Performance enhancement of procalcitonin by high-sensitivity C-reactive protein at the optimal cutoff in predicting bacteremia in emergency department adult patients.

Pathogenic bacteremia portends a high mortality risk in adult patients admitted to an Emergency Department (ED). This study aims to investigate the effect of adding high-sensitivity C-reactive protein (hs-CRP) to procalcitonin (PCT) and lactate in predicting bacteremia, Gram-negative (GNB) and Gram-positive bacteremia (GPB), using the optimal cutoff derived from the receiver operating characteristics analysis. We evaluated the diagnostic measures, including the positive-test likelihood (LR+), the negative-test likelihood (LR-), and the diagnostic odds ratio (DOR) using a single-center retrospective analysis design. This Standards for Reporting Diagnostic-compliant study comprised 886 consecutive adults who were admitted to the ED in 2010; to this cohort, a 22.2% prevalence of true bacteremia was subsequently confirmed. At the cutoff of 3.9 μg/L, PCT had a DOR of 5.3 (95% confidence interval [CI]: 3.76-7.61) and LR + of 2.8 (95% CI: 2.3-3.4) in predicting overall bacteremia. Elevated PCT and lactate (cutoff at 2 mmol/L), increased the DOR and LR + to 6.3 (95% CI: 4.27-9.29) and 4.0 (95% CI: 3.1-5.2). The DOR and LR + were further improved to 7.1 (95% CI: 4.2-11.95) and 5.6 (95% CI: 3.7-8.6), respectively, when hs-CRP at the cutoff of 1238 nmol/L was added to PCT plus lactate. High-sensitivity CRP at the cutoff of 1,255 nmol/L can enhance the discriminative power raising DOR and LR + values for GPB. The elevation of hs-CRP at the optimal cutoff might improve the diagnostic performance to predict unspecified bacteremia and GPB, but not GNB.