Sunday, April 27, 2014

The Real Food Guidelines and Saturated Fat in the Media

Things are happening in New Zealand 

The Real Food Guidelines 

Grant Schofield and several others have put together a document called the Real Food Guidelines to argue against the Dietary Guidelines.  You can view their analysis here 

In it they argue against guidelines to 

  • To be a healthy weight, balance your intake of food and drinks with your activity levels
  • Eat plenty of breads and cereals, preferably wholegrain
  • Prepare foods or choose pre-prepared foods, drinks and snacks: (1) with minimal added fat, especially saturated fat (and therefore also the emphasis on low fat dairy and lean meats); (2) that are low in salt; if using salt, choose iodised salt 

It’s great to see a variety of diet dogma being challenged at once (SFA >> CVD, fat >> obesity, salt >> hypertension, grains are nutrient rich, grain fibre is healthy).  Regarding the first point, my take away wasn’t that they are CICO deniers but rather think weight loss is more sustainable if calories aren’t restricted (and instead are spontaneously reduced) 

I left some comments here and have some further thoughts: (1) it’s good they used RCTs to justify LCHF for weight loss (and CHD risk factors) rather than the CIH of obesity (which I don’t believe in, here’s why); (2) the traditional Chinese cooking of using the whole animal could be argued more strongly for with the nutrient density of organ meats, the importance of glycine and sustainability, but I appreciate that it’s a minor point) 

Saturated Fat's Health Benefits under Scrutiny 

This is followed by a segment on New Zealand TV (3rd Degree) called Saturated fat's health benefits under scrutiny.  I thought it was fairly well done.  It was presented in a kind of debate style and was fairly balanced in screen time and questioning towards the experts (unlike Catalyst).  Basically the show was Grant Schofield and Caryn Zinn arguing that LCHF is better for weight loss and SFA doesn’t cause CHD, against Jim Mann and Rod Jackson arguing the standard SFA >> LDL-C >> CHD 

Some thoughts: 

  • Caryn’s trolley looked quite nice and surely couldn’t be too objectionable
  • I can understand the emphasis on berries as they have a health halo as superfoods, but IMO it sounds kind of elitist and faddy (don’t take it personally, just something that’s been bugging me for a while).  Berries are very expensive per kg or per calorie, while other fruit like apples, oranges and bananas are quite cheap
  • Referring to fat is nutrient dense is generally incorrect because fat only contains fat soluble nutrients (the vitamins, CoQ10, CLA, LCO3), and perhaps choline, so added fat is generally nutrient poor per calorie.  High fat whole foods can be nutrient dense, such as avocado, fatty meat, brain, eggs, milk and soybeans, but it’s easy to see in the case of meat and milk that fat reduces nutrient density per calorie
  • There have been some comments about the excessive amount of butter used.  I don’t know if it’s just there for shock value.  (If you eat > 60-70% fat you probably do need to do things like that) 

I was disappointed with the opposition 

"There is no evidence whatsoever that a high intake of saturated fat is good. And in fact there is a lot of evidence that a high intake of saturated fat is bad”
"There are a very limited number of people that understand the literature who would actually say we’ve got it wrong in terms of saturated fat, and what is more not only is there a general consensus about that…” 

Perhaps it’s just that there are a limited number of people who understand the literature.  The meta-analyses used by the mainstream against SFA suggest there’s no benefit in replacing SFA with MUFA or carbs, and that there’s only a benefit in reducing SFA if it’s replaced with PUFA.  So if SFA is bad (or poisonous), then MUFA and carbs must be bad (or poisonous) too.  On closer inspection, I learned three main things about the trials that replaced SFA with PUFA (see The DHH Part 1-4 and the individual trials) 

  • The high PUFA group always received advice to reduce TFA and/or junk food.  Often the high PUFA group also received advice to increase whole plant foods, fish/omega 3 and/or lose weight, etc and/or there were other differences between the groups (cardiotoxic medicine, smoking and cooking methods for example).  Therefore there were many differences between the SFA and PUFA groups that put the PUFA group at an advantage
  • The trials that were better controlled (where the only confounding variable was advice to reduce TFA and/or junk food) showed either no difference or harm in replacing SFA with PUFA, with the only benefit coming from a reduction in ‘probable’ + ‘possible’ events in a non-blinded trial
  • The trials that showed consistent benefits (reductions in CHD events, CHD mortality and total mortality) were the most poorly controlled 

The show may lead people to: (1) make false dichotomies of SFA being either unhealthy or healthy and couldn’t possibly just be ‘ok’; (2) think that we just need to compromise, that the middle ground is the truth and we should focus on the things they agree on (like junk food and sugar).  But at least this gets the debate out there 

* Someone may object to Rod Jackson saying we can isolate SFA.  I agree with him, and that it would be done sufficiently in a margarine vs. butter study (for example).  Even though there would be differences besides SFA and PUFA it would have greater external validity.  It’s too bad that the trials did such a poor job of controlling for things in general 

** Jim Mann: "There is one thing I would never concede, and that is that it's okay to have as much saturated fat as you'd like,".  Enough said, unless ‘as much as you’d like’ means overeating, in which case overeating anything is generally bad 

*** "Frankly I think this is the biggest red herring to be polite – garbage to be a bit less polite," says Professor Mann. "Your ancestors and mine, wherever they came from, were jolly lucky if they lived to be 40”.  HG lifespan is grossly underestimated.  Most HG’s who make it to 15 live to 45 and 45 year old HGs can expect to live another ~20-25 years [1] 

**** Jim Mann: "They need to get on and produce that evidence, and I would argue producing that evidence before they try to persuade the world that they’ve got the answer."  I agree, dietary recommendations should be held to that standard.  It’s a shame they aren’t

Monday, April 21, 2014

High This, Low That

“The beginning of wisdom is the definition of terms”
- Socrates
High carb, low fat, moderate protein.  These are words to describe diets, but what do they mean?  The problem is that it really depends on how ‘high’, ‘low’ and ‘moderate’ are defined*.  For example: the Zone (30P:30F:40C) has been called a low carb diet, but is 40% carbohydrate really low?  Or is it simply ‘lower’ than the conventional guidelines and so should be described as a 'lower carb' diet. 

I’m going to propose some definitions for ‘high’, ‘moderate’ and ‘low’ to clear this up and base it on simplicity and some observations 

  • ~10% protein is essentially the RDI for the average person, which is the sufficient intake to meet nutrient requirements of 97-98% of the population [1]
  • Modeling has suggested that at least 15% protein is required to obtain sufficient micronutrients [2]
  • The recommended protein intake for chronic disease is 15-25% [1]
  • Protein intakes of 30-40% (depending who you speak to) and greater are thought to exceed the body’s ability to metabolise protein and/or get rid of urea
  • The average protein intake in Australia in about 16.5% [3].  You could argue that protein intake would be slightly higher if 35% [4] of calories didn’t come from junk food (high in refined grains (~10%), added fats and added sugar) 

Very Low
Very High

  • If moderate protein is 20% then it would be nice if moderate carbs + fat = 80%
  • There is an ongoing debate as to whether fat or carbohydrate is generally healthier.  At this point I don’t know, and probably no one does
  • Eating ~0% fat or carbs is doable (no need for dietary carbs, PUFA requirement is extremely low), although very restrictive and probably not too healthy.  Anyway, it would be more accurate to refer to such a diet as zero fat/carb
  • In PHD, the Jaminets calculated carbohydrate requirements (dietary or not, in the absence of ketosis) to be approximately 30% [5] 

Very Low
Very High

There are two main issues with this kind of classification: 

1.     A 1% increase in carbs from 47% to 48% magically results in the diet changing from a moderate carb diet to a high carb diet.  This is silly.  Think of the cut-off points not as thresholds, but as gradients
2.     These numbers are proportions, not absolute amounts.  The absolute amounts are more important to determine whether you’re getting enough of something.  In some circumstances the % can be misleading.  For example: an athlete on a low carb diet may eat more carbs than an inactive elderly person on a high carb diet like the dietary guidelines (45-65%) 

Finally ‘high’, ‘low’ and ‘moderate’ aren’t some judgment of whether a diet is ‘good’ or ‘bad’ 

* The field of diet is full of people using ambiguous labels to promote an agenda, for example: veg*ns say the SAD is high in animal protein and fat; mainstream dieticians say the SAD is high in red meat, SFA and salt; low carbers say the SAD is low fat and high carb; and everyone says the cliché of the SAD being high in sugar, fat and salt (if they choose their words poorly they might imply salt contains calories).  I heard a dietician say Australians eat a high meat diet and handed out some information suggesting we should eat 1 serve of meat per day.  You shouldn’t be surprised when you conclude that Australians eat a ‘high meat diet’ if you define ‘high meat’ as > 1 serve/day

Wednesday, April 16, 2014

The Women's Health Initiative

Studies Associated with the Trial 

The evolution of the Women's Health Initiative: perspectives from the NIH (1995) [1] (no access)
Design of the Women's Health Initiative clinical trial and observational study (1998) [2]
Low-fat dietary pattern and risk of cardiovascular disease: the Women's Health Initiative Randomized Controlled Dietary Modification Trial (2006) [3] (this post only used this study) 

Participants and Diets 

48835 post-menopausal women, between 50-79, of diverse backgrounds and ethnicities were randomised to either a low fat diet or a control group. 

The low fat diet was intended to reduce total fat intake to 20% of calories and increase intakes of vegetables/fruits to 5 servings/d and grains to at least 6 servings/d”.  The comparison group received diet-related education materials”.  The low fat diet wasn’t calorie restricted, didn’t set weight loss goals, and didn’t try to alter fat composition. 

The groups were very similar at baseline, which is what you would expect with that many participants.  There was a significant difference in aspirin use, but I don’t think 17.4 vs. 18.1 is very noteworthy, which goes to show how much easier it is to attain statistical significance with a larger sample.  This leads me to the next point 


There were a few statistically significant differences between the groups in how the CVD risk factors changed over the 3 years, all in favour of the low fat group.  Once again, most of them aren’t that noteworthy as you can see below.  There was no difference in SBP, trigs, markers of IR and fibrinogen* 

3 Years
Low Fat
Low Fat
Physical Activity (METs/week)
Diastolic Blood Pressure
Total Cholesterol (mg/dl)#
LDL-C (mg/dl)#
Total Carotenoids (µg/mL)
Factor VIII* (%)
* Increases blood coagulation in response to injury
# P < 0.001 

Nutrient intake changed the way you would expect.  The low fat group increased protein, carbs, fibre, folate and servings of fruits and vegetables and grains; while decreasing total fat (and all types roughly in proportion), dietary cholesterol and nuts.  They achieved the target for fruit and vegetables and were close to the fat target after 1 year, but compliance got worse as the study went on and they actually consumed fewer grains at 6 years than at baseline.  The control group also made similar changes in the diet but to a much lesser extent** 

1 Year
6 Years
Total Fat (%)
(Expected: 7.0)
Serves of Fruit & Vegetables
Serves of Grains
(No Target)

The low fat group tended to have fewer CVD endpoints, but none of the differences were significant
After doing a subgroup analysis the researchers found that those with hypertension were more likely to benefit from the LF diet and those with a history of CVD (3.4% of the women) were more likely to have an CVD endpoint on the LF diet (HR=1.26, CI=1.03-1.54, P<0.006) (I’ve got no idea why that’s the case) 

In summary, a low fat diet with more fruits and vegetables had a minimal effect on CVD risk factors and didn’t significantly reduce CVD in post-menopausal women 

* The low fat group had a modest weight loss and improvement in blood lipids which is contrary to ideas that low fat is responsible for the obesity epidemic and ruins blood lipids (mainly triglycerides).  It depends on what the fat in replaced with, and in this case the fat seems to be replaced by fruits and vegetables at least to some extent 

** Since refined grains are nutrient poor and most of the grains eaten by these women were refined, it’s surprising that there was no specification for whole grains 

*** In the subgroup analysis they also found that those who consumed less SFA and TFA and/or more fruits and vegetables had lower LDL-C levels and rates of CHD.  As I usually say, associations of this kind are probably just markers of compliance

Sunday, April 6, 2014

Low-Fat Diet in Myocardial Infarction (Ball, et al)

Low-fat diet in myocardial infarction (1965) 

Participants and Diets 

264 men, under 65, who had recently had their first heart attack, were randomised to one of two diet groups: a low fat diet or the control diet.  The low fat diet group were advised to eat only 40g of fat per day.  “The daily allowance included 14 g. (1/2 oz.) butter, 84 g. (3 oz.) of meat, 1 egg, 56 g. (2 oz.) cottage cheese, and skimmed milk. The nature of the fat consumed was not altered, nor were any additional unsaturated fats given”
We have a limited amount of data on what they actually ate, but we’re told the low fat group had a large reduction in fat (~ 45g vs. 110-130g*) and ate fewer calories (~ 1900-2000 vs. 2300-2600) compared to the control group

Those who were overweight (21% of controls and 15% of LF group) were put on calorie restricted diets.  In the control group the calories that were restricted mostly came from carbohydrate instead
* Even though they didn’t meet the target of 40g, it’s a good effort considering “the diet was often unpleasant” and there was a large difference between the groups
Both groups, particularly the low fat group, lost weight and lowered their cholesterol.  (The reduction of cholesterol in the control group may have been due to weight loss) 

Low Fat Diet
Control Diet
Total Cholesterol
6 Months
4 years
6 Months
4 years
* This difference was almost significant (01 > p > 005)
** Unfortunately we aren’t told whether the differences were significant, but they probably were
Ultimately though, despite the large difference in fat consumption and the reduction in cholesterol levels, there was no difference in either relapses or death, both in the general study population (table VII), and among those with severe CHD (table VIII).  (Fortunately the people who assessed the CHD events in this trial were blinded)

Also, “the mean follow-up time to first relapse or end of trial was 3-04 years for the low-fat group, and 3-05 years for the control group”.

One might argue that the lack of difference was because the low fat group almost doubled their intake of added sugar, but the researchers found that carbohydrate and added sugar consumption wasn’t associated with relapses.  Although if you ignore the 0-5 g/day group in table VI (which I find difficult to believe) there’s a trend, but it’s not that exciting