Don’t judge the book by its cover when managing young infant with fever

March 2005

NEW YORK — When it comes to deciding management of a young infant with fever, you really can’t tell a book by its cover. The child often will look well yet still have a bacterial disease, warned M. Douglas Baker, MD, at the 17th Annual Infectious Diseases in Children Symposium here.

“If you are going to practice by looking at the child first and then deciding what you do afterward, please take this to heart: It doesn’t matter how many years you’re in practice. I’ve been in practice now for 22 years, and I can’t pick out the kids who have bacterial disease by looking at them. It’s clear that infants younger than 2 months of age with fever and bacterial disease can look good. They might not feel good, but they look good. And I think that our practice demands that you still go hunting for a bacterial source of fever,” said Baker, who is professor of pediatrics at Yale University School of Medicine in New Haven, Conn.

“If you go hunting in infants younger than 2 months of age with fever, about 10% of the time you will come up with something that resembles a bacterial infection, be it bacteremia, bacteriuria, pneumonia or bacterial meningitis,” added Baker, who is also chief of pediatric emergency medicine at Yale-New Haven Children’s Hospital.

In young children, the approach to fever management tends to be conservative, especially regarding infants younger than 1 month of age.

“It is important to be aggressive when investigating the cause of fever, and carefully consider hospitalization and treatment with empirical antibiotic therapy for children in this very young age range, as opposed to those who are older,” said Baker, a member of the editorial advisory board of Infectious Diseases in Children.

Appearances can be deceiving

A recent article by the PROS Network said that the clinical appearance of the child makes a difference. Baker does not agree: “It does if they’re sick, but it doesn’t if they’re well appearing.”

	“I’ve been in practice now for 22 years, and I can’t pick out the kids who have bacterial disease by looking at them.” — M. Douglas Baker, MD

Data from Philadelphia that were published in 1993 looking at about 750 infants with fever younger than 2 months of age found that two-thirds of infants with serious bacterial disease proven by positive cultures looked absolutely well.

Most pediatricians want to know what the work-up of these children should be and whether they can go home or must be hospitalized. Large prospective, well-designed, well-executed studies done in Boston, Philadelphia and Rochester have attempted to tease out a way to predict which child will have serious disease, but they haven’t been entirely successful.

“If you’re compulsive about looking at 1- to 3-month-old infants with fevers of 38°C or higher, these are the definitions acceptable for discharge home: a white blood cell (WBC) count less than 20,000, urinalysis less than 10 cells per high-powered field on a spun specimen, spinal fluid that was clear and a normal-looking chest x-ray. If you met all those criteria, then you were low risk,” Baker said of the study out of Boston. “You went home; you were treated with ceftriaxone. Ninety-five percent of the time, those infants whom they thought were low risk were indeed free of bacterial disease. Only 5% of the time were they wrong about being low risk for bacterial disease; those infants came back for treatment, did well and subsequently recovered uneventfully.”

The Philadelphia study was done in 1- to 2-month-olds with documented temperatures of 38.2°C or higher. Acceptable was defined as WBC count less than 15,000, a band-to-neutrophil ratio of less than 20%, a urinalysis much like the Boston criteria, spinal fluid much like the Boston criteria and a clean chest x-ray.

“If you use those low-risk criteria, which differed from the Boston criteria primarily because the WBC in the peripheral smear was lower at 15,000 and the band-to-neutrophil ratio was added, you’ll find that in those with bacterial disease, most (99%) of the time they were identified prospectively as low risk,” he said.

Some investigators say a chest x-ray is not needed if the infant with fever is young and is free of signs or symptoms of lower airway disease.

“Free of signs or symptoms of lower airway disease means that you ... have no cough, no rhonchi and no wheezes on auscultation. You have no rales or crackles. You have no grunting, no nasal flaring. You have no cough and you have no runny nose. Does anybody know an infant who looks like that?” Baker asked.

In one study of 12 children with positive chest x-rays, five had none of those signs or symptoms. In another study of 27 children with positive chest x-rays, two had none of those signs or symptoms. And finally, in another of 36 children with positive chest x-rays, five had none of those signs or symptoms.

“Now, I’m not about to tell you that the five, the two and the five were straightforward bacterial diseases. They probably were not. They might have been either, or a combination of viral and bacterial disease,” he said.

LP need?

One of the procedures that physicians would like to avoid is a lumbar puncture (LP), and the criteria do try to assess the need for an LP.

“If you’re in private practice, and you see over the course of an entire career 200 young infants, younger than 2 months of age, with fever, the chances are that you will not come across a case of bacterial meningitis. If you see 201, you’re going to come across one, and that’s about the odds. About one out of 200 or maybe one out of 250 in that age range with fever will have bacterial meningitis,” Baker said.

“If you’re looking for aseptic meningitis, you’re going to find it quite commonly — 10%, 11%, 12%, 13% of big cohorts have aseptic meningitis. The Rochester low-risk criteria call for a history and physical exam, a peripheral white blood cell count with a differential count, a stool smear and a urinalysis, and if all those things look like they’re acceptable to you, then you need not get the LP.”

The peripheral WBC count is one of the key elements for determining the need for an LP, according to the criteria. A well-appearing child with fever but a complete blood count (CBC) of between 5,000 and 15,000 is probably not going to get an LP.

The Boston study looked at 5,300 consecutive cerebrospinal fluid samples in children between 3 and 89 days of age with fever. They found 22 cases of bacterial meningitis in the group: four per 1,000 or one per 250. Forty percent of the cases had a WBC count between 5,000 and 15,000.

“This is a red flag. If you have an infant who has fever, and you do a peripheral WBC count that comes back lower than 5,000, beware. Look at that child very carefully before you consider not doing an LP, because in fact one-third of the cases of bacterial meningitis in this data set had children who had CBCs with less than 5,000 peripherally.”

New markers

Two new laboratory markers are being investigated that may be better at indicating the likelihood of serious bacterial disease in febrile infants: C-reactive protein and procalcitonin. These are both acute-phase reactants that are generated in the liver. Early studies are finding elevations of these markers in infants with fever.

“C-reactive protein takes about four or six hours or so after this stimulus to be produced,” Baker said. “In fact, it does not peak until about 36 hours out, which is bad for us in the neonatal fever business, because typically we’re seeing these infants well in advance of a 36-hour duration of fever at home.”

Procalcitonin reacts and peaks faster than C-reactive protein, Baker said, “so that might be where some of the money will lie in terms of usefulness in predicting bacterial disease, at least invasive bacterial disease, in these very young infants with fever.”

For more information:

• Baker MD. Common questions regarding the management of fever in infants. Presented at the 17th Annual Infectious Diseases in Children Symposium. Nov. 20-21, 2004. New York.