Sepsis is a frequent and sometimes fatal bodily own-goal: a massive over-reaction in which chemicals released into the bloodstream to fight infection instead trigger widespread inflammatory responses.
Just how common sepsis is remains unknown, but one 2017 estimate in the New England Journal of Medicine suggested 30 million annual cases worldwide, resulting in six million deaths. This, however, noted the authors of a subsequent commentary, was likely to be a “significant underestimate” because the authors “could find no data from the low- and middle-income countries where 87% of the world’s population lives”.
Just as disturbing is the finding that in countries where adequate data do exist, sepsis is misdiagnosed in about 30% of cases. Even in the instances where initial diagnosis is correct, reaching the conclusion can take several days.
Now, however, researchers led by Daniel Irimia from Massachusetts General Hospital in the US have invented a novel diagnostic platform that can produce a result in just hours from a single drop of blood.
The method combines a device full of microscopic channels – known as a microfluidic array – with a machine-learning algorithm.
The algorithm assesses the activity of neutrophils – the most abundant type of white blood cells – to calculate a “sepsis score”, allowing for a much more rapid diagnosis.
To test the efficiency of the new method, Irimia and colleagues tried it out on 42 patients divided into two cohorts. It returned results with more than 95% accuracy.
The researchers now intend to test the method on a larger population of at-risk patients to better assess its viability.
The research is published in the journal Nature Biomedical Engineering.
