Methicillin resistant Staphylococcus aureus infections following cardiac surgery: incidence, impact and identifying adverse outcome traits☆

Abstract

To assess the incidence and impact of Methicillin-resistant Staphylococcus aureus (MRSA) infections on cardiac surgery outcomes and to identify adverse outcome traits. Retrospective analysis of prospectively collected data from cardiac surgical and microbiology databases between April 2000 and March 2005. The overall and yearly incidence of positive MRSA cultures was examined along with the distribution of clinical infections and the associated mortality. Pre-operative patient characteristics were analysed between non-survivors and survivors of MRSA infections. Multivariate logistic regression was used to assess the relationship between pre-operative patient characteristics and in-hospital mortality in patients with MRSA. A comparison of post-operative outcomes between non-survivors and survivors of MRSA infections was also carried out and included in the logistic regression analysis. There were 319 patients with positive MRSA cultures during the study period with an overall incidence of 3.9%. Yearly incidence ranged from 2.4% to 5.2%. There were 120 carriers with pre-operative positive cultures of which 25 developed clinical surgical infections leaving 224 patients as the study group. Overall mortality in patients with MRSA during the study period was 12.9%(41/319). Mortality in the study group was 17.8% (40/224). Mortality comparison between MRSA and non-MRSA mediastinitis was 26.7%(8/30) and 17.1%(13/76), respectively (p=0.26). Mortality between MRSA and non-MRSA septicaemia was 46.9% (15/32) and 52.9% (37/70) (p=0.57). Applying the logistic EuroSCORE to the MRSA patients revealed that non-survivors had a significantly higher pre-operative risk of 10.4% compared to survivors with a pre-operative risk of 6.2% (p=0.003). Renal dysfunction and poor ejection fraction were found to be pre-operative factors associated with mortality in MRSA patients following the multivariate logistic regression analysis. Non-survivors had longer stays on intensive care, longer ventilation times, and were more likely to require support with balloon pumps and haemofiltration. MRSA septicaemia and length of ventilation were significantly associated with mortality in MRSA patients ahead of pre-operative characteristics. The incidence of MRSA is low, but carries a high mortality. MRSA septicaemia and mediastinitis have the highest associated mortality; however, this is not significantly different from non-MRSA infections. Patients with MRSA who die have higher pre-operative risk and have a poorer post-operative course than survivors.