RESUMO
BACKGROUND: Sepsis is a serious medical condition requiring timely administered, appropriate antibiotic therapy. Blood culture is regarded as the gold standard for aetiological diagnosis of sepsis, but it suffers from low sensitivity and long turnaround time. Thus, nucleic acid amplification tests (NAATs) have emerged to shorten the time to identification of causative microbes. The aim of the present study was to evaluate the clinical utility in everyday practice in the emergency department of two commercial NAATs in patients suspected with sepsis. METHODS: During a six-week period, blood samples were collected consecutively from all adult patients admitted to the general emergency department for suspicion of a community-onset sepsis and treated with intravenous antibiotics. Along with conventional blood cultures, multiplex PCR (Magicplex™) was performed on whole blood specimens whereas portions from blood culture bottles were used for analysis by microarray-based assay (Prove-it™). The aetiological significance of identified organisms was determined by two infectious disease physicians based on clinical presentation and expected pathogenicity. RESULTS: Among 382 episodes of suspected sepsis, clinically relevant microbes were detected by blood culture in 42 episodes (11%), by multiplex PCR in 37 episodes (9.7%), and by microarray in 32 episodes (8.4%). Although moderate agreement with blood culture (kappa 0.50), the multiplex PCR added diagnostic value by timely detection of 15 clinically relevant findings in blood culture-negative specimens. Results of the microarray corresponded very well to those of blood culture (kappa 0.90), but were available just marginally prior to blood culture results. CONCLUSIONS: The use of NAATs on whole blood specimens in adjunct to current culture-based methods provides a clinical add-on value by allowing for detection of organisms missed by blood culture. However, the aetiological significance of findings detected by NAATs should be interpreted with caution as the high analytical sensitivity may add findings that do not necessarily corroborate with the clinical diagnosis.
