It's Just a Virus! Viral Etiologies of Fever Without a Source

Abstract

Colvin JM, Muenzer JT, Jaffe DM, et al. Detection of viruses in young children with fever without an apparent source. Pediatrics. 2012; 130(6): e1455-e1462; doi: 10.1542/peds.2012-1391Investigators from Washington University, St. Louis and Biological Mimetics, Inc., Frederick, MD conducted a study to better understand the viral etiologies of fever without an apparent source (FWS) in young children. Children aged 2 to 36 months who presented to the emergency department (ED) at St. Louis Children’s Hospital with a temperature of ≥38°C without an apparent source and who had blood obtained for clinical management were enrolled. Children with symptoms suggestive of a viral respiratory infection or who had a positive rapid test for influenza were excluded. Two comparison groups of children were also enrolled: (1) children aged 2 to 36 months who presented to the ED with a temperature of ≥38°C but had a probable or definite bacterial infection, and (2) afebrile children aged 2 to 36 months who were having an ambulatory surgical procedure. Blood and/or nasopharyngeal (NP) specimens were collected from study children. Polymerase chain reaction (PCR) testing was used to identify viruses. The main outcome was the type and frequency of viruses detected in each study group. Secondary outcomes were antibiotic utilization among children with FWS and demographic and clinical variables associated with viral infection.A total of 206 children were enrolled: 75 with FWS, 15 with fever and a probable or definite bacterial infection, and 116 who were afebrile. All children with bacterial infection and 29% of children with FWS were hospitalized. Viruses were detected in 76% of children with FWS, 40% with bacterial infection, and 35% who were afebrile (P < .001). The viruses identified most frequently in children with FWS were adenovirus, human herpesvirus-6 (HHV-6), and enterovirus. Each was detected significantly more frequently in those with FWS (identified in 57%) than in those with bacterial infection (13%) or who were afebrile (7%). Epstein-Barr virus was the most commonly detected virus in children with bacterial infection (23%) and rhinovirus was the most commonly detected in afebrile subjects (16%). Children with FWS were significantly more likely to have >1 virus identified and to have a virus detected from their blood or NP specimen than children in the bacterial or control groups. Among children with FWS, 45% received antibiotics, including 91% of those who had adenovirus. The only significant demographic or clinical characteristic of children who had adenovirus, HHV-6, or enterovirus infection compared to children with no virus detected was that the white blood count was more likely to be >15,000 cells/mm3 in children who had adenovirus (P = .003).The authors conclude that viruses are frequently detected in children aged 2 to 36 months with FWS and that use of blood specimens for virus detection could increase virus identification and decrease empiric antibacterial treatment.This important study has several limitations. First, the investigators did not enroll all eligible children with FWS but only those who presented to the ED when study personnel were working. This introduces selection bias. Second, the inclusion criterion requiring that blood work be ordered by the treating provider as part of the clinical work-up introduces additional bias, as those with FWS for whom no blood work was ordered were excluded. Third, detection of a virus does not mean that the virus was responsible for the child’s fever, although the viruses most frequently identified in the febrile children are typically thought to be pathogenic (and many of those found in the afebrile cohort are thought to have low pathogenicity). Finally, virus detection was limited to blood and NP specimens in this study. Notably, stool specimens, which may increase virologic yield, were not included.The management of FWS has evolved along with the vaccination landscape: occult bacteremia has essentially disappeared in highly-vaccinated populations.1 Increasingly, febrile infants and young children with an unrevealing physical examination may have either a urinary tract infection2 or “a virus.” While worried parents may balk at the “it’s just a virus” diagnosis, this study gives it credence. Of note, the yield of virus-specific diagnoses increased from 49% with the use of PCR on NP specimens (which is the standard of care in most pediatric hospitals) to 76% when both NP and blood were assayed by PCR. These findings suggest, among other things, that the application of new technologies will improve our ability to render a specific diagnosis for febrile illnesses in young children. (See also AAP Grand Rounds, June 2012;27[6]:65.3)