Randomized trials substituting polyunsaturated fat for saturated fat and their effect on coronary heart disease (CHD). A closer look.

Results from randomized trials are often used to prove a causal relation between consumption of saturated fat and CHD. But can results from these trials be attributed to the changes in saturated fat intake, or could other dietary and non-dietary changes have contributed to these results?
Also, changes in serum cholesterol levels are often considered to be a reliable marker for subsequent changes in CHD risk. But can changes in serum cholesterol - which are caused by changes in consumption of dietary fat - predict subsequent changes in CHD in a reliable way?
Until now, no review tried to answer these questions.

I used articles about the randomized trials included by the systematic reviews from Mozaffarian D (2010) Ramsden CE (2010) and Skeaff CM (2009) [references 10-12, respectively] and collected data about changes in cholesterol caused by the interventions and looked for correlations with changes in CHD risk. Also, I summarized both the dietary, and non-dietary changes which occured during these trials and evaluated whether or not effects on CHD risk can be attributed to changes in saturated fat consumption in a reliable way. As always, more details about the data extracted from the original articles can be found in the extended tables on the bottom of this page.
It should be noted that there were possibly more publications presenting additive details about these trials, so the findings in the tables should be considered a starting point, and not as detailed findings from a systematic literature review.


Substitution of polyunsaturated fat for saturated fat and the predictive effect of serum cholesterol on coronary heart disease (CHD).

The dietary changes decreased levels of serum cholesterol in all 8 trials. The percentage change in cholesterol correlated fairly well with the predicted decrease in CHD risk in 6 trials (2-5, 7-8), though the predicted effect size on CHD was smaller in one trial (3), and larger in another one (8). In 2 trials, the decrease in serum cholesterol was followed by an increase in CHD rates (1, 6).
Some information should be taken into account when this data is interpreted:

  • 2) The LA Veterans Study Changes in cholesterol levels among the experimental subjects proved to be poorly correlated with adherence to the diet. The greatest difference between incidence rates of control and experimental subjects occured in the stratum with lowest adherence to the study diet. Also, within the experimental diet group alone, incidence rates were highest in the stratum with highest adherence to the study diet.
  • 5) The Oslo Diet-Heart Study The overall reduction in cholesterol levels correlated well with the overall reduction in CHD. But among subjects who died at age < 60, cholesterol levels were significantly higher among subjects who died from CHD, than among the survivors, while no significant differences in CHD mortality rates were found between these groups (36.5%, and 40.7% for the experimental and control group, respectively). Also, among subjects who died at age > 60, cholesterol levels did not differ significantly between subjects who died from CHD and the survivors, while significantly less subjects in the experimental group died from CHD (41.7%, and 53.0% for the experimental and control group, respectively).
  • 7) The DART Study The small significant decrease in serum cholesterol correlated with the small nonsignificant decrease in CHD risk. But the trial included 2 other dietary interventions. Increased fiber consumption was not related to serum cholesterol, while a nonsignificantly increased risk of CHD was found (RR = 1.23; 95% CI = 0.97-1.57). And increased fish consumption was related to a small significant increase of serum cholesterol, while a nonsignificant protective effect against CHD risk was found (RR = 0.84; 95% CI = 0.67-1.07).
  • 8) The STARS Included subjects in the trial only when their cholesterol levels decreased in response to use of the drug cholestyramine. In addition, further dietetic counseling, and if requested, suitable foodstuffs, were given to participants in the exerpimental group who did not achieve or maintain a cholesterol reduction of 15%.

Conclusion: Randomized studies consistently showed that dietary changes aimed at substitution of polyunsaturated fat for saturated fat decreased serum cholesterol levels. A global look at the subsequent effect on CHD suggests that these decreased levels of serum cholesterol, often correspond fairly well with the predicted decrease in CHD rates. But when more detailed findings are taken into account, little evidence remains that changes in serum cholesterol created by changes in dietary fat intake are predictive of CHD rates.

Table 1. Substitution of polyunsaturated fat for saturated fat and the predictive effect of changed serum cholesterol on CHD:
AuthorCohort nameDecrease of serum cholesterol in relation to dietary change
(difference between experimental group and control group)
Effect dietary change on CHD risk
8) Watts GF (1992)The STARS-12.2%RR = 0.41 (0.09-1.96) **
7) Burr ML (1989)The DART-3.5%RR = 0.91 (0.72-1.16)
6) Frantz ID Jr (1989)The Minnesota Coronary Survey-13.8%RR = 1.08 (0.84-1.37) **
5) Leren P (1970)The Oslo Diet-Heart Study-13.9%RR = 0.75 (0.57-0.99) **
4) Turpeinen O (1979)
Miettinen M (1983)
The Finnish Mental Hospital StudyWomen: -12.8%

Men: -15.5%
Women: RR = 0.64 (0.41-1.00) **

Men: RR = 0.55 (0.34-0.88) **
3) Report of a Research Committee (1968)MRC Soybean Oil- 13.5% *RR = 0.86 (0.61-1.22) **
2) Dayton S (1969)LA Veterans-12.7%RR = 0.74 (0.53-1.03) **
1) Rose GA (1965)Rose Corn Oil Trial-8.8% *RR = 1.86 (0.63-5.44) ***
* = Percentage from Ravnskov U. 1992 [9].
** = RR from Mozaffarian D. 2010 [10].
*** = RR from Ramsden CE. 2010 [11].

Can effects from randomized trials aimed at reducing intake of saturated fat solely be attributed to changes in intake of saturated fat?

No randomized trials exist where the only dietary difference existed of a reduction of saturated fat. Instead all randomized trials aimed at replacing saturated fats by polyunsaturated fats.
If effects from saturated fats on serum cholesterol and CHD do exist, these effects may be counterbalanced by the content - and possibly even the composition (e.g., amino acid profile of the protein) - of other nutrients in whole foods. Both diets rich in saturated fat and diets rich in polyunsaturated fats correlate with specific nutrient profiles. Therefore, substitution of polyunsaturated fats for saturated fats will also change the nutrient profile of the entire diet.
Until now, no review was created where both the dietary goals, and the actual dietary changes occuring during the trials were described.

An attempt was made to summarize both the dietary, and non-dietary changes which occured during these trials. They are presented in the following tables.

Dietary differences: Table 2 describes dietary differences - other than the decreased intake of saturated fat- occuring during the trial period between the experimental groups and the control groups. Detailed differences occurring during the trial period were provided by different reports about only two trials (4, 8), and to a lesser extend about two other trials (2, 3). Little to no data was provided about the other trials.
Results show that most diets from the experimental groups were higher in polyunsaturated fat intake and lower in dietary cholesterol intake. Also, results clearly show that the dietary changes in fat intake correlated to lots of other changes in dietary intakes of both foods and nutrients.
Nondietary differences: Table 3 describes nondietary changes occuring during the trial period between the experimental groups and the control groups. Detailed findings were provided about two trials only (4, 8).
Both trials were seriously contaminated:
-In the Finnish Mental Hospital Study (4), two psychotropic drugs were used more abundantly in the control group. And during the first half of the follow-up period, the more chronic cases were transferred from the hospital of the control group to another hospital. Ramsden CE (11) pointed out that one of these drugs 'thioridazine' is significantly associated with risk of sudden death (OR = 5.3; 95% CI = 1.7-16.2; P = 0.004). Also, this drug causes T-wave distortions, QRS changes, ST elevations and other electrocardiogram changes and clinical presentations both with therapeutic administration and overdoses. These electrocardiogram changes and clinical presentations overlap with those seen in MI and sudden death and may have been counted as CHD events. Furthermore, the drug may also induce weight gain.
-In the STARS Study (8), overweight patients in the experimental group only, were put on a co-intervention aimed at losing weight. And patients in the control group who died had severe CHD at entry.

Results from which trials can not only be attributed to the effect of subtituting polyunsaturated fats for saturated fats?

  • 8) The STARS Study In addition to the two nondietary effects mentioned above, the experimental group consumed several foods/nutrients which could have favourably influenced the effect on CHD (13): more fruits, vegetables, and EPA/DHA.
  • 5) The Oslo Diet-Heart Study All control groups consumed trans fats, usually varying from 1.3-2.4 g/day. But in this control group trans fat consumption was exceptionally high: Ramsden CE (11) pointed out that subjects in the control group consumed 65 g/d of partially hydrogenated fish and vegetable oil margarine, delivering 9.6 g trans fat/day. While these foods were restricted to the experimental group. In addition, the experimental group consumed several foods/nutrients which could have favourably influenced the effect on CHD (13): more vegetables, fruits, nuts, fish, and whole grains
  • 4) The Finnish Mental Hospital Study In addition to the two nondietary effects mentioned above, diets from the experimental group and the control group differed not mildly, but distinctly for a large amount of dietary items during the first half of the follow-up period (Turpeinen O. 1968). But the most important reason for exclusion would be, that the study was not randomized!
  • 2) The LA Veterans Study The control group was distinctly deficient in vitamin E. Dietary vitamin E has been linked to CHD (13)

The remaining four trials (1, 3, 6, 7) were most likely to isolate the effect of replacing saturated fats by polyunsaturated fats. The RR was < 1 in two trials (3, 7), and RR's were > 1 in the other two trials (1, 6). Of these four trials, one aimed at reducing dietary cholesterol (6), and a reduction in dietary cholesterol intake during the trial period was found in two other cohorts (3, 7). Though dietary cholesterol was not mentioned in the fourth cohort (1), the dietary goals will most likely have led to a reduction in intake. Polyunsaturated fat intake increased in all 8 trial included in this analysis.
Ramsden CE (11) extensively examined data about specific dietary fats and contacted authors to obtain additional information. He found that non-hydrogeneted study oils were substituted for trans fatty acids-containing fats, oils and foods (e.g., common hard margarines, shortenings, pastries, fried foods in all eight trials in this analysis.

Conclusion: Various confounders were able to influence the effect of replacing saturated fats by polyunsaturated fats. Sometimes, experimental groups were randomized to a complete mediterranean diet, including increased consumption of vegetables and fruits. The effect of reduced trans fat intake in the experimental groups, is likely to have attributed to the protective effects against coronary heart disease in all 8 trials. Saturated fat was replaced by polyunsaturated fat in all trials, and dietary cholesterol intake correlated strongly to saturated fat intake. Subsequently, none of the trials involved in the analysis was able to isolate the effect from saturated fats on CHD.

Table 2. Substitution of polyunsaturated fat for saturated fat and correlations with other dietary changes:
AuthorCohort nameDietary changes other than a decrease in saturated fat
8) Watts GF (1992/1996)The STARSDietary intervention:
-Reduced intakes of animal protein, meat, cheese, fish, margarine, oils, cookies, pastry, and cakes.
-Increased intakes of bread, potatoes, pasta, vegetables, fruits, legumes, and oats.
-Limit intake of alcohol.

Dietary changes during trial:
-Reduced intakes of energy, total fat, monounsaturated fat, trans fat, alpha-linolenic acid, and dietary cholesterol.
-Increased intakes of linoleic acid, EPA, DHA, carbohydrate, and fiber.
7) Burr ML (1989)The DARTDietary intervention:
-Reduced intake of total fat.
-Increased intake of polyunsaturated fat.

Dietary changes during trial:
-Reduced intakes of total fat, trans fat*, and dietary cholesterol.
-Increased polyunsaturated fat:saturated fat ratio.
6) Frantz ID Jr (1989)The Minnesota Coronary SurveyDietary intervention:
-Reduced intake of dietary cholesterol.
-Increased intake of polyunsaturated fat, liquid corn oil*, corn oil polyunsaturated margarine*, and possibly safflower oil*.

Dietary changes during trial:
-Reduced intake of trans fat*.
5) Leren P (1968/1970)The Oslo Diet-Heart StudyDietary intervention:
-Reduced intake of dietary cholesterol, refined grains*, and sugar*.
-Increased intakes of polyunsaturated fat, soya bean oil, cod liver oil*, fish*, shellfish*, nuts*, fruits*, vegetables*, and whole grains*.

Dietary changes during trial:
-Reduced intake of trans fat*.
-Increased intake of linoleic acid*, linolenic acid*, EPA/DHA*, and vitamin D*.
4) Turpeinen O (1979)
Miettinen M (1983)
The Finnish Mental Hospital StudyDietary intervention:
-Reduced intakes of dairy fat, butter, ordinary margarine.
-Increased intakes of soybean oil, soft margarine, polyunsaturated fat.

Dietary changes during trial:
-Reduced intakes of potatoes, cream, butter, common margarine, meat (products), eggs, trans fat*, dietary cholesterol and sucrose.
-Increased intakes of root vegetables, buttermilk, monounsaturated fat, total polyunsaturated fat, linolenic acid, and linolenic acid.
3) Report of a Research Committee (1968)MRC Soybean OilDietary intervention:
-Reduced intakes of butter, other margarines, cooking-fat, other oils, fat meat, whole milk, cheese, egg yolk, and most biscuits and cakes.
-Increased intake of soya bean oil and fruit juice

Dietary changes during trial:
-Reduced intakes of trans fat*, dietary cholesterol, calcium, and vitamin A.
2) Dayton S (1969)LA VeteransDietary intervention:
-Reduced intakes of butterfat and meat fat.
-Increased intakes of corn oil, soybean oil, safflower oil, cottonseed oil, and unsaturated margarine

Dietary changes during trial:
-Reduced intakes of total fat, trans fat*, and dietary cholesterol .
-Increased intakes of linoleic acid, alpha-linolenic acid, and plant sterols.

Note:
The control group was distinctly deficient in vitamin E.
1) Rose GA (1965)Rose Corn Oil TrialDietary intervention:
-Reduced intakes of fried foods, fatty meat, sausages, pastry, ice-cream, cheese, cakes, milk, eggs, and butter.
-Increased intake of corn oil.

Dietary changes during trial:
-Reduced intake of trans fat*.
-Increased intake of corn oil.
* = Data from Ramsden CE. 2010 [11].
Table 3. Substitution of polyunsaturated fat for saturated fat and correlations with non-dietary changes:
AuthorCohort nameNon-dietary changes differing between the experimental group and the control group
8) Watts GF (1992)The STARS-Overweight patients in the experimental group, but not control group, were described a diet that contained 1,000-1,200 kcal/day.
-No significant differences in the demographic and clinical characteristics were found, but subjects in the control group tended to be older, drink more alcohol, and were more likely to have angina. On the other hand subjects in the control group were less likely to have previous MI.
-The patients in the control group who died had severe CHD at entry.
7) Burr ML (1989)The DARTNo differences were found.
6) Frantz ID Jr (1989)The Minnesota Coronary SurveyNo data was provided.
5) Leren P (1970)The Oslo Diet-Heart StudyNo differences were found.
4) Miettinen M (1972)
Turpeinen O (1979)
The Finnish Mental Hospital Study-Two psychotropic drugs were used somewhat less abundantly in the experimental group: 0.63 vs 0.97 dosis/day for thioridazine, and 1.61 vs 1.92 dosis/day for total phenothiazines.
-During the first half of the follow-up period, the more chronic cases were transferred from the hospital of the control group to another hospital.
3) Report of a Research Committee (1968)MRC Soybean OilNo differences were found.
2) Dayton S (1969)LA VeteransBody weight declined in the control group, while it rose in the experimental group.
1) Rose GA (1965)Rose Corn Oil TrialNo data was provided.

References.

1) Rose GA. Corn oil in treatment of ischaemic heart disease. Br Med J. 1965 Jun 12;1(5449):1531-3. Full text
2) Dayton S. Composition of lipids in human serum and adipose tissue during prolonged feeding of a diet high in unsaturated fat. J Lipid Res. 1966 Jan;7(1):103-11. Full text
+ Dayton S. Controlled trial of a diet high in unsaturated fat for prevention of atherosclerotic complications. Lancet. 1968 Nov 16;2(7577):1060-2. Abstract
+ Dayton S. Diet high in unsaturated fat. A controlled clinical trial. Minn Med. 1969 Aug;52(8):1237-42. Abstract
+ Dayton S. A controlled clinical trial of a diet high in unsaturated fat in preventing complications of atherosclerosis. Circulation. 1969;40(suppl 2):1-63. Abstract
+ Dayton S. Diet and atherosclerosis. Lancet. 1970 Feb 28;1(7644):473-4. Abstract
3) Report of a research committee to the medical research council. Controlled trial of soya-bean oil in myocardial infarction. Lancet. 1968 Sep 28;2(7570):693-9. Abstract
4) Turpeinen O. Dietary prevention of coronary heart disease: long-term experiment. I. Observations on male subjects. Am J Clin Nutr. 1968 Apr;21(4):255-76. Full text
+ Miettinen M. Effect of cholesterol-lowering diet on mortality from coronary heart-disease and other causes. A twelve-year clinical trial in men and women. Lancet. 1972 Oct 21;2(7782):835-8. Abstract
+ Turpeinen O. Dietary prevention of coronary heart disease: the Finnish Mental Hospital Study. Int J Epidemiol. 1979 Jun;8(2):99-118. Abstract
+ Miettinen M. Dietary prevention of coronary heart disease in women: the Finnish mental hospital study. Int J Epidemiol. 1983 Mar;12(1):17-25. Abstract
5) Leren P. The effect of plasma-cholesterol-lowering diet in male survivors of myocardial infarction. A controlled clinical trial. Bull N Y Acad Med. 1968 Aug;44(8):1012-20. Full text
+ Leren P. The Oslo diet-heart study. Eleven-year report. Circulation. 1970 Nov;42(5):935-42. Full text
6) Frantz ID Jr. Test of effect of lipid lowering by diet on cardiovascular risk. The Minnesota Coronary Survey. Arteriosclerosis. 1989 Jan-Feb;9(1):129-35. Full text
7) Burr ML. Diet and reinfarction trial (DART): design, recruitment, and compliance. Eur Heart J. 1989 Jun;10(6):558-67. Abstract
+ Burr ML. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: diet and reinfarction trial (DART). Lancet. 1989 Sep 30;2(8666):757-61. Abstract
+ Ness AR. The long-term effect of dietary advice in men with coronary disease: follow-up of the Diet and Reinfarction trial (DART). Eur J Clin Nutr. 2002 Jun;56(6):512-8. Full text
8) Watts GF. Effects on coronary artery disease of lipid-lowering diet, or diet plus cholestyramine, in the St Thomas' Atherosclerosis Regression Study (STARS). Lancet. 1992 Mar 7;339(8793):563-9. Abstract
+ Watts GF. Dietary fatty acids and progression of coronary artery disease in men. Am J Clin Nutr. 1996 Aug;64(2):202-9. Abstract
9) Ravnskov U. Cholesterol lowering trials in coronary heart disease: frequency of citation and outcome. BMJ. 1992 Jul 4;305(6844):15-9. Full text
10) Mozaffarian D. Effects on coronary heart disease of increasing polyunsaturated fat in place of saturated fat: a systematic review and meta-analysis of randomized controlled trials. PLoS Med. 2010 Mar 23;7(3):e1000252. Full text
11) Ramsden CE. n-6 fatty acid-specific and mixed polyunsaturate dietary interventions have different effects on CHD risk: a meta-analysis of randomised controlled trials. Abstract
12) Skeaff CM. Dietary fat and coronary heart disease: summary of evidence from prospective cohort and randomised controlled trials. Ann Nutr Metab. 2009;55(1-3):173-201. Abstract
13) Mente A. A systematic review of the evidence supporting a causal link between dietary factors and coronary heart disease. Arch Intern Med. 2009 Apr 13;169(7):659-69. Full text