Krejci‐Manwaring 2008.

Methods	Parallel‐group, randomised controlled study This study was conducted in a secondary care setting at a single centre in the USA
Participants	32 participants with moderate to severe hand dermatitis who did not use topical tacrolimus during the previous 28 days and did not use topical corticosteroids, non‐steroidal immunosuppressants, or light treatment during the last 7 days Dropouts: 13 (including 1 participant who dropped out before the intervention was started) Inclusion criteria of the trial Adults 18 years of age or older Hand eczema with a combined severity score of 5 to 16 Exclusion criteria of the trial Pregnancy Use of topical tacrolimus 28 days before the study Use of topical corticosteroids, immunosuppressants, or light treatments to the hand 1 week before the study Use of systemic corticosteroids Study population Gender: 24 female, 8 male Age: mean 46 years, range 20 to 70 years
Interventions	Intervention • Topical tacrolimus twice daily for 12 weeks in addition to a daily dose of prednisone during 3 weeks; 30 mg in week 1, 20 mg in week 2, 10 mg in week 3 in 14/21 participants Control intervention • Vehicle ointment applied twice daily for 12 weeks; in addition, a daily dose of prednisone during 3 weeks: 30 mg in week 1, 20 mg in week 2, 10 mg in week 3 in 6/11 participants Participants were followed up at 5‐week intervals until week 14 after initial treatment Duration 14 weeks (3 weeks active treatment, 11 weeks follow‐up)
Outcomes	Primary outcomes of the trial Not defined Other outcomes Observer‐rated reduction in severity based on symptom grading scale for erythema, scaling, induration, and fissuring (5‐point scale; 0 = none to 4 = marked/intense) at weeks 1, 4, 8, 12 (end of treatment), and 14 (end of study) Investigator's global assessment at weeks 1, 4, 8, 12, and 14 Participant‐rated visual analogue scale (VAS) of pruritus Participant‐rated improvement Adverse events
Notes	Pilot study on maintenance therapy. Exact numbers of results for main outcomes not given ‐ only whether there was a statistically significant difference between the 2 interventions The study did include the secondary outcomes reduction in severity, investigator and participant‐rated and time until relapse, but we were unable to reproduce these data Declarations of interest: study authors received research, speaking, and/or consulting support from various pharmaceutical companies Funding: the study was supported by a grant from Astellas Pharma Inc. Sample size rationale: not stated
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "the randomization code list, correlating the kit number with the content of each kit, was kept on file at Fuijsawa Healthcare Medical Information Department until the time of analysis" Comment: reference to a randomisation code list
Allocation concealment (selection bias)	Low risk	Quote: "the vehicle and tacrolimus ointments were packaged in identical containers labelled with the subject number, so neither the subject, coordinator, nor the investigator knew which treatment the patient received" Comment: randomisation was remote from the participant‐recruitment centre. Vehicle and tacrolimus were packaged in pre‐labelled identical containers corresponding to a participant number
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "the vehicle and tacrolimus ointments were packaged in identical containers labelled with the subject number, so neither the subject, coordinator, nor the investigator knew which treatment the patient received" Comment: double‐blinded; this is considered an adequate blinding method
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Quote: "the vehicle and tacrolimus ointments were packaged in identical containers labelled with the subject number, so neither the subject, coordinator, nor the investigator knew which treatment the patient received" Comment: by the use of pre‐labelled and identical containers, observers were blinded
Incomplete outcome data (attrition bias) All outcomes	High risk	No intention‐to‐treat analysis but per protocol (20 of 33 = less than 80%)
Selective reporting (reporting bias)	Low risk	No trial registration found. All relevant outcomes described in the Materials section are described in the Results section
Other bias	Unclear risk	Baseline comparisons: at baseline, significant difference between groups for demographic characteristics was given. All participants had a combined symptom severity score of 5 to 16 Diagnostic certainty: yes The study was completed