OVERVIEW: What every practitioner needs to know
Are you sure your patient has fever of unknown origin? What are the typical findings for this disease?
Although no precise definition of fever of unknown origin (FUO) in children is uniformly accepted, the following criteria have been proposed:
Body temperature is more than 38oC (100.4oF) for more than 8 consecutive days.
No cause for fever is evident on repeated physical examination.
No cause for fever is evident after basic laboratory evaluation.
Conventionally, FUO refers to fever that is continuous, or nearly so, as distinct from periodic fever, in which by definition there are cycles of remission and relapse.
How to approach the evaluation of a patient with Fever of Unknown Origin
FUO is a diagnosis of exclusion.
FUO is more likely to result from an uncommon presentation of a common infection than the reverse.
Take a thorough medical history, with particular attention to travel, lifestyle habits, exposure to animals and insects, and exposure to other ill individuals.
The history should be taken repeatedly, as critical details may emerge on repeated questioning.
A complete and comprehensive physical examination of the entire body is essential and should be repeated frequently as long as fever is present. Subtle findings may be important.
The following physical findings may easily be missed if a thorough examination is not performed carefully and often, as new findings may appear during the illness:
Rashes, petechiae, and evidence of microembolic phenomena
Lymph node enlargement
Enlargement of the liver or spleen
Abdominal masses or tenderness
Lung crackles, dullness to percussion, or decreased breath sounds
Inflammation or ulcerations of mucous membranes
Heart murmurs, friction rubs, displacement of the apical impulse
Muscle or bone tenderness
Joint effusion or synovial thickening, or limitation of motion
Genital ulcerations or discharge
Rectal tenderness or ulceration, presence of occult blood in stool
Subtle neurologic findings
Always confirm that fever is actually present:
Measure using a reliable route with proper technique.
Rectal temperatures are least dependent on technique.
Temporal and tympanic measurements are reliable if done properly.
Oral and axillary measurements are highly dependent on technique and are less reliable.
Oral temperature can be affected by warm liquids, chewing, and smoking.
Make sure thermometer is properly calibrated.
Observe the measurement of temperature directly.
Consider normal diurnal variation of body temperature, which may reach 37.8°C (100°F) in late afternoon.
Exercise and dehydration may elevate body temperature significantly.
Suspect factitious fever if fever is only reported but not observed, or if fever is not accompanied by tachycardia, chills, or diaphoresis. If factitious fever is suspected, the temperature of an observed, freshly voided urine sample cannot be falsified.
Consider hospitalization for direct observation if factitious fever is suspected.
What other disease/condition shares some of these symptoms?
Most cases of FUO in previously healthy children in whom a cause is identified are due to infections. With increasing age, collagen vascular disease becomes increasingly common. The differential diagnosis varies widely with age, exposure and travel history, and underlying medical conditions.
Common infections in the United States that may present with fever as the only initial symptom include:
Urinary tract infection
Bartonellosis (cat scratch disease)
Epstein-Barr virus infection
Less common infectious causes of FUO include:
Visceral larva migrans
Primary HIV infection
Immunocompromised hosts may be at risk for even less common opportunistic invasive fungal and chronic viral and parasitic infections. Consultation with an infectious disease specialist is strongly advised for such patients.
Infections of implanted medical devices should always be suspected as potential sources of unexplained fever, including:
Central venous catheters
Travel or residence in certain tropical regions may be associated with:
Yellow fever and other arboviral diseases
Virtually all of the above infections can be identified through a careful history and physical examination and only rarely cause fever in the absence of any other symptoms or physical findings.
Noninfectious causes of FUO also vary with age and include:
Collagen vascular disease (particularly juvenile idiopathic [rheumatoid] arthritis)
Malignancy (commonly lymphoma/leukemia or neuroblastoma)
Drug hypersensitivity reactions
Inflammatory bowel disease
Hereditary periodic fever syndromes
Dehydration (especially in diabetes insipidus)
Deep vein thrombosis
Infantile cortical hyperostosis
Chronic recurrent multifocal osteomyelitis
In most reported series, fever remits before a specific diagnosis is made in a significant proportion of cases.
What laboratory studies should you request to help confirm the diagnosis? How should you interpret the results?
Laboratory evaluation should be guided by the history and physical findings.
A “shotgun” approach (ordering multiple tests simultaneously) may result in unnecessary expense and may mislead the clinician if incidental unrelated findings are revealed.
Consultation with an infectious disease specialist should be considered to facilitate a targeted, efficient approach.
The pace of the laboratory evaluation should be dictated by the severity of illness.
Fever in a stable patient who is not critically ill can be evaluated in a logical, stepwise fashion.
A critically ill or unstable patient may require a more rapid and broadly directed approach.
Complete blood count with differential is usually performed but only occasionally suggests a specific diagnosis.
Leukocytosis with left shift or thrombocytosis may suggest a bacterial cause or Kawasaki syndrome.
Leukopenia or thrombocytopenia may suggest a viral or rickettsial cause.
Peripheral smear should be carefully examined for toxic granulation, evidence of hemolysis, blast forms, and (if travel history is suggestive) parasitic forms (malaria and filaria) or bacterial forms (Babesia, Ehrlichia/Anaplasma).
Urinalysis with culture and microscopic evaluation is essential.
Children unable to provide a high-quality clean-catch specimen should be catheterized.
Bag urine specimens should never be used.
Chemistry profile including liver enzyme, fractionated bilirubin, and amylase levels may be helpful.
Markedly elevated erythrocyte sedimentation rate, C-reactive protein and/or procalcitonin levels may suggest a bacterial or collagen vascular disease or Kawsaki syndrome.
Blood for at least two blood cultures should be obtained from separate venipuncture sites, using the largest volume consistent with the child’s size and the blood culture system in use.
The laboratory should be alerted in advance if fungi or bacteria are suspected that may require special media or prolonged incubation times.
For certain infections associated with intermittent bacteremia, additional samples or special incubation techniques may be needed.
Use adult blood culture bottles for children older than 2 years of age.
Depending on age, history, physical findings and travel/exposure history, serologic testing for certain specific pathogens may be appropriate.
It is often useful to obtain several serum tubes early in the course of illness to be stored as acute specimens for later comparison with samples obtained during convalescence.
Skin testing for tuberculosis is indicated if any risk factors are present.
Depending on age and clinical findings, serologic markers of autoimmune disease may be appropriate (e.g., antinuclear antibodies, rheumatoid factor).
Bone marrow biopsy is appropriate in instances when leukocyte, reticulocyte, and platelet counts are all severely depressed, principally to rule out lymphoid malignancy.
If a bone marrow sample is obtained, cultures and histologic stains for infectious pathogens should also be performed on the specimen.
Would imaging studies be helpful? If so, which ones?
Radiographic evaluation should be guided by the history and physical findings: a “shotgun” approach of ordering multiple imaging studies simultaneously is rarely helpful and may confuse the clinician with incidental findings.
A chest radiograph is appropriate in most cases even in the absence of cough, particularly in young children in whom physical findings of pneumonia can be easily missed.
Abdominal imaging (computed tomography [CT] with contrast or ultrasonography) may be appropriate if there is leukocytosis/left shift, pyuria with negative urine cultures, jaundice, or abdominal pain.
Pelvic imaging may be appropriate in female patients who are sexually active or have symptoms compatible with pelvic inflammatory disease, tubo-ovarian abscess or septic abortion. Screening CT examination of the sinuses
Screening CT examination of the sinuses may be appropriate in selected instances when upper respiratory symptoms or facial tenderness is present, although sinusitis is an uncommon cause of FUO in the absence of such symptoms.
Technetium-pyrophosphate bone scans, gallium scintiscans, or indium-labeled white blood cell scans are only occasionally helpful in patients without an identified source of fever after other laboratory and imaging modalities have proved unhelpful.
Confirming the diagnosis
There is no “one size fits all” diagnostic algorithm appropriate for children presenting with FUO.
The following is a partial list of historical clues that may identify patients at increased risk for certain causes of FUO (note that some animal and insect-borne infections are also geographically restricted):
Exposure to animals
Cats: bartonellosis (cat-scratch disease), toxoplasmosis
Rodents: Rat bite (streptobacillary) fever, hantavirus, lymphocytic choriomengitis virus, plague, leptospirosis
Goats and sheep: brucellosis, Q fever
Exposure to insects
Ticks: Rocky Mountain spotted fever, ehrlichiosis, anaplasmosis, babesiosis, Lyme disease, Colorado tick fever, tularemia
Body lice: bartonellosis (Bartonella quintana)
Mosquitos: West Nile fever, equine encephalitis viruses
Exposure to foods
Unpasteurized milk/cheese: brucellosis, listeriosis
Processed luncheon meats/dry sausage: listeriosis
Imported fresh fruits and vegetables, unpasteurized fruit juice: salmonellosis
Raw/undercooked eggs and meats: salmonellosis
Undercooked shellfish: hepatitis A, Vibrio vulnificus
Exposure to needles or blood products
Acute hepatitis A or B
Primary HIV infection
Primary HIV infection
Acute hepatitis A or B
Tropics: malaria, dengue, arboviruses, salmonellosis
Eastern Europe: tick-borne encephalitis, Lyme disease
Africa: malaria, leishmaniasis, trypanosomiasis, schistosomiasis
South America: malaria, leishmaniasis, trypanosomiasis, bartonellosis (B. bacilliformis)
Southeast Asia: malaria, dengue, Japanese encephalitis virus, melioidosis
All of the above: tuberculosis, hepatitis A, typhoid
Southwestern United States: coccidioidomycosis, hantavirus, plague
Mississippi and Ohio River valleys: histoplasmosis, blastomycosis
Northeastern/Middle Atlantic States and Wisconsin: Lyme disease, babesiosis
Southeastern/South Central United States: Rocky Mountain spotted fever, ehrlichiosis/anaplasmosis, tularemia
Ashkenazi Jews: familial dysautonomia
Armenians, Sephardic Jews, Turks, Arabs: familial Mediterranean fever
Ulster Scots: tumor necrosis factor receptor–associated periodic syndrome
Dutch, French: hyper-IgD syndrome
If you are able to confirm that the patient has fever of unknown origin, what treatment should be initiated?
Avoid giving antibiotics in the absence of a specific diagnosis, unless the patient is critically ill.
In the stable patient, antibiotics should be administered only after a specific cause is identified, targeted toward the causative organism.
Antipyretic therapy may be administered for purposes of comfort. When given on an as-needed basis, the presence of fever will not be obscured.
Antipyretics should not be administered “around the clock” or on a schedule.
Combination antipyretic therapy (e.g., coadministration or alternating administration of acetaminophen and ibuprofen) is almost never appropriate and may increase the risk of toxicity resulting from confusion regarding dosing schedules.
Ibuprofen should be administered with caution if renal disease or dehydration may be present or if there is evidence of coagulopathy.
Acetaminophen should be administered with caution if there is evidence of liver dysfunction.
Adequate hydration should always be ensured.
What are the adverse effects associated with each treatment option?
Initiation of antibiotics in a stable patient before a definitive diagnosis is made is generally unwise:
It may prevent isolation of the causative organism from blood, cerebrospinal fluid, urine, or stool.
It may suppress symptoms of a focal infection that may then only come to light after antibiotics are stopped.
It may result in side effects such as rashes, diarrhea, liver enzyme elevation, renal dysfunction, or drug-induced fever that may be misconstrued as symptoms of the original illness.
Some antibiotics may have nonspecific antiinflammatory effects that may mask diagnostically important symptoms.
What are the possible outcomes of fever of unknown origin?
Outcome of FUO depends on cause. Published series of FUO in the modern era have varied widely as to the spectrum of causative diseases, depending on the age of the population studied and exclusion criteria. The following was found in most series:
Infectious causes account for 40%-70% of cases.
Collagen vascular disease accounts for 10%-30% of cases.
Malignancy is found in 5%-15% of cases.
In 10%-25% of cases, fever remits before a cause is identified.
In children whose fever has lasted for many weeks with no identifiable cause despite an extensive and careful evaluation, the long-term prognosis is generally favorable. In most such patients, fever eventually remits without a specific diagnosis. Collagen vascular disease is the most common diagnosis made at a later time.
Are additional laboratory studies available; even some that are not widely available?
Consultation with an infectious disease specialist is advised before unusual or infrequently ordered tests are obtained.
How can fever of unknown origin be prevented?
Generally speaking, fever of unknown origin cannot be prevented.
The risk may be reduced by the following:
Appropriate vaccination of children
Good sanitation and hand hygiene
Avoidance of exposure to animals and insects
Appropriate precautions regarding food and water when traveling in developing countries
Appropriate food preparation in the home
Use of barrier contraception during sexual activity
Use of antiplasmodial prophylaxis during travel to areas where malaria exists
What is the evidence?
Tolan, RW. “Fever of unknown origin: a diagnostic approach to this vexing problem”. Clin Pediatr. vol. 49. 2010. pp. 207-13. (A thoughtful review article outlining a basic diagnostic framework.)
Miller, ML, Szer, I, Yogev, R. “Fever of unknown origin”. Pediatr Clin North Am. vol. 42. 1995. pp. 999-1015. (Comprehensive review article with extensive bibliography.)
McClung, HJ. “Prolonged fever of unknown origin in children”. Am J Dis Child. vol. 124. 1972. pp. 544-50. (Eleven-year series [1959-1969] of 99 patients at University of Wisconsin Children's Hospital, aged 16 years and younger with 3 weeks of outpatient or 1 week of inpatient fever >38.9°F without an identified cause; 30% had infections, 15% had collagen vascular disease, 9% had neoplasms, 10% were "normal," 21% had no diagnosis.)
Pizzo, PA, Lovejoy, FH, Smith, DH. “Prolonged fever in children: review of 100 cases”. Pediatrics. vol. 55. 1975. pp. 468-73. (Six-year series at Boston Children's Hospital, fever >38.5°C on at least 4 days over a 2-week period without an identified cause; 52% had infections, 20% had collagen vascular disease, 12% no diagnosis. Cause was suggested by history and physical examination alone in 65% of cases.)
Lohr, JA, Hendley, JO. “Prolonged fever of unknown origin”. Clin Pediatr. vol. 16. 1977. pp. 768-73. (Fifty-four patients, spanning a 7.5-year period were seen at the University of Virginia Hospital. They had fever >38.3oC for at least 3 weeks and were seen either on an outpatient basis or admitted for 1 week as an inpatient. They were aged 5 months-4 years. Thirty-three percent had infections, 20% had collagen vascular disease, 13% had malignancies, and 19% had no diagnosis. Twenty-eight percent were treated for an incorrect diagnosis and 24% received inappropriate antibiotic therapy. In 17%, an incomplete history led to delay in diagnosis. In 19%, basic laboratory tests were either not performed or were misinterpreted. In 7%, positive physical findings were ignored. All malignancies were identified within 6 days, and all infections within 6 weeks.)
Steele, RW, Jones, SM, Lowe, BA. “Usefulness of scanning procedures for diagnosis of fever of unknown origin in children”. J Pediatr. vol. 119. 1991. pp. 526-30. (One hundred nine patients, spanning a 5-year period at Arkansas Children's Hospital, were seen with fever >38°C for 3 weeks without a diagnosis; only 36% ultimately had a confirmed diagnosis. Yield from imaging studies and scans was poor and only rarely suggested a diagnosis not already suspected based on other findings. Most of the undiagnosed patients did well subsequently.)
Miller, LC, Sisson, BA, Tucker, LB. “Prolonged fevers of unknown origin in children: patterns of presentation and outcome”. J Pediatr. vol. 129. 1996. pp. 419-23. (Forty patients, spanning a 10-year period, were referred to the pediatric rheumatology clinic at The Floating Hospital for Children, Boston. Seventy-three percent had periodic fever and 27% had continuous fever. Most fever ultimately remitted without a final diagnosis. A significant number of the patients with periodic fever were later diagnosed with attention deficit hyperactivity disorder or developmental delays.)
Talano, JM, Katz, BZ. “Long-term follow-up of children with fever of unknown origin”. Clin Pediatr. vol. 39. 2000. pp. 716-7. (Nineteen patients with fever lasting more than 2 weeks were seen at Children's Memorial Hospital, Chicago over a 5-year period. Sixteen were well on follow-up; two acquired juvenile idiopathic (rheumatoid) arthritis, and one manifested recurrent intussusception.
Ongoing controversies regarding etiology, diagnosis, treatment
There is no consensus regarding a precise definition of FUO. The criteria shown above are consistent with those proposed in standard textbooks of pediatric infectious diseases.
Copyright © 2017, 2013 Decision Support in Medicine, LLC. All rights reserved.
No sponsor or advertiser has participated in, approved or paid for the content provided by Decision Support in Medicine LLC. The Licensed Content is the property of and copyrighted by DSM.
- OVERVIEW: What every practitioner needs to know
- Are you sure your patient has fever of unknown origin? What are the typical findings for this disease?
- What other disease/condition shares some of these symptoms?
- What laboratory studies should you request to help confirm the diagnosis? How should you interpret the results?
- Would imaging studies be helpful? If so, which ones?
- Confirming the diagnosis
- If you are able to confirm that the patient has fever of unknown origin, what treatment should be initiated?
- What are the adverse effects associated with each treatment option?
- What are the possible outcomes of fever of unknown origin?
- Are additional laboratory studies available; even some that are not widely available?
- How can fever of unknown origin be prevented?
- What is the evidence?
- Ongoing controversies regarding etiology, diagnosis, treatment