Cerebrospinal fluid cytochemical parameters as predictors of FilmArray Meningitis/Encephalitis panel positivity: a diagnostic stewardship approach in immunocompetent and immunocompromised patients

Abstract

Infectious meningoencephalitis requires prompt diagnosis to reduce mortality and neurological sequelae. The FilmArray Meningitis/Encephalitis (FAME) panel has improved pathogen detection, but its indiscriminate use raises cost-effectiveness concerns. Because FAME is itself an imperfect reference standard, this study was framed as a diagnostic stewardship evaluation of CSF cytochemical parameters as predictors of FAME positivity, rather than as an assessment of diagnostic accuracy for confirmed central nervous system infection. We aimed to evaluate cytochemical thresholds that could guide test indication in immunocompetent and immunocompromised patients. A retrospective diagnostic accuracy study was conducted in adults with suspected acute meningoencephalitis who underwent FAME testing at two tertiary hospitals in Medellín, Colombia (2019–2023). CSF cytochemical parameters were evaluated as the index test against FAME as the reference standard; the target condition was therefore FAME positivity, not confirmed central nervous system infection. Of 533 records reviewed, 478 patients were included (median age 51 years; 55% male; 353 immunocompetent, 125 immunosuppressed). Overall FAME positivity was 14.8%. FAME-positive patients had higher median CSF leukocytes (189 vs. 3 cells/mm³) and protein (167 vs. 47 mg/dL), and lower glucose (31 vs. 62 mg/dL). CSF protein showed the greatest discriminative capacity (AUC 0.793, 95% CI 0.71–0.85), opening pressure the lowest (AUC 0.666, 95% CI 0.55–0.73). Overall, 287 tests (60%) were performed on samples without pleocytosis (< 10 leukocytes/mm³), with 5.5% positivity; adding protein < 50 mg/dL identified 192 samples (40.1%) with only 2.6% positivity. Clinically relevant results occurred in 0.5% of these (narrow definition: management change) and 1.0% (broad definition: any confirmed true positive). Negative predictive value of the combined threshold was 97.4% (95% CI 94.0–98.9%) overall and 96.0% (95% CI 86.5–98.9%) in immunosuppressed patients. CSF cytochemical parameters were associated with FAME positivity; protein concentration had the greatest discriminative capacity. Restricting testing to samples with ≥ 10 leukocytes/mm³ or ≥ 50 mg/dL protein would have increased diagnostic yield by 55.4% and reduced low-yield testing by 40.1%. This threshold might also apply to immunosuppressed patients, historically excluded from such rules, although one clinically significant case would have been missed in this subgroup under the broader definition. These findings are hypothesis-generating and require prospective validation.