Objective: To assess clinical safety and accuracy of qualitative blood culture drawn through the hub for ruling out catheter-related infection (CRI).
Design and setting: Prospective observational study in a surgical intensive care unit.
Patients: All patients with sepsis of unknown origin and possibly due to a CRI.
Interventions: Blood culture drawn through a central venous catheter (CVC) just before the catheter was cultured.
Measurements and results: In 126 patients we investigated 135 cases of sepsis of unknown origin. Using a clinical and bacteriological approach as the reference, the performance of the CVC blood culture was evaluated by the calculation of sensitivity, specificity, and positive and negative predictive values. When CVC blood culture was positive, the time to positivity was considered. Using standard definitions, seven CRIs were diagnosed. CVC blood culture identified five CRIs including the three episodes of catheter-related bacteremia. The method missed two coagulase-negative staphylococcus CRIs without bacteremia. Thirteen false-positive results occurred, including seven bacteremias from a distant source. The CVC blood culture had a sensitivity of 71% (CI 30-95%), specificity of 90% (CI 83-94%), negative predictive value of 98% (CI 93-100%), and positive predictive value of 28% (CI 11-54%). In cases of catheter-related bacteremia the time to positivity of CVC blood culture was 24 h or less.
Conclusions: Negative CVC blood culture at 24 h seems useful for management of CVC in selected critically ill surgical patients developing a clinical sepsis. The subsequent risk of catheter-related bacteremia cannot be excluded advocating for an uninterrupted clinical and bacteriological survey.