Search This Blog

Tuesday, 7 October 2014

Saturated Fat intake equivalent to 130g Butter a day is not significantly associated with higher heart disease risk in Finland.

[edited 4/11/14]

If you saw Professor Rod Jackson on NZ TV the other day, he was arguing, ably it has to be said, that the whole Big Fat Surprise / Time magazine cover, and by extension LCHF, risks reversing gains made against heart disease in recent years.
If you're advocating a high fat diet and saying saturated fat is not that important, if it's important at all, that's a serious charge, and worth taking seriously.

For the defence:
Another epidemiological analysis, a long-term follow up (21.4 year) of a population (1,981 men) was released last week.

Dietary Fatty Acids and Risk of Coronary Heart Disease in Men

The Kuopio Ischemic Heart Disease Risk Factor Study

During the average follow-up of 21.4 years, 183 fatal and 382 nonfatal CHD events occurred. SFA or trans fat intakes were not associated with CHD risk. In contrast, monounsaturated fat intake was associated with increased risk and polyunsaturated fat intake with decreased risk of fatal CHD, whether replacing SFA, trans fat, or carbohydrates. The associations with carotid atherosclerosis were broadly similar, whereas the associations with nonfatal CHD were weaker.
The association between MUFA and CHD mortality was, though statistically significant, and I believe one that has turned up before, small enough that I am not losing any sleep over it.
The interesting feature of this study was the high level of saturated fat in the diet, as seen in this PDF of supplementary data.

You will notice that calories increase stepwise with SFA, and it occurred to me that in real, gram, amounts, the difference in SFA intake between quartiles is even greater than that of SFA as %E. The conversion is easily done (Kcal ÷ 100 × %E ÷ 9 - being mathematically challenged, I am as
pleased as a dog with two tails to have thought of that on my own).

Here are the "real" median daily SFA consumption figures by quartile:

Q1 32g
Q2 42.5g
Q3 52g
Q4 67g

Now, 67g of saturated fat is a lot. It's equivalent to more than 130g of butter per day. Fat was 45.4%E in quartile 4. No more cardiovascular mortality than people eating half as much.
Some might say that 32g, or 13.4%E in the lower quartile is already too much saturated fat, but there are plenty of epidemiological studies showing the same flat line at lower intakes. No dose-response.

SFA really does stand for SFA.

There are no total mortality stats. I assume (dangerously of course, it would be better to know, but in line with similar studies) there was no difference there either, even in the higher PUFA group who had somewhat lower IHD mortality (which they could equally get by replacing carbohydrate, TFA, SFA or MUFA with PUFA). The PUFA intake associated with protection wasn't high, both the 4.8%E and 6.3%E quartiles did equally well. They ate more margarine, more fish, and more meat, for fewer calories than the lower PUFA quartiles. Nuts weren't a food measured, and vegetable oils made a minute contribution (olive oil had nothing to do with the small correlation between MUFA and IHD - I'm willing to bet that olive oil consumption would have cancelled out or reversed this association, which seems to belong to margarine and meat, probably pork, without extra fish).
PUFA from margarine plus fish, probably wasn't associated with reduced all-cause mortality, or we'd have heard about it. Nuts, on the other hand, are regularly associated with lower IHD and all-cause mortality.
There's a difference between trading one cause of death for another (what seems to happen when we increase PUFA promiscuously, including oil and margarine) and reducing mortality from all causes (what seems to happen when we eat more PUFA as nuts and fish).
High PUFA consumers were much less likely, high SFA consumers much more likely, to live in a rural area. BMI stayed fairly consistent across SFA quartiles, except the lowest quartile (with lowest dairy consumption) was a little heavier, despite the increasing calories and decreasing leisure-time exercise across quartiles. Rural life seems to be a fair substitute for leisure-time exercise.
Eating more SFA is associated with higher LDL in this study, but not with greater IHD mortality.
Eating more PUFA is associated (less consistently) with lower LDL, and with (somewhat) lower IHD mortality.
Go figure.

One limitation of this study is that four-day food diaries taken in the 1980s were the only food data used. The Finnish diet changed from the 1980s to the present day; did the participants' diets change enough to affect the outcome? The authors tested this by separating the first 10 years, which showed the same trend, but I can think of a better test; if the diets had changed significantly by the high-SFA consumers beginning to eat less SFA and substituting PUFA (the only change to SFA that would have made a difference), the protective association with PUFA would have been lost. It wasn't. Which doesn't mean that smaller effects weren't swallowed up in the changes of time; but anything major, I think, would still have stood out. The PUFA association is still much what we would predict from other studies in other countries with stable food habits.

If you are interested in the background to the KIHD study, Uffe Ravnskov has a chapter on Finland in The Cholesterol Myths, which is available for free here. Kuopio is part of the "rest of Finland", used as control in 1972, then brought into the national risk factor reduction program in, I believe, the 1980s.

CHD mortality in Finland was declining before the cholesterol and other risk-factor lowering program began

Another interesting feature of the Finnish story is the role of coffee, especially boiled coffee, in elevating LDL, and the association of boiled coffee alone with IHD mortality. Again there is no linear association between LDL elevation and risk of IHD; it is the quality of the coffee that matters. A switch from boiled to filtered coffee was part of the 1980's risk reduction program.

From Coffee consumption and death from coronary heart disease in middle aged
Norwegian men and women
, by A. Tvderdal et al (link)

It makes no difference whether coffee is boiled or filtered if the aim is preventing liver disease.