Objectives To assess the blood loss accompanying TURF and investigate its association with the resected weight of prostatic tissue, type of anaesthesia, type of presentation, histology, different surgeons and their differing techniques, and thus to reduce the morbidity associated with brood loss and transfusion, Patients and methods All prostatectomies carried out in a district general hospital were audited prospectively, recording the pre- and post-operative haemoglobin concentration (Hb), blood transfusions and the variables listed above. The audit was repeated a year later. Results The peri-operative blood loss, as assessed by various indicators, was equivalent to a decrease in Hb of 10-15 g/L (8-11%). The weight of the resected prostatic tissue was the most important measurable factor in determining blood loss, Regional anaesthesia was associated with less blood loss than general anaesthesia. The added use of a suprapubic catheter for irrigation appeared to have a marginal advantage in large resections. The type of presentation, elective or otherwise, and the histological nature of the prostate did not influence blood loss. Smaller transfusions were probably avoidable in patients having smaller resections and a normal pre-operative Hb. On re-auditing, the overall transfusion rate was reduced from 10.8% to 8.2% and from 4.4% to 1% in patients having resections of <30 g. Conclusion Blood transfusion can be reduced and rationalized. Patients with a normal pre-operative Hb and undergoing resections of <30 g do not usually require transfusion. Regional anaesthesia is associated with less blood loss but its advantage is overshadowed in practice by the weight of the resected tissue. An audit of this type is repeatable and is useful in raising awareness, objectively assessing differences and advocating and assessing any changes made.