Aim of the research: To find out why antibiotics work differently in certain patients with severe pneumonia and sepsis.

Background: Individuals can become very unwell from pneumonia, sometimes requiring admission to hospital or even the intensive care unit (ICU). In some cases, pneumonia can lead to a condition called sepsis, which can be deadly if not treated quickly. In the UK, approximately 30,000 patients die from pneumonia every year. Clinicians use antibiotic injections to treat life-threatening infections such as severe pneumonia. After being injected into the bloodstream, antibiotics quickly spread throughout the body, attacking the infection. Antibiotics are eventually broken down and removed from the body by the kidneys and other organs. However, antibiotics fail to achieve the same consistent result for every patient. This may be to do with the way the antibiotics travel through and are removed from the body, leading to different antibiotic levels in the blood at any one time. Low antibiotic levels can result in worse outcomes and antibiotic resistance. Patients can be grouped based on how their immune system reacts to infections. The SIPRES Study aims to explore if these previously described groups explain the difference in antibiotic levels in patients with severe pneumonia and sepsis.

Procedures: We will study how adult patients with severe pneumonia respond when treated with the most commonly used antibiotic in the ICU called piperacillin/tazobactam. Alongside information on how quickly patients get better and how long they need to stay in hospital or in ICU, we will collect blood samples to measure antibiotic levels and assess each patient's immune system at two time points during their treatment. This will allow us to measure antibiotic levels in blood at different times and group patients based on their immune system reaction to infection. We will describe the range of antibiotic levels seen in the different immune system reaction groups using mathematical and statistical models.

Patient involvement: We are working closely with people who have experienced severe pneumonia and will work with two patient partners and a patient advisory group to help shape this research. Patient contributors have already shaped the development of the funding application and identified important study outcomes. Patients we have spoken to are concerned over the appropriate dosing of antibiotics and appreciate the need for improved and precise approaches to treating severe infections. Moving forward, patient partners will help finalise the protocol, develop patient and public facing materials, provide their perspective on the study results and shape plans to share the outcomes of the study more broadly.

Potential impact: The SIPRES Study will help identify a group of patients at risk of low antibiotic levels in blood, who are less likely to improve with treatment and more likely to develop antibiotic resistance. Mathematical models that can help clinicians personalise antibiotic dosing for each critically ill patient with severe pneumonia will be developed. Findings have the potential to limit the development of antibiotic resistance and help patients survive and get better faster so that they can return to their normal daily lives. Individualised dosing for patients with low antibiotic levels, as opposed to 'one size fits all' prescribing, also has the potential to more efficiently allocate scarce resources to those who will benefit the most.

Inclusion Criteria:

• age ≥ 18 years;
• admitted to intensive care and receiving at least one-organ supportive care;
• treated for presumed or confirmed lower respiratory infection;
• receiving or about to receive piperacillin/tazobactam as part of standard clinical care;
• valid informed consent or enrolment through deferred consent pathway appropriate.

Exclusion Criteria:

• unlikely to survive 24 hours as judged by the treating physician;
• study antimicrobial started more than 24 hours prior.

Hospitalised critically ill patients admitted to the intensive care unit

The following anonymised data will be shared on an online accessible repository in line with Findability, Accessibility, Interoperability, and Reuse (FAIR) principles:

• Baseline clinical characteristics without identifiable data
• Outcome data
• Pharmacokinetic data
• Transcriptomic data
• Immune cytokine data

Beginning one year after publication and indefinitely

A limited anonymised dataset will be shared openly, allowing secondary analyses; more detailed IPD may be shared with researchers upon reasonable request.