Horn 2005.

Study characteristics
Methods	This study was carried out in a secondary care setting. The participants were blinded, but not the investigators. Intention‐to‐treat analysis was not carried out. This study was conducted in the USA.
Participants	233 participants were recruited: 32 dropped out. Inclusion criteria of the trial The age group was not specified Mostly refractory warts The site was not specified
Interventions	Intralesional skin test antigens versus Intralesional interferon‐alpha 2b versus Antigen plus interferon‐alpha 2b versus Placebo (saline) All subjects received injections every 3 weeks into the same wart until complete clearing of the treated wart was achieved or for a maximum of 5 treatments.
Outcomes	Outcomes of the trial > 75% reduction in surface area of warts during the trial only
Notes	Only 1 index wart was treated per participant. Participants were evaluated at each episode of treatment, and there was no long‐term follow up. Quote: "The initial design included treatment arms using GM‐CSF instead of interferon alfa‐2b. Because of serious adverse events experienced by subjects receiving GM‐CSF, these arms were discontinued and the trial proceeded using interferon alfa‐2b in place of GM‐CSF."
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was computer‐generated, using randomised blocks of random block sizes. However, 1 arm of the study was discontinued.
Allocation concealment (selection bias)	Unclear risk	Quote (page 591): "The sequence was provided to the investigators in sealed envelopes." It was unclear if they were sequentially numbered and opaque.
Blinding of participants and personnel (performance bias) All outcomes	High risk	The study was single‐blind in that the participants, but not the study investigators, were blinded to an individuals' treatment assignment. Comment: Personnel were not blinded.
Blinding of outcome assessment (detection bias) All outcomes	High risk	The investigators (outcome assessors) were not blinded.
Incomplete outcome data (attrition bias) All outcomes	High risk	233 participants were randomised: 201 were available for analysis. 23 were not included in the analysis because the trial arm was discontinued because of safety concerns. 9 were randomised but did not receive treatment. Comment: This was judged as high risk of bias as data were not presented for the group that received GM‐CSF, because of discontinuation of this arm.
Selective reporting (reporting bias)	Low risk	The only specified outcome was reported.