Skip navigation

NSAIDs, tonsillectomy and postoperative bleeding

 

Clinical bottom line

For every 100 patients undergoing tonsillectomy and treated with NSAID, 2 more will need a reoperation because of bleeding but 9 fewer will have PONV, compared with those treated with other analgesics, or opioids.


Reference

S Møiniche et al. Nonsteroidal antiinflammatory drugs and the risk of operative site bleeding after tonsillectomy: q quantitative systematic review. Anesthesia & Analgesia 2003 96: 68-77.

Review

The review sought studies that compared systemic NSAIDs given orally, rectally, or by intravenous or intramuscular injection with non NSAID controls in tonsillectomy. Trials had to be randomised, and could involve adults and children. Patients in the control group could receive a non-NSAID analgesic, placebo, or no treatment. There was a wide searching strategy.

The outcomes examined were intraoperative blood loss, and the incidence of postoperative bleeding. Reoperation because of bleeding was considered the most relevant end point. Postoperative nausea and vomiting was also examined.

Because many studies had no events, considerable attention was given to the definition of statistical significance.

Results

Data from 25 randomised studies were included. The median quality score was 4 on a scale of 0-5, and in which the minimum score had to be 1, and 21 trials were double-blind. The trials had 1,853 patients and the median trial size was 77 patients.

The main results are shown in Table 1. There was no significant difference for postoperative bleeding, or admission or readmission because of bleeding. There was a significantly raised rate of reoperation with NSAID, with an NNT of 60 (34-277). There was no significant difference in bleeding rates within the first 24 hours after operation, or later, or with preoperative use of NSAIDs.

Table 1: Main postoperative bleeding results for NSAIDs compared with controls in tonsillectomy

 

Event/total with

Percent with

 

Outcome

NSAID

Control

NSAID

Control

OR
95% CI

Postoperative bleeding

73/685

58/619

10.7

9.4

1.3 (0.9 to 1.9)

Postoperative admission or readmission

14/275

6/231

5.1

2.6

2.1 (0.9 to 5.2)

Reoperation

24/877

8/749

2.7

1.1

2.3 (1.1 to 4.8)

Nausea and vomiting

With placebo or no treatment PONV rates were 28%, with NSAIDs 32% and with opioids the PONV rate was 49%. Comparing NSAIDs with opioids, the NNT for one person to be nauseated or to vomit was 9 (5-19).

Comment

There was no opportunity to determine whether the bleeding episodes were major or minor, so the incidence of reoperation because of bleeding became a surrogate for this. It is an important clinical and economic endpoint. The evidence we have is that use of NSAIDs results in one additional reoperation for every 60 patients treated, but with a wide confidence interval. But this has to be balanced by fewer patients suffering PONV. The balance is about 2 more reoperations against 9 more cases of PONV.