where do I sign up?


why does an IV bag cost $500?

I want professional operators and physicians to look at my details.

What I don't want is dozens of layers of middlemen or ticket clippers (or healthcare driven by profit).


> why does an IV bag cost $500?

It doesn't outside of the US.


Doesn't at Kaiser either. You just get all that stuff under a single line item "Emergency Room visit Level 2: $650". Funny how that happens when the hospital and the insurance are the same company.


Indeed. In the US, it does because manufacturers have gamed the system. Through bundled contracting with hospitals etc, for example.


Hospitals pay a couple dollars for IV saline.

In fact, manufacturers are required to report such prices annually to the federal government, which bases Medicare payments on the average national price plus 6 percent. The limit for one liter of normal saline (a little more than a quart) went to $1.07 this year from 46 cents in 2010, an increase manufacturers linked to the cost of raw materials, fuel and transportation.

http://www.nytimes.com/2013/08/27/health/exploring-salines-s...

(which implies that manufacturers aren't the primary beneficiaries of the price that the hospitals charge)


Thanks. It does seem that, in this case, it's hospitals that are inflating costs. But even so, this is the Medicare reimbursement price. Maybe hospitals are "paying" cash price for saline, with a rebate on some product bundle, which doesn't get allocated properly to their cost basis for saline.


It doesn't but what you are charged for a medical service/medical supply has very, very little to do with what your service/supply actually costs.


It is around $2 USD here.

"Hartmanns IV Solution AU$2.73 ex GST

Baxter 0.18% SODIUM CHLORIDE & 4% GLUCOSE IV SOLN 500ML AU$3.72 ex GST " https://www.medshop.com.au/consumables/intravenous-solutions...


Because you're buying a lot of insurance and debt.

The lack of quality health insurance means that a substantial number of people finance healthcare via bankruptcy. That's all priced in. As an insured person or private payer, you pay dramatically more as a result -- it's a hidden tax.


Apparently they cost around $300 in India.


Yes eventually


I suspect that AD diagnoses are driven largely by caregiver frustration with unhappy old people. Just as ADHD etc diagnoses are arguably driven by caregiver frustration with unhappy young people. The simplest solution is drugging unhappy people to keep them manageable.


Spoken with the privilege of someone having never faced a diagnosis or had a close one diagnosed. One of the best things that happened to me was an adult diagnosis of ADD. Drugs really helped. Knowing what ailed me was a greater relief thou.


Not at all. My bipolar diagnosis was great for me, because it got my depressive episodes treated properly.

But an important distinction is whether medication is a choice, or is imposed by caregivers with legal authority, psychological influence, or whatever. I mean, I chose to take SSRIs for many years. And it made me crazy. Which I eventually realized.

But a ten year old kid who gets put on SSRIs inappropriately? They are screwed, and perhaps permanently.


I was put on ADHD meds as a 4 (almost 5) year old after my first violin lesson, this was back in the mid-90s. My mom said it was the weirdest thing she'd ever seen: I so badly wanted to sit still and listen because I wanted to learn to play the violin but I literally physically could not do it and became visibly frustrated with myself. And a bad change in meds in 1st grade (because I didn't like the bitter tablets) was enough for my 1st grade teacher to contact my mom about my change in behavior. It was only as an adult that I realized this must have been the few days where I got in trouble and had to sit out during recess, which was humiliating to me. I have anecdotes for days about this kind of stuff.

I am eternally grateful my parents took me to a psychiatrist when they saw some aspect of my behavior wasn't normal and that it was affecting my quality of life, even if I was far too young to know or understand that. I think it was the humane thing to do, 100%.

People talk a lot about how terrible it is that 5 year olds are put on this shit, but extreme cases like mine are out there and they are legitimate. Sometimes things show up early because they're severe enough. It'd be fucked up if 10 year olds on SSRIs was the norm, but I have no doubt that there are rare cases where it'd almost be cruel /not/ to do it. Bipolar can absolutely be diagnosed in children. If you can diagnose it that early, why wouldn't you start treating it? Isn't that what you SHOULD do as a parent?


I get that it helped you. But there are concerns about misguided and inappropriate medication of children. It's a complicated issue, I admit. And I admit that I tend toward cynicism. Anyway, this seems like a decent overview: https://www.nimh.nih.gov/about/directors/thomas-insel/blog/2...


> “If someone had a putative diagnosis of Alzheimer’s disease, they might be on an Alzheimer’s drug like Aricept or Namenda,” said James Hendrix,

Aricept is an expensive drug and does about nothing. I dont even know why patients with AD are taking it.

http://www.alzforum.org/therapeutics/donepezil

> The effect size of donepezil's benefits is small and the drug does not modify the underlying pathophysiology of the disease


The term 'Alzheimer's' has been politicised. Making a diagnosis requires a brain biopsy, usually done as part of an autopsy. I don't believe a PET scan is capable of making the diagnosis, despite what the article says.

To understand what is going on, we have to see the problem that hit clinicians and families around 10-20 years ago. It seemed like an epidemic of confused elderly patients landing on wards without a specific diagnosis, very little ability to describe what is wrong with them, and causing disruption to other patients, either directly, or because they required a nurse to attend them continually.

Explaining to the family why their mother/father had gone from being someone who was independent at home to someone confused and unable to tend for themselves was difficult. Usually the blame would be attributed to infections (especially urinary tract infections, which are very difficult to diagnose in a confused patient), but doctors were reluctant to start using the term 'dementia' initially, particularly since there wasn't a good test for it.

In retrospect, what was happening was that we were seeing the success of treatment for cardiovascular disease. Many patients who would normally have died in their 50s and 60s of heart disease we now making it to their 70s, 80s and 90s, albeit with arteries that were still far from ideal.

Every organ in the body degrades with age. We tolerate and compensate for it up to a point. At the age of 20 our hearts can increase output from 5l/min to 20l/min, sometimes this is necessary to survive a particular bad infection. If your heart is unable to do this, we say it has decompensated. At the age of 60, even without a diagnosis of heart failure, this is not possible, but thankfully we are rarely in a position where 20L/min is necessary. The brain is no exception. An old brain with loss of volume can handle a daily routine in a familiar environment, but throw in a cold, or another infection, then make it worse by starting drugs and abruptly changing their environment, and we end up with an acute confusional state which can take weeks to resolve, if it ever does completely.

My suspicion is that most cases of dementia (and mild cognitive impairment) is due to a general decrease in function as we get elderly, exacerbated by vascular insufficiencies. Alzheimer's is a specific histological diagnosis - it doesn't give us a cause, it's more of a histological finding. Since we don't really have any good treatments for dementia caused by Alzheimer's or vascular causes, and we don't have a way of telling them apart, you could argue whether this histological finding matters..

Except that Alzheimer's is a term the public are familiar with, and will donate to.

NB I am not a neurologist nor a dementia expert. I've worked a lot in elderly care and have had to deal with patients and their families on many occasions. I have some amusing memories including overhearing a conversation between a confused patient on a ward who had picked up the ward phone at around 1AM - it was the bed manager calling to see if their were any beds, but the patient she was still at home so was quite confused at the question.


Very interesting idea! I've never noticed an increased heart rate during an infection though. That sounds kind of strange.


It's seen as fairly basic/text-book knowledge, and finding a paper on it is pretty difficult. It was probably shown in the 19th century and isn't easily found on google scholar! I did find one reference [1] from more recently. Part of our sepsis warning scores depend on the heart rate to rate the severity of the illness. I suspect even in mild colds your cardiac output goes up, but it might be difficult to notice just from measuring the pulse (stroke volume can change).

Heart is just one example - our kidneys get worse as we get older too (see figure 4 of this paper - it's a straight line pretty much [2]). In practice, it doesn't cause us problems in day to day life, but give it a really big insult (like sepsis) and it can decompensate. The difference with kidneys is that we can support kidney failure with dialysis or filtering (if their heart can support that), whereas with cardiac failure this is not so simple.

Brain failure is probably one of the worst forms of failure. Most causes have no cure, and they often require 24hr care.

[1] http://ajpregu.physiology.org/content/246/3/R331.short

[2] https://www.spandidos-publications.com/10.3892/mmr.2015.3292


Alzheimer is just the most common dementia.

Valscular dementia is also common and Dementia with Lewy bodies are also possible. Is it possible that many of these were misdiagnosed other dementia's?


Anybody read the words sensitivity, specificity, or gold standard? Just because it doesn't show up on PET doesn't mean squat if PET has a low sensitivity.


So what's the gold standard of detecting Alzheimer's? (It is a rhetorical question.)


Brain cutting and histopathology See http://www.bu.edu/alzresearch/brain-cutting-636x433/

"one of a variety of silver impregnation staining techniques, such as the modified Bielschowski or Gallyas technique, or the fluorochrome dye thioflavin S is typically employed to visualize neurofibrillary tangles

Because these techniques require either specialized equipment (the thioflavin S stain requires the use of a fluorescence microscope with specialized excitatory and barrier filters) or experienced histotechnologists (in the case of silver impregnation stains), most anatomic pathologists in general practice lack the capability to properly evaluate brain specimens submitted for the diagnosis of Alzheimer's disease. Such specimens are best referred to specialized neuropathology laboratories where the necessary experience and facilities are in place."


Alzheimers is a tauprotein disorder. Alzheimers without plaques is called Primary Age-Related Tauopathy, aka PARTS.


So what did they have?


I may have Alzheimer's but .. uh... at least I don't have Alzheimer's.

It's an old joke, but fitting for the title.




Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: