Sunday, June 14, 2015

The Lyon Diet Heart Study

The meta-analysis by Schwingshackl and Hoffmann included the Lyon Diet Heart Study in their meta-analysis of fat modification trials but not in a separate analysis of SFA vs. PUFA trials.  But besides that, I thought it would be interesting to read the Lyon Diet Heart Study because it’s a trial that many people cite as evidence for different things, such as: ‘SFA is bad’, ‘omega 6 PUFAs are bad’, ‘MUFA is good’, ‘Mediterranean diet is good’; ‘LDL-C is irrelevant’.  For example:

Studies Associated with the Trial

Mediterranean alpha-linolenic acid-rich diet in secondary prevention of coronary heart disease (1994) [1]
Cretan Mediterranean diet for prevention of coronary heart disease (1995) [2]
Effect of a Mediterranean type of diet on the rate of cardiovascular complications in patients with coronary artery disease. insights into the cardioprotective effect of certain nutriments (1996) [3]
Control of bias in dietary trial to prevent coronary recurrences: the Lyon Diet Heart study (1997) [4]
Mediterranean dietary pattern in a randomized trial: prolonged survival and possible reduced cancer rate (1998) [5]
Mediterranean diet in secondary prevention of coronary heart disease (1998) [6] (no URL)
Mediterranean diet, traditional risk factors, and the rate of cardiovascular complications after myocardial infarction: final report of the Lyon Diet Heart study (1999) [7


605 people (~90% male) under 70 years old who had previously had a heart attack were randomised to a high (~0.6% TE) alpha-linolenic acid (ALA) Mediterranean diet (N = 302) or to the usual post-infarct prudent diet (N = 303).  There were no significant difference between the groups at baseline [1].

The Mediterranean diet used in the trial was based on the Cretan diet from the Seven Countries Study and is described as being high in ALA and ‘rich in vegetables and fruits’ [1].  The dietary advice included “more bread, more root vegetables and green vegetables, more fish, less meat (beef, lamb, and pork to be replaced with poultry), no day without fruit, and butter and cream to be replaced with margarine supplied by the study.”  The patients didn’t accept olive oil so canola oil based margarine was used instead.  This margarine was mostly MUFA, but also was relatively high in ALA and also contained 5.4% TFA and was to be a major fat source.  “Moderate alcohol consumption in the form of wine was allowed at meals. At each subsequent visit of the experimental patients, a dietary survey and further counselling were done by the research dietician. Diet evaluations comprised a 24-hour recall and a frequency questionnaire.” [1] [2]

“Control patients received no dietary advice apart from that of hospital dieticians or attending physicians” [1] [2]


Besides replacing butter and cream with margarine, there weren’t really any major differences in reported food consumption [1] [2]

Again, there are lots of differences in nutrient intake, but nothing major.  Regarding fatty acids, MUFA and omega 3 are higher while SFA and omega 6 PUFA are lower [1] [2] [7].  While SFA is lower, it isn’t discussed much and seems to just be incidental.  Some nutrients weren’t included in the table below so I included them in another table [1] [2]


Control Group
Experimental Group
0.69 ± 0.05
0.66 ± 0.02
Vitamin A (ug)
548 ± 111
279 ±72
Vitamin C (mg)*
101 ± 4
118 ± 4
Vitamin D (ug)
2.8 ± 0.6
1.6 ± 0.2
Vitamin E (ug)*
13.6 ± 0.5
12.1 ± 0.3
Carotene (ug)
5539  ± 413
6478 ± 492
* Significant differences

Differences in fatty intake were consistent with differences in circulating fatty acids.  Note that there is no difference in ‘18:1 Trans’ and that EPA is higher in the experimental group, due to the higher ALA intake (ALA to EPA is about 5% efficient, ALA to DHA is much lower) [1] [2]

 Surprisingly there were no differences (or even a hint of difference) between the groups regarding standard CHD risk factors such as blood lipids, blood pressure and BMI.  The only significant difference is a higher concentration of ascorbic acid at week 52 (unfortunately baseline data isn’t known) [1] [2]

Despite the lack of major dietary differences or differences in blood lipids, the experimental group had much fewer CHD and CVD events/deaths (note the low risk ratios) [7]

The incidence of cancer was also lower (7 vs. 17, adjusted RR = 0.39), but cancer mortality was similar (3 vs. 4).  Of these 24 cancers 6 were lung cancer, and smokers were much more likely to develop cancer [5]

The results of from the Lyon Diet Heart Study are very impressive despite minor differences in diet and no differences in standard CHD risk factors.  While there many statistically significant differences in dietary intake, neither of them of are particularly noteworthy (only differing by ±~10-20%).  This has led other researchers to question the results [8].  I am somewhat suspicious too and think that luck played some role

But let’s assume the results are an accurate reflection of the dietary intervention, particularly given that even if the low risk ratios are due to chance, the upper limits of the confidence intervals would still predict a pretty decent dietary effect.  What then is the take home message of this study?

The authors suggest the nutrients involved were increases in vitamins C and E, lower omega 6 and higher omega 3 and possibly a higher intake of certain amino acids, micronutrients and phytonutrients although the latter three weren’t measured [3].  Due to the multifactorial nature of the intervention it is clearly inappropriate to use this trial as evidence for specific foods or nutrients such as SFA, MUFA and PUFAs.  I think this trial is a good example that there is more to CHD than the standard risk factors such as HDL-C, LDL-C, BMI and blood pressure (as you don’t need changes in them to have an effect), but it shouldn’t be used to argue that those factors are irrelevant.  Perhaps it’s evidence that many minor differences can add up and produce a substantial effect, but I’m not sure about the degree of that.

* There were slightly fewer withdrawals in the experimental group (30 vs. 39) [1]

** The trial was single blinded.  The researchers made measures to check for bias and didn’t find any (see papers for more info).  Drug use was similar between groups at baseline and after the first year, suggesting against physician bias [1] [3] [4

*** FYI: for all the pro-omega 6 people who cite the Lyon Diet Heart Study, the rationale of the trial is essentially that the authors noted some failed attempts of replacing SFA with omega 6 PUFA (Minnesota Coronary Survey and DART) and their previous studies suggested some potential harm for increasing PUFAs (enhanced platelet aggregation and lipid peroxidation) [1].  Also, the lead author challenges the lipid hypothesis and the efficacy of statins

No comments:

Post a Comment