septic arthritis, and cellulitis, to potentially

life-threatening infections, including Kawasaki disease,

meningitis, bacteremia, and sepsis. Because the differential

diagnosis is so broad, and fever is so common, the

approach to the febrile child is based on several factors,

including age, clinical impression (well or ill-appearing),

physical examination (source or no source), and risk

stratification (high or low risk for SBI) (Figure 48- 1).

Febrile infants <30 days old, even if they are well

appearing, should have a full septic work-up, which includes

complete blood count with manual differential, blood culture, urinalysis, urine culture, and lumbar puncture. CSF

should be sent for cell count, protein and glucose levels,

Gram stain, and culture. Herpes simplex virus (HSV) polymerase chain reaction should be considered in neonates

and those who are high risk. Urine must be collected in a

sterile manner, using bladder catheterization or suprapubic

aspiration. Bagged specimens are not helpful and are fre ­

quently contaminated by skin flora. Based on symptoms,

additional studies may be considered; liver function testing

should be assessed if an infant presents with j aundice, and

stool studies can be sent if diarrhea is present.

Well appearing infants age 1-3 months are classified

as low or high risk for SBI. Their risk stratification is

dependent on history, physical exam, and basic initial lab

results. To qualify as low risk, an infant must be previously

healthy without any comorbidity, nontoxic appearing,

without a focus of infection (excluding otitis media), demonstrate normal lab values, and have reliable caregivers

who can ensure close follow-up. The Rochester Criteria,

the Philadelphia Protocol, and the Boston Criteria are the

IV antibiotics;

admit

PEDIATRIC FEVER

History and physical exam to identify source of fever �

LP; IV

antibiotics;

admit

+/- 1M

antibiotics;

discharge with

24-hr follow-up

(if antibiotics

g iven, perform

LP}

No antibiotics

Discharge with

24-hr follow-up

(UA in boys < 1

year and girls <

2 years)

Figure 48-1. Pediatric fever diagnostic algorithm. CXR, chest x-ray; IM, intramuscu lar; LP, lumbar

pu ncture; SBI, serious bacterial infection; UA, urinalysis.

most commonly used decision-making tools for determining management of fever in well-appearing neonates and

infants (Table 48-1 ). Although all three have limitations,

they have attempted to create sensitive, specific screening

criteria to identify infants at low risk for SBI.

Lower risk stratification can be considered if the white

blood cell (WBC) count is between 5 and 15,000, band to

neutrophil ratio is <0.2, and urinalysis with <8 WBC per

high-power field. Clinical impression alone is not suffi ­

cient to forego lumbar puncture in this age group. The

decision to perform lumbar puncture depends on several

factors, including laboratory results, urinalysis, vaccination

status, and presence or absence of viral symptoms.

Well appearing children 3-36 months are at lower risk

for disseminated infections and can generally be managed

based on the nature of the infection, without an extensive

work-up for SBI. Fever in this age group is most commonly

caused by viral infections. The incidence of occult bacteremia

in well-appearing febrile children in this age group has

steadily decreased due to routine administration of both the

Hib and pneumococcal conjugate vaccine. Centers for

Disease Control and Prevention data now reflect that the current rate of occult bacteremia is < 1%. Furthermore, approximately 80% of pneumococcal bacteremia will resolves

spontaneously without intervention. Thus several acceptable

variations in management exist for this patient cohort.

The evaluation of well-appearing children age

3-36 months includes urinalysis and urine culture for girls

<2 years of age and boys < 1 year of age, particularly if they

are uncircumcised. A CXR may be performed if there are

signs of lower respiratory infection. If no source is

identified after this focused evaluation, reassurance and

CHAPTER 48

Table 48-1. Rochester, Philadel phia, and Boston criteria comparison.