Postoperative bleeding after diathermy and dissection tonsillectomy

OBJECTIVE: To analyze the incidence and pattern of bleeding after tonsillectomy performed by either cold dissection or diathermy. DESIGN: A prospective, nonrandomized cohort study of postoperative hemorrhage after tonsillectomy. METHODS: Monthly reporting of the number of tonsillectomies and postoperative bleeds from otolaryngologists working in rural areas of Victoria, Australia over a 2.5 year period. Criteria for bleeding were either 1) repeat anesthesia and surgery because of hemorrhage (including return to theater from the recovery room), or 2) readmission to hospital because of bleeding, or 3) blood transfusion to replace blood loss. Main outcome measures were the incidence, volume, and time course of postoperative hemorrhage. RESULTS: The number of bilateral tonsillectomies with removal by cold-blunt dissection was 3,087. In this group, there were 57 (1.85%) bleeds. The number of bilateral tonsillectomies with removal by diathermy dissection was 1,557. In this group, there were 37 (2.38%) bleeds. If cold dissection is taken as the "control" and diathermy tonsillectomy as the "treatment" group, the relative risk of bleeding after diathermy tonsillectomy is 1.30 (95% confidence interval 0.88-1.93). The pattern of bleeding after each technique differs significantly over time, with more reactionary bleeds in the dissection group and more bleeds between 4 to 7 postoperative days after diathermy. When bleeding occurred, it was in excess of 500 mL in 16% of dissection cases and 43% of diathermy tonsillectomies. CONCLUSIONS: The difference in the risk of bleeding after each technique did not reach statistical significance, but the temporal pattern of hemorrhage differed, and more bleeds exceeding 500 mL were seen in the diathermy group.