Harvard School of Public Health study
Comment from Dr Carrie Ruxton, nutrition consultant to the British Egg Information Service, on the Harvard School of Public Health study:
- This is an observational study attempting to test for ‘cause and effect’ when it is not designed to do so. With this type of study, it is impossible to separate egg consumption from all the other factors which might impact on prostate cancer risk. Indeed, the researchers noted that men who consumed more eggs were fatter, less physically active, more likely to smoke and to have family histories of diabetes and prostate cancer, were less likely to eat poultry and fish, and more likely to eat red meat and dairy foods compared with those who ate few eggs. This suggests that egg consumption was a marker for other more important lifestyle factors which influence prostate cancer.
- The Food Frequency Questionnaire (FFQ) is an inaccurate tool for estimating habitual dietary intake. Men completed this every 4 years and the average from three FFQs in 1986, 1990 and 1994 was used to represent 'usual diet'. The portion sizes were not weighed, which represents another source of error. It is notable that, when the researchers took into account FFQs from later years, the association between eggs and prostate cancer mortality became non-significant (P=0.14).
- 199 men out of the total sample of 27,000 went on to die from prostate cancer with only 55 of these (0.2% of the total sample) consuming more than 2.5 eggs per week. This is a very small sample size for estimating the impact of dietary patterns on disease risk. Certainly, dietary advice should not be given to the wider population based on this sample. Interestingly, egg consumption measured after diagnosis of prostate cancer was not associated with risk of mortality.
- The statistical analysis did not take into account other dietary factors, e.g. total fat, saturated fat, sugars, which may be linked with prostate cancer development.
- As noted by the authors, only three previous studies have examined associations between mortality from prostate cancer and egg consumption. One showed a positive association, one showed no association, and a third showed an inverse association! Therefore, the evidence is highly inconsistent, suggesting that eggs in themselves are probably not related to cancer risk but are a marker for other associated factors.
- The authors blamed the cholesterol content of eggs for the association. Yet, controlled trials show that a high intake of eggs does not significantly increase blood cholesterol levels. This means that the cholesterol in eggs has a negligible impact on blood cholesterol levels, and is unlikely to impact on cancer risk. Choline was also suggested as a potential mechanism, yet there are no controlled trials to confirm this – only one observational study.