Predictors of early mortality in very elderly patients with bacteremia: a prospective multicenter cohort.

OBJECTIVES: The proportion of very elderly people in the population is increasing, and infectious diseases in this patient group may present with specific characteristics. The objective of this study was to investigate the outcome predictors of bacteremia among the very elderly. METHODS: This was a multicenter prospective cohort study of bloodstream infections (BSI) in patients ≥ 80 years old in 15 hospitals in Spain. The outcome variables were 14-day and 30-day mortality. Multivariate analysis was performed. RESULTS: One hundred and twenty episodes were included. Mortality was 22% (n = 26) on day 14 and 28% (n = 34) on day 30. In the univariate analysis, the variables associated with mortality were neutropenia, recent surgery, Pitt score ≥ 2, intensive care unit (ICU) admission, severe sepsis or shock, and abdominal, unknown, and respiratory tract sources. In the multivariate analysis, variables associated with mortality on day 14 were high-risk source (abdominal, unknown, and respiratory tract sources; odds ratio (OR) 7.9, 95% confidence interval (CI) 1.8-33.9), Pitt score ≥ 2 (OR 5.6, 95% CI 1.3-23.3), inadequate empirical treatment (OR 11.24, 95% CI 1.6-80.2), and severe sepsis or shock at presentation (OR 5.3, 95% CI 1.4-20.7); the interaction between empiric treatment and high-risk source was significant. On day 30, mortality was independently related to a high-risk source (OR 2.92, 95% CI 1.1-7.5) and presentation with severe sepsis or shock (OR 3.81, 95% CI 1.2-12.4). CONCLUSIONS: Presentation with severe sepsis or shock and a high-risk source of BSI were independent predictors of 14-day and 30-day mortality. Inadequate empirical treatment was also a predictor of early mortality in patients with a high-risk source.

Reappraisal of the outcome of healthcare-associated and community-acquired bacteramia: a prospective cohort study.

BACKGROUND: Healthcare-associated (HCA) bloodstream infections (BSI) have been associated with worse outcomes, in terms of higher frequencies of antibiotic-resistant microorganisms and inappropriate therapy than strict community-acquired (CA) BSI. Recent changes in the epidemiology of community (CO)-BSI and treatment protocols may have modified this association. The objective of this study was to analyse the etiology, therapy and outcomes for CA and HCA BSI in our area. METHODS: A prospective multicentre cohort including all CO-BSI episodes in adult patients was performed over a 3-month period in 2006-2007. Outcome variables were mortality and inappropriate empirical therapy. Adjusted analyses were performed by logistic regression. RESULTS: 341 episodes of CO-BSI were included in the study. Acquisition was HCA in 56% (192 episodes) of them. Inappropriate empirical therapy was administered in 16.7% (57 episodes). All-cause mortality was 16.4% (56 patients) at day 14 and 20% (71 patients) at day 30. After controlling for age, Charlson index, source, etiology, presentation with severe sepsis or shock and inappropriate empirical treatment, acquisition type was not associated with an increase in 14-day or 30-day mortality. Only an stratified analysis of 14th-day mortality for Gram negatives BSI showed a statically significant difference (7% in CA vs 17% in HCA, p = 0,05). Factors independently related to inadequate empirical treatment in the community were: catheter source, cancer, and previous antimicrobial use; no association with HCA acquisition was found. CONCLUSION: HCA acquisition in our cohort was not a predictor for either inappropriate empirical treatment or increased mortality. These results might reflect recent changes in therapeutic protocols and epidemiological changes in community pathogens. Further studies should focus on recognising CA BSI due to resistant organisms facilitating an early and adequate treatment in patients with CA resistant BSI.