High intake of fat and a low intake of foods rich in antioxidants and fibre are suggested to be associated with risk of colorectal adenomas. Inconsistency may, however, be due to dietary assessment problems or differences in the identification of cases and controls. We have compared 87 adenoma cases aged 50-76 years with 35 healthy controls and 35 'hospital' controls. All the controls were matched for sex and age (+/-5 years) and proven to be free of polyps. Current habitual diet was measured by a 5-day dietary record by weighing. Regarding the intakes of vegetable fat, protein, cholesterol, vitamin A, total vitamin D, edible fats, coffee and fish and fish products, the outcomes of the analysis depended upon the source of controls. However, an increased adenoma risk compared with either set of controls related to a low consumption of vegetables, cereals, iron, vitamin C and fibre and a high intake of total fat was found. This is suggestive of substantial differences, since bias due to abdominal symptoms, the dietary records or an inappropriate choice of controls, would have affected the estimations. The findings give further strength to the role of these dietary factors in the formation of precancerous lesions in the large intestine.