Big Data Reveals Hidden Subtypes of Sepsis

sepsis is not all the same

Much like cancer, sepsis isn’t simply one condition but rather many conditions that could benefit from different treatments, according to the results of a University of Pittsburgh School of Medicine study involving more than 60,000 patients. McGowan Institute for Regenerative Medicine affiliated faculty members who are co-authors on the study include:

Derek Angus, MD, MPH, Chair of the Department of Critical Care Medicine of both the University of Pittsburgh School of Medicine and the UPMC Healthcare System. At the University, he holds the rank of Distinguished Professor and the Mitchell P. Fink Endowed Chair in Critical Care Medicine with secondary appointments in Medicine, Health Policy and Management, and Clinical and Translational Science and he directs the CRISMA (Clinical Research, Investigation, and Systems Modeling of Acute Illnesses) Center. He also co-directs the UPMC ICU Service Center, responsible for the provision of ICU services across the 20-plus hospital system.
Gilles Clermont, MD, Professor of Critical Care Medicine, of Industrial Engineering, and of Mathematics at the University of Pittsburgh. He is also the Medical Director of the Center for Inflammation and Regenerative Modeling, University of Pittsburgh. Dr. Gilles also serves on the Critical Care Medicine Attending Staff at hospitals of the University of Pittsburgh Medical Center and the Veterans Affairs Medical Center, Pittsburgh.
John Kellum, MD, FACP, FCCM, Professor in the Departments of Critical Care Medicine (primary), Medicine, Bioengineering, and Clinical and Translational Science at the University of Pittsburgh. He is also the Director of the Center for Critical Care Nephrology and the Vice-Chair for Research, both appointments in the Department of Critical Care Medicine, and Associate Director for Acute Illness in the Institute for Personalized Medicine at Pitt. Dr. Kellum also serves as an Intensivist at UPMC.
Yoram Vodovotz, PhD, Professor in the Department of Surgery with secondary appointments in the Department of Computational & Systems Biology, the Department of Bioengineering, the Department of Immunology, the Department of Communication Science and Disorders (of the School of Health and Rehabilitation Science), and the Clinical and Translational Science Institute. He also is the Director of the Center for Inflammation and Regenerative Modeling at the McGowan Institute.

These findings, announced in JAMA and presented at the American Thoracic Society’s Annual Meeting, could explain why several recent clinical trials of treatments for sepsis, the No. 1 killer of hospitalized patients, have failed. Sepsis is a life-threatening condition that arises when the body’s response to an infection injures its own tissues and organs.

“For over a decade, there have been no major breakthroughs in the treatment of sepsis; the largest improvements we’ve seen involve the enforcing of ‘one-size fits all’ protocols for prompt treatment,” said lead author Christopher Seymour, MD, MSc, associate professor in Pitt’s Department of Critical Care Medicine and member of Pitt’s Clinical Research Investigation and Systems Modeling of Acute Illness Center. “But these protocols ignore that sepsis patients are not all the same. For a condition that kills more than 6 million people annually, that’s unacceptable. Hopefully, by seeing sepsis as several distinct conditions with varying clinical characteristics, we can discover and test therapies precisely tailored to the type of sepsis each patient has.”

In the “Sepsis ENdotyping in Emergency Care” (SENECA) project, funded by the National Institutes of Health (NIH), Dr. Seymour and his team used computer algorithms to analyze 29 clinical variables found in the electronic health records of more than 20,000 UPMC patients recognized to have sepsis within six hours of hospital arrival from 2010 to 2012.

The algorithm clustered the patients into four distinct sepsis types, described as:

Alpha: most common type (33%), patients with the fewest abnormal laboratory test results, least organ dysfunction and lowest in-hospital death rate at 2%;
Beta: older patients, comprising 27%, with the most chronic illnesses and kidney dysfunction;
Gamma: similar frequency as beta, but with elevated measures of inflammation and primarily pulmonary dysfunction;
Delta: least common (13%), but most deadly type, often with liver dysfunction and shock, and the highest in-hospital death rate at 32%.

The team then studied the electronic health records of another 43,000 UPMC sepsis patients from 2013 to 2014. The findings held. And they held again when the team studied rich clinical data and immune response biomarkers from nearly 500 pneumonia patients enrolled at 28 hospitals in the U.S.

In the next part of the study, Dr. Seymour and his team applied their findings to several recently completed international clinical trials that tested different promising therapies for sepsis—all of which had ended with unremarkable results.

When trial participants were classified by the four sepsis types, some trials might not have been failures. For example, early goal-directed therapy (EGDT), an aggressive resuscitation protocol that includes placing a catheter to monitor blood pressure and oxygen levels, delivery of drugs, fluids and blood transfusions was found in 2014 to have no benefit following a five-year, $8.4 million study. But when Dr. Seymour’s team re-examined the results, they found that EGDT was beneficial for the Alpha type of sepsis patients. Conversely, it resulted in worse outcomes for the Delta subtype.

“Intuitively, this makes sense—you wouldn’t give all breast cancer patients the same treatment. Some breast cancers are more invasive and must be treated aggressively. Some are positive or negative for different biomarkers and respond to different medications,” said senior author Dr. Angus. “The next step is to do the same for sepsis that we have for cancer—find therapies that apply to the specific types of sepsis and then design new clinical trials to test them.”

Additional authors on this research publication are Jason N. Kennedy, MS, Shu Wang, MS, Chung-Chou H. Chang, PhD, Zhongying Xu, MS, Hernando Gomez, MD, MPH, David Huang, MD, Qi Mi, PhD, Victor Talisa, MS, Shyam Visweswaran, MD, PhD, and Donald M. Yealy, MD, all of Pitt; Corrine F. Elliott, MS, and Scott Berry, PhD, both of Berry Consultants in Texas; Steven M. Opal, MD, of Rhode Island Hospital; Tom van der Poll, MD, PhD, of Pitt and the University of Amsterdam; Jeremy C. Weiss, MD, PhD, of Carnegie Mellon University; and Sachin Yende, MD, MS, of Pitt and the VA Pittsburgh Healthcare System.

Illustration: New findings from the University of Pittsburgh School of Medicine reveal that sepsis is not just one disease. Credit: Speed Drawing Video.