BACKGROUND Most surgical site infections after shoulder surgery are caused by Cutibacterium acnes. Topically applied benzoyl peroxide (BPO) has for years been used to decrease the skin load of C acnes in treatment of acne vulgaris. The purpose of this study was to examine this effect on bacterial colonization in patients subjected to elective shoulder surgery at different stages of the procedure. METHODS A total of 100 patients scheduled for primary elective open shoulder surgery were randomized to prepare either with BPO or according to local guidelines-with soap (control group). Four skin swabs were taken in a standardized manner at different times, before and after surgical skin preparation, 1 in dermis, and finally after the skin was sutured. Before skin incision, 5 punch biopsies (3 mm in diameter and maximum 4 mm deep) were retrieved spaced 2 cm apart in the planned skin incision. On culturing, quantification of C acnes was made by serial dilutions. RESULTS Men had a 5-fold higher amount of C acnes on untreated skin. Treatment with BPO considerably lowered this count (P = .0001) both before and after skin disinfection compared to the control group. This positive effect of BPO persisted until skin closure, the point at which some recolonization of C acnes had occurred, but to a higher degree in the control group (P = .040). CONCLUSIONS Preoperative BPO treatment of the shoulder may be an effective method to decrease bacterial skin load of C acnes from skin incision until wound closure.