The aim of the study was to determine dietary risk factors for fracture in men aged 46-68 years. Six thousand five hundred and seventy-six men were randomly invited using the Municipal Registry to a diet and health study. The diet was assessed using a combined 7-day menu book for hot meals, beverages and dietary supplements and a quantitative food frequency questionnaire for other foods. The fracture incidence was 103/10,000 person-years during a mean follow-up of 2.4 years. Zinc and phosphorus intake were associated with fracture risk and showed a threshold effect. The zinc intake in the lowest decentile, 10 mg daily, was associated with almost a doubled risk of fracture compared with the fourth and fifth quintiles (RR = 0.47; 95% confidence interval, 27-82) of zinc intake adjusted for energy, previous fractures, lifestyle factors and co-morbidity. Energy-adjusted phosphorus intake in the lowest quintile, mean level 1357 mg, was associated with an increased fracture risk compared with subjects in the second quintile. Smoking, martial status and physical activity were independently associated with fracture risk. Calcium, retinol and vitamin D showed no associations with fracture risk. We conclude that inadequate intakes of zinc and phosphorus are important risk factors for fracture.