Background The International Study Group for Liver Surgery (ISGLS) proposed a definition for bile leak after liver surgery. A multicentre international prospective study was designed to evaluate this definition. Methods Data collected prospectively from 949 consecutive patients on specific datasheets from 11 international centres were collated centrally. Results Bile leak occurred in 69 (7.3%) of patients, with 31 (3.3%), 32 (3.4%) and 6 (0.6%) classified as grade A, B and C, respectively. The grading system of severity correlated with the Dindo complication classification system (P < 0.001). Hospital length of stay was increased when bile leak occurred, from a median of 7 to 15 days (P < 0.001), as was intensive care stay (P < 0.001), and both correlated with increased severity grading of bile leak (P < 0.001). 96% of bile leaks occurred in patients with intra-operative drains. Drain placement did not prevent subsequent intervention in the bile leak group with a 5-15 times greater risk of intervention required in this group (P < 0.001). Conclusion The ISGLS definition of bile leak after liver surgery appears robust and intra-operative drain usage did not prevent the need for subsequent drain placement.