Data from an Italian multicentre case-control study on ovarian cancer were used to analyse the relationship between various types of fibres and ovarian cancer risk. The study, conducted between 1992 and 1999, included 1031 cases of incident, histologically-confirmed epithelial ovarian cancer. Controls were 2411 women admitted to the same network of hospitals for acute, non-malignant, non-hormonal-related diseases. Cases and controls were interviewed using a validated food frequency questionnaire (FFQ). Odds ratios (ORs), and the corresponding 95% confidence intervals (CI), were estimated using unconditional multiple logistic regression models. For total (Englyst) fibre, the OR for the highest versus the lowest quintile of intake was 0.68, and the continuous OR for the difference between the 80th and the 20th percentile of intake was 0.87. For most types of fibre, the continuous OR was significantly below 1. The OR was 0.83 for cellulose, 0.89 for soluble non-cellulose polysaccharides (NCPs), 0.86 for total insoluble fibre, 0.92 for insoluble NCP, and 0.95 (non-significant) for lignin. The inverse association was consistent across strata of age, family history and menopausal status, even if the association was apparently stronger in postmenopausal women. When fibre was classified according to the source, vegetable (but not grain) fibres, showed a significant protective effect, with an OR of 0.78.