The microbiology laboratory at St. Paul’s Hospital (Vancouver, BC, Canada) serves acute-care hospitals and the surrounding community in downtown Vancouver. We searched the laboratory database for S. sonnei found in feces and blood samples during January 2010–January 2023, separated into 2010–2020 (historical) and 2021–2023 (recent) periods. We reviewed medical records of patients with bacteremia and recorded demographics, symptoms, housing, sexual orientation, travel, substance use, coexisting conditions, hospitalization, antimicrobial susceptibility testing (AST), treatment, and mortality.

We processed positive blood cultures detected by BacT/AlertT system (bioMérieux, https://www.biomerieux.com) using VitekMS+ (bioMérieux) or FilmArray BCID (BioFire Diagnostics; https://www.biofiredx.com) with established microbiology protocols and identified pathogens. We identified Shigella in feces using benchtop biochemical methods, Vitek2 ID (bioMérieux) and Polyvalent Agglutination Sera (Remel, http://www.remel.com). If we suspected Shigella in blood samples, we used Vitek2 ID and polyvalent serology. No changes in laboratory testing protocols occurred during 2010–2023. We performed AST for ampicillin, trimethoprim/sulfamethoxazole, ciprofloxacin, ceftriaxone, and azithromycin according to Clinical and Laboratory Standards Institute M100 standards (https://clsi.org/standards/products/microbiology/documents/m100). We performed whole-genome sequencing (WGS) on isolates incubated in Mueller-Hinton broth, extracted on MagNA Pure 24 (Roche; https://diagnostics.roche.com), and processed on GridION R10.4 flowcells (https://nanoporetech.com). We basecalled runs with Guppy version 6.3.9 (https://github.com/nanoporetech/rerio) and uploaded to BugSeq (https://bugseq.com) for automated analysis.

Figure

Figure. Epidemic curve of Shigella sonnei cases from feces and blood, Vancouver, British Columbia, Canada, 2010–January 2023.

We identified 11 cases of S. sonnei bacteremia that occurred within the historical (n = 2) or recent (n = 9) periods during 2010–2023. We also observed a recent increase in fecal isolates with S. sonnei (Figure). Differences in the proportion of bacteremic among all shigellosis cases occurring during the recent compared with the historical period were not statistically significant (7.7% vs. 2.9%; p = 0.21 by Fisher exact test). Among recent cases, 89% (8/9) of patients were male (median 45 years of age, interquartile range 35–54 years) (Table). All but 1 were underhoused or experiencing homelessness and had polysubstance use disorder. Most (89%) inhabited Vancouver’s downtown eastside, the neighborhood with the highest density of Vancouver’s urban poor population. Unlike the historical case-patients, none were MSM or had travel histories. Most (89%) were not severely immunocompromised. Case-patient A had multiple myeloma but stable housing. Five (55%) patients presented with sepsis and 6 (67%) were hospitalized; the remaining patients declined recommended hospital admission.

Among the 2021–2023 cases, AST profiles were identical except in case A. The isolates were resistant to ampicillin, trimethoprim/sulfamethoxazole, ciprofloxacin, and azithromycin and susceptible to ceftriaxone. An isolate from case-patient A displayed ceftriaxone resistance and azithromycin susceptibility. Shigella was initially misidentified as E. coli in all 9 cases (8 by VitekMS+, 1 by FilmArray BCID). All isolates were indole negative, non–lactose fermenting (NLF) colonies, subsequently identified correctly as S. sonnei by Vitek2 ID and confirmed by polyvalent serology.

We performed WGS on all 2021–2023 S. sonnei isolates from blood samples. Using a new genotypic framework (8), we identified 8 of the 2021–2023 isolates as 3.6.1.1.2 (CipR.MSM5); the isolate from case-patient A genotyped as 3.6.3 (Central Asia III). The strain from the 2016 case genotyped as 3.7.18 (Global III); the 2013 strain did not undergo WGS. We identified mutations gyrA S83L, gyrA D87G, and parC S80I encoding ciprofloxacin resistance and plasmid AA336-borne mphA and ermB encoding azithromycin resistance in all isolates from persons experiencing homelessness (PEH). The 3.6.3 strain carried qnrS1, gyrA D87Y, gyrA S83L and parC S80I, which confers ciprofloxacin resistance, and bla_CTX-M-15, which confers ceftriaxone resistance.

The recent increase in S. sonnei bacteremia might reflect the overall increase in shigellosis, including S. sonnei isolated from feces. The increased proportion of bacteremia cases in the past 2 years compared with the 11-year historical period was not statistically significant. However, the 7.7% prevalence of bacteremia is still very high compared with a range of rates, 0.4%–7.3%, reported in the literature (9). Historically, shigellosis has occurred in British Columbia predominantly as a sexually transmitted enteric infection among MSM (10). In our study, invasive shigellosis among PEH was probably transmitted fecal-orally through contaminated environment and hands. An outbreak among PEH in Oregon was similarly believed to have resulted from inadequate access to hygiene and sanitation (11). WHO warns of outbreak risk from S. sonnei being introduced into areas with suboptimal water, sanitation, and hygiene standards (12). Because patient immunosuppression fails to explain increased bacteremia, other factors, such as drug use, malnutrition, and high inoculum dose, should be considered.

A study in Seattle, Washington, USA described a contemporary increase in MDR S. sonnei cases among PEH (13). Although the report did not comment on bacteremia, it highlighted the circulation of extensively-drug resistant S. sonnei carrying the extended-spectrum β-lactamase CTX-M-27. Although extensively-drug resistant S. sonnei was rare in our review, increased hospitalization and the need for parenteral therapy because of MDR S. sonnei bacteremia still substantially affected the healthcare system. One third of our patients did not complete treatment, potentially leading to ongoing transmission and illness. In addition to finding and treating cases, essential community control measures include working with housing providers to promote handwashing and sanitation practices, providing advice on recognizing and controlling infectious diarrhea, and developing pathways to reengage those who refuse hospitalization to complete parenteral antimicrobial treatment as outpatients.

Isolates from 8/9 PEH were genotype 3.6.1.1.2 (CipR.MSM5/BAPS3), epidemiologically distinct from the single isolate typed as 3.6.3 (Central Asia III). The isolate from the 2016 case had a different genotype, 3.7.18, matching that from an outbreak reported in California, USA (14). Despite its geographic proximity to Vancouver, the recent S. sonnei outbreak among PEH in Seattle involved yet another different genotype, 3.7.29.1.4.1 (global III VN2.KH1.Aus) (14). The 3.6.1.1.2 strain predominant in our study has previously been described in Australia, England, and the United States (8).

Our study included analysis of data only from patients seeking treatment at a hospital where blood cultures were collected, possibly representing the sickest cohort of patients, leading to underestimation of actual infections. Not all cases with S. sonnei isolated from feces have yet been analyzed to determine wider epidemiologic and antimicrobial resistance trends. Further studies are needed to elucidate whether the high rate of bacteremia reflects increased virulence of this strain, higher inoculum size, or host determinants.

Because current laboratory methods can misidentify Shigella as E. coli in bloodstream infections, laboratories must scrutinize diagnoses of E. coli bacteremia in high-risk patients with infectious diarrhea and sepsis. Clinicians and public health officials should be made aware of MDR S. sonnei bacteremia as a cause of increased illness among PEH and the need for parenteral therapy in the event of resistance to first- and second-line antimicrobial agents.