Thursday, June 16, 2011

Diet and Dementia - A New Study

Alrighty.  You might have noticed I've slowed down on the blogging a tad this month.  Well, there's sunshine and vitamin D to collect, and lots of social engagements (here in the Northeast U.S. we seem to do everything for the whole year in the summer).  Also, ever since I started doing crossfit, I have needed more time to admire myself in the mirror (see Melissa's comment).

Today's music is on the… cheesy classical side.  I'm feeling a little Spanish summer on the Mediterranean right now.  Just go with it (as always, right click in new tab): Concerto De Aranjuez Adagio.  

Paleoland events:  Turns out I'm the blog of the week at Latest In Paleo!  Also, "fantastic."  (Surely you jest, I mean, really?  Tell me more…thanks Angelo!)   AND very excited that  next week I will join the ranks of Stephan Guyenet, Kurt Harris, and Chris Masterjohn to participate in a Healthy Skeptic Podcast with Chris Kresser and Danny Roddy.  Chris asks me to ask you if you have any questions for me.  You can post them here on my blog, or if the Blogger comment system is too much of a deterrent (you are not the first - 20/20 hindsight would lead me to Wordpress…), then I'm told Chris is also going to post and you can ask a question over there.  If there is something I can't answer, I will surely dodge or turn it around on you just like any practiced therapist.

Now, the Science.  A new paper came out in the JAMA Archives of Neurology.  The Archives are always going to be pretty much a first tier journal in any specialty, and it is important to keep abreast of them.

Here's the paper:  Diet Intervention and Cerebrospinal Fluid Biomarkers in Amnestic Mild Cognitive Impairment.

The design:  20 healthy 69 y/os and 29 68 y/os with amnestic-type mild cognitive impairment were randomized to either a HIGH diet or an isocaloric LOW diet for four weeks.  The HIGH diet was 45% fat (25% saturated fat), 35-40% high glycemic carbohydrates, and 15-20% protein.  The LOW diet was 25% fat (7% saturated fat), 55-60% carbohydrate, and 15-20% protein.  The diets were designed by a nutritionist and the food was delivered twice weekly to the participants.  There is no more information about the diets, but I have emailed the corresponding author to ask for more details (um, were the deliveries kinda like Jenny Craig mirowaveable meals or whole food?  Is it sugar or starch?  Is the saturated fat animal fat, coconut oil?  What about the trans fats? You know, all those gory but vitally important details.)  All participants were free of major psychiatric or neurologic disorders other than mild cognitive impairment, and were free of kidney disease, diabetes, COPD, cardiac disease, alcoholism, and no one was on cholesterol-lowering medication (where did they find these 68 year/olds not on statins? Oh, Seattle.)

I quote: "The length of time participants consumed the HIGH diet was restricted due to safety considerations." My goodness!  Safety considerations!  What were they giving them?  Surely not something so horrific as coconut cream or steak… well, thus far we don't know what they were giving them.  I mean, I have to give the researchers huge kudos for examining diet and the brain at all!  But let's have a look at the results, shall we?

The participants received cognitive testing, oral glucose tolerance testing, blood collection, and a lumbar puncture before and in the fourth week of the diet.  Insulin, glucose, and blood lipids were measured, as well as CSF amyloid beta protein (a couple of types), tau protein, apolipoprotein E, and F2-isoprostanes (omega-6 derived marker of inflammation and oxidative stress).

Okay.  So the HIGH diet increased insulin resistance (measured as insulin area under the curve, which is nice), and the LOW diet decreased it (far more so in the healthy controls than the mild cognitive impairment folks, which is interesting).  On the HIGH diet, LDL went up by 20 points in the MCI patients and less than 10 points in the healthy controls.  LDL dropped 20 points on the LOW diet  in MCI folks and 15 points in the controls.  Not surprisingly, HDL went down on the LOW diet (by 4 points in controls and 8 points in MCIs) and up on the HIGH diet (about 4 points for both).   For some weird reason, the researchers used LDL/HDL ratio instead of the more standard total cholesterol/HDL ratio, and that went down in the LOW folks and up in the HIGH folks.

With respect to cognitive testing, everything was a wash except some mild improvement in "delayed visual memory" scores that were better on the LOW group than in the HIGH group.

In controls, CSF amyloid protein went down on the LOW group and up in the HIGH group.  In the MCI folks, amyloid went up in the LOW group (which would fit the general model that amyloid beta is a last ditch desperate substitute for cholesterol in those with a dementing process going on) and remained the same in the HIGH group (the authors interpreted this as pathology was already advanced in the MCI group so dietary interventions couldn't make it much worse).  CSF ApoE in controls stayed the same in the control LOW group and increase in the MCI group (again - ApoE drags cholesterol into the brain - likely to increase on a low fat diet if the brain is fat-starved) - ApoE went down in both control and MCI groups on the HIGH diet (the authors considered this to be a bad sign for the HIGH diet.  I'm not so sure.  If you need more ApoE, the inefficient ApoE4 will be found wanting…)  CSF insulin dropped in the controls on the LOW and HIGH diets - in the MCI folks insulin increased on the LOW and stayed the same on the HIGH.  Since low CSF insulin is associated with Alzheimer's, the researchers considered low insulin bad.  F2-Isoprostanes decreased in both groups on the LOW diet and increased in controls on the HIGH diet (but not in MCI folks).

Yeah. I don't know what to make of it either.  The authors seemed to think it was a ringing endorsement for the LOW diet.  "For healthy adults, the HIGH diet moved CSF biomarkers in a direction that may characterize a presymptomatic stage of [Alzheimer's disease]… the AD biomarkers were unaffected by the HIGH diet for adults with [mild cognitive impairment], possibly because more extreme intervention is needed to exacerbate already-exant pathologic processes."  And, shucks, if the HIGH diet was all industrially-fried pork rinds and chocolate cake with hydrogenated frosting, I wouldn't doubt it.

We'll see if more information comes from my email (I hope so!).  Otherwise, I think this study serves as another example to show that diet is important, but that research is pointed in the direction of macronutrients rather than the quality of the food, thus we can't really make useful conclusions.

(This paper also states that cholesterol does not cross the blood brain barrier in a healthy brain - Seneff et al suggest that transcytosis of LDL across the BBB is conclusively proven in this paper.)


  1. The diets were designed by a nutritionist--I'm surprised that either would promote health.

  2. Since IR increased in HIGH, I bet that the other 20% of fat was from healthy vegetable oil...

    QUESTION for your for this that will be the HOTTEST podcast EVER! As a husband of a brazilian psychiatrist, I see my wife struggling with patients. Almost none of them, even one with cancer, are willing to make changes in their diet. They all seek easy paths: pills. So, my question is: what are the medicine in psychiatry that you find is useless and the ones that you think can help (not a easy question, I suppose!).

  3. I have been doing CrossFit for three years now. I'm in the best shape of my life. It seems that I am meeting more and more doctors who have discovered the benefits of high-intensity cross training. The CrossFit community is also largely Paleo, which is where I discovered it.

  4. I did not understand. A high carb diet is better?. If cholesterol does not cross the BBB how is it that some people sa that high fat is good because the brain is made of cholesterol?

    I think the best blogging service is but blogger is supposed to last.

    Sorry to insist, but some bold words or sentences will help reading. At least those with bad sight and non native english speakers.


  5. Js290 - hah! Fortunately I have nothing to prove - as one of the shortest, smallest, and newest crossfitters at the box, my scores will never be competitive (except with myself). BUT, prior to having kids, I spent many, many years working out very regularly, several times with personal trainers, and never have I been able to build strength so quickly or get such obvious changes in muscles so quickly as with crossfit. While I had worked out with BFL nutrition before, I've never worked out very seriously eating Paleo before joining crossfit, so that may be the difference. For the past 5 years I've been doing sprints and weights at home, but rarely lifting anything heavier than 20 lbs (except the kids - now 40 and 32 lbs). I believe I've been crossfitting six weeks at this point - and a total of 9 workouts, and it's really quite amazing the strength gains and differences in arms and toning of the midsection and whatnot.

    Allison - maybe in three years I'll be able to do a ring dip :-)

    Mario - Thanks for your question - many of my patients are not interested in dietary changes either - I'm really looking forward to the podcast. The HIGH diet was 35-40% carb and 45% fat, high GI carbs at that - I'm guessing you could very readily increase IR with enough sugar in there.

    Gobbledy-gook- The authors certainly thought the LOW diet (high carb, low fat) was better. I wanted to explore this paper for a couple of reasons - 1) It does show some interesting biomarker changes with different diets. 2) I'm hopeful the corresponding author gets back to me with more information about the actual diets, as it is, who knows what they were eating? I mean, palmitic is not stearic is not myristic is not lauric acid, yet all of those are dietary saturated fats. (See Jamie's recent post:"The failings of gross biochemistry" ) We do ourselves a HUGE disservice trying to sort out something as complex as dementia by throwing mere macronutrient ratios at it and trying to say - oh, yeah, saturated fat is bad and a low GI high carb diet is good. A low GI high carb diet will pretty much have to be quite a lot of "real" food - or a lot of fiber - and very readily drops the HDL and LDL within the month as you can see. 3) I saw this paper tweeted many times, and many links to news stories, given the press, I thought I ought to go more deeply into it. 4) The differences in how the aMCI patients responded to the dietary changes compared to the controls are interesting, demonstrating a biochemistry that has been altered in unexpected ways - perhaps a more brittle and extreme biochemistry that is very vulnerable.

    As it is, this paper has a lot of links that are interesting that I will be following, and if we do get more detailed info back about the diets, we can possibly learn something.

    As it is I feel the biomarkers show a mixed picture of the diets' effect on the MCI patients - is decreasing ApoE really a good thing? And the authors have a graph in the paper showing an "inflection point" prior to developing MCI symptoms - some of the biomarkers are low in young adults, then go up in the years leading to dementia, but once dementia starts, the biomarkers drop again. The authors seemed to feel that with some biomarkers, increasing would be "bad" in controls (putting you closer to the inflection point) and "good" in dementia (pushing you backwards closer to the inflection point) - and, sure, I guess you could make a case for that. Or you could say that they tortured a model to fit their data so the LOW diet comes out swimmingly.

    In the paper there is no detailed discussion of why a high sat fat diet would lead to the pathophysiology of dementia - they link to several papers (some of which I've already explored) but I will hunt them all down and post anything new. They seem to think sat fat is inflammatory. Weird, huh? :-)

  6. Hello! First-time reader! I really love your blog!

    QUESTION: I've heard there are links between diet and depression. If someone appears to have a genetic disposition to depression, what are some dietary changes that seem to help? I've heard that boosting protein (with all those amino acids) can help, but is this true and are there more changes you would recommend? Also, are there any supplements you would recommend?


Tired of receiving spam comments! Sorry, no new comments on the blog

Note: Only a member of this blog may post a comment.