CSF mNGS is a diagnostic tool designed to overcome the limits of traditional diagnostics by sequencing all nucleic acids in the CSF and comparing results to a microbial database. The method is especially useful in difficult and critical clinical presentations when infection is suspected or discovering novel pathogens when traditional studies fail. However, the optimal scenarios in which to deploy this unbiased sequencing method are not well-defined.

To evaluate the impact of CSF metagenomic next-generation sequencing (mNGS) on the diagnosis and treatment of neurological syndromes.

This is a retrospective study of 94 mNGS results processed at Johns Hopkins Hospital between 2020-2024. Demographic and clinical data were collected via electronic medical record review.

Among all processed mNGS samples, 27 were positive and 67 were negative. 19 (72%) positive results had a diagnostic impact in identifying infection and 20 (30%) negative results in ruling out infection. 50 samples were collected after the initiation of antimicrobials. Overall, there was treatment change in 6 cases based on the mNGS result. The average time to CSF collection was 5.5 days among negative results and 6.7 among positive results. The average length of stay among all inpatients tested was 22 days. 49 (54%) patients admitted to the hospital were discharged to home.

There was no difference between groups in time to CSF sampling, reflective of the need for collection for traditional testing. However, with several days between admission and sampling, there is likely decreased diagnostic yield in mNGS. Nearly half the patients had poor functional outcomes on discharge and were unable to be discharged to home, evidence of the bias of using mNGS on sicker patients. Patients often received full courses of antimicrobials prior to mNGS results and many samples were collected after initiation, which affects diagnostic yield and increases the risk of unnecessary antimicrobial use.