Papastavros 1989.

Methods	Three‐arm, non‐blinded, parallel‐group RCT, with 3‐week duration of treatment and follow‐up
Participants	Location: Greece, 1 site Setting of recruitment and treatment: Department of Otolaryngology, General Hospital, published in 1989 Sample size: for whole trial 90 patients (119 ears) = 60 patients with one discharging ear and 29 patients with discharge from both ears Number randomised: 21 (ears) in hydrogen peroxide group, 27 (ears) in borax powder group Number completed: 21 (ears) in hydrogen peroxide group, 27 (ears) in borax powder group Participant (baseline) characteristics: (only available for the whole study including the systemic antibiotics arm ‐ 90 participants) Age: median 49 years; range 11 to 79 years Gender (F/M): 49 (54%)/41 (46%) Main diagnosis: patients with discharging ears High‐risk population: no Cleft palate (or other craniofacial malformation): Not reported (NR) Down syndrome: NR Indigenous groups (Australian Aboriginals/Greenland natives): NR Immunocompromised: NR Diagnosis method: Confirmation of perforated tympanic membrane: unclear Presence of mucopurulent discharge: NR Duration of symptoms (discharge): at least 6 months Other important effect modifiers: Alternative diagnosis of ear discharge: of the 119 participants included across the 3 treatment groups (systemic antibiotics, hydrogen peroxide and borax powder), it is reported that 28 (24%) had a diagnosis of cholesteatoma and/or osteitis Number who have previously had grommets inserted: NR Number who have had previous ear surgery: 68, 4 had mastoidectomy Number who had previous antibiotic treatment for CSOM: NR Inclusion criteria: Chronic suppurative otitis media, either with: persistent drainage for at least the previous 6 months; OR drainage at the first visit and a history of at least three recurrences during the previous 12 months. Exclusion criteria: Non‐suppurative cases and cases of questionable chronicity
Interventions	Hydrogen peroxide (n = 21): hydrogen peroxide as ear drops. Dose and frequency of administration unknown. Duration: 10 days if unsuccessful or additional 10 days after successful outcome. Borax powder (n = 27): borax powder insufflation. Dose, frequency and method of insufflation were not given. Duration: 10 days if unsuccessful or additional 10 days after successful outcome The mean duration of treatment for treatment was approximately 19 days. Concurrent treatment: "toileting and debridement of ear was given as necessary"
Outcomes	Outcomes of interest in the review: Primary outcomes: Resolution of ear discharge ("dry ear"), measured at between 1 week to 2 weeks. Unclear if otoscopically confirmed.
Funding sources	No information provided
Declarations of interest	No information provided
Notes	Unit of randomisation: not specified Methods for including patients with bilateral disease: not specified. There were 90 patients (119 ears) but results were presented by ear. This was a 3‐arm trial designed to compare systemic antibiotics (based on culture and sensitivity) versus "ototopical agents" which were randomly divided between hydrogen peroxide and borax powder. The study is not included in the systemic antibiotics reviews because it allocated patients with high risk of complications to this group (i.e. not randomised).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quote: "In the ototopical group, each patient was started either on instillation of H2O2 drops or insufflation of borax powder. The determination was again made at random… Some patients, assigned initially to one treatment group were transferred to the other as new cases if the therapeutic outcome was unfavourable." Comment: no information about method of sequence generation. However, there is clear evidence that the study did not follow the principle of randomisation by only allocating patients with more serious risk of complications to the systemic antibiotic group. The study also allowed patients who did not respond to be transferred to another group – which breaks randomisation. Unit of randomisation was not specified, and the number of patients in each intervention not specified.
Allocation concealment (selection bias)	High risk	Quote: "The determination was again made at random." "Unsuccessful outcomes: patients were either transferred to the other treatment modality or were released from the study." Comment: no information about method of allocation concealment in the initial treatment allocation, and how many patients were transferred to other treatment modality
Blinding of participants and personnel (performance bias) All outcomes	High risk	Comment: the study was not blinded. No mention of placebo.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: open‐label study. No details on assessor blinding.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Comment: no statements about loss to follow‐up. Unclear whether there were losses ‐ number of patients per treatment arm not reported; number of ears between treatment arms not balanced.
Selective reporting (reporting bias)	High risk	Comment: no protocols were found on the World Health Organization clinical trials registry. The study stated several important criteria and assumptions for the measurement of success/failures in the methods section, but no information was provided in the results section. Not all criteria for responses to treatment were fully reported, i.e. recurrence.