Unselected preoperative coagulation testing is known to have low positive yield. However, no study has specifically evaluated neurosurgical patients. A retrospective study of 1211 patients having neurosurgery over a one-year period was therefore conducted. Preoperative test results (activated partial thromboplastin time [aPTT], prothrombin time [PT] and platelet count) and historical factors indicating a potential bleeding tendency were recorded. Abnormality was defined as a test result outside the normal range for our laboratory. Seventeen per cent of all test results were abnormal. However, if abnormality was redefined as a test result indicating potential bleeding tendency (low platelet count, prolonged aPTT and/or PT), only 7.2% of results were abnormal. Many patients had factors on history indicating a potential bleeding tendency, but only a prolonged aPTT, cranial surgery and the use of anti-hypertensive and anaesthetic drugs preoperatively predicted postoperative bleeding. Prolonged aPTT was predictable on history in most patients. We conclude that routine screening of all preoperative neurosurgical patients in our hospital is unnecessary.