27 October 2015
The recent World Health Organization (WHO) report concerning processed meats and their links with colorectal cancer has been much in the news over the last few days.
Processed meats are those that have been modified to preserve it or change the taste via methods such as smoking, curing, or adding salt or preservatives. The report found that processed meats such as bacon, ham and sausages increased the risk of colorectal cancer, citing the example of a daily intake of 50g which was reported to increase the risk of colorectal cancer by 18%. It also concluded that there was less supporting evidence but that red meats were ‘probably carcinogenic’. At a population level, these results are clearly correct. But what do they mean?
Commentators have spoken about the confusion between strength of evidence and degree of risk, and between relative and absolute risk. These points are not just important but essential, and they are extremely well made in these two blogs. They illustrate the difficulty of communicating risk effectively to a lay public.
But, these points aside, there is a third point that will become of increasing importance as personalised medicine begins to impact on clinical practice and public health. It is the understanding that populations are heterogeneous, that literally each member of that population is biologically different from every other member. As a consequence, an individual’s risk of developing a disease could therefore be very different from the average risk as determined from population based epidemiological studies. Indeed, for some individuals the risk factor in question may not have any adverse effect at all, whereas for some, the sensitivity to that factor may be much higher than that of the average person.
The results from epidemiological studies such as those that have informed the WHO report are the prime source of evidence that a risk factor X contributes to the risk of developing disease Y. The statement that processed meats increase the risk of colorectal cancer is a statement that the average risk of colorectal cancer in a population that consumes processed meats is greater than the average risk in a population that does not. However, not everyone exposed to processed meats will develop colon cancer.
As yet, we are unclear about the predisposing factors that determine why some exposed to such foods will develop cancer and others similarly exposed will not. But from this comes the realisation that the forces that determine whether someone in a population highly exposed to X will or will not develop Y, what eminent epidemiologist Geoffrey Rose called the ‘causes of incidence’ ,are different to the forces that explain why those exposed to X have a higher chance of developing Y, Rose’s ‘causes of cases’.
We can at present only understand this point at a conceptual level, but in time biomarkers and other factors will be used to segment populations into different subgroups that will allow the risk of these sub-populations to be estimated. We may eventually be able to establish accurately who may eat processed meats with impunity, and who may not because their biology renders them particularly sensitive to the oncogenic components of such foods. This would permit much more accurate and personalised approach to disease prevention.
For the time being, however, we can but think on these matters and based on the best evidence decide what we chose to eat or not. Moderation in all things is perhaps still the best health advice that we can give to date, but in thinking about the future, the heterogeneity of populations is not to be ignored.