Member
Frailty and Alzheimer’s
New study suggests that “frailty” increases the risk of Alzheimer’s Disease for people who have AD changes in the brain:
https://www.thelancet.com/journals/l...371-5/fulltext
One of the difficult things with understanding Alzheimer’s is that we have a hard time correlating the severity of the disease with the brain changes (“plaques and tangles”) that we assume are the cause of the disease. Some people with relatively little p&t have bad AD, some people with very bad p&t have mild AD.
There was a study discussed on the boards a few months or so ago that did not detect any improvement in AD with physical exercise. This was disappointing, but most of us around here and in the medical community believe that physical activity, of which strength training is possibly the most robust form, is helpful or maybe preventive of dementia. And maybe people who have AD already may be beyond the point at which training can reverse or slow the dementia, but people who don’t have it yet can be protected by training.
This new study supports that, and sheds some light: if you have AD pathology in your brain, it will more likely be expressed (ie you have dementia) if you are more frail, which they defined as “joint, heart, or mobility problems.”
So, keep training! (But we knew that already.)
TMPHBITEU
Perhaps this is backwards. Perhaps if you are more frail, you are more likely to develop or have developed AD pathology. This error is common in the medical literature, and especially in the media reporting of the medical literature.if you have AD pathology in your brain, it will more likely be expressed (ie you have dementia) if you are more frail,
Member
Not in this case. I will try to post details from the study when I am back at work and can get behind the paywall.
What this study says is that if two people have the same degree of Alzheimer’s-type brain pathology, the person who is frailer is more likely to display symptoms of dementia. The brain changes don’t correlate with frailty one way or the other, the observable dementia does.
We have no idea how to prevent the brain changes seen in AD, or frankly whether those brain changes are the cause of the disease or a symptom. Pretty much every 80yo brain at autopsy will show some AD-type pathology, whether the person had dementia or not.
Personally, if my brain is full of plaques and tangles when I’m older (or even now), but I don’t show signs of or experience symptoms of dementia, then I don’t care. Training looks like one way to achieve that.
Member
First, let's look at what claims the author is making:
1. Frailty is positively correlated with the [phenotypic, non-invasive-diagnostic] expression of Alzheimer's disease dementia.
2. People with lower frailty index scores were better able to tolerate neuropathic load.
2.b. This claim held true even after vascular risk factors were controlled for.
3. The study suggests a complex aging process increases disease risk, rather than genetic risk.
Next, let's look at the most important of their defined terms.
Neuropathological load: amount of amyloid plaques as assessed by silver stain. - I see no major problems here, though they could have done immunoassay as the author recognized.
Frailty: A multi point assessment of general physical deficits including common disease states. -Most of the items in the frailty index are positively correlated with each other. No issue, as long as they control for many disease states commonly associated with alzheimer's. However, I would have liked to also see a physical strength index or estimated body composition index, because the general perception of frailty is much more highly associated with physical strength than the FI (frailty index the authors used) is.
Next, let's look at their data analysis methods and controls.
n=653
Logistic regression Bayesian statistical analysis.
Fraility tertials defined in 1SD intervals.
General procedure for selecting study participants based on availabilty of sufficient information.
Quick scan of their methods, I see no major issues.
Controls:
Control for vascular disease, age, sex, gender.
*** Major problem. Not enough or appropriate controls to support the claims they are making in this paper.
I see no controls for family history or genetic risk. I also only see two vascular risk factors from within the FI controlled for and analyzed separately. It leads me to believe they found some data correlations that did not support their claims. ie. perhaps diabetes showed a higher positive correlation than the overall frailty index or a grouping of vascular or metabolic risk factors showed similar or higher correlation.
Finally, let's look at if the data supports the claims and other notes.
1. Frailty is positively correlated with the [phenotypic, non-invasive-diagnostic] expression of Alzheimer's disease dementia. -Due to lack of appropriate controls, the claim was at best partially supported. The p value was .03; there is a 1/33 chance the conclusions in this paper are completely random. In addition, the frailty index could be considered misleading jargon in this context.
2. People with lower frailty index scores were better able to tolerate neuropathic load.
2.b. This claim held true even after vascular risk factors were controlled for.
Yes, claim 2 was supported. However, too few vascular risk factors were controlled for. Why not control for diabetes? Hypertension?... etc.
3. The study suggests a complex aging process increases disease risk, rather than genetic risk.
- No, the third claim was not supported. There were no controls for family history or genetic disease risk.
Concluding remarks:
This paper is another one of those "Water is Wet" type papers.
Common sense has and would tell you that increasing amounts of wide ranging disease states or their indicators would be positively correlated with the outward appearance or diagnosis of dementia regardless of underlying pathophysiology.
The author's needed more published papers, citations or to generate media buzz to further their careers.
Member
I think you're a little hard on the paper. I also think you are better equipped to delve into the statistics than I am, but I happen to look at a lot of studies specifically in the neurology/psychiatry literature, and this one is not bad.Originally Posted by 凤凰来仪
A few points:
It is well known that vascular risk factors, including diabetes, increase risk of AD. It's not totally clear how or why. This paper was more focused on the frailty phenotype, which may be a softer concept and a bit of jargon, but is clinically meaningful. It is also meaningful to the layperson: "be less frail to protect yourself from dementia" is an actionable recommendation. This is consistent with the message of The Barbell Prescription, and comports with much of what we assume about the long-term benefits of strength training. "Reduce your blood pressure and control your diabetes to protect yourself from dementia" is also valid advice, but would certainly not be new, and wouldn't have merited a mention on this board.
Diabetes, hypertension, and a few other medical conditions accounted for 1 each of the 41 items on the "frailty index" in this study. If the frailty effect was mostly driven by those conditions, it was only because those conditions led to significant other signs of frailty, e.g. worsened walking speed, grip strength, ADLs, etc. Someone with DM, HTN, CAD who is well-treated and in good general health would score as non-frail and would have weakened the association and the paper's conlusion, if these conditions themselves led to worsened AD, regardless of their effect on general frailty.
I'm not bothered by their lack of family history and "genetic" controls: mature onset AD is almost never related to presenilin or amyloid genes (those familial AD types tend to present much earlier). Familial cases are not the patients in this cohort. They did look at APOE, and the APOE4 variants (which raise risk of AD) were about equally distributed in the frail/non-frail cohorts.
I think that a p=0.03 is convincing: this is debated of course, but <0.05 is the acceptable standard, and saying there is a 97% chance that the results in this study were NOT coincidental is good enough for me. Especially since it does not seem to contradict with the entire body of AD/lifestyle/health research done before, and the recommendation to "not be frail" is harmless at worst.
I agree that "sicker people are sicker" is not exactly an earth-shattering conclusion, but I think the fact that the expression of AD brain changes into clinical Alzheimer's disease is mediated by a modifiable risk factor, frailty, and that frailty is something that people around here know how to work on, makes it newsworthy here.
Anyway, I just posted here so that people would see it and think about if, if so inclined, so mission accomplished.
Sully
Member
Using my own admittedly limited experience, I'm inclined to think, at least in the situation I'm familiar with, that the dementia preceded the frailty.
My father-in-law died of Alzheimer's a little over a year ago at age 88. The last couple years of his life he was very frail, and also pretty far gone with the dementia, but before that he was strong. He never, to my knowledge, lifted, but he did gymnastics as a younger man and even into his 70s would do hand-stands without difficulty. He went white-water rafting when he was in his early 70s. He was also mentally active, he had a number of inventions he was trying to find buyers for.
It seemed to me that the dementia took away his desire, or ability, to do the things he once did and that caused the decrease in his physical condition. The issue was exacerbated by hip troubles (bone-on-bone) for which he couldn't have surgery, because dementia patients often take a steep downturn after anesthesia, so his hip problems were mostly handled with pain management protocols.
Admittedly my sample size is one, which doesn't really prove anything, except that it IS possible for someone who's both physically and mentally active to develop dementia.
Member
Yes, very true, nothing can fully protect us from dementia. What this study would suggest is that if your father-in-law were less fit and active, the dementia would have shown up earlier.
Member
I think you are right. I was a little too harsh with my evaluation.
There is certainly valuable information presented in the paper. Without revision, I do not think it belongs in a journal with an impact factor of 27 (53 for the main Lancet Journal, which I believe it is still the second highest impact factor among medical journals) I have seen peer reviewers send back papers for revision for more minor issues than this one had.
Given the popular conception of frailty as a largely physical-strength related condition rather than disease-state, I still would have liked to at least see them separate out a "strength index" of some sort to further bolster their primary claim.
I would agree that family history and genetic controls would be unnecessary had they not gone on to make the unsupported third claim. I am not familiar enough with the "Rush Memory and Aging" study design and cohort to know if or how they screened for familial history. Given the large number of genes associated with the development of AD, the only feasible way to screen for the genetic component would be family history. If I am wrong, send me some citations, but as far as I know there are not any data disproving an association between family history (or a complex genetic panel) and tolerance of neuropathic load as defined in this article. The tolerance of neuropathic load could be more strongly correlated with genetics than frailty even if the APOE component is not. In my opinion, they have not proven or substantially supported the claim that the tolerance of neuropathic load is environmental rather than genetic. [It is also interesting that many of these genes could be factors in the development of various vascular diseases - remains to be seen.]
If they had provided a more robust comparative analysis of their data and had moderated or further supported their third claim, I would support its publication in this journal.
It is very possible if someone here is in an academic position and looking to publish, an email or two could lead to a paper in a very reputable journal. They have a large amount of data that could be analyzed further to produce a possible "strength index" and body composition or muscle mass association with AD and neuropathic load tolerance.
I think Mark is making a good point. Unless I have missed something in the study design, we cannot reasonably exclude the possibility that dementia precedes frailty rather than proceeding from it.
And thank you for posting the paper. I think it is worthy of discussion and I like to see these things. I might otherwise miss them.
Member
Thanks for your reply. I enjoy the dialogue, and you make some good points that are thought-provoking. I agree, and I think the paper's authors would too, that we cannot yet conclude that frailty is an independent cause of neuropathology being expressed as dementia. We already knew that some non-frail people become demented, and some frail people don't. And the study here, as constructed, cannot prove that frailty is a modifiable cause rather than itself a product of an underlying dementia. But.... the results are consistent with the POSSIBILITY that frailty is a modifiable risk factor for dementia. And we are very much in need of modifiable risk factors for AD, since we don't have any disease-modifying treatments.
I disagree with your focus on genetics. Doubtless there are dozens, probably hundreds, of genes that have some bearing on an individual's risk of developing AD, most of which we don't know much or anything about. In the clinical world, we think about 4 genes for AD. Three of these are related to presenilin or amyloid, and the mutations are autosomal dominant and lead to early onset familial AD. None of the subjects in this study had early AD, we assume, because the mean age at baseline of these subjects was 83. It's fair to assume that none of them had presenilin or amyloid mutations.
The other gene that modifies AD risk is APOE, which they did look at, and it did not appear to have an effect on frailty (APOE4 was almost evenly distributed among the frail and non-frail subjects, though with a slight trend to associating with frailty). So I don't think there was another underlying genetic factor (that is a known AD risk) that was driving these results and producing the frailty.
If you know of other specific genes that would be relevant here, please share.
Regardless of the weaknesses of the study, I was pleased to see that it fits with what we tend to believe around here: stronger people do better. And I don't think the flaws of the study are fatal: at worst, being stronger doesn't have an effect on whether you get AD, but there is some evidence here that leads me to believe that it should help. There is nothing here to suggest that getting strong and less frail would be harmful!
As I mentioned in my response above: if frail people are more likely to show dementia when they have Alzheimer's brain changes, then I am going to try not to be frail. That probably means there is a delay in onset of symptoms, rather than avoiding it entirely, but that would be just fine.
Member
If I had been one of the reviewers, typical protocol is to make note of the perceived weaknesses of the paper or where there are issues with the claims, data, design, methodology etc. and if what has been proven is novel or impactful enough to justify being published in a journal with the impact factor score they submitted it to. Then, along with the evaluation of the weak points, you either recommend the paper is not published, is published or is sent back for revision with suggestions of what needs to be revised or strengthened. I would have sent this one back for revision with a few suggestions and without some of the harsher words. If something is perceived as a problem by only one of the three reviewers, it may not be that big of a deal. If a pattern emerges among all reviewers, it should be fixed.
Yes, APO/APOE are considered the best current marker. It is also known to have flaws when evaluated by itself in the context of AD pathogenesis. Looks like you have a good handle on literature search/review methods, so I will let you do the searching; There are quite a few papers on it at this point. When there is a complex genetic profile associated with development of the disease, family history still trumps any kind of genetic panel.
Stronger people are more useful! A simple, educated guess will often get us to the best conclusion!
Posting Permissions
