Combining demographic shifts with age-based resistance prevalence: a modelling estimate of future AMR burden in Europe and implications for targets

Background

Antimicrobial Resistance (AMR) is a global public health crisis, with a range of potential interventions that must be optimised to achieve global targets. Evaluating the interventions requires accurate estimates of how the AMR burden of infection will change over time, especially given likely demographic shifts. This study aimed to provide an estimate of future burden in Europe, investigating resistance variation by age and sex and the impact of interventions to achieve the scale of proposed UN political declaration targets.

Methods

Using data from 12,807,473 bloodstream infection (BSI) susceptibility tests from routine surveillance in Europe, we estimate age- and sex-specific rates of change in BSI incidence for 8 bacteria over 2015- 2019. This was used to project incidence rates by age and sex for 2022-2050 and, with demographic projections, to generate estimates of BSI burden (2022–2050). Bayesian hierarchical models were fitted across 38 bacteria-antibiotic combinations to the 2010-2019 resistance proportion of BSI at the country-level with and without age/sex disaggregation. Inputting the incidence estimates into the “agesex” and “base” model respectively, we sampled 1,000 model estimates of resistant BSI burden by age, sex and country to determine the importance of age and sex disaggregation. We explored Intervention scenarios consisting of a 1, 5, or 20 per 100,000 per year reduction in infection incidence rate of change or an age-targeted 5 per 100,000 per year reduction in those aged 65 years and older.

Results

Overall, in Europe, BSI incidence rates are predicted to increase more in men than women across 7 of the 8 bacteria ( P. aeruginosa was the exception) and are projected to increase more dramatically in older age groups (74+ years) but stabilise or even decline in younger age groups. We project huge country-level variation in resistance burden to 2050, with opposing trends visible in different countries for the same bacteria-antibiotic combinations (e.g. aminoglycoside resistant Acinetobacter spp. ranged from a relative difference of 0.34 to 15.38 by 2030).

Not accounting for age and sex results in differing resistance burden projections, with 47% of bacteria- antibiotic combinations resulting in an overall under-estimate of resistant BSIs by 2030. Not including age/sex resistance patterns underestimates male cases for 76% (29/38) of the combinations compared to 11% (4/38) for women. We also often see strong age-based associations in resistant BSIs projections with bigger differences at older ages.

Achieving a 10% reduction in resistant BSI incidence by 2030 (equivalent to the UN 10% mortality target) was possible only for 57% (26/38) of bacteria-antibiotic combinations even with large reductions in BSI incidence rate of change of -20 per 100,000 per year. In some cases despite achieving a 10% reduction, the resistant BSI burden bounces back to above previous levels by 2050.

Conclusions

Including country-specific age- and sex-specific resistance levels alongside projected demographic shifts has a large impact on resistant BSI infection burden projections in Europe even to 2030. Reducing this AMR infection burden by 10% will require substantial reductions in infection incidence rates.