Fever of Unknown Origin

Last week Friday at Rheumatology Morning Report we discussed a patient with a fever of unknown origin (FUO).

This is a challenging Internal Medicine presentation with limited high quality evidence and a wide differential diagnosis.

A number of reviews have been published in recent years offering varied approaches to making a diagnosis. Take a look at our own Toronto approach published by Drs. Mourad, Detsky and Palda in the Archives of Internal Medicine. Two other reviews include one from AFP and another from Postgraduate Medicine.

A few points about FUO:

1) The classical definition of FUO is a temperature greater than 38.3C on multiple occasions over a period of 3 weeks with no diagnosis found after 1 week of inpatient investigations.

While this definition is likely outdated in the era of expanded outpatient management, the bottom line is that there must be a documented fever for a prolonged period of time. This is done in an attempt to exclude any transient viral infections. One modern definition suggests that the fever should persist for more than 3 weeks despite 3 outpatient visits and/or 3 days of in-hospital investigations to be considered an FUO.

The classic definition and approach to FUO typically excludes patients with HIV, immunosuppression or neutropenia.

2) The incidence of FUO has been estimated in one series to be less than 3% of admitted medical patients. The frequency of FUO continues to decline with the advent of increasingly sophisticated investigative modalities (cultures, serology, radiology).

3) The etiologies of FUO can be grouped into 5 categories: infection (16%), inflammation (22%), malignancy (7%), miscellaneous (4%) and undetermined (51%).

4) As always, the investigation should start with a focused history and physical examination. Remember to pay special attention to recent travel, medication use (including anti-pyretics), possible immunosuppression, the pattern of fever and any localizing symptoms.

5) A number of initial investigations have been suggested. One approach is to ensure the patient has a minimum set of investigations before proceeding down a FUO algorithm.

These might include:
i) CBC and differential
ii) Blood film
iii) 3 set of blood cultures (separated in space and time)
iv) Lytes, Cr, HCO3, LDH, liver enzymes and function tests
v) Urinalysis and culture
vi) CXR

From here if a source of fever has not be elucidated, further minimum investigations might include:

i) ANA and RF
ii) HIV testing
iii) CMV IgM or heterophile testing (if suggested clinically)
iv) Hepatitis serologies (if liver enzymes elevated)
v) ESR or CRP
vi) One article suggests Q-fever serology (if suspected clinically)
vii) One article suggests an SPEP (if MM suspected clinically)
viii) CK
ix) A TB skin test has been suggested in some articles

If this second set of investigations fails to yield of a diagnosis, all non-essential medications should be discontinued had imaging should be pursued:

i) CT chest/abdomen - assess for lymphadenopathy or collections
ii) One article suggests a technetium scan in addition to the above CTs.

Following these investigations, if no etiology is apparent, then further testing should increasingly be focused towards the most likely diagnosis. This may include:

i) Echocardiography - assess for valvular vegetations
ii) Leg dopplers - assess for a silent DVT
iii) Temporal artery biopsy - in patients greater than 50 years old to assess for TA
iv) Liver biopsy - if clinically deteriorating
v) Bone marrow biopsy - if abnormalities are seen in the CBC and blood film

6) Therapeutic Trials
In general, therapeutic trials of antibiotics or corticosteroids are discouraged unless a thorough hunt for an etiology has been completed. This is out of concern of exacerbating an undiagnosed infection (steroids), sterilizing cultures (antibiotics) or suppressing a fever while the etiology remains.

7) Prognosis
If no diagnosis is found after an extensive evaluation, the prognosis of these patients is generally good.