FEVER OF UNKNOWN ORIGIN
In 1961, Petersdorf and Beeson defined the clinical syndrome of fever of unknown origin (FUO) as continuous fever of at least 3 weeks’ duration, with temperatures of 101°F or more, that remains unexplained despite 1 week of complete investigation. Using this definition, the authors studied 100 patients at a university teaching hospital. This study, which divided the causes of FUO into infections, malignancy, and rheumatic diseases and described drug fevers, factitious fevers, and cranial arteritis as additional diagnoses, demonstrated that patients may often present with atypical manifestations of common diseases. The authors recommended no one battery of tests and discouraged therapeutic trials unless treatment was directed at a particular disease. Biopsy specimens were frequently required for specific diagnosis.
Petersdorf and colleagues looked again at FUO between the years 1970 and 1980. This second study, of 105 patients, was somewhat different from the initial one of 20 years earlier:
Neoplastic disease became the most common cause of FUO. Most neoplastic cases were secondary to Hodgkin’s or non-Hodgkin’s lymphoma. In these cases, peripheral lymphadenopathy was absent, and disease was confined to retroperitoneal nodes or the liver. Patients were usually elderly. Not all recent studies support the high incidence of malignancy. Knockaert et al., in a study from Belgium, found tumor to be a less important cause of FUO in a 1980–1989 prospective series.
Bacterial endocarditis, which caused 5 of 100 cases of FUO in the first study, did not occur in any patient during the 1970–1980 study. Better culture methods, echocardiography, and appreciation of the causes of culture-negative endocarditis may explain this change. Tuberculosis also became less common. Biliary tract infection continued to be an important cause of FUO, even in an era of better imaging techniques.
Rheumatic fever was a cause of FUO in only one patient in the second study, compared with six patients in the first. However, rheumatic fever has reemerged in some parts of the United States in the past several years; young clinicians who are unfamiliar with the disease might overlook it as a cause of fever.
Systemic lupus erythematosus is no longer a cause of FUO owing to better serologic methods.
Cytomegalovirus caused FUO in four patients in the second study and in none in the first. Patients presented with a mononucleosis-like syndrome, had self-limited disease, and had often undergone transfusion or coronary artery bypass surgery.
Continued technologic improvements make some types of FUO less common. Progress in transesophageal echocardiography has made endocarditis increasingly less common as a cause of FUO. Improved methods of computed tomography (CT) and magnetic resonance imaging (MRI) make it possible for most intraabdominal abscesses to be diagnosed in less than 3 weeks. Better serologic tests have made most collagen vascular diseases less likely to become FUOs.
There are relatively few important new causes of FUO. Cunha describes histiocytic necrotizing adenitis (Kikuchi’s syndrome) and hypergammagobulinemia IgD syndrome, but these diseases are rare.
As noted by Cunha, the workup of a patient with FUO is now usually performed on an outpatient basis. Although procedures such as bone marrow biopsy, transesoph-ageal echocardiography, liver biopsy, and CT or MRI are now easily performed in an outpatient setting, the overall evaluation must be organized in a logical order and completed in a timely manner. Factors such as the overall toxicity of the patient, vital signs during febrile episodes, and rate of weight loss may help determine whether the workup requires a period of hospitalization. In general, basic aspects of the evaluation, such as history and physical examination, CBC count, erythrocyte sedimentation rate, blood cultures, liver function tests, and urinalysis, should be completed before hospitalization.
The rising incidence of tuberculosis in the United States will likely be reflected in more cases of this disease presenting as FUO. The enhancing or boosting phenomenon in serial tuberculin testing will help exclude false-negative tests. Elderly patients in particular may have a positive purified protein derivative (PPD) test result when given a second test 1 week after a negative result on the first test. When tuberculosis presents as an FUO, it will usually be associated with miliary disease. Disseminated pulmonary lesions may be so small that they are not appreciated on chest x-ray films. This is particularly likely to occur in the elderly.
Extrapulmonary tuberculosis may also present as FUO, although better imaging techniques should help in the diagnosis. Sterile pyuria in a patient with an FUO should suggest renal tuberculosis. In some cases, the urinary sediment may be unremarkable but the urine culture is positive. Tuberculosis of mesenteric lymph nodes may present without abdominal pain but can be diagnosed by abdominal CT scan. Tuberculosis of the endometrium and fallopian tubes usually causes fever and menstrual irregularities. Tuberculous pericarditis should be suggested by changing cardiac silhouette on chest x-ray films, with confirmation of pericardial effusion on echocardiogram.
The literature on the FUO syndrome has recognized that the problem varies with the patient population studied. Durack and Street have redefined FUO to include (a) classic FUO, (b) nosocomial FUO, (c) neutropenic FUO, and (d) HIV-associated FUO. FUO that develops in a hospitalized patient, although not well studied, requires a different approach than does classic FUO. The history and physical examination, particularly in the ICU setting, may be less reliable, making clinical diagnosis more difficult. Diseases causing FUO may include pulmonary embolism, transfusion-related hepatitis or cyto-megalovirus infection, drug fever, IV line infection, intubation-related sinusitis, or Clostridium difficile colitis. FUO in neutropenic patients requires a modification in definition. These fevers of unknown cause occur abruptly and tend not to be prolonged, as patients usually respond to treatment or succumb to infection. Unlike the treatment of classic FUO, empiric treatment of neutropenic FUO must be begun quickly. Most cases of FUO in neutropenic patients are bacterial infections without an obvious source. Disseminated candidemia and aspergillosis are also common. Careful, repeated physical examination, including examination of the anorectal area, is important. FUO is extremely common in AIDS patients, although it has not been studied well as an entity. Fever may be caused by HIV infection itself. Mycobacterium avium-intracellulare infection, tuberculosis, and cytomegalovirus infection are among the common causes of FUO in AIDS patients, often taking weeks to declare themselves as clinical entities. Salmonella infection, fungal infection, and lymphoma are also well described.
Several studies have assessed FUO in the elderly patient population (Table 54-1). Giant-cell arteritis was responsible in 16% of elderly patients with classically defined FUO in one study. This syndrome may include headaches, transient visual loss, and polymyalgia rheumatica symptoms. Respiratory tract symptoms of cough, sore throat, and hoarseness have been more recently described. The temporal artery may be tender, nodular, and pulseless. Often in the FUO patient, these symptoms are not prominent and may be absent altogether. The erythrocyte sedimentation rate is almost always very high in giant-cell arteritis; however, an elevated erythrocyte sedimentation rate is common in many cases of FUO. Temporal biopsy is the diagnostic intervention of choice when the disease is suspected. Results of empiric therapy with steroids may be dramatic and is indicated when vision is threatened. However, a definitive diagnosis is necessary to justify long-term steroid therapy.
Table 54-1. Causes of fever of unknown origin in the elderly
Biliary tract and intraabdominal infection are also common causes of FUO in the elderly. Hepatic abscess, cholecystitis, and empyema of the gallbladder may present with fever alone, particularly in the debilitated patient.
Weinstein has described the clinically benign FUO syndrome. These patients may have persistent fever that is clinically unimportant. Temperature elevation may be minimal or reach 102°F to 104°F. Despite persistent fever, patients generally appear well nourished. Malaise and aching joints are common for patients with low-grade fever. Diurnal variations, ovulation, use of tobacco and chewing gum, and exercise can cause these low-grade fevers. In patients who appear well but have a high-grade fever, drugs, occupational exposures, and metabolic disorders may be causative.
Another subcategory of FUO has been defined by Knockaert. Some patients have recurrent or episodic fevers that may last months or years. Subacute cholangitis, chronic prostatitis, adults Still’s disease, and familial Mediterranean fever are common diagnoses in these patients.
Mackowiak has reviewed the characteristics of drug-induced fever, an increasingly common problem. Fifty-one episodes of drug fever and 97 cases from the literature were analyzed. Hectic fever patterns and shaking chills were common. Rashes were seen in only 18% of patients and eosinophilia in 22%. The time between initiation of the drug and development of fever averaged 8 days. Patients tended to tolerate drug fever well, and exaggerated or dangerous reactions did not develop when the patients were rechallenged. Antimicrobials appear to be becoming a more common cause of drug fever in several studies.
When FUO is studied in a community hospital, the etiologies may be somewhat different than in a referral hospital. Pulmonary emboli and alcoholic hepatitis may cause prolonged fever in this setting, whereas illnesses such as familial Mediterranean fever will be rare. One recent study finds AIDS and Lyme disease to be causes of FUO in a community hospital.
Recent literature on the FUO syndrome has focused on the role of imaging in diagnosis. Schmidt et al., using granulocyte scintigraphy with indium 111, studied 32 patients who had at least 3 weeks of unexplained fever. Focal infections were identified in five patients—four with abdominal infection and one with dental abscess. Intestinal activity was observed in one patient found to have Whipple’s disease. McNeil et al. prospectively compared CT, ultrasound, and gallium imaging in patients thought to have a focal source of sepsis. The diagnoses of the patient population studied were uncertain after standard diagnostic evaluation. All three modalities had a similar ability to detect sepsis, but sensitivity was increased by the use of any two. Rowland and Del Bene studied the impact of CT in the workup of FUO. They found that CT reduced the incidence of biopsy of normal tissue, with CT of the abdomen often correctly directing the investigation to laparotomy.
Several authors have proposed algorithms and standard workup for FUO. However, history, physical examination, and clinical judgment will continue to dictate a patient-oriented approach to the diagnosis of this syndrome. (S.L.B.)
Brusch JL, Weinstein L. Fever of unknown origin. Med Clin North Am 1988;72:1247.
Basic review of FUO that includes many clinical insights and contains an interesting section on miscellaneous causes.
Chang JC. NSAID test to distinguish between infectious and neoplastic fever in cancer patients. Postgrad Med 1988;84:71.
Nonsteroidal antiinflammatory drugs are helpful in distinguishing FUO of malignancy from that of infection; they have an antipyretic effect only in tumor patients.
Chang JC. Neoplastic fever: a proposal for diagnosis. Arch Intern Med 1989;149:1728.
Proposes criteria for the diagnosis of fever caused by cancer. Some would disagree with the use of empiric antimicrobial therapy in the FUO setting.
Chang JC, Gross HM. Utility of naproxen in the differential diagnosis of fever of undetermined origin in patients with cancer. Am J Med 1984;76:597.
Of 15 patients with neoplastic fever, 14 responded to naproxen. None of five cases of infectious fever resolved.
Cunha BA. Fever of unknown origin. Infect Dis Clin North Am 1996;10:111.
Updates the causes of FUO and describes the ambulatory workup appropriate for most patients. Categorizes diseases by laboratory abnormality, history, and physical examination clues.
Dinarello CA, Wolff SM. Molecular basis of fever in humans. Am J Med 1982;72:799.
Insights into the pathogenesis of fever are useful in the understanding of FUO.
Drenth JPH, et al. Hyperimmunoglobulinemia D and periodic fever syndrome. Medicine 1994;73:133.
This syndrome presents with prolonged fevers, arthritis of large joints, and rash. Serum levels of IgD are greater than 100 U/mL.
Durack DT, Street AC. Fever of unknown origin reexamined and redefined. In: Remington JS, Swartz MN, eds. Current clinical topics infectious disease. New York: McGraw-Hill, 1990.
Redefines FUO based on patient group. Categories of FUO include classic, nosocomial, neutropenic, and FUO in HIV-positive patients.
Gartner JC Jr. Fever of unknown origin. Adv Pediatr Infect Dis 1992;7:1.
Reviews series of FUO cases among children.
Ghose MK, Shensa S, Lerner PI. Arteritis of the aged (giant-cell arteritis) and fever of unexplained origin. Am J Med 1976;60:429.
Giant-cell arteritis may present with prolonged fever but without headache, visual disturbance, or arthralgias.
Gleckman R, Crowley M, Esposito A. Fever of unknown origin: a view from the community hospital. Am J Med Sci 1977;274:21.
In a community hospital, pulmonary emboli and alcoholic hepatitis were common causes of prolonged fever.
Hurley DL. Fever in adults. What to do when the cause is not obvious. Postgrad Med 1983;74:232.
Provides a practical approach to evaluation of FUO and concludes that each patient’s illness will dictate a specific diagnostic approach.
Jacoby GA, Swartz MN. Fever of undetermined origin. N Engl J Med 1973;289:1407.
Good overview of etiologic classification of FUO.
Kauffman CA, Jones PG. Diagnosing fever of unknown origin in older patients. Geriatrics 1984;39:46.
Excellent review of FUO in elderly patients. Giant-cell arteritis, biliary tract infection, and drug fever are particularly important in this group.
Kazanjian PH. Fever of unknown origin: review of 86 patients treated in community hospitals. Clin Infect Dis 1992;15:968.
AIDS and Lyme disease caused FUO in a recent community hospital study.
Kerttula Y, Hirvonen P, Pettersson T. Fever of unknown origin: a follow-up investigation of 34 patients. Scand J Infect Dis 1983;15:185.
Five-year follow-up of 34 patients with FUO from a hospital in Finland.
Knockaert DC. Diagnostic strategy for fever of unknown origin in the ultrasonography and computed tomography era. Acta Clin Belg 1992;47:100.
Attempts to provide a standard approach to the FUO workup.
Knockaert DC, et al. Fever of unknown origin in the 1980s. An update of the diagnostic spectrum. Arch Intern Med 1992;152:51.
Large, prospective study of FUO from Belgium, 1980–1989. Malignancy occurred in only 7% of patients.
Knockaert DC, Vaneste LJ, Bobbaers HJ. Recurrent or episodic fever of unknown origin. Review of 45 cases and review of the literature. Medicine (Baltimore) 1993;72:184.
Long-standing recurrent fevers have a smaller differential diagnosis. Bacterial infections include chronic biliary tract or prostatic infection. Still’s disease, familial Mediterranean fever, Crohn’s disease, and lymphoma are examples of this type of FUO.
Larson EB, Featherstone HJ, Petersdorf RG. Fever of undetermined origin: diagnosis and follow-up of 105 cases, 1970–1980. Medicine (Baltimore) 1982;61:269.
Follow-up study by Petersdorf group (see also Petersdorf RG, Beeson PM, 1961). Neoplastic disease was the most common cause of FUO.
Larson TS, et al. Respiratory tract symptoms as a clue to giant-cell arteritis. Ann Intern Med 1984;101:594.
Sixteen patients with giant-cell arteritis had symptoms of cough, sore throat, or hoarseness.
Mackowiak PA. Southwestern Internal Medicine Conference. Drug fever: mechanisms, maxims and misconceptions. Am J Med Sci 1987;294:275.
Excellent detailed review of the clinical picture of drug fevers. Some findings go against standard teachings on drug fever.
McNeil BJ, et al. A prospective study of computed tomography, ultrasound, and gallium imaging in patients with fever. Radiology 1981;139:647.
Prospective study compared three imaging modalities in workup of unexplained fever and found all equally sensitive in diagnosis of focal disease.
Mellors JW, et al. A simple index to identify occult bacterial infection in adults with acute unexplained fever. Arch Intern Med 1987;147:666.
Several features in patients with unexplained fever predict bacterial infection: age above 50 years, diabetes, WBC above 15,000/mm3, and band count above 1,500/mm3.
Murray HW, et al. Urinary temperature: a clue to early diagnosis of factitious fever. N Engl J Med 1977;296:23.
Urinary temperature can be used to diagnose factitious fever.
Musher DM. Fever of unknown origin: diagnostic principles. Hosp Pract 1982;17:89.
Excellent description of the traditional case-by-case approach to the diagnosis of FUO.
Musher DM, et al. Fever patterns: their lack of clinical significance. Arch Intern Med 1979;139:1225.
Fever patterns were not helpful in assessing the etiology of fever on an infectious disease consultation service.
Petersdorf RG. Fever of unknown origin. Arch Intern Med 1992;152:21.
Petersdorf’s most recent reflections on FUO (see also Petersdorf RG, Beeson PM, 1961; Larson EB, Featherstone HJ, Petersdorf RG, 1982.) Patients with a diagnosis of FUO should rarely die.
Petersdorf RG, Beeson PM. Fever of unexplained origin: report of 100 cases. Medicine (Baltimore) 1961;40:1.
Initial, classic paper that defined FUO, its etiologies, and diagnostic approach.
Pizzo PA, Lovejoy FH, Smith DH. Prolonged fever in children: review of 100 cases. Pediatrics 1975;5:468.
Children are more likely to have viral and collagen inflammatory causes of FUO than are adults.
Quinn MJ, et al. Computed tomography of the abdomen in evaluation of patients with fever of unknown origin. Radiology 1980;136:407.
Results of 29% of 78 CT scans were positive in FUO patients.
Rowland MD, Del Bene VE. Use of body computed tomography to evaluate fever of unknown origin. J Infect Dis 1987;156:408.
Documents role of CT in the workup of FUO.
Schmidt KG, et al. Indium 111 granulocyte scintigraphy in the evaluation of patients with fever of undetermined origin. Scand J Infect Dis 1987;19:339.
Indium scintigraphy was very useful in 5 of 32 patients with unexplained fever.
Smith JW. Southwestern Internal Medicine Conference. Fever of undetermined origin: not what it used to be. Am J Med Sci 1986;292:56.
Includes 80 cases seen at Dallas Veterans Administration Medical Center between 1979 and 1985. Solid tumors, the most common cause of FUO, responded to nonsteroidal antiinflammatory agents.
Weinberger A, Kesler A, Pinkhas J. Fever in various rheumatic diseases. Clin Rheumatol 1985;4:258.
Describes the pathophysiology of fever in patients with rheumatologic diseases.
Weinstein L. Clinically benign fever of unknown origin: a personal retrospective. Rev Infect Dis 1985;7:692.
Description of both low-grade and high-grade fever with a benign course. Fever resulting from occupational exposure, such as polymer fume fever, can be high-grade.
Welsby PD. Pyrexia of unknown origin 60 years on. Postgrad Med 1985;61:887.
British perspective on FUO in the elderly, including role of foreign travel.
Wolff SM, Fauci AS, Dale DC. Unusual etiologies of fever and their evaluation. Ann Rev Med 1975;26:277.
Discussion includes brief overview of each major category of disease, including granulomatous hepatitis.
Young EJ, Fainstain V, Musher DM. Drug-induced fever: cases seen in the evaluation of unexplained fever in a general hospital population. Rev Infect Dis 1982;4:69.
Antimicrobial agents were responsible for most cases of drug fever.