Establishing a Nucleated Count Reflex Trigger for Performing Cell Differential on Cerebrospinal Fluid in Pediatric Patients

Abstract

Abstract Introduction Analysis of cerebrospinal fluid (CSF) assists in the diagnosis of several disease processes affecting the central nervous system (CNS). Nucleated cell (NC) count and cell differential are two important components of laboratory testing performed to analyze CSF samples. These tests can help diagnose conditions such as infection, inflammation, and malignancy involving the CNS; however, the cell differential, performed manually on a cytospin, is labor-intensive and may divert laboratory resources. Currently, the laboratory at our tertiary care pediatric hospital performs both NC count and cell differential on all specimens submitted for CSF cell count analysis. We hypothesized that when the NC count is low (<10/mcl), performing manual cell differential does not provide clinically meaningful information and leads to unnecessary testing, increasing the turnaround time and decreasing precision. In this study, we investigated the utility of performing a manual differential on CSF when the NC count is ≤10/mcl and the need to establish reflex criteria for performing a cell differential. Methods Data were extracted from our electronic medical record database. We searched for all CSF samples that were submitted to our pediatric hospital’s laboratory for cell count analysis from September 2019 to January 2020. NC count, red blood cell (RBC) count, and number of cells available for differential (€œcells diffed) on cytospin, to a maximum of 100 cells per specimen, were obtained for each sample. Results A total of 577 CSF samples from 332 patients were submitted and analyzed for cell counts, of which 471 (82%) had NC count ≤10/mcl. There was a significant moderate positive correlation between NC count and cells diffed (Pearson correlation coefficient, 0.45, P-value <0.001). Of these 471 samples, 24 (5%) showed evidence of peripheral blood contamination (>500/mcl RBCs in samples with NC = 0/mcl, >500 x (NC count)/mcl RBCs in samples with NC > 0/mcl). A total of 4 cases from two patients with NC ≤10/mcl (0.8%) showed blasts on manual differential. Both patients had B-lymphoblastic leukemia/lymphoma and were being evaluated for CNS involvement with concurrent cytology. One patient was diagnosed as negative for CNS involvement by cytology. The sample for manual differential was contaminated by peripheral blood, hence the presence of blasts did not represent CNS disease. The second patient had the remaining 3 positive samples and was diagnosed with CNS involvement by cytology on multiple follow-ups. Conclusion Performing CSF manual differential when NC < 10/mcl does not provide additional clinically actionable information and is associated with a lower cell yield on the cytospin. Establishing a NC count cutoff for performing manual differential will prevent unnecessary testing, decrease turnaround time, and increase precision. Exceptions may be necessary for oncology patients at high risk for CNS involvement.