PREreview of Frequency of antimicrobial resistance among fecal bacteria pathogens in an informal settlement in Nairobi, Kenya, 2018 to 2020

This review is the result of a virtual, collaborative live discussion hosted by one PREreview Champion as part of the 2025 PREreview Champions Program. The session was joined by eight participants, including one facilitator who is a member of the 2025 cohort. Participants contributed actively to the discussion and later collaborated asynchronously over two weeks to finalize this report using notes from the session. We are grateful for everyone’s thoughtful contributions to this effort.

Summary of the study

Research question

What is the frequency, antimicrobial resistance profile, and presence of resistance genes among Salmonella, Shigella, Vibrio, and Campylobacter isolated from patients with diarrhea in an informal settlement in Nairobi, Kenya, between November 2018 and January 2020?

Research approach

This research employed a cross-sectional study in the Kawangware informal settlement, recruiting patients with diarrhea from November 2018 to January 2020. Stool samples were collected and analysed for bacterial pathogens using culture methods and the BD Phoenix M50 system. Antimicrobial susceptibility testing was conducted on isolates, with multidrug resistance identified based on resistance to three or more antibiotic classes. Resistant isolates were further tested by PCR to detect specific resistance genes. Demographic and risk factor data were gathered through questionnaires, and results were summarized using descriptive statistics. Ethical approvals were secured, and findings were shared with clinicians to support patient care.

Research finding(s)

Among 1,209 patients with diarrhea in Kawangware, Shigella flexneri was the most commonly isolated pathogen, followed by Salmonella Typhi. Vibrio and Campylobacter were not detected. Several uncommon commensal bacteria were also identified. High antimicrobial resistance was observed, particularly to ampicillin (83%) and tetracycline (50%), while all isolates remained susceptible to ciprofloxacin. Multidrug resistance was noted among Salmonella and Shigella isolates. PCR analysis of 60 resistant isolates detected resistance genes such as tet(A/B), sul1, sul2, aac, and ampC, which generally matched the resistance patterns observed; however, blaTEM was less common. Some isolates carried resistance genes without corresponding phenotypic resistance.

Main strengths and weaknesses

The study addressed antibiotic resistance in a high-risk population with limited access to healthcare. The study design was appropriate, and the combined use of phenotypic and molecular methods enhanced the reliability of the findings. The BD Phoenix M50 platform facilitated accurate and efficient bacterial identification and antimicrobial susceptibility testing. Ethical approval was duly obtained, and the authors acknowledged certain limitations, including the absence of E. coli pathotyping and the focus on a single geographic location. The use of multiple selective media increased the diversity of pathogen detection. Additionally, the authors responsibly communicated findings back to patients. The study effectively linked results to the broader context through strong integration of molecular and phenotypic data, complemented by relevant regional comparisons.

Major issues  

• Inconsistency Between Abstract, Results Section, and Table 1 Data: The results reported in the abstract conflict with those presented in the Results section and Table 1. For instance, the abstract lists Shigella flexneri at 58.3% and Salmonella enterica serovar Typhi at 8.3%, while the Results section reports Shigella flexneri at 2.7%, alongside other Shigella species, and Salmonella Typhi at 0.4%. This inconsistency compromises the credibility and clarity of the findings and must be resolved to ensure accuracy and coherence throughout the manuscript.

• Methodological Inconsistency in Phenotypic vs Genotypic Testing: There is a discrepancy between the methodology and results regarding genotypic characterization. The Methods state that only phenotypically resistant isolates underwent PCR screening; however, the Results report that some isolates (e.g., S. boydii and S. flexneri) carried ampicillin resistance genes despite phenotypic susceptibility. Authors should clarify whether phenotypically susceptible isolates were also screened by PCR or explain this inconsistency to maintain methodological rigor and credibility.

• Logical Flow, Paragraph Structure, and Content Justification in the Introduction: The Introduction contains all the essential background elements; however, the current paragraph arrangement weakens the logical progression of ideas. Reorganizing the paragraphs to improve coherence would ensure that each section builds effectively on the previous one, enhancing clarity and strengthening the narrative flow from the global antimicrobial resistance (AMR) context to the local focus of the study. Additionally, the Introduction would benefit from incorporating specific AMR statistics from national or global sources (e.g., WHO, CDC) to better illustrate the scale and urgency of the problem. Including a brief reference to recent AMR prevalence trends in Sub-Saharan Africa or Kenya, if available, would help contextualize the study and justify its relevance. Furthermore, the rationale for focusing on fecal bacteria or Enterobacteriaceae is not explicitly stated; briefly explaining their selection (whether due to their prevalence in diarrheal diseases, resistance profiles, or public health importance in informal settlements) would further strengthen the Introduction.

• Sample Transport, Culture Methods, and Identification Details: The phrase “samples were shipped immediately” is vague and requires a more precise description of transport logistics, including the estimated time between sample collection and processing, as well as confirmation of whether cold-chain conditions were maintained. Additionally, the culture conditions for Campylobacter are inaccurately described; this bacterium requires incubation at 42°C under microaerophilic conditions for 48 hours, and failure to meet these criteria could explain the absence of Campylobacter-like colonies (see DOI: 10.1002/9780471729259.mc08a01s10). Furthermore, the observation “No Campylobacter-like or Vibrio-like colonies were observed” should be moved from the Methods section to the Results section, as observations belong there. Finally, the manuscript should clarify which colony morphology characteristics were used for presumptive identification and explicitly state the plating method employed for primary isolation, whether streak, pour, spread plating, or filtration.

• Bacterial Suspension Preparation and Antibiotic Selection: The method used to prepare bacterial suspensions from overnight cultures is insufficiently detailed. The authors should specify the type of growth media used, incubation conditions, and how the suspensions were standardized (such as using OD600 measurements or McFarland standards) prior to downstream processes. Additionally, while CLSI guidelines are referenced, the rationale for narrowing the antibiotic panel from 30 to the final 11 agents is unclear. The manuscript should clarify whether this selection was based on clinical relevance, local prescription patterns, or resistance trends to ensure transparency and relevance of the antimicrobial susceptibility testing.

• Participant Criteria and Ethical Clarity issues: While the inclusion criteria for participants are described, the manuscript lacks explicit mention of any exclusion criteria, which is important to clarify as it directly impacts the generalizability of the findings. Additionally, the total number of participants planned or enrolled is not reported, and a clear explanation of how the sample size was determined (whether through statistical calculation or convenience sampling) is necessary to assess the robustness of the study design. Furthermore, the timeline for obtaining informed consent is ambiguous; for ethical clarity, it should be clearly stated that written consent was obtained prior to participant enrollment and before any data or sample collection took place.

• Incomplete Molecular Detection Protocol: The molecular detection protocol lacks critical methodological details, which limits reproducibility and clarity. The manuscript should specify the type of PCR performed (e.g., conventional, multiplex, or qPCR), and include a clearly labeled table outlining primer sequences, target genes, amplicon sizes, cycling conditions, and relevant references. Details of the PCR reaction mix, including final concentrations and volumes, must also be provided. Additionally, the authors should explain the post-PCR analysis procedures, such as the gel concentration, voltage used, and visualization method (e.g., ethidium bromide staining and UV transillumination). The DNA extraction method (whether a commercial kit or manual approach) should be described, along with any quality control steps like NanoDrop quantification or gel-based purity checks. DNA quantification and purity assessment prior to PCR should also be included. To improve readability and logical flow, the entire section should be reorganized chronologically, starting with DNA extraction, then quantification, PCR setup, amplification, and finally analysis.

• Regional Comparisons and Discussion Expansion: The comparison to other informal settlements is currently underdeveloped and abrupt, as it merely lists regions without discussing how the findings align with or differ from those in similar settings. Providing a brief summary of similarities or discrepancies would offer valuable context and enhance the geographical relevance of the study. Additionally, the Discussion should elaborate on the practical implications of the observed resistance patterns, specifically how these findings could inform local treatment guidelines, support antibiotic stewardship efforts, or shape public health interventions. Expanding on these aspects would significantly improve the study’s translational value and real-world applicability.

• Study Limitations and Future Directions: While ethical and logistical limitations reflecting field challenges are acknowledged, this section would be strengthened by explicitly suggesting future research directions. For example, incorporating E. coli pathotyping to fill current surveillance gaps would be valuable. Additionally, the impact of limited sample recruitment on generalizability should be more clearly specified.

Minor comments 

• Clarity of the Abstract: The abstract provides useful data; however, several clarifications could enhance its clarity and clinical relevance. The exact study location (Kawangware) should be specified, and the study design clearly stated; this was a cross sectional study. The criteria used to classify multidrug resistance (MDR) should be defined, and the percentage of MDR isolates reported. If applicable, other major MDR pathogens found or screened in the study area may be mentioned. It is important to note that all isolates were susceptible to ciprofloxacin, given its clinical significance. Including raw counts alongside percentages for bacterial isolates (e.g., Shigella flexneri (58.3%, N=35)) would further strengthen the abstract’s clarity and scientific value.

• Keywords and Abbreviations: The manuscript lacks keywords and a list of abbreviations. Adding 4 to 6 keywords and an abbreviations list would improve focus, accessibility, and reader understanding.                          

• Lack of Quantification and Clarity in Data Reporting: The manuscript does not quantify the prevalence of detected resistance genes, missing percentages or frequencies that would aid reader comprehension, and lacks graphical representation of this data. Furthermore, the reported percentages for bacterial isolates are unclear regarding their reference population (whether they relate to total patients or only positive samples) and absolute counts are not provided, reducing the completeness and clarity of the data presentation.

• Visual Data Presentation and Table Citation: Results section lacks visual aids such as images of bacterial colonies on selective media and representative PCR product gels, which would improve clarity and reader engagement. Although demographic data are well-organized, they would benefit from summary tables and graphical representation to enhance readability. Additionally, Table 1 is not referenced in the text, which needs correction to ensure proper integration of data. The manuscript also lacks a comprehensive table summarizing colony morphology characteristics for each bacterial isolate. Including such a table would aid readers in understanding isolates identification.

• Missing Questionnaire Details, Unclear Risk Factors, and Inconsistent Nomenclature: The manuscript lacks details on questionnaire administration, including whether it was face-to-face or self-administered, the language used, and whether it occurred before or after participant enrollment, information crucial for assessing potential bias and participant comprehension. It should also clearly specify the key risk factors investigated through the questionnaire or clinical records to clarify the study’s scope and epidemiological relevance. Additionally, scientific nomenclature is inconsistently used (e.g., “Salmonella enterica serovar Typhi” versus “Salmonella Typhi”) and should be standardized throughout for clarity and accuracy.

• Terminology, Consistency, and Tense Usage: The manuscript requires improvements in terminology and consistency. The term “correlation” should be replaced with “linkage” or “alignment” unless supported by statistical analysis. Use “fecal” consistently instead of “enteric”, and prefer “resistance genes” over “resistance determinants” for clarity. Additionally, the phrase “enhanced detection” should be revised to “enhanced surveillance” to better reflect the study’s objectives. The final paragraph of the Introduction should consistently use past tense, in line with standard scientific writing conventions. Furthermore, clarify key concepts and terminology throughout the manuscript. For example, specify the type of equipment that was lacking (e.g., diagnostic machines, PCR kits, refrigeration), explain “first line antibiotics” in simple terms (such as “the most commonly used antibiotics for initial treatment”), and elaborate on the AMR threat, clearly stating whether it poses a public health, economic, or clinical treatment risk.

• Paragraph Structure and Focus: Some paragraphs in the discussions are dense and attempt to cover multiple themes at once (e.g., prevalence, resistance frequency, sanitation, and antimicrobial stewardship). This reduces readability and clarity. I suggest splitting such paragraphs into two distinct sections: one focused on comparing resistance prevalence with prior regional studies, and another exploring possible explanations such as hygiene practices and water quality. This will improve flow and reader comprehension.”

• Interpretation in Results Section: Interpretative statements, such as “The resistance to this antibiotic was likely not expressed phenotypically or was weakly expressed,” should be avoided in the Results section and relocated to the Discussion, where analysis and explanations are appropriate. Similarly, observations interpreting the detection of commensal flora should also be moved to the Discussion to maintain a clear distinction between objective findings and their interpretation. The sentence ‘These organisms most likely represented commensal flora and not true pathogens and their identification demonstrates the potential of the BD Phoenix to identify fewer common organisms’ presents an interpretation and should be relocated from the Results section to the Discussion section.

• Paragraph and Sentence Length: Several paragraphs, particularly in the Introduction and Discussion sections, are too lengthy and contain multiple ideas, affecting clarity and flow. Additionally, the study duration is mentioned repeatedly, especially in the patient recruitment section, without adding new information. There are also redundant descriptions of patient recruitment and stool collection methods across different subsections (specifically, the sample collection process is explained both in the study design and eligibility sections). These issues collectively reduce readability and coherence, and they should be addressed by splitting long paragraphs and sentences, eliminating unnecessary repetition, and consolidating overlapping content.

• Study Setting Description: The description of the informal settlement includes highly specific details (e.g., housing materials such as corrugated metal sheets and polywood) that are not directly linked to the study outcomes. Unless these environmental factors relate to infection risk or microbial transmission, this section should be condensed to focus on variables relevant to public health or disease epidemiology.

• Clarify conclusion scope: The conclusion effectively summarizes the findings and clearly states that they are based on a localized study setting and may not be generalizable nationwide. Adding a forward-looking recommendation, such as expanding the study to multiple regions or incorporating broader surveillance, would improve the relevance and impact of the findings for future research.

Comments on availability of data.

• The manuscript did not include a data availability statement indicating where the underlying data can be accessed, whether in the appendix, an external repository, or fully embedded within the manuscript. Including this information is essential to enhance the transparency, reproducibility, and credibility of the study. The authors are encouraged to clearly state the availability and location of all relevant datasets used in the analysis.

Concluding remarks

We thank the authors of the preprint for posting their work openly for feedback. We also thank all participants of the Live Review for their time and for engaging in the lively discussion that generated this review.

Competing interests

The authors declare that they have no competing interests.