CME: CLINICAL PRACTICE AND ITS BASIS uted to infection, neoplasm, collagen vascular disease and miscellaneous Infectious diseases causes. Arnow and Flaherty have sug- gested a minimum set of investigations to be performed prior to so defining a Edited by Andrew Freedman MD FRCP, patient (Table 1).5 Adoption of these Senior Lecturer in Infectious Diseases and Honorary Consultant Physician, criteria is recommended to attempt to Cardiff University School of Medicine establish greater comparability between future publications on FUO. It has been suggested that FUO can be subclassified into four different types: Fever of unknown origin • classic FUO • nosocomial FUO • immune-deficient FUO John Williams MRCP DTMH, Infectious ness of at least three weeks’ duration, with HIV-related FUO. Diseases Specialist Registrar repeated temperature measurements of • Richard Bellamy MRCP DPhil MSc MMedEd, 38.3°C or higher, that defies diagnosis Only the definition of classic FUO Infectious Diseases Consultant after one week of hospital inpatient evalu- specifies that the patient must have had 4 6 Department of Infection and Travel Medicine, ation. This has subsequently been modi- fevers for at least three weeks. This may James Cook University Hospital, fied to remove the requirement that potentially cause confusion; it is prob- Middlesbrough investigations are performed in the inpa- ably preferable to restrict the term FUO tient setting because most patients can be to those cases where FUO has been pre- Clin Med 2008;8:526–30 evaluated as outpatients. sent for at least three weeks and to One problem with the definition of exclude HIV and immune deficiency. FUO is that it does not specify which Unfortunately, many clinicians use the Definitions investigations should have been per- term loosely to include any fever for formed prior to stating that a patient ful- which a cause has not been found. This Fever fils the definition. This is probably should be strongly discouraged because responsible for many of the reported dif- The definition of fever is: the most likely causes of short-term ferences in the frequency of FUO attrib- fevers which have not been thoroughly a state of elevated core temperature often, but not necessarily, part of the defensive resistance of multicellular organisms Table 1. Minimum investigations which should have been performed (host) to the invasion of live (microorgan- without a diagnosis being established prior to qualifying as fever of isms) or inanimate matter recognised as unknown origin.5 Reproduced from the Lancet with permission from pathogenic or alien by the host.1 Elsevier. The term ‘pyrexia’ is often used synony- • Comprehensive history mously with fever in the UK. Fever should • Repeated physical examination be distinguished from hyperthermia, • Full blood count which is an unregulated rise in body tem- • Routine biochemical profile perature due to a failure of temperature • Urinalysis and microscopic examination homeostasis.2 Over 100 years ago • Chest X-ray Wunderlich recorded multiple tempera- • Erythrocyte sedimentation rate (and/or C-reactive protein) ture measurements on around 25,000 • Antinuclear antibodies people. He defined normal body temper- • Rheumatoid factor ature as 37°C and suggested that the • Angiotensin-converting enzyme upper limit of the normal range should be • Routine blood cultures (three sets) while not receiving antibiotics regarded as 38°C.3 This has formed the • Cytomegalovirus IgM antibodies or virus detection in blood basis for subsequent definitions of fever. • Heterophile antibody test in children and young adults • Mantoux skin test • CT scan of abdomen (or radionuclide scan) Fever of unknown origin • HIV antibody test • Further evaluation of any abnormalities detected by the above tests Petersdorf and Beeson provided the classic definition of ‘fever of unknown CT = computed tomography; Ig = immunoglobulin. origin’ (FUO) as a prolonged febrile ill- 526 Clinical Medicine Vol 8 No 5 October 2008 © Royal College of Physicians, 2008. All rights reserved. CME Infectious diseases investigated are very different from those the thoroughness of the investigations under 18 years with FUO compared with of FUO. performed prior to classifying a patient less than 5% for adults over 65 years.9 as having FUO and the diagnostic tests However, joint disease in children with Causes of fever of unknown subsequently available to investigate it. FUO suggests a serious underlying dis- origin For example, the availability of highly order such as connective tissue disease sensitive blood culture techniques and (CTD), endocarditis or leukaemia.10 The causes of FUO in Europe and the high quality echocardiography means Still’s disease, an important cause of USA, reviewed by Armstrong and that bacterial endocarditis is now a less FUO in children, can also affect young Kazanjian7 in 14 papers published since common cause of FUO because the con- adults and is a condition often neglected 1930 (Table 2), were found to be: dition can be diagnosed relatively easily during the search for a cause. Malig- • infection: 30% of cases and is therefore unlikely to meet the nancy is a relatively uncommon cause in • neoplasms: 18% FUO criteria. children8 and young adults,11 but lym- collagen vascular diseases: 12% phoma is a potential diagnosis which is • important to exclude because delay in miscellaneous causes: 14% Children and young adults • diagnosis may adversely affect prognosis. • undiagnosed: 26%. The relative likelihood of FUO being due There is substantial variation between to specific causes is related to the age of The elderly the different series. In particular, col- the patient. In children, infections are the lagen vascular diseases were less frequent most common cause. Jacobs and Schutze Haematologic malignancies and solid causes of FUO in earlier case series than found that 64 of 164 children with FUO tumours are more common causes of presently. In resource poor countries had infections including Epstein-Barr FUO in elderly patients than in younger FUO is more frequently due to infections virus infection, osteomyelitis, bartonel- adults.11,12 Infections and CTDs are also than in established market economies.6 losis and urinary tract infections.8 Many frequent causes in the elderly. Temporal This may partly represent differences in children with FUO recover without a arteritis and polymyalgia rheumatica are the geographic and temporal distribu- diagnosis ever being established. A cause particularly common in this age group. tion of diseases, but is also explained by is not found in over 40% of children Symptoms of temporal arteritis may be Key Points Table 2. Causes of fever of unknown origin in Europe and the USA and their relative The term ‘fever of unknown origin’ frequency reported in the literature (1930–97). Adapted from Reference 7. (FUO) should be used only if the patient has had a febrile illness for Patients with attributed diagnosis (%) more than three weeks and a thorough history, examination and Collagen appropriate investigations have Country and year vascular failed to elucidate the cause of publication Infection Neoplasm disease Miscellaneous Undiagnosed Careful history taking and repeated, USA 1930 24 11 0 2 63 thorough physical examinations are more likely to establish the USA 1936 59 22 0 0 19 diagnosis than a ‘scatter-gun USA 1939 64 24 0 12 0 approach’ to performing complex Sweden 1953 24 16 6 6 49 investigations USA 1959 21 30 0 29 20 FUO can be caused by a wide range of USA 1961* 36 19 15 23 7 conditions including infections, Finland 1962 19 6 6 0 69 connective tissue diseases and USA 1963 20 18 13 10 38 malignancies Sweden 1966 13 32 3 7 45 The older the patient the more likely USA 1973 40 20 15 17 8 malignancy is to be the cause of USA 1982 30 31 9 17 12 FUO Belgium 1992 23 7 21 24 26 A therapeutic trial of empirical USA 1992 33 24 21 13 9 antibiotics is usually inappropriate Netherlands 1997 26 13 20 12 30 Total 30 18 12 14 26 KEY WORDS: antipyretic drugs, body temperature, fever, fever of unknown * The classic paper of Petersdorf and Beeson.4 origin, pyrexia Clinical Medicine Vol 8 No 5 October 2008 527 © Royal College of Physicians, 2008. All rights reserved. CME Infectious diseases Table 3. Clues to the cause of a fever of unknown origin suggested by the history. Adapted from Reference 13. Medication or exposure Drug fever Recent history of stroke Culture-negative endocarditis to toxic substances Fume fever Takayasu arteritis Polyarteritis nodosa History of tick bite Relapsing fever Cough with sputum Tuberculosis Lyme disease production Coxiella infection Psittacosis Typhoid Lung cancer Acute rheumatic fever Contact with animals Psittacosis Visual disturbance or Temporal arteritis or birds Leptospirosis eye pain Culture-negative endocarditis Brucellosis Brain abscess Toxoplasmosis Takayasu arteritis Bartonellosis Coxiella infection Rat-bite fever Complaint of myalgia Trichinosis Complaint of fatigue Carcinoma Culture-negative endocarditis Lymphoma Polyarteritis nodosa Cytomegalovirus infection Rheumatoid arthritis Typhoid Familial Mediterranean fever Systemic lupus erythematosus Polymyositis Rheumatoid arthritis Toxoplasmosis Complaint of headache Relapsing fever Abdominal pain Polyarteritis nodosa Rat-bite fever Familial Mediterranean fever Chronic meningoencephalitis Relapsing fever Malaria Brucellosis CNS malignancy Cognitive function Sarcoid meningitis Complaint of back pain Brucellosis slowing/confusion Tuberculous meningitis Culture-negative endocarditis Vertebral