OBJECTIVE: To establish whether there is an association between subsequent fracture of the outlet strut and welder identity and the weld date of Björk-Shiley convexo-concave (BScc) valves. DESIGN: Cohort study. PATIENTS: All Dutch BScc valve recipients (n = 2266). MEAN OUTCOME MEASURES: Documented outlet strut fracture during follow up. RESULTS: Weld dates were known for 97.0% of all BScc valves (n = 2534) implanted in Dutch patients (n = 2266) and welder identity was known for 52.2%. During a mean follow up of 9.4 years, 46 fractures were documented. For 60 degrees valves welded from 1981 to 1984 the fracture rate (0.22 per 100 person-years (95% CI 0.13 to 0.34)) was higher than that for valves welded before 1981 (0.04 (95% CI 0.01-0.10)). When all fracture related risk factors were taken into account, fracture rates per welder did not show any statistically significant differences. CONCLUSIONS: Welder identity does not contribute to the risk of strut fracture. Other factors in the production of BScc valves may explain the increased risk for valves welded from 1981 to 1984.