Ericsson 2009.

Study characteristics
Methods	Non‐blinded, parallel‐group randomised controlled surgical trial with 2‐year duration of follow‐up
Participants	Setting: 1 university clinic, 2 county council hospitals, Sweden Sample size: Number randomised: 118 randomised (36 declined participation, 4 excluded due to randomisation error, 7 excluded due to exclusion criteria, 71 children included in study) Number completed: 67 children Participant (baseline) characteristics: Age: 4.5 to 5.5 years Gender: 58% males, 42% females Inclusion criteria: tonsil hypertrophy and sleep disordered breathing with or without recurrent tonsillitis as determined by otolaryngologist and listed for tonsil surgery Exclusion criteria: antibiotics for throat infection in past 3 months, prior treatment for peritonsillitis, record of small tonsils, complicating disease requiring special care, inability to speak Swedish, obesity and bleeding disorder
Interventions	Intervention group: cold knife tonsillectomy (TE) with blunt dissection (n = 32) Comparator group: radiofrequency tonsillotomy (TT) of protruding part of tonsils, removed to region parallel to tonsillar pillars (n = 35) Use of additional interventions: 80% had concurrent adenoidectomy; TE 78%, TT 80%
Outcomes	Short‐term outcomes (6 months): postoperative pain, postoperative bleeding, analgesic use, return to normal diet, return to normal activity, disease‐specific quality of life (OSA‐18), behaviour (Child Behaviour Checklist, CBCL), general health (questionnaire), oSDB symptom recurrence, infection rate Long‐term (2 years): disease‐specific quality of life (OSA‐18), behaviour (Child Behaviour Checklist, CBCL), general health (questionnaire), oSDB symptom recurrence, infection rate, dentofacial morphology and growth
Notes	Participants lost pre‐trial total: 40% pre‐trial dropout (method of Zelen; 36 declined participation, 4 excluded due to randomisation error, 7 excluded due to exclusion criteria) Participants lost to short‐term follow‐up: 6% dropout before 6 months (3 spontaneous recovery and 1 declined surgery). Participants lost to long‐term follow‐up: 4% dropout at 2 years (3 did not attend ENT examination)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "The children were initially randomized from the existing ordinary waiting list for tonsil surgery either to tonsillotomy using radiofrequency technique or to regular Tonsillectomy (TE), according the method of Zelen"
Allocation concealment (selection bias)	Low risk	Quote: "The randomization procedure was implemented using a computer generated sequentially numbered list. An independent person drew from this list and assigned even numbers to TT and odd numbers to TE."
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not blinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias) All outcomes	High risk	40% pre‐trial dropout (method of Zelen; 36 declined participation, 4 excluded due to randomisation error, 7 excluded due to exclusion criteria) 6% loss to short‐term (6 months) follow‐up 4% loss to long‐term (2 year) follow‐up
Selective reporting (reporting bias)	High risk	11‐item questionnaire not fully reported
Other bias	Unclear risk	Baseline characteristics: not balanced for gender in tonsillectomy group (M 69%:F 31%) ITT analysis: performed Formal sample size calculations: performed Use of co‐interventions: similar across groups (concurrent adenoidectomy performed in 78% TE, 80% TT)