January 10, 2022
Sepsis is one of the world’s leading killers, accounting for one in every five deaths worldwide, including those caused by severe COVID-19 disease, but it is notoriously difficult to detect early. It is defined as the body’s dysfunctional response to infection and manifests itself through various symptoms, including fever, fatigue, hyperventilation, and a rapid heart rate.
“This novel technique provides unprecedented insight into the biological processes underlying sepsis of any type, including COVID-19,” says Arjun Baghela, a graduate student in the Hancock Lab who led the analysis. “While most people are unaware of sepsis, the number of deaths from life-threatening sepsis is likely to be much higher than one in five in 2020, as almost everyone who has died from COVID-19 has died from sepsis.”
Typically, physicians and healthcare providers cannot know a patient has sepsis until 24-48 hours have passed. The risk of death increases by up to 7.6 per cent for every hour that treatment is delayed, underlining the critical need for quick detection.
“Typically, a patient presents to the emergency room feeling deeply ill,” explains Dr Bob Hancock, a UBC Killam professor. “The physician looks at the patient and says, ‘This is a patient who may have sepsis,’ but only with some certainty can they begin treating them immediately. They’re playing a waiting game for the first 24-48 hours.”
The study showed that severe sepsis could be recognised when a patient first seeks medical attention. The researchers identified sets of genes that indicate whether a patient will develop severe sepsis and made sense of the five various ways in which sepsis shows itself using machine learning, often known as artificial intelligence.
It is now possible to rapidly detect sepsis using machine learning thanks to groundbreaking work done by researchers from UBC’s Hancock Lab and Microbiology and Immunology Department. The work “Predicting sepsis severity at first clinical presentation: the role of endotypes and mechanistic signatures” was published in EBioMedicine.
Dr Nivash Jeevanandam PhD,
Researcher | Senior Technology Journalist