Gene Expression Signatures May Help Differentiate Sepsis From Non-Infectious Inflammation Interview with:
Timothy E Sweeney, MD PhD
Resident, General Surgery
Postdoc, Khatri Lab, Bioinformatics
Stanford University

Medical Research: What is the background for this study? What are the main findings?

Dr. Sweeney: Sepsis is defined as the presence of systemic inflammation due to infection. Systemic inflammation can be caused from many things, such as trauma, surgery, thrombosis, autoimmunity, etc. It can also be caused by infection. On the other hand, infection does not necessarily cause systemic inflammation, either:  a person can get a minor infection, like strep throat, and not have a systemic response. It’s the intersection of severe inflammation (a syndrome called SIRS) with infection that defines sepsis.

In general surgery, we frequently see patients after traumatic injury or surgery who are having an inflammatory response (ie, fevers, fast heart rate, high white blood cell count, etc). But it’s not clear whether this inflammatory response is a reaction to the trauma or surgery, or whether there might be an infection brewing that is causing the reaction. Identifying the inflammatory response doesn’t require many special tests– it’s easy to spot. So we know which patients have inflammation and which do not. What is difficult is determining the root cause of the inflammation, and, in particular, whether there is an infection present that needs treatment with antibiotics.

Current diagnostics for infection (not sepsis) are either slow (like blood cultures, which can take 24-72 hours to return) or not highly accurate (like procalcitonin). We sought to define a better test that could specifically differentiate between people with sterile inflammation, and people with inflammation due to infection (sepsis). By integrating gene expression data from multiple publicly available cohorts, we were able to find a set of 82 genes that are significantly differently expressed between these two groups. We then used an algorithm called a greedy forward search to find a subset of 11 genes that were most diagnostic for sepsis.

Medical Research: What should clinicians and patients take away from your report?

Dr. Sweeney: These findings aren’t yet available in a clinical form. We are currently in the process of prospectively testing the diagnostic power of the 11 genes, and of their protein products. Eventually we hope to make a test that a doctor could use to help determine whether a patient has sepsis.

Medical Research: What recommendations do you have for future research as a result of this study?

Dr. Sweeney: One of our key findings is that gene expression is heavily influenced by the recovery process. Thus, future trials should be designed to compare people with and without infections who are at the same point in their hospital course (in other words, compare patients with and without infection at admission, or patients who acquire an infection at hospital day 3 to patients without infection at hospital day 3, etc). In addition, we’ve shown that gene expression signatures in the blood do carry useful diagnostic information even for acute illness. Further work will be necessary to develop robust point-of-care tests that can utilize this information.


A comprehensive time-course–based multicohort analysis of sepsis and sterile inflammation reveals a robust diagnostic gene set

Timothy E. Sweeney1,2,*Aaditya Shidham2 Hector R. Wong34and Purvesh Khatri

Sci Transl Med 13 May 2015:
Vol. 7, Issue 287, p. 287ra71
Sci. Transl. Med. DOI: 10.1126/scitranslmed.aaa5993

[wysija_form id=”3″] Interview with: Timothy E Sweeney, MD PhD (2015). Gene Expression Signatures May Help Differentiate Sepsis From Non-Infectious Inflammation

Last Updated on May 15, 2015 by Marie Benz MD FAAD