During October 2010–January 2011, a total of 122 clinical isolates of laboratory-confirmed V. cholerae O1 were recovered by the National Public Health Laboratory in Haiti and submitted to the Centers for Disease Control and Prevention (CDC; Atlanta, GA, USA) for characterization. Disk-diffusion antimicrobial drug susceptibility testing was performed at the National Public Health Laboratory and CDC. MICs were determined by broth microdilution at CDC by using Sensititer plates (CAMPY and CMV1AGNF; Trek Diagnostics, Cleveland, OH, USA) according to the manufacturer’s instructions with the following modifications: Mueller-Hinton broth without blood was used on the CAMPY plate, and for both plates, a final inoculum concentration of 5 × 10⁴ to 5 × 10⁵ CFU/mL was targeted. Escherichia coli American Type Culture Collection (ATCC; Manassas, VA, USA) 25922, Staphylococcus aureus ATCC 29213, Enterococcus faecalis ATCC 29212, and Pseudomonas aeruginosa ATCC 27853 were used for quality control testing. Where available, Clinical and Laboratory Standards Institute criteria specific for V. cholerae were used (7). For drugs lacking such criteria, manufacturers’ criteria, Clinical and Laboratory Standards Institute criteria for Enterobacteriaceae, or consensus breakpoints used by the National Antimicrobial Resistance Monitoring System were applied (8,9). Furazolidone was tested only by disk diffusion, and azithromycin was tested only by broth microdilution.

Results for all 122 outbreak isolates were similar. They showed susceptibility to azithromycin and tetracycline, reduced susceptibility to ciprofloxacin (MIC 0.25–1.0 mg/L), and resistance to furazolidone, nalidixic acid, sulfisoxazole, streptomycin, and trimethoprim/sulfamethoxazole.

With a common susceptibility pattern among all outbreak isolates, 1 isolate, 2010EL-1786 (deposited under ATCC BAA-2163), was chosen for molecular characterization. PCR was used to screen the isolate for the following resistance genes: strA, strB, sul1, sul2, dfrA1, dfrA10, and dfrA12 (10). In addition, the gyrA and parC genes were sequenced to identify quinolone resistance–determining region mutations. PCR was performed according to standard protocols by using the HotStarTaq PCR Master Mix (QIAGEN, Valencia, CA, USA). DNA sequencing was performed by using a 3730 DNA Analyzer (Applied Biosystems, Foster City, CA, USA).

The isolate 2010EL-1786 contained strA/B, sul2, and dfrA1, which mediate resistance to streptomycin, sulfisoxazole/sulfamethoxazole, and trimethoprim, respectively. Nalidixic acid resistance and decreased susceptibility to ciprofloxacin were attributed to mutations in gyrA (Ser83Ile) and parC (Ser85Leu). The mechanism responsible for furazolidone resistance was not identified. Mutations in the nfsA and nfsB genes are associated with furazolidone resistance in E. coli, but inspection of the 2010EL-1786 sequence failed to identify these genes.

Location of the resistance genes was analyzed by whole-genome sequencing of 2010EL-1786. Single-end 454 pyrosequencing (GS FLX-Titanium; Roche Diagnostics, Indianapolis, IN, USA) reads and single-end 36-bp Illumina Solexa (GAIIe; Illumina, San Diego, CA, USA) reads were assembled de novo by using Newbler (Roche Diagnostics) and CLC Genomics Workbench (CLC bio, Cambridge, MA, USA) software. Sequence finishing was performed by using Sanger sequencing of fosmid clones (11).

Figure 1

Figure 1. Genetic organization of the 2010 Haiti Vibrio cholerae O1 integrating conjugative element (ICE), ICEVchHai1. The ICE contained 97,915 bp and 95 open reading frames. Coding sequences were identified and manually annotated...

Whole-genome sequencing identified an ICE inserted in the prfC gene. This ICE, designated ICEVchHai1, was 97.9 kb and contained 95 open reading frames (Figure 1). All resistance genes identified were located on ICEVchHai1. In addition, floR, a chloramphenicol resistance gene, was detected. The strA, strB, sul2, and floR genes were part of an ≈17-kb fragment inserted into the rumB gene, whereas the dfrA1 gene was located ≈70 kb further downstream. Whole-genome sequencing also indicated a chloramphenicol acetyltransferase gene, catB9, that was not part of the ICE.

Figure 2

Figure 2. Phylogenetic tree illustrating the genetic relatedness between the Haiti integrating conjugative element (ICE) ICEVchHai1 and other ICEs described in Vibrio cholerae (ICEVchBan1, ICEVchBan9, ICEVchInd1, ICEVchInd4, ICEVchInd5, ICEVchMex1, and SXT). Each ICE...

The genetic relatedness of ICEVchHai1 was assessed by comparison with 7 other ICE sequences (4). Sequence alignments were performed by using Progressive Mauve (http://asap.ahabs.wisc.edu/mauve/download.php) and visualized with PHYLIP version 3.69 (distributed by J. Felsenstein, Department of Genome Sciences, University of Washington, Seattle, WA, USA). ICEVchHai1 showed highest homology to ICEVchInd5, an ICE derived from a V. cholerae isolate from India (Figure 2). These ICEs differed by only 5 single-nucleotide polymorphisms.

In October 2010, an epidemic caused by toxigenic V. cholerae O1, serotype Ogawa, biotype El Tor strain, was reported from Haiti. We confirmed that the outbreak strain was multidrug resistant and displayed resistance to furazolidone, nalidixic acid, sulfisoxazole, streptomycin, and trimethoprim/sulfamethoxazole and decreased susceptibility to ciprofloxacin. Genetic mechanisms responsible for resistance to 5 of these drugs were identified. Sequencing also detected floR, a gene commonly associated with chloramphenicol resistance in Enterobacteriaceae (MICs >32 mg/L) (12). However, in this study, floR was not associated with resistance; isolates from Haiti displayed chloramphenicol MICs of 4–16 mg/L (7). Clinical non-Haiti isolates lacking the floR gene displayed MICs <1 mg/L. Why the floR gene did not confer resistance in V. cholerae remains to be investigated but might be because of lower expression levels. Sequencing also showed a chloramphenicol acetyltransferase gene, catB9. How this gene affects chloramphenicol MICs in V. cholerae remains to be determined.

Most of the acquired resistance genes were located on an ≈97-kbp ICE termed ICEVchHai1. The presence of an ICE in V. cholerae from Haiti was documented by Chin et al. in 2011 (13). ICEVchHai1 showed high homology to ICEVchInd5, an ICE first identified in a V. cholerae isolate from Sevagram, India, in 1994. Since then, ICEVchInd5 has persisted among clinical isolates in India; a recent study of O1 strains isolated in India during 1994–2005 confirmed that ICEVchInd5 was the only ICE that persisted during the study period (14).

Drug-resistant V. cholerae is a global health concern because resulting infections can be more severe and difficult to treat. Infections with drug-resistant V. cholerae can result in higher case-fatality rates, prolonged hospitalizations, more secondary infections, and increased health care costs. During an outbreak in Guinea-Bissau, case-fatality rates increased from 1% to 5.3% after the outbreak strain acquired multidrug resistance (15). To limit development and spread of antimicrobial drug resistance among V. cholerae, treatment with antimicrobial agents should be restricted to patients with severe dehydration or other conditions that truly warrant their use. Surveillance should continue for antimicrobial drug resistance among V. cholerae isolates from Haiti.