• 
• 

Three Part Question

[In febrile adult patients presenting as a medical emergency] what is [the sensitivity and specificity] of [CRP as a tool for diagnosing bacteraemia]?

Clinical Scenario

An adult patient presents to the emergency department unwell with a temperature but no specific signs of a focus of infection. You wonder how useful a measurement of C Reactive Protein (CRP) will be in making a diagnosis of severe bacterial sepsis.

Search Strategy

Medline 1946- February week 3 2012 using the OVID interface
[{exp sensitivity and specificity} OR {Predictive adj value$} OR {sensitivity.mp} OR {specificity.mp} OR {di.xs} OR {du.fs} OR {exp mass screening} OR {exp diagnostic errors}] AND {exp C Reactive Protein OR c reactive protetin.mp OR CRP.mp} AND {bacteremia.mp OR bacteraemia OR exp septicaemia}
LIMIT TO humans and English language and all adults (19 plus years)

Search Outcome

Altogether 79 papers were found and 3 of these were found to be relevant

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Adams 2005 Australia	1214 patients presenting to ED acutely unwell (Good gold standard – blood culture positive)	Retrospective Review	To assess CRP as a marker of bacteraemia in ED patients	Elevated CRP was 94% sensitive 18% specific for concurrent bacterial infection.Positive likelihood ratio for bacteraemia with an elevated CRP was 1.15, and the negative likelihood ratio was 0.33	Retrospective. Only included patients who had a CRP done in ED at presentation.
Wyllie et al 2005 UK	6234 patients admitted as an emergency to the acute medical or infectious diseases services	Cohort study	Bacteraemia risk associated with CRP concentrations	Likelihood ratio for bacteraemia with elevated CRP (> 285mg/l) 4.0 Area under ROC curve for CRP 0.72
Aalto et al 2004 Finland	92 patients with infection on admission to hospital	Clinical Study	Diagnosis of infection	Positive predictive value of CRP, measured on admission, was low	Small study size. Poorly defined groups in abstract

Comment(s)

Severe bacterial sepsis is a very important diagnosis for the Emergency Department clinician, as patients with this condition can deteriorate very quickly if not promptly treated with the appropriate antibiotics. Such patients often present with non-specific signs and symptoms, making diagnosis difficult. C Reactive Protein is an acute phase protein which is said to be raised in acute infection. The purpose of this BET is to assess its efficacy in diagnosing severe bacterial infection (bacteraemia). We have considered positive blood cultures, positive urine cultures or definite consolidation on a chest X-Ray as appropriate gold standards against which to define the sensitivity and specificity of CRP as a diagnostic tool. Clinical examination was not considered to be an appropriate gold standard. The relevant studies showed a wide range of values for sensitivity and specificity of CRP. Many of these studies did not have an adequate gold standard, and many of the patient groups were not specifically ED presentations. Although many of the studies are of poor quality, on the whole the area under the ROC for CRP is low indicating it is a poor diagnostic test. The best study found (Adams 2005) also suggests CRP is a poor diagnostic tool with a specificity for concurrent bacteremia of only 18%.

Clinical Bottom Line

The evidence is not of high quality, but suggests that CRP is not a useful tool in the initial diagnosis of severe bacterial infection.

References

1. Adams, N.G. Diagnostic use of C-reactive protein in bacteraemic emergency department patients. Emergency Medicine Australasia 2005; 17 371-375
2. Wyllie DH, Bowler IC, Peto TE. Bacteraemia prediction in emergency medical admissions: role of C reactive protein. Journal of Clinical Pathology 2005; 58 352-356
3. Aalto H, Takala A, Kautiainen H, et al. Laboratory markers of systemic inflammation as predictors of bloodstream infection in acutely ill patients admitted to hospital in medical emergency. European Journal of Clinical Microbiology & Infectious Diseases 2004; 23 699-704.