Thursday, July 9, 2009

The Finnish Mental Hospital Trial

This diet trial was conducted between 1959 and 1971 in two psychiatric hospitals near Helsinki, Finland. One hospital served typical fare, including full-fat milk and butter, while the other served "filled milk", margarine and polyunsaturated vegetable oils. Filled milk has had its fat removed and replaced by an emulsion of vegetable oil. As a result, the diet of the patients in the latter hospital was low in saturated fat and cholesterol, and high in polyunsaturated fat compared to the former hospital. At the end of six years, the hospitals switched diets. This is known as a "crossover" design.

The results were originally published in 1972 in the Lancet (ref), and a subset of the data were re-published in 1979 in the International Journal of Epidemiology (ref). They found that during the periods that patients were eating the diet low in saturated fat and cholesterol, and high in vegetable oil, male participants (but not females) had roughly half the incidence of heart attack deaths. There were no significant differences in total mortality in either men or women. The female data were omitted in the 1979 report.

This study is often cited as support for the idea that saturated fat increases the risk of heart attack. The reason it's cited so often is it's one of a minority of trials that came to that conclusion. The only other controlled trial I'm aware of that replaced animal fat with polyunsaturated vegetable oil (without changing other variables at the same time) and found a statistically significant decrease in cardiovascular deaths was the Los Angeles Veterans' Administration study. However, there was no difference in total mortality, and there were significantly more heavy smokers in the control group. The difference in heart attack deaths in the V.A. trial was 18%, far less than the difference seen in the Finnish trial.

I can cite three controlled trials that came to the opposite conclusion, that switching saturated fat for vegetable oil increases cardiovascular mortality and/or total mortality: the Anti-Coronary Club Trial (4 years), the Rose et al. corn oil trial (2 years), and the Sydney Diet-Heart trial (5 years). Other controlled trials found no difference in total mortality or heart attack mortality from this intervention, including the National Diet-Heart Study (2 years) and the Medical Research Council study (7 years). Thus, the Finnish trial is an outlier whose findings have never been replicated by better-conducted trials.

I have three main bones to pick with the Finnish trial. The first two are pretty bad, but the third is simply fatal to its use as support for the idea that saturated fat contributes to cardiovascular risk:

1) A "crossover" study design is not an appropriate way to study a disease with a long incubation period. How do you know that the heart attacks you're observing came from the present diet and not the one the patients were eating for the six years before that? The Finnish trial was the only trial of its nature ever to use a crossover design.

2) The study wasn't blinded. When one wants to eliminate bias in diagnosis for these types of studies, one designs the study so that the physician doesn't know which group the patients came from. That way he can't influence the results, consciously or unconsciously. Obviously there was no way to blind the physicians in this study, because they knew what the patients in each hospital were eating. I think it's interesting that the only outcome not susceptible to diagnostic bias, total mortality, showed no significant changes in either men or women.

3) The Finnish Mental Hospital trial was not actually a controlled trial. In an editorial in the November 1972 issue of the Lancet, Drs. John Rivers and John Yudkin pointed out, among other things, that the amount of sugar varied by almost 50% between diet periods. In the December 30th issue, the lead author of the study responded:
In view of the design of the experiment the variations in sugar intake were, of course, regrettable. They were due to the fact that, aside from the fatty-acid composition and the cholesterol content of the diets, the hospitals, for practical reasons, had to be granted certain freedom in dietary matters.
In other words, the diets of the two hospitals differed significantly in ways other than their fat composition. Sugar was one difference. Carbohydrate intake varied by as much as 17% and total fat intake by as much as 26% between diet periods (on average, carbohydrate was lower and total fat was higher in the polyunsaturated fat group). The use of psychiatric drugs with known cardiovascular side effects differed substantially between groups and could have accounted for some of the difference in cardiovascular events.  

The definition of a controlled trial is an experiment in which all variables are kept reasonably constant except the one being evaluated. Therefore, the Finnish trial cannot rightfully be called a controlled trial. The fact that the result has never been replicated casts further doubt on the study.
I could continue listing other problems with the study, such as the fact that the hospital population included in the analysis had a high turnover rate (variable, but as high as 40%), and patients were included in the analysis even if they were at the hospital for as little as 50% of the time between first admission and final discharge (i.e., they came and went). But what's the use in beating a dead horse?


30 comments:

Jim Purdy (Glucose Experiment) said...

You said:
"The Finnish trial is still very useful, however. I use it as a litmus test ..."

Very interesting way of looking at it. I often look at things the same way. Anybody whose credibility is damaged in one area is likely to not be credible in others.

Robert McLeod said...

Stephan:

I think Table 14 in the second study (International Journal of Epidemiology) is the real kicker. Drug doses were vastly different both between hospitals and also between the cross-over periods for each hospital. I think this is the reason they went for the cross-over design: the administration at each institution had different policies on drug use.

According to the study, they were treating their patients with: thioridazine, various phenothiazines and psychostimulants. Read some pharmokinetics papers on these classes of drugs, look for the section on drug-diet interactions, and go hmmm...

Stephan Guyenet said...

Robert,

Good point, that's another major uncontrolled variable.

randy said...

Hello Stephan,

Please note that I had break my reply in two post for it to fit

Stephan Wrote:
The study wasn't blinded. When one wants to eliminate bias in diagnosis for these types of studies, one designs the study so that the physician doesn't know which group the patients came from. .

Reply:
This is not true. The study was blinded. The phycians did not know from which hospital or diet group the subjects came from. This is clearly discussed in the 1979 paper.

Stephan Wrote:
In an editorial in the November 1972 issue of the Lancet, Drs. John Rivers and John Yudkin pointed out, among other things, that the amount of sugar varied by almost 50% between diet periods. In the December 30th issue, the lead author of the study responded:

Reply:
This is not accurate. The sugar/sucrose intake was <15% different between the two diets. This is clearly reported in the 1979 paper.

Stephan Wrote:
How do you know that the heart attacks you're observing came from the present diet and not the one the patients were eating for the six years before that?

Reply:
Easy. There were two diets tested. The "Normal Diet" which is what ALL subjects were eating previously and the SCL diet where PUFs were substituted for SFAs.
Since both groups were eating the same diet previously and the only change was the group eating the SCL diet any changes would be due to the SCL diet. This is specifically discussed in the paper.

Stephan Wrote:
In other words, the diets of the two hospitals differed significantly in ways other than their fat composition.

Reply:
This is both untrue and misleading on a couple of levels.
First of all, the carb differences were <4% different between the two diets in each hospital.
More importantly, since this was a cross over design both carb and fats Amounts were removed as confounder. That is, the subjects had less (~50%) CVD events on the SCL diet irregardless of the carb or fat (higher or lower) amounts.

Robert Wrote:
According to the study, they were treating their patients with: thioridazine, various phenothiazines and psychostimulants.

Reply:
The authors agree and do discuss this as a possible confounder.
As discussed in the 79 paper, the hospital in which use of these drugs doubled from the first time period to the second, but the incidence of CHD was much reduced. This is the opposite effect one would see if these drugs were skewing results.

randy said...

Hello Stephan,

Here's the continuation of my reply:

Stephan Wrote:
I can cite three controlled trials that came to the opposite conclusion, that switching saturated fat for vegetable oil increases cardiovascular mortality and/or total mortality:

Reply:
As regards the Anti-Coronary Club Trail your remarks are very misleading. The study was too short and had too few participants to provide any info on mortality. What the study did find and report was a significant 30% reduction CVD events with a p value of .01.

The Oslo Diet-Heart Study (1) , which used a simular diet to the Finnish study, was followed up for 11 years, (long enough the show mortality data) show reduced CVD mortality with a p<.004.

The Rose Corn Oil was almost completely lacking in dietary controls compared to the Finnish study. In the Finnish study, all meals were prepared to conform to the study arm the men were in. We know exactly what they eat as regards macro and micro nutrients. The subjects in the Rose study, on the other hand, were "advised" to dietary content. Food recall records or logs were not used.

Also I'd like to point a paper (1) that fulfilled the Cochrane review criteria for a meta-analysis of clinical trials looking at dietary fat manipulaton and CVD. They used alot more studies than either of us has provided and put alot more brainpower to come to a conclusion. The results showed "Lifestyle advice to all those at high risk of cardiovascular disease (especially where statins are unavailable or rationed), and to lower risk population groups, should continue to include permanent reduction of dietary saturated fat and partial replacement by unsaturates."
Another paper here (3,7), contends that the above results would been greater if less PUFAs(and a higher n3/n6) and SFAs were consumed.

They contend, amoung other things, that increased PUFAs (especially with sub optimal N3/N6 ratios) used in so many of the other studies, acutally reduced the positive benenfits in the older studies that we both quoted. So, to achieve the best results reduce SFAs without a great increase in PUFA and pay attention to the N6/N3
ratios.

Another recent metabolic ward study (4) that drastically redcuced both SFAs and PUFAs showed a dramatic reduction in suite of inflammatory markers in 3 weeks with little or no reduction in weight or calories.
Following this approach, the Oslo Study(2), reduce incidence of heart attachs by 47% (both fatal and non fatal) by reducing SFAs, but not substiting large quantities of PUFAs.

This data together with the new studies (5,6) inditing SFAs are a factor in deterioating endothelial health supports the view that SFAs should be decreased without increasing total PUFAs and mantaining an optimal N3.N6 ratil.
Regards
Randy

1. Dietary fat intake and prevention of cardiovascular disease: systematic review.
http://www.ncbi.nlm.nih.gov/pubmed/11282859

2. Dietary fats and prevention of cardiovascular disease Conclusion may have been underplayed
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=1121475
3. Effect of diet and smoking intervention on the incidence of coronary heart disease. Report from the Oslo Study Group of a randomised trial in healthy men. http://www.ncbi.nlm.nih.gov/pubmed/6118715
4.
Effect of a short-term diet and exercise intervention on oxidative stress, inflammation, MMP-9, and monocyte chemotactic activity in men with metabolic syndrome factors
5.
Flow-Mediated Dilatation Is Impaired by a High–Saturated Fat Diet but Not by a High-Carbohydrate Diet

6.
Comparative effects of three popular diets on lipids, endothelial function, and C-reactive protein during weight maintenance.
Miller M, Beach V, Sorkin JD, Mangano C, Dobmeier C, Novacic D, Rhyne J, Vogel RA.
Divisionof Cardiology, University of Maryland Hospital, Room S3B06, 22 S. Greene St, Baltimore, MD 21201, USA. of variance).

7. Lipoproteins, nutrition, and heart disease

mark said...

Randy wrote:

Stephan Wrote:
In an editorial in the November 1972 issue of the Lancet, Drs. John Rivers and John Yudkin pointed out, among other things, that the amount of sugar varied by almost 50% between diet periods. In the December 30th issue, the lead author of the study responded:

Reply:
This is not accurate. The sugar/sucrose intake was <15% different between the two diets. This is clearly reported in the 1979 paper.

=============================

I'm looking at the paper,
Dietary Prevention of Coronary Heart Disease: The Finnish Mental Hospital Study.

International Journal of Epidemiology

Vol 8, No 2

1979.


This paper deals only with the male population. Female population is in another paper, and the authors do not give data for sucrose intake for the females.


SCL (soy)Diet in Hospital N 1959-1965 has a sucrose intake of 68.5grams per day.

NORM (saturated) Diet in Hospital K from 1959-1965 was 102 grams per day.

The difference is 49%

The diet higher in saturated fat had 49% more sucrose.

The SCL Diet in Hospital K 1965-1971 has a sucrose intake of 86.7

The NORM Diet in Hospital N 1965-1971 has a sucrose intake of 78.5

This time, the diet higher in saturated fat had 10% LESS sucrose intake.

So Yudkin is right about 1959-1965. 50% different.

All those numbers were pulled from Table 3.

Mark.

Stephan Guyenet said...

Randy,

Your comments contain a number of misleading or flatly incorrect statements.

First of all, it was definitely not blinded. Blinded means the physicians have no knowledge whatsoever of who's in which group, I quote from the 1972 paper:

"The house physicians who wrote the death certificates and the pathologists who did the necropsies were not taking part in the study and were only vaguely aware of its design and objectives"

In other words, they were aware of it to some degree and we have to take the authors' word for it that their knowledge didn't influence the results. The idea that the doctors doing the necropsies wouldn't be aware of a massive 12 year diet-heart trial going on in their own hospital is clearly a stretch. The only thing that was blinded was the interpretation of the death certificates, which is what you referred to in the 1979 trial.

How can you say that I'm exaggerating the differences between diet periods when you have the full text of the study right in front of you? Here's a little bit of the data, tell me if I'm making this up (from table 3, 1979 paper). Total sucrose:

Hospital N, SCL diet: 68.5 g/day
Hospital N, NORM diet: 78.5 g
Hospital K, SCL diet: 86.7 g
Hospital K, NORM diet: 102 g

As I said, and as Mark confirmed above, the amount of sucrose varied by nearly 50% between diet periods.

You referred to the numbers where the investigators averaged SCL with SCL and NORM with NORM. The problem is, you can't simply take two diets that are clearly different in many ways and average them together! That is in the realm of observational studies, not controlled trials.

One NORM diet was preceded by normal fare, and the other was preceded by 6 years of vegetable oil! Yet the investigators averaged those two different conditions together to try to eliminate statistically significant differences between the diets. Again, that is not what a controlled trial is, and it's the reason why the crossover design was never used again after the Finnish trial.

As for the Anti-Coronary club trial, how about we just state the results and let the readers decide what they think? Significant reduction in soft "cardiovascular events" in the vegetable oil group. 9 deaths from coronary heart disease in the vegetable oil experimental group, none in the saturated fat control group. Total mortality more than doubled in the vegetable oil group. The study wasn't blinded, so total mortality was the only measure not susceptible to diagnostic bias. Which group would you rather be in?

You said "The Rose et al trial was almost completely lacking in dietary controls compared to the Finnish study". That is flatly incorrect. The study included a normal diet control group. But again, why don't we let the readers decide who's correct? Here's the free full text. And by the way, the physicians were actually blinded in the Rose et al. trial, as opposed to the Finnish trial.

Randy, I'm not going to continue this discussion if you continue misrepresenting the evidence.

randy said...

Hi Mark,

Thanks for pointing that out, I agree.

I was using the 15% figure comparing the differences between study groups in the same hospital.

Because this was a crossover study comparing the same folks in the same hospital and different folks from different hospitals. We can tell by looking at the data if sucrose was a factor.

Remember there are 4 comparison groups. Three of these groups consume surcose less than 15% of each other, one group, as you point out consumes 50%(or 33%).

If sucrose intake has a (significant) effect the group with the higher intake should have higher (or lower) disease markers compared to the other groups.

Is this the case? If yes, sucrose is a good canidate of a confounder,if no then sucrose is not a factor in the results.

Thanks for the comment (based on facts) and let me know if I'm missing something

Regards
Randy

randy said...

Hello Chris,

Chris Wrote:

First of all, it was definitely not blinded. Blinded means the physicians have no knowledge whatsoever of who's in which group, I quote from the 1972 paper:

Reply,
I stated in my last post that my info was coming from the 1979 paper, but you quote from the 1972 paper.

According to the 1979 paper the evaluations were properly blinded.

Here's two quotes directly from the 1979 paper.

Regards
Randy

Quote 1:

All electrocardiograms
in this study were read and coded by the same member of the team (MM). This was done 'blindly', i.e. without knowledge from which
hospital the tracing came.

Quote 2:

Criteria for coronary death. In addition to the above
electrocardiographic patterns, one further end-point was used, i.e. the coronary death. For assessing the cause of death, that is, whether it had been coronary
or non-coronary, the following 'blind' procedure
was observed. Summaries including all relevant information on the illnesses of the subjects were
prepared on the basis of the hospital case records.
These summaries were identified by code numbers
only and contained nothing to indicate from which
hospital they came. To determine whether a death
had been coronary or non-coronary, 3 of us (OT,
MM and MJK) evaluated the summaries, first
individually and then collectively. In most cases the
agreement between the individual evaluations was
good.

randy said...

Stephan Wrote:

As I said, and as Mark confirmed above, the amount of sucrose varied by nearly 50% between diet periods.

Reply:

There were 4 comparison groups. Three of these differed by less than 15% sucrose intake. One of the 4 differed by 50 or 33 percent depending how you do the math.

If sucrose was a confounder the high sucrose comparison would differ (significantly) more than the other 3.

Regards
Randy

randy said...

Stephan Wrote:

As for the Anti-Coronary club trial, how about we just state the results and let the readers decide what they think? Significant reduction in soft "cardiovascular events" in the vegetable oil group. 9 deaths from coronary heart disease in the vegetable oil experimental group, none in the saturated fat control group. Total mortality more than doubled in the vegetable oil group. The study wasn't blinded, so total mortality was the only measure not susceptible to diagnostic bias. Which group would you rather be in?

Reply:
Normally I'd say to take a course in stats and read the study. I've assumed been through more than stat course..

For anyone that wants to read it:
http://tiny.cc/xpNq6

Comparing hard numbers without statistical analsys
is invalid. For one thing, the group size differs by 100% (or 50%).

Here are the results reported in the paper:

No differences in mortality.
Exp: 689/100,100 person years
Con: 666/100,000 person years

Significant differences in CVD events:
Exp: 196/100,000 person years
Con: 642/100,000 person years
p<.04

Regards
Randy

Stephan Guyenet said...

Randy,

You missed the point of the sucrose comment. It wasn't to say sucrose was the cause of the difference in CHD deaths, it was to say that the study wasn't properly controlled.

CHD deaths and total mortality were the primary study outcome. The study authors made no effort to blind the physicians conducting the autopsies, that is clearly stated in the 1972 paper. So here's the result: no significant difference in total mortality in either men or women. A significant reduction in CHD events in men only, based on autopsies that were not blinded. No amount of wiggling will get you around those facts.

The Finnish trial result is an outlier from all the other controlled trials that tested the same intervention. If you want to see another trial that was actually double-blinded, well controlled, had a large difference in saturated fat intake between the two groups (9 vs 18%), and was even done in a mental hospital, check this out
. This trial was superior to the Finnish trial in almost every way. I think you already know what the outcome was.

Stephan Guyenet said...

Randy,

You really have an amazing way of cherry picking data that confirm your biases and discarding data that don't. In the Anti-Coronary Club trial, the numbers you cherry picked for total mortality were for the age group 50-59 years only. The overall mortality was 3.3% in the PUFA group and 1.4% in the control group. It's a simple calculation you can do by adding the numbers for cardiac and non-cardiac mortality they presented in the paper. 27 dead in the PUFA group, 6 dead in the control group. They don't present that in any of their tables but you can draw it from the text body.

As I said before, there were 9 coronary heart disease deaths in the experimental group and zero in the control group.

I'm getting really tired of your shenanigans Randy. You are polluting my blog with your misleading but superficially knowledgeable-sounding comments.

Melchior Meijer said...

Randy and Stephan,

If memory serves, the intervention in the Finnish Mental Hospital Trial was soy oil. Soy oil has quite a lot of ALA. These folks were mentally ill in Finland, which means that they were almost certainly extremly n-3 depleted. It wouldn't be too far fetched to call this study a secundary prevention study. It's hardly surprising that throwing some n-3 into a severely n-3 deficient population produces a significant cardiovascular benefit, regardless of SFA's and n-6.

randy said...

Stephan Wrote:

You missed the point of the sucrose comment. It wasn't to say sucrose was the cause of the difference in CHD deaths, it was to say that the study wasn't properly controlled.

Reply:
Its quite clear from the results that the higher sucrose did not have increased (or decreased) CVD. Why is this relavent.

I find it disturbing that you critsize the Finnish study on this insignificant point, but laude the Rose study when it suffered from a real lack of reporting.

Atleast the Finnish kept detailed records on what and how much the subjects ate. The Rose study didn't provide any nutritional follow up information at all.

You give the Rose study a pass when we don't even know what these guys were eating, but critisize the Finnish study which, unlike the Rose study, provided detailed nutritional data.

You seem to have differnt standards for study that support your view as opposed to those that don't.

Your not being fair.

Regards
Randy

randy said...

Stephan Wrote (on the Finnish Study):

A significant reduction in CHD events in men only, based on autopsies that were not blinded.

Reply:
This is getting tiring Stephan. Your not playing fair.

I repeat for the 3rd time the autopsies were blinded! I repeat the autopsies were blinded.

Here's a quote from the 1979 paper:
//********************************
"Summaries including all relevant information on the illnesses of the subjects were
prepared on the basis of the hospital case records.
These summaries were identified by code numbers
only and contained nothing to indicate from which
hospital they came."
//********************************

How is this not blinded?

Regards
Randy

randy said...

Stephan Wrote:

If you want to see another trial that was actually double-blinded, well controlled, had a large difference in saturated fat intake between the two groups (9 vs 18%), and was even done in a mental hospital, check this out

Reply:
My response to this study is the same that the authors of paper presented in their closing remarks.

Namely, postive results for this dietary approach don't start to show up until 2 years. The study authors cite 3 papers to support this contention including the Finnish study. Its also interesting that they seem to accept the Finnish results with no reference to any of your critisms.

Here's the quote:
//*****************************
Although this study did not show a statistically significant
reduction in cardiovascular events or total deaths
from the treatment diet, the authors suspect that it might
have shown such a reduction if the period of treatment
had been longer in persons in the age range likely to
benefit. We included persons of all ages, both men and
women, with an average cholesterol concentration of 207
mg/dl, compared to about 290 mg/dl for the participants in
the Helsinki and Upid Research Clinics trials. When the
study was first proposed, very lengthy stays in mental
hospitals were common. By the end of the initial 3-year
pre-treatment observation period, the practice of vigorous
drug treatment and early discharge to the community was
in full swing.
Table 6, in which the analysis is confined to persons
on diet for at least 2 years and in the age groups chosen
//*********************************

Stephan Guyenet said...

Randy,

I'm getting the impression you aren't really reading my comments. I'll repeat this one last time. The autopsies were not blinded. The only thing that was blinded was the interpretation of the death certificates, which is what the quote you posted was referring to. Autopsy, not blinded. Interpretation of death certificates, blinded. That does not equal a blinded study.

The study I linked to was essentially equivalent to the Finnish trial but better-conducted. Your speculations and those of the authors do not change the fact that a 4.5 year double-blind trial found no difference in coronary events, coronary deaths or total mortality between groups eating very different amounts of PUFA and saturated fat. This is consistent with most of the other controlled trials that have evaluated the same intervention.

You are clinging to straws here Randy.

Robert Andrew Brown said...

Stephan said

"I could continue listing other problems with the study, such as the fact that the hospital population included in the analysis had a high turnover rate (variable, but as high as 40%), and patients were included in the analysis even if they were at the hospital for as little as 50% of the time between first admission and final discharge (i.e., they came and went)."

This struck me too when I read it.

Does this at least not put a big question mark over the result.

Melchior's comments made a deal of sense too

"If memory serves, the intervention in the Finnish Mental Hospital Trial was soy oil. Soy oil has quite a lot of ALA. These folks were mentally ill in Finland, which means that they were almost certainly extremly n-3 depleted. It wouldn't be too far fetched to call this study a secundary prevention study. It's hardly surprising that throwing some n-3 into a severely n-3 deficient population produces a significant cardiovascular benefit, regardless of SFA's and n-6."

Another question mark?

Robert McLeod sagely suggested the drugs used may have a bearing so as suggested I went and looked up the first thioridazine and this is what the warning side effect list included:

http://www.rxlist.com/thioridazine-drug.htm

"Cardiovascular System: Thioridazine produces a dose related prolongation of the QTc interval, which is associated with the ability to cause torsade de pointes-type arrhythmias, a potentially fatal polymorphic ventricular tachycardia, and sudden death (see WARNINGS). Both torsade de pointes-type arrhythmias and sudden death have been reported in association with thioridazine. A causal relationship between these events and thioridazine therapy has not been established but, given the ability of thioridazine to prolong the QTc interval, such a relationship is possible. Other ECG changes have been reported"

Another ?

I did not bother to check the rest of Robert's list.

randy said...

Stephan Wrote:
The autopsies were not blinded. The only thing that was blinded was the interpretation of the death certificates, which is what the quote you posted was referring to. Autopsy, not blinded. Interpretation of death certificates, blinded. That does not equal a blinded study.

Reply:
Point taken, sort of.

Your assuming the coroners knew which study arm the subjects came from. This may or may not be true. Its not unreasable, I admit, to assume it might be true.

But the according to the paper autopsy results had to report the following:

"Death associated with autopsy findings of occluding or nearly occluding thrombi in coronary arteries, or of recent myocardial infarction."

That's a pretty objective finding, and I doubt that the coroner could be biased to misinterpert.

Also the bulk of the data were from EKG reports that were blinded and these data show the same results.

But your right, its potentially conceiveable that the death reports could be biased. I just think it a real strech.

Regards
Randy

LeenaS said...

To Randy:
Remember, Finland is a very small country - and was smaller still in seventies. Non-biased research in a tiny pool and on a politically hot issue such as nutrition was in seventies... Dream on :)

Melchior Meijer said...

Dear all,

Please take note of the fact that many mental illnesses are associated with a hugely increased risk of CHD-events (most people reading this blog would agree that they share a common etiology). This was probably a very incident prone population. That's why I consider this a 'secundary prevention trial in disguise'. Psychiatric patients are often very n-3 depleted (especially in Finland in the seventies, before the dawn of the canola epoque) and thus prone to arrythmia's. Is it so strange that you see a reduction in fatal incidents in n-3 depleted people after you have partly correct their depletion, albeit with a relatively inferior intervention oil? ALA's anti-arrythmic properties are huge and probably solely responsible for the spectacular decrease in (only) fatal MI's in Lyon.

randy said...

Stephan Wrote:
Your speculations and those of the authors do not change the fact that a 4.5 year double-blind trial found no difference in coronary events, coronary deaths or total mortality between groups eating very different amounts of PUFA and saturated fat.

Reply:
It is NOT a speculation (according to the authors) that the study didn't have enough subjects in the appropiate age groups to make a determination after 2 years. Its after the two year period when results of the Finnish and Lipid Lab study start accumulating.

They DO speculate that is this might be the case as is indicated by the trend in the data and results of three other studies (including the Finnish study).

In effect, this study didn't have the power (admitted by the authors) to make a call after two years and its after the 2 year period when the Finnish and other studies start showing results.

Regards
Randy

Stephan Guyenet said...

Randy,

I'm not giving the Rose et al. trial a free pass at all. It wasn't a bad study, but it was underpowered and it wasn't asking quite the same question we're debating. And I'm certainly not giving the Anti-Coronary Club trial a free pass, that one wasn't even randomized! My point in citing those is simply that for every bad study with a weird result (the Finnish trial), there's another bad study that found the opposite result.

In this case, there are three trials that found the opposite result, one of which was bad (Anti-Coronary Club), one of which wasn't asking quite the same question (Rose et al), and one of which was actually large and well-conducted as far as I know (Sydney diet-heart trial). The Sydney trial found an increase in total mortality after 5 years of replacing animal fat with PUFA oil, but didn't report heart attack numbers.

I agree with your point about the Minnesota trial; the turnover rate was its main weakness. That same criticism also applies to the Finnish trial. They were both conducted in mental hospitals with a substantial turnover rate. The patients in the Finnish trial were not all on the diets for 6 years continuously, the population turned over and patients checked in and out.

Melchior,

You made some good points about the possible role of omega-3 in the Finnish result. I agree that mental patients are likely to be more omega-3 depleted than the general population. I don't think we can read too far into the Finnish trial result though. There's also the problem that there was another trial that replaced animal fat with soybean oil for up to 7 years, the 1968 Medical Research Council trial, which found no difference in total mortality or CHD mortality.

randy said...

Stephan Wrote:
You really have an amazing way of cherry picking data that confirm your biases and discarding data that don't.

Reply:
I did not and do not cherry pick data to make a point and I take offense at the accusation.

I'll assume we are looking at two different reports.

The data I reported were the primary conclusion of the paper.

Here's the link:
http://tinyurl.com/lnp63r

Here's the data I reported:
//*******Reported on page 309******
No differences in mortality.
Exp: 689/100,100 person years
Con: 666/100,000 person years
//*********************************
//**********Reported on page 313**
Significant differences in CVD events:
Exp: 196/100,000 person years
Con: 642/100,000 person years
//****************************
These statements were primary conclusions of the study authors.

I agree with you that this is a lousy study, and only remarked on it because you used it to support your position earlier.

Also Stephan I think its important when one references a study to support a conclusion that's opposite of the paper's conclusion, you state that as such. Many folks that read your stuff are not going to read the full and will assume your conclusion is the conclusion of the paper.

And if your going to disagree with the conclusion of the paper atleast acknowlege the importance of statistical analysis.

Sorry, but I'm little skeptical of immediately throwing out conclusions of a paper that has gone through any peer review, with an arugment based on numbers that has not been checked by a neutral 3rd party

Finally I do acknowledge the importance of what you doing, it could have important implications for one's health, just slow down please, especially when your throwing numbers around. I'm sure you've got some academic stat guys to overlook some of your conclusions.

Regards
Randy

Stephan Guyenet said...

Randy,

You are dead wrong about the Anti-Coronary Club trial. You said:

"No differences in mortality.
Exp: 689/100,100 person years
Con: 666/100,000 person years"

This is exactly what I'm talking about when I say you cherry pick. If you read the paper carefully, you'll see that those numbers you quoted were for the 50-59 year age group only. You ignored the total mortality numbers for all age groups and focused only on one age group, which gives a misleading impression.

The study authors did their best to conceal the total mortality numbers, as they didn't include it in any of their charts. But you can add it up from the numbers they list in the text of this paper
. I quote, from page 600, section "Deaths From Coronary Heart Disease":

"Of the eight subjects with new coronary disease events in the fully participating experimental group, three
died of coronary heart disease, one died of other causes, and the other four were still alive at the end of the observation period. Of the seven detected cases with new events in the inactive experimental group, all of which were definite myocardial infarctions, five died of coronary heart disease and the other two were still alive at the end of the observation period...
Of the 12 cases with new coronary disease events in the control group, all were still alive at the end of the observation period."

From the section "Deaths From Other Causes", page 601:

"Among the 814 original experimental group subjects, there have been 18 known deaths from causes other than coronary heart disease... compared to six such deaths out of the 463 individuals in the control group."

Next, comes the sentence you picked up on about the death rate not being different in the 50-59 age group.

Allow me to summarize. 9 deaths from CHD in the experimental group, zero in the control group. 27 total deaths in the experimental group, 6 in the control group. Overall mortality rate, 3.3% in the experimental group, 1.4% in the control group. This is the last time I'm going to correct you on this Randy.

randy said...

Stephan Wrote:
This is exactly what I'm talking about when I say you cherry pick.

Reply:
I did't cherry pick, that's what the authors presented. I didn't know about the other paper and have the numbers. It's totally appropiate to point this out but your not justified in calling me a "cherry picker". Do you want to have a serious conversation or not.

Stephan Wrote:
Allow me to summarize. 9 deaths from CHD in the experimental group, zero in the control group. 27 total deaths in the experimental group, 6 in the control group. Overall mortality rate, 3.3% in the experimental group, 1.4% in the control group. This is the last time I'm going to correct you on this Randy.

Reply:
I already explained that I hadn't the second paper and I resent you snotty tone. I repeat Do you want to have a reasonable converstaion or not, I do.

The ommisions in your report creates a false impression. This makes it hard for me to trust your reports without double checking your sources.

Your death reports in the experimental group include lots of folks that had dropped out of the experimental protocal. Also it's important to keep in mind that the had 814 subjects vs 463 in the control group.

Here's the the total picture:
//
8 new CHD cases in the participating exp group and 3 CHD deaths. (514 subejcts)

12 new CHD cases in the control and no CHD deaths. (463 subjects)

7 new CHD cases and 5 deaths in experiemental dropouts. (290 subjects)
//
Also this was an expectional crappy study that was widely critisized. I think you admitted this in an earlier post. I only reponded to it because you used it to support your postion. The control group was selected inappropiately.

Here's a qoute from Steinberg on the study:

"The control group was recruited from a quite different
population—men who had volunteered for examination
at a cancer clinic".

Finally Stephan I do urge you again to clearly state when your conclusions differ from the conclusion paper your quoting. This can't hurt, and only increases transparency.

Regards
Randy

randy said...

Stephan Wrote:
We could have avoided this entire debate if you had simply read the papers carefully and commented on them accurately.

Reply:
Initially when you referenced that paper you did not provide links. I did my own search and referenced the paper that didn't detail the death rates. I didn't even know the other paper existed. It would be great if you could provide links to your references so this won't happen again

But, unlike you, when I read the other paper, I did accurately report the data. You conviently lumped the data for the active members with the inactive members of the experimental groups. That skewed the data to make a better case for your view point.

Stephan Wrote:
Overall mortality rate, 3.3% in the experimental group, 1.4% in the control group.

Reply:
Whats the p value of that calculation. What's the percentage that these results would happen by chance. Is there really enough subjects in the groups to make this claim. Would any of your stat profs OK those numbers.

And why, in the name of reasonableness, are you deriving mortality numers for a study you admit sucks.

Stephan, when I say I want to have a conversation, I mean it. Your name calling serves no purpose.

Regards
Randy

Melchior Meijer said...

Thanks for mentioning that 1968 Medical Research Council Study, Stephan! I'll dig it up.

Lillian said...

you have nice collection of information. You focus this issue nicely.

regards
Laura brasnon
Zellweger Syndrome