We have performed a prospective single-blinded randomised study to evaluate the role of antibiotic-impregnated cement in the prevention of deep infection at primary total knee arthroplasty (TKA) in patients with diabetes mellitus. We studied prospectively 78 arthroplasties performed for osteoarthritis in such patients. They were randomly separated into two groups. In group 1 (41 knees), cefuroxime-impregnated cement was used while in group 2 (37 knees) cefuroxime was not added to the cement. The preoperative, intraoperative and postoperative management was the same for both groups. The mean follow-up was 50 months (26 to 88). There were no cases of deep infection in group 1, but five (13.5%) occurred in group 2 (p = 0.021). We conclude that cefuroxime-impregnated cement is effective in the prevention of deep infection at primary TKA in patients with diabetes mellitus.