Procedure- and Hospital-Level Variation of Deep Sternal Wound Infection From All-Japan Registry.

BACKGROUND: The outcome of cardiovascular surgery has been improving over time, but the treatment of postoperative complications such as deep sternal wound infection (DSWI) still needs critical attention. A nationwide surgical registry was analyzed for procedural details and hospital factors related to DSWI. METHODS: The study used the Japan Adult Cardiovascular Surgery Database, which captured data from 82% of all the hospitals performing cardiac surgery in Japan. A total of 109,717 surgical cases (34,980 coronary artery bypass grafting, 43,602 valve operations, 31,135 thoracic aortic operations) were included in the study. RESULTS: The overall incidence of DSWI was 1738 (1.6%). The 30-day mortality and operative mortality were 3311 (3.0%) and 5155 (4.7%), respectively. Across the 3 procedures, thoracic aortic operation showed the highest odds ratio (2.61; 95% confidence interval [CI], 2.32 to 2.94) for operative mortality but the lowest (0.91; 95% CI:,0.73 to 1.13) for DSWI incidence. Conversely, coronary artery bypass grafting showed the lowest odds ratio (1.36; 95% CI, 1.24 to 1.49) for operative mortality but the highest (1.52; 95% CI, 1.32 to 1.76) for DSWI. There was also hospital-level variation: Correlation was statistically significant between the observed-to-expected ratio of DSWI incidence and the observed-to-expected mortality ratio of cardiovascular procedures across the hospitals, but the coefficient was small (r = .24, P < .001). CONCLUSIONS: Hospitals that have a lower risk-adjusted mortality rate of cardiovascular procedures do not always have a lower risk-adjusted DSWI occurrence rate. In addition, the incidence of DSWI varies across hospitals. We need to consider DSWI independently of surgical mortality, whereas for treatment we should consider both the specific hospital environment and the multidisciplinary care.