J Korean Soc Emerg Med.  2013 Dec;24(6):694-701.

The Correlation between Infection Probability Score and Procalcitonin in Emergency Department Patients

Affiliations
  • 1Department of Emergency Medicine, Kangbuk Samsung Hospital, Sungkyunkwan University School of Medicine, Seoul, Korea. jiung.na@samsung.com

Abstract

PURPOSE
Procalcitonin is a well-established biochemical marker for bacterial infection. We conducted this study to analyze the correlation between procalcitonin and Infection Probability Score (IPS), a recently introduced scoring system to predict bacterial infection in intensive care unit patients. The cutoff value of IPS corresponding to procalcitonin cutoff values was determined for procalcitonin-guided antibiotic therapy in emergency department (ED) patients.
METHODS
A retrospective observation study was conducted on adult ED patients who simultaneously underwent an IPS-required blood test and procalcitonin treatment from January 1, 2012 to June 30, 2012. Based on their diagnosis at discharge, patients were grouped into a lower respiratory infection group or an "other" diagnosis group. The correlation between IPS and procalcitonin was analyzed by correlation and linear regression analysis. The IPS value corresponded to 0.25 ng/mL procalcitonin (in the lower respiratory infection group) and 0.5 ng/mL (in the other diagnosis group) as inferred by ROC curve analysis. A total of 722 cases (lower respiratory infection group: 258, other diagnosis group: 464) were included in the final analysis.
RESULTS
In correlation analysis, the IPS showed a significant correlation with procalcitonin level in both groups (r=0.26, p<0.01, r=0.25, p<0.01, respectively). In ROC curve analysis, IPS 14 could predict procalcitonin> or =0.25 microg/L in the lower respiratory infection group (area under curve: 0.783 [95% CI, 0.724-0.841], sensitivity: 77.8%, specificity: 72.3%). Also, IPS 14 could predict procalcitonin> or =0.5 microg/L in the other diagnosis group (area under curve: 0.764 [95% CI, 0.717-0.810], sensitivity: 70.1%, specificity: 74.2%).
CONCLUSION
The IPS had a significant correlation with procalcitonin level and IPS> or =14 corresponded to the procalcitonin cut-off value to predict bacterial infection in ED patients. Thus, IPS> or =14 may be used to predict bacterial infection and can guide early anti-microbial therapy in ED patients when procalcitonin is not readily available.

Keyword

Bacterial infections; Procalcitonin; Predictive Value of Tests

MeSH Terms

Adult
Bacterial Infections
Biomarkers
Diagnosis
Emergencies*
Hematologic Tests
Humans
Intensive Care Units
Linear Models
Predictive Value of Tests
Retrospective Studies
ROC Curve
Sensitivity and Specificity
Full Text Links
  • JKSEM
Actions
Cited
CITED
export Copy
Close
Share
  • Twitter
  • Facebook
Similar articles
Copyright © 2024 by Korean Association of Medical Journal Editors. All rights reserved.     E-mail: koreamed@kamje.or.kr