Outbreak of Sexually Transmitted Nongroupable Neisseria meningitidis–Associated Urethritis, Vietnam

We report on an outbreak of nongroupable Neisseria meningitidis–associated urethritis, primarily among men who have sex with men in southern Vietnam. Nearly 50% of N. meningitidis isolates were resistant to ciprofloxacin. This emerging pathogen should be considered in the differential diagnosis and management of urethritis.

characteristics, medical and immunization history, access to social networks, and sexual and drug use behaviors.The controls were also interviewed using the same questionnaire.

Isolate Collection
At the HHDV, men presenting with urinary discharge were initially suspected to have gonococcal urethritis.However, 19 of the collected specimens tested negative for Gonococci using real-time PCR and subsequent bacterial culture yielded 19 isolates of Gram-negative diplococci, which could not be identified as gonococci.These isolates were then transferred to Pasteur Institute of Ho Chi Minh for further laboratory examination to confirm the actual etiology.

Identification and Characterization of N. meningitidis
To identify N. meningitidis, two methods were used: 1) a biochemical test kit of Analytical Profile Index Neisseria-Haemophilus (API-NH, bioMérieux); and 2) species-specific superoxide dismutase gene (¬sodC) real-time PCR (Quanta Bioscience).A slide agglutination serogrouping (SASG) kit of N. meningitidis Antiserum (Difco) and serogroup-specific (cps) gene rt-PCR were used to determine the serogroup of the samples, among A, B, C, W, X, Y or nongroupable (NG).

Antimicrobial Susceptibility Testing
Antimicrobial susceptibilities for the urethral isolates were determined through MIC with E-test, using six antibiotics: penicillin, cefotaxime, meropenem, azithromycin, ciprofloxacin, and rifampin (Liofilchem).The Clinical and Laboratory Standard Institute (CLSI) 2017 was used to determine breakpoints for susceptible, intermediate, or non-susceptible/resistant (1).

Whole-Genome Sequencing
Whole-genome sequencing of 19 urethral N. meningitidis isolates was performed at the PIHCM, Vietnam, from which DNA was extracted using the QIAmp DNA mini kit (QIAGEN).
Libraries were then constructed using the Nextera XT DNA Library Prep Kit v2.0 (Illumina).
Multilocus sequence typing (MLST), fine-typing antigens, targeted outer-membrane proteins (OMV), and antibiotic resistance genes were obtained through BIGSdb (4).All draft genomes were submitted to the National Center for Biotechnology Information (NCBI) under BioProject no.PRJNA672782 and were also uploaded to the PubMLST database.

Phylogenetic Analysis
Phylogenetic inferences based on differences in the gene-by-gene comparison of 1,605 defined loci of the core-genome MLST were made.The genomes of the isolates in Vietnam were compared against those in the United States (n = 209) and United Kingdom (n = 2), using an invasive isolate, M21273, as a reference (5,6).The phylogeny was inferred using the PubMLSTbased Genome Comparator tool.
A subset of 81 genomes of the US NmNG UC isolates and 7 of invasive CC-11 as an outgroup was inferred with a time-measure phylogeny with BEAST v1.10 (7).A concatenated core alignment of 1,275,571 bps was generated through the genome-comparator and aligned with MAFFT.Subsequently, it was masked with Gubbins v2.2.1, leaving 1,081 bp of polymorphic sites (8).The Gubbins-masked alignment was generated using maximum-likelihood phylogeny with IQ-Tree, replicating 1,000 ultrafast-bootstrap (9).
To examine temporal signals, a root-to-tip linear regression was constructed with TempEst v1.5.3, estimating correlation coefficient was 0.91 and R 2 was 0.83 (10).A timemeasured phylogeny was inferred by BEAST/BEAGLE, using the general time-reversible (GTR) substitution model and gamma heterogeneity site.The uncorrelated exponential relaxed clock (UCED), which allows each branch to evolve independently, was selected over the other molecular clocks, such as relaxed uncorrelated lognormal (UCLN), random local clock (RLC), and strict clock.The model was selected by using generalized stepping-stone sampling and estimation of the most recent common ancestor (TMRCA) of the subset.The marginal likelihood estimation was also used to evaluate the flexible nonparametric Skygrid model and other parametric models: constant, expansion, logistic, exponential growth, and GMRF Skyride.The Bayesian Skygrid was performed with 87 parameters corresponding to 88 sequences, and 17 grid points as the length of time from the newest to the root of the tree, running 250 million steps for two separate times.
Urethral swab specimens were collected following the hospital routine protocol to test for gonorrhea, Chlamydia, mycoplasma, and ureaplasma.From each participant 3 mL of venous blood was drawn to test for HIV and syphilis.
HIV was screened using a rapid test (Determine HIV-1/2, Alere Medical Co., Matsudo, Japan) at HHDV.Positive specimens were sent to Center for Disease Control of Ho Chi Minh City to confirm positivity with three different tests: Advia Centaur HIV Ag/Ab Combo (CHIV) Assay, Siemens Healthcare Diagnostics Inc., U.S.; SD.HIV 1/2 3.0, SD.Korea Co., Ltd, Korea; and Determine HIV, Alere Medical Co., Matsudo, Japan.Testubg followed the national HIV testing algorithm.

Data Management and Analysis
All interview answer sheets were reviewed by the investigators for any missing information.These sheets were stored in locked cabinets in HHDV.Data were entered using Epi-Data version 3.1 (EpiData Association, Odense, Denmark), and all statistical analyses were conducted using Stata version 14.0 (StataCorp, Station, TX, USA).Continuous variables were described using mean, median and range.Categorical variables were presented as proportions.
Conditional univariate and multivariate logistic regressions were used to assess risk factors for US NmNG urethritis.Given the small sample size, we used forward selection to add in variables one by one: those variables in univariate analysis giving p values lower than the others were included in a multivariate model.A log likelihood-ratio test was used to select the better fit model between the former (without the add-in variable) and the current (with the add-in variable) models.If the test gave p<0.1, the newly added variable was retained in the model.If any two variables were thought to be highly correlated with each other, only one was included in multivariate analysis.

Table 1 .
Coinfection with other sexually transmitted infections among 19 cases of United States Neisseria meningitidis urethritis clade*

Table 2 .
Genome typing of isolates from 19 cases of United States Neisseria meningitidis urethritis clade, Vietnam*