Ellis 1991

Methods	RCT, active, controlled, double blind Date of study: not stated Location: single‐centre (University of Michigan Medical Center, Ann Arbor)
Participants	Randomised: 85 participants (mean age 46 years (cyclosporine 3), 42 years (cyclosporine 5), 46 years (cyclosporine 7.5), 43 years (placebo), 66 male) Inclusion criteria Participants with moderate‐severe psoriasis (BSA ≥ 25) Non‐response to phototherapy Non‐response to conventional systemic treatment Failure to at least 1 line Exclusion criteria Pregnancy Dropouts and withdrawals Not stated
Interventions	Intervention A. Ciclosporin (Sandimmun) (n = 15), orally, 7.5 mg/kg, 8 weeks Control intervention B. Ciclosporin (Sandimmun) (n = 20), orally, 5 mg/kg, 8 weeks C. Ciclosporin (Sandimmun) (n = 25), orally, 3 mg/kg, 8 weeks D. Vehicle (Sandimmun oral olive oil) (n = 25), orally, 8 weeks
Outcomes	Assessment at 8 weeks Primary or secondary outcomes not stated Outcomes Target lesions PASI Urinary creatinine clearance Blood count Blood pressure
Notes	Funding (p 277): Sandoz research Institute, the Babcock Dermatologic Endowment (Ann Arbor) and a Clinical research centre grant (M01‐RR‐00042) from the National Institutes of Health Declarations of interest: not stated (p 277) "Drs Ellis and Voorhees are consultants to Sandoz Pharmaceuticals corporation (the manufacturer of cyclosporine).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote (p 278): "patients were assigned numbers in consecutive order; each number had been preassigned to one of four treatments groups by means of a computer generated random code in blocks 17" Comment: probably done
Allocation concealment (selection bias)	Unclear risk	Comment: no description of the method used to guarantee allocation concealment
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote (p 278): "The preparation of cyclosporine and vehicle were identical …patients were blinded to their treatment" Comment: probably done
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote (p 278): "Other physician who were blinded to group assignment and laboratory findings evaluated the patient" Comment: probably done
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Randomly assigned 85, analysed not stated Dropouts and withdrawals not stated Quote (p 279): "In the primary, intention‐to‐treat analysis" Management of missing data: no description of the method used to guarantee management of missing data
Selective reporting (reporting bias)	Unclear risk	Comment: no protocol was available. The pre‐specified outcomes mentioned in the methods section appeared to have been reported