This article is more than 5 years old. Information may no longer be current.
T2Bacteria Panel demonstrates ‘excellent’ sensitivity, specificity
Click Here to Manage Email Alerts
ATLANTA — Researchers reported at ASM Microbe that the T2Bacteria Panel “demonstrated excellent performance” in detecting bacteremia in patients compared with blood culture.
“The T2Bacteria Panel had excellent sensitivity and specificity at detecting leading causes of bloodstream infection,” Cornelius (Neil) J. Clancy, MD, assistant professor of medicine and director of the mycology program at the University of Pittsburgh, told Infectious Disease News.
T2Biosystems announced in late May that it received clearance from the FDA to market the test in the United States. Researchers said the T2Bacteria Panel — an automated, rapid, culture-independent diagnostic test — can detect bacteremia from whole blood in 3 to 5 hours, whereas blood culture can take 2 to 3 days to turn over results.
Researchers said the test targets five organisms that are responsible for at least 50% of bloodstream infections, including Escherichia coli, Enterococcus faecium, Klebsiella pneumoniae, Pseudomonas aeruginosa and Staphylococcus aureus. It runs on the FDA-cleared T2Dx Instrument, which can also be used to test and monitor Candida auris infections with the T2Candida Panel.
In a prospective study conducted in 11 hospitals across the U.S. — headed by Eleftherios Mylonakis MD, of Brown University, and M. Hong Nguyen MD, of the University of Pittsburgh — researchers compared the T2Bacteria Panel with blood culture for diagnosing bacteremia using 1,427 blood samples from patients aged at least 18 years. Based on blood culture results, the test’s overall sensitivity was 90%. Its specificity was 96% to 99% when blood culture was used as the gold standard comparator, and 98% to 100% when a composite of clinical and microbiological criteria was used.
Based on blood culture results, the false-positive rate for the T2Bacteria Panel was 11%. However, blood culture is a suboptimal gold standard diagnostic, according to Clancy, who was a co-investigator on the study.
“The sensitivity of blood cultures is 50% or less,” he said. “If you know your blood culture is a suboptimal gold standard, how do you interpret your new diagnostic test-positive, blood culture-negative results? Are they true negatives — suggesting that you’re getting false-positive results with your diagnostic — or are they in fact true positives?”
Clancy said tests based on nucleic acid amplification and detection (NAATs) like the T2Bacteria and T2Candida panels work particularly well on patients who are already receiving antimicrobial therapy. He said nonviable or latent organisms that are being killed or suppressed by an antimicrobial can still be detected by these assays.
“My guess is that a significant percentage of those T2-postivie, blood culture-negative results are in fact true positives,” he said. “We know from the emerging T2Candida literature that patients with these discordant results have outcomes that are similar to those of patients with blood culture-positive candidemia, suggesting the results have biologic significance. On balance, patients with discordant T2Bacteria-positive/blood culture-negative results probably merit antimicrobial therapy against the pathogen detected, if they have signs and symptoms that are consistent with bloodstream infection.”
Among patients with these types of discordant results in the study, 70% had findings consistent with infection, 52% had previously received antibiotics, and 40% had the bacteria identified by T2bacteria recovered from a site other than the blood or from the blood at a different time point.
Among patients with samples that tested positive using both methods, 66% would have potentially benefited from earlier initiation of antibiotics if the T2Bacteria test had been used, according to the researchers.
“I think the challenge for this, and for all of these new rapid diagnostics coming out, is translating the nuts and bolts of sensitivity, specificity and clinical trial performance into algorithms that you can actually use and interpret for patients and drive care with them,” Clancy said. “The panel and machine work. They work very well for what they do. The question is now, how can clinicians use it and make a difference in the care for their patients?” – by John Schoen
Reference:
Nguyen M, et al. Abstract CPHM LB6. Presented at: ASM Microbe; June 7-11, 2018; Atlanta.
Disclosures: The authors report no relevant financial disclosures.