M ost episodes of fever in humans are short-lived and do not require almost all fall within one of five categories: infections, neoplasms, con-
diagnostic investigation or specific therapy. Some are manifestations nective tissue diseases, miscellaneous other disorders, and undiag-
of more serious illnesses, most of which can be readily diagnosed and nosed illnesses. The relative frequencies of individual diagnoses within
effectively treated. However, a small but important subgroup of fevers these five categories vary depending on the geographic region, ages of
are both persistent and difficult to diagnose. Such puzzling fevers have the patients, type of medical practice, and other factors (Fig. 51-1).
fascinated and frustrated clinicians since the earliest days of clinical In most series, infections have comprised the largest category,
thermometry,1 resulting in a welter of clinical publications. The two accounting for 25% to 50% of cases. However, in patients older than 65
most important of these, from a historical perspective, are the classical years, infections become less common, falling into second or third place
treatises, Prolonged and Perplexing Fevers, published by Keefer and as a cause of classical FUO.8,9 In the series of Knockaert and associates,9
Leard in 1955,2 and Fever of Unknown Origin: Report on 100 Cases, by infection was the cause of FUO in only 25% of cases 65 years of age or
Petersdorf and Beeson in 1961.3 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
Terminology and Definitions surveys involving younger adults, in which as many as 30% of cases
In the United States, the term fever of unknown origin (FUO) is gener- remain undiagnosed.16 The longer the duration of fever before medical
ally used.4 In other countries an alternative term, pyrexia of unknown consultation, the less likely that a definitive diagnosis will be made.17
origin (PUO), is often used.5 Among the infections responsible for classical FUO, abscesses,
The first formal definition of FUO to gain broad acceptance was endocarditis, tuberculosis, and complicated urinary tract infections
proposed by Petersdorf and Beeson nearly five decades ago: fever have consistently been among the most important. These tend to vary
higher than 38.3 C (101 F) on several occasions, persisting without in incidence according to locale. Visceral leishmaniasis, for example,
diagnosis for at least 3 weeks in spite of at least 1 weeks investigation although absent from most series of classical FUO, accounted for 8%
in hospital.3 Later investigators have modified and extended this clas- of cases in a study reported in 1997 from Spain.18 Other examples of
sical definition to reflect evolutionary changes in clinical practice.6,7 causes of classical FUO in distinct populations include familial Medi-
These changes include the conduct of most diagnostic tests in the terranean fever in Ashkenazi Jews; Kikuchis disease, an unusual form
outpatient setting rather than in the hospital, the increasing number of necrotizing lymphadenitis seen primarily in Japan19; and TRAPS
of immunocompromised patients (especially those with neutropenia), (TNF-receptor associated periodic fever), formerly called familial
the proliferation of increasingly complex surgical and intensive care Hibernian fever, an inherited periodic fever syndrome described origi-
treatment protocols, and the advent of human immunodeficiency nally in Ireland.20 The miscellaneous category contains both varied and
virus (HIV) infection leading to the acquired immunodeficiency syn- individually rare causes of classical FUO (Table 51-2).
drome (AIDS). In response to this evolving environment, cases of FUO Of the connective tissue diseases responsible for classical FUO, Stills
are currently codified into four distinct subclasses of the disorder: disease (juvenile rheumatoid arthritis), other variants of rheumatoid
classical FUO, health careassociated FUO, immune-deficient FUO, arthritis, and systemic lupus erythematosus predominate in younger
and HIV-related FUO (Table 51-1).Computed axial tomography, patients, whereas temporal arteritis and polymyalgia rheumatica syn-
magnetic resonance imaging, ultrasound imaging, nucleic acidbased dromes are more common in elderly patients.
diagnostic testing, and rapid tests for pathogens have changed the Malignant neoplasms, another important cause of FUO, can induce
landscape of FUO. One could argue that the definitions of types of fever directly through the production and release of pyrogenic cyto-
FUO have much overlap and are not as clinically useful as they were kines, as in the case of certain lymphomas. They can also generate
in the past. 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
Classical Fever of Unknown Origin as a notable cause of FUO,3,22 in reality this tumor only rarely occurs
Classical FUO refers to the type of FUO defined by Petersdorf and with fever.23
Beeson in 1961.3 The only alteration to their definition required to The relative frequency with which the major diagnostic categories
conform to modern medical practice is to incorporate investigation in are represented in series of classical FUO varies according to both the
the outpatient setting, which today has become the preferred venue era in which the series was published6,13 and its country of origin (see
for evaluation and treatment. Most patients with classical FUO have Fig. 51-1).8,15,24,25 Since the mid-1900s, the frequency with which infec-
subacute or chronic symptoms and therefore can be safely investigated tions and malignant neoplasms have been identified as causes of clas-
as outpatients. In a series of 53 such patients, for example, the median sical FUO has fallen steadily, whereas the proportion of miscellaneous
duration of fever before diagnosis was 40 days.1 Even though the causes and undiagnosed conditions has risen.24 However, in develop-
tempo and accuracy of clinical investigation have improved in recent ing countries, the frequency with which infections are diagnosed has
years, many cases of classical FUO continue to elude diagnosis for changed little.15 Consequently, in these countries malignant neoplasms
extended periods, as reflected in the findings of a Japanese survey and connective tissue disorders are comparatively less important as
conducted from 1982 to 1992, in which an average of 49 days (range causes of classical FUO than in developed countries (see Fig. 51-1).15
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
INFANTS AND CHILDREN
have dealt with classical FUO rather than with the three other more
recently defined subclasses listed earlier.7 Over the years a recurrent The diseases responsible for classical FUO in infants differ from those
theme has become clear: of the myriad disorders causing classical FUO, 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)
TABLE
51-2 Examples of Rare Miscellaneous Causes of Fever
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 TABLE Possible Causes of Fever in Neutropenic Patients
fever, however, is associated with a worse prognosis. Health care 51-5 Not Responding to Broad-Spectrum Antibiotics
associated sinusitis, often a complication of mechanical ventilation Approximate Frequency
arising from supine positioning and the use of endotracheal, gastric, in High-Risk Patients
and feeding tubes is common39 and should always be considered when Causes (%)
evaluating FUO in intensive care unit patients. Fungal infections susceptible to empirical therapy 40
Fungal infections resistant to empirical antifungal 5
therapy
STROKE PATIENTS
Bacterial infections (with cryptic foci, biofilms, and 10
In patients with a recent stroke, fever is usually the result of an infec- resistant organisms)
tion, most commonly a urinary tract infection related to urinary cath- Toxoplasma gondii, mycobacteria, or fastidious 5
pathogens (legionella, mycoplasma,
eterization. However, in some cases, a focus of infection cannot be Chlamydophila pneumoniae, bartonella)
identified, and when the fever does not respond to empirical antibiotic Viral infections (herpesviruses, cytomegalovirus, 5
treatment it is presumed to be due to the stroke itself. In a study of Epstein-Barr virus, human herpesvirus 6,
330 patients hospitalized for acute stroke, Georgilis and associates40 varicella-zoster virus, herpes simplex virus,
observed that noninfective fevers were most often associated with parainfluenza virus, respiratory syncytial virus,
influenza viruses)
intracranial mass effects and tended to occur earlier after the onset of
Graft-versus-host disease after hematopoietic 10
stroke than fevers due to infection. stem cell transplantation
Undefined (e.g., drug fever, toxic effects of 25
chemotherapy, antitumor responses, undefined
Immune-Deficient Fever of pathogens)
Unknown Origin 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.
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 Human Immunodeficiency
of 116 hematology-oncology patients, for example, Engelhart and
associates41 observed 33 FUOs in 28 patients, for an overall rate of 8.2
VirusRelated Fever of Unknown Origin
episodes per 1000 patient days. Because of impaired immune responses, Episodes of fever are commonplace in patients infected with HIVa
signs of inflammation other than fever are notoriously absent or special subgroup of immunodeficient patients, now so numerous that
diminished in such patients, leading to atypical clinical manifestations a separate category of FUO is justified.44-46 The primary phase of HIV
and absence of radiologic abnormalities in what otherwise would be infection is characterized by a mononucleosis-like illness in which
readily diagnosed infections.42 fever is a prominent feature. All too often, primary HIV infection
In patients with impaired cell-mediated immunity, FUO is often eludes diagnosis because the illness is nonspecific and precedes sero-
due to conditions other than pyogenic bacterial infections, as illus- conversion. For this reason it represents an important cause of HIV-
trated in a recent survey of transplant recipients by Chang and co- associated FUO. Once symptoms of the primary phase of the HIV
workers.43 In that series, infections were the cause of 7 of the 12 infection resolve, HIV-infected patients enter a long period of sub-
(58%) episodes of fever in which a diagnosis could not be established clinical infection during which they are usually afebrile.47 However, in
during 3 days of intensive investigation; these included five cases of the later phases of untreated HIV infection, episodes of fever become
human herpesvirus 6 infection, one of varicella-zoster virus infection common, often signifying a superimposed illness. Many of these
with atypical skin lesions, and one of pneumonia due to Serratia are potentially devastating opportunistic infections, which tend to
marcescens (in which the pulmonary infiltrate was initially inapparent present in atypical fashion owing to the tendency of the disordered
radiographically). Three (25%) cases were due to noninfectious con- immune response or prior therapy, or both,48 to distort their clinical
ditions: malignant neoplasm, drug-induced fever, and adrenal insuf- manifestations.
ficiency. In two (17%) cases, the cause of the fever could not be Once highly active antiretroviral therapy (HAART) has been started
determined. 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
NEUTROPENIC FEVER OF UNKNOWN ORIGIN
receiving therapy.49 Mycobacterial infections have been the common-
Neutropenia is a dangerous condition that can be considered a special est cause of FUO in such patients; collagen-vascular diseases have been
subclass of immunodeficiency. The number of patients with episodes distinctly uncommon.50 In a recent series reported by Armstrong and
of neutropenia resulting from cytotoxic therapy or from various colleagues,51 the etiology of the FUO was identified in 56 of 70 (80%)
malignancies affecting the bone marrow is rising, even as the median patients. A single etiology was identified in 43; 3 distinct causes of FUO
duration of neutropenia in such patients is falling, owing to were identified in 3 cases, and 2 causes in 10 others. The 70 causes of
the increasing application of treatment with colony-stimulating FUO in 56 patients in whom the etiology was determined are listed in
factors. Table 51-6. Most cases of FUO in HIV-infected patients are the result
Episodes of fever are common in patients with neutropenia. Many of opportunistic infections, the specific frequencies of which are dic-
such episodes are short-lived, because they either respond quickly to tated, at least in part, by geographic variation in the prevalence of these
treatment or are manifestations of rapidly fatal infections. Because infections.52
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. Clinical Evaluation of Fever of
However, only about 35% of prolonged episodes of febrile neutropenia
respond to broad-spectrum antibiotic therapy. Although practitioners
Unknown Origin
often assume that if fever does not respond promptly to antibacterial The evaluation of a patient with FUO typically includes a comprehen-
therapy, fungal infection must be responsible, other causes are equally sive history, verification that the patient actually has fever, consider-
likely to be identified (Table 51-5).42 ation of the fever pattern, repeated physical examinations, a host of
51 Fever of Unknown Origin 783
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
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.)
limited spectrums of activity (e.g., antimycobacterial drugs) continue tomatic relief. In this series, the 5-year mortality rate for undiagnosed
to be an acceptable means of strengthening a presumptive diagnosis FUO was only 3.2%.
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 Selected Causes of Fever of
ineffective therapy, giving the false impression of success. Further-
more, empirical treatment is rarely specific. Rifampin, for example, is
Unknown Origin
likely to be included in empirical therapeutic regimens for tuberculo- The following conditions are discussed briefly because they represent
sis, but is highly active against numerous bacterial species other than some of the more important causes of FUO.
Mycobacterium tuberculosis. Conversely, fluoroquinolones given for
other reasons may have a beneficial effect on tuberculosis or Q fever.
DISSEMINATED GRANULOMATOSES
Similarly, fevers caused by malignant neoplasms have been reported
to respond better to nonsteroidal anti-inflammatory agents such as Tuberculosis, histoplasmosis, and sarcoidosis are all potential causes
naproxen than fevers of infectious origin,86,87 but the action of naproxen of FUO. The diagnosis of these granulomatous diseases can be difficult,
is nonspecific; the ability of the so-called naproxen test to differentiate especially when the infection is disseminated but does not show a clas-
malignant from nonmalignant causes of FUO remains unvalidated. sical miliary pattern or other abnormalities on chest radiographs. An
For these reasons, therapeutic trials, even when successful in reducing epididymal nodule or an oral mucosal ulcer in patients with FUO may
fever, may delay the correct diagnosis and thus the appropriate treat- be the only clues to disseminated tuberculosis and disseminated his-
ment of FUO. Therefore, empirical therapeutic trials should be toplasmosis, respectively, and should be diligently sought on physical
reserved for those very few patients in whom all other approaches have examination.
failed or those so seriously ill that therapy cannot be withheld for a The disseminated granulomatoses, although potentially fatal, are
further period of observation, or both. In practice, this occurs most among the most readily treatable causes of FUO, therefore meriting
often in the case of suspected tuberculosis. 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
MANAGEMENT
skin test has been reported to be nonreactive in as many as half of
A fundamental principle in the management of classical FUO is that patients with disseminated tuberculosis, and sputum smears reveal
therapy should be withheld, whenever possible, until the cause of the acid-fast bacilli in only one fourth to one half of cases. Biopsies of the
fever has been determined, so that treatment can be tailored to a spe- lung and liver each demonstrate granulomas in 80% to 90% of cases
cific diagnosis.1 This approach is based upon the oft-repeated observa- of miliary tuberculosis. Bone marrow biopsy demonstrates granulo-
tion that nonspecific treatment rarely cures FUO, and has the potential mas in only half of the patients, unless anemia, leukopenia, or mono-
to delay reaching a specific diagnosis. This ideal is, however, frequently cytosis is present, in which case the likelihood of finding granulomas
ignored in clinical practice because the road to diagnosis of FUO is, exceeds 80%. Bronchoalveolar lavage fluid is often culture positive, but
by definition, often long and frustrating. As a result, clinicians may feel rarely reveals acid-fast bacilli on microscopic examination. Tests based
compelled to treat symptoms empirically, even though the agents used upon nucleic acid amplification can provide earlier detection of
may obscure the very signs and symptoms upon which the diagnosis M. tuberculosis.
depends. An important exception is that empirical treatment with
corticosteroids may be appropriate in patients with suspected tempo-
LYMPHOMA
ral arteritis to prevent vascular complications such as blindness or
stroke. Primary care physicians, who have learned from experience Fever is a well-recognized manifestation of some malignant neoplasms,
that the most cost-effective approach to many acute febrile illnesses is especially those originating in the hematopoietic system and those with
to try empirical antimicrobial therapy before undertaking expensive metastases to the liver.3,14,86,90 A number of mechanisms have been
diagnostic exercises, should appreciate that this approach is less likely postulated for such fevers, including necrosis or inflammation in or
to succeed in patients with FUO. around the tumor, generating inflammatory cytokines, and heat gener-
In febrile neutropenic patients, the principles of treatment are ated by highly metabolically active tumor cells themselves.86 Current
entirely different. Because of the relatively high prevalence of serious evidence, however, indicates that tumor-associated fever most often
bacterial infections responsible for these fevers, febrile neutropenic results from the production of pyrogenic cytokines, such as tumor
patients should generally receive broad-spectrum antimicrobial necrosis factor- and interleukin-1, either by the tumor cells them-
therapy immediately after samples for appropriate cultures have been selves or by mononuclear cells that have infiltrated the tumors.91-93
obtained (see Chapter 308).88 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
PROGNOSIS
lymphomas, especially Hodgkins disease.95 However, such periodic
The prognosis of FUO is determined by the cause of the fever and by fevers are rarely reported in current reviews of FUO.
the nature of any underlying disease or diseases. The time required to
establish the diagnosis is less important. Elderly patients and those
THROMBOEMBOLIC DISEASE
with malignant neoplasms have the poorest prognosis.1 Diagnostic
delay affects the prognosis adversely in intra-abdominal infections, Thromboembolic disease has only occasionally been reported as a
miliary tuberculosis, disseminated fungal infections, and recurrent cause of FUO. However, a series by AbuRahma and co-workers96 sug-
pulmonary emboli.74 gests that in contrast to the findings of earlier surveys, lower extremity
Patients in whom FUO remains undiagnosed after extensive evalu- venous thrombosis and recurrent occult pulmonary embolism may be
ation generally have a favorable outcome, characteristically with reso- responsible for as many as 6% of todays cases of FUO. Thromboem-
lution of their fever in 4 or more weeks without sequelae.3,8,83,89 In a bolic disease is important as a cause of FUO despite its relatively low
study of 61 patients followed long term for undiagnosed FUO, incidence, because it is potentially fatal and can be effectively treated.
Knockaert and associates89 were largely unsuccessful in identifying This diagnosis is most likely to be overlooked as the cause of FUO
specific etiologies. Most cases eventually resolved spontaneously, gen- when emboli are unaccompanied by pulmonary abnormalities, and
erally obviating the need for corticosteroid therapy. Some patients, when the attending physician fails to recognize that thrombophlebitis
however, required nonsteroidal anti-inflammatory drugs for symp- (even in the absence of focal signs of inflammation) and pulmonary
51 Fever of Unknown Origin 787
emboli (even in the absence of radiologically apparent pulmonary artery, extensive (bilateral) segments of the temporal artery may need
infiltrates) are capable of inducing high fevers of long duration. 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.
ENDOCARDITIS
Microorganisms causing endocarditis are generally easily identified,
ADULT STILLS DISEASE (JUVENILE
because in many cases they circulate continuously in the bloodstream,
RHEUMATOID ARTHRITIS)
and because todays blood culture techniques are highly efficient in
isolating the bacterial species most often responsible for the infection. Stills disease is a diagnosis based solely on clinical findings, because
For these reasons, endocarditis is less common as a cause of FUO today definitive serologic markers and other diagnostic tests for the disorder
than in earlier times. Most of those few endocarditis cases that remain do not exist.104 Its clinical features include high spiking fever, arthral-
in the category of FUO belong to the special subgroup comprising gias or arthritis, a transient maculopapular rash, lymphadenopathy,
culture-negative endocarditis (see Chapter 77). Rarely, FUO patients hepatosplenomegaly, serositis, and sore throat. Leukocytosis is gener-
with vegetations detected on a heart valve by echocardiography and ally marked, and rheumatoid factor and antinuclear antibody tests are
negative blood cultures prove to have nonbacterial thrombotic negative.
endocarditis (marantic endocarditis) associated with an undetected The fever of Stills disease is characteristically high and spiking, with
cancer. temperatures reaching as high as 41.6 C. The fever pattern may be
The reported incidence of culture-negative endocarditis has varied either intermittent (quotidian) or remittent. Although most patients
widely, between 2.5% and 31% of cases of infective endocarditis.97 The seek medical attention within 2 weeks of the onset of symptoms, some
high rates of negative blood cultures reported in some previous case 25% suffer for more than 4 weeks before doing so. A distinctive, eva-
studies most likely reflect the use of less effective culture techniques nescent, salmon-pink macular or maculopapular rash is typically
and less rigidly defined criteria for the diagnosis of endocarditis than present during the early course of the illness.
those employed in more recent series. Today, culture-negative cases Hepatomegaly or abnormal liver function tests, or both, are common
comprise 8% to 15% of the total. The most common cause of negative in Stills disease. Severe liver failure, however, is seen almost exclusively
blood cultures in infective endocarditis is prior antibiotic therapy; if in conjunction with aspirin or other nonsteroidal anti-inflammatory
such cases are excluded, less than 5% of cases of infective endocarditis drug therapy. In a third of patients, the disease has a self-limited
should exhibit negative blood cultures.98 course, in a quarter an intermittent course, and in a third a chronic
Some species of microorganisms that cause endocarditis can be course. Rarely, Stills disease has a fatal outcome.104 Three predictors
difficult to isolate using standard blood-culture techniques.97,99,100 of an unfavorable outcome include root joint (shoulder and hip)
A few important species, such as Bartonella spp., may require arthritis on presentation, polyarthritis, and rash. No other clinical or
prolonged-incubation for 2 to 3 weeks, or special culture media. If laboratory manifestations, including human leukocyte antigen (HLA)
blood cultures remain sterile after 5 to 7 days in cases in which a tests, predict the outcome.104
diagnosis of endocarditis seems highly likely, blind subculture from
liquid onto solid media may shorten the time required to isolate the
DRUG FEVER
etiologic microorganism.
Blood cultures are usually sterile in cases of endocarditis caused by Fever may be the sole or most prominent feature of an adverse drug
filamentous fungi. Both yeasts and filamentous fungi (especially Asper- reaction.105 This disorder is generally not accompanied by other signs
gillus spp.) should be considered as a cause of endocarditis responsible of drug allergy, notwithstanding statements to the contrary in numer-
for FUO in intravenous drug users, patients with prosthetic heart valves, ous textbooks and review articles. Although occasionally present,
and patients who have received prolonged intravenous antibiotic neither rash nor eosinophilia is common. Relative bradycardia is, like-
therapy.101 Serologic studies, and nucleic acid amplification tests per- wise, uncommon. Several authors have suggested that there might be
formed on vegetations or emboli, may be helpful in diagnosing cases of a characteristic drug-induced fever pattern. However, no such pattern
endocarditis due to Coxiella, Bartonella, Legionella, and Tropheryma has yet emerged, perhaps because antipyretics and external cooling
whipplei. These organisms should be considered in patients with FUO measures are so frequently used to treat drug-induced fever. Consider-
and suspected endocarditis in whom blood cultures are sterile.102 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
TEMPORAL ARTERITIS AND
taking a certain medication for a long time, it cannot be the cause of
POLYMYALGIA RHEUMATICA
FUO. Contrary to some early reports, there does not appear to be any
Temporal arteritis, also known as giant cell arteritis, and polymyalgia clear associations between drug fever and systemic lupus erythemato-
rheumatica are diseases with protean manifestations.103 No single clini- sus, atopy, female sex, or advanced age.105
cal symptom, sign, or noninvasive laboratory test can be relied on to
establish the diagnosis. More often than not, physical findings indica-
FACTITIOUS FEVER
tive of active arteritis are absent, and a history consistent with arteritis
is obtained only after careful questioning by a clinician fully versed in Although in most series, factitious fever is a relatively uncommon
the varied symptoms and signs of the disorder.103 In geriatric patients, cause of FUO, in one report it was diagnosed in 9% of cases of pro-
temporal arteritis may manifest only as unexplained fever or pro- longed FUO, suggesting that it might be more common than generally
longed malaise, depression, or anemia. appreciated.106 It has two formsa fraudulent form and a self-induced
The erythrocyte sedimentation rate is usually markedly elevated, form.107 Instances of patients manipulating a mercury thermometer no
usually more than 80 to 100 mm/hour, during active arteritis and is longer occur with the advent of electronic and infrared thermometry.
thus both useful as an aid to diagnosis and a clinical marker of disease Genuine fever can be induced by the injection or ingestion of pyro-
activity. Rarely, the erythrocyte sedimentation rate is normal during genic materials. Data falsification has also been used to create fraudu-
active arteritis. For this reason, symptoms or signs, or both, consistent lent fever.108
with the disease should be pursued even if the sedimentation rate is Patients with self-induced fever may also cause real fevers through
normal. the self-administration of pyrogenic substances, most often bacterial
Nodules or diminished temporal artery pulsations need not be suspensions.
present for an arterial biopsy to reveal active temporal arteritis. Because Patients with factitious fever are generally young women, approxi-
pathologic abnormalities may be confined to short segments of the 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- PERIODIC FEVERS
der. In the rare male patient who manifests factitious fever, the fraudu-
lent variety is most frequent. In women, the self-induced variety Many of the diseases mentioned earlier can cause periodic fevers,
predominates. In fact, all 12 of 12 cases of self-induced infection including juvenile rheumatoid arthritis, adult-onset Stills disease,
reported by Reich and Gottfried109 occurred among women. Factitious Crohns disease, bartonellosis, and Behets syndrome. There are also
fever, especially the fraudulent variety, appears to be increasing some- several rare, hereditary periodic fever syndromes included on the long
what in frequency among older patients.110,111 list of causes of FUO.112 These include familial Mediterranean fever,
Obviously, factitious fever must have some psychogenic basis. hyper-IgD syndrome, and TRAPS. Although genetic defects respon-
However, a survey of 41 cases of factitious disorders by Reich and sible for these three types of hereditary periodic fever have been
Gottfried109 found no evidence of major psychiatric diagnoses among identified,113 the striking periodicity of the clinical attacks remains
patients with either self-induced infection or simulated illnesses. Nor unexplained. They should be suspected in cases of FUO in which the
did these investigators find evidence of psychosis on projective testing fever is recurrent (i.e., febrile periods separated by symptom-free inter-
among the few patients who underwent detailed psychological vals of variable duration), persists for more than 2 years, and in which
analysis. there is a family history of similar attacks.
REFERENCES
1. Durack DT. Fever of unknown origin. In: Mackowiak PA, ed. 29. Zenone T. Fever of unknown origin in adults: evaluation of 144 54. Aduan RP, Fauci AS, Dale DC, et al. Prolonged fever of unknown
Fever. Basic Mechanisms and Management. 2nd ed. Philadelphia: cases in a non-university hospital. Scandinavian J Infect Dis. origin (FUO): A prospective study of 347 patients. Clin Res.
Lippincott-Raven; 1997:237-249. 2006;38:632-638. 1978;26:558A.
2. Keefer CS, Leard SE. Prolonged and Perplexing Fevers. Boston: 30. Ghose MK, Shensa S, Lerner PI. Arteritis of the aged (giant cell 55. Mackowiak PA. Commentary: Fever patterns. Infect Dis Clin
Little, Brown; 1955:1-248. arteritis) and fever of unexplained origin. Am J Med. Pract. 1997;6:308-309.
3. Petersdorf RG, Beeson PB. Fever of unexplained origin: Report 1976;60:429-436. 56. Woodward TE. The fever pattern as diagnostic aid. In: Mack-
on 100 cases. Medicine. 1961;40:1-30. 31. Tal S, Guller V, Gurevich A, et al. Fever of unknown origin in the owiak PA, ed. Fever: Basic Mechanisms and Management. 2nd
4. Alt HL, Barker MH. Fever of unknown origin. JAMA. elderly. J Intern Med. 2002;252:295-304. ed. Philadelphia: Lippincott-Raven; 1997;215-235.
1930;94:1457-1461. 32. Tal S, Guller V, Gurevich A. Fever of unknown origin in older 57. McCarthy PL, Dolan TF. Hyperpyrexia in children. Am J Dis
5. Wong SY, Lam M. Pyrexia of unknown originapproach to adults. Clinics Geriatr Med. 2007;23:649-668. Child. 1976;130:849-851.
management. Singapore Med J. 1995;36:204-208. 33. Speil C, Mushtaq A, Adamski A, et al. Fever of unknown origin 58. McCarthy PL, Jekel JF, Dolan TF. Temperature greater than or
6. Smith JW. Fever of undetermined origin: Not what it used to be. in the returning traveler. Infect Dis Clinics North Am. equal to 40 C in children less than 24 months of age: A prospec-
Am J Med Sci. 1986;292:56-64. 2007;21:1091-1113. tive study. Pediatrics. 1977;59:663-668.
7. Durack DT, Street AC. Fever of unknown origin reexamined and 34. Konecny P, Davidson RN. Pyrexia of unknown origin in the 59. Pizzo PA, Lovejoy FH, Smith DH. Prolonged fever in children:
redefined. Curr Clin Top Infect Dis. 1991;11:35-51. 1990s: Time to redefine. Br J Hosp Med. 1996;56:21-24. Review of 100 cases. Pediatrics. 1975;55:468-473.
8. Iikuni Y, Okada J, Kondo H, et al. Current fever of unknown 35. Neuhoff FJ, Brewer JE, Castaneda T, et al. Frequency and yield 60. Lohr JA, Hendley JO. Prolonged fever of unknown origin: A
origin 1982-1992. Intern Med. 1994;33:67-73. of postoperative fever evaluation. Infect Dis Obstet Gynecol. record of experiences with 54 childhood patients. Clin Pediatr.
9. Knockaert DC, Vanneste LJ, Bobbaers HJ. Fever of unknown 1998;6:252-255. 1977;16:768-773.
origin in elderly patients. J Am Geriatric Soc. 1993;41: 36. Anand VT, Phillipps JJ, Allen D, et al. A study of postoperative 61. Sheon RP, Van Ommeu RA. Fever of obscure origin: Diagnosis
1187-1192. fever following paediatric tonsillectomy. Clin Otolaryngol Allied and treatment based on a series of sixty cases. Am J Med.
10. Chantada G, Casak S, Daza Plata J, et al. Children with fever of Sci. 1999;24;360-364. 1963;34:486-499.
unknown origin in Argentina: An analysis of 113 cases. Pediatr 37. Shaw JA, Chung R. Febrile response after knee and hip arthro- 62. Vickery DM, Quinnell RK. Fever of unknown origin. An algo-
Infect Dis J. 1994;13:260-263. plasty. Clin Orthop Relat Res. 1999;367:181-189. rithmic approach. JAMA. 1977;238:2183-2188.
11. Knockaert DC, Vanneste LJ, Laurent J, et al. Recurrent or epi- 38. Circiumaru B, Baldock G, Cohen J. A prospective study of fever 63. Esposito AL, Gleckman RA. A diagnostic approach to the
sodic fever of unknown origin: Review of 45 cases and survey of in the intensive care unit. Intensive Care Med. 1999;25: adult with fever of unknown origin. Arch Intern Med.
the literature. Medicine. 1993;72:184-196. 668-673. 1979;139:575-579.
12. Knockaert DC. Diagnostic strategy for fever of unknown origin 39. Hall J. Assessment of fever in the intensive care unit. Is the 64. Hurley DL. Fever in adults. What to do when the cause is not
in the ultrasonography and computed tomography era. Acta answer just beyond the tip of our nose? Am J Respir Crit Care obvious? Postgrad Med. 1983;74:232-244.
Clin Belg. 1992;47:100-116. Med. 1999;159:693-694. 65. Gries E, Hoensch H, Ohnhaus EE. Differential diagnosis in fever
13. Kazanjian PH. Fever of unknown origin: Review of 86 patients 40. Georgilis K, Plomaritoglou A, Dafni U, et al. Aetiology of fever of unknown origin: significance of concomitant clinical symp-
treated in community hospitals. Clin Infect Dis. 1992;15: in patients with acute stroke. J Intern Med. 1999;246:203-209. toms [in German]. Klin Wochenschr. 1986;64:307-313.
968-973. 41. Engelhart S, Glasmacher A, Exner M, et al. Surveillance for noso- 66. Nolan SM, Fitzgerald FT. Fever of unknown origin: The general
14. Larson EB, Featherstone HJ, Petersdorf RG. Fever of undeter- comial infections and fever of unknown origin among adult internists approach. Postgrad Med. 1987;81:190-205.
mined origin: Diagnosis and follow-up of 105 cases, 1970-1980. hematology-oncology patients. Infect Control Hosp Epidemiol. 67. Koch H, Jenke V. Fever of unknown origina basic diagnostic
Medicine. 1982;61:269-292. 2002;23:244-248. program [in German]. Z Gesamte Inn Med. 1990;45:560-562.
15. Sharma BK, Kumari S, Varma SC, et al. Prolonged undiagnosed 42. Corey L, Boeckh M. Persistent fever in patients with neutrope- 68. Fischer JL, Henselmann L. Diagnostic steps in fever. Fever of
fever in Northern India. Trop Geogr Med. 1992;44:32-36. nia. N Engl J Med. 2002;346:222-224. unknown origin [in German]. Med Klin. 1987;82:229-235.
16. De Kleijn EMHA, Vandenbroucke JP, van der Meer JWM, et al. 43. Chang FY, Singh N, Gayowski T, et al. Fever in liver transplant 69. Palazzi DL, McClain KL, Kaplan SL. Hemophagocytic syndrome
Fever of unknown origin (FUO). I. A prospective multicenter recipients: Changing spectrum of etiologic agents. Clin Infect in children: An important diagnostic consideration in fever of
study of 167 patients with FUO, using fixed epidemiologic entry Dis. 1998;26:59-65. unknown origin. Clin Infect Dis. 2003;36:306-312.
criteria. Medicine. 1997;76:392-400. 44. Regnery RL, Anderson BE, Clarridge JE III, et al. Characteriza- 70. Kuse ER, Langefeld I, Jaeger K, et al. Procalcitonin in fever
17. Barbado FJ, Vazquez JJ, Pena JM, et al. Fever of unknown origin: tion of a novel Rochalimaea species, R. henselae sp nov, isolated of unknown origin after liver transplantation: A variable to
A survey on 133 patients. J Med. 1984;15:185-192. from blood of a febrile, human immunodeficiency virus positive differentiate acute rejection from infection. Crit Care Med.
18. Benito N, Nunez A, de Gorgolas M, et al. Bone marrow biopsy patient. J Clin Microbiol. 1992;30:265-274. 2000;28:555-559.
in the diagnosis of fever of unknown origin in patients with 45. Miralles P, Moreno S, Perez-Tascon M, et al. Fever of uncertain 71. Rowland MD, Del Bene VE. Use of body computed tomography
acquired immunodeficiency syndrome. Arch Intern Med. origin in patients infected with the human immunodeficiency to evaluate fever of unknown origin. J Infect Dis. 1987;156:
1997;157:1577-1580. virus. Clin Infect Dis. 1995;20:872-875. 408-409.
19. Norris AH, Krasinskas AM, Salhany KE, et al. Kikuchi-Fujimoto 46. Bissuel F, Leport C, Perrone C, et al. Fever of unknown origin 72. Quinn MJ, Sheedy PF II, Stephen DH, et al. Computed tomog-
disease: A benign cause of fever and lymphadenopathy. Am J in HIV-infected patients: A critical analysis of a retrospective raphy of the abdomen in evaluation of patients with fever of
Med. 1996;101:401-405. series of 57 cases. J Intern Med. 1994;236:529-535. unknown origin. Radiology. 1980;136:407-411.
20. McDermott EM, Smillie DM, Powell RJ. Clinical spectrum of 47. Wheeler DA, Call S, Wasserman SS, et al. Relationship between 73. Knockaert DC, Vanneste LJ, Vanneste SB, et al. Fever of unknown
familial Hibernian fever: A 14-year follow-up study of the index basal body temperature and stage of disease in asymptomatic origin in the 1980s: An update of the diagnostic spectrum. Arch
case and extended family. Mayo Clin Proc. 1997;72:806-817. HIV-infected men. Infect Dis Clin Pract. 1997;6:47-50. Intern Med. 1992;152:51-55.
21. Klastersky J, Weerts D, Hensgens C, et al. Fever of unexplained 48. Sepkowitz KA. Effect of prophylaxis on the clinical manifesta- 74. Arnow PM, Flaherty JP. Fever of unknown origin. Lancet.
origin in patients with cancer. Eur J Cancer. 1973;9:649-656. tions of AIDS-related opportunistic infections. Clin Infect Dis. 1997;350:575-580.
22. Hamman L, Wainwright CW. The diagnosis of obscure fever. 1998;26:806-810. 75. Knockaert DC, Mortelmans LA, De Roo MC, et al. Clinical value
II. The diagnosis of unexplained high fever. Johns Hopkins Hosp 49. Lozano F, Torre-Cisneros J, Santos J, et al. Impact of highly active of gallium-67 scintigraphy in evaluation of fever of unknown
Bull. 1936;58:307-331. antiretroviral therapy on fever of unknown origin in origin. Clin Infect Dis. 1994;18:601-605.
23. Berger L, Sinkoff MW. Systemic manifestations of hyperne- HIV-infected patients. Eur J Clin Microbiol Infect Dis. 2002;21: 76. Syrjala MT, Valtonen V, Liewendahl K, et al. Diagnostic signifi-
phroma: A review of 273 cases. Am J Med. 1957;22:791-796. 132-139. cance of indium-111 granulocyte scintigraphy in febrile
24. Shoji S, Imamura A, Imai Y, et al. Fever of unknown origin: A 50. Barbado FJ, Gomez-Cerezo J, Pena JM, et al. Fever of unknown patients. J Nucl Med. 1987;28:155-160.
review of 80 patients from the Shinetsu area of Japan from origin: Classic and associated with human immunodeficiency 77. Lavender JP, Lowe J, Barker JR. Ga-67 citrate scanning in neo-
1986-1992. Intern Med. 1994;33:74-76. virus infection. A comparable study. J Med. 2001;32:152-162. plastic and inflammatory lesions. Br J Radiol. 1971;44:361-366.
25. Saxe SE, Gardner P. The returning traveler with fever. Infect Dis 51. Armstrong WS, Katz JT, Kazanjian PH. Human immunodefi- 78. Hilson AJW, Maisey MN. Gallium-67 scanning in pyrexia of
Clin North Am. 1992;6:427-439. ciency virus-associated fever of unknown origin: A study of 70 unknown origin. BMJ. 1979;II:1330-1331.
26. Jacobs RF, Schutze GE: Bartonella henselae as a cause of pro- patients in the United States and review. Clin Infect Dis. 79. Sfakianakis GN, Al-Sheikh W, Heal A, et al. Comparison of scin-
longed fever and fever of unknown origin in children. Clin Infect 1999;28:341-345. tigraphy with 111In leukocytes and 67Ga in the diagnosis of occult
Dis. 1998;26:80-84. 52. Hot A, Schmulewitz L, Viard JP, et al. Fever of unknown origin sepsis. J Nucl Med. 1982;23:618-626.
27. Gleckman RA, Esposito AL. Fever of unknown origin in the in HIV/AIDS patients. Infect Dis Clin North Am. 2007;21: 80. Meller J, Altenvoerde G, Munsel U, et al. Fever of unknown
elderly: Diagnosis and treatment. Geriatrics. 1986;41:45-52. 1013-1032. origin: Prospective comparison of [18F] FDG imaging with a
28. Kauffman CA, Jones PG. Diagnosing fever of unknown origin 53. Knockaert DC. Recurrent fevers of unknown origin. Infect Dis double-bead coincidence camera and gallium-67 citrate SPECT.
in older patients. Geriatrics. 1984;39:46-51. Clin North Am. 2007;21:1189-1211. Eur J Nucl Med. 2000;7:1617-1625.
51 Fever of Unknown Origin 789
81. Blockmans D, Knockaert D, Maes A, et al. Clinical value of [18F] 92. Dinarello CA. Interleukin-1. Rev Infect Dis. 1984;6:51-95. 105. Mackowiak PA, LeMaistre CF. Drug fever: A critical appraisal of
fluoro-deoxyglucose positron emission tomography for patients 93. Lachman LB, Moore JO, Metzger RS. Preparation and charac- conventional concepts. Ann Intern Med. 1987;106:
with fever of unknown origin. Clin Infect Dis. 2001;32: terization of LAF from acute monocytic and myelomonocytic 728-733.
191-196. leukemic cells. Cell Immunol. 1978;41:189-206. 106. Aduan RP, Fauci AS, Dale DC, et al. Factitious fever and self-
82. Lorenzen J, Buchert R, Bohuslavizki KH. Value of FDG PET in 94. Esposito AL, Gleckman R. Fever of unknown origin in the induced infection, a report of 32 cases and review of the litera-
patients with fever of unknown origin. Nucl Med Comm. elderly. J Am Geriatr Soc. 1978;26:498-505. ture. Ann Intern Med. 1979;90:230-242.
2001;22:779-783. 95. Reimann HA. Periodic (Pel-Ebstein) fever of lymphomas. Ann 107. Sarwari AR, Mackowiak PA. Factitious fever. Curr Clin Topics
83. Larson EB, Featherstone HJ, Petersdorf RG. Fever of undeter- Clin Lab Sci. 1977;7:1-5. Infect Dis. 1997;17:88-94.
mined origin: Diagnosis and follow-up of 105 cases, 1970-1980. 96. AbuRahma AF, Saiedy S, Robinson PA, et al. Role of venous 108. Petersdorf RG, Bennet IL. Factitious fever. Ann Intern Med.
Medicine. 1982;61:269-292. duplex imaging of the lower extremities in patients with fever 1957;46:1039-1062.
84. Solis-Herruzo JA, Benita V, Morillas JD. Laparoscopy in fever of of unknown origin. Surgery. 1997;121:366-371. 109. Reich D, Gottfried LA. Factitious disorders in a teaching hospi-
unknown originStudy of seventy cases. Endoscopy. 97. Cannady PB Jr, Sanford JP. Negative blood cultures in infective tal. Ann Intern Med. 1983;99:240-247.
1981;13:207-210. endocarditis: A review. South Med J. 1976;69:1420-1424. 110. Potin M, Regamey C, Glauser MP. Factitious fever as a cause of
85. Mitchell DP, Hanes TE, Hoyumpa AM Jr, et al. Fever of unknown 98. Von Reyn CF, Levy BS, Arbeit RD, et al. Infective endocarditis: prolonged fevers. Apropos of 5 clinical cases [in French].
originAssessment of the value of percutaneous liver biopsy. An analysis based on strict case definitions. Ann Intern Med. Schweiz Med Wochenschr. 1983;113:1534-1539.
Arch Intern Med. 1977;137:1001-1004. 1981;94:505-518. 111. Marcus EL, Van Dijk JM. Fever of unknown origin [letter]. J Am
86. Chang JC, Hawley BH. Neutropenic fever of undetermined 99. Van Scoy RE. Culture-negative endocarditis. Mayo Clin Proc. Geriatr Soc. 1991;36:637-638.
origin (N-FUO): Why not use the naproxen test? Cancer Invest. 1982;57:149-154. 112. Drenth JPH, van der Meer JWM. Hereditary periodic fever. N
1995;13:448-450. 100. Washington JA II. The role of the microbiology laboratory in Engl J Med. 2001;345:1748-1757.
87. Rolston KVI. Neoplastic fever: All who shiver are not infected. the diagnosis and antimicrobial treatment of infective endocar- 113. Simon A, Mariman EC, van der Meer JWM, et al. A founder
Support Care Cancer. 2005;13:870-877. ditis. Mayo Clin Proc. 1982;57:22-32. effect in the hyperimmunoglobulinemia D and periodic fever
88. Hughes WT, Armstrong D, Bodey GP, et al. Guidelines for the 101. Rubinstein E, Noriega ER, Simberkoff MS, et al. Fungal endo- syndrome. Am J Med. 2003;114:148-152.
use of antimicrobial agents in neutropenic patients with unex- carditis: Analysis of 24 cases and review of the literature. Medi- 114. MacLean JD, Lalonde RG, Ward B. Fever from the tropics. Travel
plained fever. J Infect Dis. 1990;161:381-396. cine (Baltimore). 1975;54:331-334. Med Advisor. 1994;5:27.
89. Knockaert DC, Dujardin KS, Bobbaers HJ. Long-term follow-up 102. Tompkins LS, Roessler BJ, Redd SC, et al. Legionella prosthetic- 115. Doherty JF, Grant AD, Bryceson ADM. Fever as the presenting
of patients with undiagnosed fever of unknown origin. Arch value endocarditis. N Engl J Med. 1988;318:530-535. complaint of travellers returning from the tropics. QJM.
Intern Med. 1996;156:618-620. 103. Goodman BW. Temporal arteritis. Am J Med. 1979;67: 1995;88:277-281.
90. Bodel P. Tumors and fever. Ann N Y Acad Sci. 1974;230:6-13. 839-852.
91. Bodel P, Ralph P, Wenc K, et al. Endogenous pyrogen production 104. Pouchot J, Sampalis JS, Beaudet F, et al. Adult Stills disease:
by Hodgkins disease and human histiocytic lymphoma cell lines Manifestations, disease course, and outcome in 62 patients.
in vitro. J Clin Invest. 1980;65:514-518. Medicine. 1991;70:118-136.