Tag Archives: Observational studies

Is It Time To Stop Eating Deadly Eggs Or Time To Stop Putting Nutritional Epidemiology In The Headlines?

After carefully ingesting the new JAMA egg study that has gotten the media and many patients in a tizzy I consumed a three egg (of course with yolk) omelette.

I am happy to report that I survived the incident and am not concerned at all that my longevity has been compromised.

My 2013 summary of eggs, dietary cholesterol and heart disease (see here) is still valid and I highly recommend patients and readers read that post plus my updates on eggs with newer data (see here and here) rather than information related to the new egg study.

Although CNN and other news outlets lead with an inflammatory headline suggesting that eating those 3 eggs increased my risk of heart disease the new egg study could not possibly prove causation because it was an observational study.

Nutritional epidemiology has come under considerable criticism in the last few years for churning out these weak observational studies .John Ionaddis has been particularly vocal about these limitations, writing:

A large majority of human nutrition research uses nonrandomized observational designs, but this has led to little reliable progress. This is mostly due to many epistemologic problems, the most important of which are as follows: difficulty detecting small (or even tiny) effect sizes reliably for nutritional risk factors and nutrition-related interventions; difficulty properly accounting for massive confounding among many nutrients, clinical outcomes, and other variables; difficulty measuring diet accurately; and suboptimal research reporting. Tiny effect sizes and massive confounding are largely unfixable problems that narrowly confine the scenarios in which nonrandomized observational research is useful

This egg study contains the usual flaws that render it inconclusive:

First, the study relies on data collected from a food frequency questionnaire. Have you ever sat down and tried to recall exactly what you ate in the previous week? How accurate do you think your estimate of specific food items would be?

Ed Archer has written about the inaccuracy of the food frequency questionairres extensively. Here’s a sample from one of his devastating critiques;

In lieu of measuring actual dietary intake, epidemiologists collected millions of unverified verbal and textual reports of memories of perceptions of dietary intake. Given that actual dietary intake and reported memories of perceptions of intake are not in the same ontological category, epidemiologists committed the logical fallacy of “Misplaced Concreteness.” This error was exacerbated when the anecdotal (self-reported) data were impermissibly transformed (i.e., pseudo-quantified) into proxy-estimates of nutrient and caloric consumption via the assignment of “reference” values from databases of questionable validity and comprehensiveness. These errors were further compounded when statistical analyses of diet-disease relations were performed using the pseudo-quantified anecdotal data. These fatal measurement, analytic, and inferential flaws were obscured when epidemiologists failed to cite decades of research demonstrating that the proxy-estimates they created were often physiologically implausible (i.e., meaningless) and had no verifiable quantitative relation to the actual nutrient or caloric consumption of participants.

In addition to unreliable initial data the subjects were followed up to 30 years without any update on their food consumption. Has your food consumption remained constant over the last 30 years? Mine hasn’t. I went from avoiding eggs to eating them ad lib and without concern for my cardiovascular health about 5 years ago after looking at the science related to dietary cholesterol.

It’s Hard To Get Around Confounding Variables

Observational studies like this one try to take into account as many factors as they can which might influence outcomes. Invariably, however, there are factors which are unaccounted for, indeed unknowable, which could be influencing the results.

Individuals who were avidly trying to follow a healthy lifestyle in 1985 likely had drummed into their heads the message when these questionnaires were filled out that they needed to limit egg consumption. These individuals were also likely following other healthy habits, including exercising more, smoking less, and eating more fruits and vegetables and less junk food.

Observational studies cannot account for all these confounding variables.

At science-media centre.org they do a fantastic job of having independent experts in the field present their evaluation of scientific studies which have been popularized in the media. For the JAMA egg study their analyses can be found here.

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University emphasized the problem with residual confounding :

That’s because, for instance, there will be many other differences between people that eat many eggs and people that eat few other than their egg consumption.  These other differences might be what’s causing higher death rates in people who eat a lot of eggs, rather than anything to do with the eggs themselves.  The researchers point out that this has been a particular problem in some previous studies, and that this may have been a reason for inconsistency in the results of those studies.  They have made considerable efforts to allow statistically for other differences in the new study.  But they, correctly, point out that their own study is still not immune from this problem (known as residual confounding), and that therefore it’s impossible to conclude from this new study that eating eggs, or consuming more cholesterol in the diet, is the cause of the differences in cardiovascular disease rates and overall death rates that they observed.

For observational epidemiological studies like this egg study which show increased risks that are only “modest” it is highly likely that the next such study will find something different.

Eggs Are Not Eaten In Isolation

Finally, It’s important to remember that eggs, like most foods, are rarely consumed without accompanying food. This accompaniment is often bacon in the US. Eggs are often cooked in oil or butter and unless you cook them yourself you are unlikely to know the nature of the oil.

Eggs are frequently components of recipes.

We have no idea how these factors play into the results of the egg study.

So, rather than drastically cutting egg consumption I propose that there be a drastic cut in the production of weak observational nutrition studies and a moratorium on inflammatory media coverage of meaningless nutritional studies.

Eggsponentially Yours,


N.B. For a good article on this topic read Julia Volluz of Vox on why nutritional science is so messy

Breakfast is Not The Most important Meal of the Day: Feel Free to Skip it

It always irritates me when a friend tells me that I should eat breakfast because it is “the most important meal of the day”. Many in the nutritional mainstream have propagated this concept along with the idea that skipping breakfast contributes to obesity. The mechanism proposed seems to be that when you skip breakfast you end up over eating later in the day because you are hungrier.

The skeptical cardiologist is puzzled. Why would i eat breakfast if I am not hungry in order to lose weight? What constitutes breakfast? Is it the first meal you eat after sleeping? If so, wouldn’t any meal eaten after sleeping qualify even it is eaten in the afternoon? Is eating a donut first thing in the morning really healthier than eating nothing? Why would your first meal be more important than the last? isn’t it the content of what we eat that is important more than the timing?

The 2010 dietary guidelines state

eat a nutrient-dense breakfast. Not eating breakfast has been associated with excess body weight, especially among children and adolescents. Consuming breakfast also has been associated with weight loss and weight loss maintenance, as well as improved nutrient intake

The US Surgeon General website advises that we encourage kids to eat only when they are hungry but also states

Eating a healthy breakfast is a good way to start the day and may be important in achieving and maintaining a healthy weight

A recent study anayzes the data in support of the “proposed effect of breakfast on obesity” (PEBO) and found them lacking.
This is a fascinating paper that analyzes how scientific studies which are inconclusive can be subsequently distorted or spun by biased researchers to support their positions. It has relevance to how we should view all observational studies.

Observational studies abound in the world of nutritional research. The early studies by Ancel Keys establishing a relationship between fat consumption and heart disease are a classic example. These studies cannot establish causality. For example, we know that countries that consume large amounts of chocolate per capita have large numbers of Nobel Prize winners per capitaChocolate Consumption and Nobel Laureates
Common sense tells us that it is not the chocolate consumption causing the Nobel prizes or vice versa but likely some other factor or factors that is not measured.

Most of the studies on PEBO are observational studies and the few, small prospective randomized studies don’t clearly support the hypothesis.

Could the emphasis on eating breakfast come from the “breakfast food industry”?
I’m sure General Mills and Kellogg’s would sell a lot less of their highly-processed, sugar-laden breakfast cereals if people didn’t think that breakfast was the most important meal of the day.

My advice to overweight or obese patients:
Eat when you’re hungry. Skip breakfast if you want.
If you want to eat breakfast, feel free to eat eggs or full-fat dairy (including butter)
These foods are nutrient-dense and do not increase your risk of heart disease, even if you have high cholesterol.
You will be less hungry and can eat less throughout the day than if you were eating sugar-laden, highly processed food-like substances.