NEW YORK (GenomeWeb) – A trio of pediatricians and emergency medicine doctors has received a five-year, approximately $5.8 million grant from the National Institute of Child Health and Human Development to develop a faster and more precise method to rule out bacteria as the cause of fever in young babies.
The physician-investigators hope to eventually develop the method into a PCR-based diagnostic test that could possibly be used at the point of care.
Every baby less than one month old brought to the emergency room with a fever of 100.4 degrees Fahrenheit or higher gets a lumbar puncture to sample spinal fluid for meningitis causing-bacteria. About half of all babies younger than two months with fevers will also be admitted to the hospital for at least the 48 hours it takes to demonstrate by culture methods that they don't have bacterial infections.
Yet, of the 500,000 or so babies in the US per year subjected to this invasive and costly fever protocol, fewer than 10 percent actually wind up having bacterial infections, and the majority of those are of the urinary tract.
The grant recipients  — Prashant Mahajan at Wayne State University School of Medicine; Octavio Ramillo at Nationwide Children's Hospital and Ohio State University; and Nathan Kuppermann at University of California Davis — have been working together for more than a decade to develop a method that might help reduce such unnecessary testing and possibly identify the cause of fever.
In the initial phases of their study, the investigators also collaborated with 21 pediatric emergency rooms that support a research consortium called the Pediatric Emergency Care Applied Research Network, or PECARN, to collect and analyze gene expression in samples from thousands of babies with fevers.
"What we demonstrated was that infants with a bacterial infection had an over-expression of the inflammatory genes while infants with viral infections had an over-expression of the interferon genes," Wayne State's Mahajan said.
The details of the genes and gene networks are the subject of a journal article currently under review, he noted.
The three co-PIs published a Viewpoint article in JAMA Pediatrics two years ago arguing the advantages of host biosignatures for diagnosing infection in febrile infants. Ramillo also published last year in Current Opinion in Infectious Diseases on the "paradigm shift" of detecting specific types of infections using host responses; and the researchers together also published the results of a prospective feasibility study earlier this year in Pediatric Emergency Care, detailing how they managed to get 4,000 good quality RNA samples from the PECARN sites.
In an interview Ramillo also noted that a host biosignature which the group previously obtained in children with influenza was matched to a signature in experimentally infected adults and used to develop an RT-PCR assay, as covered by GenomeWeb.
In the context of clinical infectious disease diagnosis, the idea of looking at the host and not the pathogen is new, but the method is already used successfully in the fields of rhematology and oncology, he added.
The current funding is for bacterial infections, but Ramillo noted that the collaborators have been exploring this in other clinical situations as well.
"Even in the context of viral infections that don't cause viremia, we can see very good patterns — for instance with influenza, [respiratory syncytial virus], or rhinovirus," Ramillo said.
"We can see the common patterns and distinguish [infection with] viruses or bacteria, but then when you go in-depth you can see differences between individual viruses and bacteria as well," he said.
The potential diagnostic test the team is developing will initially tell a clinician whether a baby has an infection that is bacterial or non-bacterial.
"The implications for invasive procedures and hospitalization of these vulnerable infants is substantial," Kuppermann told GenomeWeb in an email.
Mahajan concurred, suggesting that, with a relatively fast test to rule in, or out, bacterial infection, "the parents are not so stressed and they don't have to take time off, and kids don't have to get invasive procedures or antibiotics."
Mahajan said that the team is considering developing the assay into a point-of-care or bedside test that is PCR-based, possibly a microfluidic chip of some variety, and that they aspire to a turn-around time of four to six hours.
Distinguishing the biosignature would require relatively small numbers of gene probes, perhaps 15 or 20, Ramillo said.
And in the future the method may be useful to separate asymptomatic carriers of a particular pathogen or to confirm suspected contaminants in PCR. It might also be used to predict severity of disease, as Ramillo and other co-authors showed in a recent PLoS Medicine study of RSV.
At the moment the researchers are also still characterizing the biosignatures, using Illumina gene expression arrays and whole-genome sequencing, and not in real time.
The three researchers asserted that the validation of previous results in a larger sample is the critical next step. The current accuracy is as high as 95 percent compared to blood or urine culture. But they would like it to be even higher.
"Meningitis is very uncommon, but the consequences of missing even one case are devastating," Mahajan said.
The researchers also emphasized the collaborative nature of the funding. "Everything is about 'team' and 'us,'" Ramillo said, adding that the collaboration of all the staff in the PECARN network is also a very important part, since "enrolling these babies, so young, with such stressed parents in the emergency department, requires a lot of effort by everybody."
"It is a huge investment by NIH, but also a huge investment by many individuals to make this happen," he said.