13.1 Fever Without Source, Serious Bacterial Infection & Kawasaki

Infectious diseases is clinical category #6 by exam volume on the PNCB CPNP-PC outline. Domain II wants a fever sorted by age and appearance, not by which virus is circulating at daycare. Domain III wants an AAP-aligned plan: the neonate goes to emergency care, the well-appearing older young infant is risk-stratified, urinary tract infection is actively included or excluded, and Kawasaki disease leaves primary care the same day. You are not running a pediatric infectious-disease service. You are deciding who cannot wait until morning.

Clinic opening. A 16-day-old has a rectal temperature of 38.3°C. The infant is feeding, has a slightly stuffy nose, and looks "a little warm." The caregiver wants acetaminophen and a phone check tomorrow. If you diagnose a viral upper respiratory infection and send that 2-week-old home on antipyretics, you missed this section.

Fever without source is an age problem first

In young infants, fever is typically ≥38.0°C (100.4°F) measured rectally. Bundling can warm the skin; it does not create a sustained rectal fever you can ignore. The first branch is age plus appearance, not the presence of a little rhinorrhea.

Do not send a 2-week-old home on antipyretics. A 3-week-old with fever is not a viral URI until a hospital evaluation says so. Diagnosing "a cold" because the nose is stuffy is the classic trap. Neonates have immature immune responses, poorly localized signs, and a high cost of missed meningitis, bacteremia, UTI, or HSV.

Ill appearance at any age — lethargy, poor perfusion, inconsolability, petechiae or purpura, meningismus, hypoxia, or a parent who says "this is not my child" — bypasses outpatient algorithms. That child is an emergency, not a risk-stratification worksheet.

Infants in the first weeks of life sit outside any "looks good, home with acetaminophen" plan, including those younger than the 8-day start of some well-appearing febrile-infant pathways. You do not improvise a Rochester score in the parking lot for a 10-day-old.

Risk stratification is a concept, not a lab-cutoff flash card

Historical Rochester, Philadelphia, and Boston criteria, the PECARN febrile-infant rule, and the AAP step-by-step pathway all try to identify well-appearing febrile infants at low enough risk that some invasive testing or hospitalization can be deferred. PNCB will not fairly test every milligram-per-liter CRP cutoff from last year's app. Know the idea:

• Shared ingredients: appearance, age band, urine (UTI dominates culture-proven SBI), blood culture plus inflammatory markers (procalcitonin, CRP, ANC — names matter more than memorized thresholds), and CSF when age or markers have not excluded meningitis.
• Inflammatory markers do not replace clinical judgment. A well-appearing 35-day-old with a reassuring procalcitonin is not the same patient as a mottled 35-day-old with the same number.
• Respiratory viral testing (influenza, SARS-CoV-2, RSV) can modify risk in older young infants inside a protocol. A positive RSV in a 10-day-old does not let you skip the neonatal workup.
• You use the current AAP febrile-infant guideline and your receiving hospital's pathway. You do not freelance.

The CPNP-PC's job in clinic is to recognize the age band, decide ED versus protocolized same-day evaluation, obtain or arrange a proper urine specimen, and not pretend that a well appearance in a 2-week-old is a discharge criterion.

Occult bacteremia: history versus the conjugate-vaccine era

In the 1980s–1990s, well-appearing febrile children 3–36 months had a meaningful rate of occult bacteremia, mostly Streptococcus pneumoniae. Automatic CBC, blood culture, and empiric antibiotics for a high WBC were common.

Conjugate vaccines changed the epidemiology. Haemophilus influenzae type b vaccine and pneumococcal conjugate vaccine (PCV) dropped occult bacteremia in immunized children to a small fraction of a percent. The CPNP-PC does not draw a blood culture on every 14-month-old with a fever of 39.2°C who is playing on the floor, fully immunized, and has no UTI risk. You still think about:

• Under-immunized or unimmunized children
• Ill appearance or a child who will not be observable
• Focal infection (bone, joint, pneumonia, skin, mastoid)
• UTI
• Persistent fever without source that is not following a typical viral course

Do not pretend occult bacteremia vanished. Do not practice 1994 laboratory medicine on a 2026 toddler.

UTI is the most common SBI in febrile young children

When a young child has fever without another adequate source, urinary tract infection is the SBI you must actively include or exclude. Highest yield: girls, uncircumcised boys, younger age (especially under 12 months), fever without source, and prior UTI. Mild tympanic redness is not an adequate source that lets you skip the urine.

Collection method is a diagnostic decision, not a convenience:

Culture is the gold standard. UA — leukocyte esterase, nitrite, microscopy for WBCs or bacteria — is a rapid screen that supports starting therapy in a high-risk febrile infant after a proper specimen, not a substitute for culture. Nitrite is specific but insensitive in infants who void frequently. Leukocytes on a bag from a febrile girl is how you both overtreat contamination and miss a real UTI you never confirmed.

Treat young febrile infants with UTI as pyelonephritis until proven otherwise (fever implies upper tract). Outpatient oral therapy is for selected well-appearing older infants and children with reliable follow-up. Neonates and ill infants are admitted. Imaging (renal and bladder ultrasound, VCUG) follows current AAP UTI guidance by age and recurrence — look it up rather than reciting every VCUG indication.

Kawasaki disease: count the days and refer

Kawasaki disease (KD) is a medium-vessel vasculitis of childhood. Coronary-artery complications are why primary care does not watch another three days of "viral fever" when the criteria are met.

Classic (complete) KD: fever for ≥5 days plus 4 of 5:

1. Bilateral bulbar conjunctival injection without exudate (nonexudative)
2. Oral changes: cracked red lips, strawberry tongue, injected pharynx — not isolated exudative tonsillitis
3. Cervical lymphadenopathy, often unilateral, ≥1.5 cm (the least commonly present criterion)
4. Polymorphous rash (not vesicular)
5. Extremity changes: palmar/plantar erythema, dorsal hand/foot edema; later periungual peeling

Incomplete KD is the infant trap: prolonged fever, fewer than four criteria, plus compatible labs (elevated CRP/ESR, then supplemental findings such as anemia, hypoalbuminemia, elevated ALT, thrombocytosis after day 7, sterile pyuria, leukocytosis) and echocardiography. Infants can present with fever and almost no mucocutaneous findings and still have coronary involvement.

Primary-care action: REFER. Same-day pediatric hospital or cardiology evaluation. IVIG and aspirin are hospital treatments. You do not start high-dose aspirin from a sample closet and send the child home. You do not wait for the fingers to peel to "confirm" the diagnosis — peeling is late.

Differential includes viral exanthem, measles, scarlet fever, drug reaction, and systemic-onset JIA. Nonexudative conjunctivitis plus cracked lips plus swollen hands after five days of fever is KD until a hospital team says otherwise.

Clinic vignette. A 2-year-old has had fever for 6 days. Eyes are red without discharge. Lips are cracked. A blotchy rash is on the trunk. Hands look puffy. A cervical node is 2 cm. Do not give another day of watchful waiting for roseola. Refer. Do not tell the family you will start aspirin tonight at home.

Pitfalls the exam writes on purpose

1. Sending a 2- or 3-week-old with fever home as a viral URI.
2. Treating neonatal fever with acetaminophen counseling and a next-day visit.
3. Memorizing every Rochester/PECARN lab cutoff instead of the risk-stratification concept and the age bands.
4. Ordering automatic blood cultures for occult bacteremia in a fully immunized, well-appearing toddler.
5. Missing UTI in a febrile girl or uncircumcised boy because "the ears look a little red."
6. Diagnosing complete Kawasaki and managing it in primary care with aspirin samples.
7. Waiting for peeling skin before referring incomplete KD in an infant.

Clinic close. Neonate with fever: ED, not antipyretics. Young infant: protocolized risk stratification, urine first among SBI sources. Toddler: conjugate-vaccine era plus UTI vigilance. Five days of fever plus four KD signs, or incomplete KD with labs/echo concern: hospital for IVIG and aspirin. You are the person who does not miss the infant who cannot wait.