Performance of Metagenomic Next-Generation Sequencing for the Diagnosis of Viral Meningoencephalitis in a Resource-Limited Setting

Abstract

Abstract Background Meningoencephalitis is a devastating disease worldwide. Current diagnosis fails to establish the cause in ≥50% of patients. Metagenomic next-generation sequencing (mNGS) has emerged as pan-pathogen assays for infectious diseases diagnosis, but few studies have been conducted in resource-limited settings. Methods We assessed the performance of mNGS in the cerebrospinal fluid (CSF) of 66 consecutively treated adults with meningoencephalitis in a tertiary referral hospital for infectious diseases in Vietnam, a resource-limited setting. All mNGS results were confirmed by viral-specific polymerase chain reaction (PCR). As a complementary analysis, 6 viral PCR-positive samples were analyzed using MinION-based metagenomics. Results Routine diagnosis could identify a virus in 15 (22.7%) patients, including herpes simplex virus (HSV; n = 7) and varicella zoster virus (VZV; n = 1) by PCR, and mumps virus (n = 4), dengue virus (DENV; n = 2), and Japanese encephalitis virus (JEV; n = 1) by serological diagnosis. mNGS detected HSV, VZV, and mumps virus in 5/7, 1/1, and 1/4 of the CSF positive by routine assays, respectively, but it detected DENV and JEV in none of the positive CSF. Additionally, mNGS detected enteroviruses in 7 patients of unknown cause. Metagenomic MinION-Nanopore sequencing could detect a virus in 5/6 PCR-positive CSF samples, including HSV in 1 CSF sample that was negative by mNGS, suggesting that the sensitivity of MinION is comparable with that of mNGS/PCR. Conclusions In a single assay, metagenomics could accurately detect a wide spectrum of neurotropic viruses in the CSF of meningoencephalitis patients. Further studies are needed to determine the value that real-time sequencing may contribute to the diagnosis and management of meningoencephalitis patients, especially in resource-limited settings where pathogen-specific assays are limited in number.