Staphylococci in cardio-thoracic surgery : Epidemiological and clinical studies

Sammanfattning: Mediastinitis complicating cardio-thoracic surgery causes suffering to the patient and huge costs to the society. It is often caused by methicillin resistant S. epidermidis (MRSE) and S. aureus. Bacterial findings in wound-samples from patients reexplored due to sternal wound complication (SWC) in 1984-1995 (n = 65) and 1998-2000 (n = 43) were compared. The proportion of Coagulase Negative Staphylococci (CoNS) remained stable: 65% and 60%, with a decrease of methicillin resistant CoNS from 45% to 33%. The proportion of S. aureus increased from 12% to 30%. Transmission of staphylococci from the patients skin or from staff to the wound occurs during operation by contact or by the airborne route. Staff as a possible source of MRSE was investigated with respect to carriage and dispersal of MRSE. Carriage occurred in 42% and dispersal in 33%. Dispersal and the possibility to reduce total bacterial air-counts and air-counts of MRSE by wearing tightly woven scrub suits compared to conventional scrub suits were studied experimentally in a test chamber and during 65 elective cardio-thoracic operations. Total bacterial air-counts were reduced by 60% when wearing tightly woven scrub suits. Air-counts of MRSE were low and not reduced. The results from the test chamber and from the operations were consistent. Patients' skin, staff hands and air as sources for MRSE and S. aureus found in the surgical wound were investigated. Pulsed field gel electrophoresis (PFGE) was used for strain typing. The patients' skin was the most probable source for MRSE irrespective of air-counts. For S. aureus the airborne route of infection was of some importance but when total air-counts were lowered the patients' skin was the probable source. As infected cases can be clinically difficult to identify among patients with SWC we applied a microbiological definition of diagnosis. Standardized tissue sampling, optimal culturing methods and strain typing were used. 51 cases were followed prospectively. 30% of infections caused by CoNS were clinically misinterpreted as non-infected. All S. aureus-infections and 86% of CoNS- infections were identified from primary cultures on agar plates which made it possible to deliver a fast result. CoNS-infection was significantly correlated to sternal dehiscence.