Between June 1978 and June 1989, superficial or deep mediastinitis (or both) developed in only five (0.16%) of 3118 consecutive patients. All patients studied underwent cardiac procedures through a median sternotomy and survived more than 7 postoperative days. The surgical team disciplined itself to divide presternal soft tissues with a scalpel and used electrocautery for pinpoint hemostasis only. This 0.16% infection rate was statistically significantly lower than those in 28 previously published studies (Pearson's chi 2 test, p less than 0.05). Twenty-four predisposing factors were evaluated by Fisher's exact test. Among these only an operating time longer than 3 hours is related to sternotomy infections (p = 0.0208), and this effect was not a strong one. Statistical evidence strongly suggests that discriminate use of electrocautery is a major reason for the lowest median sternotomy infection rate reported to date.