“I know the numbers seem small, but we did a sample-size power calculation and it’s just fine,” according to the researcher.
The initial study goal was to confirm earlier promising findings from a study of 370 febrile children in the United Kingdom, Spain, and the United States, conducted by the Immunopathology of Respiratory, Inflammatory and Infectious Disease Study (IRIS) Consortium, a study in which several of Ms. Barral-Arca’s senior coinvestigators participated. The IRIS investigators demonstrated that the combined expression pattern of two genes – IFI44L and FAM89A – distinguished the bacterial from viral infections with impressive sensitivity and specificity (JAMA. 2016 Aug 23-30;316[8]:835-45).
The two-gene signature performed similarly well in Ms. Barral-Arca’s study. However, when she and her coinvestigators tested the discriminatory power of the two genes individually, they got a surprise: The real-time PCR analysis assessing expression of IFI44L alone performed even better than the two-gene combination, discriminating viral from bacterial infections with 91% sensitivity, 93% specificity, and an area under the curve of 94%. In contrast, the two-gene signature based upon IFI44L and FAM89A had a sensitivity of 91%, a specificity of 86%, and an area under the curve of 92%. While those differences in performance are small, a single-gene assay saves time, work, and cost, according to Ms. Barral-Arca.
Her group then validated their findings regarding the performance of the IFI44L single-gene signature in two independent cohorts: stored blood samples from the children in the earlier IRIS study, and a group of children with diarrhea of viral or bacterial etiology.
“One gene seems to be enough,” she said. “We have demonstrated in a real-life scenario that host gene expression microarray data can be successfully translated into a fast, highly accurate, and relatively inexpensive in vitro assay that could be implemented in the clinical routine.”
Planned future work includes investigation of how the gene expression evolves over time from fever onset, the possible utility of the assay in noninfectious febrile illnesses such as rheumatoid arthritis, and whether the test discriminates viral from bacterial infection in adults.
Ms. Barral-Arca reported having no financial conflicts regarding her study, supported by institutional funding.