Sepsis - Projects and Studies
The AI Clinician
Sepsis, or overwhelming infection, is a common reason for intensive care unit (ICU) admission and death throughout the world. Despite advances in the way we treat this condition, it remains a significant economic and healthcare burden. A key part of the treatment of sepsis is the administration of IV fluids and blood pressure medication. However, there is huge uncertainty around dosing of these drugs in an individual patient.
The AI Clinician is a study which explores whether the use of artificial intelligence applied to large medical databases can be used to tackle this issue by recommending the correct medication doses for individual patients to doctors. Our previous work has shown it has the potential to improve patient survival rates. The tool will be capable of processing patient data within the electronic patient record of NHS hospitals in real-time to suggest a course of action.
The long-term expected benefits of this project are numerous: improved patient survival, reduced use of precious intensive care resources and reduction in healthcare costs.
More information can be found on the AI Clinician webpage.
Gene-expression responses to corticosteroids in septic shock
Septic shock is when severe infection leads to low blood pressure and organ failure. It is a life-threatening condition. Corticosteroids are often used in the sickest patients with sepsis to try to improve blood pressure and outcomes for patients. Although we know that corticosteroids can shorten the duration of septic shock, their effect on patient survival rates is less clear. Some trials have shown a potential improvement with others showing no difference.
Previous research has suggested that two groups of patients with sepsis exist based on the way in which their genetic information is switched on or off during their illness. One group has poorer immune function and worse outcomes than the other. However, our research has suggested that when the effects of corticosteroids are examined in these two groups of patients, the responses are very different. The group whose immune function normally works well and who normally have good outcomes appear to come to harm with these drugs, an effect not seen in the other group. The reasons for these differences in treatment effect are not known but we propose that the corticosteroids affect the way that patients’ genes are switched on or off and may affect the way the immune system works.
In this study we are measing the way that patients’ genes are switched on or off in the two sepsis groups and in response to these drugs. Better understanding the mechanisms by which corticosteroids have their effects will help doctors to better target these treatments in the future.
Genetic influences on treatment response in sepsis
Severe infection, also known as sepsis, is a life-threatening condition that has an enormous impact for patients and society. The response to severe infection and its treatment are very variable from one patient to another. However, the reasons for this variability are not fully known. We propose that some of this variation is accounted for by genetic predisposition.
In this study we are identifying changes in genes that could reduce the effectiveness of treatments for sepsis. Once these have been identified we will measure them in blood samples collected during clinical trials of drugs used in sepsis. We will then be able to test if these genetic changes influence which patients respond to the treatments tested in these trials. If our hypothesis is proven, in the future, it will enable us to measure patients’ genes to determine who will get most benefit from sepsis treatments.
Molecular Phenomics of Septic Shock
The measurement of hundreds and thousands of small molecules, known as molecular fingerprints, can provide rich information about patients. Following how these fingerprints change through time can provide an indication of progression of disease or response to treatment. Analysis of the variation in these molecular fingerprints will allow the development of models to predict treatment responders and non-responders.
In critical care medicine the health of the patient can change rapidly, and early markers of organ dysfunction are essential. We are applying molecular fingerprinting techniques to blood samples collected as part of two clinical trials in septic shock. We aim to determine how monitoring the trajectories of molecular fingerprints over time can better determine those patients who are deteriorating on their current treatments and detect responders and non-responders to the treatments used in the trials.
Current Clinical Trials
Community-acquired pneumonia (CAP) is when individuals who have not been recently hospitalised develop an acute infection of the lungs. CAP that is severe enough to require admission to an Intensive Care Unit (ICU) is associated with substantial risk of dying, with pneumonia being responsible for half of all cases of sepsis and septic shock. Bacterial and viral infections are responsible for the vast majority of CAP, and this includes the COVID-19 virus.
All patients who are treated in an ICU, receive care that consists of multiple different treatments, which act together to treat both the infection and its effects on the body. REMAP-CAP uses an innovative trial design to efficiently evaluate multiple interventions simultaneously, reviewing the many treatment options that are available.
The REMAP-CAP trial provides a global research platform that can adapt to efficiently evaluate multiple treatment options for patients who are critically ill due to a global respiratory pandemic. During the COVID-19 pandemic, REMAP-CAP has swiftly adapted to meet the growing demands in ICU care and provided major breakthroughs in the treatment of severe COVID-19.
Sepsis is the most common cause of death in critically ill patients, with a quarter of those progressing to develop severe sepsis during hospitalisation. These hospital acquired infections are recognised as an increasing public health problem causing more than a million deaths worldwide. Although there has been improvement in outcomes of patients with severe sepsis, the number of patients with sepsis has increased. Currently there is no effective treatment for severe sepsis caused by bacteria other than prompt administration of fluids, antibiotics and supportive care.
It is already well established that administration of beta-lactam antibiotics is beneficial to patients with severe sepsis as these antibiotics treat the most common Gram +ve/-ve bacteria causing infection. In animal and in vitro models it has been shown that continuous infusion of these antibiotics is superior to intermittent infusion, however human trials have been under-powered to provide significant evidence of this. The BLING III trial aims to determine whether giving a patient a continuous infusion of a beta lactam compared to an intermittent beta lactam infusion would affect patient outcomes and the vital status of patients at hospital discharge. Preceded by BLING I and BLING II, BLING III is a multicentre randomised, controlled trial designed to provide necessary evidence on whether there is a survival and economic benefit associated with continuous beta-lactam infusion in patients critically ill with sepsis.
The IMPACCT study will investigate whether it is possible to identify and stratify patients with sepsis who are at higher risk of poor outcomes and developing new infections. IMPACCT is an observational study with no intervention or novel treatment introduced and no change to patients’ standard care and treatment when taking part. A clinical diagnostic test will be conducted on participants’ blood samples and the study team will collect and analyse data on the patient’s medical condition throughout their treatment in hospital and after they have been discharged. Participants will be adult patients in intensive care who are being, or have been, treated for suspected sepsis. Over a 2-year period the study will recruit 600 patients in hospitals in the UK, France and Sweden.
The test performed on the blood samples is a rapid RNA-based diagnostic test that produces an Immune Profiling Panel (IPP). The IPP can then be used to categorise and stratify participants. Several outcomes will be collected and assessed for the participants, within hospital and following discharge. At the end of the study the analysis will examine any association between the outcomes and the stratified groups determined by the IPP.
Why it is important
Future patients could benefit from the findings of the study as a result of improvements to clinical care and potentially tailored personal treatments. At present there is no clinical test to assess the immune function of patients who have sepsis. If the study is able to demonstrate, through the use of the RNA-based diagnostic test, that it is possible to identify patients with sepsis who are at risk of poor outcomes then that could enable a more personalised medicine approach to the treatment of sepsis.
Visit the IMPACCT Study page to find out more about IMPACCT and how it aims to investigate identifying patients at higher risk of developing new infections.
The SuDDICU study is examining the effects on patients who receive selective decontamination of the digestive tract (SDD) and determining whether this preventative measure is effective in reducing new ICU acquired infections and reducing hospital mortality.
SDD is an infection control approach designed to reduce mortality through preventing sepsis. This study aims to determine if the systematic delivery of SDD to critically ill patients within an ICU is clinically effective, and cost-effective, at reducing hospital mortality without increasing antibiotic resistance.
Septic shock is when severe infection leads to low blood pressure and organ failure. It is a life-threatening condition. Drugs called corticosteroids are often used in the sickest patients to try to improve blood pressure and outcomes for patients. Although we know that corticosteroids can shorten the duration of septic shock, their effect on patient survival rates is less clear. Some trials have shown a potential improvement with others showing no difference.
We have shown that patients with septic shock can be divided into two groups based on the patterns in which their genetic information is expressed in blood samples. These patterns have been called Sepsis Response Signatures or SRS. Normally patients with the first pattern (SRS1) of gene expression are at a higher risk of dying than those with the second (SRS2). This maybe because patients with the SRS1 pattern may have a weakened immune system and maybe less able to fight infection.
We have divided patients from a previous trial, where patients with septic shock were randomly allocated to get corticosteroids or a placebo, into SRS1 and 2 groups. Patients in both SRS groups had a shorter duration of septic shock if given corticosteroids. However, those in the SRS2 group, who normally have better survival, were more likely to die if they received corticosteroids than if they received placebo. Although corticosteroids may appear beneficial in all patients, by making shock improve more quickly, they may unintendedly cause harm to one group of patients.
These findings are from a small number of patients so need to be confirmed in a larger trial.
To determine if patients with severe septic shock have improved survival if corticosteroids are given based on their SRS pattern compared to the standard approach of giving corticosteroids to all patients.
Patients with severe septic shock will be randomly allocated to receive corticosteroids either using a ‘standard approach’ or an ‘SRS approach’. In the ‘standard approach’ corticosteroids will be given whatever the patient’s SRS pattern, as occurs now. In the ‘SRS approach’ treatment will be targeted. This approach aims to target treatment to give most benefit and avoid possible harm. The doctors looking after the patients will not know the patients’ SRS group or which treatment approach they are receiving.
Patients’ SRS patterns will be determined by measuring seven genes. This will be done using a blood test analysed by a machine on the intensive care unit. This award will allow us to confirm the accuracy of this blood test to determine SRS groups.
We will compare survival rates between the ‘standard approach’ and the ‘SRS approach’. Rates of new infection, duration of shock and quality of life will also be recorded.
This project will deliver the first trial to find out if targeting corticosteroid treatment is better than giving them to all severely ill patients. This type of “Personalised Medicine” study has been seen as a priority for critically ill patients.
Closed Clinical Trials
Levosimendan for the Prevention of Acute oRgan Dysfuction in Sepsis
Sepsis is a major health problem, with some patients developing septic shock. This occurs when infection causes the blood pressure to fall to dangerous levels, reducing blood flow to vital organs. Adrenaline-like drugs are commonly used to improve a patient's blood pressure, but they can have serious side effects.
Levosimendan is a new type of drug that works in a different way to adrenaline-like drugs, potentially avoiding these side effects. It may also decrease inflammation and improve blood flow to vital organs. In small studies in patients who had sepsis, significant improvements were seen in the function of the heart, kidneys and other organs when levosimendan was given. It has therefore been suggested that it may be a useful drug to help patients who have septic shock.
The LeoPARDS trial investigated whether using levosimendan to treat patients who had septic shock improved the function of different organs and improved recovery. The study was funded by the NIHR and the trial was overseen by a Trial Management Committee and an Independent Steering Group.
Thirty-four hospitals across the UK took part in the trial from January 2014 to December 2015. Patients who had septic shock were randomised (assigned by chance) to have either a 24-hour dose of levosimendan or placebo (a liquid that looked like levosimendan, but with no drug in it), as well as normal standard care for sepsis. A total of 515 patients took part; 258 patients in the levosimendan group and 257 in the placebo group. Patients were followed up daily within the ICU for up to 28 days and standard medical records were checked to see how these patients were 3 and 6 months after their ICU stay.
The LeoPARDS Trial found that levosimendan did not improve organ function or any other outcomes when added to standard care in septic shock. Levosimendan is known to dilate blood vessels and this might improve the blood supply to various organs. However, in patients receiving levosimendan higher doses of adrenaline-like drugs had to be given to maintain blood pressure due to the dilation of the blood vessels. The patients then had a faster heart rate and more irregular heartbeats. Patients who required help with breathing from a ventilator required this support for longer.
We cannot recommend the use of levosimendan to treat all cases of septic shock. Although it is disappointing that this study has not shown a benefit from using levosimendan in septic shock, the trial has still provided useful information to help doctors and nurses who look after patients with sepsis. We are very grateful to all the patients, and the families, who took part in the study. Knowing when not to use drugs is important to reduce the risk of side-effects, to avoid wasting precious resources, and to give patients the best chance of recovery from their illness. The results of this study have also highlighted the potential importance of controlling heart rates in sepsis. We are working with colleagues in Birmingham who now plan to study the effects of drugs called beta-blockers in sepsis, to see if it will help patients if we can slow down very fast heart rates.
Vasopressin vs. Noradrenaline as Initial therapy in septic Shock
Sepsis is the body’s response to severe infection. In severe cases of sepsis a drop in blood pressure occurs (septic shock) and this results in poor blood flow to the kidney and other vital organs.
Vasopressin and steroids are both naturally produced hormones that are released during times of severe illness. However when blood pressure drops due to infection, these compensatory mechanism often fail. Studies have shown that administering both of these drugs can help restore blood pressure and reduce the use of other adrenaline-type drugs.
Previous studies found that vasopressin may be most effective if used earlier and for less severe drops in blood pressure and may have a specific role in preventing kidney failure. It may also work best if given in combination with steroids.
We tested these ideas in a clinical trial in 18 general adult intensive care units in the United Kingdom between February 2013 and May 2015. 409 adult patients (average age 66 years) who had septic shock took part.
We found that patients treated with vasopressin required less adrenaline-type drugs to improve their blood pressure.
Our main analysis looked at a combined measure of survival and duration of kidney failure.
We saw that results were very similar between all the different treatment groups. However, in patients treated with vasopressin there was less use of dialysis for acute kidney failure. There was no further improvement when steroids were also added.
Other results, including tests of safety were similar between the different treatments. As expected some patients had poor blood flow to their fingers and toes. This was most common in patients treated with high doses of vasopressin and adrenaline-like drugs.
Vasopressin can be used to treat septic shock. It should not replace the current standard treatment of adrenaline-like drugs but may be useful in some patients to prevent the use of dialysis. Caution should be observed when using high doses of both vasopressin and adrenaline-type drugs and doctors may need to change the blood pressure targets to reduce the risk of side-effects.