Biomarkers improve the ability of clinical scores to predict intensive care unit admission in patients with community-acquired pneumonia
Abstract number: O129
Albrich W.C., Schuetz P., Batschwaroff M., Christ-Crain M., Zimmerli W., Bucher H.C., Müller B.
Objectives: A clinical prediction tool (SMART-COP), based on eight clinical, laboratory and radiologic parameters, compared favourably with the PSI and CURB65 scores to predict intensive respiratory and vasopressor support in patients with community-acquired pneumonia (CAP) (Charles et al. Clin Inf Dis 2008;47:37584). We compared the diagnostic accuracy to predict Intensive Care Unit (ICU) requirements in CAP between clinical scores and biomarkers.
Methods: In patients with CAP enrolled in the Swiss multicentre Pro-HOSP study, receiver operating characteristic curves (ROC) to predict ICU admission were compared between the clinical scores SMART-COP, PSI and CURB65 and the biomarkers Pro-Adrenomedullin (Pro-ADM) and Endothelin-1 precursor peptides (Pro-ET1).
Results: In preliminary analysis, a SMART-COP score of 3 was present in 30 of 83 (36.2%) patients requiring ICU admission compared with only 64 of 842 (7.6%) without ICU admission (p 0.0001). The ROC values of ICU admission were highest for SMART-COP score (0.75, 95% CI: 0.690.81), Pro-ET1 (0.73, 95% CI: 0.680.79) and Pro-ADM (0.72, 95% CI: 0.660.77), and lower for PSI score (0.68, 95% CI: 0.630.73) and CURB65 score (0.64, 95% CI: 0.580.70). Combining Pro-ET1 and the SMART-COP score (AUC=0.80, 95% CI: 0.750.85) in a combined logistic model significantly improved the diagnostic accuracy for ICU admission over each individual predictor (p < 0.01, for each comparison). To predict the outcome of death (n T50) and ICU admission or both (n T118), the combination of Pro-ADM and the SMART-COP score was significantly better (AUC=0.80; p < 0.05 for each comparison) than any predictor individually (AUC: Pro-ADM 0.72, SMART-COP 0.73, PSI 0.73, CURB65 0.68, Pro-ET1 0.73).
Conclusion: As biomarkers Pro-ADM and Pro-ET1 are prognostic markers for a severe course of CAP, combining biomarkers and clinical severity assessment tools might become useful measures for more appropriately triaging CAP patients in an era of limited healthcare resources. This hypothesis has to be tested in intervention studies.
|Session name:||19th European Congress of Clinical Microbiology and Infectious Diseases|
|Location:||Helsinki, Finland, 16 - 19 May 2009|
|Back to top|