Evalutation of Fever without Identifiable Source in a Previously Healthy Child Less Than 3 Years of Age


  1. Fever is generally defined as a rectal temperature > 38.00C (100.40F). Typically oral and axillary temperatures are about 0.60C (1.00F) and 1.10C (2.00F) lower than rectal, respectively.
  2. Of major importance is to identify the children with serious bacterial infections = SBI (bacteremia, urinary tract infection, meningitis, bacterial gastroenteritis or pneumonia) or serious viral illness for which treatment is available (i.e. herpes simplex infection in neonates)
  3. Approximately 13% of infants < 28 days with a temp > 38.10C will have a serious bacterial infection (SBI).
  4. 8-9 % of infants < 3 mo of age with a temp > 38.10C will have a SBI.
  5. Among children > 3 months, SBI is usually associated with a temp > 39.00C.  In children 3-24 months who are nontoxic, temp > 39.40C (1030F) and without an identifiable focus of infection, about 3% will be bacteremic (may be less in the present "post-Haemophilus influenzae type b era").



A.  Rochester Criteria (Dagan et al.: J. Pediatr., 1985)

  1. Studied 233 previously healthy infants < 3 mo:
  2. Considered as "low-risk": infants who did not have any infectious focus on physical examination, WBC between 5000 and 15,000/per mm3, had < 1500 bands/ per mm3, and a normal U/A and stool < 5 WBC/hpf (if diarrhea).
  3. 89 (38%) infants did not meet one or more of these criteria and were classified as being at "high risk" for serious bacterial infection.
  4. 22 (25%) of the 89 infants in the high risk group and one of the 144 (0.7%) "low-risk" infants had a serious bacterial infection.

B.  Subsequent studies using the Rochester criteria:

  1. One low risk infant was hospitalized for Neisseria meningitidis bacteremia (McCarthy et al.: PIDS, 1990).
  2. 13 of 25 infants 28 to 89 days old with SBI that could be evaluated were identified as high risk by the Rochester criteria (sensitivity=0.52)  (Baskin et al.: J. Pediatr., 1992)
  3. Of 437 infants < 60 days who met the low risk criteria, 5 (1.1%) infants had a SBI including 2 infants with bacteremia (negative predictive value of 98.9%)  (Jaskiewicz et al.: Pediatrics, 1994)
  4. Of 134 infants < 29 days old studied retrospectively, 48 infants met the low risk criteria.  3 (6.3%) of these infants had SBI (negative predictive value 93.8%) (Ferrera et al,: Am J Emerg Med, 1997).
  5. Of 254 febrile 1 to 2-month old infants evaluated, 109 (42.9%) would have been identified as meeting the low risk criteria.  5 (4.6%) of the 109 infants had a SBI (Baker et al,: Arch Pediatr Adolesc Med, 1999)

C.  Infant (Yale) Observation Scale (McCarthy et al,: Pediatrics, 65: 1090-1095, 1980 and Pediatrics 70: 802-809, 1982)

  1. Includes quality of cry, reaction to parent stimulation, state variation, color, hydration, and response to social overtures (Score range 6-30)
  2. One study (Teach et al: J. Pediatr. 126: 877-881, 1995) applied the criteria to 6611 children 3-36 mo of age with a temp > 39.00C who appeared non-toxic.  The median YOS score for both patients with bacteremia (n=192) and patients without bacteremia (n=6419) was 6, but the mean rank among patients with bacteremia was significantly higher.  The sensitivity, specificity and positive and negative predictive values for a YOS score greater than 10 were 5.2%, 96.7%, 4.5% and 97.1%, respectively.

D.  Virtually all studies have shown that even the most experienced physicians are unable to reliably identify those infants < 2 months with SBI on the basis of history and physical examination alone



(Baraff et al..: Pediatrics 92: 1-12, 1993)

A.  General comments

  1. The practice guidelines below published in 1993 were consensus recommendations from a panel of experts in pediatrics, infectious diseases and emergency medicine after review of the relevant literature available.
  2. When surveyed or presented with pertinent clinical scenarios primary care physicians often manage febrile infants with fewer laboratory tests than the practice guidelines below suggest (Young: Pediatrics 95: 623-627, 1995 and Baucher and Pelton: Pediatrics 100: 137-138, 1997).
  3. The most variance in regards to the management of febrile children and the practice guidelines appears to be with the 28 to 90 day old age group.
  4. The reasons for the lack of adherence to the practice guidelines among primary care physicians is unclear but are likely multifactorial.  One study found that specialty,  years since graduation from medical school, perceived comfort in diagnosing serious bacterial illness and knowledge of the practice guidelines were predictors of compliance (Zerr et al: PIDS  18: 232-8, 1999)
  5. Another consideration may be that while patients with occult H. influenzae bacteremia have a substantial risk of remaining febrile or developing an invasive focus of infection, the majority of children with S. pneumoniae bacteremia will clear the bacteremia without treatment.  Thus, with the dramatic decrease in cases of invasive H. influ disease since the approval of H. influ. conjugate vaccine, some physicians may feel that the practice guidelines are too conservative in regards to the laboratory tests suggested.

B.  First month of life-Published practice guidelines (Pediatrics 92: 1-12, 1993)

  1. All febrile infants < 28 days of age should have a "septic work-up" (CBC with differential, blood culture, urine culture, LP +/- CXR), be hospitalized, and receive parenteral antibiotics.
  2. Empiric antibiotics:  Ampicillin + Gentamicin, Cefotaxime/Ceftriaxone (if LP abnormal)

C.  28 to 90 days of age-Published practice guidelines (Pediatrics 92: 1-12, 1993)

  1. In addition to a careful history and physical examination, the following laboratory evaluations are recommended:  CBC with differential, blood culture, urinalysis and urine culture.
  2. "Low-risk" infants can be safely treated as outpatients if the following criteria are met:
    1. previously healthy
    2. Non-toxic appearance (Infant Observation Scale score < 10)
    3. No identifiable focus of infection
    4. WBC 5000 to 15,000 cells per mm3
    5. Band/neutrophil ratio < 0.2
    6. Normal urinalysis
    7. Reliable, consistent caretaker with a telephone and transportation readily available
  3. If the infant does not meet the low risk criteria should have blood culture, urine culture, LP and consider CXR.
  4. Hospitalize pt and give IV antibiotics pending cultures
  5. If meets the low risk criteria, recommend:
    1. Urine culture for all infants
    2. Consider LP and blood culture
    3. Outpatient follow-up at 24 and 48 hours
    4. Consider no antibiotics pending cultures OR Ceftriaxone 50 mg/kg IM at initial visit and at 24 hr. pending cultures.
    5. If Ceftriaxone is given, a blood culture and LP should be done prior to antibiotics.

D.  3 to 36 mo of age-Published practice guidelines (Pediatrics 92: 1-12, 1993)

  1. Evaluation relies more on history and physical examination
  2. Previously healthy children in this age group with recognizable viral syndromes (bronchiolitis, croup, herpes stomatitis) are unlikely (<1%) to have SBI  (Greenes and Harper: Pediatr Infect Dis 18: 258-261, 1999).
  3. If no identifiable focus and child with temp > 39.00C, consider:
    1. CBC with differential and blood cultures in children with WBC < 5000 or >15,000 cells per mm3
    2. Urine culture in females < 2 years and males < 6 mo
    3. Careful follow-up without antibiotic treatment OR Ceftriaxone 50 mg/kg IM pending cultures if WBC < 5000 or >15,000 cells/mm3.
    4. If ill enough to be hospitalized: empiric treatment with Cefotaxime or Ceftriaxone (+Vancomycin if LP abnormal and bacterial meningitis suspected)




  1. Few febrile children who present without specific signs and symptoms of lung disease have abnormalities on CXR.  Thus for children > 2 months of age the vast majority of physicians would not obtain a CXR in a child without pulmonary symptoms.
  2. However, whether to obtain a CXR on all febrile infants < 2 months remains controversial.  Some experts recommend documenting a negative CXR before sending home a "low risk" febrile infant in this age group.

B.  CSF examination

  1. All febrile infants less than 28 days of age should have an LP.
  2. Some experts feel that a normal CSF examination (WBC< 8, normal protein and glucose) should be documented before sending home a "low risk" 1-2 month old febrile infant.
  3. An LP should be done in febrile infants 1-3 months of age to whom Ceftriaxone is to be administered because of a concern of SBI.

C.  Urine culture (UC) (Pediatrics 103: 843-852, 1999)

  1. The prevalence of UTI in children 2 months to 2 years of age who have no fever source evident by history or physical examination is about 5%.
  2. The prevalence of UTI in febrile girls is higher than boys (6.5% vs. 3.3% in children < 1 yr. and 8.1% vs. 1.9% in children 1-2 yr. old).  The rate in uncircumcised boys is 5 to 20 times higher than in circumcised boys.
  3. Young children with UTI are at higher risk than are older children for incurring acute renal injury and subsequent scarring than older children if left untreated (although the extent of the risk of subsequent hypertension or significant renal dysfunction is unclear).
  4. Urinalysis is done to identify children who are at high risk for UTI and merit empiric antibiotics pending the results of a UC.
    1. >5 WBC/hpf in a centrifuged urine specimen (standard urinalysis) fails to identify a significant number of children with UTIs (about 70% sensitive).
    2. Using either a positive leukocyte esterase or nitrite test is a more sensitive screen for UTI:  sensitivity 90-100% and specificity 58-91% (tends to be less sensitive in children < 2 years of age)
    3. The "enhanced urinalysis", a combination of a gram stain and manual cell count using an uncentrifuged urine specimen, appears to provide an excellent combination of sensitivity and specificity but is technically more time-consuming.
  5. Thus a UC should be considered in febrile children without a source, especially in girls and uncircumcised boys.
  6. If antibiotics are to be initiated because of an abnormal urinalysis or as empiric treatment for fever without a source, then a UC should be obtained via suprapubic aspiration of the bladder or bladder catherization.

D.  Stool examination

  1. Infants who present with watery or bloody diarrhea should have a stool examination for blood and fecal leukocytes.  Stool cultures should be sent when blood or >5 WBC/hpf are present in the stool.
  2. In infants with bacterial gastroenteritis the band to neutrophil ratio is often elevated (>0.2).



A.  The most common pathogens currently cultured from children in this age group with occult bacteremia are Streptococcus pneumoniae (90%), Salmonella species (5%) and Neiserria meningitidis (1%).

B.  Most children with pneumococcal bacteremia improve spontaneously but about 25% of untreated patients have persistent bacteremia or develop new focal symptoms, including 3% to 6% who develop meningitis.

C.  Without treatment, children with H. influenzae bacteremia have a substantial risk of remaining bacteremic or developing complications.

D.  Occult meningococcal bacteremia has frequent complications, including meningitis in about 40% and death in 4%.

E.  Empiric antibiotic treatment of children with occult bacteremia decreases the rate of complications including meningitis.

F.  The height of the fever (> 390C) and results of the CBC (WBC >15,000 or <5000 per mm3) and differential (ANC >10x109 cells/L) may suggest an increased risk of occult bacteremia.  For example, if WBC count > 15,000 cells per mm3 the risk of SBI rises from 3% to about 10-13%.

G.  A recent study (Finkelstein et al, Pediatrics 105 (Suppl 1): 260-266, 2000) of the pediatric offices of a HMO found that among 15 children treated with bacterial meningitis or meningococcal sepsis, 5 had an office visit for fever in the week prior to hospitalization but only 1 had documented fever > 390C and did not have laboratory testing.

H.  The availability now of a conjugate pneumococcal vaccine will likely decrease the prevalence of invasive pneumococcal disease and undoubtedly influence future practices.



  1. AAP:  [Practice parameter: the diagnosis, treatment, and evaluation of the initial urinary tract infection in febrile infants and young children. Pediatrics 103: 843-852, 1999.
  2. Anbar R, Richardson-deCorral V, and O'Malley P:  Difficulties in universal application of criteria identifying infants at low risk for serious bacterial infection.  J. Pediatr.  109: 483-485, 1986.
  3. Baker MD et al:  Failure of infant observation scales in detecting serious illness in febrile, 4- to 8-week infants.  Pediatrics 85: 1040-1043, 1990.
  4. Baker MD and Bell LM: Unpredictability of serious bacterial illness in febrile infants from birth to 1 month of age. Arch Pediatr Adolesc Med 153: 508-511, 1999.
  5. Baker MD et al:  The applicability of an established outpatient management protocol for febrile 0-1 month old infants.  Pediatrics 103: 627-631, 1999.
  6. Baker MD:  Evaluation and management of infants with fever.  Ped Clin NA 46 (6): 1061-1072, 1999.
  7. Baraff LJ, Bass JW, Fleisher GR, et al.:  Practice guidelines for the management of infants and children 0-36 months of age with fever without source.  Ann. Emerg. Med. 22: 1198-1210, 1993.
  8. Baskin MN, O'Rourke EJ, and Fleisher GR.:  Outpatient treatment of febrile infants 28 to 89 days of age with intramuscular administration of ceftriaxone.  J. Pediatr. 120: 22-27, 1992.
  9. Bass JW, Steele RW, Wittler RR, et al.:  Antimicrobial treatment of occult bacteremia: a multicenter cooperative study.  PID 12:466-473, 1993.
  10. Baucher H and Pelton SI:  Management of the young febrile child: a continuing controversy.  Pediatrics 100: 137-138, 1997.
  11. Chiu CH et al:  Identification of febrile neonates unlikely to have bacterial infections. Pediatr Infect Dis 16: 59-63, 1997.
  12. Crain EF and Gershel JC:  Urinary tract infections in febrile infants younger than 8 weeks of age. Pediatrics 86: 363-367, 1990.
  13. Dagan R, Sofer S, Phillip M, and Shachak E:  Identification of infants unlikely to have serious bacterial infection although hospitalized for suspected sepsis.  J. Pediatr. 10: 855-860, 1985.
  14. Ferrera PC et al:  Neonatal  fever: utility of the Rochester criteria in determining low risk for serious bacterial infections.  Am J Emerg Med 15: 299-302, 1997.
  15. Finkelstein JA et al:  Fever in pediatric primary care: occurrence, management, and outcomes.  Pediatrics 105 (Suppl 1): 260-266, 2000.
  16. Gorelick MH et al:  Screening tests for urinary tract infection in children: a meta-analysis. Pediatrics 104 (5): e54, 1999.
  17. Greenes DS and Harper MB: Low risk of bacteremia in febrile children with recognizable viral syndromes.  Pediatr Infect Dis 18: 258-261, 1999.
  18. Hoberman A et al:  Is urine culture necessary to rule out urinary tract infection in young febrile children? Pediatr Infect Dis 15: 304-309, 1996.
  19. Jaskiewicz JA et al:  Febrile infants at low risk for serious bacterial infection-an appraisal of the Rochester criteria and implications for management. Febrile Infant Collaborative Study Group. Pediatrics 94: 390-396, 1994.
  20. Kuppermann N et al: Risk of bacteremia and urinary tract infections in young febrile children with bronchiolitis.  Arch Pediatr Adolesc Med 151: 1207-1214, 1997.
  21. Kuppermann N:  Occult bacteremia in young febrile children.  Ped Clin NA 46 (6): 1071-1102, 1999.
  22. Lee GM and Harper MB: Risk of bacteremia for febrile young children in the post-Haemophilus influenzae type b era.  Arch Pediatr Adolesc Med 152: 624-628, 1998.
  23. Lieu TA, Baskin MN, Schwartz JS and Fleisher GR:  Clinical and cost-effectiveness of outpatient strategies for management of febrile infants.  Pediatrics  89: 1135-1144, 1992.
  24. McCarthy CA et al:  Outpatient management of selected infants younger than two months of age evaluated for possible sepsis.  PIDS 9: 385-389, 1990.
  25. McCarthy PL et al:  History and observation variables in assessing febrile children.  Pediatrics 65: 1090-1095, 1980.
  26. McCarthy PL et al: Observation scales to identify serious illness in febrile children.  Pediatrics 70: 802-809, 1982.
  27. Shaw KN et al:  Prevalence of urinary tract infection in febrile young children in the emergency department.  Pediatrics 101: 16, 1998.
  28. Teach et al:  Efficacy of an observation scale in detecting bacteremia in febrile children three to thirty-six months of age, treated as outpatients.  J. Pediatr. 126: 877-881, 1995.
  29. Teach SJ and Fleisher GR:  Duration of fever and its relationship to bacteremia in febrile outpatients three to 36 months old.  The Occult Bacteremia Study Group.  Pediatr Emerg Care 13: 317-319, 1997.
  30. Young PC: The management of febrile infants by primary-care physicians in Utah: comparison with published practice guidelines.  Pediatrics 95: 623-627, 1995.
  31. Zerr et al: Predictors of physician compliance with a published guideline on management of febrile infants.  PIDS  18: 232-8, 1999.