Procalcitonin is a Poor Predictor of Non-Infectious Fever in the Neurocritical Care Unit.

Abstract

Fever is a common occurrence in the Neurocritical Care Unit (NCCU). It is reported that up to 50% of these fevers are associated with a non-infectious source. As this is a diagnosis of exclusion, a complete fever evaluation must be done to rule out infection. Procalcitonin (PCT) has been identified as a possible biomarker to distinguish infectious from non-infectious etiologies of fever. We hypothesized that PCT could be used as a predictor of infectious fever in febrile patients with intracranial hemorrhage admitted to the NCCU. A prospective observational cohort of patients admitted to a 12-bed NCCU in a tertiary-care university hospital from January 1, 2014, to October 1, 2014, was studied. Patients with intracranial hemorrhage (aneurismal subarachnoid hemorrhage, traumatic brain injury, intracerebral hemorrhage, or non-traumatic subdural hemorrhage) and fever defined as ≥101.4°F were included. All patients had a urinalysis, chest X-ray, two sets of blood cultures, and PCT as part of their fever evaluation. Patients also had urine, sputum, CSF cultures, and Clostridium difficile toxin PCR as clinically indicated. Patients with incomplete fever evaluations were excluded. Seventy-three patients met inclusion criteria: 36 had infections identified and 37 did not. Type of intracranial hemorrhage was similar between groups. For those with identified infection, median PCT was 0.15ng/mL (IQR 0.06-0.5ng/mL). For those without identified infection, median PCT was 0.09ng/mL (IQR 0.05-0.45ng/mL), p=0.30. Analyzing subgroups of intracranial hemorrhage patients revealed no group with a significant difference in PCT values. Patients with identified infection did have higher white blood cell counts (median 14.1×109/L (11.6-17.4×109/L) compared to those without identified infection 12×109/L (9.9-14.1×109/L), p=0.02. Among patients with intracranial hemorrhage, PCT did not differentiate infectious fever from non-infectious fever.