Predictive Analytics, Blood Samples Determine COVID-19 Severity

August 09, 2021 - Medical University of South Carolina (MUSC) researchers discovered a biomarker in blood samples that with predictive analytics can determine which patients will develop COVID-19 symptoms.

It remains unknown as to why some patients infected with COVID-19 develop severe symptoms while others remain asymptomatic. This uncertainty is alarming to scientists as the Delta variant spreads across the United States.

Through their retrospective study using predictive analytics, the team of MUSC researchers discovered that a reduced level of sphingosine is significantly associated with developing COVID-19 symptoms. Additionally, high levels of sphingosine and acid ceramidase (AC) are associated with asymptomatic infections.

“We developed this project at a time when there wasn’t a successful vaccine,” director of the Lipidomics Shared Resource at Hollings Cancer Center and leader of the Hollings Developmental Cancer Therapeutics Research Program Besim Ogretmen, PhD, said in a press release.

“We wanted to contribute to the field and know which patients who were exposed to this virus would be symptomatic versus asymptomatic.”

Over the past 16 months, the United States has seen more than 35 million cases and almost 630,000 deaths due to COVID-19. Death due to the infection is thought to result from an overactive immune response to the virus in the lungs, causing severe respiratory distress.

However, symptoms vary from person to person, with some developing severe cases and others remain asymptomatic.

Sphingolipids can regulate inflammation and the immune system in response to infections. The Ogretmen lab used their knowledge and expertise from decades of analyzing the production and processing of different lipids to study the COVID-19 patient serum samples from the MUSC COVID-19 Biorepository. The Ogretmen lab researchers looked for changes in the sphingolipid levels.

“Just by looking at the data, you can clearly separate the different patient groups, even without doing technical statistical analyses,” said Alhaji Janneh, lead author and graduate student in the Department of Biochemistry and Molecular Biology.

Asymptomatic patients who tested positive for COVID-19 had a slight increase in serum sphingosine levels compared to those who tested negative. In those who developed symptoms, there was a 15-fold reduction in sphingosine levels.

Conversely, around 75 percent of asymptomatic patients had high AC levels, while most symptomatic patients had no detectable AC. The presence of serum AC correlated with higher levels of sphingosine.

“Can this be an alternative way to predict which patients are the most vulnerable to severe disease?” asked Ogretmen, who is also a professor in the Department of Biochemistry and Molecular Biology and the SmartState Endowed Chair in Lipidomics and Drug Discovery.

“If we can separate asymptomatic patients from symptomatic patients, we can use limited remedies and resources for patients who are more vulnerable.”

Researchers determined there was a 99 percent probability of correctly predicting which COVID-19 patients would develop symptoms by using blood levels of sphingosine.

While analyzing levels of lipids from patient samples is expensive and requires sophisticated equipment, the development of an ELISA-based assay could provide a cost-effective alternative that could be widely implemented.

“There are several outstanding questions remaining. How does vaccination impact sphingosine levels? How do sphingosine levels change with the introduction of more variants? Nevertheless, the ability to identify at-risk patients quickly could vastly improve treatment of COVID-19 and allow for effective distribution of scarce resources,” the press release concluded.