Media reports this week have been buzzing over the link between egg consumption, smoking and heart disease. The study has already received much criticism, many because causation (egg intake) does not equal causation (plaque build-up), and because, as has been noted by the authors, these results do not take into account saturated fat intake (or indeed dietary measures of any other kind other than eggs), exercise, or alcohol consumption. Furthermore, only one dietary questionnaires was used (upon referral to the clinic). Yes, it was based on the participants’ memory of their food intake, which has been pounced upon, but this is typical of most nutrition studies and very difficult to get around (you can’t keep people in a lab for 2 years and feed them through test tubes, after all).
However, many studies looking at the impact of nutrition suffer from similar faults, and that doesn’t mean the study is invalid. Aside from the obvious ones mentioned above, there are other factors about this study that concern me more. David Spence and colleagues report a correlation between ‘egg-years’, defined as the number of egg yolks eaten per week times the length of time this amount had been eaten, and the size of carotid plaque area, which is a significant predictor of cardiovascular risk (1). This association held up even after authors adjusted their results for the number of cigarette years (number of packs/wk multiplied by the number of years smoking). The authors also adjusted for age, since older people have larger plaque areas. A total of 1231 individuals were considered in the analysis.
Other factors that the authors include in the multiple regression analysis that were significant were body mass index (BMI), total cholesterol, systolic blood pressure, gender, and diabetes. Triglycerides and HDL or LDL cholesterol separately were not significant.
The authors state that ‘egg-yolk years was more predictive than fasting cholesterol or BMI’; however, BMI and total cholesterol are actually negatively associated with plaque size. That is, individuals with a higher BMI have smaller plaques. This seems a bit puzzling, although the authors note in the article that the association between dietary and serum cholesterol is not always positive. “I suspect the negative associations…may be related to interactions with age and diabetes, and maybe sex” notes Spence.
The authors report that a large number (almost 40%) of all subjects did never smoked. Spence notes that even within the non- and low-smoking subjects (less than 1 pack year), “egg yolk consumption in the top quintile increased plaque area significantly”.
In the analysis, the ‘egg-years’ variable was used rather than a frequency of egg consumption. This could be problematic as egg consumption is now confounded with age, which is definitely shown to be associated with plaque size. The authors claim that the partial correlation adjusted for age was not significant; however a more satisfying comparison would have been to consider frequency of egg consumption and remove time from the variable altogether.
Perhaps no other food has been subjected to such alternating vilification and heraldry than the egg, partially because the association between cholesterol and cardiovascular disease is inconsistent among studies. In a large study published in 1999 looking at two prospective cohort studies, the Health Professionals Follow-up Study and the Nurses' Health Study (2), no association between number of eggs per week and coronary heart disease (CHD) were found in 1805 individuals. However, a mildly significant association was found between egg consumption and increased risk of CHD in diabetics. This study included 3 dietary questionnaires, but all three were early in the study and did not reflect possible changes in diet over a 20-year span. In an analysis of the Physicians Health Study cohort, again no risk was found between CHD and egg consumption frequency, but there was a significant association between mortality and egg consumption frequency in both healthy and diabetic subjects (4). This study, based on >20,000 patients, had the benefit of six dietary questionnaires over the entire course of the study, along with detailed information about many other potential co-founding factors and dietary intake. In a follow-up of the Nurses Health Study, authors found a significant association between dietary cholesterol and overall mortality among the almost 5,000 participants deceased, although they did not specifically calculate egg intake (5).
Spence and colleagues have previously stated their opinion that egg intake has a negative outcome on vascular diseases and caution that looking at serum cholesterol alone does not adequately describe the effect of dietary cholesterol and cannot be used as a surrogate, which is often done (3). Clearly the authors approached this current study with a bias towards their results, and I find it very puzzling that such a narrow investigation was undertaken without assessing additional variables given the wealth of data that could have been available. Spence noted that a follow-up piece with additional analyses would be published in the near future as a response to the apparently many Letters to the Editor this work engendered.
1. Spence JD, Jenkins DJA, Davignon J. 2012. Egg yolk consumption and carotid plaque. Ahterosclerosis 2. Hu FB, Stampfer MJ, Rimm EB, Manson JE, Ascherio A, Colditz GA, Rosner BA, Spiegelman D, Speizer FE, Sacks FM, Hennekens CH, Willett WC. 1999. A prospective study of egg consumption and risk of cardiovascular disease in men and women. JAMA 281:1387-94. 3. Spence JD, Jenkins DJA, Davignon J. 2010. Ieatary 4. Djoussé L, Gaziano JM. 2008. Egg consumption in relation to cardiovascular disease and mortality: the Physicians' Health Study. J Clin Nutr 87:964-969. 5. Baer HJ, Glynn RJ, Hu FB, Hankinson SE, Willett WC, Colditz GA, Stampfer M, Rosner B. 2011. Risk factors for mortality in the nurses' health study: a competing risks analysis. Am J Epidemiol 173:319-29.