MRSA: predictor of outcome in critically ill patients

Abstract

Methicillin resistant Staphylococcus aureus (MRSA) infection in critically ill patients is associated with considerable morbidity, mortality and cost.1 2 In 1994, we studied 29 MRSA positive patients admitted to the intensive care unit (ICU). Although the size of population studied was too small to be conclusive, there was a suggestion that MRSA did not have a significant impact on overall mortality (unpublished observation), however, there was an increase in average length of stay. In this retrospective study, microbiology records identified MRSA infected/colonised patients admitted to ICU over 36 months (1996–1998). Patient demographics, site of isolation and hospital mortality were recorded. Isolation of MRSA from blood, sterile sites or endotracheal secretions was considered to be a surrogate marker of serious infection. In those patients with MRSA isolated in multiple sites, the most likely site for serious infection was selected. One hundred and sixty seven patients were found to be infected/colonized with MRSA and of these 59 patients had positive blood or sputum cultures. Hospital mortality of MRSA positive patients was 50.9% compared to 33.8% overall hospital mortality of all ICU patients. Standardized incidence ratio for death was 1.51 (95% confidence interval 1.20–1.86) for patients with MRSA. As expected the number of deaths rises with age (≤60 yr was 34.8%, 61–75 yr 52.6%, >75 yr 64.4%. Odds ratio 2.74 (95% confidence interval 1.09–6.93) for >75 compared with ≤60 yr old). Diagnosis of MRSA from blood and sputum (group A) had a mortality of 44.1%, compared to 51.3% in those with MRSA from screening sites (group B) (P=0.403). Whilst mortality increases with age for both groups, it rises faster in patients with MRSA in group B (Table 7). There was no difference in mortality between groups. The poor association between mortality and positive blood or sputum cultures corroborates our previous observation that detection of MRSA is more likely to be an indicator of poor prognosis rather than the cause of fatal infection.