From January, 1979, to December, 1984, at the Cardiac Surgery Department of the University of Torino Medical School, major sternal wound infections developed in 48 (1.86%) of 2,579 consecutive patients. These patients underwent open-heart procedures through a midline sternotomy and survived long enough for infection to appear. Possible risk factors were evaluated by means of a multivariate analysis. For the group of patients, we considered age, sex, hospital environment (different locations of our surgical facilities over the years), interval between hospital admission and operation, antibiotic prophylaxis, type of surgical procedure, elective or emergency surgical procedure, reoperation, duration of surgical procedures, duration of cardiopulmonary bypass, amount of blood transfused, postoperative blood loss, chest reexploration, rewiring of a sterile sternal dehiscence, duration of mechanical ventilation, and days of treatment in the intensive care unit. Univariate analysis indicated that age, sex, type and mode of surgical procedure, antibiotic prophylaxis, and duration of mechanical ventilation were not significantly associated with wound infection. For all other predisposing factors, a p value of less than .05 was demonstrated. These variables were entered in a multiple stepwise logistic regression. Six emerged as significant: hospital environment (p = .0001), interval between admission and surgery (p = .041), reoperation (p less than .0001), blood transfusions (p = .031), early chest reexploration (p less than .0001), and sternal rewiring (p less than .0001). Contamination of patients may occur before, during, and after operation, and any kind of reintervention may predispose to wound infection.