We feel that the paper by Ewah and colleagues [1] should not pass without comment. Based on the prospective assessment of 100 children they suggest that if their standard paediatric day case protocol is used, children undergoing tonsillectomy will have well-controlled postoperative pain, a reduced incidence of vomiting, and minimal postanaesthetic and surgical morbidity.

Our own experience based on a prospective audit of children following tonsillectomy undertaken within the last year, using a very similar peri-operative analgesic regimen and the same pain assessment tool, gave very different results.

We prospectively studied 60 children (mean age 6.8 years, range 1–13 years) undergoing tonsillectomy (with or without adenoidectomy, and with or without grommet insertion). All children received a loading dose of diclofenac 1−1 and paracetamol 30–40−1 rectally. They then received ibuprofen 5−1 and paracetamol 15−1 orally four times a day for 1 week. The day of operation was called Day 0, and pain was assessed on a daily basis up to and including Day 5. All parents and all children aged 5 years or above were asked to score their pain each day for 5 days, once per day (first thing in the morning). They were asked to record using the Faces Scale the ‘worst’ and the ‘usual’ level of pain they had experienced in the previous 24-h period, as described by Cummings et al.[2]. The families were phoned on two occasions to check the results, to confirm analgesic consumption and to enquire if they had needed to contact any agency for advice. The extent of postoperative pain experienced is shown in Table 1.

Table 1.   Numbers of children reporting pain as severe (Faces 4 or 5).
 Postoperative day
Usual level of pain in  the previous 24 h 6% 6% 6% 6%12%
Worst level of pain in  the previous 24 h26%55%41%50%47%

Nineteen of the 60 families had sought advice from medical agencies by the end of the 5th postoperative day, with 14 families seeing their general practitioner.

Both our patients and the Epsom patients are of similar ages and underwent a ‘hot’ tonsillar technique (removal of the tonsils by dissection with haemostasis secured predominantly using diathermy). Unlike the Epsom group, a proportion of the operations in our group were undertaken by surgeons in training, but subanalysis of our data showed no difference in postoperative pain between surgeons of different experience. The only difference in analgesia prescribed was that our patients did not receive codeine. We think this is unlikely to explain the difference for two reasons. Firstly, codeine is a prodrug and a significant minority of the UK population are unable to convert this to useful levels of morphine [3]. Secondly, a prospective, randomised double blind study failed to demonstrate any difference in postoperative pain relief between paracetamol alone and paracetamol with codeine in children following tonsillectomy [4].

We believe that the difference in results has arisen because Ewah and colleagues did not adequately assess postoperative pain. The timing of the pain assessment and specific questions asked are not described. If they undertook a single point assessment of severity of pain on a daily basis, then their results are very similar to ours for ‘usual’ level of pain, with most children appearing to have reasonable pain control. It is only when a more thorough assessment of the pain experience is undertaken that the true picture emerges. Our families reported good pain relief during the day, but many children waking between 02.00 and 03.00 hours with severe pain, commonly earache. When this had occurred for two or three nights in a row, they were driven to seek advice on the pain from their general practitioner. Ewah and colleagues only assessed pain for 3 days following surgery and do not report the number of families seeking additional advice. It is recognised that pain after tonsillectomy is considerable for more than 7 days after the procedure [5]. Other studies have reported similar levels of postoperative pain as ourselves [5, 6], with similar levels of unscheduled visits to medical practitioners.

We suggest, therefore, that their main conclusion that their day case technique results in minimal postoperative morbidity is not proven, as postoperative pain assessment follow-up was inadequate and incomplete.


  1. Top of page
  2. References