Justification: tonsillectomy is one of the surgical procedures performed most frequently in children, although its effectiveness is relatively low. Most throat illnesses become less frequent with increasing age. The factors that need to be considered for the indication of TE are: potential risks and benefits compared to alternative management, the natural course of the disease, the frequency and severity of infections, the preferences of the family, antibiotic tolerance, academic performance, accessibility of health care services and heath care costs.1
Clinical practice guidelines recommend differentiating between two categories of cases: severe, in which TE would be indicated (≥ 7 episodes documented in a year, ≥ 5 episodes/year in two consecutive years, ≥ 3 episodes/year in three consecutive years), and moderate, in which TE would not be indicated. The eligibility for TE must be determined on a case-to-case basis.
Otherwise, TE would only be indicated in patients with GAS infection complicated by multiple antibiotic allergy, peritonsillar abscess, a history of rheumatic fever (RF) or contact with an individual with RF.2
Validity or scientific rigour: the population under study was well defined, although there was some imprecision as regards diagnosis, since many studies use a general clinical definition (throat pain). The intervention (tonsillectomy) was well defined, but the alternative management with which it was compared was not properly explained. The outcome measures and methodology varied from study to study, so the authors could not perform a meta-analysis. The search was restricted to works published in English and excluded studies with a high risk of bias, which happened to be the studies involving patients with more severe forms of disease. Both factors limit the external validity of the study.
Clinical relevance: the authors reported a greater decrease in the number of throat infections, clinician contacts, diagnosis of infection by group A streptococcus (GAS) and school absences in the short term in patients treated with TE. These differences, while statistically significant, seem scarcely relevant from a clinical standpoint. This is a moderate effect that, furthermore, does not persist and is only observed in the first year, which may overlap with the natural decrease in episodes of tonsillitis as years go by. In one of the studies with the lowest risks of bias, the TE group had 1.74 episodes of throat pain or infection in the first year post surgery, (95% confidence interval [95 CI]: 1.54 to 2) compared to 2.93 episodes in the control group (95 CI: 2.69 to 3.22).3
There are previous studies that have not found a reduction in the number of respiratory infections following TE4 or have found only a modest decrease in recurrent pharyngitis.5 A study of the cost-effectiveness of TE compared to medical treatment in reducing the frequency of tonsillitis in school-aged children concluded that TE can reduce the overall number of episodes of tonsillitis in the two years following the procedure (95 CI: 0.61 to 5.2) at a reasonable cost (with a reduction of £261 [95 CI: 161 to 586] per prevented episode of “tonsillitis” [throat involvement]).6 Furthermore, we must take into account that the risks of severe complications of TE could exceed the risks of rheumatic fever or local suppurative complications in conservative treatment.
Applicability to clinical practice: the methodological limitations and heterogeneity of the studies reviewed, the selection bias in favour of mild and moderate forms of disease, and the impossibility of performing a quantitative synthesis of the data pose significant challenges to drawing conclusions from this study. At any rate, it seems that the beneficial effects of tonsillectomy are unimportant from a clinical standpoint and that they do not persist, so the indication for this surgery should be assessed very carefully on a case-to-case basis.
Conflicts of interest: the authors of the commentary had no conflicts of interest to declare.