Procalcitonin in early rule-in/rule-out of sepsis in SIRS patients admitted to a medical ward.

BACKGROUND: A relevant amount of patients with clinical suspect of sepsis is admitted and treated in medical wards (MW). These patients have a better prognosis but are older and with more comorbidities compared to those admitted to intensive care units (ICU). Procalcitonin (PCT) is extensively used in emergency departments for the diagnosis of sepsis, but its accuracy in the setting of a MW has not been thoroughly investigated. Predicted low PCT levels also call for the comparison of immunomagnetic-chemiluminescent (L-PCT) and time-resolved amplified cryptate emission (TRACE, K-PCT) technologies, in PCT determination. METHODS: In 80 patients with systemic inflammatory response syndrome (SIRS) diagnostic criteria and suspect of sepsis newly admitted to a MW, PCT was determined with L- and K-PCT method. RESULTS: Sixty patients were diagnosed as sepsis (20 microbiologically and 40 clinically proven) and 20 with non-infective SIRS. The sepsis group had significantly higher levels of both PCTs, with no differences between the clinically and microbiologically proven subgroups. The areas under ROC curves for L- and K-PCT were 0.72 and 0.78 (p<0.001 for each), respectively. Based on MW customized cut-off values of 0.150 (L-PCT) and 0.143 ng/mL (K-PCT), overall accuracies were 66.8 (95% CI 58.7-78.9) and 78.2% (69.8-87.2), respectively, compared to the 55% (44.2-66) of 0.5 ng/mL canonical cut-off. Neither PCT-L nor -K held prognostic value on survival. CONCLUSIONS: In MW patients, customized PCT cut-off levels provide better accuracy than customary levels adopted from ICU, and TRACE technology seems to offer a wider analysis range.