on 03-Sep-2023 (Sun)

Occurrence of several occasions of fever higher than 38.3°C (100.9°F) with a duration greater than at least 3 weeks despite 1 week of hospital evaluation is still recognized as the classic definition for fever of unknown origin (FUO)

The frequency with which infections and neoplasms have been identified as the causes of classic FUO has decreased steadily, whereas the proportion of miscellaneous causes and undiagnosed conditions has risen in recent years.

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 classic text Prolonged and Perplexing Fevers, by Keefer and Leard (published in 1955),2 and the paper “Fever of unexplained origin: report on 100 cases,” by Petersdorf and Beeson (published in 1961).3

In response to this evolving environment, Durack and Street,7 in 1991, proposed a revised definition in which cases of FUO require 3 days of investigation and must be codified into four distinct subclasses of the disorder: classic FUO, nosocomial (health care–associated) FUO, neutropenic (immune- deficient) FUO, and HIV-related FUO

Classic 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 classic FUO have subacute or chronic symptoms and therefore can be safely investigated as outpatients. In published series of such patients, for example, the median duration of fever before diagnosis was between 40 and 44 days.1,9

Over the years a recurrent theme has become clear: of the myriad disorders causing classic FUO almost all fall within one of five categories: infections, neoplasms, connective tissue diseases, miscellaneous other disorders, and undiag- nosed illnesses.

The relative frequencies of individual diagnoses within these five categories vary depending on the decade (Fig. 56.1), geographic region, ages of the patients, type of medical practice, and other factors.

In more recent published series, infections have comprised the largest category, accounting for 16% to 55% of cases (Table 56.1).4 However, in patients older than 65 years, infections become less common, falling into second or third place as a cause of classic FUO.4,8,10

Knockaert and associates,10 infection was the cause of FUO in only 25% of patients 65 years of age or older; temporal arteritis and various connective tissue diseases accounted for 31% of cases, and tumors accounted for 12%. Only 8% of cases went undiagnosed, which was a percentage similar to that reported by Colpan and colleagues9 but one substantially lower than that reported in surveys involving younger adults, in which as many as 30% of cases remained 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 classic FUO, abscesses, endocarditis, tuberculosis, and complicated urinary tract infections have consistently been among the most important.

Visceral leishmaniasis, for example, although absent from most series of classic FUO, accounted for 8% of cases in a study reported in 1997 from Spain.18

Other examples of causes of classic FUO in distinct populations include familial Mediterranean fever in Ashkenazi Jews; Kikuchi-Fujimoto disease, an unusual form of necrotizing lymphadenitis seen primarily in Japan19; and tumor necrosis factor receptor–associated periodic syndrome (TRAPS) fever, formerly called familial Hibernian fever, an inherited periodic fever syndrome described originally in Ireland.20

The miscellaneous category contains both varied and individually rare causes of classic FUO (Table 56.2).

Of the connective tissue diseases responsible for classic FUO, Still 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 cytokines, as in the case of certain lymphomas. They may also generate fevers indirectly by undergoing induced or spontaneous necrosis and/or by creating conditions conducive to secondary infections.1

Among the malignant neoplasms responsible for FUO, hematologic cancers, hypernephromas, and gastrointestinal (mainly colorectal) and central nervous system cancers have been documented as common causes.3,4 In a recent series of 80 patients with FUO, Ergonul and colleagues8 established a malignant neoplasm etiology in 14 cases (18%).

Since the mid-1990s, the frequency with which infections and malignant neoplasms have been identified as causes of classic FUO has fallen steadily, whereas the proportion of miscellaneous causes and undiagnosed conditions has risen.24 However, in developing countries, the frequency with which infections are diagnosed has changed little.15

Whereas connective tissue diseases are rarely seen in children younger than 12 months, Still disease is a leading cause of FUO in older children and young adults. Joint involvement in children with FUO usually signifies a serious underlying disorder, such as a connective tissue disease, endocarditis, or leukemia.11

Chow and Robinson26 analyzed 18 papers concerned with pediatric FUO published between 1968 and 2008. Of 1638 children ages birth to 18 years, 832 (51%) had infections, 93 (6%) had malignant neoplasms, 150 (9%) had noninfectious inflammatory diseases, 179 (11%) had miscellaneous causes such as inflammatory bowel disease and Kawasaki disease, and 384 (23%) had no diagnosis. Although the distribution of diagnostic categories was similar among developed versus developing countries, urinary tract infections, brucellosis, tuberculosis, and typhoid fever were more common in developing countries. The most common infections diagnosed in cases of FUO in developed countries included urinary tract infections, osteomyelitis, tuberculosis, and bartonellosis (e.g., Bartonella henselae).

Although a daily rectal temperature greater than 38.3°C (100.9°F) lasting more than 2 weeks despite diagnostic evaluation has been defined as FUO in children, 9 of 18 papers analyzed by Chow and Robinson used the classic definition of FUO.26

One of the most striking features of classic FUO in patients older than 65 years is the relatively high frequency with which connective tissue diseases are identified as the cause of the illness (Table 56.3).4,10,28,29,30

In developed countries, connective tissue diseases surpass even infections as the leading cause of classic FUO in the elderly.4,10,29 This is primarily because the temporal arteritis and polymyalgia rheumatica syndromes are common in this setting.4,29,30,31

In elderly patients in whom infections are identified as the cause of FUO, intraabdominal abscesses, complicated urinary tract infections, tuberculosis, and endocarditis have predominated.10

For unclear reasons, factitious fever is a rare cause of FUO in older adults.

The occurrence of classic FUO in an elderly patient has a distinctly poorer prognosis than for the younger patient because of the relatively high incidence of malignancies in the elderly.32

Fever in returned travelers (see Chapter 319) is most often due to common infections, such as malaria and respiratory or urinary tract infections.33

Of the many febrile conditions encountered among returning travelers (Table 56.4), malaria, typhoid fever, and acute HIV infection are the ones most likely to manifest as FUO.

Nosocomial (health care–associated) FUO, as the name implies, is a condition in which patients first manifest fever during active medical treatment for some other illness.

Leading examples of causes attributable to health care– associated FUO include drug fever, postoperative complications (i.e., occult abscesses), septic thrombophlebitis, recurrent pulmonary emboli,

myocardial infarction, cancers, blood transfusion, and Clostridioides

difficile (formerly Clostridium difficile) colitis.4,7,34

Several reports have highlighted the fact that it is often difficult to identify the precise cause of postoperative fevers.

In a series of 537

consecutive patients undergoing major gynecologic surgery, 211 (39%)

developed postoperative fever.35

Of patients with evidence of urinary tract infections and pneumonia, the median length of postoperative fever was 2 days (range: 1–7 days) and 3 days (range: 2–4 days), respectively.

n a prospective study by Kendrick and colleagues36 of postoperative fever among 292 patients admitted to a gynecologic oncology service after abdominal or vaginal operation, 58 (20%) patients developed postoperative fever. Among 37 (16%) low-risk surgical patients developing postoperative fever, only 6 (3%) had an infection diagnosis. The majority of infections occurred within 4 days of the operative procedure and included pneumonia, vaginal cuff cellulitis, and urinary tract infection

Although fever is a well-recognized manifestation of some surgical procedures, most episodes are short lived and do not meet the classic definition of FUO. Postoperative fevers generally do not require extensive diagnostic investigation for unusual causes of fever

In another series concerned with the etiology of persistent post- operative fever in patients undergoing total joint arthroplasty, few definitive diagnoses were established, causing the authors to conclude that postoperative fever (postoperative days 1 through 5) is a normal component of the inflammatory response to this type of major surgery.37

Fever is common in intensive care units (ICUs), most often developing relatively early after admission to the ICU, in which case it tends to be of noninfectious 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 ICU patients.

More often than not, the causes of fever in the ICU are essentially no different from general causes of health care–associated fevers (i.e., abscesses, drug fever, postoperative complications, septic thrombophle- bitis, recurrent pulmonary emboli, myocardial infarction, cancers, blood transfusion, and C. difficile colitis).

Ventilator-associated pneumonia is frequently underdiagnosed among patients with respiratory failure, fever, and nonspecific radiographic pulmonary densities. Patients with acute respiratory distress syndrome (ARDS) tend to develop pneumonia earlier than non-ARDS patients.

In patients with a recent stroke, fever is usually the result of an infection, most commonly a urinary tract infection related to urinary catheterization or a respiratory tract infection. 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 associ- ates40 observed that noninfectious fevers were most often associated with intracranial mass effects and tended to occur earlier after the onset of stroke than fevers due to infection

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 features 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 illustrated in a prospective evaluation of patients with leukemia and lymphoma by Toussaint and coworkers.43 In that series, infections were the cause of 319 (67%) of the 477 episodes of fever. The majority of infections included respiratory tract infections (28.8%), secondary bacteremia due to gram-negative bacilli (15.7%), genitourinary tract infections (12.9%), skin and soft tissue infections (11.3%), and primary bacteremia (11.0%). One hundred nine (23%) cases were due to noninfectious conditions: malignant neoplasm, metastatic disease, and drug-induced fever. While noninfectious neoplastic-related fever was more common (41%) among nonneutropenic patients, noninfectious drug-induced fever was more common among neutropenic patients (13%). In 47 (10%) cases, the cause of the fever could not be determined

Not only are the duration and nadir of neutropenia correlated with the incidence of fever and infections, but so are outcomes related to morbidity and mortality.

The highest percentages of deaths occur among patients with persistent severe neutropenia (<3).

Furthermore, severe infections increase with the increasing proportion of time neutropenia and lymphopenia levels are less than 500 cells/mm3. However, the incidence of fever and infectious episodes decreases when levels are 1500 cells/mm3, above which level there is no associated decreased incidence, or the expected duration of neutropenia is less than 7 days.

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.

However, only about 35% of prolonged episodes of febrile neutropenia (usually defined as persistent fevers for >7 days after initiation of empirical antimicrobial therapy in association with a negative workup, and neutropenia expected to last >7 days) respond to broad-spectrum antibiotic therapy.

Practitioners often assume that if fever does not respond promptly to antibacterial therapy, fungal infection must be responsible, but other potential causes are at least as likely to be identified.42,44

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 seroconversion. 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 subclinical infection during which they are usually afebrile.48 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 oppor- tunistic infections, which tend to manifest in atypical fashion owing to the tendency of the disordered immune response or prior therapy, or both,49 to distort their clinical features.

In a report of 274 HIV-infected patients, for example, Abellan-Martinez and colleagues45 observed an incidence range of 2.57 to 3.66 FUO episodes per 100 HIV-infected patients per year prior to the initiation of HAART and an incidence range of 0.84 to 1.24 episodes after the introduction of therapy.

In another study, the frequency of FUO was 3% in untreated patients compared with 0.6% in those receiving therapy.50 Mycobacterial infections have been the most common cause of FUO in such patients; collagen vascular diseases have been distinctly uncommon.45,51

Most cases of FUO in HIV-infected patients are the result of opportunistic infections, the specific frequencies of which are dictated, at least in part, by geographic variation in the prevalence of these infections.53

A point deserving emphasis is that most patients diagnosed with “classic FUO” were not suffering from unusual or rare conditions; rather, they exhibited atypical manifestations of common illnesses in an era when advanced diagnostic laboratory or imaging studies were not available.

The most important lesson learned from the “classic FUO” concept is that, in many instances, available information from the history and physical examination should be utilized more often.

Changes to the original FUO definition by Durack and Street7 emphasize that treatment is clearly more time critical for certain classes than others.

The evaluation of a patient with FUO typically includes a comprehensive history, verification that the patient actually has fever, consideration of the fever pattern, repeated physical examina- tions, and a host of laboratory investigations, key imaging studies, and invasive diagnostic procedures (Table 56.6).

Particular attention should be given to recent travel, exposure to pets and other animals, the work environment, and recent contact with people exhibiting similar symptoms. The family history 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, or intraab- dominal disorders, complications or reactivation of which might account for the source of fever.

In cases of recurrent FUO, rare disorders are more common than in nonrecurrent cases and are also more likely to remain undiagnosed.54

Finally, a complete list of the patient’s medications must be obtained so that each may be evaluated as a potential source of drug-induced fever.

The next step in the evaluation of the patient with FUO is to verify the presence of fever. The importance of this step should be self-evident, yet it is often overlooked. In fact, in a series of 347 patients admitted to the National Institutes of Health for investigation of prolonged fever, 35% were ultimately determined either not to have significant fever at all or to have fever of factitious origin.55

Clinicians have endeavored to diagnose particular diseases by analyz- ing the associated fever patterns since the earliest days of clinical thermometry.56

A few, such as the Pel-Ebstein pattern seen in some cases of Hodgkin disease, the typhus inversus (i.e., reversal of the normal diurnal pattern) in some cases of disseminated tuberculosis, and the pulse-temperature disassociation sometimes seen in typhoid fever, have been posited as having diagnostic value (Fig. 56.2). Unfortunately, with the possible exception of the well-known periodicity of tertian and quartan malaria, these fever patterns are neither sufficiently sensitive nor specific for diagnosis of any disease.

Thus, although fever patterns per se are rarely diagnostic, they occasionally offer useful information57 and should be considered in the context of other signs, symptoms, and laboratory data.

The resolution of fever after the institution of a disease-specific therapy (e.g., empirical therapy for suspected tuberculosis) sometimes provides strong evidence to support a presumptive diagnosis

In pediatric populations, the height of a fever correlates roughly with the likelihood of bacteremia. McCarthy and coworkers58,59 reported that in young children with febrile illnesses, the likelihood of bacteremia is 7% in those with temperatures of 40°C (104°F) or less, 13% with temperatures of 40.5°C to 41°C (104.9°F to 105.8°F), and 26% with temperatures of 41.1°C (106°F) or greater.

It is commonly believed that there may be a similar relationship between high fever and the likelihood of bacteremia in adults; this belief has not been substantiated. In any case, the relationship is at best loose, even in children, with numerous examples of bacteremia in which there is little or no fever and nonbac- teremic conditions, such as drug-induced fever, thrombophlebitis, and recurrent pulmonary emboli, in which extremely high fevers sometimes are encountered.

In the investigation of FUO, several aspects of the physical examina- tion should be accorded closer scrutiny than generally given during the evaluation of other illnesses (Table 56.7). Frequently, key physical abnormalities in patients with FUO are so subtle as to require repeated examinations to be appreciated. Examples include the nodular or weakly pulsatile temporal artery of temporal arteritis, the telltale oral ulcers of disseminated histoplasmosis or Behçet syndrome, the choroid granuloma or epididymal nodule of extrapulmonary tuberculosis, the testicular nodule of polyarteritis nodosa, and the vague rectal fluctuance of a perirectal abscess.

The diagnostic yield of the physical examination alone in the evaluation of FUO has not been studied directly. Nevertheless, in two pediatric series, abnormal physical findings were reported to have contributed to the diagnosis in 60% of cases of FUO.60,61 In half of these, the abnormalities were detected only after repeated examina- tions.

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.60,62

The literature is replete with algorithms indicating which laboratory tests should be performed to evaluate FUO.4,21,29,63,64,65,66,67

When formulating a diagnostic plan for FUO patients, the clinician should remember the old saying that the cause is more often a common disease presenting in atypical fashion than a rare disease presenting in typical fashion.

Inappropriate diagnostic tests in the evaluation of FUO may not only delay identification of the correct diagnosis but also result in false-positive tests leading to misguided treatment plans.4,21,66

A well-conducted history and physical examination is perhaps more important than any other diagnostic procedure in focusing the FUO evaluation, accelerating initiation of appropriate therapy while minimizing the cost and potential toxicity of unnecessary interventions

In most series, noninvasive laboratory tests have yielded the diagnosis in approximately a fourth of the cases.10,24 The most useful of these have been serologic tests for microbial pathogens and for various rheuma- tologic disorders.

Examination of blood smears is occasionally diagnostic, especially in patients with tick-borne or louse-borne relapsing fever, anaplasmosis, and ehrlichiosis.

Paradoxically, the advent of enhanced microbial culture systems has had less of an impact on the proportion of successful diagnoses than might have been anticipated. This is because modern culture systems have become so proficient at recovering fastidious bacteria, mycobacteria, and fungi from blood that they provide the diagnosis promptly, before the time required to meet the definitions of FUO has elapsed.

Bone marrow examination should be considered for diagnosis of suspected granulomatous diseases (e.g., tuberculosis, histoplasmosis, and sarcoidosis), carcinomatosis, and hemophagocytic syndrome, especially in patients with abnormal complete blood cell counts.68,69,70 In two series, bone marrow examination contributed to the diagnosis of FUO in approximately a fourth of cases.68,69

In transplant patients, procalcitonin levels may be of use in differentiating infection from acute organ rejection, in that elevated levels are seen in the former but not the latter condition.71

In one series, more than three CT or ultrasound examinations, or both, were performed for each FUO patient evaluated.72 Neverthe- less, the diagnostic yield per test performed was low—about 10%.

Of the major diagnostic imaging modalities, another more recent series reported sensitivities of 60% for plain-film chest radiography, 82% for chest CT, 86% for abdominal ultrasound, and 92% for abdominal CT.67

False-negative CT studies have occasionally been encountered, even in cases of abscesses in solid organs, as a result of distortions of normal anatomy, small abscess size, or failure to use both oral and intravenous contrast agents.67,73

In a series of 67 patients undergoing evaluation for FUO, Wagner and colleagues74 reported that application of MRI of the aortic arch and proximal cervical arteries in selected patients improved the diagnosis of large vessel vasculitis by 20%.

Although one case did not demonstrate vasculitic changes by cervical imaging, four cases would have remained undiagnosed without the application of MRI. The most common diseases diagnosed by this testing modality were giant cell arteritis (55%), Takayasu arteritis (27%), Wegener granulomatosis (9%), and microscopic polyangiitis (9%).

In a meta-analysis reported by Dong and colleagues,83 FDG-PET/CT exhibited an overall pooled sensitivity of 98.2% and specificity of 85.9% for establishing a diagnosis in patients with FUO. Of the major diagnostic categories, the reported sensitivities were 86.7% for malignant neoplasms, 81.5% for infectious diseases, and 76.3% for noninfectious inflammatory conditions. The authors concluded that FDG-PET/CT should be considered when conventional diagnostic methods have been unsuccessful in the diagnosis of an etiology for FUO

These findings suggest that FDG imaging should be performed earlier, rather than later, in the diagnostic evaluation of the patient with FUO. Diagnostic considerations for which this technology may be most helpful included localized abscesses, osteomyelitis, sinusitis, sarcoidosis, vasculitis, adult-onset Still disease, Crohn disease, and subacute thyroiditis

Histopathologic examination of tissues obtained by excisional biopsy, needle biopsy, or laparotomy can provide a definitive diagnosis in some cases; however, in most published series of FUO patients, biopsy gave the final answer in fewer than half of cases.73

The majority of patients with FUO undergo at least one such procedure, even though the diagnostic yield is only fair, with an average of two or three, sometimes even more, separate biopsies required to establish a final diagnosis.14,84

An important exception to the injunction against blind biopsies concerns the temporal artery, which may merit biopsy in an elderly FUO patient with an erythrocyte sedi- mentation rate greater than 50 to 100 mm/h, even in the absence of localizing signs.10

In a recent retrospective study of 100 patients continuously observed for classic FUO, Mete and associates85 successfully identified specific etiologies in 61% of patients based on clinical features and noninvasive tests. While invasive procedures were performed in 79% of patients, a diagnostic benefit was obtained in only 49% of the cases. Biopsy procedures were the most common invasive procedure performed, yielding a diagnosis in 42% of cases.

The contribu- tion of laparoscopy and laparotomy to the diagnosis of FUO may be most helpful in patients with solid cancers, lymphomas, and disseminated tuberculosis.85,86

Familial Mediterranean fever (FMF) is a hereditary inflammatory disease transmitted in an autosomal-recessive pattern characterized

by short, recurrent attacks of fever with abdominal, chest, or joint

pain and erysipelas-like erythema.54,63,64,87

Fever is the main clinical manifestation, lasting up to 96 hours, followed by a relatively long interval between attacks. The disease occurs predominantly among Sephardic Jews, Armenians, Turks, and Arabs

Underlying diseases may remit spontaneously during the course of ineffective therapy, giving the false impression of success.

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 antiinflammatory agents such as naproxen than fevers of infectious origin,90,91 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.

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

Traditional classic FUO cases are much less common than the complex and multifactorial FUO cases encountered among ICU patients. The differential diagnosis for prolonged febrile episodes has expanded well beyond infectious causes to what has recently been termed fever of too many origins.92

A fundamental principle in the management of classic 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 specific diagnosis.1

As a result, clinicians may feel compelled to treat symptoms empirically, even though the agents used may obscure the very signs and symptoms on which the diagnosis depends.

An important exception is that empirical treatment with corticosteroids may be appropriate in patients with suspected temporal arteritis to prevent vascular complications such as blindness or stroke.

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 antipseudomonal antimicrobial therapy immediately after samples for appropriate cultures have been obtained (see Chapter 305).

Diagnostic delay affects the prognosis adversely in intraabdominal infections, miliary tuberculosis, disseminated fungal infections, and recurrent pulmonary emboli.73

Patients in whom FUO remains undiagnosed after extensive evaluation generally have a favorable outcome, characteristically with resolution of their fever in 4 or more weeks without sequelae.3,84,87

In this series, the 5-year mortality rate for undiagnosed FUO was only 3.2%.

In a series of 91 patients undergoing evaluation for FUO, Mansueto and colleagues84 reported 29 patients (31.8%) who were discharged without a diagnosis and followed over a period of 48 months. Although a definitive diagnosis was established in eight cases, four of such patients died as a result of noninfectious complications related to neoplastic conditions.