Although fever is common in infants and young children three to 36 months of age, its management remains challenging and controversial. In this issue of American Family Physician,¹ Luszczak provides an update of the 1993 practice guideline on the management of febrile infants and young children, which was developed by an expert panel of senior academic pediatricians.²

In the first two years of life, children have an average of four to six acute febrile episodes, with medical care being sought for two thirds of these children.^2,3 Most have a temperature lower than 39°C (102.2°F) and a viral infection or an obvious bacterial origin for the fever.⁴ The challenge in management involves determining which of the 20 percent of children with no obvious source for their fever have occult bacteremia and will subsequently develop a serious bacterial illness.⁵

Various important issues were not addressed in Luszczak's article.¹ Fever phobia (an exaggerated fear of fever) exists equally among parents and health care professionals. Results of several studies^6,7 have documented the anxiety parents experience when their young child develops a fever. Despite evidence showing that a temperature elevation up to 40°C (104°F) has a number of beneficial effects, including improved host defenses and increased susceptibility of bacteria to these defenses,⁸ many nurses and physicians are overly concerned about rare potential complications and often needlessly prescribe antipyretics.^9,10

In addition, physicians often assume that parents are able to manage their febrile child at home. However, recent evidence challenges this assumption. In one study of 92 caregivers of young children,¹¹ fewer than one third could accurately measure their child's temperature and treat the fever appropriately. In another study,¹² more than one half of infants younger than one year received an incorrect dose of antipyretic. Many instruction labels for commonly used antipyretics include language that surpasses the reading comprehension of one half of the parents in the United States.¹³ Furthermore, the common practice of using a tepid sponge bath to reduce fever offers little advantage over antipyretics and may cause the child more discomfort.^14,15 Finally, the frequently recommended practice of alternating doses of acetaminophen with ibuprofen has never been studied; in fact, this practice may lead to parental confusion and overdosing of medication.¹⁶ Clearly, education on the management of fever is needed for health care professionals and parents.

Surveys of family physicians, pediatricians and emergency medicine physicians demonstrate significant variability in approaches to the evaluation and management of infants and young children who have a fever with no apparent locus of infection.¹⁷ One reason for this variability is that most research in this area has been performed in urban pediatric emergency departments—not in physicians' offices where most of these children are initially seen. A number of the tests recommended for the initial assessment of febrile infants and young children (e.g., white blood cell count, chest radiograph) may not be readily available in many medical offices.¹⁸ Thus, physicians must decide either to follow the guidelines and send infants and young children with fever to the emergency department, or to modify their approach based on what can be done in the office.

As noted in Luszczak's article,¹ the initial approach to the febrile infant or young child involves a careful history, observation of the patient's state of well-being and a detailed physical examination. About one in five children younger than three years of age will have a fever without source. Of these children who have a temperature of 39°C (102.2°F), 2 to 5 percent will have occult bacteremia, and of these, 10 percent may develop a serious bacterial illness (e.g., meningitis or pneumonia) if untreated.^19,5

Physicians must make a series of crucial decisions when treating infants and young children who have a fever without source¹⁸: Should tests be performed to identify those who have occult bacterial infections? If no locus of bacterial infection is found, should antibiotics be given empirically? If the choice is to empirically treat with antibiotics, should oral or parenteral therapy be used, or should a broad-spectrum or narrow-spectrum antibiotic be used, and how long should treatment last? Should the patient be treated as an outpatient or in a hospital? If the patient is not hospitalized, what follow-up should be arranged? The 1993 practice guideline was developed to help physicians make these decisions.

Two noteworthy developments have occurred since the practice guideline was first published. Urinary tract infection is now known to be among the most common occult bacterial sources of fever in children younger than three years of age.^4,5 Because the risk of renal damage is greatest in this age group, the American Academy of Pediatrics recently released a practice guideline on the diagnosis, treatment and evaluation of urinary tract infections in febrile infants and young children.²⁰ The combination of a urine dipstick test and microscopic examination of a centrifuged sample appears to be the best way to screen for urinary tract infection. Controversy exists concerning the best method for obtaining a urine sample for culture when this screen is abnormal. Bag-collected urine samples are associated with a false-positive rate as high as 87 percent,⁴ but the recommended approach of transurethral bladder catheterization or suprapubic bladder aspiration may not be easily performed by most office-based physicians.

With the introduction of routine vaccination, Haemophilus influenzae type b disease in infants and children has nearly been eradicated in the United States. Results of several studies^21,22 indicate that most cases of occult bacteremia are now caused by Streptococcus pneumoniae infection, which often resolves spontaneously without treatment. A recent meta-analysis²¹ of S. pneumoniae occult bacteremia suggests that 2,910 children with fevers of 39°C and higher would have to be treated with oral antibiotics to prevent one case of meningitis. Assuming that an adverse outcome (death or neurologic disability) occurs in 33 percent of children with meningitis, 6,570 children would have to be treated to prevent a single adverse outcome. Consequently, some investigators²³ advocate that empiric antibiotic therapy not be given. Furthermore, other investigators⁵ believe that the introduction of a conjugate S. pneumoniae vaccine will make the practice guideline on fever in infants and young children obsolete within the next one to two years.

In the meantime, family physicians faced with febrile children three to 36 months of age should individualize therapy. It is impossible to eliminate all risk in medical practice. As pointed out in one recent article,²⁴ one must choose between being a “risk-minimizer” (which potentially eliminates adverse sequelae from occult infections—and liability risk), and a “test-minimizer”(which, for the sake of practicality, increases the risk of missing an occult infection).²⁴ For physicians who are risk-minimizers, the modified guideline presented in this issue can serve as a useful starting point.