Epidemiology of Dietary Fatty Acids and Its Relation to Colorectal Cancer

Fatty Acids Colorectal Cancer

Fatty Acids and Colorectal Cancer

The article titled "Dietary Fatty Acids and Colorectal Cancer: A Case-Control Study" describes a seven-year study of close to 3000 subjects studying the relationship between fatty acids and colorectal cancer. As the study involved the analysis of data collected on colorectal patients over a seven-year period, and thus started with the outcome of colorectal cancer and traced backwards to discover a relationship between exposure to fatty acids and this outcome, it was by definition a "Case-Control" study (the title of the article also presents the study as a case-control one). The researchers began with about 1500 colorectal patients and 1500 control subjects and administered weekly questionnaires, called "semi quantitative food frequency questionnaire" involving 150 food items and the frequency the participants ate them. (Theodoratou, 2007, p.182)

This type of study is particularly useful in studying the outbreak and spread of diseases in relation to those who have not been affected. In short, it compares two groups, one with the disease and one without, and attempts to discover the difference between the two. This difference may be the determining factor involved in the cause of the disease. In the case of this article, the researchers sought to determine the affect of fatty acids on the rates of colorectal cancer by studying people who had the disease and those who did not and determining if the presence of fatty acids were higher in the diets of those with cancer than those without. However, one major flaw in this design is the fact that it requires subjects to fill out weekly questionnaires earnestly and honestly, without leaving out anything that may be questionable or embarrassing. In fact, on many occasions the researchers were forced to return questionnaires to participants because there were "forms with more than a maximum acceptable number of blank entries…" (Theodoratou, 2007, p.183)

This particular study attempted to discover the relationship between fatty acids and colorectal cancer and its results were calculated as an "Odds Ratio," or a means to measure the relationship between exposure to fatty acids and the incidents of colorectal cancer. For instance, one calculation in the article assessed the "intake of total fatty acids and trans-MUFAs as well as of the individual fatty acids palmitic, stearic, and oleic…" (Theodoratou, 2007, p.189) When calculating the intake of total fatty acids and trans-MUFAs along with palmitic acid, the odds ratio was calculated to be 1.22 with a 95% confidence interval (CI). If an OR is equal to 1.0, then it can be assume that the exposure does not affect the outcome, and if less than 1.0 then the exposure can lower the odds of the outcome. But since the OR in this case was 1.22, it can be concluded that exposure to palmitic acid increased the risk of colorectal cancer. This figure was calculated using the Wilcoxon Rank Sum test, which is used to test for the difference between two samples. In this case it was the subjects with colorectal cancer compared to those without in relation to the intake of palmitic acid. The 95% confidence ratio indicates that the observed outcome, the exposure to palmitic acid increases the odds of colorectal cancer, is within 95% of the theoretical, or predicted, outcome; meaning that 95% of the time one can expect similar outcomes.

In the results of the study the researchers use the term "P-value" and "P for trend" when discussing the statistical results of the study. This term brings into the discussion the concept of the "null hypothesis," or the idea that when something cannot be proven 100%, at least the other possibilities can be eliminated. The p-value is the calculation that a study's null hypothesis is correct and involves the type of statistical distribution that was expected. In many cases the researchers simply choose a p-value of .05, or a 5% chance that the data will demonstrate the null hypothesis to be correct. If the calculated p-value is greater than .05, or it has a less than 5% chance of being correct, then the hypothesis is wrong. In the current study the researchers "adjusted the p-value level from .05 to .0035 by applying the Bonferroni correction for 14 independent tests." (Theodoratou, 2007, p.189) This meant that their chances of being wrong dropped from 5% to just .35%; due to the fact that what they really wanted was a set of p-values in relation to the intake of different levels of different fatty acids, otherwise known as a "trend test." The p-trend values are the chances that the different fatty acids do not have a direct relation to the incidents of colorectal cancer.

It may seem that the odds ratio and the p-trend are different indicators of the same data, or that they would need to display similar statistical significance, but this is not necessarily the case. In many cases it is possible for the p-trend value to be deemed statistically significant while other calculations, such as the odds ratio, are not. P-trend values do not display the chance that the input in question is responsible for the outcome, instead it calculates the chance that it is not responsible for the outcome. This would necessarily take into account the other possible influences and calculate the chance that the input in question could also be eliminated. This is quite different from the calculation involving the odds ratio, which calculates that chance that the input is responsible for the outcome. Odds ration only takes into account whether or not the input, or in the case of this article, the input of fatty acids on whether or not there is a relationship between increases in fatty acids and the prevalence of colorectal cancer. On the other hand, p-trend values calculate not only the existence of fatty acids but the rate and amounts of these acids over time and whether or not they do not have an effect. At any one point in time the odds ratio may indicate that there is little or no relationship between fatty acids and colorectal cancer but over time the different amounts of fatty acids can become statistically important when calculating the incidents of cancer.