Quantitative and molecular adjudication of BioFire FilmArray and QIAstat-Dx discordance in meningitis/encephalitis diagnostics

Résumé fourni par la source

ABSTRACT Syndromic multiplex panels enable rapid cerebrospinal fluid (CSF) diagnosis of meningitis and encephalitis, but clinically relevant inter-platform discordance persists, and its analytical basis remains incompletely understood. In this retrospective multicenter study, 182 residual CSF specimens from 182 patients at 4 Swiss tertiary-care centers and 10 external quality assessment samples were tested centrally by the QIAstat-Dx Meningitis/Encephalitis panel and compared with routine BioFire FilmArray results. Discordant results were adjudicated using pathogen-specific quantitative laboratory-developed tests (LDTs) calibrated to external standards, and enterovirus-positive specimens underwent molecular typing. Among 182 clinical CSF specimens, concordance was 87.9% across shared targets. Complete agreement was observed for herpes simplex virus type 2, Haemophilus influenzae , Neisseria meningitidis , Streptococcus agalactiae , and Cryptococcus neoformans/gattii . Across the remaining shared targets, positive percent agreement ranged from 50% to 93% and negative percent agreement from 98% to 100%. Enterovirus showed the greatest discordance, with 86% positive percent agreement and 10 FilmArray-only detections. LDT testing confirmed low-level enterovirus positivity in four discordant specimens (100‒700 copies/mL), and discordant viral detections clustered at lower target burden than concordant positives ( P < 0.001). Molecular typing identified enterovirus species A, B, and C in both concordant and FilmArray-only positive specimens. In conclusion, FilmArray and QIAstat-Dx showed high agreement for CSF pathogen detection. Residual discordance was largely confined to low-burden viral detections, particularly enterovirus, and was more consistent with near-threshold target abundance than with a major type-restricted detection gap. These findings support routine syndromic CSF testing and emphasize interpretation of borderline positives in conjunction with confirmatory LDT testing and clinical context. IMPORTANCE Syndromic cerebrospinal fluid panels are increasingly used as first-line tests for suspected central nervous system infection, yet aggregate agreement metrics can obscure clinically relevant differences near the limit of detection. In this multicenter comparison, most discordance clustered in low-burden enterovirus-positive cerebrospinal fluid (CSF) specimens and showed no clear segregation by enterovirus type, indicating that disagreement largely arose from near-threshold detection rather than a major gap in assay inclusivity. The broader implication is that performance comparisons of syndromic panels should move beyond overall agreement and specifically address how platforms handle borderline-positive results. This matters clinically because low-level viral detections can alter post-test probability, antimicrobial stewardship, CSF resampling, and follow-up microbiologic workup. More generally, the data support laboratory-integrated diagnostic strategies that combine multiplex screening with quantitative adjudication, molecular characterization, and interpretation of semiquantitative signals to strengthen the analytical and clinical utility of syndromic testing for central nervous system infections.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.