Monday, February 10, 2014

The Diet Heart Hypothesis: Part 1

The Evidence Against Saturated Fat 

I started this series by looking at the evidence against saturated fat in regards to CVD.  The evidence came from three types of studies 

1.      Observational studies
2.      RCTs examining the effect of SFA on blood lipids
3.      RCTs examining the effect of SFA on CHD events, CHD mortality and total mortality 

Of these only the third can tell us whether replacing SFA with PUFA will reduce CHD events, CHD mortality and total mortality.  The other two shouldn’t be relied upon for dietary guidelines or for establishing causality.  There are instances where things that appear favourable in observational studies and things that lower cholesterol or improve the total:HDL-C ratio have been found to have no effect or an adverse effect in RCTs (see these links) 

There were three meta-analysis done on with trials replacing SFA with PUFA, with inconsistent results.  Two found some benefit and the other found some harm if the PUFA was omega 6.  So I looked at trials in the meta-analyses and two others that weren’t included, the Anti-Coronary Club and the Saint Vincent’s Hospital Study.  Below are the links to the blog posts 

Summary of the Trials 

Below is a table that summarises the details related to the trial and the results of the trial.  I didn’t include all this information in my blog posts.
Trial Details
Result Details
1) Only looking at the two different oil groups
Y = Yes (replacing SFA with PUFA decreased or end-points)
2) It wasn’t randomised but there was stratification
- = No difference (in cholesterol or end-points between the groups)
3) FL = Free Living, DC = Domiciliary Care, MP = Mental Patients in an institution
N = No (replacing SFA with PUFA increased cholesterol or end-points)
4) Probably no age restriction except over 18 or 21
? = Don’t know as the result wasn’t reported
5) Cross-over design, so only 6 years on each diet
* = Probably significant with a larger sample
6) Turnover rate was a problem
** = Probably significant but researchers didn't say

My Meta-Analysis 

If I were to do a meta-analysis of that looked at trials replacing SFA with PUFA and have high standards, I wouldn’t include any of the trials because none of them did a perfectly simple replacement of SFA with PUFA.  There was always something extra, whether it was restriction of junk food (baked goods usually) and/or trans fats in the experimental group or more obvious things like more fruits, vegetables, fish, etc in the experimental group*. 

The next best controlled trials are those without a multifactorial intervention (deliberate or accidental) besides restriction of junk food and/or trans fat in the experimental group.  These trials are:
  • The Rose Corn Oil Trial
  • The Medical Research Council Trial
  • The Sydney Diet Heart Study
  • The Minnesota Coronary Survey 

If we to look at the results of these trials we get we find:
  • The experimental group reduced cholesterol in all trials
  • In Rose the results were consistently unfavourable for the DHH, but weren’t significant (total mortality came close (0.05 < p < 0.10)), due to the low number of participants (n=80).
  • In MRC there were fewer CHD events but no difference in CHD mortality or total mortality.
  • In Sydney we don’t know how many CHD events there were, but there were significantly more deaths from CHD in the experimental group and almost significantly more deaths from all causes in the experimental group, again probably due to the low number of participants (n=458)
  • In Minnesota we don’t know how many deaths from CHD there were, but there was a small decrease (~10%) in CHD events in the experimental group among men that was not significant and no difference in total mortality among men, although among women there was an increase in both CHD events (31.7%) and total mortality (16.4%) that may be significant, but the researchers didn’t say
Supportive of DHH for CHD Events
Supportive of DHH for CHD Mortality
Supportive of DHH for Total Mortality

So far these collective results suggest an inconsistency for men regarding CHD events, but likely a minor adverse effect for both CHD mortality and total mortality in men, and nothing but adverse effects in women.

* Except for Minnesota, but in the only paper I had to (the other 3 weren’t even on PubMed) they didn’t discuss the diet in much detail, so I can’t be sure.

The Multifactorial Trials 

Some of the multifactorial trials found either no effect or an inconsistent effect.  While it’s impossible to know the outcome of these trials if these confounding variables weren’t present (would CHD event have been higher in DART, would total mortality have been higher in LA Vets and DART, it’s impossible to know), it suggests the results would have been less favourable for the DHH 

Supportive of DHH for CHD Events
Supportive of DHH for CHD Mortality
Supportive of DHH for Total Mortality
LA Vets

The trials that that more consistently supported the DHH were the most multifactorial.  These trials were Oslo, Finnish and STARS.  The Oslo experimental group had more whole foods, LCO3, CLO and vitamin D and less hydrogenated fish oil, refined grains and sugar.  Finnish had on balance more confounders in favour of the experimental group (see the post).  The STARS experimental group lost weight, had less trans fat, more LCO3, less baked goods and potentially more whole foods.

* The ACC experimental group was instructed to eat more whole plant foods and fish and less junk food and hydrogenated fats.  The LA Vets control group had a vitamin E deficient diet due to reusing the butter for cooking and had hydrogenated fats.  The DART experimental group increased fruits, vegetables and fish and decreased biscuits and cakes. 

Next week I’ll look at differences between men and women, the effect of PUFA on other things, the lower CHD events in the experimental group and a fresh look at the meta-analyses.


  1. Steven, brilliant job on this series!

    Why did you leave out this study? -

    1. Thanks. This series was about the trials replacing SFA with PUFA, whereas that trial didn't alter fat intake

      Low fat for CVD isn't a controversial issue, there's simply a lack of evidence in favour of them whether it's from RCTs (like the one you linked) or observational studies. It's a shame that this information isn't more well known and you can read stuff like "and all that fat would worry most experts"