Antibiotic resistance (ABR) is a major public health threat, with significant geographical and bacterial species-dependent variations observed.^{1 ,2} There are currently limited studies examining the factors driving the emergence and transmission of ABR. Understanding ABR dynamics requires study at the species level, which takes into account varied resistance mechanisms, ecological niches and epidemiological sources.

Previous studies on the determinants of country-level ABR suggested a correlation between ABR rates and socioeconomic factors, in addition to antibiotic consumption.^{3, 4} This finding emphasised the need to contain ABR transmission to reduce its global burden, rather than by only restricting antibiotic use.

A recent global study of ABR dynamics was conducted through a multivariable spatial-temporal analysis utilising the Antimicrobial Testing Leadership and Surveillance (ATLAS) system, which collects clinical isolates from diverse bacterial infections, to identify determinants of ABR on a global scale.⁵ The study, published in The Lancet Planetary Health, aimed to identify the main country-year determinants of ABR across different drug-bacterium pairs of clinical interest for human health.⁵ Longitudinal surveillance data of ABR rates from infection samples for 51 countries over 14 years (2006–2019) were analysed using a mixed-effect negative binomial model to investigate associations with potential drivers of ABR.⁵

A total of 808,774 isolates from all infection sources across 13 different clinically relevant drug-bacterium pairs were analysed.⁵ Covariables such as antibiotic sales, meteorological and climatic variables, wealth and health metrics, population density and tourism were selected and tested as explanatory factors in the statistical analysis. Univariate and multivariable analyses were conducted to identify associations between ABR rates and the selected covariables. Sensitivity analyses were performed to assess the effect of potential biases on the outcomes of the model.

ABR patterns were found to be strongly dependent on the source country and drug-bacterium pairs analysed, and significant heterogeneity was observed between countries.⁵

Other key findings were:⁵

• Median ABR rates were similar between carbapenem-resistant Acinetobacter baumannii [72.3%; interquartile range (IQR), 40.5–90.7] and aminopenicillin-resistant Escherichia coli (68.0%; IQR, 57.7–75.9), with significant variation noted between countries.
• From 2006 to 2019, ABR rates showed different temporal trends between drug-bacterium pairs. Some pairs exhibited global upward trends, such as carbapenem-resistant Acinetobacter baumannii (24 of 47 countries exhibiting statistically significant positive slopes), while others showed downward trends, such as fluoroquinolone-resistant Pseudomonas aeruginosa (24 out of 50 countries exhibiting statistically significant negative slopes).
• Carbapenem resistance has been increasing in more than 60% of investigated countries between 2006 and 2019: carbapenem-resistant Klebsiella pneumoniae (61%; 13 statistically significant positive slopes); carbapenem-resistant Pseudomonas aeruginosa (76%; 14 statistically significant positive slopes); and carbapenem-resistant Acinetobacter baumannii (83%; 24 statistically significant positive slopes).

Associations between ABR rates and potential drivers varied greatly between drug-bacterium pairs.⁵

• Antibiotic sales were significantly associated with ABR rates for specific drug-bacterium pairs.
  • Carbapenem-resistant Acinetobacter baumannii: ABR rates were positively correlated with carbapenem sales but inversely correlated with global antibiotic sales.
  • Penicillin-nonsusceptible Streptococcus pneumoniae and macrolide-resistant Streptococcus pneumoniae: ABR rates were positively associated with global antibiotic sales.
• Meteorological factors were mostly found to be associated with Enterobacteriaceae and Streptococcus pneumoniae drug–bacterium pairs.
  • Average temperature was found to be significantly associated with ABR rates in Enterobacteriaceae drug-bacterium pairs, particularly Escherichia coli and carbapenem-resistant Klebsiella pneumoniae.
  • Extreme climatic events, such as flooding, were positively associated with ABR rates of Escherichia coli and macrolide-resistant Streptococcus pneumoniae, suggesting that natural disasters may contribute to high resistance rates.
  • ABR rates of penicillin-non-susceptible Streptococcus pneumoniae were inversely associated with average relative humidity.
  • Rainfall was negatively associated with ABR rates of fluoroquinolone-resistant Acinetobacter baumannii.
• Health system quality, as measured by the Global Health Security (GHS) index, has been associated with decreased ABR rates for most drug-bacterium pairs, highlighting the importance of hygiene and infection control measures.
  • The GHS index was inversely associated with ABR rates in all Escherichia coli and Klebsiella pneumoniae drug–bacterium pairs, as well as in Pseudomonas aeruginosa and Acinetobacter baumannii
• Economic factors, such as GDP per capita, have shown mixed associations with ABR rates, with carbapenem-resistant Acinetobacter baumannii positively associated with GDP per capita, while others showed an inverse association including fluoroquinolone-resistant Pseudomonas aeruginosa and macrolide-resistant Streptococcus pneumoniae.
• Tourist departures were inversely associated with ABR rates for aminopenicillin-resistant Escherichia coli and carbapenem-resistant Pseudomonas aeruginosa.

Despite considering various explanatory variables, there is high unexplained spatial random effects (RE) variance, indicating the presence of unknown factors contributing to ABR.⁵ Some countries, such as Mexico, Japan, South Korea and Czech Republic, exhibited particularly high spatial random effects estimates compared to other countries in the ATLAS database, across different drug-bacterium pairs.⁵

In conclusion, the study revealed that ABR patterns were found to be strongly dependent on the country and drug-bacterium pairs examined, indicating heterogeneity in ABR rates across space and time.⁵ The study highlights a need for tailored interventions to address bacterial resistance, given the diversity of mechanisms driving global antibiotic resistance across pathogens.⁵

 References

1. World Health Organization. Citing websites: Global action plan on antimicrobial resistance. 2015. Available from: http://www.emro.who.int/health-topics/drug-resistance/global-actionplan.html. Accessed 20 Sep 2021.
2. Organisation for Economic Co-operation and Development. Stemming the superbug tide: just a few dollars more. Paris, France: OECD Publishing; 2018.
3. Collignon P, et al. Lancet Planet Health 2018;2:e398–405.
4. Blommaert A, et al. J Antimicrob Chemother 2014;69:535–47.
5. Rahbe E, et al. Lancet Planet Health 2023;7:e547–e557.