Sunday, December 29, 2013

The Saint Vincent's Hospital Study

Studies Associated with the Trial

Blood Cholesterol and Coronary Heart Disease (1963) (no access)
The Effects of Two Low Fat Dietary Patterns on the Blood Cholesterol Level of Young Male Coronary Patients (1963) [1]
Modified-fat dietary management of the young male with coronary disease. A five-year report (1967) [2]
The 5-Year Experience of Modified Fat Diets on Younger Men with Coronary Heart Disease (1970) [3]
Ten-year experience of modified-fat diets on younger men with coronary heart-disease (1973) [4]
 
The 'Pilot Study'
 
100 men, aged 20-50 with confirmed myocardial infarction, were randomised to one of two diet groups.  Both groups were placed on a diet where 28% of calories came from fat.  “The fat content was achieved by eliminating certain dairy products, rich desserts and pastries, certain fried foods, and fatty meats” [2]
 
“Diet 1 contained 1 oz of a 50% mixture of corn-safflower oil”
“Diet 2 contained 1 oz of a 50% mixture of coconut-peanut oil” 
 
There were a few other differences between the diets: Diet 1 was allowed 8 fish and seafood meals, whereas Diet 2 was only allowed 1; and Diet 1 had a special margarine that was higher in PUFA, whereas Diet 2 had a regular margarine (there may have been differences in TFA, but I can’t tell) [1]
 
All overweight participants (73) were placed on a standard 1,200 calorie weight loss diet.  After the desired weight was attained they were randomly placed on their assigned diet [2].
 
[2]
Diet 1
Corn and Safflower
Diet 2
Coconut and Peanut
Protein (%)
20.4
20.7
Fat (%)
SFA/MUFA/PUFA
P/S Ratio
27.8
5.5/9.3/14.1
3:1
28.2
9.6/13.2/3.3
1:3
Carbohydrates (%)
51.8
51.1
Cholesterol (mg)
260
395
 
After 12 months both groups had a minor decrease in cholesterol but the difference between the groups was not significant, which wasn’t consistent with metabolic ward studies.  Thinking that the issue may have been compliance, the researchers gave the men frozen meals.  After another 12 months (2 years) the pilot study ended due to a lack of difference in cholesterol levels between the groups [2]

A New Comparison
 
The researchers then merged the two diet groups and added a matched, non-dietary managed control group of 100 participants, with fairly similar risk factors, but the control group had more heavy drinkers, smokers and heavy smokers.  This made the remainder of the study non-randomised.  None of the overweight members of the control group were placed on the 1,200 calorie weight loss diet [2] [3]
 
After 5 years the study group had an average 10% (24 mg/dl) drop in total cholesterol, whereas the cholesterol level in the control group didn’t significantly change.  There was still no significant difference in cholesterol between the two diet groups despite good dietary adherence [2] [3]
 
Interestingly, while the average cholesterol level in the control group stayed pretty constant, the cholesterol levels in the participants who died in the control group actually slightly increased, whereas the cholesterol level in the participants who survived slightly decreased.  The differences weren’t significant [2]
 
 
The study group had slightly more myocardial infarctions and slightly fewer deaths, although since the years of experience in the study group was much greater than the control group, the incidence rates per years of experience was much lower in the study group (37.7% fewer myocardial infarctions and 57.1% fewer deaths [2] 

Of the 9 deaths in the study group, 5 came from the corn + safflower oil group and 4 came from the coconut + peanut oil group, but we don’t know what the causes of death were or the differences in myocardial infarctions between the diet groups [2] 

 
However, in addition to the weight loss in the study group, smoking was another confounding variable.  Smokers had approximately twice the incidence rate of non-smokers.  The control group already had more smokers at baseline and the difference became more pronounced at the end of the study.  After 5 years there was a decrease in smokers in the study group (from 37 to 29) and an increase in the control group (from 47 to 59) [3]
 
It’s a shame, what started as a fairly good (but probably underpowered from number and age) RCT testing the diet heart hypothesis ended up as a basic, multifactorial diet intervention study, where one can’t really tease apart the effects of smoking, weight loss and perhaps other dietary factors from dietary fat reduction.  But as for the diet heart hypothesis, the researchers found that “the degree of unsaturation of the diet did not appear to influence serum cholesterol value or mortality” [2]

None of the meta-analyses included this trial.  While it started off testing the diet heart hypothesis, it ended up non-randomised, testing a broad dietary intervention with weight loss, was underpowered to compare differences between the two diet groups (which were randomised) and didn't report the causes of death or incidence of myocardial infarction in the two diet groups.
 
* Peanut oil is “unexpectedly atherogenic” in animal models of atherosclerosis, which seems to be due to peanut lectin [5]
 
** After a 5 year follow up period (10 years total) there were still significant reductions in cholesterol in the study group compared to the control group and 42.9% fewer deaths in the study group (16% vs. 28%) [4]

Sunday, December 22, 2013

The Oslo Diet Heart Study

Studies Associated with the Trial

The effect of plasma cholesterol lowering diet in male survivors of myocardial infarction. A controlled clinical trial (1966) [1] (no access)
The effect of a cholesterol lowering diet in male survivors of myocardial infarction. (A controlled clinical trial) (1967) [2] (in Norwegian)
The effect of plasma-cholesterol-lowering diet in male survivors of myocardial infarction. A controlled clinical trial (1968) [3]
The Oslo Diet-Heart Study: Eleven-Year Report (1970) [4]
Prevention of coronary heart disease: some results from the Oslo secondary and primary intervention studies (1989) [5] (no access) 

Participants and Diets 

412 men aged 30 to 64, who had previously had a heart attack, were randomised to two groups: an experimental group and a control group.  The control group consumed their regular diet, while the experimental group was instructed to consume a diet low in animal fats and cholesterol and high in vegetable oils.  Dietary information from only 17 participants in the experimental group was reported.  The groups were similar at baseline, except the experimental group had slightly higher blood pressure [3] 

The papers I had access to only had pretty basic information on the diet, which came from only 17 participants in the experimental group 

Energy (kcal)
2400
Protein (g)
92
Fat (g)
SFA/MUFA/PUFA (g)
P/S Ratio
104
22/27/55
2.4
Carbohydrate (g)
Sugar (g)
269
51
Dietary Cholesterol (mg)
264

Results 

Cholesterol levels decreased in the experimental group [3]


The number of major CHD relapses can be calculated using the table below from the total number of CHD relapses - angina.  CHD mortality can also be calculated using the table below from the number of fatal myocardial infarctions (10 vs. 23) and the number of sudden deaths (27 vs. 27) [3].  CVD mortality and total mortality are reported more simply in a paper [4] and the results are presented in my table below. 



Experimental
Control
Major CHD Events [3]
70
91
Total CHD Events (P = 0.011*) [3]
80
120
CHD Mortality [3]
37
50
CVD Mortality (P = 0.09) [4]
38
52
Total Mortality (P = 0.13) [4]
41
55

At the end of the 5 year diet period: “the surviving dieters were advised to adhere to the cholesterol lowering diet in the future. The surviving controls were informed that a reduced fat intake possibly might be beneficial, but they received no detailed dietary instructions.”The men were followed for an additional 6 years to track mortality [4].  After 11 years the mortality is as follows:

 [4]
Experimental
(n=206)
Control
(n=206)
Total Cardiovascular Mortality (p=0.13)
88
102
Other Mortality
13
6
Total Mortality (p=0.35)
101
108

So far, this looks to be one of the more favourable studies for the diet heart hypothesis, but even though I don’t have access to the data other people do.  The meta-analysis by Ramsden, et al discussed how the Oslo Diet Heart Study isn’t a simple omega 6 vs. SFA trial, but rather is multifactorial in favour of the experimental diet:

  • The experimental diet was higher in omega 3.  Meats and eggs were replaced with fish, shellfish and ‘whale beef’.  The experimental group was supplied with considerable quantities of Norwegian sardines canned in cod liver oil, which proved to be popular as a bread spread”.  And also soybean oil has some plant omega 3 
  • The experimental diet was higher in vitamin D, which is a more important source of vitamin D for Norwegians.  The fish and cod liver oil supplied 610 IU of vitamin D 
  • The experimental group was encouraged to eat more nuts, fruits, and vegetables; to limit animal fats; and to restrict their intake of refined grains and sugar 
  • The experimental diet was lower in trans fats.  Paraphrasing from Ramsden, et al: in the two decades before the ODHS, Oslo males had a 7-fold increased incidence of an initial heart attack, which coincided with a rapid rise in partially hydrogenated fish and vegetable oil margarines.  In the control group, about 6.9% of total calories came from these margarines, whereas they were “entirely restricted” in the experimental group

With the trial being so multifactorial in favour of the experimental group I don’t think one should use it as evidence for the diet heart hypothesis or for recommendations for omega 6 to replace SFA.

Sunday, December 15, 2013

The Medical Research Council Trial

Studies Associated with the Trial

Controlled trial of soya-bean oil in myocardial infarction: report of a research committee to the medical research council (1968) [1]
Dietary Aspects of a Controlled Trial of Soya-bean Oil in Myocardial Infarction (1969) [2]
Cancer among men on cholesterol-lowering diets: Experience from five clinical trials (1971) [3]
Letter: Are PUFA harmful? (1974) [4] 

Participants and Diets 

393 men who recently had a heart attack were randomised to two groups: an experimental group and a control group.  The groups were similar at baseline [1] 

The experimental group was asked to eat a diet low in SFA and take 85g of soybean oil daily, where “at least 43 g. of soya-bean oil daily had to be taken unheated, and it was often drunk with fruit juice”.  In addition “up to 35 g. of other fat per day was also allowed. 14 g. of this was taken as a moderately unsaturated margarine (’ Blue Band ’). Foods allowed daily included lean meat (up to 85 g.), any fish (which was encouraged), skimmed milk, and clear soups. Foods forbidden included butter, other margarines, cooking-fat, other oils, fat meat, whole milk, cheese, egg yolk, and most biscuits and cakes.” [1] [2] 

The control group remained on their regular diet, although they did seem to reduce fried foods after having their heart attack [1] 

73 in the experimental group and 90 in the control group were placed on a reduced carbohydrate diet to lose weight.  5 in the experimental group and 6 in the control group were placed on a very low calorie diet (1,000 calories) for weight loss. [1] 

On average, the experimental group consumed 80g of soybean oil per day and 40g from other fats.  The SFA:PUFA ratio was 1:2 in the experimental group and 6:1 in the control group.  Calories, protein, carbohydrate and sugar intakes were almost identical, and the experimental group was lower in cholesterol (258mg vs. 588mg) and slightly lower in calcium (808 mg vs. 930 mg) and vitamin A (by 1,000 IU) [1] 

Results 

Linoleic acid levels in adipose tissue were on average 19.0% among those in the experimental group with good adherence compared to 9.6 with poor adherence [2] 

Total Cholesterol (mg/dl)
Experimental
Control
Baseline
272
273
6 Months
213
259
6 Years
239
269
[1] 

The end-points in the trial were relapses, which included things like angina, heart attacks and deaths from cardiovascular disease.  There were no significant differences between the groups.  The difference in first relapses was due to non-major relapses like angina.  Unfortunately the number of major relapses wasn't reported, instead the number of 'men who had a major relapse at any time during trial' was, which is the closest measure and was slightly lower in the experimental group 

Experimental
Control
Number of men
199
194
First relapses
62
74
Major
40
39
Fatal
15
14
Non-fatal definite reinfarctions
25
25
Other non-fatal relapses
22
35
Men who had a major relapse at any time during trial
45
51
Fatal
25
25
Non-fatal definite reinfarctions
20
26
All deaths from cardiovascular disease
27*
25
* “1 death was from subarachnoid haemorrhage and 1 from cerebral embolism” [1] 

There were also 6 deaths from cancer in the control group compared to only 1 in the experimental group.  This resulted in total mortality being slightly lower in the experimental group (28 vs. 31) [1] [3]

It’s interesting that the experimental group had lower total CHD events, but not major CHD events and CHD mortality.  The authors suggested that: “One possibility is that the men in the control group themselves may have felt that they were having less active treatment, and might have been psychologically motivated both to complain more and more readily of cardiac symptoms. The other possibility, of course, is that the diet may have had a real effect in these cases of minor ischaemia.” [1]

Sunday, December 8, 2013

The Los Angeles Veterans Administration Trial

Studies Associated with the Trial

A controlled clinical trial of a diet high in unsaturated fat. Preliminary observations (1962) [1]
A palatable diet high in unsaturated fat (1962) [2] (no access)
Effect of long-term feeding of various fats on whole blood clotting times in men (1964) [3] (no access)
Vitamin E status of humans during prolonged feeding of unsaturated fats (1965) [4] (no access)
Influence of a diet high in unsaturated fat upon composition of arterial tissue and atheromata in man (1965) [5]
Composition of lipids in human serum and adipose tissue during prolonged feeding of a diet high in unsaturated fat (1966) [6]
Adipose tissue linoleic acid as a criterion of adherence to a modified diet (1967) [7]
A Controlled Clinical Trial of a Diet High in Unsaturated Fat in Preventing Complications of Atherosclerosis (1968 [8] (abstract only)
Controlled trial of a diet high in unsaturated fat for prevention of atherosclerotic complications (1968) [9]
Trial of unsaturated-fat diet (1968) [10] (letter to the editor)
Diet and cardiovascular diseases (1969) [11] (letter to the editor)
Prevention of coronary heart disease and other complications of arteriosclerosis by modified diet (1969) [12]
Diet high in unsaturated fat. A controlled clinical trial (1969) [13]
Diet and atherosclerosis (1970) [14] (letter to the editor)
Incidence of cancer in men on a diet high in polyunsaturated fat (1971) [15]
Increased prevalence of cholelithiasis in men ingesting a serum-cholesterol-lowering diet (1973) [16] 

Participants and Diets 

846 male veterans living in a domiciliary unit were randomised to a control group or an experimental group (high PUFA diet).  “Volunteers were accepted into the study regardless of possible pre-existing atherosclerotic complications.”  But were excluded if “they were known to be alcoholic, if they had presented behavior problems, if they were on therapeutic diets, if there were problems of communication, if they had disease likely to result in death within five years or if they were below fifty-five years of age” [1] 

All the subjects ate in a large facility with two dining rooms that served either the control or the experimental diet.  Adherence was tracked by the proportion of meals taken from the correct dining hall.  The meals were ad libitum [1] 

I don’t have access to the detailed data on the diet, which seems to be in this study [2], but it has been briefly described in others: 

“The control diet was a conventional food pattern containing 40 per cent fat calories, mostly of animal origin.” [1] 

“The design of the experimental diet involved substitution of vegetable oils for about two thirds of the animal fat, total fat content being kept about 40 per cent.”.  Eggs were restricted to seven per week.  “Vegetable oils were incorporated into the diet in the form of filled milk, imitation ice cream, "unsaturated" margarine, special frankfurters and filled cheeses. Vegetable oils were used liberally in cooking and baking. Meat fat was minimized by the use of lean cuts, specially trimmed” [1] 

Control
Experimental
Energy (kcal)
2400
2425
Protein (g)
94
95
Fat (g)
108
106
Fat (% of total calories)
40.5
39.3
Linoleic acid (% of fatty acids)
10.0
38.4
Linoleic Acid (% of total calories)
4.0
15.0
PUFA:SFA (ratio)
0.3
1.7
Cholesterol (mg)
721
365
Other Sterols (mg)
150
604
[1] [6] 

Micronutrient intake was not reported except for α-tocopherol (the main form of vitamin E), which was 9.4-fold higher in the experimental group (22.6mg vs. 2.4mg) [5] and only 16.0% of the RDA in the control group [101], indicating a very inadequate intake and a substantial confounding variable in favour of the experimental group.  While vegetable oils tend to be richer sources of vitamin E than animal fats, this is insufficient to explain the 9.4-fold difference between the groups and the very low intake of α-tocopherol in the control group.  An explanation may lie in [4] and while I don't have access Chris Masterjohn does, and he mentioned that: “The researchers took care not to reuse the vegetable oil after cooking but took no such precautions with the butter, resulting in butter that was very deficient in vitamin E”, which makes sense as the nutrient analysis was conducted after the food had been cooked

* 1 mg of α-tocopherol = 1.5 IU of vitamin E

Results

Baseline cholesterol levels were 231 in the experimental group and 232 in the control group.  After 5 years the average cholesterol levels decreased by 20% in the experimental group and 10% in the control group.  Some of the decrease in the control group could be attributed to age. [6] 

In the experimental diet, linoleic acid made up 40% of the total fatty acids.  As a result, among people with good adherence (≥ 80%), the average proportion of linoleic acid in adipose tissue increased from 11% to 32% at 5 years and was projected to increase to ~35%, which seems consistent with proportion of fat from linoleic acid and adherence in this sample (~90%) [6].  At 5 and 6 years there was a correlation between the linoleic acid concentration in adipose tissue and adherence to the experimental diet [7].  They also calculated that the half-time of fatty acid turnover in adipose tissue was 680 days (in other words, 50% of the fatty acids in adipose tissue are turned over after 680 days) [6]

[6]

The primary end-points were sudden death and myocardial infarction (coronary heart disease (CHD) events), and the secondary endpoints included other cardiovascular events.  Both total CHD and total CVD events were lower in the experimental group [9] [12] [13].


Experimental
Control
Definite myocardial infarction by ECG only
9
4
Definite myocardial infarct, overt
33
47
Sudden death
18
27
Definite cerebral infarct
13
25
Ruptured aneurysm
2
5
Amputation for gangrene
7
5
Miscellaneous
3
6
Total coronary heart disease events
60
78
Total cardiovascular events
85
119


CHD mortality and CVD mortality were both lower in the experimental group, with the difference between the groups being significant for CVD mortality (referred to as 'deaths due to atherosclerotic events') [8] [9] [12] [13]


Experimental
Control
Acute myocardial infarct
24
24
Sudden death
18
27
Cerebral infarct
3
9
Amputation
0
3
Ruptured aneurysm
2
5
Intestinal infarct
1
2
Mortality from coronary heart disease
42
51
Mortality from cardiovascular disease
48
70


However, total mortality was very similar, though the actual figure is a  bit ambiguous.  It has been reported as 177 vs. 174 (page 757 [12]) and 174 vs. 178 (page 464 [15]).  Figure 3 in [9] shows it to be higher in the experimental group, while figure 4 in [13] shows it to be lower.  Also, two meta-analyses (Skeaff & Miller, and Hooper, et al) have reported it as 174 vs. 177, but I'm not sure where that comes from and may just be a mistake (switching around 177 vs. 174).

Most of the difference in non-CVD mortality was due to differences in cancer mortality, which was higher in the experimental group (31 vs. 17, p = 0.06) [15].  

Left [9].  Right [13]

Other results include: (1) the average weight in the experimental group increased by 2% whereas the average weight in the control group decreased by 2% [5]; and (2) more gallstones seemed to form in the experimental group than the control group, which was also correlated with adherence to the experimental diet within the experimental group [16] 

Experimental
Control
Autopsies that looked for gallstones in men who a gallbladder
92
95
Gallstones
23
10
Gallbladder removals during the trial
0
4
Percentage (%)
25.0
14.7


Despite the randomisation, there were more heavy smokers in the control group but more moderate smokers in the experimental group.  However, the researchers stratified the incidence of CVD by packs of cigarettes and found that the “uneven distribution of cigarette-smoking habits had no net effect whatsoever on the outcome of the trial” [14]

To finish up, the Los Angeles Veterans Administration Trial was well designed, but the low vitamin E intake in the control group is a major confounding variable in favour of the experimental group and so the results shouldn't be taken at face value.  As total mortality was very similar, the difference in vitamin E intake suggests this is somewhat of an unfavourable trial for replacing SFA with PUFA

Chris Masterjohn has also discussed this trial here and here