Inman 1984.

Methods	Country where data collected: Canada Parallel‐group RCT Unit of randomisation: participant Unit of analysis: participant Duration: NR (duration of healing up to mean 26 days)
Participants	Inclusion criteria: age > 1 year; full‐thickness burns; < 24 h post injury Exclusion criteria: prior topical antibiotic treatment, pregnant, allergic to sulfa drugs Participants: 121 analysed, N randomised unclear Mean age (years): 31 ± 21 vs 33 ± 25 Male participants: NR Burn type: flame 35 vs 38; scald 8 vs 20; electrical contact 3 vs 1; other 8 vs 8 Burn degree: full‐thickness Burn size (%TBSA): full‐thickness 13 ± 16 vs 10 ± 11 Burn location: perineal 10 vs 9 (9 vs 5 full‐thickness); inhalation injury 10 vs 16 (ventilator 7 vs 9)
Interventions	Intervention arm 1: SSD (1%) plus chlorhexidine digluconate (0.2%) cream (Silvazine); "buttered on to wound and/or wound dressed with "buttered" cotton gauze. 54 participants Intervention arm 2: SSD (1%) (Flamazine) buttered on to wound and/or wound dressed with "buttered" cotton gauze. 67 participants Cointerventions: antibiotics as appropriate; daily bathing with non‐antibacterial soap and wound debridement, wound excision as appropriate
Outcomes	Primary outcome: wound healing Primary outcome: infection Secondary outcome: mortality (overall, infection‐related) Secondary outcome: adverse events
Notes	A list of exclusions are presented that appear likely to account for post‐randomisation withdrawals, number randomised unclear Funding: British Columbia Professional Firefighters Association; Smith & Nephew Canada An additional paper (Snelling 1991) reported additional participants but it appeared that these participants were not randomised to the intervention groups and so are not reported here. The reference is provided as a secondary citation for the study. Funding: British Columbia Professional Firefighters Association and Smith and Nephew Canada
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: “Patients were randomly assigned to receive either Silvazine or Flamazine”. Comment: no detail on randomisation methods
Allocation concealment (selection bias)	Unclear risk	Quote: “Patients were randomly assigned to receive either Silvazine or Flamazine”. Comment: no detail on allocation concealment
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: “Wounds were cultured with a swab once or twice weekly with twice weekly cultures being taken from most patients whose wounds involved more than 10 per cent of the body surface. Surface cultures were obtained at each culture session. Full‐thickness burn wound biopsies were also obtained, and examined for histological evidence of bacterial invasion into dermis or fat and quantitative bacterial counts determined.” Comment: no information on blinding of assessors
Incomplete outcome data (attrition bias) All outcomes	High risk	“Patients who did not survive for 7 days, who had all eschar excised before day 7, who were discharged before day 7 or who went on to heal all of what was initially diagnosed as the full‐thickness component of the burn wound were excluded from the study group.” Comment: excluded participants would more usually be handled as part of an ITT population. As such, their exclusion is a potential source of bias.
Selective reporting (reporting bias)	Unclear risk	Quote: “The clinical assessment of study wounds included inflammation (secretion, reddening, coating) and healing tendency (very good, good, moderate, none).” Comment: some uncertainty about the above statement – the word “included” implies there may possibly have been more outcomes assessed
Other bias	Unclear risk	Comment: no direct quotes but no evidence of additional sources of bias, but reporting insufficient to be certain