Staphylococcus aureus is a common cause of bloodstream infection in neutropenic patients. We analyzed trends in incidence, epidemiology, predisposing factors and outcome of S. aureus bacteremia (SAB) in febrile neutropenic patients.

We reviewed a database of febrile neutropenia at a single institution in Brazil in a 30-year period (January 1991 to December 2020). We divided in 3 periods: 1991-1994 (period 1), 2001-2010 (period 2) and 2011-2020 (period 3), and looked at incidence rates, risk factors and mortality. Risk factors and prognostic factors were evaluated by univariate (Fisher or Chi-square for categorical variables and Wilcoxon test for continuous variables) and multivariate (logistic regression) analysis.

We diagnosed 62 (3%) episodes of SAB in 2,243 episodes of febrile neutropenia, 23 (38%) of which were caused by methicillin-resistant S. aureus (MRSA). The incidence of SAB per 1,000 days of neutropenia was 1.57, 2.53 and 1.50 in periods 1, 2 and 3, respectively (p > 0.05), and of MRSA was 0.49, 0.84 and 1.07, respectively (p > 0.05). Most infections (70%) were diagnosed in the first day of fever and the clinical presentation were skin infection (19%), catheter-related infection (14%) and pneumonia (3%). Complications like endocarditis or osteomyelitis were not observed. The only factor associated with S. aureus bacteremia by multivariate analysis was cellulitis at site of catheter exit site or tunnel on day 1 of fever (odds ratio [OR] 4.34, 95% confidence interval [CI] 2.05 – 9.16), whereas quinolone prophylaxis was protective (OR 0.37, 95% CI 0.18 - 0.74). The crude mortality of SAB was 30.6%, compared 13.8% in patients without this infection (p < 0.001). However, early mortality (within 3 days of start of empiric antibiotic therapy) was not significantly different (4.8% vs. 1.9%, p = 0.08). The mortality with MRSA was higher than with methicillin-susceptible S. aureus (47.8% vs. 13.9%, p < 0.001), but no difference was found regarding early mortality (5.0 vs 1.7%, p = 0.27). Among various predictors of death during febrile neutropenia MRSA bacteremia (but not methicillin-sensitive) was associated with an increased risk of death by multivariate analysis. The median time from febrile neutropenia to death in MRSA episodes was 19 days.

The incidence of SAB over the period remained stable. Complication of SAB were not common. Although the crude mortality was higher with SAB (including MRSA), it was not associated with early deaths, thus not demanding anti-staphylococcal coverage in the first day of febrile neutropenia. MRSA bacteremia was associated with increased risk of death, with most deaths occurring after 2 weeks of febrile neutropenia.