Studies on fibre intake and incident colorectal cancer (CRC) indicate inverse associations. Differences by tumour stage have not been examined. We examined associations between fibre intake and its sources, and incidental CRC. Separate analyses were carried out on the basis of sex, tumour location and the Tumour, Node, Metastasis (TNM) classification. The Malmö Diet and Cancer Study is a population-based cohort study, including individuals aged 45–74 years. Dietary data were collected through a modified diet history method. The TNM classification was obtained from pathology/clinical records and re-evaluated. Among 27 931 individuals (60 % women), we found 728 incident CRC cases during 428 924 person-years of follow-up. Fibre intake was inversely associated with CRC risk (P
trend=0·026). Concerning colon cancer, we observed borderline interaction between fibre intake and sex (P=0·052) and significant protective association restricted to women (P
trend=0·013). Intake of fruits and berries was inversely associated with colon cancer in women (P
trend=0·022). We also observed significant interactions between intakes of fibre (P=0·048) and vegetables (P=0·039) and sex on rectal cancer, but no significant associations were seen between intake of fibre, or its sources, in either of the sexes. Except for inverse associations between intake of fibre-rich cereal products and N0- and M0-tumours, we did not observe significant associations with different TNM stages. Our findings suggest different associations between fibre intake and CRC depending on sex, tumour site and fibre source. High fibre intake, especially from fruits and berries, may, above all, prevent tumour development in the colon in women. No clear differences by TNM classification were detected.