Caitlin Farley

Background/Objectives: Infections caused by Legionella species are primarily associated with Legionella pneumophila, but non-pneumophila species are increasingly implicated in human disease. Despite this, antimicrobial susceptibility testing (AST) data for non-pneumophila species remain scarce, and standardised testing protocols or resistance thresholds have not been established. This study aimed to address this gap by evaluating and comparing AST performance for non-pneumophila Legionella species relative to L. pneumophila using three methodologies. Methods: AST was conducted on 89 Legionella isolates using LASARUS agar dilution, buffered yeast extract broth microdilution (BYE-BMD), and BCYE-α agar dilution, against ampicillin, azithromycin, chloramphenicol, doxycycline, levofloxacin, and rifampicin. Growth performance and minimum inhibitory concentrations (MICs) were assessed after a 96 h incubation. Results: MIC profiles were obtained using LASARUS and BYE-BMD for 53.9% and 93.3% of isolates, respectively. While L. pneumophila reached sufficient turbidity in BYE-BMD after a 48 h incubation, non-pneumophila species required an extended incubation (72–96 h). Non-pneumophila species displayed broader MIC ranges against azithromycin (0.016–1 mg/L) and levofloxacin (0.016–0.25 mg/L), but a narrower rifampicin range (≤0.0005–0.032 mg/L) relative to L. pneumophila. L. longbeachae exhibited a higher MIC₅₀ for rifampicin despite overlapping susceptibility ranges across all species (0.001–0.016 mg/L). Conclusions: This study demonstrates species-specific differences in Legionella susceptibility and highlights the limitations in extrapolating L. pneumophila-based AST data. Azithromycin MICs in non-pneumophila species exceeded those of L. pneumophila, raising clinical concern. While BYE-BMD was the most effective method for MIC determination, three species required BCYE-α due to poor growth. These findings support developing standardised, species-specific AST protocols and thresholds amid rising macrolide resistance and the increasing detection of non-pneumophila infections.

Background: Antimicrobial resistance (AMR) in Legionella pneumophila is emerging as a concern, particularly with resistance to macrolides and fluoroquinolones. Although clinically significant resistance in Legionella pneumophila remains uncommon, systematic genomic surveillance using whole-genome sequencing (WGS) is needed to anticipate treatment failure as metagenomic diagnostics move toward routine use. Objectives: We assessed the UK Health Security Agency AMR pipeline for predicting resistance in L. pneumophila by analysing 522 L. pneumophila isolates from England and Wales (2020–2023) together with nine database sequences that carry confirmed 23S rRNA mutations conferring high-level azithromycin resistance. The objective of the present study was to examine the presence of antimicrobial resistance genes (ARGs) in L. pneumophila isolates and to determine whether they exhibited phenotypic resistance through minimum inhibitory concentration (MIC) testing. Methods: Serogroups (sgs) were determined using an in-house qPCR assay, and L. pneumophila non-sg1 isolates were serogrouped using the Dresden monoclonal antibody (mAb) typing method. Sequence types were determined using the standard sequence-based typing method by Sanger sequencing. WGS reads were screened against standard AMR databases to identify resistance genes and resistance-mediating mutations. Agar dilution measured MICs for azithromycin, erythromycin, ampicillin, levofloxacin, tetracycline and spectinomycin in isolates possessing the bla_OXA-29, lpeAB or aph(9)-Ia gene. Results: AMR screening detected lpeAB, two allelic β-lactamase variants (bla_OXA-29 and bla_LoxA) and aph(9)-Ia in 165 of the 522 L. pneumophila isolates, while all high-azithromycin MIC reference sequences contained the expected 23S mutation. Only lpeAB was associated with a significant twofold elevation in macrolide MICs. Neither β-lactamase variant increased ampicillin MICs, and aph(9)-Ia carriage did not correlate with higher spectinomycin MICs. Conclusions: Advanced genomic analytics can now deliver timely therapeutic guidance, yet database-flagged genes may not translate into phenotypic resistance. Continuous pairing of curated mutation catalogues with confirmatory testing remains essential for distinguishing clinically actionable determinants such as 23S mutations and lpeAB from silent markers like bla_OXA-29 and aph (9)-Ia.

Background: Child survival rates have improved globally, but neonatal mortality due to infections, such as group B Streptococcus (GBS), remains a significant concern. The global burden of GBS-related morbidity and mortality is substantial. However, data from low and middle-income countries are lacking. Vaccination during pregnancy could be a feasible strategy to address GBS-related disease burden. Methods: We assessed maternal rectovaginal GBS colonization and neonatal disease rates in a prospective cohort of 6062 women–infant pairs. Surveillance for invasive infant disease occurred in parallel at 2 Kampala hospital sites. In a nested case-control study, we identified infants <90 days of age with invasive GBS disease (iGBS) (n = 24) and healthy infants born to mothers colonized with GBS (n = 72). We measured serotype-specific anticapsular immunoglobulin G (IgG) in cord blood/infant sera using a validated multiplex Luminex assay. Results: We found a high incidence of iGBS (1.0 per 1000 live births) within the first 90 days of life across the surveillance sites, associated with a high case fatality rate (18.2%). Maternal GBS colonization prevalence was consistent with other studies in the region (14.7% [95% confidence interval, 13.7%–15.6%]). IgG geometric mean concentrations were lower in cases than controls for serotypes Ia (0.005 vs 0.12 µg/mL; P = .05) and III (0.011 vs 0.036 µg/mL; P = .07) and in an aggregate analysis of all serotypes (0.014 vs 0.05 µg/mL; P = .02). Conclusions: We found that GBS is an important cause of neonatal and young infant disease in Uganda and confirmed that maternally derived antibodies were lower in early-onset GBS cases than in healthy exposed controls.

Background: Streptococcus agalactiae (Group B Streptococcus, GBS) is a leading cause of neonatal sepsis in high-income countries. While intrapartum antibiotic screening reduces this risk, increasing resistance to macrolides and lincosamides in Europe since the 1990s has limited therapeutic options for penicillin-allergic patients. Reports of reduced beta-lactam susceptibility in GBS further emphasise the need for robust antimicrobial resistance (AMR) surveillance. However, broth microdilution (BMD) methods are unsuitable for large-scale antimicrobial susceptibility testing (AST). Objective: To demonstrate that agar-dilution AST provides equivalent results to broth dilution methods, with superior capacity for high-throughput screening. Methods: Agar-dilution and microdilution AST methods were compared using a panel of 24 characterised susceptible and resistant GBS strains for benzylpenicillin, chloramphenicol, clindamycin, erythromycin, gentamicin, levofloxacin, tetracycline, and vancomycin. Minimum inhibitory concentration (MIC) agreements were evaluated, and resistance profile correlations were assessed using Cohen’s kappa values. Results: Agar-dilution demonstrated >90% agreement with BMD MIC for most antimicrobials, except vancomycin (87.5%), erythromycin (83.33%), and tetracycline (52.78%). Cohen’s kappa values indicated strong agreement (0.88–1.00) for resistance determination. Agar-dilution avoided “trailing growth” issues associated with BMD and facilitated easier detection of non-GBS contaminants. Conclusions: Agar-dilution is a valid method for high-throughput AMR surveillance of retrospective cohorts (96 isolates per plate) and is critical for identifying emerging GBS resistance trends and informing therapeutic guidelines. However, due to the large number of plates required per antimicrobial, it is impractical for routine clinical diagnostics.

Background: Streptococcus agalactiae (Group B Streptococcus, GBS) is a leading cause of neonatal sepsis in high-income countries. While intrapartum antibiotic screening reduces this risk, increasing resistance to macrolides and lincosamides in Europe since the 1990s has limited therapeutic options for penicillin-allergic patients. Reports of reduced beta-lactam susceptibility in GBS further em-phasize the need for robust antimicrobial resistance (AMR) surveillance. However, broth micro-dilution (BMD) methods are unsuitable for large-scale antimicrobial susceptibility testing (AST). Objective: To demonstrate that agar-dilution AST provides equivalent results to broth dilution methods, with superior capacity for high-throughput screening. Methods: Agar-dilution and mi-crodilution AST methods were compared using a panel of 24 characterized susceptible and resistant GBS strains for benzylpenicillin, chloramphenicol, clindamycin, erythromycin, gentamicin, levofloxacin, tetracycline, and vancomycin. Minimum inhibitory concentration (MIC) agreements were evaluated, and resistance profile correlations were assessed using Cohen’s kappa values. Results: Agar-dilution demonstrated >90% agreement with BMD MIC for most antimicrobials, except vancomycin (87.5%), erythromycin (83.33%), and tetracycline (52.78%). Cohen’s kappa values indicated strong agreement (0.88–1.00) for resistance determination. Agar-dilution avoided “trailing growth” issues associated with BMD and facilitated easier detection of non-GBS con-taminants. Conclusions: Agar-dilution is a valid method for high-throughput AMR surveillance of retrospective cohorts (96 isolates per plate) and is critical for identifying emerging GBS resistance trends and informing therapeutic guidelines. However, due to the large number of plates required per antimicrobial, it is impractical for routine clinical diagnostics.

Background: Every year an estimated 2–3 million babies are stillborn, with a high burden in Africa. Infection is an important driver of stillbirth. There is a lack of data on the bacterial causes of stillbirth in Uganda, contributing to a lack of interventions such as effective prophylaxis and development of maternal vaccine options against the most implicated pathogens. Methods: The PROGRESS study was an observational cohort study undertaken in Kampala, Uganda, between November 2018 and April 2021. If a woman delivered a stillborn baby, consent was sought for the collection of a heart-blood aspirate. One to three mL of blood was collected and sent for culture using the BD Bactec blood culture system. Organism identification was performed using biochemical testing and matrix-assisted laser desorption/ionization–time of flight mass spectrometry. Susceptibilities to appropriate panels of antimicrobials were determined by agar dilution. Results: Kawempe Hospital registered 34 517 births in the study period, of which 1717 (5.0%) were stillbirths. A total of 581 (33.8%) were recruited into the study, and heart blood aspirates were performed on 569 (97.9%). Blood samples were sufficient for analysis of 476, with a total of 108 positive cultures (22.7% of sampled stillbirths). Fifty-nine of 108 blood cultures contained organisms that were considered potential pathogens, giving a pathogen positivity rate of 12.4%. Common pathogens included Enterococcus spp. (n = 14), Escherichia coli (n = 13), viridans streptococci (n = 18), Klebsiella pneumoniae (n = 6), and group B Streptococcus (n = 5). Gram-negative organisms were frequently resistant to commonly used first-line antimicrobials. Conclusions: The high proportion of stillbirths caused by likely pathogenic bacteria in Uganda highlights the potential for prevention with prophylaxis and stresses the need for further investment in this area.

Background: Epidemiological evidence about the etiology and antimicrobial resistance of neonatal infections remains limited in low-resource settings. We aimed to describe the etiology of neonatal infections in a prospective observational cohort study conducted at two hospital sites in Kampala, Uganda. Methods: Babies admitted to either unit with risk factors or signs of sepsis, pneumonia, or meningitis had a blood culture, nasopharyngeal swab, and lumbar puncture (if indicated) collected. Basic demographics were collected, and babies were followed up until discharge or death to determine admission outcome. Blood cultures were processed using the BACTEC system and identification confirmed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Cerebrospinal fluid was processed using standard microbiological testing and swabs were processed using the multiplex real-time polymerase chain reaction assay. Antimicrobial susceptibilities of bacterial isolates to World Health Organization–recommended first-line antibiotics (ampicillin or benzylpenicillin and gentamicin) were assessed using e-tests. Results: A total of 7323 infants with signs or risk factors for sepsis had blood cultures, 2563 had nasopharyngeal swabs, and 23 had lumbar punctures collected. Eleven percent of blood cultures and 8.6% of swabs were positive. Inpatient mortality was 12.1%, with 27.7% case fatality observed among infants with Gram-negative bloodstream infections. Escherichia coli (14.8%), Acinetobacter spp. (10.3%), and Klebsiella spp. (7.6%), were notable contributors to Gram-negative sepsis, whereas Group B Streptococcus was the predominant Gram-positive pathogen identified (13.5%). Almost 60% of Gram-negative pathogens were ampicillin- and gentamicin-resistant. Conclusions: Our study demonstrates high levels of antimicrobial resistance and inpatient mortality from neonatal sepsis in the first months of life in Uganda. This underscores the pressing need for revised, context-specific antimicrobial treatment guidelines that account for the evolving landscape of antimicrobial resistance in neonatal sepsis.

Backgroud: Although not fully investigated, studies show that Legionella pneumophila can develop antibiotic resistance. As there is limited data available for Portugal, we determined the antibiotic susceptibility profile of Portuguese L. pneumophila serogroup 1 (LpnSg1) isolates against antibiotics used in the clinical practice in Portugal. Methods: Minimum inhibitory concentrations (MICs) were determined for LpnSg1 clinical (n = 100) and related environmental (n = 7) isolates, collected between 2006–2022 in the context of the National Legionnaire´s Disease Surveillance Programme, against azithromycin, clarithromycin, erythromycin, levofloxacin, ciprofloxacin, moxifloxacin, rifampicin, doxycycline, tigecycline, and amoxicillin/clavulanic acid, using three different assays. Isolates were also PCR-screened for the presence of the lpeAB gene. Results: Twelve isolates had azithromycin MICs above the EUCAST tentative highest WT MIC, 9 of which were lpeAB negative; for erythromycin and clarithromycin, all isolates tested within the susceptible range. The number of isolates with MICs above the tentative highest WT MIC for the remaining antibiotics was: ciprofloxacin: 7; levofloxacin: 17; moxifloxacin: 8; rifampicin: 11; doxycycline: 82; tigecycline: 4. EUCAST breakpoints are not available for amoxicillin/clavulanic acid. We estimated the ECOFFs and one isolate had a MIC eightfold higher than the E-test ECOFF. Additionally, a clinical isolate generated three colonies growing on the E-test inhibition zone that resulted in MICs fourfold higher than for the parental isolate. Conclusions: We report, for the first time, elevated MICs against first-line and other antibiotics (including azithromycin, fluoroquinolones and amoxicillin/clavulanic acid commonly used to treat pneumonia patients in Portugal) in Portuguese L. pneumophila strains. Results point towards decreased susceptibility in circulating strains, justifying further investigation.

Background: Although not fully investigated, studies show that Legionella pneumophila can develop antibiotic resistance. As there is limited data available for Portugal, we determined the antibiotic susceptibility profile of Portuguese L. pneumophila serogroup 1 (LpnSg1) isolates against antibiotics used in the clinical practice in Portugal. Methods: Minimum inhibitory concentrations (MICs) were determined for LpnSg1 clinical (n=100) and related environmental (n=7) isolates, collected between 2006-2022 in the context of the National Legionnaire’s Disease Surveillance Programme, against azithromycin, clarithromycin, erythromycin, levofloxacin, ciprofloxacin, moxifloxacin, rifampicin, doxycycline, tigecycline, and amoxicillin/clavulanic acid, using 3 different assays. Isolates were also PCR-screened for the presence of the lpeAB gene. Results: Twelve isolates had azithromycin MICs above the EUCAST tentative highest WT MIC, 9 of which were lpeAB negative; for erythromycin and clarithromycin, all isolates tested within the susceptible range. The number of isolates with MICs above the tentative highest WT MIC for the remaining antibiotics was: ciprofloxacin: 7; levofloxacin: 17; moxifloxacin: 8; rifampicin: 11; doxycycline: 82; tigecycline: 4. EUCAST breakpoints are not available for amoxicillin/clavulanic acid. We estimated the ECOFFs and one isolate had a MIC 8-fold higher than the E-test ECOFF. Additionally, a clinical isolate generated three colonies growing on the E-test inhibition zone that resulted in MICs 4-fold higher than for the parental isolate. Conclusions: We report, for the first time, elevated MICs against first-line and other antibiotics (including azithromycin, fluoroquinolones and amoxicillin/clavulanic acid commonly used to treat pneumonia patients in Portugal) in Portuguese L. pneumophila strains. Results point towards decreased susceptibility in circulating strains, justifying further investigation.