To investigate the association between dietary behaviours and colorectal cancer (CRC) in the context of the Mediterranean diet.Case-control study.All patients (cases) were recruited from Saint Savvas Cancer Hospital and Alexandra General Hospital in Athens, Greece. Controls were voluntarily selected from the general population and matched to cases by age group (±10 years) and sex.Two hundred and fifty cases with newly diagnosed CRC (mean age 63 (sd 12) years, 59·6 % males) and 250 controls matched on age and sex were studied. A standardized questionnaire assessing sociodemographic, clinical, lifestyle, dietary characteristics and nutritional behaviours was applied. Multiple logistic regression analysis was used to evaluate the aforementioned factors in addition to the MedDietScore (an index that evaluates adherence to the Mediterranean diet) on CRC development.The higher the daily number of meals, the lower the likelihood of having CRC (OR = 0·74, 95 % CI 0·61, 0·89); coffee drinking was associated with higher likelihood of having CRC (OR = 3·27, 95 % CI 1·09, 9·8); the use of non-stick cookware was positively associated with CRC (OR = 1·57, 95 % CI 1·02, 2·4). However, these associations slightly lost their significance when adherence to the Mediterranean diet was taken into account. Moreover, a 1/75 increase in the modified-MedDietScore plus the aforementioned nutritional behaviours was associated with 13 % lower odds (95 % CI 0·83, 0·91, P < 0·001) of having CRC.Nutritional behaviours in addition to dietary habits should be taken into account in detecting individuals prone to the development of CRC.