A genetic signature in newborns can predict neonatal sepsis before symptoms even start to show, according to a new study.
The study, led by UBC and SFU researchers in collaboration with the Medical Research Council (MRC) Unit The Gambia, has the potential to help healthcare workers diagnose babies earlier, including in lower- and middle-income countries (LMICs) where neonatal sepsis is of particular concern. The research, published today in eBiomedicine , is funded by the National Institutes of Health and the Canadian Institutes of Health Research.
"Neonatal sepsis is caused by the body's irregular response to a severe infection that occurs within the first 28 days of life. Globally, it affects around 1.3 million babies annually, and unfortunately, in LMICs, those rates are higher," said first author Andy An, a UBC MD/PhD student who completed the research as a doctoral student in the department of microbiology and immunology. "Even when treatment is successful, sepsis can have lifelong effects because it can lead to developmental delay in children, imposing cognitive deficits and long-term health issues. By recognizing it as early as possible, we can treat infants promptly and ideally, head off these harms."
Neonatal sepsis causes an estimated 200,000 deaths worldwide each year, with the highest rates in LMICs. In Canada, the risk is lower at about one in 200 live births, but higher in prematurely born babies.
Rolling the dice on health
Diagnosing sepsis is challenging for doctors and families. The symptoms can look like many other illnesses, and tests to check if sepsis is present can take several days, aren't always accurate, and are largely only available in hospitals. The uncertainty can delay urgent treatment with antibiotics.
"Knowing that sepsis is impending would also allow physicians more time to determine the appropriate treatment to use," said co-senior author Dr. Bob Hancock, professor in the UBC department of microbiology and immunology. "The consequences of neonatal sepsis are so severe in the most vulnerable individuals that providing an early diagnosis to assist and guide physicians could save lives."
Equitable access to healthcare
The researchers participated in a large study in The Gambia where blood samples were taken from 720 infants at birth. Out of this cohort, 15 babies developed early-onset sepsis.
The researchers used machine learning to map the expression of genes active at birth, in search of biological markers that could predict sepsis.
"We found four genes that, when combined in a 'signature', could accurately predict sepsis in newborns nine times out of 10," said co-senior author Dr. Amy Lee, assistant professor in the SFU department of molecular biology and biochemistry. "This was a unique opportunity where samples were available from all babies in this cohort on the day of birth, meaning we could study the genes expressed in the sepsis babies before they got sick. Most other studies have only published markers that were present when the babies were already ill, and this would therefore not be a predictive signature."
"The early recognition of sepsis is vital for infants' survival, and identifying markers that might allow us to 'predict' babies at particular risk would be an enormous advantage, since we could then target specific surveillance and treatment of such infants," said Dr. Beate Kampmann, who led the clinical component of the study at the MRC Unit in The Gambia.
The researchers hope the signature will one day be incorporated not only into PCR tests in hospitals, but also in portable, point-of-care devices.
"There are point-of-care devices available that can test for gene expression, for instance, COVID-19 and influenza, with a single drop of blood. They can operate anywhere with a power source including batteries and can be used by anyone, not just trained healthcare providers," Dr. Hancock. "These portable devices could be retooled to recognize this 'signature' relatively easily and inexpensively."
The next step for the research would involve a large prospective study to show the signature is successful at predicting sepsis in other populations and prove its methodology, and then the development of point-of-care tools for approval by relevant government bodies.
Interview language(s): English (An, Hancock, Kampmann, Lee)
