Objective:

To assess the diagnostic value of shallow-depth cerebrospinal fluid (CSF) metagenomic next-generation sequencing (mNGS) for central nervous system (CNS) infections in Lima, Peru.

Background:

CNS infections remain diagnostically challenging, particularly in low- and middle-income countries. mNGS enables broad-range pathogen detection with a single test. Shallow-depth mNGS offers a cost-effective approach that could facilitate wider adoption, but its diagnostic performance remains unknown.

Design/Methods:

Adult patients with suspected CNS infection and surplus CSF (after routine clinical testing) were enrolled at Hospital Nacional Dos de Mayo. Patients who received antibiotics for >7 days before lumbar puncture were excluded. DNA/RNA extraction, library preparation, pooling, and sequencing (Illumina NextSeq 500/550) were preformed locally at Universidad Peruana Cayetano Heredia. Sequencing data were processed using the open-source Chan-Zuckerberg ID software for pathogen detection via computational filtering and manual review.

Results:

Forty-three patients were included (32.6% female; median age 47.5). RNA libraries were sequenced in one run with mean depth of 13,033,943 reads per sample (SD 9,110,614), while DNA libraries were sequenced in two runs with mean depth of 30,203,814 reads per sample (SD 18,643,933). Pathogens were detected via mNGS in 51.1% of patients (n=22), including 41.2% diagnosed exclusively by mNGS. Nine patients with probable or possible tuberculous meningitis were found to have alternative non-tuberculous infections. Many pathogens detected exclusively via mNGS represent treatable conditions with medications available through the Peruvian public health system, including cerebral toxoplasmosis, neurocysticercosis, neurosyphilis, and varicella zoster virus.

Conclusions:

Shallow-depth CSF mNGS demonstrated substantial diagnostic utility in a resource-limited setting, identifying pathogens in over half of cases and providing exclusive diagnostic information in over 40% of patients, supporting its potential for broader implementation in CNS infection diagnosis. Further validation of these results is underway, including direct comparison to deeper sequencing depths more traditionally used in clinically available mNGS assays.