Design of randomized controlled trials in the treatment of leg ulcers: More answers with fewer patients

Objective: To illustrate the benefit of the factorial design in randomized controlled trials of leg ulcers. Design: A 2x2x2 factorial design. Setting: Hospital leg ulcer clinics in Edinburgh and Falkirk. Patients: Adults with at least one unhealed leg ulcer of determined origin, present for at least 2 months and greater than 1 cm in diameter. Interventions: Pentoxifylline (Trental) 400 mg, three times daily, versus p lacebo; single-layer versus multilayer bandage; a hydrocolloid versus knitt ed viscose dressing. Main outcome measure: Complete healing of all ulcers w ithin 24 weeks. Results: Of 525 patients screened, 200 pure venous ulcers were randomized ( 58.5% healed by 24 weeks), 45 complex venous ulcers were randomized (57.8% healed) and 41 arterial patients were randomized (excluding bandaging compa risons) (19.5% healed). There were no interactions between treatments. Conclusion: The factorial design was feasible to administer and allowed thr ee therapeutic questions to be investigated using the same resources as wou ld have been needed to answer a single question.