51
Fever of Unknown Origin
PHILIP A. MACKOWIAK | DAVID T. DURACK
M ost episodes of fever in humans are short-lived and do not require
diagnostic investigation or specific therapy. Some are manifestations
of more serious illnesses, most of which can be readily diagnosed and
effectively treated. However, a small but important subgroup of fevers
are both persistent and difficult to diagnose. Such puzzling fevers have
fascinated and frustrated clinicians since the earliest days of clinical
thermometry,1 resulting in a welter of clinical publications. The two
most important of these, from a historical perspective, are the classical
treatises, Prolonged and Perplexing Fevers, published by Keefer and
Leard in 1955,2 and Fever of Unknown Origin: Report on 100 Cases, by
Petersdorf and Beeson in 1961.3
Terminology and Definitions
In the United States, the term fever of unknown origin (FUO) is gener-
ally used.4 In other countries an alternative term, pyrexia of unknown
origin (PUO), is often used.5
The first formal definition of FUO to gain broad acceptance was
proposed by Petersdorf and Beeson nearly five decades ago: fever
higher than 38.3 C (101 F) on several occasions, persisting without
diagnosis for at least 3 weeks in spite of at least 1 weeks investigation
in hospital.3 Later investigators have modified and extended this clas-
sical definition to reflect evolutionary changes in clinical practice.6,7
These changes include the conduct of most diagnostic tests in the
outpatient setting rather than in the hospital, the increasing number
of immunocompromised patients (especially those with neutropenia),
the proliferation of increasingly complex surgical and intensive care
treatment protocols, and the advent of human immunodeficiency
virus (HIV) infection leading to the acquired immunodeficiency syn-
drome (AIDS). In response to this evolving environment, cases of FUO
are currently codified into four distinct subclasses of the disorder:
classical FUO, health careassociated FUO, immune-deficient FUO,
and HIV-related FUO (Table 51-1).Computed axial tomography,
magnetic resonance imaging, ultrasound imaging, nucleic acidbased
diagnostic testing, and rapid tests for pathogens have changed the
landscape of FUO. One could argue that the definitions of types of
FUO have much overlap and are not as clinically useful as they were
in the past.
Classical Fever of Unknown Origin
Classical FUO refers to the type of FUO defined by Petersdorf and
Beeson in 1961.3 The only alteration to their definition required to
conform to modern medical practice is to incorporate investigation in
the outpatient setting, which today has become the preferred venue
for evaluation and treatment. Most patients with classical FUO have
subacute or chronic symptoms and therefore can be safely investigated
as outpatients. In a series of 53 such patients, for example, the median
duration of fever before diagnosis was 40 days.1 Even though the
tempo and accuracy of clinical investigation have improved in recent
years, many cases of classical FUO continue to elude diagnosis for
extended periods, as reflected in the findings of a Japanese survey
conducted from 1982 to 1992, in which an average of 49 days (range
8 to 217 days) was required to complete the diagnostic evaluation.8
Of the many publications concerning the etiology of FUO,3,9-15 most
have dealt with classical FUO rather than with the three other more
recently defined subclasses listed earlier.7 Over the years a recurrent
theme has become clear: of the myriad disorders causing classical FUO,
almost all fall within one of five categories: infections, neoplasms, con-
nective tissue diseases, miscellaneous other disorders, and undiag-
nosed illnesses. The relative frequencies of individual diagnoses within
these five categories vary depending on the geographic region, ages of
the patients, type of medical practice, and other factors (Fig. 51-1).
In most series, infections have comprised the largest category,
accounting for 25% to 50% of cases. However, in patients older than 65
years, infections become less common, falling into second or third place
as a cause of classical FUO.8,9 In the series of Knockaert and associates,9
infection was the cause of FUO in only 25% of cases 65 years of age or
older; temporal arteritis and various connective tissue diseases
accounted for 31% of cases, and tumors for 12%. Only 8% of cases went
undiagnoseda percentage substantially lower than that reported in
surveys involving younger adults, in which as many as 30% of cases
remain undiagnosed.16 The longer the duration of fever before medical
consultation, the less likely that a definitive diagnosis will be made.17
Among the infections responsible for classical FUO, abscesses,
endocarditis, tuberculosis, and complicated urinary tract infections
have consistently been among the most important. These tend to vary
in incidence according to locale. Visceral leishmaniasis, for example,
although absent from most series of classical FUO, accounted for 8%
of cases in a study reported in 1997 from Spain.18 Other examples of
causes of classical FUO in distinct populations include familial Medi-
terranean fever in Ashkenazi Jews; Kikuchis disease, an unusual form
of necrotizing lymphadenitis seen primarily in Japan19; and TRAPS
(TNF-receptor associated periodic fever), formerly called familial
Hibernian fever, an inherited periodic fever syndrome described origi-
nally in Ireland.20 The miscellaneous category contains both varied and
individually rare causes of classical FUO (Table 51-2).
Of the connective tissue diseases responsible for classical FUO, Stills
disease (juvenile rheumatoid arthritis), other variants of rheumatoid
arthritis, and systemic lupus erythematosus predominate in younger
patients, whereas temporal arteritis and polymyalgia rheumatica syn-
dromes are more common in elderly patients.
Malignant neoplasms, another important cause of FUO, can induce
fever directly through the production and release of pyrogenic cyto-
kines, as in the case of certain lymphomas. They can also generate
fevers indirectly by undergoing spontaneous or induced necrosis or by
creating conditions conducive to secondary infections, such as post-
obstructive pneumonia.21 Although hypernephromas are often listed
as a notable cause of FUO,3,22 in reality this tumor only rarely occurs
with fever.23
The relative frequency with which the major diagnostic categories
are represented in series of classical FUO varies according to both the
era in which the series was published6,13 and its country of origin (see
Fig. 51-1).8,15,24,25 Since the mid-1900s, the frequency with which infec-
tions and malignant neoplasms have been identified as causes of clas-
sical FUO has fallen steadily, whereas the proportion of miscellaneous
causes and undiagnosed conditions has risen.24 However, in develop-
ing countries, the frequency with which infections are diagnosed has
changed little.15 Consequently, in these countries malignant neoplasms
and connective tissue disorders are comparatively less important as
causes of classical FUO than in developed countries (see Fig. 51-1).15
INFANTS AND CHILDREN
The diseases responsible for classical FUO in infants differ from those
in older children and adults. Respiratory infections cause classical FUO
779
780 PART II Major Clinical Syndromes

TABLE
51-1 Summary of Definitions and Major Features of the Four Subtypes of Fever of Unknown Origin (FUO)

Classical FUO Health CareAssociated FUO Immune-Deficient FUO HIV-Related FUO

Definition >38.0 C, >3 wk, >2 visits or 3 >38.0 C, >3 days, not present >38.0 C, >3 days, negative 38.0 C, >3 wk for outpatients, >3 days for
days in hospital or incubating on admission cultures after 48 hr inpatients, HIV infection confirmed
Patient Community, clinic, or hospital Acute care hospital Hospital or clinic Community, clinic, or hospital
location
Leading causes Cancer, infections, inflammatory Health careassociated Majority due to infections, HIV (primary infection), typical and
conditions, undiagnosed, infections, postoperative but cause documented in atypical mycobacteria, CMV,
habitual hyperthermia complications, drug fever only 40%-60% lymphomas, toxoplasmosis,
cryptococcosis, immune reconstitution
inflammatory syndrome (IRIS)
History Travel, contacts, animal and Operations and procedures, Stage of chemotherapy, drugs Drugs, exposures, risk factors, travel,
emphasis insect exposure, medications, devices, anatomic administered, underlying contacts, stage of HIV infection
immunizations, family history, considerations, drug immunosuppressive
cardiac valve disorder treatment disorder
Examination Fundi, oropharynx, temporal Wounds, drains, devices, Skin folds, IV sites, lungs, Mouth, sinuses, skin, lymph nodes, eyes,
emphasis artery, abdomen, lymph nodes, sinuses, urine perianal area lungs, perianal area
spleen, joints, skin, nails,
genitalia, rectum or prostate,
lower limb deep veins
Investigation Imaging, biopsies, sedimentation Imaging, bacterial cultures CXR, bacterial cultures Blood and lymphocyte count; serologic
emphasis rate, skin tests tests; CXR; stool examination; biopsies of
lung, bone marrow, and liver for cultures
and cytologic tests; brain imaging
Management Observation, outpatient Depends on situation Antimicrobial treatment Antiviral and antimicrobial protocols,
temperature chart, protocols vaccines, revision of treatment regimens,
investigations, avoidance of good nutrition
empirical drug treatments
Time course of Months Weeks Days Weeks to months
disease
Tempo of Weeks Days Hours Days to weeks
investigation
CMV, cytomegalovirus; CXR, chest radiograph; HIV, human immunodeficiency virus; IV, intravenous.
Adapted from Durack DT. Fever of unknown origin. In: Mackowiak PA, ed. Fever. Basic Mechanisms and Management. 2nd ed. Philadelphia: Lippincott-Raven; 1997:237-249.

TABLE
51-2 Examples of Rare Miscellaneous Causes of Fever

Addisons disease Infected urachal cyst

8% Adult-onset Stills disease Inflammatory bowel disease
Alcoholic hepatitis Kikuchi-Fujimoto disease
23% 27% Allergic alveolitis
Lofgren syndrome
20% 33% Aortic dissection
Lymphomatoid granulomatosis
Aortitis
Atrial myxoma Metal fume fever
21% Autoimmune cholangitis Myeloproliferative syndromes
13% 13%
Bartonellosis Pancreatitis
25% Parathyroid apoplexy
17% Behets syndrome
Paroxysmal hemoglobinurias
Carcinomatous meningitis
Pericarditis
Castlemans disease
USA Europe Periodic fever
Chronic meningitis
Pheochromocytoma
Cirrhotic fever
Polyarteritis nodosa
5% Cyclic neutropenia
Postpericardiotomy syndrome
13% Drug fever and other Pulmonary emboli
15% hypersensitivities
Resorbing hematoma
34% Erythema multiforme Retroperitoneal fibrosis
9% 50% 20% Fabrys disease Rosai-Dorfman disease
Factitious fever Sarcoidosis
Familial Hibernian fever Schnitzlers syndrome
22% 13% Familial Mediterranean fever
20% Sinusitis
Giant coronary aneurysm Serum sickness
Granulomatous hepatitis Sjgrens syndrome
Granulomatous peritonitis Subacute necrotizing lymphadenitis
India Total
Hantavirus infection Thrombotic thrombocytopenic
Hemoglobinopathies purpura
Infections Miscellaneous Undiagnosed Hemolytic anemias Thyroiditis and thyrotoxicosis
Hemophagocytic syndrome Veno-occlusive disease
Neoplasms Connective tissue diseases Histiocytosis X Vitamin B12 deficiency
Human picornavirus infection
Wegeners granulomatosis
Figure 51-1 The five main etiologic categories of fever of unknown Hypereosinophilic syndrome
Whipples disease
origin, comparing their frequency in series from three geographic Immunoblastic lymphadenopathy
regions. (Redrawn from Durack DT. Fever of unknown origin. In: Adapted from Durack DT. Fever of unknown origin. In: Mackowiak PA, ed. Fever.
Mackowiak PA, ed. Fever. Basic Mechanisms and Management. 2nd ed. Basic Mechanisms and Management. 2nd ed. Philadelphia: Lippincott-Raven;
Philadelphia: Lippincott-Raven; 1997:237-249.) 1997:237-249.
 51 Fever of Unknown Origin 781

in infants more often than in children older than 12 months or in TABLE

adults.10 The relative frequency of infections as the cause of FUO in 51-4 Causes of Fever in the Returned Traveler*
infants is high, owing to the fact that connective tissue diseases and MacLean et al114 Doherty et al115
cancers are rare in this age group. Kawasaki disease occurs predomi- Diagnosis (n = 587) (n = 195)
nantly in children younger than 5 years. Whereas connective tissue Malaria 32 42
diseases are rarely seen in children younger than 12 months, Stills Hepatitis 6 3
disease is a leading cause of FUO in older children and young adults. Respiratory infection 11 2.6
Joint involvement in children with FUO usually signifies a serious Urinary tract infection/pyelonephritis 4 2.6
underlying disorder, such as a connective tissue disease, endocarditis, Dysentery 4.5 5.1
or leukemia.10 Dengue fever 2 6.2
In a series of 146 pediatric cases of FUO, Jacobs and Schutze26 estab- Enteric fever 2 1.5
lished a specific diagnosis in only 84 (57.5%). Of these, 64 (43.8%) had Tuberculosis 1 2
infections, 11 (7.5%) autoimmune disorders, 4 (2.7%) malignant neo- Rickettsial infection 1 0.5
plasms, and 5 (3.4%) a variety of other disorders, such as drug-induced Acute HIV infection 0.3 1.0
fever, sarcoidosis, and mercury poisoning. The most common infec-
Amebic liver abscess 1 0
tious diseases diagnosed in their series were Epstein-Barr virus (EBV)
Other miscellaneous infections 4.3 9.2
infection (15%), osteomyelitis (10%), bartonellosis (5%), and urinary
Miscellaneous noninfectious causes 6 1
tract infections (4%).
Undiagnosed 25 24.6
*Values listed are percentages.
ELDERLY PERSONS

Includes upper respiratory tract infection, pneumonia, and bronchitis.
HIV, human immunodeficiency virus.
One of the most striking features of classical FUO in patients older From Suh KN, Kozavsky PE, Keystone JS. Evaluation of fever in the returned traveler.
than 65 years is the relatively high frequency with which connective Travel Med. 1999;83:997-1017.
tissue diseases are identified as the cause of the illness (Table 51-3).27,28
In developed countries, connective tissue diseases surpass even infec-
tions as the leading cause of classical FUO in the elderly.8,9,29 This is abscess or dengue, are occasionally diagnosed, especially among inter-
primarily because the temporal arteritis and polymyalgia rheumatica national travelers returning from the tropics. Of the many febrile
syndromes are common in this setting.8,30,31 Unfortunately these diag- conditions encountered among returning travelers (Table 51-4),
noses are frequently missed or delayed because their symptoms are malaria, typhoid fever, and acute HIV infection are the ones most
subacute and nonspecific. In elderly patients in whom infections are likely to manifest as FUO.
identified as the cause of FUO, intra-abdominal abscesses, complicated
urinary tract infections, tuberculosis, and endocarditis have predomi-
nated.9,14 For unclear reasons, factitious fever is a rare cause of FUO in Health CareAssociated Fever of
older adults. Relatively few cases of FUO go undiagnosed in elderly
patients (see Table 51-3). The occurrence of classical FUO in an elderly
Unknown Origin
patient has a distinctly poorer prognosis than for the younger patient Health careassociated FUO, as the name implies, is a condition in
because of the relatively high incidence of malignancies in the elderly.32 which patients first manifest fever during active medical treatment for
some other illness. Such FUO cases, as might be expected, are fre-
quently attributable to risk factors encountered in the health care
RETURNED TRAVELERS
environment, including surgical procedures, urinary and respiratory
Fever in returned travelers (see Chapter 330) is most often due to tract instrumentation, intravascular devices, drug therapy, and immo-
common infections, such as malaria and respiratory or urinary tract bilization. Leading examples include drug fever, septic thrombophle-
infections.31,33 However, exotic causes of fever, such as amebic liver bitis, recurrent pulmonary emboli, and Clostridium difficile colitis.7,34

POSTOPERATIVE PATIENTS
Several reports have highlighted the fact that it is often difficult to
TABLE Final Diagnosis in Elderly Compared with Younger identify the precise cause of postoperative fevers. In a series of 537
51-3 Patients with Fever of Unknown Origin consecutive patients undergoing major gynecologic surgery, 211
Diagnosis <65 Years (n = 152) >65 Years (n = 201) (39%) developed postoperative fever.35 Of 77 blood cultures per-
Infections 33 (21%) 72 (35%) formed on these patients, none was positive. Although 11 of 106 (10%)
Abscess 6 25 urine cultures were positive and 5 of 54 (9%) chest radiographs were
Endocarditis 2 14 abnormal, a specific pathologic process was detected in only 8% of
Tuberculosis 4 20 febrile patients. In a study of postoperative fever following tonsillec-
Viral infections 8 1 tomy in children, Anand and colleagues36 found no association between
Other 13 12 cultures of blood or the core or the surface of tonsils and the incidence
Tumors 8 (5%) 37 (19%) or severity of fever, and concluded that postoperative fever, at least in
Hematologic 3 19 the 24 hours following tonsillectomy, is rarely the result of infection.
Solid 5 18 In another series concerned with the etiology of persistent postop-
Multisystem diseases* 27 (17%) 57 (28%) erative fever in patients undergoing total joint arthroplasty, few defini-
Miscellaneous 39 (26%) 17 (8%) tive diagnoses were established, causing the authors to conclude that
No diagnosis 45 (29%) 18 (9%) postoperative fever is a normal component of the inflammatory
*Rheumatic diseases, connective tissue disorders, vasculitis (including temporal response to this type of major surgery.37
arteritis), polymyalgia rheumatica, and sarcoidosis.

Includes factitious fever (seven cases), habitual hyperthermia (five cases), and
drug-induced fever (three cases). INTENSIVE CARE UNIT PATIENTS
Adapted from Iikuni Y, Okada J, Kondo H, et al. Current fever of unknown origin
1982-1992. Intern Med. 1994;33:67-73; Knockaert DC, Vanneste LJ, Bobbaers HJ. Fever Fever is common in intensive care units, most often developing rela-
of unknown origin in elderly patients. J Am Geriatr Soc. 1993;41:1187-1192. tively early after admission to the unit, in which case it tends to be of
782 PART II Major Clinical Syndromes
noninfective origin and carries a favorable prognosis.38 Prolonged
fever, however, is associated with a worse prognosis. Health care
associated sinusitis, often a complication of mechanical ventilation
arising from supine positioning and the use of endotracheal, gastric,
and feeding tubes is common39 and should always be considered when
evaluating FUO in intensive care unit patients.
STROKE PATIENTS
In patients with a recent stroke, fever is usually the result of an infec-
tion, most commonly a urinary tract infection related to urinary cath-
eterization. However, in some cases, a focus of infection cannot be
identified, and when the fever does not respond to empirical antibiotic
treatment it is presumed to be due to the stroke itself. In a study of
330 patients hospitalized for acute stroke, Georgilis and associates40
observed that noninfective fevers were most often associated with
intracranial mass effects and tended to occur earlier after the onset of
stroke than fevers due to infection.
Immune-Deficient Fever of
Unknown Origin
Obviously, various forms of immunosuppression predispose more or
less strongly to a wide variety of infectious complications. Thus, it is
not surprising that immunosuppressed patients have perhaps the
highest incidence of FUO of any group of patients. In a recent series
of 116 hematology-oncology patients, for example, Engelhart and
associates41 observed 33 FUOs in 28 patients, for an overall rate of 8.2
episodes per 1000 patient days. Because of impaired immune responses,
signs of inflammation other than fever are notoriously absent or
diminished in such patients, leading to atypical clinical manifestations
and absence of radiologic abnormalities in what otherwise would be
readily diagnosed infections.42
In patients with impaired cell-mediated immunity, FUO is often
due to conditions other than pyogenic bacterial infections, as illus-
trated in a recent survey of transplant recipients by Chang and co-
workers.43 In that series, infections were the cause of 7 of the 12
(58%) episodes of fever in which a diagnosis could not be established
during 3 days of intensive investigation; these included five cases of
human herpesvirus 6 infection, one of varicella-zoster virus infection
with atypical skin lesions, and one of pneumonia due to Serratia
marcescens (in which the pulmonary infiltrate was initially inapparent
radiographically). Three (25%) cases were due to noninfectious con-
ditions: malignant neoplasm, drug-induced fever, and adrenal insuf-
ficiency. In two (17%) cases, the cause of the fever could not be
determined.
NEUTROPENIC FEVER OF UNKNOWN ORIGIN
Neutropenia is a dangerous condition that can be considered a special
subclass of immunodeficiency. The number of patients with episodes
of neutropenia resulting from cytotoxic therapy or from various
malignancies affecting the bone marrow is rising, even as the median
duration of neutropenia in such patients is falling, owing to
the increasing application of treatment with colony-stimulating
factors.
Episodes of fever are common in patients with neutropenia. Many
such episodes are short-lived, because they either respond quickly to
treatment or are manifestations of rapidly fatal infections. Because
bacteremia and sepsis are frequent causes, empirical broad-spectrum
antibiotics should be administered promptly, without waiting for the
results of cultures, when fever develops in neutropenic patients.
However, only about 35% of prolonged episodes of febrile neutropenia
respond to broad-spectrum antibiotic therapy. Although practitioners
often assume that if fever does not respond promptly to antibacterial
therapy, fungal infection must be responsible, other causes are equally
likely to be identified (Table 51-5).42
TABLE Possible Causes of Fever in Neutropenic Patients
51-5 Not Responding to Broad-Spectrum Antibiotics
Approximate Frequency
in High-Risk Patients
Causes (%)
Fungal infections susceptible to empirical therapy 40
Fungal infections resistant to empirical antifungal 5
therapy
Bacterial infections (with cryptic foci, biofilms, and 10
resistant organisms)
Toxoplasma gondii, mycobacteria, or fastidious 5
pathogens (legionella, mycoplasma,
Chlamydophila pneumoniae, bartonella)
Viral infections (herpesviruses, cytomegalovirus, 5
Epstein-Barr virus, human herpesvirus 6,
varicella-zoster virus, herpes simplex virus,
parainfluenza virus, respiratory syncytial virus,
influenza viruses)
Graft-versus-host disease after hematopoietic 10
stem cell transplantation
Undefined (e.g., drug fever, toxic effects of 25
chemotherapy, antitumor responses, undefined
pathogens)
From Corey L, Boeckh M. Persistent fever in patients with neutropenia. N Engl J Med.
2002;346:222-224.
Copyright 2002 Massachusetts Medical Society. All rights reserved.
Human Immunodeficiency
VirusRelated Fever of Unknown Origin
Episodes of fever are commonplace in patients infected with HIVa
special subgroup of immunodeficient patients, now so numerous that
a separate category of FUO is justified.44-46 The primary phase of HIV
infection is characterized by a mononucleosis-like illness in which
fever is a prominent feature. All too often, primary HIV infection
eludes diagnosis because the illness is nonspecific and precedes sero-
conversion. For this reason it represents an important cause of HIV-
associated FUO. Once symptoms of the primary phase of the HIV
infection resolve, HIV-infected patients enter a long period of sub-
clinical infection during which they are usually afebrile.47 However, in
the later phases of untreated HIV infection, episodes of fever become
common, often signifying a superimposed illness. Many of these
are potentially devastating opportunistic infections, which tend to
present in atypical fashion owing to the tendency of the disordered
immune response or prior therapy, or both,48 to distort their clinical
manifestations.
Once highly active antiretroviral therapy (HAART) has been started
and HIV viral load effectively suppressed, the frequency of FUO in
HIV-infected patients falls markedly. In one study, the frequency of
FUO was 3% in untreated patients compared with 0.6% in those
receiving therapy.49 Mycobacterial infections have been the common-
est cause of FUO in such patients; collagen-vascular diseases have been
distinctly uncommon.50 In a recent series reported by Armstrong and
colleagues,51 the etiology of the FUO was identified in 56 of 70 (80%)
patients. A single etiology was identified in 43; 3 distinct causes of FUO
were identified in 3 cases, and 2 causes in 10 others. The 70 causes of
FUO in 56 patients in whom the etiology was determined are listed in
Table 51-6. Most cases of FUO in HIV-infected patients are the result
of opportunistic infections, the specific frequencies of which are dic-
tated, at least in part, by geographic variation in the prevalence of these
infections.52
Clinical Evaluation of Fever of
Unknown Origin
The evaluation of a patient with FUO typically includes a comprehen-
sive history, verification that the patient actually has fever, consider-
ation of the fever pattern, repeated physical examinations, a host of
 51 Fever of Unknown Origin 783

TABLE Diseases Established as the Etiology of Fever in 70

intermittent, hectic, quotidian, picket fence, sustained, quartan, and
51-6 Cases of HIV-Associated Fever of Unknown Origin saddleback fevers. Such terms have been used to codify fever patterns
No. (%) of Times Diagnosis
into general categories in an attempt to enhance their diagnostic
Etiology Was Established utility.56 A few, such as the Pel-Ebstein pattern seen in some cases of
Infection Hodgkins disease, the typhus inversus (i.e., reversal of the normal
DMAC 22 (31) diurnal pattern) in some cases of disseminated tuberculosis, and the
PCP 10 (13) pulse-temperature disassociation sometimes seen in typhoid fever,
CMV 8 (11) have been posited as having diagnostic value (Fig. 51-2). Unfortu-
Histoplasmosis 5 (7)
nately, with the possible exception of the well-known periodicity of
Viral (not CMV)* 5 (7)
tertian and quartan malaria, these fever patterns are neither sensitive
Bacterial 4 (5)
nor specific enough for diagnosis of any disease.
Mycobacterium tuberculosis 4 (5)
Thus, although fever patterns per se are rarely diagnostic, they occa-
Fungal (not histoplasmosis) 2 (3)
sionally offer useful information56 and should be considered in the
Parasitic 2 (3)
context of other signs, symptoms, and laboratory data. The resolution
of fever after the institution of a disease-specific therapy (for example,
Mycobacterium genavense 1 (1)
empirical therapy for suspected tuberculosis) sometimes provides
Total 63 (88)
good evidence to support a presumptive diagnosis.
Neoplasia
In pediatric populations, the height of a fever correlates roughly
Lymphoma 5 (7)
with the likelihood of bacteremia. McCarthy and co-workers57,58
Kaposis sarcoma 1 (1)
reported that in young children with febrile illnesses, the likelihood of
Total 6 (8)
bacteremia is 7% in children with temperatures of 40 C or less, 13%
Miscellaneous
with temperatures of 40.5 C to 41 C, and 26% with temperatures of
Drug fever 2 (3)
41.1 C or greater. It is commonly believed that there may be a similar
Castlemans disease 1 (1)
relationship between high fever and the likelihood of bacteremia in
Total 3 (4)
adults; this belief has not been substantiated. In any case, the relation-
*Includes hepatitis C, hepatitis B, adenovirus pneumonia, herpes simplex esophagitis,
and varicella-zoster encephalitis (one case each).
ship is at best loose, even in children, with numerous examples of

Includes disseminated cryptococcosis and pulmonary aspergillosis (one case each). bacteremia in which there is little or no fever and nonbacteremic

Includes cerebral toxoplasmosis and disseminated cryptosporidiosis (one case each). conditions, such as drug-induced fever, thrombophlebitis, and recur-
CMV, cytomegalovirus; DMAC, disseminated Mycobacterium avium complex; HIV, rent pulmonary emboli, in which extremely high fevers sometimes are
human immunodeficiency virus; PCP, Pneumocystis carinii pneumonia.
From Armstrong WS, Katz JT, Kazanjian PH. Human immunodeficiency virus
encountered.
associated fever of unknown origin: A study of 70 patients in the United States and
review. Clin Infect Dis. 1999;28:341-345.
PHYSICAL EXAMINATION
In the investigation of FUO, several aspects of the physical examina-
laboratory investigations, key imaging studies, and invasive diagnostic tion should be accorded closer scrutiny than generally given during
procedures (Table 51-7). the evaluation of other illnesses (Table 51-8). Frequently, key physical
abnormalities in patients with FUO are so subtle as to require repeated
examinations to be appreciated. Examples include the nodular or
HISTORY
weakly pulsatile temporal artery of temporal arteritis, the telltale oral
It is axiomatic that a comprehensive history is a cornerstone of the ulcers of disseminated histoplasmosis or Behets syndrome, the
evaluation of any complex illness such as FUO. The history can be choroid granuloma or epididymal nodule of extrapulmonary tubercu-
especially important in determining the choice of the initial laboratory losis, the testicular nodule of polyarteritis nodosa, and the vague rectal
investigations. Particular attention should be given to recent travel, fluctuance of a perirectal abscess. The diagnostic yield of the physical
exposure to pets and other animals, the work environment, and recent examination alone in the evaluation of FUO has not been studied
contact with people exhibiting similar symptoms. The family history directly. Nevertheless, in two pediatric series, abnormal physical
should be carefully scrutinized for possible hereditary causes of fever
for example, familial Mediterranean fever. Likewise, the past medical
history must be scrutinized for prior episodes of FUO and for any
previously diagnosed conditions, such as lymphoma, rheumatic fever, TABLE General Diagnostic Evaluation of Patients with Fever
or intra-abdominal disorders, complications or reactivation of which 51-7 of Unknown Origin
might account for the source of fever. In cases of recurrent FUO, rare Comprehensive history
disorders are more common than in nonrecurrent cases and are also Repeated physical examinations
Complete blood count
more likely to remain undiagnosed.53 Finally, a complete list of the Routine blood chemistry determinations
patients medications must be obtained, so that each may be evaluated Urinalysis, including microscopic examination
as a potential source of drug-induced fever. Chest radiograph
Erythrocyte sedimentation rate
Antinuclear antibodies
VERIFICATION OF FEVER AND FEVER PATTERN Rheumatoid factor
Blood cultures: three or more separate specimens obtained in absence of
The next step in the evaluation of the patient with FUO is to verify the antimicrobial therapy
presence of fever. The importance of this step should be self-evident, Cytomegalovirus IgM antibodies or viral detection in blood
Heterophile antibody test in children and young adults
yet it is often overlooked. In fact, in a series of 347 patients admitted Tuberculin skin test
to the National Institutes of Health for investigation of prolonged Computed tomography of abdomen, pelvis, or other sites
fever, 35% were ultimately determined either not to have significant Magnetic resonance imaging
fever at all, or to have fever of factitious origin.54 Radionuclide scans
Human immunodeficiency virus antibodies or viral detection assay
Clinicians have endeavored to diagnose particular diseases by ana- Further evaluation of any abnormality detected by above tests
lyzing the associated fever patterns since the earliest days of clinical Venous duplex imaging of lower limbs
thermometry.55 These efforts have given rise to an extensive and fre- Adapted from Arrow PM, Flaherty JP. Fever of unknown origin. Lancet.
quently arcane terminology, including descriptors such as remittent, 1997;350:575-580, with permission from Elsevier.
784 PART II Major Clinical Syndromes

Day 1st 2nd 3rd 4th 5th 6th Day 1st 2nd 3rd 4th 5th 6th Day 1st 2nd 3rd 4th 5th 6th 7th
107 M E M E M E M E M E M E 107 M E M E M E M E M E M E 107 M E M E M E M E M E M E M E

105 105 105

Temperature 103 103 103
101 101 101
99 99 99
97 97 97
A Simple Tertian Simple Quartan Malignant Tertian

Date 1st week 2nd week 3rd week 4th week

Day of disease 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
Pulse Temp.
170 108
160 107
150 106 Temperature
140 105
130 104
120 103
110 102 Pulse
100 101
90 100
80 99
70 98
60 97
50 96
40 95
Respirations
B Stools

106
105
104
103
Temperature

102
101
100
99
98.6
98
97
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
C Days

Time
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
AM
PM
107
106
105
104
Temperature

103
102
101
100
99
98
97
D 96

Figure 51-2 Distinctive fever patterns. A, Malaria. B, Typhoid fever (demonstrating relative bradycardia). C, Hodgkins disease (Pel-Ebstein
pattern). D, Borreliosis (relapsing fever pattern). (From Woodward TE. The fever pattern as a clinical diagnostic aid. In: Mackowiak PA, ed. Fever.
Basic Mechanisms and Management. 2nd ed. Philadelphia: Lippincott-Raven; 1997:215-236.)

findings were reported to have contributed to the diagnosis in 60% of

cases of FUO.59,60 In half of these, the abnormalities were detected only
after repeated examinations. A vigorous search for lymphadenopathy
is recommended. When enlarged lymph nodes are detected in a patient
with FUO, lymph node biopsy is often undertaken. However, with the
exception of the lymphomas, the diagnostic yield of lymph node
biopsy in FUO is disappointingly low.59,61
LABORATORY INVESTIGATIONS
The literature is replete with algorithms indicating which laboratory
tests should be performed to evaluate FUO.62-68 Although useful as
general guides, blind application of such algorithms may result in
performance of an excessive number of tests. They should be
selectively applied using clues gleaned from the history and physical
 51 Fever of Unknown Origin 785

Examples of Subtle Physical Findings Having

studies have occasionally been encountered, even in cases of abscesses
TABLE in solid organs, as a result of distortions of normal anatomy, small
51-8 Special Significance in Patients with Fever of
Unknown Origin abscess size, or failure to use both oral and intravenous contrast
Body Site Physical Finding Diagnosis agents.74 CT pulmonary angiograms can be very helpful in diagnosis
Head Sinus tenderness Sinusitis of pulmonary emboli. Magnetic resonance imaging (MRI) is especially
Temporal Nodules, reduced Temporal arteritis useful for evaluation of the central nervous system and in the abdomen,
artery pulsations the spleen and lymph nodes.
Oropharynx Ulceration Disseminated histoplasmosis Scanning with labeled autologous leukocytes has been used for
Tender tooth Periapical abscess evaluating cases in which infections and malignant neoplasms are the
Fundi or Choroid tubercle Disseminated granulomatosis* cause of FUO. Although this scan has generally fallen out of favor as
conjunctivae Petechiae, Roths spot Endocarditis contributing very little to diagnosis, there are reports that labeled
Thyroid Enlargement, tenderness Thyroiditis leukocyte scans can provide a higher diagnostic yield than that obtained
Heart Murmur Infective or marantic by either CT or ultrasound scanning.75,76 Another technique for evalu-
endocarditis ating undiagnosed fever, especially FUO, is gallium-67 (67Ga) scan-
Abdomen Enlarged iliac crest lymph Lymphoma, endocarditis, ning. Its ability to image inflammation was first described by Lavender
nodes, splenomegaly disseminated granulomatosis* and colleagues in 1971.77 Shortly thereafter, Hilson and Maisey78
Rectum Perirectal fluctuance, Abscess reported positive results by 67Ga scanning in 50 of 67 patients with
tenderness FUO, of whom 32 had abscesses. The 67Ga scan was considered at one
Prostatic tenderness, Abscess time to be particularly effective in visualizing chronic infections and
fluctuance
lymphomas,79 but this older test for FUO has been largely displaced by
Genitalia Testicular nodule Periarteritis nodosa
CT and MRI scans. Positron emission tomography (PET) using the
Epididymal nodule Disseminated granulomatosis
positron-emitting glucose analogue, 18F-fluoro-2-deoxy-d-glucose
Lower Deep venous tenderness Thrombosis or
extremities thrombophlebitis (FDG), has generally not been helpful in the workup of patients with
Skin and nails Petechiae, splinter Vasculitis, endocarditis FUO but there are reports to the contrary.80-82 In a prospective com-
hemorrhages, parison of the two diagnostic modalities, Meller and colleagues80
subcutaneous nodules, reported a sensitivity of 81% and a specificity of 86% for PET/FDG
clubbing imaging in detecting the focus of the fever, as compared with 67% and
*Includes tuberculosis, histoplasmosis, coccidioidomycosis, sarcoidosis, and syphilis. 78%, respectively, for 67Ga scintigraphy. In another series of patients

See text for note on the nonspecificity of lymphadenopathy identified on physical
examination. with FUO reported by Lorenzen and associates,82 no patient with a
negative FDG PET had a localizable source of fever identified by any
other imaging study.
examination to direct the choice and sequence of tests. When formu-
lating a diagnostic plan for FUO patients, the clinician should remem-
INVASIVE DIAGNOSTIC PROCEDURES
ber the old saying that the cause is more often a common disease
presenting in atypical fashion, than a rare disease presenting in typical Histopathologic examination of tissues obtained by excisional biopsy,
fashion. needle biopsy, or laparotomy can provide a definitive diagnosis in
In most series, noninvasive laboratory tests have yielded the diag- some cases, but in most published series of FUO patients, biopsy gave
nosis in approximately a quarter of the cases.9,24,25 The most useful of the final answer in less than half.74 The majority of patients with FUO
these have been serologic tests for microbial pathogens and for various undergo at least one such procedure, even though the diagnostic yield
rheumatologic disorders. Examination of blood smears is occasionally is only fair, with an average of two or three, sometimes even more,
diagnostic, especially in patients with tick-borne or louse-borne relaps- separate biopsies required to establish a final diagnosis.13,83 The diag-
ing fever, anaplasmosis, and ehrlichiosis. Paradoxically, the advent of nostic yield of operative and CT-guided biopsies is higher than that of
enhanced microbial culture systems has had less of an impact on the old-style bedside biopsy procedures.13 For this reason, bedside biopsies
proportion of successful diagnoses than might have been anticipated. should rarely be performed today unless guided by localizing informa-
This is because modern culture systems have become so proficient at tion gained from imaging studies. An important exception to the
recovering fastidious bacteria, mycobacteria, and fungi from blood injunction against blind biopsies concerns the temporal artery, which
that they provide the diagnosis promptly, before the time required to may merit biopsy in an elderly FUO patient with an erythrocyte sedi-
meet the definitions of FUO has elapsed. Bone marrow examination mentation rate greater than 50 to 100 mm/hour, even in the absence
should be considered for diagnosis of suspected granulomatous dis- of localizing signs.9
eases (e.g., tuberculosis, histoplasmosis, and sarcoidosis), carcinoma- Exploratory laparotomy, once a prominent procedure in the workup
tosis and hemophagocytic syndrome, especially in patients with of FUO, is rarely performed today, unless localized abdominal physical
abnormal complete blood counts (CBCs).69 In transplant patients, signs or imaging findings, or both, are present. This is because few
procalcitonin levels may be of use in differentiating infection from abdominal anatomic abnormalities are currently missed by CT scan-
acute organ rejection, in that elevated levels are seen in the former but ning or MRI, leaving only rare cases of vasculitis, polyarteritis nodosa,
not the latter condition.70 granulomatous disease, and chronic cholecystitis to be diagnosed by
laparotomy.72 Laparoscopy, including laparoscopic liver biopsy, has
helped further obviate the need for laparotomy.84 Blind liver biopsy,
IMAGING STUDIES
even in patients with known liver abnormalities detected by laboratory
Imaging studies have been used primarily to localize abnormalities as tests or imaging, is less accurate than laparoscopic liver biopsy.85
a preamble to more definitive invasive tests. Computed tomography
(CT) of the abdomen has been especially effective in this regard.71
THERAPEUTIC TRIALS
Structures appearing abnormal on CT are almost always confirmed as
such on laparotomy or biopsy.72 As a result, abdominal CT and, to a In the past, empirical therapy with anti-inflammatory agents, such as
lesser extent, ultrasound imaging of the gallbladder and hepatobiliary corticosteroids, aspirin, or antimicrobial agents, was often given with
system have been used extensively to evaluate cases of FUO. In one the intent of providing an indirect diagnostic test in patients with
series, more than three CT or ultrasound examinations, or both, were unexplained FUO.1 In rare cases, even antineoplastic drugs were used
performed for each FUO patient evaluated.73 Nevertheless, the diag- for this purpose. Today such trials are seldom indicated. However,
nostic yield per test performed is lowabout 10%. False-negative CT in carefully selected cases, therapeutic trials employing agents with
786 PART II Major Clinical Syndromes
limited spectrums of activity (e.g., antimycobacterial drugs) continue
to be an acceptable means of strengthening a presumptive diagnosis
when all other means have failed.
The limitations and risks of empirical therapeutic trials are obvious.
Underlying diseases may remit spontaneously during the course of
ineffective therapy, giving the false impression of success. Further-
more, empirical treatment is rarely specific. Rifampin, for example, is
likely to be included in empirical therapeutic regimens for tuberculo-
sis, but is highly active against numerous bacterial species other than
Mycobacterium tuberculosis. Conversely, fluoroquinolones given for
other reasons may have a beneficial effect on tuberculosis or Q fever.
Similarly, fevers caused by malignant neoplasms have been reported
to respond better to nonsteroidal anti-inflammatory agents such as
naproxen than fevers of infectious origin,86,87 but the action of naproxen
is nonspecific; the ability of the so-called naproxen test to differentiate
malignant from nonmalignant causes of FUO remains unvalidated.
For these reasons, therapeutic trials, even when successful in reducing
fever, may delay the correct diagnosis and thus the appropriate treat-
ment of FUO. Therefore, empirical therapeutic trials should be
reserved for those very few patients in whom all other approaches have
failed or those so seriously ill that therapy cannot be withheld for a
further period of observation, or both. In practice, this occurs most
often in the case of suspected tuberculosis.
MANAGEMENT
A fundamental principle in the management of classical FUO is that
therapy should be withheld, whenever possible, until the cause of the
fever has been determined, so that treatment can be tailored to a spe-
cific diagnosis.1 This approach is based upon the oft-repeated observa-
tion that nonspecific treatment rarely cures FUO, and has the potential
to delay reaching a specific diagnosis. This ideal is, however, frequently
ignored in clinical practice because the road to diagnosis of FUO is,
by definition, often long and frustrating. As a result, clinicians may feel
compelled to treat symptoms empirically, even though the agents used
may obscure the very signs and symptoms upon which the diagnosis
depends. An important exception is that empirical treatment with
corticosteroids may be appropriate in patients with suspected tempo-
ral arteritis to prevent vascular complications such as blindness or
stroke. Primary care physicians, who have learned from experience
that the most cost-effective approach to many acute febrile illnesses is
to try empirical antimicrobial therapy before undertaking expensive
diagnostic exercises, should appreciate that this approach is less likely
to succeed in patients with FUO.
In febrile neutropenic patients, the principles of treatment are
entirely different. Because of the relatively high prevalence of serious
bacterial infections responsible for these fevers, febrile neutropenic
patients should generally receive broad-spectrum antimicrobial
therapy immediately after samples for appropriate cultures have been
obtained (see Chapter 308).88
PROGNOSIS
The prognosis of FUO is determined by the cause of the fever and by
the nature of any underlying disease or diseases. The time required to
establish the diagnosis is less important. Elderly patients and those
with malignant neoplasms have the poorest prognosis.1 Diagnostic
delay affects the prognosis adversely in intra-abdominal infections,
miliary tuberculosis, disseminated fungal infections, and recurrent
pulmonary emboli.74
Patients in whom FUO remains undiagnosed after extensive evalu-
ation generally have a favorable outcome, characteristically with reso-
lution of their fever in 4 or more weeks without sequelae.3,8,83,89 In a
study of 61 patients followed long term for undiagnosed FUO,
Knockaert and associates89 were largely unsuccessful in identifying
specific etiologies. Most cases eventually resolved spontaneously, gen-
erally obviating the need for corticosteroid therapy. Some patients,
however, required nonsteroidal anti-inflammatory drugs for symp-
tomatic relief. In this series, the 5-year mortality rate for undiagnosed
FUO was only 3.2%.
Selected Causes of Fever of
Unknown Origin
The following conditions are discussed briefly because they represent
some of the more important causes of FUO.
DISSEMINATED GRANULOMATOSES
Tuberculosis, histoplasmosis, and sarcoidosis are all potential causes
of FUO. The diagnosis of these granulomatous diseases can be difficult,
especially when the infection is disseminated but does not show a clas-
sical miliary pattern or other abnormalities on chest radiographs. An
epididymal nodule or an oral mucosal ulcer in patients with FUO may
be the only clues to disseminated tuberculosis and disseminated his-
toplasmosis, respectively, and should be diligently sought on physical
examination.
The disseminated granulomatoses, although potentially fatal, are
among the most readily treatable causes of FUO, therefore meriting
close attention.74 Serial chest radiographs may demonstrate subtle
infiltrates that increase gradually over time. The erythrocyte sedimen-
tation rate is usually raised, and anemia is common. The tuberculin
skin test has been reported to be nonreactive in as many as half of
patients with disseminated tuberculosis, and sputum smears reveal
acid-fast bacilli in only one fourth to one half of cases. Biopsies of the
lung and liver each demonstrate granulomas in 80% to 90% of cases
of miliary tuberculosis. Bone marrow biopsy demonstrates granulo-
mas in only half of the patients, unless anemia, leukopenia, or mono-
cytosis is present, in which case the likelihood of finding granulomas
exceeds 80%. Bronchoalveolar lavage fluid is often culture positive, but
rarely reveals acid-fast bacilli on microscopic examination. Tests based
upon nucleic acid amplification can provide earlier detection of
M. tuberculosis.
LYMPHOMA
Fever is a well-recognized manifestation of some malignant neoplasms,
especially those originating in the hematopoietic system and those with
metastases to the liver.3,14,86,90 A number of mechanisms have been
postulated for such fevers, including necrosis or inflammation in or
around the tumor, generating inflammatory cytokines, and heat gener-
ated by highly metabolically active tumor cells themselves.86 Current
evidence, however, indicates that tumor-associated fever most often
results from the production of pyrogenic cytokines, such as tumor
necrosis factor- and interleukin-1, either by the tumor cells them-
selves or by mononuclear cells that have infiltrated the tumors.91-93
Numerous studies have identified lymphomas as the neoplasms that
most commonly cause FUO in adults.3,13,14,27,94 Periodic (Pel-Ebstein)
fevers have been reported to be characteristic of fever associated with
lymphomas, especially Hodgkins disease.95 However, such periodic
fevers are rarely reported in current reviews of FUO.
THROMBOEMBOLIC DISEASE
Thromboembolic disease has only occasionally been reported as a
cause of FUO. However, a series by AbuRahma and co-workers96 sug-
gests that in contrast to the findings of earlier surveys, lower extremity
venous thrombosis and recurrent occult pulmonary embolism may be
responsible for as many as 6% of todays cases of FUO. Thromboem-
bolic disease is important as a cause of FUO despite its relatively low
incidence, because it is potentially fatal and can be effectively treated.
This diagnosis is most likely to be overlooked as the cause of FUO
when emboli are unaccompanied by pulmonary abnormalities, and
when the attending physician fails to recognize that thrombophlebitis
(even in the absence of focal signs of inflammation) and pulmonary

emboli (even in the absence of radiologically apparent pulmonary
infiltrates) are capable of inducing high fevers of long duration.
ENDOCARDITIS
Microorganisms causing endocarditis are generally easily identified,
because in many cases they circulate continuously in the bloodstream,
and because todays blood culture techniques are highly efficient in
isolating the bacterial species most often responsible for the infection.
For these reasons, endocarditis is less common as a cause of FUO today
than in earlier times. Most of those few endocarditis cases that remain
in the category of FUO belong to the special subgroup comprising
culture-negative endocarditis (see Chapter 77). Rarely, FUO patients
with vegetations detected on a heart valve by echocardiography and
negative blood cultures prove to have nonbacterial thrombotic
endocarditis (marantic endocarditis) associated with an undetected
cancer.
The reported incidence of culture-negative endocarditis has varied
widely, between 2.5% and 31% of cases of infective endocarditis.97 The
high rates of negative blood cultures reported in some previous case
studies most likely reflect the use of less effective culture techniques
and less rigidly defined criteria for the diagnosis of endocarditis than
those employed in more recent series. Today, culture-negative cases
comprise 8% to 15% of the total. The most common cause of negative
blood cultures in infective endocarditis is prior antibiotic therapy; if
such cases are excluded, less than 5% of cases of infective endocarditis
should exhibit negative blood cultures.98
Some species of microorganisms that cause endocarditis can be
difficult to isolate using standard blood-culture techniques.97,99,100
A few important species, such as Bartonella spp., may require
prolonged-incubation for 2 to 3 weeks, or special culture media. If
blood cultures remain sterile after 5 to 7 days in cases in which a
diagnosis of endocarditis seems highly likely, blind subculture from
liquid onto solid media may shorten the time required to isolate the
etiologic microorganism.
Blood cultures are usually sterile in cases of endocarditis caused by
filamentous fungi. Both yeasts and filamentous fungi (especially Asper-
gillus spp.) should be considered as a cause of endocarditis responsible
for FUO in intravenous drug users, patients with prosthetic heart valves,
and patients who have received prolonged intravenous antibiotic
therapy.101 Serologic studies, and nucleic acid amplification tests per-
formed on vegetations or emboli, may be helpful in diagnosing cases of
endocarditis due to Coxiella, Bartonella, Legionella, and Tropheryma
whipplei. These organisms should be considered in patients with FUO
and suspected endocarditis in whom blood cultures are sterile.102
TEMPORAL ARTERITIS AND
POLYMYALGIA RHEUMATICA
Temporal arteritis, also known as giant cell arteritis, and polymyalgia
rheumatica are diseases with protean manifestations.103 No single clini-
cal symptom, sign, or noninvasive laboratory test can be relied on to
establish the diagnosis. More often than not, physical findings indica-
tive of active arteritis are absent, and a history consistent with arteritis
is obtained only after careful questioning by a clinician fully versed in
the varied symptoms and signs of the disorder.103 In geriatric patients,
temporal arteritis may manifest only as unexplained fever or pro-
longed malaise, depression, or anemia.
The erythrocyte sedimentation rate is usually markedly elevated,
usually more than 80 to 100 mm/hour, during active arteritis and is
thus both useful as an aid to diagnosis and a clinical marker of disease
activity. Rarely, the erythrocyte sedimentation rate is normal during
active arteritis. For this reason, symptoms or signs, or both, consistent
with the disease should be pursued even if the sedimentation rate is
normal.
Nodules or diminished temporal artery pulsations need not be
present for an arterial biopsy to reveal active temporal arteritis. Because
pathologic abnormalities may be confined to short segments of the
51 Fever of Unknown Origin 787
artery, extensive (bilateral) segments of the temporal artery may need
to be examined to establish the diagnosis. Blindness, the most feared
complication of temporal arteritis, can generally be averted through
the timely use of corticosteroid therapy.
ADULT STILLS DISEASE (JUVENILE
RHEUMATOID ARTHRITIS)
Stills disease is a diagnosis based solely on clinical findings, because
definitive serologic markers and other diagnostic tests for the disorder
do not exist.104 Its clinical features include high spiking fever, arthral-
gias or arthritis, a transient maculopapular rash, lymphadenopathy,
hepatosplenomegaly, serositis, and sore throat. Leukocytosis is gener-
ally marked, and rheumatoid factor and antinuclear antibody tests are
negative.
The fever of Stills disease is characteristically high and spiking, with
temperatures reaching as high as 41.6 C. The fever pattern may be
either intermittent (quotidian) or remittent. Although most patients
seek medical attention within 2 weeks of the onset of symptoms, some
25% suffer for more than 4 weeks before doing so. A distinctive, eva-
nescent, salmon-pink macular or maculopapular rash is typically
present during the early course of the illness.
Hepatomegaly or abnormal liver function tests, or both, are common
in Stills disease. Severe liver failure, however, is seen almost exclusively
in conjunction with aspirin or other nonsteroidal anti-inflammatory
drug therapy. In a third of patients, the disease has a self-limited
course, in a quarter an intermittent course, and in a third a chronic
course. Rarely, Stills disease has a fatal outcome.104 Three predictors
of an unfavorable outcome include root joint (shoulder and hip)
arthritis on presentation, polyarthritis, and rash. No other clinical or
laboratory manifestations, including human leukocyte antigen (HLA)
tests, predict the outcome.104
DRUG FEVER
Fever may be the sole or most prominent feature of an adverse drug
reaction.105 This disorder is generally not accompanied by other signs
of drug allergy, notwithstanding statements to the contrary in numer-
ous textbooks and review articles. Although occasionally present,
neither rash nor eosinophilia is common. Relative bradycardia is, like-
wise, uncommon. Several authors have suggested that there might be
a characteristic drug-induced fever pattern. However, no such pattern
has yet emerged, perhaps because antipyretics and external cooling
measures are so frequently used to treat drug-induced fever. Consider-
able variability has been reported in the length of time between the
onset of treatment with a drug and the onset of drug-related fever.
Therefore, clinicians must not assume that because a patient has been
taking a certain medication for a long time, it cannot be the cause of
FUO. Contrary to some early reports, there does not appear to be any
clear associations between drug fever and systemic lupus erythemato-
sus, atopy, female sex, or advanced age.105
FACTITIOUS FEVER
Although in most series, factitious fever is a relatively uncommon
cause of FUO, in one report it was diagnosed in 9% of cases of pro-
longed FUO, suggesting that it might be more common than generally
appreciated.106 It has two formsa fraudulent form and a self-induced
form.107 Instances of patients manipulating a mercury thermometer no
longer occur with the advent of electronic and infrared thermometry.
Genuine fever can be induced by the injection or ingestion of pyro-
genic materials. Data falsification has also been used to create fraudu-
lent fever.108
Patients with self-induced fever may also cause real fevers through
the self-administration of pyrogenic substances, most often bacterial
suspensions.
Patients with factitious fever are generally young women, approxi-
mately 50% of whom have had training in some aspect of health care,
788 PART II Major Clinical Syndromes

often as nurses. Physicians have rarely been diagnosed with the disor-
der. In the rare male patient who manifests factitious fever, the fraudu-
lent variety is most frequent. In women, the self-induced variety
predominates. In fact, all 12 of 12 cases of self-induced infection
reported by Reich and Gottfried109 occurred among women. Factitious
fever, especially the fraudulent variety, appears to be increasing some-
what in frequency among older patients.110,111
Obviously, factitious fever must have some psychogenic basis.
However, a survey of 41 cases of factitious disorders by Reich and
Gottfried109 found no evidence of major psychiatric diagnoses among
patients with either self-induced infection or simulated illnesses. Nor
did these investigators find evidence of psychosis on projective testing
among the few patients who underwent detailed psychological
analysis.
PERIODIC FEVERS
Many of the diseases mentioned earlier can cause periodic fevers,
including juvenile rheumatoid arthritis, adult-onset Stills disease,
Crohns disease, bartonellosis, and Behets syndrome. There are also
several rare, hereditary periodic fever syndromes included on the long
list of causes of FUO.112 These include familial Mediterranean fever,
hyper-IgD syndrome, and TRAPS. Although genetic defects respon-
sible for these three types of hereditary periodic fever have been
identified,113 the striking periodicity of the clinical attacks remains
unexplained. They should be suspected in cases of FUO in which the
fever is recurrent (i.e., febrile periods separated by symptom-free inter-
vals of variable duration), persists for more than 2 years, and in which
there is a family history of similar attacks.

REFERENCES
1. Durack DT. Fever of unknown origin. In: Mackowiak PA, ed.
Fever. Basic Mechanisms and Management. 2nd ed. Philadelphia:
Lippincott-Raven; 1997:237-249.
2. Keefer CS, Leard SE. Prolonged and Perplexing Fevers. Boston:
Little, Brown; 1955:1-248.
3. Petersdorf RG, Beeson PB. Fever of unexplained origin: Report
on 100 cases. Medicine. 1961;40:1-30.
4. Alt HL, Barker MH. Fever of unknown origin. JAMA.
1930;94:1457-1461.
5. Wong SY, Lam M. Pyrexia of unknown originapproach to
management. Singapore Med J. 1995;36:204-208.
6. Smith JW. Fever of undetermined origin: Not what it used to be.
Am J Med Sci. 1986;292:56-64.
7. Durack DT, Street AC. Fever of unknown origin reexamined and
redefined. Curr Clin Top Infect Dis. 1991;11:35-51.
8. Iikuni Y, Okada J, Kondo H, et al. Current fever of unknown
origin 1982-1992. Intern Med. 1994;33:67-73.
9. Knockaert DC, Vanneste LJ, Bobbaers HJ. Fever of unknown
origin in elderly patients. J Am Geriatric Soc. 1993;41:
1187-1192.
10. Chantada G, Casak S, Daza Plata J, et al. Children with fever of
unknown origin in Argentina: An analysis of 113 cases. Pediatr
Infect Dis J. 1994;13:260-263.
11. Knockaert DC, Vanneste LJ, Laurent J, et al. Recurrent or epi-
sodic fever of unknown origin: Review of 45 cases and survey of
the literature. Medicine. 1993;72:184-196.
12. Knockaert DC. Diagnostic strategy for fever of unknown origin
in the ultrasonography and computed tomography era. Acta
Clin Belg. 1992;47:100-116.
13. Kazanjian PH. Fever of unknown origin: Review of 86 patients
treated in community hospitals. Clin Infect Dis. 1992;15:
968-973.
14. Larson EB, Featherstone HJ, Petersdorf RG. Fever of undeter-
mined origin: Diagnosis and follow-up of 105 cases, 1970-1980.
Medicine. 1982;61:269-292.
15. Sharma BK, Kumari S, Varma SC, et al. Prolonged undiagnosed
fever in Northern India. Trop Geogr Med. 1992;44:32-36.
16. De Kleijn EMHA, Vandenbroucke JP, van der Meer JWM, et al.
Fever of unknown origin (FUO). I. A prospective multicenter
study of 167 patients with FUO, using fixed epidemiologic entry
criteria. Medicine. 1997;76:392-400.
17. Barbado FJ, Vazquez JJ, Pena JM, et al. Fever of unknown origin:
A survey on 133 patients. J Med. 1984;15:185-192.
18. Benito N, Nunez A, de Gorgolas M, et al. Bone marrow biopsy
in the diagnosis of fever of unknown origin in patients with
acquired immunodeficiency syndrome. Arch Intern Med.
1997;157:1577-1580.
19. Norris AH, Krasinskas AM, Salhany KE, et al. Kikuchi-Fujimoto
disease: A benign cause of fever and lymphadenopathy. Am J
Med. 1996;101:401-405.
20. McDermott EM, Smillie DM, Powell RJ. Clinical spectrum of
familial Hibernian fever: A 14-year follow-up study of the index
case and extended family. Mayo Clin Proc. 1997;72:806-817.
21. Klastersky J, Weerts D, Hensgens C, et al. Fever of unexplained
origin in patients with cancer. Eur J Cancer. 1973;9:649-656.
22. Hamman L, Wainwright CW. The diagnosis of obscure fever.
II. The diagnosis of unexplained high fever. Johns Hopkins Hosp
Bull. 1936;58:307-331.
23. Berger L, Sinkoff MW. Systemic manifestations of hyperne-
phroma: A review of 273 cases. Am J Med. 1957;22:791-796.
24. Shoji S, Imamura A, Imai Y, et al. Fever of unknown origin: A
review of 80 patients from the Shinetsu area of Japan from
1986-1992. Intern Med. 1994;33:74-76.
25. Saxe SE, Gardner P. The returning traveler with fever. Infect Dis
Clin North Am. 1992;6:427-439.
26. Jacobs RF, Schutze GE: Bartonella henselae as a cause of pro-
longed fever and fever of unknown origin in children. Clin Infect
Dis. 1998;26:80-84.
27. Gleckman RA, Esposito AL. Fever of unknown origin in the
elderly: Diagnosis and treatment. Geriatrics. 1986;41:45-52.
28. Kauffman CA, Jones PG. Diagnosing fever of unknown origin
in older patients. Geriatrics. 1984;39:46-51.
29. Zenone T. Fever of unknown origin in adults: evaluation of 144
cases in a non-university hospital. Scandinavian J Infect Dis.
2006;38:632-638.
30. Ghose MK, Shensa S, Lerner PI. Arteritis of the aged (giant cell
arteritis) and fever of unexplained origin. Am J Med.
1976;60:429-436.
31. Tal S, Guller V, Gurevich A, et al. Fever of unknown origin in the
elderly. J Intern Med. 2002;252:295-304.
32. Tal S, Guller V, Gurevich A. Fever of unknown origin in older
adults. Clinics Geriatr Med. 2007;23:649-668.
33. Speil C, Mushtaq A, Adamski A, et al. Fever of unknown origin
in the returning traveler. Infect Dis Clinics North Am.
2007;21:1091-1113.
34. Konecny P, Davidson RN. Pyrexia of unknown origin in the
1990s: Time to redefine. Br J Hosp Med. 1996;56:21-24.
35. Neuhoff FJ, Brewer JE, Castaneda T, et al. Frequency and yield
of postoperative fever evaluation. Infect Dis Obstet Gynecol.
1998;6:252-255.
36. Anand VT, Phillipps JJ, Allen D, et al. A study of postoperative
fever following paediatric tonsillectomy. Clin Otolaryngol Allied
Sci. 1999;24;360-364.
37. Shaw JA, Chung R. Febrile response after knee and hip arthro-
plasty. Clin Orthop Relat Res. 1999;367:181-189.
38. Circiumaru B, Baldock G, Cohen J. A prospective study of fever
in the intensive care unit. Intensive Care Med. 1999;25:
668-673.
39. Hall J. Assessment of fever in the intensive care unit. Is the
answer just beyond the tip of our nose? Am J Respir Crit Care
Med. 1999;159:693-694.
40. Georgilis K, Plomaritoglou A, Dafni U, et al. Aetiology of fever
in patients with acute stroke. J Intern Med. 1999;246:203-209.
41. Engelhart S, Glasmacher A, Exner M, et al. Surveillance for noso-
comial infections and fever of unknown origin among adult
hematology-oncology patients. Infect Control Hosp Epidemiol.
2002;23:244-248.
42. Corey L, Boeckh M. Persistent fever in patients with neutrope-
nia. N Engl J Med. 2002;346:222-224.
43. Chang FY, Singh N, Gayowski T, et al. Fever in liver transplant
recipients: Changing spectrum of etiologic agents. Clin Infect
Dis. 1998;26:59-65.
44. Regnery RL, Anderson BE, Clarridge JE III, et al. Characteriza-
tion of a novel Rochalimaea species, R. henselae sp nov, isolated
from blood of a febrile, human immunodeficiency virus positive
patient. J Clin Microbiol. 1992;30:265-274.
45. Miralles P, Moreno S, Perez-Tascon M, et al. Fever of uncertain
origin in patients infected with the human immunodeficiency
virus. Clin Infect Dis. 1995;20:872-875.
46. Bissuel F, Leport C, Perrone C, et al. Fever of unknown origin
in HIV-infected patients: A critical analysis of a retrospective
series of 57 cases. J Intern Med. 1994;236:529-535.
47. Wheeler DA, Call S, Wasserman SS, et al. Relationship between
basal body temperature and stage of disease in asymptomatic
HIV-infected men. Infect Dis Clin Pract. 1997;6:47-50.
48. Sepkowitz KA. Effect of prophylaxis on the clinical manifesta-
tions of AIDS-related opportunistic infections. Clin Infect Dis.
1998;26:806-810.
49. Lozano F, Torre-Cisneros J, Santos J, et al. Impact of highly active
antiretroviral therapy on fever of unknown origin in
HIV-infected patients. Eur J Clin Microbiol Infect Dis. 2002;21:
132-139.
50. Barbado FJ, Gomez-Cerezo J, Pena JM, et al. Fever of unknown
origin: Classic and associated with human immunodeficiency
virus infection. A comparable study. J Med. 2001;32:152-162.
51. Armstrong WS, Katz JT, Kazanjian PH. Human immunodefi-
ciency virus-associated fever of unknown origin: A study of 70
patients in the United States and review. Clin Infect Dis.
1999;28:341-345.
52. Hot A, Schmulewitz L, Viard JP, et al. Fever of unknown origin
in HIV/AIDS patients. Infect Dis Clin North Am. 2007;21:
1013-1032.
53. Knockaert DC. Recurrent fevers of unknown origin. Infect Dis
Clin North Am. 2007;21:1189-1211.
54. Aduan RP, Fauci AS, Dale DC, et al. Prolonged fever of unknown
origin (FUO): A prospective study of 347 patients. Clin Res.
1978;26:558A.
55. Mackowiak PA. Commentary: Fever patterns. Infect Dis Clin
Pract. 1997;6:308-309.
56. Woodward TE. The fever pattern as diagnostic aid. In: Mack-
owiak PA, ed. Fever: Basic Mechanisms and Management. 2nd
ed. Philadelphia: Lippincott-Raven; 1997;215-235.
57. McCarthy PL, Dolan TF. Hyperpyrexia in children. Am J Dis
Child. 1976;130:849-851.
58. McCarthy PL, Jekel JF, Dolan TF. Temperature greater than or
equal to 40 C in children less than 24 months of age: A prospec-
tive study. Pediatrics. 1977;59:663-668.
59. Pizzo PA, Lovejoy FH, Smith DH. Prolonged fever in children:
Review of 100 cases. Pediatrics. 1975;55:468-473.
60. Lohr JA, Hendley JO. Prolonged fever of unknown origin: A
record of experiences with 54 childhood patients. Clin Pediatr.
1977;16:768-773.
61. Sheon RP, Van Ommeu RA. Fever of obscure origin: Diagnosis
and treatment based on a series of sixty cases. Am J Med.
1963;34:486-499.
62. Vickery DM, Quinnell RK. Fever of unknown origin. An algo-
rithmic approach. JAMA. 1977;238:2183-2188.
63. Esposito AL, Gleckman RA. A diagnostic approach to the
adult with fever of unknown origin. Arch Intern Med.
1979;139:575-579.
64. Hurley DL. Fever in adults. What to do when the cause is not
obvious? Postgrad Med. 1983;74:232-244.
65. Gries E, Hoensch H, Ohnhaus EE. Differential diagnosis in fever
of unknown origin: significance of concomitant clinical symp-
toms [in German]. Klin Wochenschr. 1986;64:307-313.
66. Nolan SM, Fitzgerald FT. Fever of unknown origin: The general
internists approach. Postgrad Med. 1987;81:190-205.
67. Koch H, Jenke V. Fever of unknown origina basic diagnostic
program [in German]. Z Gesamte Inn Med. 1990;45:560-562.
68. Fischer JL, Henselmann L. Diagnostic steps in fever. Fever of
unknown origin [in German]. Med Klin. 1987;82:229-235.
69. Palazzi DL, McClain KL, Kaplan SL. Hemophagocytic syndrome
in children: An important diagnostic consideration in fever of
unknown origin. Clin Infect Dis. 2003;36:306-312.
70. Kuse ER, Langefeld I, Jaeger K, et al. Procalcitonin in fever
of unknown origin after liver transplantation: A variable to
differentiate acute rejection from infection. Crit Care Med.
2000;28:555-559.
71. Rowland MD, Del Bene VE. Use of body computed tomography
to evaluate fever of unknown origin. J Infect Dis. 1987;156:
408-409.
72. Quinn MJ, Sheedy PF II, Stephen DH, et al. Computed tomog-
raphy of the abdomen in evaluation of patients with fever of
unknown origin. Radiology. 1980;136:407-411.
73. Knockaert DC, Vanneste LJ, Vanneste SB, et al. Fever of unknown
origin in the 1980s: An update of the diagnostic spectrum. Arch
Intern Med. 1992;152:51-55.
74. Arnow PM, Flaherty JP. Fever of unknown origin. Lancet.
1997;350:575-580.
75. Knockaert DC, Mortelmans LA, De Roo MC, et al. Clinical value
of gallium-67 scintigraphy in evaluation of fever of unknown
origin. Clin Infect Dis. 1994;18:601-605.
76. Syrjala MT, Valtonen V, Liewendahl K, et al. Diagnostic signifi-
cance of indium-111 granulocyte scintigraphy in febrile
patients. J Nucl Med. 1987;28:155-160.
77. Lavender JP, Lowe J, Barker JR. Ga-67 citrate scanning in neo-
plastic and inflammatory lesions. Br J Radiol. 1971;44:361-366.
78. Hilson AJW, Maisey MN. Gallium-67 scanning in pyrexia of
unknown origin. BMJ. 1979;II:1330-1331.
79. Sfakianakis GN, Al-Sheikh W, Heal A, et al. Comparison of scin-
tigraphy with 111In leukocytes and 67Ga in the diagnosis of occult
sepsis. J Nucl Med. 1982;23:618-626.
80. Meller J, Altenvoerde G, Munsel U, et al. Fever of unknown
origin: Prospective comparison of [18F] FDG imaging with a
double-bead coincidence camera and gallium-67 citrate SPECT.
Eur J Nucl Med. 2000;7:1617-1625.

81. Blockmans D, Knockaert D, Maes A, et al. Clinical value of [18F]
fluoro-deoxyglucose positron emission tomography for patients
with fever of unknown origin. Clin Infect Dis. 2001;32:
191-196.
82. Lorenzen J, Buchert R, Bohuslavizki KH. Value of FDG PET in
patients with fever of unknown origin. Nucl Med Comm.
2001;22:779-783.
83. Larson EB, Featherstone HJ, Petersdorf RG. Fever of undeter-
mined origin: Diagnosis and follow-up of 105 cases, 1970-1980.
Medicine. 1982;61:269-292.
84. Solis-Herruzo JA, Benita V, Morillas JD. Laparoscopy in fever of
unknown originStudy of seventy cases. Endoscopy.
1981;13:207-210.
85. Mitchell DP, Hanes TE, Hoyumpa AM Jr, et al. Fever of unknown
originAssessment of the value of percutaneous liver biopsy.
Arch Intern Med. 1977;137:1001-1004.
86. Chang JC, Hawley BH. Neutropenic fever of undetermined
origin (N-FUO): Why not use the naproxen test? Cancer Invest.
1995;13:448-450.
87. Rolston KVI. Neoplastic fever: All who shiver are not infected.
Support Care Cancer. 2005;13:870-877.
88. Hughes WT, Armstrong D, Bodey GP, et al. Guidelines for the
use of antimicrobial agents in neutropenic patients with unex-
plained fever. J Infect Dis. 1990;161:381-396.
89. Knockaert DC, Dujardin KS, Bobbaers HJ. Long-term follow-up
of patients with undiagnosed fever of unknown origin. Arch
Intern Med. 1996;156:618-620.
90. Bodel P. Tumors and fever. Ann N Y Acad Sci. 1974;230:6-13.
91. Bodel P, Ralph P, Wenc K, et al. Endogenous pyrogen production
by Hodgkins disease and human histiocytic lymphoma cell lines
in vitro. J Clin Invest. 1980;65:514-518.
92. Dinarello CA. Interleukin-1. Rev Infect Dis. 1984;6:51-95.
93. Lachman LB, Moore JO, Metzger RS. Preparation and charac-
terization of LAF from acute monocytic and myelomonocytic
leukemic cells. Cell Immunol. 1978;41:189-206.
94. Esposito AL, Gleckman R. Fever of unknown origin in the
elderly. J Am Geriatr Soc. 1978;26:498-505.
95. Reimann HA. Periodic (Pel-Ebstein) fever of lymphomas. Ann
Clin Lab Sci. 1977;7:1-5.
96. AbuRahma AF, Saiedy S, Robinson PA, et al. Role of venous
duplex imaging of the lower extremities in patients with fever
of unknown origin. Surgery. 1997;121:366-371.
97. Cannady PB Jr, Sanford JP. Negative blood cultures in infective
endocarditis: A review. South Med J. 1976;69:1420-1424.
98. Von Reyn CF, Levy BS, Arbeit RD, et al. Infective endocarditis:
An analysis based on strict case definitions. Ann Intern Med.
1981;94:505-518.
99. Van Scoy RE. Culture-negative endocarditis. Mayo Clin Proc.
1982;57:149-154.
100. Washington JA II. The role of the microbiology laboratory in
the diagnosis and antimicrobial treatment of infective endocar-
ditis. Mayo Clin Proc. 1982;57:22-32.
101. Rubinstein E, Noriega ER, Simberkoff MS, et al. Fungal endo-
carditis: Analysis of 24 cases and review of the literature. Medi-
cine (Baltimore). 1975;54:331-334.
102. Tompkins LS, Roessler BJ, Redd SC, et al. Legionella prosthetic-
value endocarditis. N Engl J Med. 1988;318:530-535.
103. Goodman BW. Temporal arteritis. Am J Med. 1979;67:
839-852.
104. Pouchot J, Sampalis JS, Beaudet F, et al. Adult Stills disease:
Manifestations, disease course, and outcome in 62 patients.
Medicine. 1991;70:118-136.
51 Fever of Unknown Origin 789
105. Mackowiak PA, LeMaistre CF. Drug fever: A critical appraisal of
conventional concepts. Ann Intern Med. 1987;106:
728-733.
106. Aduan RP, Fauci AS, Dale DC, et al. Factitious fever and self-
induced infection, a report of 32 cases and review of the litera-
ture. Ann Intern Med. 1979;90:230-242.
107. Sarwari AR, Mackowiak PA. Factitious fever. Curr Clin Topics
Infect Dis. 1997;17:88-94.
108. Petersdorf RG, Bennet IL. Factitious fever. Ann Intern Med.
1957;46:1039-1062.
109. Reich D, Gottfried LA. Factitious disorders in a teaching hospi-
tal. Ann Intern Med. 1983;99:240-247.
110. Potin M, Regamey C, Glauser MP. Factitious fever as a cause of
prolonged fevers. Apropos of 5 clinical cases [in French].
Schweiz Med Wochenschr. 1983;113:1534-1539.
111. Marcus EL, Van Dijk JM. Fever of unknown origin [letter]. J Am
Geriatr Soc. 1991;36:637-638.
112. Drenth JPH, van der Meer JWM. Hereditary periodic fever. N
Engl J Med. 2001;345:1748-1757.
113. Simon A, Mariman EC, van der Meer JWM, et al. A founder
effect in the hyperimmunoglobulinemia D and periodic fever
syndrome. Am J Med. 2003;114:148-152.
114. MacLean JD, Lalonde RG, Ward B. Fever from the tropics. Travel
Med Advisor. 1994;5:27.
115. Doherty JF, Grant AD, Bryceson ADM. Fever as the presenting
complaint of travellers returning from the tropics. QJM.
1995;88:277-281.