Implementation of an Algorithmic Approach to Molecular Infectious Disease Laboratory Testing on Cerebrospinal Fluid Samples

Abstract Rapid turnaround time for infectious disease laboratory testing is critical for the encephalitic/meningitic patient. Clinicians and emergency medicine providers often will bundle cerebrospinal fluid (CSF) studies (cell counts, glucose, protein) with molecular infectious disease laboratory (MIDL) assays, up front, to expedite clinical decision making about antimicrobial therapy and hospital admission. Many patients, however, ultimately do not have CSF pleocytosis and, predictably, have negative MIDL results. We quantified patients who had MIDL testing ordered, without concurrent CSF pleocytosis, to justify a mandated algorithm for CSF MIDL testing. Four months of MIDL CSF results were reviewed (viral monoplex PCRs); 274 patients were identified. Forty-seven percent of patients (128/274) were neonates (2 months old), who would have been expected to generate a CSF pleocytosis, 55% (81/146) had one or more MIDL studies ordered without CSF pleocytosis; 3.4% of these patients (5/146) did not have CSF cell counts ordered. Among all patients, 7.7% (21/274) had positive MIDL results (1 HSV1; 2 HSV2; 18 enterovirus). Seventeen patients had CSF pleocytosis; the 4 patients without CSF pleocytosis were neonates (3) or had no cell counts ordered (1). A ratio of 700 RBC to 1 WBC was used to approximate the expected number of CSF WBC in bloody taps (10%; 28/274). We show that over half of patients >2 months old had MIDL studies ordered without CSF pleocytosis; all patients with positive MIDL results had CSF pleocytosis or were neonates or had no cell counts ordered. An algorithm will be implemented in which all neonates with concern for CNS infection will have CSF MIDL tests ordered; among nonneonates, only those with documented CSF pleocytosis may have MIDL assays ordered (with rare exceptions; eg, leukopenia).