Watchful waiting is useful for children with recurrent throat infections

BMJ 2004; 329 doi: (Published 16 September 2004) Cite this as: BMJ 2004;329:654
  1. Paul Little, professor of primary care research (psl3{at}
  1. 1 Primary Medical Care, Community Clinical Sciences Division, Southampton University, Aldermoor Health Centre, Southampton SO16 5ST

    Van Staaij et al's study is a welcome addition to a controversial subject.1 It shows that for children with moderately frequent throat infections (on average three in the previous year) a “wait and see” approach results in acceptable control of symptoms and avoids postoperative pain and complications (1% requiring operative surgery for haemorrhage, and 2.6% having severe nausea or dehydration). The major limitation of the study is the large number of children from the watchful waiting group who had tonsillectomy (34%). Since a per protocol analysis was not done—that is, comparing those who had tonsillectomy with those who did not, controlling for severity indices—it cannot be concluded that tonsillectomy in itself is ineffective but simply that immediate tonsillectomy is not effective. The data from this trial, however, match data from a similar trial, which reported little symptomatic benefit and a significant rate of complications (7%) among children who had tonsillectomy for more severe symptoms.2

    Should children with more severe symptoms be offered surgery? With the normal caveats about subgroup analysis, there was some evidence from Van Staaij et al's trial that those more severely affected (three or more infections a year) had some benefit from immediate tonsillectomy—one less episode of sore throat. The earlier Paradise trial assessed tonsillectomy among selected children with severe symptoms3—the “Paradise” criteria of seven or more throat infections in the preceding year, or five or more a year for each of the preceding two years, or three or more a year for each of the preceding three years. This trial showed a reduction of around one episode, rated as moderate or severe (3 of 38 surgical patients v 41 of 35 controls); however, the trial was small and was criticised by the Cochrane review for imbalances of important baseline characteristics (the author argued that this was unlikely to affect inferences).4

    Given the paucity of evidence and controversy about existing evidence, more data are clearly needed on tonsillectomy among children with recurrent throat infections, and particularly data on non-surgical approaches. Until this evidence is available it would be reasonable for doctors to share with parents the probable benefits of surgery—among children with the Paradise criteria, one less episode of moderately severe or severe sore throat a year; among children with at least three infections in the past year, one less episode of sore throat a year—but also the important harms of operation—a complication rate of 4-7%. For the remaining children, doctors should probably not offer tonsillectomy.


    • Competing interests. PL has been paid for two consultancy sessions from Abbott Pharmaceuticals for antibiotics for complications of respiratory tract infections.


    1. 1.
    2. 2.
    3. 3.
    4. 4.