DOI: 10.4103/wkmj.wkmj_72_26 ISSN: 2707-6180

Temporal and Geographic Patterns of Carbapenem Resistance among Database-submitted Enterobacterales Isolates

Aziza Omari, Nurlan Sadykov, Karlygash Akpayeva, Bakhytzhan Seksenbaeyev, Assylkhan Bekbayev
Show PDF Cite

A
BSTRACT

Objective:

Carbapenem-resistant Enterobacterales are a major antimicrobial-resistance threat, but open surveillance databases are affected by uneven isolate submission. This study described temporal and geographic patterns in carbapenem resistance among database-submitted Enterobacterales isolates and evaluated how adjustment for database composition changed trend inference.

Methods:

A retrospective open-database surveillance analysis was performed using phenotypic antimicrobial-susceptibility records from the European Molecular Biology Laboratory-European Bioinformatics Institute Antimicrobial Resistance (AMR) Portal release 2025-12. Enterobacterales isolates collected from 2000 to 2024 with carbapenem susceptibility results were included. Resistance was defined as resistance to at least one tested carbapenem. Prevalence estimates were calculated with Wilson 95% confidence intervals (CIs). Temporal trends were analyzed using crude and adjusted logistic regression, country-clustered robust standard errors, country fixed effects, region-by-year interactions, and sensitivity analyses.

Results:

The isolate-level dataset included 37,501 isolates from 84 countries and 34 species. Overall, 2751 isolates were carbapenem resistant, giving an observed prevalence of 7.34% (95% CI 7.08–7.60). Regional prevalence was highest in Africa (35.96%) and Asia (30.50%) and lowest in Oceania (0.43%). Among species with at least 100 isolates, Klebsiella pneumoniae had 43.30% resistance, whereas Salmonella enterica had 0.01%. The crude temporal model suggested decreasing annual odds, whereas adjustment for region and species suggested increasing odds; country-level modeling changed the inference.

Conclusion:

Observed resistance varied substantially by species, region, country, and year. Open AMR databases are valuable for reproducible surveillance analytics, but submitted-isolate proportions should not be interpreted as population-level prevalence without adjustment and sensitivity analysis.

More from our Archive