In ancient Rome, poisoning antidotes were sometimes written to be
taken with a grain of salt (cum grano salis) – a practice which any
evidence-based clinician would currently view with skepticism.
This study regarding serious bacterial infection among children with
extremely high fever, by Dr. Barbara Trautner and colleagues, provides
very useful epidemiologic data to guide health care providers caring for
children with fever. Indeed, the noted lack of value of the white blood
cell count or absolute neutrophil count in distinguishing serious
bacterial infection in this patient population is thought-provoking.
However, the recommendation to provide “antibiotics for all children with
hyperpyrexia who do not have a confirmed viral illness” is not directly
reached from the data, and must be recognized as personal opinion.
It was difficult to discern the frequency of “occult” serious
bacterial infection from the data presented in the manuscript. That is,
among fully immunized, non-septic-appearing, previously healthy children
with normal urinalyses, normal chest radiography, and without occult blood
in stools, what was the incidence of serious bacterial infection? As the
presence or absence of “rhinorrhea” as documented by the study is quite
subjective, it would also be helpful to know the incidence of occult
serious bacterial infection among children in the midst of an obvious
acute viral upper respiratory infection (as opposed to mild, subacute
nasal congestion).
Before a scientific recommendation to administer empiric antibiotics
can be made, the true incidence of occult serious bacterial infection must
be provided. Additionally, the possibility of spontaneous clearance of
bacteremia must be considered. And then, the risks of diagnosis and
treatment must be considered. What is the false-positive blood culture
rate? How does antibiotic therapy affect subsequent diagnostic
evaluations, including subsequent lumbar puncture? What is the adverse
event rate of expectant antibiotics within the individual (allergy,
opportunistic enterocolitis, etc.), and what are the possible adverse
implications on a population-based model? Until risk-benefit modeling is
performed, the data presented should be incorporated into rational
clinical decision making, but the advisement for expectant antibiotics
should be taken with a grain of salt.
Conflict of Interest:
None declared
eLetters: ![[Read eLetters]](/icons/misc/sectionDOWN.gif){kind=icon}
