Mortality impact of positive blood cultures in patients with suspected community-acquired bacteraemia. A Danish population-based cohort study

Mette Søgaard, Mette Nørgaard, Henrik Toft Sørensen, Henrik Carl Schønheyder

Research output: Contribution to journalConference abstract in journalResearch


Objective: We examined the prognostic impact of positive blood cultures compared to negative cultures on mortality in patients, who had blood cultures obtained during the first 72 hours of admission to a medical ward.
Methods: We conducted this population-based cohort study of adults (n = 20,210) with a first-time registered blood culture during 1995 through 2006 in Northern Denmark. We obtained information on blood cultures, coexisting chronic diseases (for this study identified as the 19 chronic diseases included in the Charlson Comorbidity Index), laboratory findings, and immunosuppressive therapy from medical databases. Positive cultures were defined as those with growth of one or more pathogen given an aetiological role based on joint clinical and microbiological assessment.
Mortality within 180 days following the date of first blood culture was determined through the Danish Civil Registration System. We computed Kaplan-Meier curves and product limit estimates for the main study variables. Next, time-dependent Cox regression analyses was used to compare the risk of death in patients with positive blood cultures and patients with negative cultures at days 0-7, 8-30, and 31-180, controlling for age, gender, coexisting chronic diseases, marital status, use of immunosuppressives, and calendar period. Further, we conducted analyses restricted to patients a discharge diagnose of infectious diseases (ICD-10 codes A00-B99).
Results: In total, 1,665 (8.2%) patients had positive blood culture. Mortality among patients with positive cultures was higher than among patients with negative during the first 30 days of follow-up: 8.4% vs. 4.6% after 7 days, and 5.5% vs. 4.9% during days 8-30, corresponding to adjusted mortality rate ratios (MRRs) of 1.5 (95% CI: 1.2-1.8) and 0.9 (95% CI: 0.7-1.2), respectively. Beyond day 30, mortality was 8.2% among patients with negative culture and 10.4% among patients with positive culture (adjusted MRR 1.0, 95% CI: 0.9-1.2). 2,934 patients had a discharge diagnosis of infectious disease, of which the blood culture was positive in 646 (22.0%). Among these patients the prognostic impact of bacteraemia persisted throughout the follow-up period (0-7 days: MRR = 1.7 (95% CI 1.2-2.6); 8-30 days: MRR = 1.2 (95% CI 0.7-1.9); 31-180 days: MRR = 1.6 (95% CI 1.0-2.1)).
Conclusion: Positive blood culture is a predictor of mortality in patients with suspected bacteraemia.
Original languageEnglish
JournalClinical Microbiology and Infection
Pages (from-to)s39
Number of pages1
Publication statusPublished - 2009
Externally publishedYes
Event19th Congress of Clinical Microbiology and Infectious Diseases - Helsinki
Duration: 16 May 200919 May 2009


Conference19th Congress of Clinical Microbiology and Infectious Diseases

Cite this