Thursday, February 28, 2008

Cholesterol (the HDL sort) kills?

This article was picked up by GinnyL, a prolific reader of diabetes news and a poster on Dr Bernstein's site. There are no links to anywhere from the article, so I pubmeded the guy who got a mention by name, Dr Genest, for 2008, and came up with this ref, which I can get at the full text of.

BTW there are some interesting papers of FH he's published, maybe another day.

There's a lot of waffle but you do get the fact he is talking about EPIC-Norfolk and IDEAL. This appears to be a reasonable summary of their findings:

The more HDL particles you have the better, the less cholesterol in the particles the better. ie you want to have a high HDL-C particle count with not too much cholesterol in each particle.

Your lab work merely picks out any HDL it can find, measures the total cholesterol in this fraction and gives you that number. Compare this to the accepted wisdom re LDL-C, typically from Dr Davis:

Your lab work just picks out the LDL, measures the cholesterol and guesses (sorry, calculates) an approximate number. Actually that's completely wrong. The lab measures the TC, subtracts a few numbers, adds VAT, subtracts the first number you thought of and that's the best guess of your LDL-C concentration. Maybe.

Anyway, if your really do measure particle sizes/numbers, the fewer LDL particles you have better, the more cholesterol in each particle the better.

Are you seeing a pattern here? It all comes down to particle numbers, sizes, contents. What controls all of these? Not statin deficiencies, as in IDEAL.

Forget your cholesterol. What marker predicts heart attacks and total mortality without all of the paradoxes?

Follow EPIC and HbA1c (yes, same that same EPIC study as this one) to get some sort of clue. Control what you are doing wrong diet/lifestyle-wise to glycosylate your haemoglobin and your liver will sort out whatever cholesterol particles sizes/numbers it needs for health.

The EPIC/IDEAL paper is quite amusing for the shock horror throughout the discussion that high HDL could be killing people via heart attacks, but you get used this sort of rhetoric from cardiologists. It sort of gets summed up by Dr Genest's comments about torcetrapib:

"The second such compound, torcetrapib, proved toxic despite causing a large increase in HDL-C levels and was withdrawn from clinical use"

No, Dr Genest, think again. Perhaps it's not "despite", perhaps it's "because". Perhaps it is the artificial packing of cholesterol in to HDL particles which was toxic, not the drug per se. Treating a lab number is all torcetrapib did, it's for idiots. Sort out your diet so as to live within the limits of your own personal level of insulin resistance and leave your lipids alone. On Kitava and for the fishing Bantu this can involve 70% of calories from carbohydrate producing an HDL <1.0mmol/l with excellent cardiovascular health. For an insulin resistant Norfolk diabetic it might mean 5% of calories from carbs and the rest from saturated fat giving an HDL of 2.0mmol/l.

It's not the labels on the lipids that do the damage.


PS Also from Dr Genest

"Mutations that impair the function of cholesteryl ester transfer protein (CETP) are associated with marked elevations in HDL-C levels but not necessarily with protection against coronary heart disease (8)"

Ref (8) does not quite say that the mutation (the one which torectrapib mimicked) was neutral, it said it was bad. Naughty mis-citation. He must be a cardiologist. And people are surprised at torcetrapib killing people, using heart attacks? Torcetrapib was a success, it did exactly what it should do! The Zhong et al paper was published in 1996. It was the basis for the development of torcetrapib. You can't say Pfizer didn't know what they were doing. You can't understand why they did it either.

Wednesday, February 27, 2008

Gluten and NK cells (forget the antibodies)

I just thought I'd stick this down as it's minor and it will get forgotten if I don't say something. There are grades/types of inflammatory bowel disease. The really easy one to treat is coeliac disease, where gluten avoidance is all that is required. There are also a host of problems which are responsive to gluten avoidance which do not come up positive on antibody test or have "typical" coeliac gut biopsy results.

I stumbled across this paper which grabbed me because my wife works with Natural Killer (NK) cells and has pointed out their relevance to surviving diseases before antibodies kick in. I was lucky to get the full text.

This is from the introduction, before it discusses coeliac disease, which is the second disease it covers (rheumatoid arthritis is the first):

"MIC A/B are stress-induced molecules acting as danger signals to alert NK cells and CD8 T lymphocytes through engagement of the NKG2D activating receptor [2]. At variance with classical MHC class I molecules, MIC are specialized for reporting stress without the requirement for peptide (or other ligand) binding and can direct NK and T cells to kill transformed or infected cells."

This translates as:

Gluten stresses your gut cells. They scream on the molecular level. NK cells cells get in there and put them out of their misery. This is (auto) immunity. It does NOT REQUIRE antibody production.


Antibody production appears to be an add on to the immune system, a memory storage facility for preventing future re infection with a previous encountered bug. The acute survival of an infection, like influenza, relies on the innate immune system, based around NK cells. If you had to wait 10 days for your lymphocytes to get around to antibody production before you recovered, all viruses would be lethal. NK cells are non specific and use molecules like MIC as markers that something is seriously wrong with a cell and it's time to get killin'. MIC is a like a large target sign painted on a cell and NK cells are equipped an AK47 looking for that target. Usually a cell expressing MIC is infected, but gluten appears to do as much "stressing" as a viral infection.


Gluten reactions do not need antibody production. It is perfectly possible to have serious on going gut damage without a positive antibody test. Statistically it is quite probable that you WILL get a positve antibody test. There is so much damage going in that battlefield which is your gut under the influence of gluten that there is a trauma "soup". All you need is for an "antigen presenting cell" (APC) to collect some gluten from this soup, give it to your lymphocytes and tell them "we want this stuff recognised by an antibody, get producing". Of course the APC may pick up tissue transaminase, endomysial protein or any of a host of other proteins to push at the lymphocytes (which make the antibodies). But there is a chance you may not get either gluten or transaminase etc presented (ie no antibodies in our current lab tests), so you continue with "psychosomatic" gut problems or get labeled as inflammatory bowel disease, which is not regarded as gluten responsive. Severe IBD requires more than gluten avoidance but this basic step is probably essential.

Antibodies are a consequence, and an amplifier, of gluten toxicity, not a cause. That's intrinsic to gluten, MIC and NK cells. What brought this on?

Speculation time.

Did I have psoriasis or Dermatitis Herpetiformis (essentially coeliac disease of the skin)?

Answer: Maybe it's the same parallel as gut problems. WGA certainly gets in to the systemic circulation. I think it's a reasonable assumption gluten does too. It will be landing on vascular and/or dermal cells in the skin. Gluten induced stress here results in MIC expression. NK cells have that AK47, an itchy trigger finger and are seriously looking for MIC. Sometimes there is antibody production, sometimes there isn't. This could easily be the result of receptor subtypes (genetic) on your antigen presenting cells. Different labels, different lab results (if it ever gets that far) but a similar approach is needed. Perhaps there are a host of immune mediated problems triggered by gluten that are not antibody producing.


Saturday, February 23, 2008

What, Sweden again?

I've had this paper kicking around since it was published. There are a few things which grabbed me about it. I was initially a bit dubious; a diet containing 20% of calories as carbohydrate didn't really seem low enough to be seriously useful for a diabetic. But then I twigged that these people were using a calorie restricted diet, so the absolute amounts of carbs were actually down at 80-90g/d. That's just above Dr Lutz's suggestion of 72g/d and not hugely above Kwasniewski's recommendations. Of course it's WAY above Dr Bernstein's recommendations (30g/d), but he really is a perfectionist, a good one too. The 50% of calories from dietary fat under weight loss conditions was obviously being supplemented from stored adipose tissue (quite highly saturated I believe), hence the weight loss.

These are some of the sections of text I particularly enjoyed.

From the introduction:

"The objective of the present study, therefore, was to determine to what degree the changes among the 16 patients in the low-carbohydrate diet group at 6-months were preserved or changed 22 months after start, even without close follow-up. In addition, we report that, after the 6 month observation period, two thirds of the patients in the high-carbohydrate changed their diet. This group also showed improvement in bodyweight and glycemic control."

OK, it's a short study. A couple of years is way too quick to see changes in heart attack rate from eating all that fat. Isn't it? You know, fatty streaks to soft plaque to... (yawn)

The other feature here was the comment about "changed their diet" applied to the control group (low fat, ADA style diet). It's made clearer in the methods:

"Seven of the controls switched to a 20 % carbohydrate diet immediately after the follow-up period. For those we have data 12–14 months after the change. Three more have later sought information and have changed diet. We have no long-term data for those. Five of the original controls have not changed diet."

This was an open study by the look of it. You can't spend hours sitting in a diabetes clinic chatting to someone who is loosing weight, dropping HbA1c and binning their insulin without realising that the LC intervention diet works. Only an ADA diabetologist is stupid enough to miss this. Given the option to change, you change. Unless you really are as stupid as...

But of course, all of that fat will have the LC diet group dropping like flies from heart attacks. We all know that. We've been told.

So, from the discussion:

"We have examined the medical charts for both the original high-carbohydrate group and the low-carbohydrate group from 3 months after the initiation of the diet therapy – when an effect might be detected – and forward for episodes of cardiovascular disease. Three episodes of cardiovascular disease have occurred among the 5 patients that never changed diet. The 16 patients in the low-carbohydrate diet group (19 months observation time) and the 7 from the high-carbohydrate diet group that changed diet (10 months observation time) – totalling 23 patients – have been free of cardiovascular disease during the follow-up period (p < 0.03. Fischer Exact)."

So we have three out of five people with episodes of CV disease during about a year of ADA style low fat diet and zero out of 23 patients on Lutz style high fat.

Must be a paradox. Ha! (now think of the Queen launching a new ship) I name this paradox "The Swedish Paradox".

What, again? Ok, there are only a limited number of countries in the world and so many paradoxes looking for names...


Wednesday, February 20, 2008

Bantu Lp(a) and Swedish Lp(a)

Lipoprotein (a) is interesting enough that I might put up some more thoughts about it, but this post is based on the Bantu cross sectional study cited, as all the best cardiac studies are, by Dr Davis of Track Your Plaque.

The initial link is to this paper which explains quite a lot about Lp(a) and its genetics.

I wanted to compare values of the Bantu with an intervention trial in some Swedes, and to think about eating fish and taking fish oil, to reduce that "risk factor" for heart disease. And yes, Lp(a) is a "risk factor" for all forms of vascular disease; ischaemic heart disease, peripheral vascular disease, cerebrovascular disease and abdominal aortic aneurism.

First the Bantu. That first paper covered the Bantu genetics, apoprotein (a) sizes and Lp(a) blood levels. Basically, if you match individual gene types, the fish eating Bantu always have lower Lp(a) than their matched apo(a) size vegetarian Bantu equivalents. You get nothing about diet in this initial paper, for that you need to go to a previous publication from the same group.

Fish eaters and vegetarians were pretty well matched for calorie intake at around 2100kcal/d. The fish eaters ate 18% of calories from protein, 12% from fat and 70% from complex carbohydrate. The values for the vegetarians were 11% from protein, 7% from fat and 82% from complex carbohydrate. Dean "Mcdonalds" Ornish would be proud of the vegetarians. BTW, both groups ate 4g/day of salt, worth remembering that.

On a fully vegetarian, completely home grown, ultra low fat, low salt, complex carbohydrate diet these Bantu natives show exactly the same rise in blood pressure with age as Westeners do. Their systolic pressure approximates quite closely to that "age plus 100" rule of thumb which used to be considered normal in the UK and diastolic pressure followed systolic as you would expect. Fish eaters stuck with a systolic around 120-130mmHg and a diastolic around 70mmHg, life long. Remember both groups only ate 4g/d of salt. Now what does that tell you about salt and hypertension? Anyway, the politically correctest Bantu had significant markers of CV deterioration with age, not so the fish eaters.

BTW, as an aside, HDL figures were abysmal with both groups at under 1.0mmol/l, exactly as you'd expect on a low fat diet. They were slightly higher in the vegetarian group, presumable because they drank more alcohol. Triglycerides were higher in the veggie group, again pretty much as you'd expect from their carb intake.

The theory was/is that it was the omega three fats which were saving the fish eaters from hypertension and Lp(a) elevation. They undoubtedly did eat more omega three fats, and less omega six fats, and we know from the Lyon Heart Study that this is probably a Good Thing.

The Bantu fishermen were getting roughly 5% of calories from fish derived fat, that's 105kcal, equivalent of about 12ml of fish oil per day, as fish. This intake of fish oil is associated with a Lp(a) concentration of 201mg/dl as opposed to 321mg/dl in the vegetarians.

So what happens if we take a group of Swedes and give them more than twice that dose of fish derived lipid, as 30ml of fish oil per day, then track their Lp(a) level? Initially the subjects were on a typical Swedish diet (no information given) and were advised not to change it throughout the study. The fish oil was added at 30ml plus some vitamin E per day for three weeks, then there were two weeks off of fish oil, then three more weeks on fish oil, but without the vitamin E (for Bruce; yes, lipid peroxides rocketed on the later protocol).

Lp(a) before fish oil plus vit E: 128mg/dl, after three weeks supplemented: 125mg/dl. Back on Swedish diet for 2 weeks: 124mg/dl, after fish oil without E for three weeks: 128mg/dl.

All of these figures are essentially identical. Fish oil at 30ml per day does nothing to Lp(a) levels on a Swedish diet. It's also worth noting that, although we don't know the Swedish geneotypes for apoprotein (a) size, these values are all much lower than either of the Bantu groups. Maybe 35% of calories from fat in Sweden?

So is it really the fish oil which is saving the carnivorous Bantu from the hypertension and elevated Lp(a) which are prevalent in their vegetarian cousins?

Somehow I doubt it, based on the Swedish data and the behaviour of Lp(a).

Intervention studies show us that low fat diets increase Lp(a) and high fat diets decrease it, I put the studies here.

The fish eating Bantu got through more than half a kilo of fish on a good day. If your average calorie in take stays unchanged, that's an awful lot of complex carbohydrate you didn't eat. If I assume causality, not eating complex carbohydrate appears to be extremely good for your cardiovascular system. And your Lp(a) levels.

You don't need any "magic" fats to drop Lp(a), unless those magic fats are saturated fats, which seem to do the job best (just turn the DELTA study on its head).


BTW: No, I don't believe our liver manufactures Lp(a) to kill us. Again, more thoughts on Lp(a) to come. I might even talk hedgehogs if I can get my head around what they were doing with their lipids!

Another BTW, the Kitava studies showed these islanders ran at 70% calories from carbs but with Swedish levels of Lp(a). Kitava was lowish in fat at 20% of calories, but their dietary fats were almost all saturated (from coconut), with just a little omega 3s from fish. So it looks to be arguable whether you need to consume saturated fats to keep you Lp(a) levels down or keep carbs below the vegetarian Bantu levels. Ignoring genetics. Open to discussion on that one.

EDIT 28th Feb: Just been back to the Kitava papers for something else and noticed Lipoprotein(a) was around 300mg/dl, ie Bantu levels, bad Bantu levels. No heart disease. Bear in mind Lp(a) is notoriously variable from lab to lab. But still needs slotting in to Lp(a) and heart disease.

Fiber, inulin and cancer

Just a quickie,

It has come up off blog on a number of occasions that inulin, a fructose polymer completely indigestible to humans but edible by bacteria, is a potential carcinogen. Couple of refs here and here. I was a bit dubious of this model as it showed beef to be carcinogenic too. That upsets my biases!!!! And slightly odd as this different model showed bacon, yes you read that correctly as bacon, to be protective. Beef and chicken were neutral in this chemical carcinogen based model. I decided that playing with mice, especially bowel cancer prone mice like ApcMin/+ engineered mice, might say relatively little about humans. Still not sure.

Now this dropped out of Pubcrawler today. If you are a ApcMin/+ mouse you can choose you carcinogen from wheat bran or apple pulp (that's the healthy fiber from healthy fruit, apples):

"In conclusion, both types of resistant carbohydrates increased polyp number and tumour burden and this was associated with elevated epithelial cell proliferation and crypt fission."

Because my access account doesn't allow me the full text I can't see what the fiber did to the wild type control mice. Always the most interesting bit and not in the abstract.

Crypt fission is bad, it's what wheat germ agglutinin and insulin do. Looks like all sorts of fiber do too. Just an extreme mouse model, but next time someone extolls their belief in the benefits of resistant starches or fiber, maybe they should read and think!


Tuesday, February 19, 2008

Cat rawfeeding; scroll down the comments

Just like to say thanks to everyone who posted comments on the last blog entry.

Thanks, they're appreciated.


Best ever statin study comment?

Hi All and possibly Brian,

I'd just like to direct anyone who is remotely interested in a diametrically opposite view to my own to the comment down on this post.

This is an essential read. I think we all have to realise that there MUST be two sides to a given controversy and, while I wouldn't want to make a habit of it, both sides of the argument need an airing.

For Brian's benefit,

Yes, my diet is exactly as described, no joking.

You ask:

"Are all these experts wrong, as well as the expert advisory panels on cardiovascular disease?"


A much better question:

Which hormone converts a vascular smooth muscle cell to an osteogenic cell (calcium phosphate secreting) in the vascular media?

Answer: Insulin

The NCEP answer: It's a statin deficiency! (what was the question? Oh never mind)

Thanks for the list of references, but I already feel there is far too much cholesterol hypothesis bashing on my blog. To go through them one at a time would be tedious and far better commentators than I have done so many times.

I note in passing that you cite the PROSPER study. Which particular cancer would you prefer to trade for your heart attack? Or perhaps you'd like to choose one for a grandparent?

Your concerns for my well being are touching, but I'll stick as I am thanks.


Thursday, February 14, 2008

French paradox in Sweden

This press release came out today, can't see anything on pubmed, but the findings are what you'd expect. Nice, whatever you think of Spurlock (not alot). Quote from Nystrom and his press release:

That signs of liver damage were linked to carbohydrates was another key finding, he said.

"It was not the fat in the hamburgers, it was rather the sugar in the coke," he said.

But the most startling result implies that an intensive fast food diet might have some health benefits too, apparently from fat.

"We found that healthy HDL cholesterol actually increased over the four-week period -- this was very counter-intuitive," said.

HDL, sometimes called "good cholesterol," seems to clean the walls of blood vessels, removing excess "bad cholesterol" that can cause coronary artery disease and transporting it to the liver for processing [in the fairyland of the lipid hypothesis that is, Peter].

Nystrom has yet to publish the cholesterol findings, but said they were consistent with the so-called "French Paradox." For nearly two decades, scientists have wrestled to explain how the French can consume a diet rich in fats -- from abundant butter, cream, cheese and meat -- yet have generally low levels of heart disease and hypertension.

Quick summary of the findings:

Sugar mangles your liver (a ten fold increase is ALT is VERY scary), saturated fat improves your lipids (and lots of other things), there is no French Paradox.

I love the French "Paradox", it makes me think of reality kicking hard at a very heavily barred door.


Wednesday, February 13, 2008

Vitamin D supplementation: Bad?

There's a new discussion paper out on vitamin D which deserves serious reading:

Vitamin D discovery outpaces FDA decision making

Some of it is very interesting if you want an opinion that suggests vitamin D supplementation, at any level, is bad and that low levels of 25 OH D3 are a product of disease, not a contributor. And as such are a marker of good things happening in the immune system. He includes the changes in gut flora as a source of obesity as being triggered by D2 supplementation (not had time to read all of this in detail) and I'm not sure I agree with him on what's happening (anywhere!), but there's plenty there to think about here. You have to look at both sides of a decision making process.

No time to read it in detail today, arghhhh


EDIT later in the day

By the time you've been through enough of the easily available references to realise that they don't actually say what the author claims they say, you tend to bin the review. There's the bin, dump goes the review. Gimme the D3.

Tuesday, February 12, 2008

Single vessel disease

Once again, thanks have to go to Dr Davis of Track Your Plaque for a nice introduction to an interesting subject.

Have a read through this blog post, just the comments are enough. The second comment is from a man with all of his coronary calcium score localised in to one single coronary artery. His bemused query is:

"Doesen't the same blood with the same small particle, low HDL and everything else travel through ALL the arteries? How come it only harms me in 1 artery? This doesen't make any sense to me."

Absolutely. I cannot see how the lipid hypothesis could remotely explain this phenomenon. If our liver merely makes LDL-C as a never ending supply of arterial suicide bombs, surely that evil LDL would attack ALL blood vessels? Sometimes it does, sometimes it doesn't. Often the attack is very specifically localised.

Another comment from a further anonymous poster is equally telling:

"When nurse called with results her explaination was that I must be a non smoker, which i am, because in smokers the plaque tends to spread more evenly in the coronary arteries and in non smokers they tend to show up in 1 or 2 arteries."

So another confirmation that, in non smokers at least, plaque is often localised to one or two out of four arteries. This localisation phenomenon is real. You just have to ask a lipid believer why this should occur.

Basically it was pretty well sorted out back in 1954. Well maybe not the smoking bit. My assumption for that is that smoking just damages everything, vascular linings included. Anyway, have a read here.

I really like this paper. It's solid, basic, core hydraulic physics. It could almost have been written by an anaesthetist. Page 2 of the pdf has a neat diagram, complete with algebra, explaining how the tension in the wall of a curved artery has to be highest on the inside of the curve. If you are going to get wall damage, that's where it will show first. He goes through Laplace's general law relating vessel wall tension to hydrostatic pressure (in excess of that of the surrounding tissue), the radius of curvature of the vessel along its course and the radius of the vessel itself.

He then settles down to predict where arteriosclerosis is likely and unlikely to occur under various circumstances and backs it up with histology slides. The article is over 8 pages long and does deserve careful reading, but here is his suggestion as to why some coronary arteries are more prone to arteriosclerosis than others. He suggests it may be surrounding tissue pressure that reduces wall damage in some sections of some coronary arteries. Probably also curved arteries are more prone, tortuous arteries most prone. Wide diameter arteries are more prone. Lapalce's Law predicts this.

"Geringer has pointed out that in the instances where a portion of the anterior descending coronary artery lies buried in the myocardium, the buried segment is spared from atherosclerosis. Branches of the same artery lying in the epicardial fat, and therefore not surrounded by appreciable tissue pressure, are affected by atherosclerosis."

This ref to Geringer is from 1951 and seems to have fallen off of the bottom of pubmed! It's Geringer, E.: Am. Heart J., 41: 359, 1951 for anyone with access to a well established library and a duster.

Completely independently Wexler made this observation in a detailed analysis of spontaneous arteriosclerosis in aged breeding rats, back in 1964:

"One wonders if the physical pressure of the surrounding myocardium may decrease the need for maintaining the structural rigidity of these medium sized arteries by calcification. In the case of the larger epicardial branches, however, the presence of calcium could promote increased polymerization of mucopolysaccharides. The increased polymerization would provide the turgor or rigidity required by the exposed epicardial arteries."

What would be very interesting to see is is how protected or exposed individual calcified vs non calcified arteries are on EBCT scans...

It was obvious, at least as recently as 1964, that arteriosclerosis was a repair process. Possibly a repair process gone wrong, but a repair process never the less.

In 2008, lost in the lipid hypothesis, these oddities are inexplicable.

Arteriosclerosis today is merely our liver trying to kill us with sticky Lp(a).


Monday, February 11, 2008

High fat Dutch style

Thanks to erik for the link to the project outline for the dutch study looking at the influence of a high fat meal on cholecystokinin release. This looks to be part of the lead up to using high fat feeding to ameliorate the liver failure, loss of gut wall integrity and what looks like systemic inflammatory response syndrome in major trauma human patients.

A quick pubmed on the group leader gave this paper from when Greve was visiting Edinburgh. High fat syringe feeding by mouth before major blood loss in rats was the project. It improved all sorts of parameters, as any high fat eater would expect. But you have to be careful with these models of bled-out rats as it is hard to get a project license which allows the rats to live or die. Fair enough, I don't like rats suffering either. I'm very fond of rats. They make good pets. But ultimately improved survival is what is needed for human trauma patients and this cannot always be predicted by acute amelioration of lab findings in the first 24 hours.

But the fat will deliver! I believe... (sounding like a statinator here).

A quick look at the full text of one of the group's related papers gives an idea of what they mean by high fat feeding. It's about 50% of calories from olive oil. I feel 85% from beef dripping would be even better, but just try and get that through the nozzle of a 5ml syringe at room temperature!

Usually when the Dutch talk high fat they mean HIGH fat, just look at this paper on diabetes. Human clinical type 2 diabetics on 89% of calories from fat. They improve. Mind you, if you've eaten as much Utrecht cheese cake as I have, you'd realise why the Dutch are not afraid of fat.

It was a crossover study so you can see what the same patients were like on 89% of calories from carbohydrate. Just look at the basal figures on the left hand side of table 2. As a diabetic which would you prefer to be? Check the fasting glucose, look at the lipids if you care about them.

Just compare this to the Philadelphia approach to diabetes. I hit on this paper as it's the first related link to the Dutch one. It grabbed my attention as the high protein diet did badly! Luckily I have access to the full text and in Philly they essentially starved two groups of diabetics on 1300kcal/d (for people with BMI well over 30). Both groups were low fat (45g/d) and they were either high carb (138g/d) or very high carb (HC,170g/d). The merely high carb (138g/d) group was describes as a high protein diet (HP). Whatever. The high protein group did worse on an HbA1c basis, slightly surprisingly. This looks to have been because both groups ended up with similar final values (final HbA1c for HC of 6.9%, HP gave 6.6%). The higher protein group had a non significant drop from baseline as their starting value was lower (Initial HbA1c for HP 7.6%, for HC 8.2%). Would you have guessed that from the abstract?

BTW the HDL-C deteriorated in both groups, as you would expect on 45g of fat per day. No surprise there. Only the HC group had a fall which reached statistical significance. Bet you didn't see that in the abstract either.

What both Philadelphia groups needed was a Dutch 89% fat milkshake six times a day, with enough calories for weight stability. Of course, at that level of fat intake they wouldn't be hungry, so skipping the occasional shake would allow weight loss if so desired. Even without weight loss they'd still really get their HbA1c down.

And actually improve their lipids.


Tracking plaque 1950s style

Before EBCT scanners arrived on the scene, this is how you assessed plaque in peripheral arteries. Certainly back in the 1950s this was cutting edge. Coronary arteries are a non starter using this technique!

"With both needles securely in place [one in each femoral artery, no local block, ouch! My comment], 20 c.c. of 35%, Diodrast [an old fashioned type of iodine contrast, very irritant, again my comment] was injected simultaneously into both arteries from 50 c.c. syringes equipped with stopcocks, polyethylene tubing and Luer locks. The injection lasted about 7 seconds and was accompanied by fairly severe but transient pain followed by a burning sensation passing down into the legs and feet and lasting 20 to 30 seconds. This was followed by visible flushing of the skin of the legs in some cases. Meanwhile 5 to 7 x-ray films 14" x 17" (35 x 42.5 cm.) were exposed with a cassette changer at 1.5 second intervals, starting when about 10 c.c. of the dye had been injected."

No one died from reaction to the dye and all of them came back for a repeat arteriogram using the same technique. Some even came back twice!

Assessing the radiographs in the paper is quite difficult. Getting X rays in to journals while trying to maintain diagnostic quality is never easy, even today. When it's a pdf of a scan of a print of an X ray in a journal published in 1954 which you have in front of you, you need the eye of faith to see what the authors saw. However, with it being a three author paper this probably increases the likelihood of plaque regression cited as being real.

Yes, it seems they had plaque regression in 1954. It was a very small series (reading through the technique, I'm not surprised!) and the changes are not particularly convincing when viewed from 53 years later. But they do claim some degree of reversal over about a 3 month period. The radiologist scoring the radiographs for plaque changes was blinded as to treatment vs control when he assessed the arteriograms.

Apart from Professor Yudkin's throw away comment about ascobate and heart failure, I don't think I've seen a lot of information on ascorbate for atherosclerosis in the literature. Obviously the Vitamin C Foundation makes lots of claims nowadays, but I don't think I've seen anything serious in print. Dr Davis has only commented once on the Pauling/Rath therapy (ascorbate/lysine/proline) for IHD, and that was to say it doesn't drop Lp(a). Fair enough, but Lp(a) is another story. Obviously if anyone is eating modest fat, lots of fruit and vegetables and avoiding saturated fat there is no way they will drop their Lp(a), whatever vitamins they take!

The group doing the arteriograms were giving ascorbate at 500mg three times a day as their treatment, without change in diet. This strikes me as a miniscule dose. I just wonder how much ascorbate Prof Yudkin's elderly heart failure patients were taking in the 1970s. Oh, and whether Dr Willis' patients (it's a Canadian study) were getting more sunshine during those two to six months of his study period........ Vitamin D again? The paper was published in December 1954, it all depends on how long it took them to gather and publish their data. A modern day research scientist can get a paper from results to e-pub in about six weeks. Modern clinicians can take >2 years!

But however it is achieved, I don't think it's arguable; atherosclerosis is reversible.


Sunday, February 10, 2008

Lipoprotein(a) and DELTA

This has got to be one of the strangest abstracts I've read. Because it's only an abstract most of the really interesting information, like insulin levels, lies hidden in the American Journal of Clinical Nutrition.

It's a crossover design so all people acted as their own control group. The study participants were chosen as having low HDL cholesterol and/or elevated triglycerides, two classic markers of insulin resistance. The purpose of the study was to see which of two interventions decreased the participants "risk factors" for heart disease (ie it was a short study with no body count). Both interventions reduced some of the dreaded saturated fat from the of 35% of total fat in the Average American Diet (usually termed the Standard American Diet or SAD, but...). Excellent move to improve lipids wouldn't you think? Except replacing the saturated fat with carbohydrate decreased HDL by 7.2% and using monounsaturated fat decreased HDL by a mere 4.3%.

How about triglycerides? Well the MUFA substitution did nothing to triglycerides (the phrase used is "tended to be lower", meaning no change unless viewed through rose tinted spectacles. I hope their statistician resigned over this phrase, or has since been sacked). The use of carbohydrate cranked up triglycerides, but by how much it doesn't say, just "significantly".

But the best results were lipoprotein(a). You know, that BAD lipoprotein with the genetically pre programmed blood level. It got un-genetically reprogramed upward by 20% in the increased carbohydrate period. The MUFA feeding did far better on this front, MUFA replacing saturated fat only increased lipoprotein(a) by 11%.

In the murky world of lipid belief these changes are, I understand, considered to be BAD. All of them, except perhaps that non-changed triglyceride value in the MUFA feeding period.

Luckily the calculated LDL-C went down, although goodness knows what happened to lipid particle size and numbers (actually you can pretty well assume they got worse, especially the small dense LDL during the carbohydrate period). The abstract certainly makes no suggestion the researchers considered anything other than calculated LDL. It was enough that this reduced a little to allow the DELTA researchers to claim improved CV risk for both interventions.

Actually they claim that MUFA "improved" cardiac "risk" more than carbs. Even if you believe their LDL stupidity, can anyone really believe that a 6.3% reduction is either statistically or biologically "better" than a 7.0% reduction? In your dreams, cardiologist.

What came out rather well was the Average American Diet, SAD or not.


Saturday, February 09, 2008

What do I eat (2) recipes

I mentioned a few posts ago that I had some typical recipes on my hard drive, well here they are. I've tidied them up a bit and put a rough analysis on the end of each one. Salt and pepper can be added as preferred, It strikes me that none of them are special, this is just the sort of meals I've always cooked, but now we eat the meat portion, often with enough veg to keep carb intake reasonable and no bulk carbs. Obviously anyone needing bulk calories can simply add more butter to any recipe. Muscle building might need a little extra protein. You can improve the flavour by browning the onions and meat first if you like. I also cook a fair few Ken Hom wok recipes too, replace conservative vegetable oil with bulk butter or coconut oil and be careful to keep the fat...

So here are some typical main meal recipes. Some we eat with a vegetable, some not. Depends on the rest of the day...

Bolognese sauce.

1 lb 21% fat beef.
1 rasher bacon, chopped.
1 small can Sainsbury's chopped black olives (about 0.5 cupful), drained.
Butter 50g
Olive oil 1 tablespoon (20ml)
Tin tomatoes.
1 medium onion chopped.
1 sweet green pepper de seeded and chopped.
30 ml Pesto sauce (olive oil, not sunflower based).
2 cloves garlic crushed.
Splash red wine.

Place all ingredients in a saucepan, boil, stir, simmer for an hour, reduce to a thick sauce. Cheddar or parmesan grated on top to serve. You need a spoon for the fat.

This probably serves three. Eating half is a challenge. More than half is a fat overload. I've tried, don't go there.

Total estimated food in the pan: 2300kcal, Fat 193g, carbs 38g, protein 99g.

African beef stew, serves 2, maybe 3...

1 lb diced beef
Tin tomatoes.
Medium carrot, sliced.
Medium onion, chopped.
50-75g butter, depends on how fatty the meat is.
50g peanut butter.
About 200ml water, to just cover meat.
Salt and pepper to taste.
Fresh root ginger, however much you like.
3 cloves garlic, crushed
Pinch Cayenne pepper
Pinch ground cloves
Tablespoon vinegar or lemon juice.

Place all ingredients in a casserole, bring to boil, stir well, cover, place in oven at gas mark four for 2-3 hours until meat melts in the mouth. Stir every half hour.

Can be cooked very slowly on top of the cooker.

Total estimated casserole: 1600-2000kcal (extra butter), fat >125g, carbs 33g, protein 100g

Goulash. Serves 2.

1 lb diced pork or beef.
medium onion, chopped
sweet red pepper, deseeded and chopped.
Butter 70g
Tin tomatoes.
Full packet (20g?) paprika.
5fl oz soured cream.

Place all ingredients except soured cream in a casserole. Bring to boil, stir well, place in oven around gas mark 4 for 2-3 hours until meat melts in the mouth. Just before serving add soured cream.

Total in casserole 2100kcal, fat 170g, carbs 35g, protein 100g

Chili Mix - serves 2

450g economy mince. Can use pork mince.
1 medium onion, chopped.
1 red pepper, de seeded and chopped.
2 cloves of garlic, crushed.
400g tin tomatoes.
Splash red wine.
50g butter, more if your economy beef is now reduced fat.
1 fresh red chili pepper, finely chopped or 1/2 teaspoon chili powder or both.
1 tablespoon ground cumin
cayenne pepper to taste (not essential)

Prepare ingredients and place all in saucepan. Bring to boil, simmer for 40 mintues, stirring occasionally. Reduce to a thick sauce. Serve with grated cheese on top.

Very aprox per casserole : 1600kcal, 130g fat, 30g carbs, 90g protein

Any cook book will provide a host of recipes. Nourishing Traditions is good too, I just avoid grains and increase fat.

Oh, never forget the ice cream!

Six egg yolks, 1 pint double cream, 20g glucose (can use sucrose), 10g honey, capful of vanilla extract. Blend and freeze.



Tuesday, February 05, 2008

Fructose and Gout

Many thanks to Stan for pointing me to this BBC article and the study it relates to.

Obviously anyone making a habit of soft drink consumption, be that sucrose or HFCS sweetened, is in trouble metabolically. The joy of this report is the association between not only orange juice and gout, but also with apples and oranges as whole fruit. Observational perhaps, but the biochemistry is all there to support causality.

Ultimately fructose is fructose is fructose. Eat an orange and your liver will immediately set about defending you as best it can. When it fails or goes wrong, you get gout.

I like the BBC's picture. The chap with metabolic syndrome and gout is probably wealthy enough in the 1700s to buy sugar and to drink port, possibly the sweetest bulk consumption form of alcohol. Obesity and gout.

A few choice fruity quotes:

"I can think of some situations, for example in severe treatment failure gout, where reducing sweet fruits, such as oranges and apples could help," he added.

Translation, take drugs, more drugs and even more drugs. In absolutely dire straits, dump fruit. Why not just dump the fruit in the first place? Gout drugs do not look to be much fun.

"But this finding needs to be balanced against the benefit of fruit and vegetables in preventing other chronic disorders like heart disease and stroke."

Show me the data!

"...and fructose rich fruits (apples and oranges) were associated with a higher risk, the researchers said."

Quite a lot of fructose in grapes and plums and and kiwis and peaches and bananas and and and...

The nice thing about science is that eventually the truth will out.