Researchers have shown that it is possible to diagnose a bacterial infection from a small sample of blood—based on the immune system’s response to the bacteria—in infants with fevers who are 2 months of age or younger.

With additional research, the new technique could be an improvement over the standard method, which requires isolating live bacteria from blood, spinal fluid, or urine, and growing them in a laboratory culture. The study, funded in part by the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), appears in the August 23 issue of the Journal of the American Medical Association.1

Healthcare providers who evaluate young infants with fevers have limited means to diagnose quickly and accurately whether an illness results from a bacterial infection. Determining if an illness is caused by bacteria may involve complicated medical procedures, such as a lumbar puncture. While they wait for the test results, physicians also may need to admit the infant for a lengthy hospital stay or prescribe antibiotics, which may later turn out to be unnecessary.

“The development of a fast and noninvasive diagnostic tool holds promise for better outcomes and lower treatment costs for young infants with fevers of unknown cause,” says Valerie Maholmes, PhD, chief of NICHD’s pediatric trauma and critical illness branch.

With advances in genetic sequencing technology, researchers have explored alternative approaches to diagnosing infections, such as assessing the body’s immune response. When the immune system wards off an infection, immune cells activate certain genes, depending on whether the infection results from bacteria or viruses. Collectively, these distinct genes form what is called a “biosignature.”

“Previous studies have suggested that analyzing immune cell biosignatures can distinguish between bacterial and viral infections in children and adults,” says Ruth Brenner, MD, MPH, an NICHD program officer. “However, scientists were uncertain if this approach would work in young infants because of their immature immune systems.”

In the current study, researchers in the Pediatric Emergency Care Applied Research Network (PECARN) sought to define a group of genes that could serve as biosignatures in infants. The study team enrolled infants 2 months of age or younger from 22 emergency rooms. More than 1,800 infants with fevers were enrolled. For the preliminary study, the researchers tested samples from 279 randomly selected infants (89 with bacterial infections, 190 without bacterial infections), along with samples from 19 healthy infants without fevers.

The PECARN team identified a biosignature of 66 genes, measured in a small blood sample, from which they could distinguish between newborns with or without bacterial infections. These serious bacterial infections included blood infections (bacteremia), brain or spinal fluid infections (bacterial meningitis), and urinary tract infections, all of which are typically diagnosed by growing bacteria from samples of body fluids. In the study, the biosignature was 87% sensitive—correctly identifying a sample as positive. The researchers also found a group of 10 genes signifying bacteremia, a very serious condition. This biosignature correctly identified a positive sample 94% of the time.

When compared to the traditional culture method, preliminary findings indicate that biosignatures could ultimately lead to a fast and noninvasive test for diagnosing bacterial infections in infants with fevers. More work is needed, however, to optimize the diagnostic approach.

The study was led by investigators from the Children’s Hospital of Michigan, Nationwide Children’s Hospital, and UC Davis Medical Center. In addition to NICHD, funding was provided by the Health Resources and Services Administration and the Maternal and Child Health Bureau within the US Department of Health and Human Services.


  1. Mahajan P, Kuppermann N, Mejias A, et al. Association of RNA biosignatures with bacterial infections in febrile infants aged 60 days or younger. JAMA. 2016;316(8):846–857; doi: 10.1001/jama.2016.9207.