Life-threatening systemic inflammation known as sepsis can follow infection with SARS-CoV-2 (shown in green in this colorized electron micrograph), the virus that causes COVID-19. Image: NIAID
New research suggests that the virus responsible for COVID-19 was a more common and deadly cause of sepsis early in the pandemic than previously assumed — accounting for about one in six cases of sepsis from March 2020 to November 2022.
The results, published online Sept. 29 in JAMA Network Open, suggest that clinicians should rethink how they treat sepsis while also providing a framework for future surveillance of viral sepsis.
Sepsis is a serious, sometimes fatal overreaction of the immune system to an infection. Doctors and researchers don’t know as much about sepsis that occurs in response to viral infection as they do about sepsis that arises from bacterial infection.
“Most people, including medical professionals, equate sepsis with bacterial infections,” said first author Claire Shappell, HMS instructor in medicine at Brigham and Women’s Hospital. “This is reflected in treatment guidelines and quality measures that require immediate antibiotics for patients with suspected sepsis.”
“However,” she said, “viral infections, including SARS-CoV-2, which causes COVID-19, can trigger the same dysregulated immune response that leads to organ dysfunction as in bacterial sepsis.”
To capture a fuller and more accurate picture of viral sepsis cases, investigators from Harvard Medical School and Brigham and Women’s analyzed electronic health record (EHR) data from five Mass General Brigham hospitals to track the rate of sepsis associated with SARS-CoV-2 during the pandemic.
Previous efforts to quantify SARS-CoV-2-associated sepsis were limited by inconsistent definitions of viral sepsis and under-recognition of cases, said senior author Chanu Rhee, HMS associate professor of population medicine and of medicine at Brigham and Women’s.
Rhee and colleagues knew from their previous work that EHRs could provide more accurate estimates of sepsis incidence and outcomes than administrative data sets. Rhee said the new study marks the first time researchers have used the method to study viral sepsis.
COVID and sepsis by the numbers
The team quantified the incidence and mortality for SARS-CoV-2-associated sepsis using clinical criteria adapted from the U.S. Centers for Disease Control and Prevention, which incorporated positive SARS-CoV-2 tests and clinical signs of organ dysfunction.
Using EHR data from the first two and a half years of the pandemic, the team identified 431,017 hospitalizations from 261,595 individuals.
During that time, 5.4 percent of hospitalizations were due to SARS-CoV-2 infections. Of those, 28.2 percent had SARS-CoV-2-associated sepsis.
The mortality rate for patients with SARS-CoV-2-associated sepsis was initially high: 33 percent over the first three months of the pandemic.
However, the rate declined over time and eventually became similar to the mortality rate for presumed bacterial sepsis: about 14.5 percent. That rate then remained stable throughout the study period.
Health record analysis as a model for studying viral sepsis
The study’s design and use of EHR data provides a framework for future research into sepsis associated with other viruses, including influenza (flu) and respiratory syncytial virus (RSV).
The team hopes to apply this method to larger and nationally representative data sets to report generalizable epidemiologic data on viral sepsis.
“Our study draws attention to the high burden and poor outcomes associated with viral sepsis, while demonstrating the utility of using EHR-based algorithms to conduct surveillance for both viral and bacterial sepsis,” said Shappell.
“We also hope our findings highlight that sepsis is not a ‘one-size-fits-all’ entity, but one that requires clinicians to tailor their diagnosis and treatment strategy to each patient’s syndrome and probable pathogen,” she said.
Authorship, funding, disclosures
Additional authors are Michael Klompas, Christina Chan, Tom Chen, Sanjat Kanjilal, and Caroline McKenna.
This study was funded by the CDC (grant U54CK000484) and National Institutes of Health (grant 1F32GM143862).
Klompas reported receiving grants from the Agency for Healthcare Research and Quality and personal fees from UpToDate outside the submitted work. Rhee reported receiving personal fees from UpToDate and Cytovale and grants from the Agency for Healthcare Research and Quality outside the submitted work.
