Metagenomic Identification of Fusarium solani Strain as Cause of US Fungal Meningitis Outbreak Associated with Surgical Procedures in Mexico, 2023

Abstract

We used metagenomic next-generation sequencing (mNGS) to investigate an outbreak of Fusarium solani meningitis in US patients who had surgical procedures under spinal anesthesia in Matamoros, Mexico, during 2023. Using a novel method called metaMELT (metagenomic multiple extended locus typing), we performed phylogenetic analysis of concatenated mNGS reads from 4 patients (P1-P4) in parallel with reads from 28 fungal reference genomes. Fungal strains from the 4 patients were most closely related to each other and to 2 cultured isolates from P1 and an additional case (P5), suggesting that all cases arose from a point source exposure. Our findings support epidemiologic data implicating a contaminated drug or device used for epidural anesthesia as the likely cause of the outbreak. In addition, our findings show that the benefits of mNGS extend beyond diagnosis of infections to public health outbreak investigation.

Keywords: Fusarium solani; Mexico; Nectria haematococca; SURPI+ computational pipeline for pathogen detection; United States; agnostic pathogen detection; fungi; mNGS; meningitis/encephalitis; metaMELT; metagenomic next-generation sequencing; multiple extended locus typing; outbreak investigation.

Figure 1

Identification of Fusarium solani in CSF from a patient with fungal meningitis associated with surgical procedures in Matamoros, Mexico, 2023. A) Heat map of 13 mNGS CSF reads from patient P1 showing reads aligning to Nectria hematococca, the anamorph of F. solani. That initial detection triggered a Centers for Disease Control and Prevention notification recommending testing for all patients in the outbreak who might have been exposed (14). B) Mapping of the 18 total Fusarium spp. mNGS reads recovered from patient P1 to the F. solani reference genome (ATCC MYA-4622). Arrows indicate the mapped positions of the 18 reads. CSF, cerebrospinal fluid; mNGS, metagenomic next-generation sequencing; RPM(pp), reads per million (preprocessed).

Figure 2

Flowchart showing the metaMELT analysis workflow used for identification of Fusarium solani strain as cause of fungal meningitis US outbreak associated with surgical procedures in Mexico, 2023. metaMELT (metagenomic multiple extended locus typing), is a novel analytic technique for simultaneously diagnosing the infection and characterizing the interrelatedness of F. solani strains. metaMELT uses the following steps: A) perform mNGS analysis of patient samples (i.e., cerebrospinal fluid, plasma, or brain tissue), using the SURPI+ computational pipeline (https://github.com/chiulab/SURPI-plus-dist) to identify pathogens; B) identify F. solani reads; C) map reads to the F. solani reference genome and then extract and concatenate; D) perform phylogenetic analysis on concatenated sequences. SURPI, sequence-based ultra-rapid pathogen identification.

Figure 3

Diagram of concatenation and alignment step of the metaMELT procedure (metagenomic multiple extended locus typing, a novel analytic technique for simultaneously diagnosing the infection and characterizing the interrelatedness of Fusarium solani strains) used for identification of F. solani strain as cause of fungal meningitis US outbreak associated with surgical procedures in Mexico, 2023. The diagram shows concatenated metagenomic next-generation sequencing reads from 4 patients and the corresponding regions extracted from reference genomes, which are aligned to by using MAFFT version 7.388 (22). This diagram demonstrates the steps shown in Figure 2, panels C–D.

Figure 4

Phylogenetic analysis of concatenated metagenomic next-generation sequencing reads from US patients from a fungal meningitis outbreak associated with surgical procedures in Mexico, 2023. A) Phylogenetic trees showing clustering of strains from patients P1–P5 (pink shaded region) within a subclade that also includes fungal genomes unrelated to the outbreak (dotted rectangle). B–E) Phylogenetic trees of individual patients exhibiting similar topologies: B) P1; C) P2; D) P3; E) P4. Each patient is positioned in a cluster containing the same reference genomes, including the 2 outbreak genomes recovered from patients P1 and P5. Outbreak reads were mapped to corresponding regions from Fusarium solani reference genomes by using metaMELT (metagenomic multiple extended locus typing, a novel analytic technique for simultaneously diagnosing the infection and characterizing the interrelatedness of F. solani strains). Scale bars indicate nucleotide substitutions per site. P1–P5, patients 1­–5.

Figure 5

Effect of number of reference genomes on performance of metaMELT (metagenomic multiple extended locus typing, a novel analytic technique for simultaneously diagnosing the infection and characterizing the interrelatedness of Fusarium solani strains) for identification of F. solani strain as cause of fungal meningitis US outbreak associated with surgical procedures in Mexico, 2023. A, B) metaMELT phylogenetic trees that include mNGS reads from patients P1–P4 are shown with and without outbreak-related genomes: A) only 1 outbreak-related reference genome from P5; B) tree without any outbreak-related reference genomes. Note the clustering of patients P1–P4 (pink shaded regions) even in the absence of an outbreak-related reference genome. C, D) metaMELT phylogenetic trees that include mNGS reads from patients P1–P4 are shown with only 10 reference genomes, including patient P5 (C); and only 5 reference genomes, including patient P5 (D). Scale bars indicate nucleotide substitutions per site. P1–P5, patients 1­–5.