Comparison of the accuracy of procalcitonin, neutrophil CD64, and C-reactive protein for the diagnosis and prognosis of septic patients after antibiotic therapy.

Background: The performance of the inflammatory biomarkers in the management of septic patients who received antimicrobial therapies is largely neglected. This study aimed to compare the accuracy of procalcitonin (PCT), neutrophil CD64 (CD64), and C-reactive protein (CRP) for the diagnosis and prognosis of septic patients after antimicrobial therapy. Methods: This study prospectively recruited consecutive patients without infection and those diagnosed with infection but had received initial antimicrobial therapies. Sepsis was diagnosed according to sepsis-3 criteria. Serum PCT, CD64 and CRP levels were measured upon entry to the ICU. We also collected each patient's baseline characteristics. The diagnostic and prognostic performance of these parameters was evaluated from the area under the receiver operator characteristic curve (AUC). Results: A total of 635 consecutive ICU patients were screened for eligible and 289 (45.5 %) patients were diagnosed with sepsis upon entry to the ICU. The area under the curve (AUC) for PCT, CD64 and CRP in the identification of sepsis is 0.726, 0.692 and 0.719, respectively. Neither PCT (p = 0.587) nor CD64 (p = 0.373) is superior to CRP in the diagnosis of septic patients who received antimicrobial therapies. The AUC for PCT, CD64 and CRP in the prediction of ICU mortality in these sepsis patients is 0.702, 0.637 and 0.593, respectively. The prognostic performance of PCT (p = 0.006) rather than CD64 (p = 0.509) is better than CRP. Conclusions: Both PCT and CD64 are not superior to CRP in the identification of septic patients who received antimicrobial therapies. However, PCT instead of CD64 has a better prognostic accuracy than CRP for the prediction of ICU mortality of these septic patients.