Perioperative use of laboratory coagulation assays is limited by the delay in obtaining results. The management of haemostasis during major surgical procedures requires rapid and accurate measurement of the prevailing coagulation status. In this prospective study, we have evaluated the reliability of on-site prothrombin time assessed by the portable coagulation monitor CoaguChek-Plus compared with standard laboratory assays during elective non-cardiac surgery. Sixty-two patients were assigned to one of three groups: group A = normal preoperative coagulation where minor intraoperative blood loss is expected; group B = normal preoperative coagulation where major intraoperative blood loss is expected; and group C = preoperative anticoagulation and minor intraoperative blood loss expected. On-site prothrombin time and laboratory prothrombin time showed poor correlation in group A (r2 = 0.24; bias (2 SD) 1.80 (3.34) S) and group B (r2 = 0.30; 1.43 (3.12) S). The correlation in group C was better (r2 = 0.71; 1.41 (1.92) S). We conclude that prothrombin time measured with the CoaguCheck-Plus monitor did not appear to be suitable for the management of haemostasis.