Background: Median sternotomy is the most frequently used incision for cardiac procedures but carries a substantial risk for deep sternal wound infections and/or sternal dehiscence. In contrast to previous studies that examined risk factors for sternal infections this study evaluates factors that lead to poor outcome after surgical revision of the non healing sternum. Methods: Between 1985 and 1999, 193 adults (mean age 64 ± 9 years, m/f = 3/1) necessitated sternal revisions (incidence 1.93%). Pre-, intra- and post-operative risk factors were evaluated for their influence on the outcome after sternal revision. Results: 65 of the 193 patients had a complicated course: ten (5.2%) died due to sepsis/multi organ failure (n = 6) or cardiac causes (n = 4). 32 patients (16.6%) needed several revisions, 17 (9%) were discharged with sternal instability, 5 (3%) with chronic fistula and one with persistent osteomyelitis. Univariate and multivariate analysis identified cardiopulmonary resuscitation (odds ratio (OR) = 11.188, p = 0.010), corticoid treatment (OR = 7.043, p = 0.0055), diabetes (OR = 4.130, p = 0.0128), smoking history (OR = 2.996, p = 0.0041), renal insufficiency (hazard ratio (HR) = 1.884), old age (OR = 1.108, p = 0.0266), high body mass (HR = 1.06), ECC time (p = 0.023), cross clamp time (p = 0.028), systemic hypothermia (p = 0.016), non-use of IMA (p = 0.042) or prolonged ventilation as risk factors for mortality or poor outcome. No correlation between sternal closure technique, mediastinal irrigation or antibiotic therapy and outcome after mediastinal revision could be found. Conclusions: To avoid disappointing results after sternal revision one should aim to preoperatively identify high-risk patients and aggressively address risk factors. This rather than modifications of the surgical and medical approach might improve the outcome of patients with mediastinal complications.