Fever of Unknown Origin
Andrew P. Steenhoff
Fever of unknown origin (FUO ) is a diagnostic dilemma for pediatricians because it is often difficult to distinguish clinically between benign and potentially life-threatening causes. Pediatricians face the important challenge of not missing the diagnosis of a serious illness or an easily treatable condition that can result in increased morbidity. Fortunately, FUO is usually an uncommon presentation of a common disease, with most of these common diseases being easily treatable.
The classification of FUO is best reserved for children with a temperature >38°C (100.4°F) documented by a healthcare provider and for which the cause could not be identified after at least 8 days of evaluation (Table 204.1 ). It is important to differentiate FUO from fever without a source ; FWS is fever where the source has not yet been identified and is differentiated from FUO by the duration of the fever. FWS can progress to FUO if no cause is elicited after 7 days of evaluation.
Table 204.1
Summary of Definitions and Major Features of 4 Subtypes of Fever of Unknown Origin (FUO)
| FEATURE | CLASSIC FUO | HEALTHCARE-ASSOCIATED FUO | IMMUNE-DEFICIENT FUO | HIV-RELATED FUO |
|---|---|---|---|---|
| Definition | >38°C (100.4°F), >3 wk, >2 visits or 1 wk in hospital | ≥38°C (100.4°F), >1 wk, not present or incubating on admission | ≥38°C (100.4°F), >1 wk, negative cultures after 48 hr | ≥38°C (100.4°F), >3 wk for outpatients, >1 wk for inpatients, HIV infection confirmed |
| Patient location | Community, clinic, or hospital | Acute care hospital | Hospital or clinic | Community, clinic, or hospital |
| Leading causes | Cancer, infections, inflammatory conditions, undiagnosed, habitual hyperthermia | Healthcare-associated infections, postoperative complications, drug fever | Majority caused by infections, but cause documented in only 40–60% | HIV itself, typical and atypical mycobacteria, CMV, lymphomas, toxoplasmosis, cryptococcosis, immune reconstitution inflammatory syndrome (IRIS) |
| History emphasis | Travel, contacts, animal and insect exposure, medications, immunizations, family history, cardiac valve disorder | Operations and procedures, devices, anatomic considerations, drug treatment | Stage of chemotherapy, drugs administered, underlying immunosuppressive disorder | Drugs, exposures, risk factors, travel, contacts, stage of HIV infection |
| Examination emphasis | Fundi, oropharynx, temporal artery, abdomen, lymph nodes, spleen, joints, skin, nails, genitalia, rectum or prostate, lower-limb deep veins | Wounds, drains, devices, sinuses, urine | Skin folds, IV sites, lungs, perianal area | Mouth, sinuses, skin, lymph nodes, eyes, lungs, perianal area |
| Investigation emphasis | Imaging, biopsies, sedimentation rate, skin tests | Imaging, bacterial cultures | CXR, bacterial cultures | Blood and lymphocyte count; serologic tests; CXR; stool examination; biopsies of lung, bone marrow, and liver for cultures and cytologic tests; brain imaging |
| Management | Observation, outpatient temperature chart, investigations, avoidance of empirical drug treatments | Depends on situation | Antimicrobial treatment protocols | Antiviral and antimicrobial protocols, vaccines, revision of treatment regimens, good nutrition |
| Time course of disease | Months | Weeks | Days | Weeks to months |
| Tempo of investigation | Weeks | Days | Hours | Days to weeks |
CMV, Cytomegalovirus; CXR, chest radiograph; HIV, human immunodeficiency virus; IV, intravenous line.
Adapted from Mackowak PA, Durack DT: Fever of unknown origin. In Mandell GL, Bennett, JE, Dolin R, editors: Mandell, Douglas, and Bennett's principles and practice of infectious diseases , ed 7, Philadelphia, 2010, Elsevier (Table 51-1).
Etiology
The many causes of FUO in children are infectious, rheumatologic (connective tissue or autoimmune), autoinflammatory, oncologic, neurologic, genetic, factitious, and iatrogenic processes (Table 204.2 ). Although oncologic disorders should be seriously considered, most children with malignancies do not have fever alone. The possibility of drug fever should be considered if the patient is receiving any drug. Drug fever is usually sustained and not associated with other symptoms. Discontinuation of the drug is associated with resolution of the fever, generally within 72 hr, although certain drugs, such as iodides, are excreted for a prolonged period, with fever that can persist for as long as 1 mo after drug withdrawal.
Table 204.2
Diagnostic Considerations for Fever of Unknown Origin in Children
| ABSCESSES |
| BACTERIAL DISEASES |
| LOCALIZED INFECTIONS |
| SPIROCHETES |
| FUNGAL DISEASES |
| RICKETTSIAE |
| VIRUSES |
| PARASITIC DISEASES |
| RHEUMATOLOGIC DISEASES |
| HYPERSENSITIVITY DISEASES |
| NEOPLASMS |
| GRANULOMATOUS DISEASES |
| FAMILIAL AND HEREDITARY DISEASES |
|
Anhidrotic ectodermal dysplasia Familial Mediterranean fever, many other autoinflammatory diseases (Chapter 188 ) |
| MISCELLANEOUS |
Most fevers of unknown origin result from atypical presentations of common diseases. In some cases, the presentation as an FUO is characteristic of the disease (e.g., JIA), but the definitive diagnosis can be established only after prolonged observation, because initially there are no associated or specific findings on physical examination, and all laboratory results are negative or normal.
In the United States the systemic infectious diseases most commonly implicated in children with FUO are salmonellosis, tuberculosis, rickettsial diseases, syphilis, Lyme disease, cat-scratch disease, atypical prolonged presentations of common viral diseases, Epstein-Barr virus (EBV) infection, cytomegalovirus (CMV) infection, viral hepatitis, coccidioidomycosis, histoplasmosis, malaria, and toxoplasmosis. Less common infectious causes of FUO include tularemia, brucellosis, leptospirosis, and rat-bite fever. Acquired immunodeficiency syndrome alone is not usually responsible for FUO, although febrile illnesses often occur in patients with AIDS as a result of opportunistic infections (see Table 204.1 ).
Juvenile idiopathic arthritis (JIA) and systemic lupus erythematosus (SLE) are the connective tissue diseases most often associated with FUO. Inflammatory bowel disease (IBD) and Kawasaki disease are also frequently reported as causes of FUO. If factitious fever (inoculation of pyogenic material or manipulation of the thermometer by the patient or parent) is suspected, the presence and pattern of fever should be documented in the hospital. Prolonged and continuous observation of the patient, which can include electronic or video surveillance, is imperative. FUO lasting >6 mo is uncommon in children and suggests granulomatous, autoinflammatory, or autoimmune disease. Repeat interval evaluation is required, including history, physical examination, laboratory evaluation, and imaging studies.
Historically, 90% of pediatric FUO cases in the United States had an identifiable cause: approximately 50% infectious, 10–20% collagen vascular, and 10% oncologic. Later studies from the 1990s had variable results: 20–44% infectious, 0–7% collagen vascular, 2–3% oncologic, and up to 67% undiagnosed. The reason for the paradoxical increase in undiagnosed cases of FUO ironically is likely caused by improved infectious and autoimmune diagnostic techniques. The advent of polymerase chain reaction (PCR), improved culture techniques, and better understanding of atypical viral and bacterial pathogenesis and autoimmune processes likely contribute to earlier diagnosis of these conditions and fewer children with these conditions advancing to the category of FUO. By contrast, causes of FUO remain primarily infectious in developing settings where there is a higher infectious disease burden, and advanced diagnostics techniques are more limited.
Diagnosis
The evaluation of FUO requires a thorough history and physical examination supplemented by a few screening laboratory tests and additional laboratory and imaging evaluation informed by the history or abnormalities on examination or initial screening tests (see Table 204.2 ). Occasionally the fever pattern helps make a diagnosis (Fig. 204.1 ). Nonetheless, most diseases causing an FUO do not have a typical fever pattern.
History
A detailed fever history should be obtained, including onset, frequency, duration, response or nonresponse to therapy, recurrence, and associated symptoms. Repetitive chills and temperature spikes are common in children with septicemia (regardless of cause), particularly when associated with kidney disease, liver or biliary disease, infective endocarditis, malaria, brucellosis, rat-bite fever, or a loculated collection of pus.
The age of the patient is helpful in evaluating FUO. Children >6 yr old often have a respiratory or genitourinary tract infection, localized infection (abscess, osteomyelitis), JIA, or rarely, leukemia. Adolescent patients are more likely to have IBD, autoimmune processes, lymphoma, or tuberculosis, in addition to the causes of FUO found in younger children.
A history of exposure to wild or domestic animals should be solicited. The incidence of zoonotic infections in the United States is increasing, and these infections are often acquired from pets that are not overtly ill. Immunization of dogs against specific disorders such as leptospirosis can prevent canine disease but does not always prevent the animal from carrying and shedding leptospires, which may be transmitted to household contacts. A history of ingestion of rabbit or squirrel meat might provide a clue to the diagnosis of oropharyngeal, glandular, or typhoidal tularemia . A history of tick bite or travel to tick- or parasite-infested areas should be obtained.
Any history of pica should be elicited. Ingestion of dirt is a particularly important clue to infection with Toxocara canis (visceral larva migrans) or Toxoplasma gondii (toxoplasmosis).
A history of unusual dietary habits or travel as early as the birth of the child should be sought. Tuberculosis, malaria, histoplasmosis, and coccidioidomycosis can reemerge years after visiting or living in an endemic area. It is important to identify prophylactic immunizations and precautions taken by the patient against ingestion of contaminated water or food during foreign travel. Rocks, dirt, and artifacts from geographically distant regions that have been collected and brought into the home as souvenirs can serve as vectors of disease.
A medication history should be pursued rigorously. This history should elicit information about nonprescription preparations and topical agents, including eyedrops, that may be associated with atropine-induced fever.
The genetic background of a patient also is important. Descendants of the Ulster Scots may have FUO because they are afflicted with nephrogenic diabetes insipidus. Familial dysautonomia (Riley-Day syndrome), a disorder in which hyperthermia is recurrent, is more common among Jews than among other population groups. Ancestry from the Mediterranean region should suggest familial Mediterranean fever . Both familial Mediterranean fever and hyper-IgD syndrome are inherited as autosomal recessive disorders. Tumor necrosis factor receptor–associated periodic syndrome and Muckle-Wells syndrome are inherited as autosomal dominant traits.
Pseudo-FUO is defined as successive episodes of benign, self-limited infections with fever that the parents perceive as 1 prolonged fever episode. This needs to be carefully ruled out before undertaking an unnecessary evaluation. Usually, pseudo-FUO starts with a well-defined infection (frequently viral) that resolves but is followed by other febrile viral illnesses that may be less well defined. Diagnosis of pseudo-FUO usually requires a careful history, focusing on identifying afebrile periods between febrile episodes. If pseudo-FUO is suspected and the patient does not appear ill, keeping a fever diary can be helpful.
Physical Examination
A complete physical examination is essential to search for any clues to the underlying diagnosis, and often it is worthwhile to repeat a detailed examination on different days to detect signs that may have changed or been missed (Tables 204.3 and 204.4 ). The child's general appearance, including sweating during fever, should be noted. The continuing absence of sweat in the presence of an elevated or changing body temperature suggests dehydration caused by vomiting, diarrhea, or central or nephrogenic diabetes insipidus. It also should suggest anhidrotic ectodermal dysplasia, familial dysautonomia, or exposure to atropine. The general activity of the patient and the presence or absence of rashes should also be noted.
Table 204.3
Subtle Physical Findings with Special Significance in Patients with Fever of Unknown Origin
| BODY SITE | PHYSICAL FINDING | DIAGNOSIS |
|---|---|---|
| Head | Sinus tenderness | Sinusitis |
| Temporal artery | Nodules, reduced pulsations | Temporal arteritis |
| Oropharynx | Ulceration | Disseminated histoplasmosis, SLE, IBD, Behçet syndrome, periodic fever syndromes |
| Tender tooth | Periapical abscess, sinus referred pain | |
| Fundi or conjunctivae | Choroid tubercle | Disseminated granulomatosis* |
| Petechiae, Roth spots | Endocarditis | |
| Thyroid | Enlargement, tenderness | Thyroiditis |
| Heart | Murmur | Infective or marantic endocarditis |
| Relative bradycardia | Typhoid fever, malaria, leptospirosis, psittacosis, central fever, drug fever | |
| Abdomen | Enlarged iliac crest lymph nodes, splenomegaly | Lymphoma, endocarditis, disseminated granulomatosis* |
| Audible abdominal aortic or renal artery bruit | Large vessel vasculitis such as Takayasu arteritis | |
| Costovertebral tenderness | Chronic pyelonephritis, perinephric abscess | |
| Rectum | Perirectal fluctuance, tenderness | Abscess |
| Prostatic tenderness, fluctuance | Abscess | |
| Genitalia | Testicular nodule | Periarteritis nodosa, cancer |
| Epididymal nodule | Disseminated granulomatosis | |
| Spine | Spinal tenderness | Vertebral osteomyelitis |
| Paraspinal tenderness | Paraspinal collection | |
| Lower extremities | Deep venous tenderness | Thrombosis or thrombophlebitis |
| Upper or lower extremities | Pseudoparesis | Syphilitic bone disease |
| Skin and nails | Petechiae, splinter hemorrhages, subcutaneous nodules, clubbing | Vasculitis, endocarditis |
* Includes tuberculosis, histoplasmosis, coccidioidomycosis, sarcoidosis, granulomatosis with polyangiitis, and syphilis.
Adapted from Mackowak PA, Durack DT: Fever of unknown origin. In Mandell GL, Bennett, JE, Dolin R, editors: Mandell, Douglas, and Bennett's principles and practice of infectious diseases, ed 7, Philadelphia, 2010, Elsevier (Table 51-8).
Table 204.4
Adapted from Wright WF, Mackowiak PA: Fever of unknown origin. In Bennett JF, Dolin R, Blaser MJ, editors: Mandell, Douglas, and Bennett's principles and practice of infectious diseases, ed 8, Philadelphia, 2015, Elsevier (Table 56-9).
A careful ophthalmic examination is important. Red, weeping eyes may be a sign of connective tissue disease, particularly polyarteritis nodosa. Palpebral conjunctivitis in a febrile patient may be a clue to measles, coxsackievirus infection, tuberculosis, infectious mononucleosis, lymphogranuloma venereum, or cat-scratch disease. In contrast, bulbar conjunctivitis in a child with FUO suggests Kawasaki disease or leptospirosis. Petechial conjunctival hemorrhages suggest infective endocarditis. Uveitis suggests sarcoidosis, JIA, SLE, Kawasaki disease, Behçet disease, and vasculitis. Chorioretinitis suggests CMV, toxoplasmosis, and syphilis. Proptosis suggests an orbital tumor, thyrotoxicosis, metastasis (neuroblastoma), orbital infection, Wegener granulomatosis (granulomatosis with polyangiitis), or pseudotumor.
The ophthalmoscope should also be used to examine nail-fold capillary abnormalities that are associated with connective tissue diseases such as juvenile dermatomyositis and systemic scleroderma. Immersion oil or lubricating jelly is placed on the skin adjacent to the nail bed, and the capillary pattern is observed with the ophthalmoscope set on +40.
FUO is sometimes caused by hypothalamic dysfunction . A clue to this disorder is failure of pupillary constriction because of absence of the sphincter constrictor muscle of the eye. This muscle develops embryologically when hypothalamic structure and function also are undergoing differentiation.
Fever resulting from familial dysautonomia may be suggested by lack of tears, an absent corneal reflex, or a smooth tongue with absence of fungiform papillae. Tenderness to tapping over the sinuses or the upper teeth suggests sinusitis. Recurrent oral candidiasis may be a clue to various disorders of the immune system, especially involving the T lymphocytes. Hyperactive deep tendon reflexes can suggest thyrotoxicosis as the cause of FUO.
Hyperemia of the pharynx, with or without exudate, suggests streptococcal infection, Epstein-Barr virus infection, CMV infection, toxoplasmosis, salmonellosis, tularemia, Kawasaki disease, gonococcal infection, or leptospirosis.
The muscles and bones should be palpated carefully. Point tenderness over a bone can suggest occult osteomyelitis or bone marrow invasion from neoplastic disease. Tenderness over the trapezius muscle may be a clue to subdiaphragmatic abscess. Generalized muscle tenderness suggests dermatomyositis, trichinosis, polyarteritis, Kawasaki disease, or mycoplasma or arboviral infection.
Rectal examination can reveal perirectal lymphadenopathy or tenderness, which suggests a deep pelvic abscess, iliac adenitis, or pelvic osteomyelitis. A guaiac test should be obtained; occult blood loss can suggest granulomatous colitis or ulcerative colitis as the cause of FUO.
Laboratory Evaluation
The laboratory evaluation of the child with FUO and whether inpatient or outpatient are determined on a case-by-case basis. Hospitalization may be required for laboratory or imaging studies that are unavailable or impractical in an ambulatory setting, for more-careful observation, or for temporary relief of parental anxiety. The tempo of diagnostic evaluation should be adjusted to the tempo of the illness; haste may be imperative in a critically ill patient, but if the illness is more chronic, the evaluation can proceed in a systematic manner and can be carried out in an outpatient setting. If there are no clues in the patient's history or on physical examination that suggest a specific infection or area of suspicion, it is unlikely that diagnostic studies will be helpful. In this common scenario, continued surveillance and repeated reevaluations of the child should be employed to detect any new clinical findings.
Although ordering a large number of diagnostic tests in every child with FUO according to a predetermined list is discouraged, certain studies should be considered in the evaluation. A complete blood cell count (CBC) with a white blood cell (WBC) differential and a urinalysis should be part of the initial laboratory evaluation. An absolute neutrophil count (ANC) of <5,000/µL is evidence against indolent bacterial infection other than typhoid fever. Conversely, in patients with a polymorphonuclear leukocyte (PMN) count of >10,000/µL or a nonsegmented PMN count of >500/µL, a severe bacterial infection is highly likely. Direct examination of the blood smear with Giemsa or Wright stain can reveal organisms of malaria, trypanosomiasis, babesiosis, or relapsing fever.
An erythrocyte sedimentation rate (ESR) >30 mm/hr indicates inflammation and the need for further evaluation for infectious, autoimmune, autoinflammatory, or malignant diseases, tuberculosis, Kawasaki disease, or autoimmune disease. A low ESR does not eliminate the possibility of infection or JIA. C-reactive protein (CRP) is another acute-phase reactant that becomes elevated and returns to normal more rapidly than the ESR. Experts recommend checking either ESR or CRP, because there is no evidence that measuring both in the same patient with FUO is clinically useful.
Blood cultures should be obtained aerobically. Anaerobic blood cultures have an extremely low yield and should be obtained only if there are specific reasons to suspect anaerobic infection. Multiple or repeated blood cultures may be required to detect bacteremia associated with infective endocarditis, osteomyelitis, or deep-seated abscesses. Polymicrobial bacteremia suggests factitious or self-induced infection or GI pathology. The isolation of leptospires, Francisella, or Yersinia requires selective media or specific conditions not routinely used. Therefore, it is important to inform the laboratory what organisms are suspected in a particular case. Urine culture should be obtained in all cases.
Tuberculin skin testing (TST) should be performed with intradermal placement of 5 units of purified protein derivative that has been kept appropriately refrigerated. In children >2 yr old, it is reasonable to test for tuberculosis using an interferon-γ release assay (IGRA).
Imaging studies of the chest, sinuses, mastoids, or GI tract may be indicated by specific historical or physical findings. Radiographic evaluation of the GI tract for IBD may be helpful in evaluating selected children with FUO and no other localizing signs or symptoms.
Examination of the bone marrow can reveal leukemia; metastatic neoplasm; mycobacterial, fungal, or parasitic infections; histiocytosis; hemophagocytosis; or storage diseases. If a bone marrow aspirate is performed, cultures for bacteria, mycobacteria, and fungi should be obtained.
Serologic tests can aid in the diagnosis of EBV infection, CMV infection, toxoplasmosis, salmonellosis, tularemia, brucellosis, leptospirosis, cat-scratch disease, Lyme disease, rickettsial disease, and on some occasions JIA. The clinician should be aware that the reliability and the sensitivity and specificity of these tests vary; for example, serologic tests for Lyme disease outside of reference laboratories have been generally unreliable.
Radionuclide scans may be helpful in detecting abdominal abscesses as well as osteomyelitis, especially if the focus cannot be localized to a specific limb or multifocal disease is suspected. Gallium citrate localizes inflammatory tissues (leukocytes) associated with tumors or abscesses. Technetium-99m phosphate is useful for detecting osteomyelitis before plain radiographs demonstrate bone lesions. Granulocytes tagged with indium or iodinated IgG may be useful in detecting localized pyogenic processes. 18 F-fluorodeoxyglucose positron emission tomography (PET) is a helpful imaging modality in adults with an FUO and can contribute to an ultimate diagnosis in 30–60% of patients. Echocardiograms can demonstrate vegetation on the leaflets of heart valves, suggesting infective endocarditis. Ultrasonography (US) can identify intraabdominal abscesses of the liver, subphrenic space, pelvis, or spleen.
Total body CT or MRI (both with contrast) is usually the first imaging study of choice; both permit detection of neoplasms and collections of purulent material without the use of surgical exploration or radioisotopes. CT and MRI are helpful in identifying lesions of the head, neck, chest, retroperitoneal spaces, liver, spleen, intraabdominal and intrathoracic lymph nodes, kidneys, pelvis, and mediastinum. CT or US-guided aspiration or biopsy of suspicious lesions has reduced the need for exploratory laparotomy or thoracotomy. MRI is particularly useful for detecting osteomyelitis or myositis if there is concern about a specific limb. Diagnostic imaging can be very helpful in confirming or evaluating a suspected diagnosis. With CT scans, however, the child is exposed to large amounts of radiation. PET-CT or MRI may help localize an occult tumor.
Biopsy is occasionally helpful in establishing a diagnosis of FUO. Bronchoscopy, laparoscopy, mediastinoscopy, and GI endoscopy can provide direct visualization and biopsy material when organ-specific manifestations are present. When employing any of the more invasive testing procedures, the risk/benefit ratio for the patient must always be considered before proceeding further.
Management
The ultimate treatment of FUO is tailored to the underlying diagnosis. Fever and infection in children are not synonymous, and antimicrobial agents should only be used when there is evidence of infection, with avoidance of empirical trials of medication. An exception may be the use of antituberculous treatment in critically ill children with suspected disseminated tuberculosis. Empirical trials of other antimicrobial agents may be dangerous and can obscure the diagnosis of infective endocarditis, meningitis, parameningeal infection, or osteomyelitis. After a complete evaluation, antipyretics may be indicated to control fever associated with adverse symptoms.
Prognosis
Children with FUO have a better prognosis than adults. The outcome in a child depends on the primary disease process. In many cases, no diagnosis can be established, and fever abates spontaneously. In as many as 25% of children in whom fever persists, the cause of the fever remains unclear, even after thorough evaluation.
In a series of 69 patients referred for “prolonged” unexplained fever, 10 were not actually having fever, and 11 had diagnoses that were readily apparent at the initial visit. The remaining 48 were classified as having FUO. The median duration of reported fever for these patients was 30 days. Fifteen received a diagnosis, and 10 (67%) had confirmed infections: acute EBV or CMV infection (n = 5; with 1 patient developing hemophagocytic lymphohistiocytosis); cat-scratch disease (3); and histoplasmosis (2). The other 5 patients had inflammatory conditions (systemic JIA, 2; IBD, 1), central fever (1), or malignancy (acute lymphoblastic leukemia, 1).