Fever of unknown origin (FUO) was first described by Dr. Petersdorf and Dr. Beesom in 1961. FUO was defined as a temperature of 101 F (38.3 C) or higher with a minimum duration of 3 weeks without an established diagnosis after an intensive 1-week investigation in the hospital.
Today, due to technological advances allowing for sophisticated outpatient evaluations, the 1-week inpatient investigation is no longer required; however, experts suggest certain initial tests should be performed, and these tests should be unrevealing to establish the diagnosis of FUO. These tests are comprehensive history, repeated physical examination, complete blood count (CBC) with differential, 3 sets of blood cultures (from different sites, several hours apart, and before initiation of antibiotic therapy, if indicated), chest radiograph, complete metabolic panel (including liver function tests), which may be followed by hepatitis serologies, if abnormal, urinalysis with microscopy and urine culture, erythrocyte sedimentation rate (ESR), antinuclear antibodies (ANA), rheumatoid factor (RA), cytomegalovirus IgM antibodies or virus detection in blood, heterophile antibody test, tuberculin skin test, human immunodeficiency virus (HIV) antibody test, and computed tomography (CT) scan of abdomen.
Over 200 malignant/neoplastic, infectious, rheumatic/inflammatory, and miscellaneous disorders can cause FUO. Clinicians often order non-clue-based imaging and specific testing early in the FUO work-up, which may be inefficient or misleading and not cost effective. Despite extensive workup and diagnostic advances, up to 51% of FUO cases remain undiagnosed. In modern medicine, FUO remains one of the most challenging diagnoses.
It is important to note that immunocompromised and HIV patients may require an entirely different approach in diagnosing and treatment of recurrent fevers. This article focuses on FUO in immunocompetent adult patients.
Over the years researches narrowed down FUO etiologies to the following categories: infectious, rheumatologic/inflammatory, malignant/neoplastic, and miscellaneous.
Epidemiology of FUO is broad and depends on the etiology, age group, geography, environmental exposure, and immune/HIV status. In developing countries, infectious etiology of FUO is most prevalent whereas in developed countries FUO is likely due to non-infectious inflammatory disease.
There is no clear-cut diagnostic approach to FUO. Thorough history with a focus on most probable etiology based on patient’s symptoms is the key to pinpoint the origin of FUO. Information about previous illnesses, localizing symptoms, alcohol intake, home medications, occupational exposures, pets, travel, and familial disorders should not be overlooked. Constellation of patient-reported symptoms should help clinician narrow down etiology of FUO to 4 categories as mentioned above, as each of these has clinical hallmarks. For example, if a patient presents with B-symptoms, early satiety, and significant weight loss, a clinician should pursue neoplasm/malignancy as the most probable cause of FUO. On the other hand, if a patient presents with rigors, an infectious etiology should be considered, while joint involvement is a hallmark of rheumatologic disorders.
If an infectious etiology is likely, history of present illness should include prior invasive procedures/surgeries, dentition, tuberculosis (TB) exposure, pet contacts, mosquito/tick bites, rodent exposure, history of blood transfusions, and immunosuppressive drugs. Fever pattern analysis, for example, morning spikes versus twice daily versus semi-weekly fevers, provides additional clues to specific infectious culprits, such as malaria. Important physical exam findings include heart murmur, spinal tenderness, hepatomegaly, splenomegaly, and epididymal nodule. A new murmur could signify subacute bacterial endocarditis. Spinal tenderness could be a sign of subacute vertebral osteomyelitis. Isolated hepatomegaly could signify Q fever, typhoid/enteric fever, and brucellosis. Isolated splenomegaly could be a clue for miliary TB, Epstein-Barr virus (EBV), and cytomegalovirus (CMV). If an epididymal nodule is palpated, consider EBV infection, renal TB, and brucellosis.
Similarly, when considering malignancy/neoplasm, it is important to inquire about unintentional weight loss, age-appropriate cancer screening, family history of cancer, smoking, and alcohol use. On physical exam, one could notice relative bradycardia suggestive of lymphoma/ central nervous system (CNS) malignancy, a new heart murmur pointing toward atrial myxoma, or sternal tenderness making myeloproliferative disorder a possibility. Isolated hepatomegaly and FUO is a clue for hepatoma, renal cell carcinoma, or liver metastases.
When considering rheumatologic disorder and FUO, ask about muscle and joint pain and stiffness, oral ulcers, and family history of autoimmune conditions. If a patient is reporting rigors and chills, rheumatologic etiology of FUO is less likely. Fever distribution analysis could differentiate periarteritis nodosa (morning fevers) vs. adult Still’s disease (double quotidian). On physical exam, it is important to look for oral ulcers (Behcet disease, systemic lupus erythematosus [SLE]), unequal pulses (Takayasu arteritis), lymphadenopathy (SLE, RA, sarcoidosis), and rashes (sarcoidosis, SLE, adult Still disease). An epididymal nodule is a clue for periarteritis nodosa, SLE, and sarcoidosis whereas hepatomegaly without splenomegaly argues against rheumatologic disorders.
Cirrhosis and Crohn’s disease are often overlooked as miscellaneous causes of FUO. If suspected, it is important to inquire about past medical history, history of alcohol intake, intravenous drug use, non-alcoholic hepatosteatosis (NASH), and hepatitis. On physical examination, splenomegaly is an important diagnostic clue for Crohn’s disease and liver cirrhosis.
When working up the differential diagnosis for FUO, it is important to remember that the cause is more likely a subtle and /or uncommon manifestation of common disease rather than a rare disease. Diagnosing a cause of FUO can be a cumbersome task and requires repeated diligent and thorough history taking and physical examination.
To diagnose FUO, basic initial diagnostic tests outlined above should have been inconclusive. At this point, a clinician should exclude patient’s manipulation with a thermometer and have them stop taking as many medications as possible to minimize the possibility of drug-induced fevers.
In the past, nuclear medicine testing was generally reserved for cases that remain undiagnosed after thorough initial evaluation. Recent European studies suggest utilizing fluorodeoxyglucose positron emission tomography (FDG-PET)/CT scan earlier in FUO work up, if available. FDG-PET/CT and FDG/PET are highly sensitive and non-invasive diagnostic techniques for anatomic localization of infectious, inflammatory or neoplastic processes and, although nonspecific, can guide further definitive tests such as biopsy or aspiration.
If FDG-PET is not available, labeled leukocyte studies could be used as an alternative; however, they might have a lower diagnostic yield. Gallium- and indium-labeled leukocyte studies are highly sensitive but not specific enough for establishing a diagnosis. However, these tests are helpful to localize involved site for a targeted evaluation with a CT scan. Note that indium scans have a high rate of false negatives with bone infections. Positron emission tomography can aid in detecting obscure infections or malignancies.
Cardiac echocardiography can be helpful if culture-negative endocarditis or atrial myxoma is suspected.
The most common invasive tests associated with FUO are biopsies of lymph nodes, liver, bone marrow epididymal nodule, and temporal artery. These tests are performed only if the clinical picture or initial tests reveal findings that require histopathological evaluation. Biopsies are most commonly used to diagnose malignancy, certain infections, myeloproliferative disorders, and inflammatory conditions causing FUO. For example, temporal artery biopsy should be considered in a patient older than 60 years old and with significantly elevated ESR, particularly, if there are other symptoms suggestive of giant cell arteritis. Additionally, if the physical examination reveals lymphadenopathy in a patient with FUO, lymph node biopsy is recommended and may reveal definitive etiology of fevers.
There is no single standard FUO management protocol given the variety of possible etiologies. The most important thing is to investigate and/or rule out all possible diagnosis. Please note that empiric antibiotics are not indicated unless the patient with FUO is neutropenic. Antibiotics may delay the diagnosis of some occult infections. Empiric glucocorticoids are also not indicated unless there is strong clinical suspicion for specific rheumatologic diagnosis. However, in patients whose condition is deteriorating empiric therapeutic trials of antibiotics, steroids, or antituberculous agents may be considered.
Naprosyn test can be performed to differentiate infectious and neoplastic etiologies of FUO. The test is conducted over 3 to 4 days, during which patient temperatures are trended while a patient is given naproxen. If temperatures decrease substantially, malignant/neoplastic etiology is likely. However, if temperatures remain the same or only minimally decrease, the FUO is likely of infectious origin. The utility of naproxen test is not well studied, and at this time, experts believe the test is not specific enough to be useful for the individual patient.
It is important to remember that up to 51% of cases remain undiagnosed. However, the prognosis for these patients is generally good, and it is highly probable that FUO will spontaneously resolve in weeks to months. In stable patients without a diagnosis, non-steroidal anti-inflammatory drugs could be used for symptomatic management.
The differential diagnosis for FUO is massive but can be grouped into 4 following categories based on etiology: infectious, rheumatologic/inflammatory, malignant/neoplastic, and miscellaneous.
Infection accounts for about a third of cases of FUO. The most common infections causing FUO are miliary tuberculosis (TB), brucellosis, and Q fever, followed by intraabdominal, pelvic, intranephric and perinephric abscesses, typhoid/enteric fevers, toxoplasmosis, cat scratch disease (CSD), HIV, cytomegalovirus (CMV), Epstein-Barr virus (EBV), and extrapulmonary (renal, central nervous system [CNS]) TB. Note that in HIV population 75% causes of FUO are infectious but rarely due to HIV itself.
Another third of FUO cases is due rheumatologic and inflammatory disorders, such as adult Still disease, giant cell/temporal arteritis, periarteritis nodosa, microscopic polyangiitis, rheumatoid arthritis (RA), and systemic lupus erythematosus (SLE). It could be rarely associated with Takayasu's arteritis, Kikuchi disease, sarcoidosis (CNS), Felty syndrome, Gaucher disease, polyarticular gout, pseudogout, antiphospholipid syndrome (APS), Behcet disease, and Marshall syndrome.
Neoplasms and malignancies account for up to 18% of FUO etiologies. The most common neoplasms associated with FUO are lymphoma and renal cell carcinoma, followed by acute myeloid leukemia and myeloproliferative disorders. FUO is rarely associated with atrial myxoma, multiple myeloma, colon carcinoma, pancreatic carcinoma, hepatoma, CNC metastasis, liver metastasis, and systemic mastocytosis.
The remainder of FUO etiologies is classified as miscellaneous. These include drugs induced fevers, liver cirrhosis, subacute thyroiditis, and Crohn’s disease. Less commonly FUO is caused by deep vein thrombosis, pulmonary embolus, hematomas, familial Mediterranean fever, hypothalamic dysfunction, hypertriglyceridemia (type V) and fictitious fever.
There are many possible etiologies of FUO. The key to establishing the diagnosis is a thorough history, repeated physical exam, and clue-driven diagnostic tests.
The experts recommend beginning the FUO work up with the following investigations:
Additionally, further evaluation of any abnormalities detected by these tests could be performed.
Utilizing FDG-PET/CT early in FUO workup may provide anatomic localization of areas with an increased metabolic process to guide more definitive investigation such as biopsy or aspiration for culture.
Routine use of antibiotics and steroids to treat FUO is not recommended. Symptomatic treatment with antipyretic agents is reasonable, especially if a patient is suffering discomfort. However, it is important to remember that these medications can distort fever patterns, which could be potentially helpful in deriving FUO etiology.
If FUO remains undiagnosed after a thorough investigation, there is up to 75% chance that fevers will spontaneously resolve.
In modern medicine, FUO remains one of the most challenging diagnoses. It may be caused by over 200 malignant/neoplastic, infectious, rheumatic/inflammatory, and miscellaneous disorders. Diagnosing FUO requires a thorough history, repeated physical examinations and selective clue-based non-specific laboratory testing. Clinicians often take “shot-gun” approach early in the FUO work-up, which may be inefficient/ misleading and not cost effective. Despite extensive workup and available diagnostic advances, up to 51% of FUO cases remain undiagnosed. For patients whose FUO remains undiagnosed after thorough investigation prognosis is generally good and FUO spontaneously resolves in several weeks/months.