Antibiotic-resistant bacteria pose a growing public health threat. As those bacteria resist all antibiotics referred to as pan-resistance antibiotic treatment will become extremely difficult. Researchers hypothesized that a steep increase in mortality rates would accompany the emergence of pan-resistant bacteria. To test this hypothesis, they developed a numerical and computer model to simulate the impact of a hypothetical pan-resistant bacterium on death rates in the United States.
According to our simulations, during the first five years after the emergence of a single pan-resistant common bacterium, the number of deaths from infection in the general population could increase dramatically. Failure to prevent pan-resistance emergence will spell doom to public health.
Estimating sepsis incidence and mortality was conducted using published data on hospital discharge surveys, electronic health records (EHR), and insurance claims. Patient consent was not required as the data were publicly available and fully de-identified, ensuring privacy and confidentiality. Ethical approval from institutional review boards was not required because the study utilized publicly accessible, de-identified data.
Owing to variable diagnosis and coding practice in claims databases, claims-based data usually underestimate the incidence of sepsis. Hence, researchers estimated a correction factor for claims-based estimates using data from 2009 to 2012, a period when both data sources provided overlapping information. The mean difference in total U.S. sepsis cases between the electronic health records of Rhee et al. and the claims-based data of Stoller et al. was computed. The correction factor was then applied to all claims-based incidence estimates to enhance accuracy.
Likewise, mortality rates due to sepsis were estimated from the claims data and electronic health records without consideration of death certificate data, especially those from the CDC (2023). It was assumed that death certificate data did not adequately capture sepsis or organ dysfunction as contributing factors to death at the time of death and, therefore, were likely to produce mortality estimates that were too low.
No adjustments were made for variability in diagnostic and coding practices needed for sepsis mortality rates. To put into definition the recent path of cases of sepsis and mortality thereof in the US, incidence estimates were aggregated from peer-reviewed literature. From 1979 to 2014, a collective amount of nearly 100 million hospitalizations was documented in several studies. Sepsis incidence estimates were derived from hospital discharge surveys, electronic health records, and insurance claims.
Over this period, coding practices for medical records related to sepsis have changed, an artifact most likely contributing to the observed trends over time. However, these data showed a concurrent increase sepsis incidence rates alongside a decline in sepsis case mortality rates from 1979 to 2014.
Authors modeled the emergence of the theoretical pan-resistant strain of E. coli according to natural history on an already well-characterized multidrug-resistant E. coli strain-the H30R subclones-of multi-locus-sequence types 131, ST131-H30R. The biological basis of the rapid success of ST131-H30R is still under investigation and tentatively ascribed to the acquisition of fluoroquinolone resistance-determining mutations & CTX-M-15 beta-lactamase gene during the period of high prescription rates for extended-spectrum cephalosporins and fluoroquinolones.
Alarmingly, H30R has recently developed carbapenemase-mediated resistance to carbapenems last traditional line of defense against extremely resistant E. coli strains-and is now the predominant, and in some places the only, recognized carbapenem-resistant E. coli lineage. Further probably associated reasons with the impressive ascent of ST131-H30R are found in the acquisition of several accessory genes that increase fitness, unique patterns of methylation, and mutations within regulatory regions that might well enhance virulence, commensalism, and/or transmissibility, phenotypes for which ST131-H30R is known.
All models indicate a significant rise in annual sepsis-related deaths with the emergence of a pan-resistant E. coli strain projected for 2040. With even the most conservative of estimates, our models predict that the appearance of just one pan-resistant strain along with the relative success of E. coli ST131-H30R would more than multiply E. coli sepsis deaths 18 times (an average of 7240 per year) in just five years. These data suggest that not preventing extraintestinal pathogenic E. coli from switching to a pan-resistant state will have serious consequences for both public health and clinical medicine.
A diverse range of microbial species contributes to the growing global threat of antibiotic resistance. Findings on antibiotic resistance based, however, on one strain of E. coli, specifically from the United States, should give rise to data-driven perspectives on the already incomplete picture of the global crisis. Modeling studies along the same lines for other important diseases could more accurately warn clinicians, policy makers, and the educated public on how to avoid antibiotics in the future. The accuracy of such a forecast should make prioritizing antibiotic resistance over today’s competing demands more effective.
Reference: Koch BJ, Park DE, Hungate BA, et al. Predicting sepsis mortality into an era of pandrug-resistant E. coli through modeling. Commun Med. 2024;4:278.
https://doi.org/10.1038/s43856-024-00693-7
