Conquering Sepsis Clinical Trial Design

Designing and operating clinical trials for diagnosing sepsis can be a challenge, as the gold standard for sepsis detection offers a very low prediction rate due to the nature of how sepsis evolves in patients. FDA gold-standard septic blood cultures predict sepsis only 14% of the time, which makes it a very challenging baseline for generating strong predictive data to prove novel diagnostics’ efficacy. However, Immunexpress persisted and overcame these challenges through robust study design. Dr. Rolland Carlson, CEO of Immunexpress, discusses how the company overcame the challenges of implementing sepsis clinical trials. 

Moe Alsumidaie: Can you tell me about the design of the SeptiCyte validation study? What were the study's endpoints?

Dr. Rolland Carlson

Dr. Rolland (Rollie) Carlson: The primary endpoint of the validation study of SeptiCyte RAPID is to differentiate infection-positive (sepsis) from infection-negative systemic inflammatory response (SIRS) for patients suspected of sepsis. In other nomenclature, sepsis is sometimes called IPSI (infection positive systemic inflammation) versus INSI (infection negative systemic inflammation).

Patients suspected of sepsis have to have at least two signs of sepsis, and that can include a fever, a tachycardia, or other clinical assessments. Patients who are suspected of sepsis need to have their blood drawn within 24 hours of an ICU admission order. 

We then measure the blood sample with our SeptiCyte technology on the Biocartis Idylla platform. Our SeptiCyte technology measures the patient’s genomic response to the infection-rather than attempting to find and characterize the invading pathogen via blood culture. By detecting the genetic signal produced by the immune system’s response to infection, our test can quickly and reliably differentiate a patient responding to an infection from someone who is not. The test can be drawn in the emergency department in the hospital. But ultimately the intent is to determine if a patient should be admitted into the ICU. The Idylla platform produces a clinically actionable result from the blood draw in about an hour. 

We then compare our result with the result obtained from a blood culture of the same patient. When there was a positive blood culture result, our results always matched the blood culture. However, due to the imperfections in blood cultures, only 14% of patients with sepsis result in positive blood culture tests. Because of this, we gathered an expert panel of KOLs and physicians from preeminent universities to diagnose patients based on their initial clinical impression. Even though the FDA considered this expert panel to be the "gold standard" for our study, only 68% of the time did all three of the experts concur as to what the result was, whether it's sepsis or SIRS and none of them agreed on the outcome about 10% of the time. 

Additionally, we added some exploratory components to the study, such as looking at the clinical performance of SeptiCyte with or without other biomarkers, and we measured other clinical demographic and lab parameters. 

MA: What clinical trial challenges have you encountered during the validation studies of SeptiCyte RAPID?

RC: The study was designed as a prospective observational, non-randomized clinical study. The challenge of conducting the study is that we were comparing our diagnostic tool against an imperfect standard. 

The “gold standard” of sepsis diagnostics is viewed as a positive blood culture, which can be obtained 24 to 48 hours later. Additionally, even if the patient is septic, a positive blood culture is only observed about 14% of the time. 

In order to overcome the challenge of inaccurate blood cultures, we compiled an expert panel of physicians in the field of sepsis, what we call the retrospective physician diagnosis (RPD) panel. This independent panel (not onsite physicians) assed the diagnosis of sepsis by considering the physician recommendation, as well as other clinical parameters. 

So, there really was no gold standard. We didn’t know what was truly sepsis positive or sepsis negative for patients in the ICU. Even with strong suspicion of sepsis by physicians, major epidemiological studies show that about 44% of them do not have sepsis.

Our goal, in conjunction with the FDA, was to prove that our diagnostic tool was not only a comparable, but a better alternative. So, that was our primary challenge: we were comparing against a standard that was imperfect. We completed this study in 2015 and obtained clearance of SeptiCyte LAB in 2017. 

The goal of the validation study was to replicate the results of SeptiCyte LAB on a faster platform. SeptiCyte RAPID utilizes our SeptiCyte technology on the Biocartis Idylla platform, enabling a sample-to-answer diagnosis in about an hour. 

MA: How were you ensuring there was sufficient demographic diversity in your study?

RC: Our secondary challenge was correctly representing the U.S. demographic by enrolling sufficient minorities in our clinical trial. We conducted a follow-on study to ensure the proper representation of African Americans in our study. Our goal was to accrue samples from additional clinical sites that would have a more ethnic diversity. Those sites included the University of Southern California in Los Angeles, Rush University in Chicago and Grady Memorial Hospital/Emory University in Atlanta. With the addition of these sites, which we refer to as the NEPTUNE study, the mix of patients was representative of the U.S. demographic. 

Our original clinical trial was restricted to adult patients that had been admitted to the ICU within 24 hours. Due to the frantic nature of an ICU patient’s treatment in the first 24 hours of admission, we simplified this sampling by combining the blood draw for the test with the normal draw of the patient for blood culture. We were then able to perform the SeptiCyte test post hoc. 

MA: How did you evaluate and mitigate study risks?  

RC: One of the most important parts of risk mitigation is ensuring there was proper onsite training, as well as trained backup operators for each of these sites. The training and cross-training of operators and principle investigators was led by the director of clinical operations, Dr. Sylvia Cermelli. She was responsible for the decision tree and risk mitigation onsite. She handled all the onsite training and testing with this new platform. 

Due to the simplicity of this new platform, our operator training was much more efficient compared to the old platform. On SeptiCyte LAB, there were at least 40 hands-on steps to run the test. With SeptiCyte RAPID on the Idylla platform, there are only two hands-on steps: the operator pipettes blood into the cartridge, inserts the cartridge into the instrument tray, then walks away and returns in about an hour to receive the results. 

Our first platform required a large training component for operators and resulted in a large potential for operator error. Our second-generation platform basically obviated that, although we still trained each site operator.

MA: What is the value of this diagnostic to physicians and patients? Especially during times, like now with COVID and increasing patient volumes in the ICU?

RC: Our test is designed for people with a strong suspicion of sepsis. However, 44% of the time, those patients suspected of sepsis don’t actually have sepsis. Regardless, all patients suspected of sepsis are admitted to the ICU, and there just aren't enough ICU beds to compensate for these patients. This problem is now exacerbated by the COVID pandemic. So, the utility of our test is extremely valuable right now.

Additionally, patients with COVID are ultimately dying from a virally-induced systemic inflammatory response: this is viral sepsis. When a COVID-positive patient comes into the ER, they are assessed over a series of hours and many times are sent home. This is where our tests could have high utility. Our test can detect whether or not a patient is starting to go down that true sepsis cascade due to the virus. 

Our test is designed to pick up viral induced sepsis, bacterial induced sepsis and fungal induced sepsis. For COVID patients, we're seeing that the cascade starts with viral sepsis, but could lead to a bacterial coinfection. Because we measure the patient’s gene activation to a potential sepsis event, we are not limited to only finding the invading pathogen. Of course, if a COVID-positive patient is determined to have sepsis, he or she would not be discharged, but admitted into the ICU. 

We are modifying our clinical trials right now at the request of the U.S. government to make sure that we can include COVID-positive patients. We've submitted a grant to the U.S. government for accelerated emergency use authorization of our test for COVID triage.

Moe Alsumidaie, MBA, MSF, is a thought leader and expert in the application of business analytics toward clinical trials, and Editorial Advisory Board member for and regular contributor to Applied Clinical Trials.