RESUMO
BACKGROUND: Cancer patients are highly prone to develop bacterial bloodstream infections (BSI) and are also at risk of neutropenia. Knowledge of the prevalence of these infections and whether neutropenia is associated with a change in mortality is important to more effective management and reducing mortality and morbidity. OBJECTIVES: Estimate the prevalence of bacterial BSI among oncology inpatients and assess the associations of 30-day mortality with Gram stain results and neutropenia. DESIGN: Retrospective cross-sectional SETTING: University hospital in Saudi Arabia. PATIENTS AND METHODS: We retrieved records of oncology inpatients at King Khalid University Hospital, excluding patients without malignancy and with non-bacterial BSI. The number of records included in the analysis was reduced based on a sample size calculation and systematic random sampling used to select patients to include in the study. MAIN OUTCOME MEASURES: Prevalence of bacterial BSI and association between neutropenia and 30-day mortality. SAMPLE SIZE: 423. RESULTS: The prevalence of bacterial bloodstream infections was 18.9% (n=80). Gram-negative bacteria were more prevalent (n=48, 60.0%) than gram-positive bacteria, with the most common being Escherichia coli (n=20, 25.0%). The 23 patients (28.8%) who died included 16 (69.6%) with gram-negative infections and 7 (30.4%) with gram-positive infections. There was no statistically significant association of bacterial BSI-related 30-day mortality with Gram stain (P=.32). Of 18 patients (22.5%) who were neutropenic, only one (5.6%) died. Sixty-two (77.5%) patients were non-neutropenic, of whom 22 (35.50%) died. We found a statistically significant association between the presence of neutropenia and bacterial BSI-related 30-day mortality (P=.016), with mortality being lower among neutropenic patients. CONCLUSIONS: Gram-negative bacteria are more prevalent in bacterial BSI than gram-positive bacteria. No statistically significant association of Gram stain result with mortality was found. However, the 30-day mortality rate was lower among neutropenic patients than among non-neutropenic patients. We recommend further investigation with a larger sample size in multiple regions to further unravel the association of neutropenia with bacterial bloodstream infection-related 30-day mortality. LIMITATIONS: Lack of regional data and sample size. CONFLICT OF INTEREST: None.
