Monday, 20 August 2012
Eggceptional news: Its no yolk. The myth has been laid
In just four days, this posting has become the 7th most visited post on this blog - a measure of the interest and likely concern about the original report. Please forward the link to colleagues, and follow drphealth.blogspot.com regularly, or on Twitter @drphealth.
The recent news on the supposed finding that egg yolks are as bad for you as smoking has spread through the media faster than avian influenza. No doubt some folks have sworn off egg yolks and modified their egg cracking behaviours, the egg marketing people are plucking their feathers in disgust and the battle lines are being drawn for what should be an eggciting debate.
The article prepublished on-line in Atherosclerosis, followed a group of 1262 healthy people attending a prevention clinic in London, Ontario. The author, a neurologist, has argued repeatedly for the value of following carotid plaque formation as a risk factor for strokes for which there is reasonable evidence of a relationship - however showing causation and definition of the correlation is still open for some debate. The abstract is available and those that can access the full article are highly recommended to do so. Atherosclerosis article on egg yolks
The weak study design was based on a single recall questionnaires of behavior looking at egg eating habits and using this to define into two groups, those with two or less yolks a week, and those with 3 or more. After carotid plaque formation was then measured and simple analysis of smoking and egg consumption adjusting for age and gender against the outcome of plaque size. Most first year students would immediately flag numerous major design flaws in such an approach. Not bad for a fishing expedition, but wanting in terms of drawing conclusions.
Of the two groups apparently those with low egg consumption (n=388) had a mean of 125 ± 129 mm2, and those of the egg eating group (n= 603) 132 ± 142. Oops, somewhere we lost nearly 300 study participants. Assuming that the written range is based on a 95% confidence interval around the standard deviation (the most conservative of the assumptions), the t-test of these means has a probability of 0.15 – or not significant. Granted, the presented data don’t allow adjustment for age, gender and smoking status – but the lack of raw statistical significance against the findings when supposedly only age was adjusted and where the probability is apparently p<.0001 should raise some feathers. (also technically since the confidence intervals include zero, it is questionable whether the appropriate statistical approach (ie. using a Poisson distribution) might have been better).
Are you sensing something rotten in these eggs?
Noting as well that the absolute difference between the two groups is only a 6% margin of difference. The conclusion is that this difference is comparable to that of two-thirds of that of plaque formation for tobacco users. In fairness to the author, they make no blatant claims that egg consumption is as bad as smoking, in fact they recommend prospective studies and being sure to adjust for some measure of weight (which seems to be a critical covariant in this debate). While the full article is not the most eloquent of scientific writing, and critical appraisal gurus will have a hay day de-constructing innumerable problems, the most grievous of transgressions occurred after the publication was printed on-line.
The problem seems to have arisen when the University of Western Ontario communications folks took a crack at releasing to the lay media UWO communications release crowing in the headline that research finds egg yolks almost as bad as smoking. Misrepresentation? Possibly. Certainly a spin on the questionable facts.
The damage is done, just google “egg yolks as bad as smoking”. Newspapers across the globe have picked up the coverage of the press release and are running with the story without the scientific community having any opportunity to pull in the reins. Like the benefits of oat bran, vaccines and autism or a hundred other scientific myths – this one has become entrenched as an urban reality before its time, and the mythbusters will take years to establish its credibility or not as media retractions are as rare as hen’s teeth. By then, we will have genetically modified chickens producing yolkless, or at least cholesterol reducing yolked eggs and an industry that can propagate the myth for its own benefit.
Marginal science compounded by the new wave of high tech communications to support researchers in getting higher scores on referencing and citing for performance assessments. The question, will this be used to penalize appropriately both the researcher and the communications people? It would be a progressive day if there was a public statement on how this story was not what it was cracked up to be.