Multiple introductions of multidrug-resistant typhoid associated with acute infection and asymptomatic carriage, Kenya

  1. Samuel Kariuki
  2. Zoe A Dyson
  3. Cecilia Mbae
  4. Ronald Ngetich
  5. Susan M Kavai
  6. Celestine Wairimu
  7. Stephen Anyona
  8. Naomi Gitau
  9. Robert Sanaya Onsare
  10. Beatrice Ongandi
  11. Sebastian Duchene
  12. Mohamed Ali
  13. John David Clemens
  14. Kathryn E Holt
  15. Gordon Dougan

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Understanding the dynamics of infection and carriage of typhoid in endemic settings is critical to finding solutions to prevention and control.

Methods:

In a 3-year case-control study, we investigated typhoid among children aged <16 years (4670 febrile cases and 8549 age matched controls) living in an informal settlement, Nairobi, Kenya.

Results:

148 S . Typhi isolates from cases and 95 from controls (stool culture) were identified; a carriage frequency of 1 %. Whole-genome sequencing showed 97% of cases and 88% of controls were genotype 4.3.1 (Haplotype 58), with the majority of each (76% and 88%) being multidrug-resistant strains in three sublineages of the H58 genotype (East Africa 1 (EA1), EA2, and EA3), with sequences from cases and carriers intermingled.

Conclusions:

The high rate of multidrug-resistant H58 S . Typhi, and the close phylogenetic relationships between cases and controls, provides evidence for the role of carriers as a reservoir for the community spread of typhoid in this setting.

Funding:

National Institutes of Health (R01AI099525); Wellcome Trust (106158/Z/14/Z); European Commission (TyphiNET No 845681); National Institute for Health Research (NIHR); Bill and Melinda Gates Foundation (OPP1175797).

Article activity feed

  1. Version published to 10.7554/elife.67852 on eLife
  2. eLife

    Evaluation Summary:

    This study contributes a significant advance to the field in terms of detailed genomic epidemiology of the introductions of drug resistant typhoid into Kenya, and Africa in general. The findings highlight the role of asymptomatic carriage in the spread, drug-resistance mechanisms and the emergence of typhoid strains in Kenya. The authors also contribute a small number of sequences from both carriage and acute disease, most sequenced cases are associated with acute disease, making this deposit of publicly available carriage sequences extremely valuable. This work will be of interest to a broad readership, including but not limited to clinicians, biologists, and public health experts.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #2 and Reviewer #3 agreed to share their names with the authors.)

  3. eLife

    Reviewer #1 (Public Review):

    Kariuki et al. report a clinical, microbiological, and genomic epidemiology analysis of multidrug-resistant typhoid strains sampled from a large case-control study of children living in an informal settlement in Kenya. Using whole-genome sequencing, the authors demonstrate that the frequency and distribution of S. Typhi genotypes among cases and controls is similar, highlighting the importance of asymptomatic carriage as a reservoir for transmission. The authors also describe the presence of drug-resistant genes and their genomic localisation. Further phylodynamic and geographic analysis defined clades associated with the Kenyan S. Typhi isolates in a global context to determine potential importation and time of emergence.

    The participant inclusion and exclusion criteria are sound, and the microbiological and bioinformatics analysis methods are robust and well-described.

    Specific comments:

    - Although this is mentioned in the referenced papers, it would be great to briefly mention the microbiological methods used for serotyping in the microbiology laboratory.

    - It would be great to mention if, in general, there are any discrepancies in the isolation frequency of S. Typhi isolation between stool samples and rectal swabs. If the isolation frequency differs between the sample types, then biased sampling favouring one of these samples could affect the estimated carriage frequency.

    - Details on how other bacterial species were identified are not included. It would be great to mention whether Kraken or other tools were used.

    - Considering that some isolates were mistyped in the laboratory, it would be great to include some discussion on this. It seems that the inclusion criteria required all the sequencing of S. Typhi and not all other Salmonella serotypes (see lines 132 & 147). I wondered whether broadening the criteria for selecting the isolates to undergo genome sequencing would somewhat change the results, specifically the carriage frequency and distribution of genotypes (i.e., identifying more S. Typhi that would be mistyped otherwise).

    - It's not clear why the IS1 insertion sequences were identified in the genomes. S. Typhi probably contain other IS elements, therefore, a brief explanation of why on the IS1 elements were identified should be provided.

    - The link between the phylogenetic branch lengths and tree tips suggesting a more prolonged duration from acquisition to sampling in the clinic is not clear. Since the isolates were sampled cross-sectionally, it's not clear how this information was inferred in the absence of data on within-host diversity of the isolates.

    - Some information is hidden in columns of Table S1. It would be great to present this data as a separate Excel spreadsheet file.

  4. eLife

    Reviewer #2 (Public Review):

    The authors present a detailed and well-written investigation of the genomic epidemiology of MDR Typhoid in Kenya from both carriage and acute case sequences. The authors found phylogenetically interspersed carriage and acute cases that indicate a reservoir role for asymptomatic carriers. The Phylogenetic results also indicated that there wasn't a point-source short-term outbreak of MDR typhoid but rather 3 well-established east african lineages that are associated with continuous community transmission and presence in both carriage and acute cases for a long period of time.

    The dataset is small but the authors account for this and include publicly available previously sequenced isolates for phylogenetic context. The phylogenetic conclusions are well supported by the presented data and appropriate methodology has been applied.

    Overall, the paper is a useful addition to our knowledge on Typhoid and is of considerable interest to the field.

  5. eLife

    Reviewer #3 (Public Review):

    The authors set out to describe the genomic epidemiology of Salmonella Typhi causing diarrhea among children in an informal Kenyan setting between 2013 and 2016. A unique strength of this paper is the they also analyzed contemporaneous carriage isolates from age matched controls in the same locality. Based on the phylogenetic overlap between carriage and invasive isolates, they deduce a potential role for carriage driving disease transmission community. This finding has important implications, since it places the onus on health officials to devise control strategies that encompass surveillance and potentially interventions in healthy carriers. It also signals the need to study the role of alternative reservoirs of S. typhi like the environment and animals.
    The authors went further and probed the historical context of the isolates. They leveraged epidemiological data dating back to the 1990's and genome sequences from previous studies in Kenya to determined that most of the isolates causing disease were the offspring of multidrug resistant lineages that replaced the drug susceptible lineages in the 1990's. Their dating places the introduction of these lineages around 1990, which correlated with the first reports of these lineages in Kenya. This historical context is important in understanding the evolution of drug resistance and how the genomic landscape has changed with changing antibiotic usage patterns. This dataset adds context on the evolutionary trajectory of S. typhi in Kenya, providing a broader baseline for future studies to track the emergence of novel lineages as well as the evolution of existing lineages. Future studies will have a wider temporal spread and, with that, likely improved confidence in inferences from phylogenetic dating.

    Finally, the authors studied patterns of mutations to identify genetic signatures that differentiated carriage from invasive disease isolates. They showed that a higher proportion of non-synonymous mutations likely meant positive selection and a longer duration of colonization in carriage than disease. While they identify important gene classes that had a higher proportion of non-synonymous mutations in carriers, there is little discussion on the biological implication of their findings. It is also unclear whether these findings may be biased by clonal inheritance i.e., if most a given mutation occurs on a branch or clade that is dominated by carriage isolates, the signal may be due to phylogenetic placement and not necessarily a biological adaption process. One way to overcome this would be a genome wide association study that looks at mutations, unitigs as well as whole genes, and takes phylogenetic placement into account. However, the authors rightly acknowledged that their study yielded a low number of genomes, which may not be powered for such analyses.
    A major concern was the high proportion of isolates that were initially designated as S typhi but confirmed by genomics to be another species (34.6%). Although this could be due to low specificity of the serotyping assays or randomly picking different colonies in a mixed infection, it can also indicate sample mix ups in the dataset. It is important for the authors to clarify what steps were taken to rule out sample mix-ups and to share details on how reliable their initial serotyping was. This is compounded by the low number of genomes, which the authors acknowledge, as well as the geographic localization, which raises questions about how generalizable these findings are to the wider population in Kenya.

  6. Version published to 10.1101/2021.03.10.434750v1 on bioRxiv