Breaking down the microbiology world one bite at a time

Custom Cures for a Superbug Showdown

Written by

Imagine you’re at a doctor’s office, feeling unwell with a stubborn cough. The doctor prescribes you antibiotics, hoping to knock out the bacteria causing the infection. But what if the bacteria have outsmarted the antibiotics? Over time, they’ve developed resistance, like a clever burglar who knows all the security codes. The antibiotics that once worked no longer do, leaving you vulnerable to even tougher, more dangerous infections. This is the growing problem of antimicrobial resistance (AMR). This makes it harder to treat infections, stretch out illnesses, and increase the risk of death, even from routine procedures. Traditionally, the doctors diagnose and prescribe selected antibiotics but those can often be inefficient and slow, leading to overuse of broad-spectrum antibiotics.

However, the new approach of personalised antimicrobial susceptibility testing (AST) and clinical prediction models are promising solutions to this growing problem. Personalised medicine is a tailored medical treatment plan for each patient based on their genetic makeup, medical history, and even microbial resistance profiles of pathogens. By using the personalised AST, physicians can choose the most effective antibiotic based on the patient’s unique microbial resistance patterns, reducing the use of broad-spectrum antibiotics and the risk of further resistance development. A recent advancement in the field includes the current study incorporating machine learning prediction models which are faster and minimise chances of developing resistance. This study focuses on urinary tract infection (UTI) and demonstrates that personalised AST could significantly improve treatment. Specifically, the personalized approach provided a 32% increase in susceptible results for WHO Access agents compared to the standard approach while maintaining overall efficacy in selecting effective treatments. WHO Access agents are a classification in the WHO AWaRe framework (Access, Watch, Reserve), a tool developed by WHO to support the combat of antimicrobial resistance. Antibiotics classified under the “Access” group are effective against a wide range of common infections and have lower resistance potential.

Additionally, physicians can also factor in a wide range of variables like patient health status (for example, if the patient is diabetic), previous treatments, genomic data like resistance profiles from NGS/WGS etc to guide treatment decisions. Primarily, this study explores how clinical prediction models can enhance effectiveness of AST. These models combine microbial resistance data with clinical information to recommend personalised antibiotic therapies. For example, a patient’s test results (for example- NGS data) are input into the model, which then predicts the most effective treatment based on the patient’s history and the pathogen’s genetic resistance profile.

While traditional AST methods still provide essential data, machine learning can process this data to provide more accurate, patient-specific predictions. It also helps reduce time in urgent clinical scenarios, guiding clinicians towards the right antibiotic faster. However, one limitation is that these prediction models rely on accuracy of patient history like their previous antibiotic use to work effectively. An incomplete or inaccurate history could lead to suboptimal predictions.

Another study points to the role of our microbiome (the community of bacteria living in our bodies) in resistance development. The study points out that alterations to the microbiome can make a patient more susceptible to resistant infections. By examining the patient’s microbiome alongside resistance data, clinicians can tailor antibiotic treatment more precisely, ensuring that therapy doesn’t just target the pathogen but also maintains the healthy microbiota found in the body.

By combining traditional diagnostic methods with cutting edge technologies like NGS, machine learning, and microbiome analysis, clinicians can make better informed decisions and tailor treatments to individual patients. In future, we can expect integration of these advanced tools into routine clinical practice to reverse the AMR crisis, ensuring that antibiotics remain effective for future generations.

Link to the original post: Howard, A., Hughes, D.M., Green, P.L. et al. Personalised antimicrobial susceptibility testing with clinical prediction modelling informs appropriate antibiotic use. Nat Commun 15, 9924 (2024). https://doi.org/10.1038/s41467-024-54192-3

Featured image: Bing Image Creator

Additional references: Sirwan Khalid Ahmed, Safin Hussein, Karzan Qurbani, Radhwan Hussein Ibrahim, Abdulmalik Fareeq, Kochr Ali Mahmood, Mona Gamal Mohamed, Antimicrobial resistance: Impacts, challenges, and future prospects, Journal of Medicine, Surgery, and Public Health, Volume 2, 2024, 100081, ISSN 2949-916X, https://doi.org/10.1016/j.glmedi.2024.100081