CURB65 may predict 30-day mortality in patients with bacteraemia

Abstract number: 1732_264

Barlow G., Lillie P., Parsonage M., Adams K., Thaker H., Moss P., Meigh J., Meigh R., Mawer S., Wilson J., Dibb W., Baruah J.

Objective: Bacteraemia is an important cause of mortality in hospitalised patients. CURB65 has been validated as a predictor of mortality in community-acquired pneumonia (CAP) (Lim et al. Thorax 2003), but not in other causes of sepsis. The aim of this study was to assess the performance of CURB65 in patients with bacteraemia.

Methods: A retrospective cohort study was performed using data routinely collected in the delivery of the Hull Bacteraemia Service. As part of this service, patients with confirmed bacteraemia are seen at the bedside by an infectious diseases physician. A typed report, which includes physiological data, is then sent to the patient's physician. Patients were included if they had all CURB65 criteria recorded at the point of bacteraemia. 30-day mortality was established by hospital database and stratified by CURB65 score. A receiver operating curve (ROC) was produced and area under the curve and 95% confidence intervals (CI) calculated.

Results: Of 151 patient reports, 61 patients (62% male) had a full set of CURB65 criteria. 49% of patients were over 65 years old and 34% were being managed on a renal ward. The most common bacteria were: MSSA (41% of patients), MRSA (23%); various Gram-negatives (16%); thought to be significant coagulase-negative staphylococci (15%); Group A–G streptococci (10%); and enterococci (8%). The most common sources of bacteraemia were: central venous line (31% of patients); intravenous drug use (15%); urinary tract (13%); contamination (10%); and skin/soft tissue (8%). 9% of patients were receiving discordant therapy prior to review. Overall, 30-day mortality was 25%. Recrudescence within 90 days occurred in 5.5% of patients. The table shows 30-day mortality stratified by CURB65 score. The area under the ROC was 0.73 (95% CI: 0.6–0.86).

30-day mortality stratified by CURB65 score in patients with bacteraemia

CURB65 score30-day mortality

Conclusions: This is the first study to assess the performance of CURB65 in a non-respiratory infection. Although preliminary, CURB65 appeared to stratify 30-day mortality in patients with bacteraemia. The cut-off for severe illness may need to be lower, however, than in CAP. There is the potential, therefore, to identify a low-risk cohort of patients (i.e. CURB65 = 0) who may be appropriate for an early switch to oral therapy and discharge from hospital. Likewise, patients at high risk (i.e. CURB65 ≥ 2) are more likely to need aggressive therapy according to the principals of the surviving sepsis campaign. A large prospective study is warranted.

Session Details

Date: 31/03/2007
Time: 00:00-00:00
Session name: European Society of Clinical Microbiology and Infectious Diseases
Location: ICC, Munich, Germany
Presentation type:
Back to top