A Better Approach for Pediatric Fever—Finally
BY JIJOE JOSEPH, DO, MBA; NANCY NGUYEN, DO; DEAN OLSEN, DO; & LYNCEAN UNG, DO
The literature on pediatric fever is vast—a PubMed search yields more than 14,000 results—but that volume of evidence has still not yielded a unified clinical practice guideline or consensus addressing the clinical approach, appropriate workup, and disposition for febrile babies. Even so, the best evidence indicates that a fever in neonates and infants is no longer an automatic ticket to a hospital stay, and new studies permit risk-stratification that actually sends some babies home.
A lot of hype surrounds pediatric fever, but the amplitude of fever is not as important as the child's appearance and level of activity (playful vs. lethargic) in predicting bacteremia. (Pediatrics 1989;83:1016.) Neonates (infants under 28 days old) and young infants (those 29-90 days old) have immature immune systems, and the risk of serious bacterial infections, such as urinary tract infection, bacterial meningitis, sepsis, and pneumonia, is significant (about 20%) and must be ruled out to prevent morbidity or mortality. (Pediatrics 1989;83:1016.)
Body temperature should be measured rectally in children younger than 2 because it is the closest to core temperature. Oral temperatures are typically 1°C (1.8°F) lower than rectal ones, and axillary temperatures 2°C (3.6°F) lower, but these conversions should not be used as substitutes for rectal temperatures in the ED despite their convenience. (Barren JM, et al. Pediatric Emergency Medicine. Philadelphia: Saunders/Elsevier. 2008: 291-8.) A 2015 meta-analysis determined the accuracy of peripheral temperatures was unacceptable for making clinical decisions. Tympanic and temporal thermometry also have poor sensitivity for fever detection in pediatric patients. (Ann Intern Med 2015;163::768.)
We should definitely treat fever in the ED, not just because it makes patients more comfortable, but also because it allows the physician to re-evaluate a child's behavior when afebrile, risk-stratify for serious bacterial infections, and counsel parents accordingly. A serious bacterial infection should be suspected if the patient's vitals don't improve or his clinical picture worsens after correction of fever. Elevations in body temperature may alarm parents and caregivers, but fever is the body's way of fighting infection. Don't worry until the child's temperature reaches 41.5°C (106.7°F). (Pediatrics 2011;127:580.) Heart rate increases by approximately 10 beats/minute and respiratory rate by five breaths/minute for every degree of Celsius over 38°C.
The absence of fever in the emergency department at the time of presentation, however, should not dissuade a physician from performing the appropriate workup in a child with a history of fever reported by the caregiver. Treatments for alleviating fever are not created equal. Despite confusion on this subject, data suggest that ibuprofen is superior to acetaminophen for treating fever and pain in children, probably because of its more potent anti-inflammatory mechanism of action. (Arch Pediatr Adolesc Med 2004;158:521.) Using both drugs in an alternating pattern may be more effective, but is not recommended because the combination increases the potential for errors. (Paediatr Child Health 2007;12:127.) Most fevers are caused by viruses, but Group B streptococci, Escherichia coli, and Listeria monocytogenes are the most common causes of serious bacterial infections in neonates. UTIs are the most common occult serious bacterial infection, especially in children. (Pediatrics 2003;111[5 Pt 1]:964.)
Before the 1980s, neonates and young infants with fever would undergo a full sepsis workup, including studies for blood, urine, and cerebrospinal fluid, and be hospitalized for 48 hours pending bacterial culture results. A 1993 review-based clinical guideline emphasized that a full sepsis workup consisting of CBC, blood culture, urinalysis (noting that UTI may not present with pyuria), chest x-ray, HSV PCR if at risk, and CSF cultures for infants under 28 days old is recommended despite the low probability of a serious bacterial infection. (Pediatrics 1993;92:140.)
This guideline was set forth to achieve a greater than 92% sensitivity in diagnosing a serious bacterial infection in a neonate, but it did not take into consideration the procedural complications and trauma involved in obtaining a lumbar puncture. This led clinicians to suggest the Rochester Criteria for risk-stratifying neonates in diagnosing serious bacterial infections.
Two studies by Dagan, et al., demonstrated that an infant was at low risk for serious bacterial infection if he met certain criteria. (J Pediatr 1985;107:855 and 1988;112:355.) These became known as the Rochester criteria, which along with other criteria, became ways to identify febrile young infants at low risk for serious bacterial infection who may be eligible for discharge home without hospitalization. (Pediatrics 1994;94:390.)
The Rochester criteria are the only ones that include neonates and young infants in risk-stratifying fever. A febrile infant with all of the Rochester criteria has a less than one percent risk of serious bacterial infection and a negative predictive value of 98.9% supporting discharge home with close follow-up and no empiric antibiotics. (Pediatrics 1994;94:390.)
The Philadelphia criteria also may be used to risk-stratify patients between 29 and 59 days old presenting with fever. These criteria have a sensitivity of 98% and a higher NPV of 100%, but they mandate a lumbar puncture to classify the patient as low risk. A lumbar puncture is not taken into account in the Rochester criteria. (N Engl J Med 1993;329:
The Boston criteria include infants up to age 89 days and also mandate a lumbar puncture. (J Pediatr 1992;120:22.) They utilize a higher WBC threshold of 20,000 WBC/mm, but had the lowest NPV at 94.6%.
The Rochester and Philadelphia criteria permit infants at low risk to be discharged without antibiotics with reliable home observation and follow-up within 24 hours. (Pediatrics 1994;94:390; N Engl J Med 1993;329:1437.) The Boston criteria are the only of the three where the patient receives antibiotics at discharge. (J Pediatr 1992;120:22.) (See a comparison of all three criteria: Table 1.)
Here's your best bet: A recent study by Borgia Gomez, et al., introduced a step-by-step algorithm to risk-stratify patients and compared the results to the Rochester criteria. (Pediatrics 2016;138; Epub 2016 Jul 5.) The algorithm uses laboratory tests such as C-reactive protein and procalcitonin. Infants 90 days or younger with an objective temperature over 38°C at home or in the ED and a fever without a source were assigned 0-9 points. (Table 2.)
The study showed a sensitivity of 92% and a NPV of 99.9%, which surpasses the Rochester criteria. (Tables 3-5.) (Pediatrics 2016;138; Epub 2016 Jul 5.) The availability of procalcitonin is not currently universal, however, limiting the applicability of the study.
Low-risk neonates are discharged with close follow-up based on the Rochester criteria, and high-risk neonates get admitted with antibiotics to treat their potential septicemia after a full workup, according to all three criteria. (Pediatrics 1994;94:390; N Engl J Med 1993;329:1437; J Pediatr 1992;120:22.)
The recommended antibiotic regimen includes an aminopenicillin (e.g., ampicillin 50 mg/kg IV) and an aminoglycoside (e.g., gentamicin 2.5 mg/kg) or a cephalosporin (e.g., cefotaxime). Data show that survival in neonatal sepsis may be increased with empiric use of gentamicin over cefotaxime. (Pediatrics 1994;94:390.)
Ceftriaxone use in this age group is discouraged because neonates cannot metabolize or excrete it well, especially with preexisting hyperbilirubinemia. Acyclovir should also be given if the neonate is at risk for herpes infection. Well-appearing infants aged 29-59 days should undergo the full sepsis evaluation unless they are considered low risk per the Rochester criteria. A urinalysis with culture should always be performed (again noting that pyuria may not be present with UTI), as well as a CBC, blood culture, and chest x-ray. (Pediatrics 1994;94:390.) Infants who meet the low-risk criteria and have reliable home observation and follow-up within 24 hours can be discharged without antibiotics, according to the Rochester and Philadelphia criteria.
Infants in this age group who meet high-risk criteria should be admitted pending culture results for antibiotics and a lumbar puncture for cell count and culture. The antibiotic of choice is ampicillin 50 mg/kg IV plus ceftriaxone 50 mg/kg IV (for normal CSF) or 100 mg/kg (for abnormal CSF). (Pediatrics 1994;94:390; N Engl J Med 1993;329:1437; J Pediatr 1992;120:22.) Most causes of infection in the 60-90-day-old group tend to be viral, especially if vaccinations are up to date. (N Engl J Med 1993;329:1437.) All infants in this age group should receive a urinalysis and culture. Further septic workup may be tailored to the appearance of the child. (N Engl J Med 1993;329:1437.) Routine CBC and blood cultures are not indicated in a vaccinated, well appearing child. (N Engl J Med 1993;329:1437; J Pediatr 1992;120:22.)
Chest x-rays may not be necessary in a child who is well-appearing and does not exhibit any respiratory symptoms such as tachypnea, tachycardia, wheezing, cough, or other focal symptoms. Lumbar puncture is only indicated in this age group if the infant looks toxic or is irritable. (N Engl J Med 1993;329:1437; J Pediatr 1992;120:22.)
Dr. Joseph is a fourth-year emergency medicine resident, Dr. Nguyen is a third-year emergency medicine resident, and Drs. Olsen and Ung, clockwise from top left, are emergency physicians, all at Nassau University Medical Center in East Meadow, NY.
P.S. If you liked this post, we think you'd
enjoy the EMN enews,
which delivers breaking news and online exclusives right to your inbox. Click the box below.