Prolonged, Recurrent, and Periodic Fever Syndromes

Definition

The 3 criteria for fever of unknown origin (FUO) in Petersdorf and Beeson’s landmark study of 100 adults in 1961 were illness of >3 weeks duration, fever higher than 38.3°C (101°F) on several occasions, and an uncertain diagnosis after 1 week of study in the hospital. In the 1990s, a prospective, population-based study conducted in adults in the Netherlands used fixed epidemiologic entry criteria and a specific diagnostic protocol. ^, In their study, the classic definition of FUO was adjusted in that immunocompromised patients were excluded and 1 week of hospitalization was replaced with 1 week of intensive evaluation (usually including computed tomography [CT] of the abdomen) and pursuit of additional diagnostic clues without arriving at a diagnosis.

In pediatrics, studies of FUO have used various definitions, with temperature thresholds between 38.0°C (100.4°F) and 38.9°C (102°F) and durations of fever ranging from 14 to 21 days. Some required the presence of daily fever and others only on multiple occasions. Reported etiologies vary broadly depending on the cohort’s definition as FUO. For purposes of this chapter, FUO in children and adolescents is considered only after a minimum of 14 days of daily documented temperature ≥38.3°C without apparent cause after performance of repeated physical examinations and screening laboratory tests and chest radiograph.

More than half of children referred for FUO do not meet these fever and exclusion criteria. The method of temperature measurement and frequency and timing of fever are ascertained and interpreted with the use of norms for age and sex ( Fig. 15.1 ). At 18 months of age, the mean rectal temperature for boys is 37.7°C with a standard deviation (SD) of 0.38°C; a rectal temperature of 38.5°C is <2 SD above the mean. Normal body temperature fluctuates within 24 hours an average by 0.5°C in adults and as much as 1.5°C in children, with temperature typically peaking in the late afternoon or evening. ^, A toddler with low-grade “fever” solely in the evenings likely has a normal body temperature. Additionally, patients whose sole complaint is a temperature ≤38.3°C do not merit additional pursuit as it is unlikely that a serious etiology will not be found, even if the finding is protracted or the patient is said to normally carry a “low” temperature.

Etiology

In Petersdorf and Beeson’s original publication, a high incidence of serious disease was found, with the big three categories of infections accounting for 36% of cases, neoplasms for 19%, and collagen vascular diseases for 13%; miscellaneous conditions were found in 35%, and no diagnosis was found in 7% of cases. In recent studies in adults, probably because many diagnoses are made earlier with more complex imaging, serologic, or molecular diagnostic tests, a larger proportion of the smaller cadre of patients who fulfill criteria for FUO have no identified diagnosis. In a systematic review of 3164 adults with FUO investigated since 2000 (comprising 18 series, all from outside of the US), the categories of confirmed diagnoses were significantly associated with geography, and the category of no confirmed diagnosis (23% of cases) with higher income settings. Additional skewing of categories of diagnoses occurs as more patients are evaluated and managed solely as outpatients. In a study of 84 adult patients with FUO referred to a single outpatient site in Japan, 67% of patients had no diagnosis, 95% of whom had fever resolution at follow up.

In studies of FUO in children, decreasing stringency of entry criteria correlates with increased likelihood of self-limited conditions. ^{, ,} In some series performed 4–5 decades ago, when fever approached 4 weeks duration without a definable source, the rate of life-altering diagnoses such as malignancy and autoimmune or autoinflammatory disease rose to 40%. ^{, ,} A systematic review published in 2011 of 18 pediatric studies involving 1638 children divided studies into those performed in developed countries (i.e., 8 studies from 3 countries, all published >1 decade ago) and developing countries (i.e., 10 studies, with 8 published since 2000). Overall, categories of diagnoses in developed and developing countries were remarkably similar: infection in 51%, autoimmune or autoinflammatory diseases in 9%, malignancy in 6%, miscellaneous noninfectious conditions in 11%, and no diagnosis in 23% of cases. Infectious causes were slightly more common in developing countries (56% vs. 42%), whereas finding no cause was more common in developed countries (31% vs. 18%). The infections diagnosed in 832 children included in the aforementioned systematic review are shown in Table 15.1 . Studies in children during the last decade include small numbers of hospitalized cases, and outpatient cases likely biased by subspecialty referral. In 2 recent series of children hospitalized for FUO at 2 different US medical centers, representations of categories shifted by age group and duration of fever. However, oncologic diagnoses were least represented overall. ^, Among 120 children evaluated for unexplained fever in a single outpatient pediatric infectious diseases referral site who received no definitive diagnosis or were thought to have recurrent self-limited illnesses, only 3 patients were diagnosed with a fever-related condition after 8 years of follow up. Regardless of setting of evaluation of children with FUO, finding no diagnosis has risen to represent almost one-half of cases, ^, which typically predicts spontaneous resolution without sequelae. ^{, ,}

Children <6 years of age are more likely to have an infectious cause of fever than older children and adolescents, who are more likely to have autoimmune and autoinflammatory disorders. Box 15.3 summarizes noninfectious causes of FUO. Leukemia and lymphoma are the most common malignancies diagnosed, and juvenile idiopathic arthritis, inflammatory bowel disorders, and systemic lupus erythematosus are the most common autoinflammatory and autoimmune disorders, but they may be variously classified. Kawasaki disease is the most common miscellaneous diagnosis in developed countries. Table 15.2 shows discriminating features of infectious and noninfectious causes of FUO. Epstein-Barr virus (EBV), Bartonella henselae, infections related to the kidney and urinary tract, and osteomyelitis (especially of spine and pelvic bones) are the most common infections identified in recent studies. ^{, ,} Agents that produce granulomatous infections (i.e., Bartonella, Mycobacterium, Salmonella, Histoplasma, Brucella, and Francisella spp.) disproportionately cause FUO and frequently involve the visceral organs, reticuloendothelial system, and bone marrow. In addition, FUO also can be the manifestation of a heralding unusual infection (such as disseminated Candida, Cryptococcus , or Aspergillus infection) due to congenital immunodeficiency.

Approach to the Evaluation

Patients with FUO are typically evaluated in phases, the pace of which is determined by the progression and severity of symptoms and laboratory test results. Screening tests are performed on all patients, with specific serologic tests and imaging studies guided by exposure histories and examination findings ( Table 15.2 ). Serology for EBV and B. henselae are typically sent early in the workup in patients with compatible clinical presentations due to the high frequency of these diagnoses. ^{, ,} Bone marrow biopsy is rarely required to confirm infections. Untargeted screening for endocrine disorders is rarely fruitful.

A chest radiograph should be performed for all patients regardless of their symptoms. Abdominal ultrasound can identify organ enlargement in patients with abdominal symptoms, elevated levels of serum hepatic enzymes, or in whom hepatosplenic cat-scratch disease or paranephric infection is suspected. ^, CT of the sinuses is viewed as an early study for patients with symptoms of sinusitis. Chest CT may be indicated when granulomatous or embolic disease is suspected. Abdominal CT or magnetic resonance imaging (MRI) study is helpful in patients with features directing suspicion to the abdomen or sometimes when a cryptogenic malignancy is strongly suspected. In a series of 28 children with prolonged fever and abdominal symptoms or signs, an abdominal CT was abnormal in 86%. However, the yield of CT in others was low. In another study of 109 children with FUO without localization, multiple modalities, including radioisotope scans and bone marrow biopsies, rarely led to a diagnosis. The interpretation of imaging studies as “definitely normal” is unusual, which often leads to additional studies, diverting the focus from the correct diagnosis.

¹⁸ F-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET)/CT performed in the diagnosis of both infectious and noninfectious etiologies of FUO in adults is comparable with other molecular imaging techniques and radiolabeled leukocyte scans. American and European nuclear medicine societies have included FUO among indications for PET/CT. Inflammatory cells have high avidity for ¹⁸ F-fluoro-2-deoxy-D-glucose (18F-FDG) and thus FDG-PET/CT can help detect and differentiate infection, malignancy, and inflammatory conditions. Performing FDG-PET or FDG-PET/CT has been reported cumulatively in approximately 200 European children with FUO, with diagnosis results in approximately 50%. ^, In the most recent study among children in the Netherlands, FDG-PET/CT identified a cause of fever (true positive) in 48% (most commonly endocarditis, systemic-onset juvenile idiopathic arthritis, and inflammatory bowel disease). Added value is purported for true negative results as well, which among FUO cases carries a favorable prognosis for resolution.

Most patients with FUO can be managed solely as outpatients with a planned interval return visit to obtain an interim history, to repeat an examination and screening tests (if persistently febrile), and to await resolution or the appearance of diagnostic clues. If there is progression of symptoms, weight loss, a new concerning symptom, or an abnormality on examination or screening tests, then the need for hospitalization, further testing, special imaging, or acquisition of tissue for diagnosis are more strongly considered.

Approach to Recurrent Fever

When a history is carefully obtained, most children who are referred to subspecialists for recurrent fever have multiple self-limited infections. Healthy children 2–3 years of age can have up to 10 self-limited illnesses per year, and children who attend organized childcare can have more.

In the absence of respiratory tract symptoms, fever that abates repeatedly during antibiotic therapy occasionally can result from endocarditis, renal infection, central nervous system ventricular shunt infection, or pneumonia in predisposed hosts. Relapsing infections such as Borrelia recurrentis, Brucella, and malaria are unusual or rare causes of recurring fevers distinguished by their unique exposures. Autoimmune disorders and certain autoinflammatory disorders (e.g., inflammatory bowel disease, systemic juvenile idiopathic arthritis) can also manifest with recurrent fever. Fig. 15.2 shows an approach to analyzing patients with recurrent fever.