Classification of the components of the MedDietScore in relation to colorectal cancer prevalence.

1593 Background: Mediterranean diet has been proposed to have a protective effect towards chronic diseases, including various types of cancer. This study aimed to evaluate whether Mediterranean diet as a whole or its specific components are associated with colorectal cancer, which increases worldwide. Methods: This is a case-control study. Between December 2009 and July 2010, 234 cases (aged 63±11, 60% males) with first diagnosis of colorectal cancer were consecutively selected from our hospitals. At the same period 250 subjects (controls) (aged 55±12, 45% males) with no clinical symptoms, signs or suspicious of any type of cancer in their medical history, were group matched to the patients by age (±5 years) and gender. The number of the enrolled subjects was decided through power analysis. Adherence to the Mediterranean diet was accessed using MedDietScore; a dietary index, validated for cardiometabolic diseases, that incorporates inherent characteristics of this pattern (theoretical range 0-55). Multi-adjusted logistic regression and linear discriminant analysis using Wilks’ lambda (ranges 0-1, the lower values, the better), applied. Results: Adherence to the Mediterranean diet was associated with lower likelihood of colorectal cancer (for 1-unit increase in the MedDietScore, OR==0.87, 95%CI 0.83-0.92). Linear discriminant analysis revealed that poultry consumption (Wilks’ lambda == 0.938, p<0.001), followed by potatoes (λ== 0.951, p<0.001), alcohol (λ== 0,968, p<0.001), cereals (λ== 0,976, p==0.001), red meat (λ== 0,979, p==0.001), vegetables (λ== 0.984, p==0.005) and dairy products (λ== 0.988, p==0.016) had the best discriminating ability in predicting colorectal cancer cases. Fruit, fish, legume and olive oil consumption were not found to have any significant discriminating ability. Total score, i.e. the MedDietScore, had λ== 0.960 (p<0.001). Conclusions: The Mediterranean diet has a protective effect against colorectal cancer, whereas focusing on its components meat, potatoes, cereals, vegetables, dairies and alcohol consumption had the best discriminating ability; these findings may help to further explore the association of this dietary pattern with the pathophysiology of colorectal cancer.