Sunday, October 21, 2012

The Diet Heart Hypothesis

The Diet Heart Hypothesis

The diet heart hypothesis is the most common application of the lipid hypothesis.  The version using total cholesterol goes like this:

  • A higher cholesterol level is associated with an increased risk of CVD
  • SFA (and dietary cholesterol) raises cholesterol and PUFA lowers cholesterol
  • Therefore SFA (and dietary cholesterol) increases the risk of CVD and PUFA decreases the risk of CVD. 

A possible new version of the lipid hypothesis using the Total:HDL-C ratio goes like this:

  • A higher Total:HDL-C ratio is associated with an increased risk of CVD
  • SFA raises the Total:HDL-C ratio and PUFA lowers the Total:HDL-C ratio
  • Therefore SFA increases the risk of CVD and PUFA decreases the risk of CVD. 

While SFA raises cholesterol and the Total:HDL-C ratio in the short term (see graphs below), it doesn’t seem to in the long term [1] [2]

 * If we go off the short term studies, replacing 10% carbohydrate with SFA will increase total cholesterol by about 0.4 mmol/l (15.2 mg/dl) and increase the Total:HDL-C ratio by 0.03.  For reference a good Total:HDL-C ratio might be between 3:1 to 4:1 a bad ratio around 5:1 or 6-1.  In other words replacing 10% carbohydrate with SFA increases the Total:HDL-C by less than 1%.  Then if we assume we can use the Total:HDL-C ratio as a risk factor we have increased our risk of CVD by less than 2.1%.  Then consider that no food in 100% SFA.  As we can see in the second graph all the fats are expected to reduce the Total:HDL-C ratio. 

Saturated Fat in Observational Studies

Cohort study: In women with CHD, the highest quartile of SFA intake had less atherosclerosis, and even slightly reversed it.  This initial result was not yet adjusted for confounding variables such as the higher rates of smoking, diabetes, trans fats, alcohol and less exercise, fibre and drug use.  Interestingly they had slightly lower LDL-C (ns).  They also had higher HDL-C, lower triglycerides, a lower total:HDL-C ratio, lower ApoB (ns) and less hypertension (ns).  Overall SFA and MUFA was associated with less progression of atherosclerosis, while PUFA and carbohydrates was associated with a greater progression of atherosclerosis [4]

SFA and CHD by European country [5]

And a meta-analysis found: 

“…insufficient evidence from prospective epidemiologic studies to conclude that dietary saturated fat is associated with an increased risk of CHD, stroke, or CVD” [6] 

Saturated Fat in Clinical Trials

Most of the clinical trials in this area are poorly designed.  Some aren’t randomised, most aren’t isolating one variable (SFA vs linoleic acid) and are instead replacing SFA and artificial TFA with omega 6 PUFA and omega 3 PUFA, and some change the whole diet along with SFA for the intervention group. 

“The American Heart Association (AHA) and individual scientists advise consumption of at least 5–10 % of energy as n-6 PUFA to reduce CHD risk. They note that randomised controlled trials (RCT) of CHD outcomes are considered to be the ‘gold-standard’ for guiding clinical practice decisions. Individual RCT, and two meta-analyses combining seven RCT, are cited as providing ‘the most convincing’ and ‘decisive’ evidence-base, with ‘immediate implications’ for ‘population and individual level recommendations’ to substitute n-6 PUFA-rich vegetable oils for SFA. However, the conclusions of these meta-analyses have been questioned due to their (1) omission of relevant trials with unfavourable outcomes; (2) inclusion of trials with weak design and dominant confounders; (3) failure to distinguish between trials that selectively increased n-6 PUFA, from trials that substantially increased n-3 PUFA; (4) failure to acknowledge that n-6 and n-3 PUFA replaced large quantities of trans-fatty acids (TFA), in addition to SFA, in several trials.” [7] 

Dietary recommendations to include linoleic acid have been based on data showing benefits when SFA and TFA have been replaced with both omega 3 and 6 PUFA, usually along with other inventions that improve food quality.  Replacing SFA and artificial TFA with both omega 6 and omega 3 PUFA decreased all-cause mortality by 22% [7] [8] 

However, replacing SFA and TFA with just linoleic acid increases CVD by 16% and all-cause mortality by 13% [7] [8].  Artificial TFA has been well established to promote CVD, so one would expect the effect of replacing SFA with omega 3 and 6 PUFA would be smaller than a 22% decrease in mortality and the effect of replacing SFA with linoleic acid would be greater than a 16% increase in CVD and a 13% increase in total mortality. 

“Unfortunately this potential confounding role of TFA was not appreciated. Similarly, the displacement of TFA, rather than the substitution of mixed n-3/n-6 PUFA for SFA, may account for some or all of the 22 % reduction in non-fatal MI+CHD death in our meta-analysis. By contrast, the increased CHD risks from n-6 specific PUFA diets in our meta-analysis may be underestimated as n-6 PUFA also replaced substantial quantities of TFA” [7] 

How is it that PUFA consumption is associated with lower CVD in observational studies, yet in clinical trials replacing SFA and TFA with LA increases CVD mortality?  There’s something called the healthy observer effect where those who observe officially endorsed health practices (such as replacing SFA with LA) have a lower risk of disease and death.  The healthy observer effect can explain why hormone replacement therapy was associated with lower CVD in observational studies but higher CVD in trials [8]

I’ll discuss in a later post a reason why replacing SFA and TFA with LA resulted in increased mortality.  If you would like to read more about the clinical trials comparing SFA with PUFA you can find that in most of the further reading and in ‘Perfect Health Diet’ by Paul and Shou-Ching Jaminet and ‘The Great Cholesterol Con’ by Anthony Coplo 

Further Reading:
(1) Diet-Heart Controlled Trials: a New Literature Review
(2) New Review of Controlled Trials Replacing Saturated fat with Industrial Seed Oils
(3) Good Fats, Bad Fats: Separating Fact from Fiction
(4) Precious Yet Perilous
(5) Myth: One High-Saturated Fat Meal Can Be Bad
(6) The Diet-Heart Hypothesis: Stuck at the Starting Gate
(7) Does Dietary Saturated Fat Increase Blood Cholesterol? An Informal Review of Observational Studies
(8) Saturated Fat Is Not Associated With CVD, Evidence of Publication Bias
(9) The Cholesterol Wars: Steinberg Strikes Back
(10) How Conflating the Lipid Hypothesis With the Diet-Heart Hypothesis Led to the Public Condemnation of Bacon, Butter, and Eggs

1 comment: