During February 2018–January 2020, NICD received 102 V. cholerae isolates for testing; 9 were identified as V. cholerae O1. We characterized the bacteria by whole-genome sequencing, comparative genomics, and phylogenetic analysis (Appendix 1). The Human Research Ethics Committee of the University of the Witwatersrand (Johannesburg, South Africa) provided ethics approval for this study (protocol no. M160667).

Of 9 V. cholerae O1 isolates tested, we identified 2 ST69 (7PET) and 7 ST75 isolates. The ST69 isolates were collected in October 2018 from 2 cholera patients in a family cluster. The index case-patient had traveled to Zimbabwe, where an outbreak was ongoing (8), within the 7-day cholera incubation period before symptom onset. We confirmed these ST69 isolates belonged to the previously described highly antimicrobial-resistant Zimbabwe outbreak strain (8). The 7 ST75 isolates originated from KwaZulu-Natal and Limpopo Provinces. Five isolates were collected from patients with cholera, all adults 37–57 years of age; 2 isolates were from environmental samples collected during case investigations, 1 from sewage in Limpopo Province and 1 from river water in KwaZulu-Natal Province (Table 1). The 3 KwaZulu-Natal cases occurred ≈200–600 km apart; the first occurred in February 2018 and the last in January 2020. The 2 Limpopo cases occurred ≈70 km apart in the same district during November 2018. The Limpopo cases were >900 km from the KwaZulu-Natal cases. Epidemiologic investigations involved interviewing case-patients by using a standard case investigation form; visiting case-patients’ residences to inspect water and sanitation services and interview other household members; collecting stool samples from household members; and collecting environmental samples when indicated. Investigators found no evidence of importation from another country, epidemiologic links between cases, or secondary transmission.

The 7 ST75 isolates showed notable features (Table 2). In particular, all carried the cholera toxin (CTX) prophage resembling CTX-2 with ctxB1 genotype; Vibrio pathogenicity island 1 (VPI-1) encoding the toxin co-regulated pilus; and a variant form of Vibrio pathogenicity island 2 (VPI-2). However, isolates did not contain Vibrio seventh pandemic island I (VSP-I) and VSP-II. We noted several genomic islands (GIs), including VC-GI 119, but GI-05 was not present (Appendix 2).

The only antimicrobial-resistance determinant found in all ST75 isolates was the qnrVC4 gene, located in the chromosomal superintegron. Various qnrVC alleles previously have been reported in the Vibrionaceae family and sometimes are associated with fluoroquinolone resistance (10,11). However, all ST75 isolates we analyzed showed fluoroquinolone susceptibility, MIC of ciprofloxacin 0.06 µg/mL, and susceptibility to all other tested antimicrobial drugs. This pansusceptibility sharply contrasts antimicrobial resistance trends observed in 7PET isolates from Africa, which reportedly became increasingly antimicrobial resistant over time; after the 2000s, none were susceptible to antimicrobial agents (5).

Figure

Figure. Maximum-likelihood phylogenomic tree for Vibrio cholerae O1 sequence type (ST) 75 isolates collected from South Africa, 2018–2020. The tree represents phylogeny for 7 V. choleraeO1 ST75...

We further compared the ST75 isolates from South Africa with a larger global collection of 144 ST75, or closely related ST169, ST170, and ST182, genomes (Appendix 2), and constructed a maximum-likelihood phylogeny by using 49,540 SNPs (Figure). Our phylogenetic analysis showed that the 7 isolates from South Africa clustered in the L3b.1 clade, defined by H. Wang et al. (9), with a maximum pairwise distance of 22 SNPs. Isolates from Limpopo Province had a maximum pairwise distance of 1–6, but KwaZulu-Natal Province isolates had no SNP differences. Core-genome MLST showed Limpopo Province isolates differed from the KwaZulu-Natal Province isolates by 4–5 alleles (Appendix 1 Figure). The closest related isolates were collected in Russia from Rostov Oblast in 2005 and Republic of Kalmykia in 2011 and from Turkmenistan in Central Asia in 1965, but none of those isolates contained the CTX prophage. L3b.1 isolates from Taiwan containing the CTX prophage ctxB3 allele were more distant.

Emergence of ST75 L3b.1 clade in South Africa is cause for concern. Recent studies on V. cholerae O1 isolated in Taiwan (12) and China (13) reported emerging and potential toxigenic ST75. Genomic signatures of these ST75 isolates closely resembled the US Gulf Coast V. cholerae O1 clone that emerged in 1973 (14). In particular, an investigation of V. cholerae O1 isolated during 2002–2018 in Taiwan showed that ST75 emerged there in 2009 and now is more prevalent than the ST69 pandemic clone (12). Our findings from South Africa align with the findings from Taiwan, showing that ST75 isolates outnumber ST69 isolates.

One limitation of our study is that we used reference laboratory data and a review of published V. cholerae O1 data to conclude that all previous cholera isolates in South Africa characterized by MLST were V. cholerae O1 biotype El Tor ST69. However, we cannot exclude the possibility that V. cholerae O1 isolates not characterized by MLST, particularly those from environmental samples, could have been non-ST69.

Epidemic 7PET lineage cholera demands an aggressive public health response to prevent outbreaks. In contrast, sporadic V. cholerae O1 infections mediated by other lineages, including those carrying toxin co-regulated pilus and CTX genes, typically are not epidemic-prone; most are associated with sporadic cases that rarely lead to secondary transmission (15). Tailoring the public health response to the degree of epidemic risk would be invaluable, especially in resource-limited settings. In countries that are not cholera-endemic but are at high risk for cholera introductions, conventional laboratory determination of V. cholerae O1, even complemented by identifying ctxA or ctxB genes, might be insufficient. Typing resolution of genomics, which distinguishes between 7PET and nonepidemic lineages, can elucidate the local and global epidemiology of cholera and inform public health decisions.

The emergence and dominance of nonepidemic, non-7PET, V. cholerae ST75 L3b.1 in South Africa requires close monitoring. The spatiotemporal pattern suggests local spread, possibly indicating a geographically widespread risk for sporadic disease from this strain. South Africa should strengthen its disease and environmental surveillance systems to identify non-pandemic ST75 strains, define local epidemiology, and inform an appropriate public health response.