Ability to predict the development of surgical site infection in cardiac surgery using the Australian Clinical Risk Index versus the National Nosocomial Infections Surveillance-derived Risk Index

Abstract

Surgical site infection (SSI) is a major infectious complication that increases mortality, morbidity, and healthcare costs. There are scores attempting to classify patients for calculating SSI risk. Our objectives were to validate the Australian Clinical Risk Index (ACRI) in a European population after cardiac surgery, comparing it against the National Nosocomial Infections Surveillance-derived risk index (NNIS) and analyzing the predictive power of ACRI for SSI in valvular patients. All the patients that who underwent cardiac surgery in a tertiary university hospital between 2011 and 2015 were analyzed. The patients were divided into valvular and coronary groups, excluding mixed patients. The ACRI score was validated in both groups and its ability to predict SSI was compared to the NNIS risk index. We analyzed 1,657 procedures. In the valvular patient group (n: 1119), a correlation between the ACRI score and SSI development (p < 0.05) was found; there was no such correlation with the NNIS index. The area under the receiver-operating characteristic curve (AUC) was 0.64 (confidence interval [CI] 95%, 0.5-0.7) for ACRI and 0.62 (95% CI, 0.5-0.7) for NNIS. In the coronary group (n: 281), there was a correlation between ACRI and SSI but no between NNIS and SSI. The ACRI AUC was 0.70 (95% CI, 0.5-0.8) and the NNIS AUC was 0.60 (95% CI, 0.4-0.7). The ACRI score has insufficient predictive power, although it predicts SSI development better than the NNIS index, fundamentally in coronary artery bypass grafting (CABG). Further studies analyzing determining factors are needed