Outcome of inappropriate initial antimicrobial treatment in patients with methicillin-resistant Staphylococcus aureus bacteraemia.

To determine the impact of delayed treatment of methicillin-resistant Staphylococcus aureus bacteraemia (MRSA bacteraemia) on mortality. We used two different and complementary approaches: a retrospective cohort study and a matched case-control study. Of the total of all patients with S. aureus bacteraemia in a tertiary hospital over a 4 year period, 127 patients with MRSA bacteraemia were enrolled in the cohort study. The cases for the case-control study were defined as patients who received appropriate empirical treatment for MRSA bacteraemia; the controls, who were patients who received inappropriate empirical treatment for MRSA bacteraemia, were selected according to the matching variables of age, sex, severity of underlying illness, classification of main underlying disease and prior hospital stay. On the 14-point matching scale that was used to select the best controls, the average score (+/-S.D.) of the 30 controls was 11.2 (+/-2.0). In the cohort study, the difference in S. aureus bacteraemia-related mortality between appropriate (30%, 9/30) and inappropriate (39%, 38/97) empirical treatment was not significant (P=0.36). In addition, multivariate analysis did not indicate that inappropriate empirical treatment was associated with mortality when independent predictors for mortality were considered (adjusted OR 1.1, 95% CI 0.4-3.1). In the case-control study, S. aureus bacteraemia-related mortality in case patients was 30% (9/30) and in control patients 33% (10/30) (P>0.99, McNemar's test). In four of the nine discordant pairs, the case patients died whereas the control patients lived. Our data suggest that an initial delay of 2 days in the use of appropriate antibiotics, especially of vancomycin or other glycopeptides, before the preliminary microbiological report may not adversely affect the outcome in patients with MRSA bacteraemia.