We performed a retrospective study of a large cohort of patients who had episodes of Staphylococcus aureus bacteremia (SAB) from January 1995 through February 1999 at 1 medical center to identify predictors of 30-day mortality in SAB. Among 293 patients with episodes of SAB, 68 died (23.2%) within 30 days of onset. There was no significant difference in 30-day mortality associated with treatment with vancomycin, a beta-lactam, or a miscellaneous group of antimicrobial agents (P=.180). By logistic regression, an acute physiology score (a component of the acute physiology and chronic health evaluation [APACHE III]) >60 at onset of SAB was the most important predictor of 30-day mortality (odds ratio [OR], 15.7). Other significant predictors were lung (OR, 5.8) or unknown (OR, 4.1) focus of SAB, age > or =65 years (OR, 2.0), and diabetes mellitus (OR, 2.4). Future investigators of SAB should take into consideration acute severity of illness at onset as well as other factors when evaluating or comparing outcomes.

A severity of NHAP model was derived from a large cohort of episodes in multiple facilities. The model had reasonable discriminatory power in the derivation cohort. The model may aid clinicians in making treatment decisions in the nursing home setting and in making hospitalization decisions. Although prepneumonia functional status provides a reasonable estimate of NHAP severity and prognosis, the severity of NHAP model permitted further refinement of these estimates. The severity of NHAP model requires validation before it can be recommended for general use.

During the past two decades, there have been important advances in blood culture methodology. These advances have resulted in earlier detection and identification of pathogens causing bloodstream infections. However, there are many facets of the blood culture as a diagnostic test that are not affected by new culture methods or systems that continue to cause problems with interpretation of results. The objective of this review is to focus on those factors influencing the results of blood cultures that have clinical relevance. Such factors include skin preparation, timing, procurement techniques, volume of blood obtained, number of cultures, anaerobic blood cultures, and contamination. In addition, bacteremia prediction models are discussed and suggestions are provided as to how these models could be of greater clinical use. Blood culture methods and systems are not discussed in this review.

This study sought to reevaluate the epidemiology of bloodstream infection in nursing home residents. The records of 166 nursing home residents admitted to an urban, public, university-affiliated hospital with 169 episodes of bloodstream infection between January 1997 and April 2000 were retrospectively reviewed. The most common organisms isolated were Escherichia coli (27% of isolates), Staphylococcus aureus (18%; 29% were methicillin-resistant strains), and Proteus mirabilis (13%). There was minimal resistance to quinolones and third-generation cephalosporins among aerobic gram-negative bacilli. The most common sources were the urinary tract (51% of episodes) and the lungs (11%); a source was not identified in 22% of episodes. Hospital mortality was 18%. Independent predictors of hospital mortality were a pulmonary source of infection, systolic blood pressure <90 mm Hg, and leukocytosis >20,000 cells/mm3. Compared with other studies published in the past 2 decades, mortality was lower. The most common resistant organism was methicillin-resistant S. aureus.