Thursday, January 29, 2015

Observational Studies Regarding Saturated Fat and Coronary Heart Disease

There are a few main meta-analyses that pooled the results of observational studies.  I won’t go into much detail or list all the observational studies, though I may look into them later

The Jakobsen, et al meta-analysis included 11 cohort studies (4-10 years of follow up) with a total of 344,696 people, 5249 CHD events and 2155 CHD deaths.  The hazard ratios (HR) come from substitution models to see the whether replacing 5% of calories from SFA with MUFA, PUFA or carbohydrate (CHO) is associated with CHD (taken from data in table 2).  See figure 1 for the plots on the individual studies.

CHD Events
1.19 (1.00-1.42)
1.15 (0.84-1.58)
1.23 (0.98-1.55)
0.87 (0.77-0.97)
0.85 (0.68-1.06)
0.87 (0.76-1.01)
1.07 (1.01-1.14)
1.00 (0.89-1.12)
1.11 (1.02-1.20)

CHD Mortality
1.01 (0.73-1.41)
0.88 (0.51-1.54)
1.10 (0.71-1.69)
0.74 (0.61-0.89)
0.61 (0.37-1.01)
0.80 (0.64-0.99)
0.96 (0.82-1.13)
0.86 (0.65-1.13)
1.03 (0.86-1.24)

Their conclusion was: “The associations suggest that replacing SFAs with PUFAs rather than MUFAs or carbohydrates prevents CHD over a wide range of intakes.”  However, their results suggest that replacing MUFA with PUFA may be more beneficial*

* An odd explanation they provide for SFA>MUFA being generally unfavourable is that: “the main source of MUFAs was animal fat, whereby confounding from other dietary components in meat and dairy products cannot be excluded.”  Because (1) animal fat is apparently low in SFA; (2) we just know meat and dairy is bad; and (3) there’s definitely no confounding going on in the SFA>PUFA results

The Mente, et al meta-analysis used the used the Bradford Hill criteria to “derive a causation score based on 4 criteria (strength, consistency, temporality, and coherence)”, see table 1A score of 4 was considered strong evidence of a cause-and-effect relationship between the dietary exposure and disease. A score of 3 was deemed to indicate moderate evidence of causation. A score of 2 or less was considered a reflection of weak evidence of causation”.  They pooled the results of many observational studies looking at a large variety of dietary factors and these results can be seen in table 2.  The RRs came from comparing the highest quintile/quartile with the lowest.

Total N
No. of Studies
No. of Criteria Met (of 4)
Total Fat
(1) Consistency in Coronary Risk of Mortality
(2) Consistency in Coronary Risk, Mortality or MI
* “Bradford Hill score is 3 when restricting analyses to cohort studies of high methodologic quality (low risk of bias)”

The Skeaff & Miller meta-analysis pooled studies that compared the highest vs. lowest quintile/quartile and those with 1% and 5% increments in calorie intake from SFA.  The summary for all fats is presented below in table 3.  The forest plots for the SFA analysis are presented in the paper table 6, 7 and 8 (paper is open access).  They found SFA was not significantly associated with CHD events or CHD mortality.

The Siri-Tarino meta-analysis included 11 cohort studies (5-23 years of follow up) with a total of 347,747 people and 11,006 CVD events.  The RRs came from comparing the highest quintile/quartile with the lowest.  The characteristics of the included studies are presented in table 2 and table 3.  They found SFA wasn’t significantly associated with CHD or stroke

The recent Chowdhury meta-analysis included 20 cohort studies with a total of 283,963 people and 10,518 CHD events.  The RRs for the cohort studies (figure 1) and circulating fatty acids (figure 2) came from comparing the top third with the bottom third.  They found SFA wasn’t significantly associated with CHD in their analysis of cohort studies and those measuring circulating fatty acids.  It’s interesting that: (1) stearic acid had the highest RR among circulating SFA even though unlike other SFAs, stearic acid doesn’t increase LDL-C (it actually slightly lowers it) [1]; and (2) margaric acid and pentadecanoic acid (from dairy fat) had some of the strongest inverse associations with CHD in the circulating fatty acid analysis 

The most recent meta-analysis by Farvid, et al included 13 cohort studies (5.3-30 years of follow up), with a total of 310,602 people, 12,479 CHD events and 5882 CHD deaths.  RRs for linoleic acid (LA) came from: (1) comparing the highest vs. lowest tertile/quartile/quintile; (2) 5% energy increments from LA; (3) substituting 5% energy from CHO with LA; and (4) substituting 5% energy from SFA with LA

CHD events
CHD mortality
LA (high vs. low)
0.85 (0.78-0.92)
0.79 (0.71-0.89)
LA (+5%)
0.90 (0.85-0.94)
0.87 (0.81-0.93)
CHO>LA (5%)
0.89 (0.80-0.98)
0.87 (0.81-0.94)
SFA>LA (5%)
0.90 (0.80-1.01)
0.87 (0.82-0.94)

Similar to Jakobsen, the Farvid meta-analysis isn’t really anti-SFA, but rather is pro-LA.  They found LA is inversely associated with CHD and association was the same whether CHO or SFA replaced LA


In summary, all the meta-analyses agree that SFA isn’t significantly associated with CHD.  The Jakobsen and Farvid meta-analyses found that replacing 5% of total calories from SFA with PUFA (which means almost halving current SFA intake and almost doubling current PUFA intake) would reduce both CHD events and CHD mortality, but the replacement doesn't seem to be specific for SFA and appears to be an independent effect of PUFA.  These meta-analyses, even Jakobsen and Farvid, should not be used to argue that SFA are toxic (etc) as you must then come to the conclusion that MUFA (particularly in Jakobsen) and carbohydrate are equally as toxic (which is pretty rediculous).

* When looking at the Forest plots you can see how it’s easy it is to cherry pick studies to support any preconceived notions, which is a major reason why meta-analyses are important

Tuesday, January 27, 2015

Saturated Fat, Blood Lipids and Coronary Heart Disease

The Lipid Hypothesis

Total cholesterol (total-C) was one of earliest risk factors identified for coronary heart disease (CHD) and formed the basis of the lipid hypothesis, which is that “measures used to lower the plasma lipids in patients with hyperlipidemia will lead to reductions in new events of coronary heart disease” [1].  Later research identified the total-C:HDL-C ratio to be the measure of blood lipids most predictive of CHD, being twice as informative as total-C* [2].  With this information the original lipid hypothesis should be modified to ‘measures used to lower the plasma total-C:HDL-C ratio in patients with hyperlipidemia will lead to reductions in new events of coronary heart disease’.  Or simply put, a hypothesis that in a population: ↓ total-C:HDL-C >> ↓ CHD

* It really shouldn’t be surprising that total-C is far from being the most predictive of CHD, as total-C includes HDL-C which is inversely associated with CHD

** Using the total-C:HDL-C ratio or non-HDL-C:HDL-C ratio avoids the issues of (1) calculating LDL-C; and (2) ignoring cholesterol in other Apo B containing lipoproteins such as IDL and VLDL

The Diet Heart Hypothesis

The lipid hypothesis logically led to figuring out how to reduce cholesterol* through diet and drugs as a means of reducing CHD.  The fatty acid composition of the diet was identified early as having a major influence on cholesterol and led to the Keys equation**, which predicted the change in total-C based on changes in SFA and PUFA intake [3].

The knowledge that dietary factors affect cholesterol levels led to the diet heart hypothesis (or diet heart question as it’s sometimes called [4]), which takes many forms***, but the one driven by Keys and more commonly tested in future clinical trials is that: by reducing cholesterol, decreasing intake of SFA and increasing intake of PUFA (higher PUFA:SFA ratio) would be expected to reduce CHD.

When cholestyramine (a cholesterol lowering drug – bile acid sequestrant) first appeared successful in clinical trials for CHD the results were often generalised to cholesterol lowering diets and used evidence in support of the diet heart hypothesis.  But the results shouldn’t be generalised to another drug, yet alone a dietary interventions [1] [4] [5].  Each target of the lipid hypothesis has to be tested in its own right.

Fortunately there has been further research on the diet heart hypothesis: associations between intakes of SFA, MUFA and PUFA with CHD have been explored in many observational studies and the diet heart hypothesis has been tested in several clinical trials, although often quite poorly.

* Different people will use total-C, LDL-C or total-C:HDL-C, etc

** Δ total-C (mg/dl) = 2.74 Δ SFA (%) – 1.31 Δ PUFA (%).  Note that dietary cholesterol isn’t considered a factor in the equation “Such low-fat diets usually contain less cholesterol and animal proteins, but the change in the serum-cholesterol level in man does not depend on this fact” [3]

*** May also include total fat and/or dietary cholesterol, but neither of these originated from Keys.

**** The diet heart hypothesis was initially about total-C because that’s all that could be measured.  This explains why MUFA wasn’t a factor in the Keys equation because even though MUFA also reduces LDL-C and total-C:HDL-C, it doesn’t reduce total-C [6].  With the total-C:HDL-C ratio being the most predictive measure of blood lipids for CHD, the diet heart hypothesis should be concerned with total-C:HDL-C rather than total-C

A Problem with Mechanisms

The diet heart hypothesis is not an unreasonable hypothesis, but like any other hypothesis it needs to be tested.  Relying on mechanisms or risk factors as surrogate measures of CHD (or other hard endpoints) is a leap of faith that the incidence of CHD will be consistent in that direction.  Unfortunately the majority the conventional dietary advice stops there, simply stating that SFA increases LDL-C and that higher LDL-C increases the risk of CHD, then either implying or just outright saying ‘therefore SFA increases the risk CHD’.

For the above to be correct it assumes either that ‘SFA only increases LDL-C*’ or that ‘LDL-C is the only risk factor for CHD’.  However, SFA, PUFA, etc may affect the incidence of CHD independent of cholesterol for better or worse, and without testing you wouldn’t know.  There are good examples where the results of clinical trials were contrary to risk factors and mechanisms:

  • Carnitine increases TMAO and TMAO is associated with CHD [7], but carnitine supplementation reduces CHD events and total mortality [8] (post)
  • CETP increases total-C:HDL-C, but torcetrapib, a CETP inhibitor, significantly increased CHD events and total mortality in clinical trials [9] 

So far the discussion has been on the effects of single nutrients/chemicals, but it gets more complex when discussing food as food is comprised of many nutrients and other chemicals, that each has their own effects, not to mention nutrient interactions.  For example: does meat increase CHD because of SFA (or something else like Neu5gc), or decrease CHD due to carnitine?  This isn't to say that mechanisms and risk factors are useless.  It's just that it's important to test things and not make assumptions

* Whether it’s LDL-C or total-C:HDL-C, etc the point is still the same.  I originally used LDL-C in this section because most conventional dietary advice focusses on LDL-C

** See Just Kale Me

A Problem with Blood Lipids 

A further problem with relying on blood lipids as risk factors is that reducing cholesterol may reduce CHD events and CHD mortality, but may not reduce total mortality.  For example, a meta-analysis found the relationship between total-C and total mortality followed a U-shaped curve such that both high and low total-C was associated with higher total mortality [10].  That being said, there are probably many confounding variables in this type of research, such as low total-C being the result of disease or statin use in secondary prevention.  But it’s important to remember that cholesterol has many beneficial functions and lowering cholesterol too far will probably impair some of those functions.

* Cholesterol is essential to life, but this fact shouldn’t be used as an argument that it’s not capable of promoting pathology.  For example: SOCS3 is essential to life but is one of the major signalling molecules responsible for leptin resistance.

Tuesday, January 13, 2015

The US News Diet Rankings: Part 2

Health Risks

Are there health risks?
Possibly. By shunning dairy and grains, you’re at risk of missing out on a lot of nutrients. Also, if you’re not careful about making lean meat choices, you’ll quickly ratchet up your risk for heart problems.
While there are no specific dieter restrictions, you’ll want to consider talking with your doctor before making changes to your meal plans.

Calcium is probably the only nutrient in a Paleo diet* where the intake won’t meet the RDI because of the lack of dairy (meanwhile conventional dietary advice is often low in choline).  There are really no other nutrients in dairy and grains that can’t be found in meat, fruit and vegetables, which is generally true of most foods.  For example, despite people describing whole grains as good sources of B vitamins, meat is actually a better source of most of the B vitamins [1].

In the Paleo RCTs the Paleo diet is more nutrient dense than conventional dietary advice, which is not surprising considering that meat, fruit and vegetables is more nutrient dense per calorie than grains and dairy.

      Nutrient intake (% of RDA)                  Nutrient intake (% of RDA)**
      Average: PD 173%, DHCG 117%        Average: PD 196%, ADA 141%
      Median: PD 143%, DHCG 118%         Median: PD 156%, ADA 137%

As for calcium, there are other factors of a Paleo diet that probably makes up for the lower intake of calcium:

  • Vitamin D increases calcium absorption, and sunlight (or vitamin D supplements for those in higher latitudes) is widely recommended in the Paleo community
  • At most calcium/calcium + vitamin D supplementation reduces fracture rates by 24% [2], while vitamin K2 supplementation reduces fracture rates by 60% [3].  Conventional dietary advice recommends reducing animal fat, which probably explains how vitamin K2 is far more successful than calcium (the nutrient most of us are lacking for bone health is probably vitamin K2, not calcium)
  • By reducing grains and legumes, Paleo diets have a lower amount of phytic acid (which reduces calcium absorption) 

Lastly as for SFA and CVD, observational studies find no benefit when SFA is replaced with MUFA or carbohydrate [4] [5], and while replacing SFA with PUFA is associated with reduced CHD in observational studies [4], in the better controlled RCTs replacing SFA with PUFA results in either: lower CHD events, but not CHD mortality and total mortality [6] [7]; or higher CHD events and total mortality [7] [8] [9]

* Unless someone often regularly includes bulletproof coffee, ‘Paleo’ junk food or huge amounts of added fat in their Paleo diet, in which case all of those things would replace nutrient dense foods, thereby reducing nutrient intake.  There are bad implementations of all diets

** Nutrient intake adjusted to 2000kcal.  ‘RDA’ used came from average of men and women Australian DRI

*** In the section below they say the macronutrient ratio (P:F:C) of “a sample Paleo menu” is 38:39:23.  I don’t know where they got those figures from and I’m not too keen on 38% of calories from protein.  Cordain, et al estimated hunter-gatherers got 19-35% of calories from protein 28-48% from fat and 22-40% from carbohydrate, which averages to 27:43:31 (I know, it equals 101%)


The first page is what you expect, a whole bunch of diets which conform to dietary guidelines, with a little variation.  Page two is where you start to see some wacky diets with gimmicks, dubious claims and/or meal replacements.  Paleo was equal last with the Dukan diet and beating Paleo was:

  • Other diets that restricted food groups like the vegetarian diet #11 and the vegan diet #19.  Because restricting food groups is bad when it’s grains, but is fine when it’s meat or all animal foods (which do have unique nutrient profile unlike grains)
  • Diets based on dubious weight loss claims or outrageous justifications such as the Flat Belly Diet #16, Abs Diet #19, Medifast Diet #26, Acid-Alkaline diet #28, the Supercharged Hormone Diet #28 and Body Reset Diet #30

The rankings lose all credibility with their fairly high ranking of Slim-Fast #13, a diet based on poor quality, highly refined meal replacements.  Slim-Fast ranks more highly than several other diets based on food such as Zone #23 and Low GI #24 and is equal #13 with the Anti-Inflammatory diet

I’ve noticed a few things on how diets get a high rank:

  • Conform to the dietary guidelines, regarding macronutrient ratios and food groups (ignore that food groups are largely a human construct and often have little to do with nutrient profiles)
  • Emphasise the importance of reducing SFA, salt and dietary cholesterol (ignore that there’s no good evidence for this)
  • Only bother doing research if you followed the previous two points, otherwise it won’t count for much and probably won’t get cited anyway.  Even if you have no research supporting your diet you can piggyback on other people’s research and say that your diet is similar to theirs
  • Promote weight loss myths that the authors believe such as eating 6 small meals a day
  • Your diet should have lots of handholding.  Meal replacements or delivering meals is preferable to your dieters having to shop and cook their own food.  And certainly don’t require that your dieters have to do something crazy like research the diet to understand it
  • Your diet should be flexible so that it’s sustainable and easy to follow (ignore that mediocre effort often just produces mediocre results).  Diets that do enough to actually be effective are definitely unsustainable and there’s no way your dieters will stick with it long term
  • Leading on from the previous point, for the long term weight loss criteria your diet’s delivered meals or meal replacements should create a minor calorie deficit like 20 calories a day, not to be that effective [10], but just to tick boxes in the rankings.

To wrap up, a positive is that I don’t expect I’ll need to respond to the rankings again.  If history is anything to go by, the rankings and their justifications won’t change that much, even with new evidence.

Sunday, January 11, 2015

The US News Diet Rankings: Part 1

The annual circus of the US News Diet Rankings is back for a repeat performance this year.  Not one to disappoint old fans, they’re performing almost the exact same act as last year.  I usually don’t deal with comedy on this blog and leave that kind of thing to The Onion.  But I’m writing about this today because some poor person might get the idea that the rankings are supported by evidence and logic, and not simply written by clowns.

In this post I’ll be discussing their representation of the research regarding the Paleo diet (see here).  In the next post I’ll discuss the rest and look where some of the diets were ranked.

There are now 4 trials comparing the Paleo diet to convention dietary advice:

·         Lindeberg, et al (2009).  Paleo vs. Mediterranean
·         Jonnson, et al (2009).  Paleo vs. American Diabetes Association
·         Mellberg, et al (2014).  Paleo vs. Nordic Nutrition Recommendations
·         Boers, et al (2014).  Paleo vs. Dutch Health Council Guidelines

There are also 4 uncontrolled Paleo trials which I won’t discuss here besides saying that the first 3 were positive and the last one was a mix, but had some issues (see here)

·         Osterdahl, et al (2007)
·         Frassetto, et al (2009)
·         Ryberg, et al (2013)
·         Smith, et al (2014)

Weight Loss

“Will you lose weight?”
“No way to tell. Paleo diets haven’t yet drawn the attention of many researchers. One tiny study that looked at weight loss found that 14 participants lost an average of about 5 pounds after three weeks on a Paleo regimen. (But even the researchers called their study “underpowered.”) Still, if you build a “calorie deficit” into your Paleo plan – eating fewer calories than your daily recommended max, or burning off extra by exercising – you should shed some pounds. How quickly and whether you keep them off is up to you.”

Other Group
P value between groups
(12 weeks)
Δ Weight (kg)
Δ Waist (cm)
Δ Fat mass (%)
(3 months)*
Waist (cm)
(24 months)
Δ Weight (kg)
Δ Fat mass (kg)
(2 weeks)
Δ Weight (kg)
CI = (-2.3; -0.3)
Δ Waist (cm)
Bold: p < 0.05.  Underline: p = 0.05-0.10
* Jonnson, et al (2009) used a crossover design

The rankings mention the importance of a calorie deficit.  In the first three trials both groups spontaneously reduced calorie intake on both diets (both diets were ad libitum), though to a greater extent in the Paleo diet.  Satiety is often a limiting factor in weight loss using simple calorie restriction.  In the studies that measured satiety the Paleo diet has been found to be more satiating per calorie than the Mediterranean diet [1] and the American Diabetes Association diet [2]

The rankings misrepresent the literature on Paleo diets regarding weight loss.  They only reference one uncontrolled trial and none the four RCTs, which found: (1) you probably will lose weight on the Paleo diet; (2) the Paleo diet is equal to or better than conventional dietary advice for weight loss (etc); and (3) the Paleo diet is more satiating per calorie than conventional dietary advice

* The Boers study was different as it maintained set calorie intakes and the diets were made isocaloric

Cardiovascular Disease

“Does it have cardiovascular benefits?”
“Unknown. While some studies have linked Paleo diets with reducing blood pressure, bad “LDL” cholesterol and triglycerides (a fatty substance that can raise heart disease risk), they have been few, small and short. And all that fat would worry most experts.”

Other Group
P value between groups
(3 months)*
(24 months)
Δ Triglycerides
(2 weeks)**
Δ total-C:HDL-C
Δ Triglycerides
Bold: p < 0.05.  Underline: p = 0.05-0.10
* Jonnson, et al (2009) used a crossover design
** P value after adjustment for weight loss

I would argue that the cardiovascular benefits of almost every diet ranked is unknown without long term RCTs measuring CVD events, CVD mortality and total mortality, but I’ll settle for risk factors.

Three of the controlled trials found that: (1) the Paleo diet improves blood pressure and blood lipids; and (2) the Paleo diet is equal to or better than conventional dietary advice for improving blood pressure and blood lipids.

The statement “and all that fat would worry most experts” is among the more idiotic.  Why would they be worried about a higher fat diet?  Current evidence doesn’t support higher fat being an issue for CVD or low fat diets being therapeutic.  Isocaloric replacement of carbohydrate with any fat reduces the total-C:HDL-C ratio [3], the best predictor of CHD [4].  In observational studies, total fat [5] or replacing SFA with carbohydrate has no effect on CHD risk [6] [7].  Clinical trials have found that low fat diets don’t reduce CHD [8] [9], and just to make a point: the low fat, low SFA diet in the WHI, which included more fruits and vegetables, increased the incidence of CVD events among women with pre-existing CVD (HR=1.26, CI=1.03-1.54, P<0.006) [9].  Besides, the Mediterranean diet is ~35-40% of calories from fat and Cordain, et al estimate ~28-58% of calories came from fat in hunter-gatherer diets (average of 28% and 58% = 43%) [10].  Do you think a few % of calories from fat makes a difference?  If not, no double standards please.


Can it prevent or control diabetes?
Prevention: Being overweight is one of the biggest risk factors for Type 2 diabetes. If reverting back to the Paleo era helps you lose weight and keep it off, you’ll stand a better chance of staving off the disease.
Control: One small study comparing a Paleo and a traditional diabetes diet in 13 Type 2 diabetics showed the Paleo diet resulted in lower levels of hemoglobin A1C, a measure of blood sugar over time. The approach needs to be studied more before strong conclusions can be drawn, but most diabetes experts recommend a diet that includes whole grains and dairy products.

Other Group
P value between groups
(12 weeks)
Δ HbA1c
Δ 30m glucose
Δ 120m glucose
Δ AUC glucose
Normal glucose levels***
Diabetic glucose levels***
(3 months)*
Fasting glucose
AUC glucose
Diabetic OGTT (y/n)
(24 months)
Δ Fasting glucose
Δ Fasting insulin
(2 weeks)**
Δ Fasting glucose
Δ Fasting insulin
Δ AUC glucose
Δ AUC insulin
Bold: p < 0.05.  Underline: p = 0.05-0.10
* Jonnson, et al (2009) used a crossover design
** P value after adjustment for weight loss
*** No statistics on change from baseline within groups.  Paleo had 2 normal and 10 diabetic at baseline.  Mediterranean had 2 normal 9 diabetic at baseline

Weight loss is important regarding type 2 diabetes and as discussed in the first section the Paleo diet is good for weight loss, being equal to or better than conventional dietary advice.  These trials also provide a good example of how calorie intake/weight loss effects blood glucose:

·         The Lindeberg trial had more weight loss in both groups and a greater difference between the groups and consequently glucose tolerance improved and to a greater extent in the Paleo group
·         Compliance was poor in the Mellberg trial, which lead to only minor weight loss and little improvement in fasting glucose and insulin at 24 months
·         The Boers trial was isocaloric and when adjusted for weight loss there were no significant differences between the groups regarding glucose and insulin

But again the rankings only cite one trial when there are 4 controlled trials measuring glucose and insulin levels, and the one trial they cited wasn’t the more appropriate and very favourable Lindeberg trial.  Once again the controlled trials find that: (1) the Paleo diet can improve glucose levels and insulin resistance; (2) the Paleo diet is equal to or better than conventional dietary advice for improving glucose levels and insulin resistance; and (3) poor compliance = poor results, regardless of diet (the Mellberg trial)


The rankings consistently misrepresent the research regarding the Paleo RCTs, by citing inappropriate studies and only ever one study per section, making it seem like there’s less research on the Paleo diet than there is.  The Paleo RCTs find that the Paleo diet can promote weight loss and improve CVD risk factors, glucose levels and insulin sensitivity, and that the effect of the Paleo diet is equal to or better than conventional dietary advice.

I acknowledge that 4 RCTs with a total of 146 participants isn’t much and obviously isn’t enough justification to completely overhaul dietary guidelines, but the results of these trials should surely spark some interest and be replicated on a larger scale.