Approach to the Adult Patient

with Fever of Unknown Origin

ALAN R. ROTH, D.O., and GINA M. BASELLO, D.O., Jamaica Hospital Medical Center,
Mount Sinai School of Medicine Family Practice Residency Program, Jamaica, New York

Fever of unknown origin (FUO) in adults is defined as a temperature higher than

38.3°C (100.9°F) that lasts for more than three weeks with no obvious source despite
appropriate investigation. The four categories of potential etiology of FUO are classic,
nosocomial, immune deficient, and human immunodeficiency virus–related. The four
subgroups of the differential diagnosis of FUO are infections, malignancies, autoim-
mune conditions, and miscellaneous. A thorough history, physical examination, and
standard laboratory testing remain the basis of the initial evaluation of the patient
with FUO. Newer diagnostic modalities, including updated serology, viral cultures,
computed tomography, and magnetic resonance imaging, have important roles in the
assessment of these patients. (Am Fam Physician 2003;68:2223-8. Copyright© 2003
American Academy of Family Physicians.)

A
dult patients frequently pre- and changes in disease states, such as the emer-
sent to the physician’s office gence of human immunodeficiency virus
with a fever (temperature (HIV) infection and an increasing number of
higher than 38.3°C [100.9°F]).1 patients with neutropenia. Others contend that
Most febrile conditions are altering the definition would not benefit the
readily diagnosed on the basis of presenting evaluation and care of patients with FUO.4
symptoms and a problem-focused physical The four categories of potential etiology of
examination. Occasionally, simple testing FUO are centered on patient subtype—clas-
such as a complete blood count or urine cul- sic, nosocomial, immune deficient, and HIV-
ture is required to make a definitive diagnosis. associated. Each group has a unique differen-
Viral illnesses (e.g., upper respiratory infec- tial diagnosis based on characteristics and
tions) account for most of these self-limiting vulnerabilities and, therefore, a different
cases and usually resolve within two weeks.2 process of evaluation (Table 1).5
When fever persists, a more extensive diagnos-
tic investigation should be conducted. CLASSIC
Although some persistent fevers are manifes- The classic category includes patients who
tations of serious illnesses, most can be readily meet the original criteria of FUO, with a new
diagnosed and treated. emphasis on the ambulatory evaluation of
these previously healthy patients.6 The revised
Definitions and Classifications criteria require an evaluation of at least three
The definition of fever of unknown origin days in the hospital, three outpatient visits, or
(FUO), as based on a case series of 100 pa- one week of logical and intensive outpatient
tients,3 calls for a temperature higher than testing without clarification of the fever’s
38.3°C on several occasions; a fever lasting cause.5 The most common causes of classic
more than three weeks; and a failure to reach a FUO are infection, malignancy, and collagen
diagnosis despite one week of inpatient investi- vascular disease.
gation. This strict definition prevents common
and self-limiting medical conditions from NOSOCOMIAL

See page 2113 for being included as FUO. Some experts have Nosocomial FUO is defined as fever occur-
definitions of strength- argued for a more comprehensive definition of ring on several occasions in a patient who has
of-evidence levels. FUO that takes into account medical advances been hospitalized for at least 24 hours and has

Downloaded from the American Family Physician Web site at www.aafp.org/afp. Copyright© 2003 American Academy of
Family Physicians. For the private, noncommercial use of one individual user of the Web site. All other rights reserved.
TABLE 1
Classification of Fever of Unknown Origin (FUO)

Category of FUO Definition Common etiologies

Classic Temperature > 38.3°C (100.9°F) Infection, malignancy, collagen vascular disease
Duration of > 3 weeks
Evaluation of at least 3 outpatient visits or 3 days
in hospital
Nosocomial Temperature > 38.3°C Clostridium difficile enterocolitis, drug-induced,
Patient hospitalized ≥ 24 hours but no fever or pulmonary embolism, septic thrombophlebitis,
incubating on admission sinusitis
Evaluation of at least 3 days
Immune deficient Temperature > 38.3°C Opportunistic bacterial infections, aspergillosis,
(neutropenic) Neutrophil count ≤ 500 per mm3 candidiasis, herpes virus
Evaluation of at least 3 days
HIV-associated Temperature > 38.3°C Cytomegalovirus, Mycobacterium avium-intracellulare
Duration of > 4 weeks for outpatients, > 3 days complex, Pneumocystis carinii pneumonia,
for inpatients drug-induced,Kaposi’s sarcoma, lymphoma
HIV infection confirmed

HIV = human immunodeficiency virus.

Adapted with permission from Durack DT, Street AC. Fever of unknown origin—reexamined and redefined. Curr Clin Top Infect Dis 1991;11:37.

not manifested an obvious source of infection that could assessed for three days without establishing an etiology for
have been present before admission. A minimum of three the fever.5 In most of these cases, the fever is caused by
days of evaluation without establishing the cause of fever is opportunistic bacterial infections. These patients are usu-
required to make this diagnosis.5 Conditions causing noso- ally treated with broad-spectrum antibiotics to cover the
comial FUO include septic thrombophlebitis, pulmonary most likely pathogens. Occult infections caused by fungi,
embolism, Clostridium difficile enterocolitis, and drug- such as hepatosplenic candidiasis and aspergillosis, must
induced fever. In patients with nasogastric or nasotracheal be considered.9 Less commonly, herpes simplex virus may
tubes, sinusitis also may be a cause.7,8 be the inciting organism, but this infection tends to present
with characteristic skin findings.
IMMUNE DEFICIENT
Immune-deficient FUO, also known as neutropenic HIV-ASSOCIATED
FUO, is defined as recurrent fever in a patient whose neu- HIV-associated FUO is defined as recurrent fevers over a
trophil count is 500 per mm3 or less and who has been four-week period in an outpatient or for three days in a hos-
pitalized patient with HIV infection.5 Although acute HIV
infection remains an important cause of classic FUO, the
virus also makes patients susceptible to opportunistic infec-
The Authors tions. The differential diagnosis of FUO in patients who are
ALAN R. ROTH, D.O., is chairman and program director of the Jamaica HIV positive includes infectious etiologies such as Mycobac-
Hospital Medical Center, Mount Sinai School of Medicine Family Prac- terium avium-intracellulare complex, Pneumocystis carinii
tice Residency Program, Jamaica, N.Y. He is also associate professor of
community and preventive medicine at Mount Sinai School of Medi- pneumonia, and cytomegalovirus. Geographic considera-
cine. Dr. Roth received his medical degree from the New York College tions are especially important in determining the etiology of
of Osteopathic Medicine, Old Westbury, N.Y., and completed a family FUO in patients with HIV. For example, a patient with HIV
medicine residency at the Jamaica Hospital Medical Center.
who lives in the southwest United States is more susceptible
GINA M. BASELLO, D.O., is assistant director of the Jamaica Hospital to coccidioidomycosis. In patients with HIV infection, non-
Medical Center, Mount Sinai School of Medicine Family Practice Resi-
dency Program, and clinical instructor of community and preventive med- infectious causes of FUO are less common and include lym-
icine at the Mount Sinai School of Medicine. She received her medical phomas, Kaposi’s sarcoma, and drug-induced fever.9,10
degree from the New York College of Osteopathic Medicine and com-
pleted a family medicine residency at Jamaica Hospital Medical Center.
Differential Diagnosis
Address correspondence to Alan R. Roth, D.O., Jamaica Hospital Med- The differential diagnosis of FUO generally is broken
ical Center, Family Practice Residency Program, 89-06 135th Street,
Suite 3C, Jamaica, NY 11418 (e-mail: [email protected]) Reprints into four major subgroups: infections, malignancies,
are not available from the authors. autoimmune conditions, and miscellaneous (Table 2). Sev-

2224 AMERICAN FAMILY PHYSICIAN www.aafp.org/afp VOLUME 68, NUMBER 11 / DECEMBER 1, 2003
 Fever of Unknown Origin
TABLE 2
Common Etiologies of Fever of Unknown Origin

Infections Autoimmune conditions

Tuberculosis (especially Adult Still’s disease
extrapulmonary) Polymyalgia rheumatica
Abdominal abscesses Temporal arteritis
Pelvic abscesses Rheumatoid arthritis
Dental abscesses Rheumatoid fever common in this population, malignancy has become a com-
Endocarditis Inflammatory bowel disease mon etiologic consideration in elderly patients. Malignan-
Osteomyelitis Reiter’s syndrome cies that sometimes are difficult to diagnose, such as chronic
Sinusitis Systemic lupus erythematosus leukemias, lymphomas, renal cell carcinomas, and meta-
Cytomegalovirus Vasculitides static cancers, often are found in patients with FUO.12
Epstein-Barr virus Miscellaneous
Human immunodeficiency virus Drug-induced fever AUTOIMMUNE CONDITIONS
Lyme disease Complications from cirrhosis
Rheumatoid arthritis and rheumatic fever are inflam-
Prostatitis Factitious fever
Sinusitis Hepatitis (alcoholic,
matory diseases that used to be commonly associated with
Malignancies granulomatous, or lupoid) FUO, but with advances in serologic testing, these condi-
Chronic leukemia Deep venous thrombosis tions usually are diagnosed more promptly. At this time,
Lymphoma Sarcoidosis adult Still’s disease and temporal arteritis have become the
Metastatic cancers most common autoimmune sources of FUO because they
Renal cell carcinoma remain difficult to diagnose even with the help of labora-
Colon carcinoma tory testing.
Hepatoma Multisystem inflammatory diseases such as temporal
Myelodysplastic syndromes arteritis and polymyalgia rheumatica have emerged as the
Pancreatic carcinoma
autoimmune conditions most frequently associated with
Sarcomas
FUO in patients older than 65 years.15 Elderly patients who
present with symptoms consistent with temporal arteritis
associated with an elevation of the erythrocyte sedimenta-
eral factors may limit the applicability of research literature tion rate should be referred for temporal artery biopsy.16
on FUO to everyday medical practice. These factors
include the geographic location of cases, the type of insti- MISCELLANEOUS
tution reporting results (e.g., community hospital, univer- Many unrelated pathologic conditions can present as
sity hospital, ambulatory clinic), and the specific subpopu- FUO, with drug-induced fever being the most com-
lations of patients with FUO who were studied. Despite mon.11,14 This condition is part of a hypersensitivity reac-
these limiting factors, infection remains the most common tion to specific drugs such as diuretics, pain medications,
cause of FUO in study reports.3,11,12 antiarrhythmic agents, antiseizure drugs, sedatives, certain
antibiotics, antihistamines, barbiturates, cephalosporins,
INFECTIONS salicylates, and sulfonamides (Table 3).
Of the many infectious diseases that are associated with Complications from cirrhosis and hepatitis (alcoholic,
FUO, tuberculosis (especially in extrapulmonary sites) and granulomatous, or lupoid) are also potential causes of
abdominal or pelvic abscesses are the most common.13 FUO.12,13 Deep venous thrombosis, although a rare cause
Intra-abdominal abscesses are associated with perforated of FUO, must be considered in relevant patients, and
hollow viscera (as occurs in appendicitis), diverticulitis, venous Doppler studies should be obtained.17 Factitious
malignancy, and trauma. Other common infections that fever has been associated with patients who have some
should be considered as the source of FUO include suba- medical training or experience and a fever persisting
cute bacterial endocarditis, sinusitis, osteomyelitis, and longer than six months.18 Failure to reach a definitive diag-
dental abscess.11,13 As the duration of fever increases, the nosis in patients presenting with FUO is not uncommon;
likelihood of an infectious etiology decreases. Malignancy 20 percent of cases remain undiagnosed. Even if an exten-
and factitious fever are more common diagnostic consid- sive investigation does not identify a cause for FUO, these
erations in patients with prolonged FUO.14 patients generally have a favorable outcome.19

MALIGNANCIES Evaluation of the Patient with FUO

Because of a substantial increase in the elderly population, The initial approach to the patient presenting with fever
as well as advances in the diagnosis and treatment of diseases should include a comprehensive history, physical examina-

DECEMBER 1, 2003 / VOLUME 68, NUMBER 11 www.aafp.org/afp AMERICAN FAMILY PHYSICIAN 2225
TABLE 3
Agents Commonly Associated
with Drug-Induced Fever

Allopurinol (Zyloprim) Meperidine (Demerol)

Captopril (Capoten) Methyldopa (Aldomet)
Cimetidine (Tagamet) Nifedipine (Procardia)
Clofibrate (Atromid-S) Nitrofurantoin (Furadantin)
Erythromycin Penicillin a thoughtful and focused investigation, inappropriate tests
Heparin Phenytoin (Dilantin) might be performed.
Hydralazine (Apresoline) Procainamide (Pronestyl) The preliminary evaluation helps in the formulation of
Hydrochlorothiazide (Esidrix) Quinidine a differential diagnosis and guides further studies that are
Isoniazid more invasive or expensive. These preliminary investiga-
tions should include a complete blood count, liver func-
tion test, erythrocyte sedimentation rate, urinalysis, and
tion, and appropriate laboratory testing. As the underlying basic cultures. Simple clues found during initial testing
process develops, the history and physical assessment often will guide the clinician toward one of the major sub-
should be repeated. The first step should be to confirm a groups of FUO. The decision to obtain further diagnostic
history of fever and document the fever pattern. Classic studies should be based on abnormalities found in the ini-
fever patterns such as intermittent, relapsing sustained, tial laboratory work-up and not represent a haphazard use
and temperature-pulse disparity may prove to be useful of costly or invasive modalities.
but rarely are diagnostic.20 Skin testing for tuberculosis with purified protein deriv-
In taking a history from a patient with FUO, particular ative (PPD) is an inexpensive screening tool that should be
attention should be given to recent travel, exposure to pets used in all patients with FUO who do not have a known
and other animals, the work environment, and recent con- positive PPD reaction. However, a positive PPD reaction
tact with persons exhibiting similar symptoms. In patients alone does not prove the presence of active tuberculosis. A
returning from areas where tuberculosis and malaria are chest radiograph also should be obtained in all patients to
common, the index of suspicion for these diseases should screen for possible infection, collagen vascular disease, or
be elevated. In patients who have had contact with pets or malignancy. If this initial assessment does not disclose the
other animals, diseases common to animal handlers must source of fever, more specific investigatory techniques,
be suspected. such as serology, sonography, computed tomography
The family history should be carefully scrutinized for (CT), magnetic resonance imaging (MRI), and nuclear
hereditary causes of fever, such as familial Mediterranean medicine scanning should be conducted, based on clinical
fever. The medical history also must be examined for con- suspicion.
ditions such as lymphoma, rheumatic fever, or a previous Abdominal sonography, pelvic sonography, or CT scan-
abdominal disorder (e.g., inflammatory bowel disease), the ning should be performed early in the diagnostic process to
reactivation of which might account for the fever. Finally, rule out such common causes of FUO as intra-abdominal
drug-induced fever must be considered in patients who are abscess or malignancy, depending on the primary evalua-
taking medications.19 tion.17 This testing, including directed biopsies, has greatly
Diagnostic clues often are not readily apparent on phys- reduced the need for more invasive operative studies.23
ical examination; repeated examination may be essential. MRI should be reserved for clarifying conditions found
Careful attention to the skin, mucous membranes, and through the use of other techniques or when the diagnosis
lymphatic system, as well as abdominal palpation for remains obscure. The use of radionucleotide scanning,
masses or organomegaly, is important. The physician’s such as gallium 67, technetium Tc 99m, or indium-labeled
choice of imaging should be guided by findings from a leukocytes, is warranted for detecting inflammatory condi-
thorough history and physical examination21 (e.g., a car- tions and neoplastic lesions that often are underdiagnosed
diac murmur in the presence of negative blood cultures by CT scans; however, these tests tend not to detect colla-
should be investigated with a transthoracic echocardio- gen vascular disease and other miscellaneous conditions24
gram or, if needed, transesophageal echocardiogram) (Table 4).
rather than strictly following the stepwise approach out- Endoscopic procedures may be helpful in the diagnosis
lined in Figure 1. Also, Duke’s clinical criteria include two of disorders such as inflammatory bowel disease and sar-
major and six minor criteria that help determine the like- coidosis. The newest diagnostic technique in the evaluation
lihood of endocarditis.22 [Evidence level A, validated clini- of the patient with FUO is positron emission tomography
cal decision tool] A cost-effective individualized approach (PET). This modality appears to have a very high negative
is essential to the evaluation of these patients, and without predictive value in ruling out inflammatory causes of fever.

2226 AMERICAN FAMILY PHYSICIAN www.aafp.org/afp VOLUME 68, NUMBER 11 / DECEMBER 1, 2003
 Fever of Unknown Origin

Diagnosis of Fever of Unknown Origin

Complete history and physical assessment

Positive findings Yes Order appropriate and specific

diagnostic testing.

No

CBC, electrolytes, LFT, blood culture, urinalysis,

urine culture, ESR, PPD skin test, chest radiograph

Positive results Yes Order appropriate follow-up

diagnostic testing.

No

CT of abdomen/pelvis with contrast

Assign to most likely category.

Infection Malignancies Autoimmune conditions Miscellaneous

Urine and sputum cultures Rheumatoid factor, ANA Order appropriate
for AFB, VDRL, HIV test; diagnostic tests based
Diagnosis clear?
serology for CMV, EBV, on information from
ASO titer (geographically the history.
Hematologic Nonhematologic
specific testing)
Peripheral smear, Mammography, chest
Diagnosis clear? No
serum protein CT with contrast,
electrophoresis upper/lower
endoscopy, bone
Diagnosis clear? Temporal artery biopsy,
No scan, gallium 67 scan
lymph node biopsy
Diagnosis clear?
No
TTE/TEE, lumbar puncture,
gallium 67 scan, sinus No
films (radiography or CT) Bone marrow biopsy

MRI of the brain, biopsy of suspicious

skin lesions or lymph nodes, liver
biopsy, diagnostic laparoscopy

FIGURE 1. Algorithm for the diagnosis of fever of unknown origin. (CBC = complete blood count; LFT = liver function test;
ESR = erythrocyte sedimentation rate; PPD = purified protein derivative; CT = computed tomography; AFB = acid-fast
bacilli; HIV = human immunodeficiency virus; CMV = cytomegalovirus; EBV = Epstein-Barr virus; ASO = antistreptolysin-O
antibodies; ANA = antinuclear antibody; TTE = transthoracic echocardiography; TEE = transesophageal echocardiography;
MRI = magnetic resonance imaging)

DECEMBER 1, 2003 / VOLUME 68, NUMBER 11 www.aafp.org/afp AMERICAN FAMILY PHYSICIAN 2227
Fever of Unknown Origin

5. Durack DT, Street AC. Fever of unknown origin—reexamined and

redefined. Curr Clin Top Infect Dis 1991;11:35-51.
TABLE 4
6. Durack DT. Fever of unknown origin. In: Mackowiak PA, ed. Fever:
Diagnostic Imaging in Patients with FUO basic mechanisms and management. 2d ed. Philadelphia: Lippin-
cott-Raven, 1997:237-49.
7. Konecny P, Davidson RN. Pyrexia of unknown origin in the 1990s:
Imaging Possible diagnoses time to redefine. Br J Hosp Med 1996;56:21-4.
8. Kountakis SE, Burke L, Rafie J J, Bassichis B, Maillard AA, Stiern-
Chest radiograph Tuberculosis, malignancy,
berg CM. Sinusitis in the intensive care unit patient. Otolaryngol
Pneumocystis carinii pneumonia Head Neck Surg 1997;117:362-6.
CT of abdomen or pelvis Abscess, malignancy 9. Hughes WT, Armstrong D, Bodey GP, Feld R, Mandell GL, Meyers
with contrast agent JD, et al. From the Infectious Diseases Society of America. Guide-
Gallium 67 scan Infection, malignancy lines for the use of antimicrobial agents in neutropenic patients
Indium-labeled leukocytes Occult septicemia with unexplained fever. J Infect Dis 1990;161:381-96.
10. Armstrong WS, Katz JT, Kazanjian PH. Human immunodeficiency
Technetium Tc 99m Acute infection and inflammation virus–associated fever of unknown origin: a study of 70 patients
of bones and soft tissue in the United States and review. Clin Infect Dis 1999;28:341-5.
MRI of brain Malignancy, autoimmune 11. De Kleijn EM, Vandenbroucke JP, van der Meer JW. Fever of unknown
conditions origin (FUO). I. A prospective multicenter study of 167 patients with
PET scan Malignancy, inflammation FUO, using fixed epidemiologic entry criteria. The Netherlands FUO
Study Group. Medicine [Baltimore] 1997;76:392-400.
Transthoracic or Bacterial endocarditis
12. Knockaert DC, Vanneste L J, Bobbaers HJ. Recurrent or episodic
transesophageal fever of unknown origin. Review of 45 cases and survey of the lit-
echocardiography erature. Medicine [Baltimore] 1993;72:184-96.
Venous Doppler study Venous thrombosis 13. Kazanjian PH. Fever of unknown origin: review of 86 patients
treated in community hospitals. Clin Infect Dis 1992;15:968-73.
14. Aduan RP, Fauci AS, Dale DC, Wolff SM. Prolonged fever of
FUO = fever of unknown origin; CT = computed tomography; MRI = unknown origin (FUO): a prospective study of 347 patients. Clin
magnetic resonance imaging; PET = positron emission tomography. Res 1978;26:558.
15. Knockaert DC, Vanneste LJ, Bobbaers HJ. Fever of unknown ori-
gin in elderly patients. J Am Geriatr Soc 1993;41:1187-92.
16. Epperly TD, Moore KE, Harrover JD. Polymyalgia rheumatica and
However, because of its limited availability it is too early to temporal arteritis. Am Fam Physician 2000;62:789-96.
17. Mourad O, Palda V, Detsky A. A comprehensive evidence-based
determine if PET scans will prove to be a useful diagnostic approach to fever of unknown origin. Arch Intern Med 2003;
tool in the evaluation of these patients.25,26 163:545-51.
More invasive testing, such as lumbar puncture or 18. Mackowiak P, Durack D. Fever of unknown origin. In: Mandell GL,
Douglas RG, Bennett JE, Dolin R, eds. Mandell, Douglas, and Ben-
biopsy of bone marrow, liver, or lymph nodes, should be nett’s Principles and practice of infectious diseases. 5th ed.
performed only when clinical suspicion shows that these Philadelphia: Churchill Livingstone; 2000:623-31.
tests are indicated or when the source of the fever remains 19. Arnow PM, Flaherty JP. Fever of unknown origin. Lancet 1997;
350:575-80.
unidentified after extensive evaluation. When the definitive 20. Mackowiak PA. Commentary. Fever patterns. Infectious Disease
diagnosis remains elusive and the complexity of the case Clinical Practice 1997;6:308-9.
increases, an infectious disease, rheumatology, or oncology 21. Kupferwasser LI, Darius H, Muller AM, Martin C, Mohr-Kahaly S,
Erbel R, et al. Diagnosis of culture-negative endocarditis: the role
consultation may be helpful. of the Duke criteria and the impact of transesophageal echocar-
diography. Am Heart J 2001;142:146-52.
The authors indicate that they do not have any conflicts of inter- 22. Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of
ests. Sources of funding: none reported. infective endocarditis: utilization of specific echocardiographic
findings. Am J Med 1994;96:200-9.
REFERENCES 23. Hirschmann JV. Fever of unknown origin in adults. Clin Infect Dis
1997;24:291-300.
1. Cherry DK, Woodwell DA. National Ambulatory Medical Care Sur- 24. Suga K, Nakagi K, Kuramitsu T, Itou K, Tanaka N, Uchisato H, et
vey: 2000 summary. Adv Data 2002;328:1-32. al. The role of gallium-67 imaging in the detection of foci in recent
2. Dykewicz MS. Rhinitis and sinusitis. J Allergy Clin Immunol cases of fever of unknown origin. Ann Nucl Med 1991;5:35-40.
2003;111(suppl 2):S520-9. 25. Lorenzen J, Buchert R, Bohuslavizki KH. Value of FDG PET in
3. Petersdorf RG, Beeson PB. Fever of unexplained origin: report on patients with fever of unknown origin. Nucl Med Commun
100 cases. Medicine 1961;40:1-30. 2001;22:779-83.
4. Cunha BA. Fever of unknown origin. Infect Dis Clin North Am 26. Knockaert DC, Vanderschueren S, Blockmans D. Fever of unknown
1996;10:111-27. origin in adults: 40 years on. J Intern Med 2003;253:263-75.

2228 AMERICAN FAMILY PHYSICIAN www.aafp.org/afp VOLUME 68, NUMBER 11 / DECEMBER 1, 2003