Evidence based clinical practice guideline of fever of uncertain source. Outpatient evaluation and management for children 2 months to 36 months of age. SOURCE(S): Cincinnati (OH): Children's Hospital Medical Center (CHMC); 2000. 10 p. [36 references] MAJOR RECOMMENDATIONS: Evidence strengths are given in brackets. Definitions of evidence grades are repeated following the recommendations. Clinical Assessments The Telephone Encounter 1. When contacted by telephone, it is recommended that a practitioner consider physically assessing any febrile child when verbal descriptions are unclear and the degree of illness uncertain. Temperature and Fever 2. Fever is usually defined as a temperature of at least 100.4 degrees F (38 degrees C) rectally (see Table 1 in the original guideline document for definitions used in the guideline). Although rectal temperatures are more accurate, it is recommended that a practitioner give credence to a parent's verbal report of a child's fever measured by any method, including when detected only by touch. A fever detected only by touch is reported to have a sensitivity of 84% and specificity of 76% [Evidence Grade: C]. Assessing the Degree of Illness 3. Both subjective and objective measures are recommended to estimate the degree of illness. * Note 1: The following features are included in the objective Yale Observation Scale. They are recommended for consideration in this assessment to supplement practitioners’ subjective impressions (see Table 2 in the original guideline document for the Yale Observation Scale) [Evidence Grade: A, C]. * Cry quality * Reaction to parent * State of alertness * Color * Hydration * Interactive behaviors * Note 2: The sicker a febrile child appears, the more likely the fever is associated with a serious bacterial infection [Evidence Grade: A, C]. Well appearing <3% chance of serious bacterial infections Ill appearing 26% chance of serious bacterial infections Toxic 92% chance of serious bacterial infections * Note 3: Response to antipyretics is not a reliable predictor of illness severity [Evidence Grade: S, E]. Identifying Overt Focal Infections History 4. It is recommended that history be targeted to determine the child's immunization status and exposures to known infectious agents. This information, with knowledge of disease prevalence in the community, will help establish the likelihood of specific causes of fever, and particularly those of viral origin. * Note 1: Less than 1 out of 100 viral infections is associated with a bacterial infection [Evidence Grade: C, D]. 5. It is recommended that history also be targeted to determine if behavioral changes caused by some focal infections have been observed. These include, but are not limited to, the ear pulling of otitis media, coughing of pneumonia, vomiting of gastroenteritis, or crying with voiding associated with some urinary tract infections. Physical Examination 6. Because occult bacteremia can occur with focal infections, it is recommended that when a source of infection is identified on physical examination, further evaluations be considered whenever the practitioner judges that focal findings are insufficient to explain the degree of the child's fever and illness. Identifying Occult Focal Infections 7. Using current technology and evidence, there are no perfect methods for detecting all possible occult infections. There also are no methods for identifying with absolute certainty which specific child is among the very few with a serious bacterial infection. It is recommended that the extent of evaluation to determine if a particular child has an occult infection, be based in part on each individual practitioner's level of comfort with the child's degree of illness. These approximations result from subjective and objective physical findings as supplemented by selected laboratory results. In addition, it is appropriate to include in the decisions a family's apparent tolerance for risk, the likelihood that identification of the source of fever will modify management, and the likelihood of good outpatient follow-up [Evidence Grade: Local expert opinion]. Urinary Tract Infections (UTIs) 8. When evaluating infants presenting with fevers of uncertain source, it is recommended that a practitioner have a low threshold for obtaining a urinalysis and urine culture using sterile collection technique [Evidence Grade: C]. This recommendation is strongest for the youngest infants. It also applies to older children with clinical signs and symptoms suggesting a urinary tract infection. * Note 1: The prevalence of urinary tract infections in infants and young children 2 months to 2 years of age who have no fever source evident from history or physical examination is estimated to be about 5% [Evidence Grade: E, S]. * Note 2: Consistent clinical findings range from being generally nonspecific in the infant, to very evident voiding dysfunction in the older child (see Table 3 in the guideline document for clinical signs and symptoms of urinary tract infection). * Note 3: The prevalence of urinary tract infection in febrile girls age 2 months to 2 years is more than twice that in boys (relative risk, 2.27). The prevalence of urinary tract infection in girls younger than 1 year of age is 6.5%; in boys, it is 3.3%. The prevalence of urinary tract infection in girls between 1 and 2 years of age is 8.1%; in boys, it is 1.9%. The rate in circumcised boys is low, 0.2% to 0.4%. The rate in uncircumcised boys is 5 to 20 times higher than in circumcised boys [Evidence Grade: E, S]. 9. It is recommended that any positive urinalysis result be considered consistent with a presumptive diagnosis of urinary tract infection and an indication to initiate antibiotic therapies and other measures fully described in the Children’s Hospital Medical Center guideline titled "Evidence Based Clinical Practice Guideline for Patients 6 Years of Age or Less with a First Time Acute Urinary Tract Infection" <http://www.guideline.gov/FRAMESETS/guideline_fs.asp?guideline=1196> (March, 1999). * Note 1: Any one of the following study results defines a positive urinalysis. These tests have sensitivities varying between 30-50% and specificities of 75-85%. A positive culture on urine collected using sterile technique remains the only standard for diagnosing a definite urinary tract infection [Evidence Grade: C]. * Positive nitrite screen. * Positive leukocyte esterase. * Positive microscopic examination. Positive if >10 white blood cells/hpf (spun sample) or organisms visible in unspun sample or on Gram stain. More sensitive and specific than other urine screening tests but not readily available [Evidence Grade: C]. Pneumonia 10. Pneumonia is seldom occult. It is recommended that pneumonia be considered when a fever exceeds 39 degrees C (102.2 degrees F) or a white blood cell count exceeds 20,0000 mm3 [Evidence Grade: C]. * Note 1: In children with lower temperatures and cell counts, the absence of respiratory distress, tachypnea, rales (crackles), or decreased breath sounds reduces the likelihood of pneumonia [Evidence Grade: Canadian guideline]. * Note 2: Age is a predictor of the cause of pneumonia. Viral pneumonia is most common during the first 2 years of life [Evidence Grade: Canadian guideline]. * Note 3: Chest x-rays do not often help in the choice of appropriate pneumonia therapy [Evidence Grade: C]. * Note 4: Pneumonia and bacteremia are infrequently associated (<3%) [Evidence Grade: C]. Bacterial Gastroenteritis 11. Gastroenteritis due to any cause is rarely occult and almost always signaled by some combination of diarrhea and vomiting. * Note 1: Rotavirus is the most common cause of acute gastroenteritis in children [Evidence Grade: S, E]. * Note 2: Bloody or mucoid stools usually herald bacterial gastroenteritis. There is also increased likelihood if there is a history of foreign travel or occurrence during a specific pathogen community outbreak [Evidence Grade: S, E]. 12. Typical gastroenteritis usually responds well to supportive care. The management and specific recommendations for children with gastroenteritis are detailed in the Children’s Hospital Medical Center guideline titled "Evidence Based Guidelines and Recommendations for the Community Evaluation and Medical Management of Children with Acute Gastroenteritis" <http://www.guideline.gov/FRAMESETS/guideline_fs.asp?guideline=657> (1997). Meningitis 13. Although bacterial meningitis is sometimes occult in the ill-appearing febrile infant or very young child, bacterial meningitis is usually suspected by clinical exam (e.g., neck stiffness, lethargy, vomiting), and diagnosed by lumbar puncture and the characteristics of the spinal fluid. Identifying Occult Bacteremias 14. It is recommended that febrile children be assessed to approximate their risk of being one of the <3% with occult bacteremia [Evidence Grade: C]. * Note 1: Risks are low but modified by height of fever and patients age. The risk for occult bacteremia is increased if more than one risk factor is present [Evidence Grade: A, C]. See the original guideline document for a table of fever as a predictor of occult bacteremia and a graph of age as a predictor of pneumococcal disease in a febrile child. Laboratory Studies and Detection of Occult Bacteremia 15. In spite of well appearance, a small proportion of febrile children might have bacteremia, it is not always necessary to obtain laboratory studies immediately for children appearing well and with clinical characteristics consistent with a low risk for bacteremia. If the family understands, agrees, and has the resources to assure medical follow-up, a low risk child might be considered a candidate for observation at home with daily reassessments until the fever resolves [Evidence Grade: D, S, E, Local expert opinion]. 16. When a child appears ill, or there is any uncertainty about a child’s condition or risk status, laboratory studies are recommended. Studies strongly recommended for consideration include urinalysis, sterile urine culture, complete blood count, white blood cell count, and absolute neutrophil count [Evidence Grade: S, E, Local expert consensus] (see the original guideline document for a graph of white blood cell count as a predictor of occult bacteremia). * Note 1: A white blood cell count of >15,000/mm3 raises risk for bacteremia to 3-4%. If >20,000/mm3, the risk is 8-10%. * Note 2: An absolute neutrophil count is more sensitive and specific than a white blood cell or absolute band count for occult bacteremia detection. An absolute neutrophil count of >10,000/mm3 raises risk, to 8-10% [Evidence Grade: A, C]. 17. Because 6% to 10% of children with bacteremia may develop serious bacterial infections, it is recommended that a blood culture also be considered especially if the child is ill appearing or the degree of illness is uncertain. This recommendation is particularly strong if antibiotic therapy is to be started. * Note 1: Blood culture of a single large blood volume is more likely to grow organisms than multiple cultures on smaller volumes [Evidence Grade: S, E, C]. 18. There is no published evidence demonstrating that chest x-rays, stool cultures, and lumbar puncture are helpful as "routine" studies. No specific recommendations are made other than to consider these studies when there are specific indications that the child is likely to have occult or complicated pneumonia, gastroenteritis, or meningitis [Evidence Grade: S, E]. Treatment for Occult Bacteremia 19. Children who appear well, are judged to be at sufficiently low risk to preclude the need for laboratory studies, and have a high likelihood of excellent follow-up can often be considered candidates for observation at home without starting antibiotic therapy [Evidence Grade: Local expert opinion]. 20. If because of history, physical, and laboratory assessments, a child is judged to be at high risk of being among the <3% of febrile children with occult bacteremia, starting empiric antibiotic therapy is considered a reasonable option after obtaining appropriate samples for culture [Evidence Grade: C, Local expert opinion]. * Note 1: This recommendation is strongest for ill appearing children pending clinical responses and return of bacterial culture results. This is especially recommended for those with white blood cell count >15,000/mm3 or an absolute neutrophil count >10,000 /mm3. * Note 2: Although complications are rare and up to 75% of occult pneumococcal bacteremia resolve spontaneously, children with occult bacteremia and treated with antibiotics clinically improve earlier and are less likely to be bacteremic at follow-up. An element of uncertainty is acknowledged, however, and it is estimated that a practitioner would need to treat 19 patients with suspect bacteremia for each patient subsequently documented to have actual bacteremia [Evidence Grade: C, S, E, Local expert opinions]. * Note 3: A decision to use antibiotics and the specific choice of antibiotic must be balanced against the increasing emergence of bacterial resistance. Also, in spite of a prevalent practice of starting empiric oral or parenteral antibiotics in febrile patients to try to prevent complications, the efficacy for this practice has never been documented in a randomized controlled fashion [Evidence Grade: M, E, S]. 21. Based on the 90% predominance of Streptococcus pneumoniae as a likely organism, it is recommended that initial antibiotic choices preferentially include only amoxicillin or ceftriaxone, or both in combination. There are no consistent evidences that one or the other of these choices is more efficacious [Evidence Grade: A]. Note: See the original guideline document for a table of antibiotic dosing, and a discussion of the issue of resistance and the use of "high dose" amoxicillin. * Note: For patients with likely allergies to the preferred antibiotics, no alternative antibiotics have been documented to be more efficacious. It is recommended that alternative drug choices be based on bacterial cultures and antibiotic sensitivities. Consultation with a specialist in pediatric infectious diseases might also be considered. 22. It is recommended that deferring antibiotic treatment pending a period of observation at home also be considered a valid management option for selected patients who look well, have normal laboratory studies, and with high likelihood of excellent follow-up. 23. It is recommended that, with the exception of special situations, if a decision is made to treat with antibiotics, the therapy be discontinued if cultures are reported as being negative. 24. If a blood culture is returned as being positive or fever is persistent, reexamination is recommended to reassess for the possibility that bacteremia was due to a previously undetected focal infection such as meningitis. It is recommended that the decision about whether to start or change antibiotics and whether to follow the child as an inpatient or outpatient be based on the current condition of the child and the organism isolated [Evidence Grade: S, E]. Evidence Based Grading Scale: A: Randomized controlled trial: large sample B: Randomized controlled trial: small sample C: Prospective trial or large case series D: Retrospective analysis E: Expert opinion or consensus F: Basic laboratory research S: Review article M: Meta-analysis Q: Decision analysis L: Legal requirement O: Other evidence X: No evidence Edward E. Rylander, M.D. Diplomat American Board of Family Practice. Diplomat American Board of Palliative Medicine.