Background: Tonsillectomy is one of the most frequent surgical procedures. "Hot" tonsillectomy techniques (eg, diathermy and coblation) have become well established, but there is little evidence to suggest that they are better than traditional "cold steel" dissection with only packs or ties for haemostasis.
Methods: We obtained data for the occurrence of complications after all tonsillectomies done since July, 2003, in England and Northern Ireland. We recorded postoperative haemorrhages and other complications, taking place within 28 days of surgery, that led to delayed discharge, return to theatre, or re-admission. By February, 2004, the audit had included 13554 patients, of whom 11796 (87%) in 252 hospitals consented to electronic submission of their tonsillectomy data. Thus, 75% of the hospitals where tonsillectomies are done submitted data for 60% of all eligible patients.
Findings: Haemorrhage occurred in 389 patients (3.3%). 59 patients (0.5%) had a primary haemorrhage (during initial stay), 337 (2.9%) a secondary haemorrhage (after discharge), and seven had both. The overall haemorrhage rate was 3.1 times (95% CI 1.9-5.0) higher with bipolar diathermy tonsillectomy than with cold steel tonsillectomy without any use of diathermy (p<0.001). The corresponding relative risk for coblation tonsillectomy was 3.4 (1.9-6.2; p<0.001). When cold steel was used for dissection and diathermy only for haemostasis the relative risk was 2.2 (1.3-3.7; p=0.002).
Interpretation: The use of techniques such as diathermy and coblation increased postoperative haemorrhage. These methods should therefore be used with appropriate caution and only after proper training.