Investigating the impact of sepsis phenotypes on antibiotic treatment in patients with severe pneumonia and sepsis

Research question: Can sepsis phenotypes help explain antibiotic blood concentration variability in critically ill patients with severe pneumonia and sepsis?

Background: Pneumonia is a significant health concern in the UK, with high mortality rates, particularly when complicated by sepsis. Specific immune phenotypes have been associated with worse outcomes in patients with pneumonia and sepsis.

Achieving adequate antibiotic concentrations in blood is crucial for treatment success and limiting antimicrobial resistance.

Standardised antibiotic dosing leads to inadequate antibiotic concentrations in approximately 30% of critically ill patients.

My recent systematic review identified that the impact of immune phenotypes, and the associated physiological disruption, on antibiotic concentrations is unknown.

Aims and objectives: My aim is to explore the pharmacokinetic variability of antibiotics in different sepsis phenotypes.

The research objectives are to (1) describe the pharmacokinetics of a cohort of critically ill patients with severe pneumonia receiving beta-lactam antibiotics; (2) characterise sepsis phenotypes in the same population; and (3) explore the interaction between sepsis phenotypes and antibiotic pharmacokinetics using mathematical modelling.

Methods: This prospective observational cohort study has been designed to enrol 119 critically ill adults with severe pneumonia treated with piperacillin/tazobactam in the first 24 hours of treatment.

I will collect blood samples and measure concentrations of piperacillin/tazobactam on four occasions over two separate dosing intervals.

Sepsis phenotypes will be characterised using biomarkers, cytokines and RNA sequencing, according to the previously published and validated Sepsis Response Signature and hypo- or hyperinflammatory phenotypes.

Alongside regression analyses, I will use the phenotype classifications as covariates in nonlinear mixed-effects models of piperacillin/tazobactam, followed by Monte Carlo simulations.

I will look at differences in serum piperacillin/tazobactam concentrations between groups as the primary outcome, and Sequential Organ Failure Assessment scores, all-cause mortality at 28 days, length of stay, clinical cure, treatment target attainment, safety and toxicity as secondary outcomes. The Strengthening the Reporting of Observational Studies in Epidemiology Statement will be used for study reporting.

Timelines for delivery: The first year will be dedicated to project setup, an updated systematic review of the relevant literature, and an educational overseas research visit as part of my initial research training.

Once the study opens, I will recruit participants over 18 months with an anticipated study close in the second quarter of 2027. A further six months will be allocated to analyses and write-up. Study results will be published by the end of 2027.

Anticipated impact and dissemination: Sepsis phenotypes may explain some of the observed antibiotic blood concentration variability in critically ill patients.

My research findings will inform future mathematical models that could impact dosing schedules of commonly used antibiotics in patients with severe pneumonia and sepsis.

Outputs from this fellowship will be disseminated through multiple peer-reviewed manuscripts and conference proceedings in collaboration with patient partners.

Individualised dosing for patients with low antibiotic levels, as opposed to 'one size fits all' prescribing, has the potential to improve patient outcomes, limit the development of antimicrobial resistance, and more efficiently allocate scarce resources to those who will benefit the most.