[cycomiko]

Originally Posted by Lost Dog

You're missing the point. There's nothing wrong with the egg or it's yolk. The study is just crap.

What's likely happening is that people who tend to eat 7+ eggs a week are the types (like my Dad) who eat hash browns, toast, sausage, pancakes, etc. with those eggs. They also probably have crappy diets, overall. Look at the circumstantial evidence that I quoted. The ate fewer vegetables and ate more cereal. They drank a lot and smoked, too. Except for the smoking, none of those things are show stoppers in moderation, but we're likely looking at excessive eating of crap and a non-exercising, lifestyle.

if you take the raw data, uncontrolled for any of these other factors (age only) , egg consumption had no association with MI or stroke, but a modest association with all cause mortality.

However, when statistically adjusted to remove the biases of the vast majority of the above items

(Adjusted for age, BMI (continuous), smoking (never, past, or current smoker), and history of hypertension, vitamin intake, alcohol consumption (1, 1–4, 5–6, or7 drinks/wk), vegetable consumption (3, 3–4, 5– 6, 7–13, servings/wk), breakfast cereal (0, 1, 2–6, or7 servings/wk), physical activity (1, 1, 2–4, or5 times/wk), treatment arm (4 groups), atrial fibrillation diabetes mellitus (yes or no), hypercholesterolemia (yes or no), and parental history of premature myocardial infarction (yes or no).)

it confirmed the lack of association with MI or Stroke, and weakened the assocaition with all-cuase mortality (from a hazard ratio of 1.41 down to 1.23, but still statistically significant)

Without diabetes the association was at 1.22 hazard ratio, where as those with diabetes the hazard ratio was around 2-2.27 depending on the number of eggs.


So "independant" of the above items, >7/week egg intake was still associated increased risk of all cause mortality, but that was not from MI or stroke, so they died from 'something'...

Buy 20% increased risk is nothing in the big picture. For diabetics, 2x the risk 'might' be worth noting.

But as Alan has already noted, its observation, not causation. Even tho cohorts like this allow for some termporal ideas to be formed, its not providing enough information to be trying to hand out "proof of causation".