Figure 2

Figure 2. Phylogenomics of the Vibrio choleraeO1 El Tor isolates from CAR, 1997–2016. A) Maximum-likelihood phylogeny for 1,215 7PET genomic sequences. A6 was used as an outgroup. The last 9...

The 30 V. cholerae O1 isolates were received at the Institut Pasteur, Paris, France. We performed antimicrobial susceptibility testing, whole-genome sequencing, comparative genomics, and phylogenetic analyses by using methods previously described (3,5–11) (Appendix 2). We then contextualized these 30 V. cholerae O1 isolates within a global collection of 1,185 seventh pandemic V. cholerae El Tor (7PET) genomic sequences and constructed a maximum-likelihood phylogeny of 1,215 genomes by using 9,964 single-nucleotide variants (SNVs) evenly distributed over the nonrepetitive, nonrecombinant core genome (Figure 2, panel A). Our phylogenomic analysis (Appendix 1 Tables 2–4) showed that all CAR isolates belonged to the 7PET lineage (12,13).

Previous genomic studies described 12 introductions of the 7PET lineage from Southern Asia into Africa during 1970–2016 (3,6). The introduced sublineages were called T1 and T3–T13. The 2 cholera outbreaks in CAR in 1997 were caused by sublineage T7, which had been introduced into West Africa during the early 1980s, and T9, which was introduced in the late 1980s, according to Weill et al. (3). T9 isolates were identified in neighboring countries such as Chad and Cameroon (particularly northern Cameroon) before they were detected in northwest CAR, but T7 isolates were identified in Gabon and the western part of DRC, along the Congo River in Kinshasa, before being identified in CAR along the Oubangui River, a tributary of the Congo River (Figure 2, panel B; Appendix 2 Figures 1, 2).

The 2011 and 2016 outbreaks were caused by closely related bacterial populations from the same sublineage, T10, introduced into East Africa during the 1990s and later detected in the African Great Lakes Region (AGLR) (3) (Figure 2, panel B; Appendix 2 Figure 3). The prevalent T10 sublineage has several clades, and the 2011 and 2016 CAR isolates are characterized by multilocus sequence type (ST) 515, a single-locus variant of ST69; ST69 is the predominant ST among 7PET isolates. In addition, these isolates have an alteration to the wbeT gene, a 4-nucleotide deletion called B01 (Tables 1,2), that underly the Inaba serotype. Phylogenetic data showed that the 7PET strains causing the 2011 and 2016 cholera outbreaks in CAR spread from AGLR to the western part of the DRC and CAR (Figure 2, panel B; Appendix 2 Figure 3). These genomic data are consistent with results from a recent epidemiologic study showing the spatiotemporal distribution of cholera cases during the 2011–2012 and 2015–2017 outbreaks in DRC, suggesting a spread from cholera hotspots in AGLR to major cities in the upstream section of the Congo River, followed by a downstream spread toward the densely populated CAR capital, Kinshasa, and then to the mouth of the Congo River, which opens into the Gulf of Guinea (14).

Except during the outbreak in northwest CAR in 1997, all cholera cases were reported along or close to the Oubangui River, suggesting that 7PET strains probably moved from area to area along the river and with the displacement of human populations. The risk factors in these remote areas are unknown, but the prevailing conditions, such as poor hygiene and sanitary conditions, overcrowding, lack of latrines, and drinking water from the Oubangui River, likely would increase the risk for transmission via the fecal–oral route, as evidenced by the high attack rates observed at several sites when the 2011 outbreak began (15). Nevertheless, since the declaration of the first case in 1997, the small number (<1,000) of cholera cases in CAR contrasts with the much larger numbers in central Africa (2). The low population density of CAR, its poor transport infrastructure, and poor trading links are probably key factors limiting disease spread (15). Phylogenetic analyses showed no other isolates from Africa were derived from CAR isolates in the aftermath of the 4 outbreaks, which also suggests that the transmission of cholera is impeded in this country. Of note, all 4 outbreaks were caused by serotype Inaba 7PET strains. This serotype has a nonmethylated form of lipopolysaccharide caused by an alteration to the wbeT gene (3) (Table 2). The implication of this serotype in all 4 outbreaks suggests that these 7PET sublineages circulated regionally for some time, long enough to acquire this alteration to the wbeT gene, before reaching CAR (Appendix 2 Figures 1–3).

All CAR isolates in this study displayed resistance to polymyxin B, consistent with the susceptibility pattern reported for the El Tor biotype until recently (6). All but 1 of the isolates collected along the Oubangui River in 1997 were susceptible to all other antimicrobial drugs tested; the outlying isolate contained the extended-spectrum β-lactamase bla_SHV-2a gene on an IncA/C2 plasmid (Table 1; Appendix 1 Table 3). No susceptible isolates have been collected in CAR since. All the other isolates display mutations of the VC_0715 and VC_A0637 genes, conferring nitrofuran resistance, and carry an SXT/R391 genomic element called ICEVchInd5, encoding resistance to streptomycin (strAB), sulfonamides (sul2), trimethoprim, the O/129 vibriostatic agent (dfrA1), and trimethoprim-sulfamethoxazole (sul2 and dfrA1). The 2011 and 2016 CAR isolates also recently acquired gyrA mutations (Table 1), resulting in resistance to nalidixic acid (Appendix 1 Table 3).

Strains from 3 7PET sublineages caused 4 cholera outbreaks identified in CAR during 1997–2016. The southern and southeastern parts of CAR are higher risk areas for cholera outbreaks, particularly when cases are reported in the western part of DRC. These findings highlight the need for an effective surveillance system, and for coordinated communication actions on cholera that target healthcare professionals and the populations living along the Oubangui River, to prevent and control cholera outbreaks in CAR.