Does celecoxib prescription for pain management impact post-tonsillectomy hemorrhage requiring surgery? A retrospective observational cohort study.

BACKGROUND: Adenotonsillectomy and tonsillectomy (now referred to as tonsillectomy thereafter) are common pediatric surgeries. Postoperative complications include hemorrhage requiring surgery (2-3% of cases) and pain. While non-steroidal anti-inflammatory drugs are commonly administered for post-surgical pain, controversy exists regarding bleeding risk with cyclo-oxygenase-1 inhibition and associated platelet dysfunction. Preliminary evidence suggests selective cyclo-oxygenase-2 inhibitors, for example celecoxib, effectively manage pain without adverse events including bleeding. Given the paucity of data for routine celecoxib use after tonsillectomy, we investigated the association between post-operative celecoxib prescription and post-tonsillectomy hemorrhage requiring surgery using chart-review data from the Children's Hospital of Eastern Ontario, Canada.

METHODS: After ethics approval we performed a retrospective single-center observational cohort study in children <18 yrs undergoing tonsillectomy from January 2007 to December 2017. Cases of adenoidectomy alone were excluded due to low bleed rates. The primary outcome was the proportion of patients with post-tonsillectomy hemorrhage requiring surgery. The association between a celecoxib prescription and post-tonsillectomy hemorrhage requiring surgery was estimated using inverse probability of treatment weighting based on propensity scores and using generalized estimating equations to accommodate clustering by surgeon.

RESULTS: An initial patient cohort of 6468 was identified and 5846 children with complete data were included in analyses. Median (interquartile range) age was 6.10 (4.40, 9.00) years and 46% were female. In our cohort, 28.1% (n=1644) were prescribed celecoxib. Among the 4996 tonsillectomy patients, 1.7% (n=86) experienced post-tonsillectomy hemorrhage requiring surgery. The proportion with post-tonsillectomy hemorrhage requiring surgery among patients who had a tonsillectomy and were or were not prescribed celecoxib was 1.94% (30/1548; 95% CI: 1.36-2.75) and 1.62% (56/3448; 95% CI: 1.25-2.10), respectively. Modelling did not identify an association between celecoxib prescription and increased odds of post-tonsillectomy hemorrhage requiring surgery (OR=1.4, 95% CI: 0.85-2.31, p=0.20).

CONCLUSIONS: Celecoxib does not significantly increase the odds of post-tonsillectomy hemorrhage requiring surgery, after adjusting for covariates. This large pediatric cohort study of celecoxib administered after tonsillectomy provides compelling evidence for safety but requires confirmation with a multi-site randomized controlled trial.