To determine the cause of a coagulase-negative staphylococcal outbreak and to identify risk factors for surgical-site infections among patients following Medtronic Freestyle bioprosthesis implantation. Retrospective case-control study. An 800-bed university referral center. The cohort of 64 patients undergoing Freestyle valve replacement from September 1998 to December 1998. Seven patients developed infection (10.9% vs 1.1% during the preceding 8 months), including two with mediastinitis and five with endocarditis. There were no statistically significant differences between cases and controls with respect to age, gender, weight, underlying illness, preoperative hospital stay, duration of surgery, time on bypass, central venous catheter duration, National Nosocomial Infection Surveillance risk index, New York Heart Association class, albumin, or antibiotic prophylaxis. However, only three cases were documented to have received vancomycin prophylaxis. Of all staff evaluated, only surgical resident A was significantly associated with infection (odds ratio, 7.68; 95% confidence interval, 1.3-44.1; P=.02) Pulsed-field gel electrophoresis patterns on Staphylococcus epidermidis isolates from four of the six cases were identical. These cases were performed on different days. Surgical resident Awas the only staff member present in the operating room for all cases caused by the epidemic strain. This S epidermidis strain, however, was not isolated from operating room staff. A surgical resident was significantly associated with infection. However, the cause of this outbreak was likely multifactorial. Changes occurring during the investigation included institution of vancomycin as routine prophylaxis and modification of surgical technique, which contributed to the resolution of the outbreak.