Investigating the rising incidence of Gram-negative bloodstream infections in England using large linked epidemiology, clinical and genomic datasets

Bacterial infections of the blood are serious, life-threatening events which cost the NHS huge amounts of money. One particular group, so-called 'Gram-negative bacteria' (including eg. E. coli), have become an increasingly common cause over the past decade. More concerningly, recent research has shown that antibiotic resistance in these bacteria is becoming a large and rapidly increasing problem in some parts of the country. Effective antibiotics are vital to the safety of many aspects of healthcare including cancer treatment and surgery, and this particular group of bacteria make a very important contribution to the global issue of antibiotic resistance. The World Bank predicts that by 2050, antibiotic resistance could push an additional 28.3 million people, particularly those in low-income countries, into extreme poverty.

The UK government has placed antibiotic resistance on the national risk register and aims to halve by 2021 the number of these 'Gram-negative' bloodstream infections. It is currently difficult to see how this can be achieved because we don't have a good understanding of the reasons why these infections are becoming more common. In this project, I will explore regional variations in how commonly these infections, and particularly antibiotic-resistant infections, occur. By doing this I plan to understand the extent to which different factors thought to cause these infections are contributing to their increase. Researchers have recently shown that in Oxfordshire, bloodstream infections caused by one type of these bacteria which arise out of hospital are increasing faster than those following hospital admission. One explanation for this might be that antibiotic prescribing by GPs is an important risk factor for serious bloodstream infections in some patients. I will investigate if the findings in Oxfordshire are also seen nationally and in other types of 'Gram-negative' bacteria. If there truly is a link between antibiotics consumed in the community and serious bloodstream infections, this would be an important finding and a powerful message to help GPs reduce the number of potentially unnecessary antibiotic prescriptions.

I will use the latest DNA sequencing technologies to look in detail at how similar the bacteria causing infections in different patients are. This will help me to investigate whether these infections may be spreading between patients, for example in hospitals or nursing homes. 'Gram-negative' bacteria have the interesting property of being readily able to acquire new genes from their surroundings. These are found on bits of DNA called 'mobile genetic elements' which provide useful extra functions for bacteria like resistance to antibiotics. The technology to study these in detail has only recently emerged and I will use it to understand more about the ways in which they contribute to the spread of antibiotic resistance. I will also look at whether there is any evidence that they are able to move between different members of the 'Gram-negative' family of bacteria.

Finally, I will explore whether there are specific features in the DNA of 'Gram-negative' bloodstream infections that occur out of hospital which explain why they are becoming more common. I will also investigate whether we might be able to use the genetic code of these bacteria to explain why some patients become more unwell than others when they develop these infections. In future, this may help doctors to quickly identify which patients are likely to need specialist care.

This study will help scientists and healthcare professionals to better understand why these infections are becoming more common and hopefully develop new strategies to bring this emerging problem under control as a result.

Gram-negative bloodstream infections (GNBSI) are associated with significant morbidity and mortality and represent a significant factor in the global problem of antimicrobial resistance (AMR). GNBSI have been increasing in incidence in England and worldwide over the past decade although the precise reasons for this are unknown. Recent research in Oxford has suggested that for E. coli, the increase is primarily driven by community rather than nosocomial onset infections and that antibiotic resistant infections are increasing faster than susceptible ones. Identifying causal risk factors for acquisition and transmission is a significant research priority.

In this study I will leverage large national databases to systematically identify healthcare exposures which are driving the increasing numbers of GNBSI cases and associated AMR across England. I will investigate in detail the role of the mobilome in the dissemination of AMR within and also between species. My initial analysis suggests that a subset of E. coli bloodstream infections are genetically highly similar to isolates from different patients which may suggest transmission or point source acquisition plays a role in some of these infections; I will explore the evidence for this. I will conduct genome wide association studies to search for pathogen genomic factors which may be contributing to the increasing incidence of community associated infections or are predictive of clinical severity.

Interrogating such vast and rich datasets provides an opportunity to identify epidemiological exposures driving the increase of GNBSIs and the underlying pathogen biology which these may be selecting for, aiding the identification of potential future interventions. In addition, a more robust evidence base regarding the potential for iatrogenic factors such as excess antibiotic prescribing and urinary tract instrumentation to cause severe invasive infections would be a powerful incentive for better stewardship.

Research to combat antimicrobial resistance is critical to the future of human health as well as the global economy and there are therefore likely to be many interested stakeholders in this project including:

Patients
- Patients are generally well aware of the potential benefits of antibiotics but often feel they are not associated with significant harm. Improving the evidence base for the harm associated with unnecessary antimicrobial prescribing will empower informed patients to take in active role in reducing their exposure and that of their families in partnership with healthcare professionals. (timescale 3-5 years)

Policymakers/politicians
- The World Bank predicts that by 2050, global GDP would fall by 1.1% in a low-impact AMR scenario and 3.8% in a high-impact one, with low-income countries being affected in a disproportionate manner. Healthcare costs are estimated to significantly increase and livestock would also be affected by the 'One Health' nature of AMR. Research developing methodologies to identify the emergence and spread of AMR is therefore likely to be important to better inform future models and allow policymakers to prepare (timescale 15-30 years)
- Policy makers need actionable insights on which to focus limited resources. The UK government has set a target to halve healthcare-associated GNBSI by 2021 which will be impossible without an evidence base for what is driving the increasing incidence of these infections. My research project is aimed at directly addresses this need and will help policymakers decide to what extent interventions should be focussed on particular healthcare settings/exposures. (timescale 3 years)

Public Health Practitioners
- The National Infection Service is PHE's scientifically-led service with a remit to reduce the burden of infectious disease in the English population. Gram-negative bloodstream infections are a key aspect of this and development of methodology and experience of analysis of multiple routinely collected datasets will be very beneficial. (timescale 1-3 years)
- The application of whole genome sequencing for surveillance of infections as well as routine diagnostics is a key area of development in PHE, with Gram-negatives presenting a particular challenge. Insights and methodology gained from this project will play an important role in the development of this over the next few years. (timescale 3-5 years)
- A more robust evidence for the role of antibiotic prescribing/medical procedures in severe invasive infections will help PHE to persuade clinicians to change their practice
- Development of methodologies for the analysis of linked genomics and epidemiological datasets on a large scale will help to improve the utilisation of routinely collected data for the benefit of public health.

Clinicians
- Knowledge of the role of nosocomial environments in GNBSI will help infection control teams design new clinical facilities and improve existing ones. (timescale 3-5 years)
- A more robust evidence base around the individual and population level harms caused by excess antibiotic prescribing is likely to aid clinical decision making and offer protection to clinicians deciding that potential risks outweigh the benefits. (timescale 3-5 years)

Charities:
- Charities can use research such as this to highlight the impact and importance of research into infection and antimicrobial resistance.

Schools:
- Inspiring the next generation to be interested in STEM subjects and actively seek to gain skills, particularly in coding, at an early stage is vital to the economic future of the country. Demonstrating practical 'real world' uses of these skills to tackle major global problems will hopefully inspire school children to become interested in these subjects.