Login / Signup

Assessment of a Multiplex LAMP Assay (Eazyplex® CSF Direct M) for Rapid Molecular Diagnosis of Bacterial Meningitis: Accuracy and Pitfalls.

Anne-Gaëlle LeroyElise PersynSophie-Anne GibaudLise CrémetPaul Le TurnierMyriam BenhamidaElise LaunayAurélie GuillouzouicPascale BemerStéphane Corvecnull On Behalf Of The Western French Study Group On Early Bacterial Meningitis
Published in: Microorganisms (2021)
Background: Automated molecular panels are attractive tools for improving early meningitis diagnosis. This study assessed the Eazyplex® CSF direct M panel (EP), a multiplex real-time Loop-Mediated Isothermal Amplification assay. Methods: From December 2016 to December 2019, cerebrospinal fluid (CSF) samples were routinely tested with the EP V1.0. CSF parameters and microbiological and clinical data were retrospectively collected. Results: Out of 230 CSF samples, the EP yielded positive, negative, and invalid results for 32 (13.9%) (16 N. meningitidis, nine S. pneumoniae, two S. agalactiae, two E. coli, two H. influenzae, one L. monocytogenes), 182 (79.1%), and 16 (7%) samples, respectively. Among the positive samples, 14 (44%) remained negative in culture (antibiotic therapy before lumbar puncture (n = 11), meningococcal meningitis (n = 3)). High CSF protein concentrations and cellularity were associated with LAMP inhibition, counteracted by centrifugation. The automated software yielded 13 false positive and five false negative results. Amplification curve analysis was necessary and enabled the attainment of positive (PPA) and negative percentage agreement and positive and negative predictive values of 91.4%, 100%, 100%, and 98.3%. Three false negative results remained (two E. coli and one N. meningitidis). E. coli presented the poorest PPA (50%). Conclusion: This work confirms the strong performance of the EP, of particular interest in cases of antibiotic therapy before lumbar puncture.
Keyphrases
  • cerebrospinal fluid
  • loop mediated isothermal amplification
  • high throughput
  • escherichia coli
  • machine learning
  • minimally invasive
  • deep learning
  • ultrasound guided
  • binding protein
  • quantum dots